volume 4: issue 2 april 2011 - the michigan journal …michiganjb.org/issues/42/text42.pdfkush patel...

The Michigan Journal of Business (ISSN# 1941-5745) is published semian-nually by undergraduate business students on behalf of the Stephen M. Ross School of Business at the University of Michigan. Communication should be addressed to:

Michigan Journal of Business701 Tappan StreetAnn Arbor, MI 48109E-mail: [email protected]: www.michiganjb.org

EDITORIAL OBJECTIVEThe Michigan Journal of Business intends to provide undergraduate students worldwide with a platform for exceptional work in the field of business. The Journal seeks to publish distinguished theses, empirical research, case studies, and theories in issues relating to areas of Accounting, Economics, Finance, Marketing, Management, Operations Management, Information Systems, Business Law, Corporate Ethics, and Public Policy. The Journal is distributed and cataloged in prestigious university libraries around the world, and is en-listed in the Directories of Open Access Journals (DOAJ), a scholarly journal database that enlists more than 3000 of the world’s leading publications. The contemporary business environment is exceedingly complex. Analyzing this real world phenomenon through traditional applications of theories often yield a suboptimal understanding of the world. The Journal, accordingly, encour-ages work that takes an interdisciplinary approach to understanding a topic and emphasizes the importance of incorporating the knowledge of liberal arts into an area of interest. By providing a venue to recognize high quality work, the Journal gives an incentive for students to explore their area of interest, reward-ing them with the experience to share the power of knowledge with others. The Journal’s mission and philosophy parallel the mission of the University of Michigan, the premier research university in the United States.

VOLUME 4: ISSUE 2 APRIL 2011

THE MICHIGAN JOURNAL OF BUSINESS2

CONTRIBUTOR INFORMATIONThe Journal only accepts works from undergraduate students or works com-pleted during undergraduate study. Each manuscript submitted should include a short abstract, author information, and any acknowledgements. Papers will be evaluated based upon sound analysis, originality of argument, and novelty of research. For more information on submitting article for publication, please visit www.michiganjb.org.

REVIEW PROCESSThe organization is entirely student-run, with an editorial staff of about 20 of the top students at the Department of Economics and the Stephen M. Ross School of Business at the University of Michigan. Each semester, the Michi-gan Journal of Business calls for papers from undergraduate students around the world. Throughout the semester, the editorial board carefully reviews, se-lects, and edits exceptional work for publication. Faculty willing to advise the Journal is formed from each department to give minor oversight for the proj-ect. Throughout the process, a blind review process is implemented to ensure an impartial review of all submissions.

EDITORIAL BOARDThe editorial board is composed of the top undergraduate students at the Ste-phen M. Ross School of Business and the Department of Economics at the University of Michigan. They are selected through a competitive applica-tion process based on academic merit and aptitude in writing. Past and pres-ent members of the MJB editorial board include a former Ross valedictorian who now attends Yale Law School, an editor who now attends Michigan Law School, and editors with professional experience at firms including Goldman Sachs, McKinsey & Co., Morgan Stanley, and JPMorgan.

3

WINTER 2011 EDITORIAL BOARD

EDITOR-IN-CHIEFDominic Spadacene

SENIOR EDITORSStephanie BirndorfJonathan BrowalskiVictoria GreensteinBrandon Lebowitz

Ifat RibonShuhan Wang

MANAGING EDITORSKaty Beth Deschenes

Scott Suh

DIRECTOR OF FINANCEShaun Yu

DIRECTOR OF PRODUCTION AND MEDIAHabib Khan

ASSOCIATE EDITORSBen CharoenwongStephanie Chueh

Kurtis DrogeDaniel Grossman

Michael HuDino Kaknjo

Gregory KohlerAJ MalhotraAnkur PatelKush Patel

Ritika SinghKabir Sodhi

Adam SwiecickiNathan Torreano

Shuhan WangRitika SinghYahya SyedDadong YanRoger Zhong


FACULTY SUPERVISIONThe Journal is supervised by Professor Tammy Feldman, Lecturer of Business Economics and Public Policy, and Professor Scott Moore, Arthur F. Thurnau Professor of Business Information Technology and BBA Program Director at the Ross School of Business. The following members of the Ross School of Business faculty also provide minor oversight to the editorial board during the review process:

Anocha Aribarg Ph.D., University of Wisconsin Assistant Professor of Marketing

George Cameron III J.D., Ph.D., University of Michigan Professor Emeritus of Business Law

Tammy FeldmanPh.D, Harvard UniversityLecturer of Business Economicsand Public Policy

David Hess J.D., Iowa, Ph.D., University of Pennsylvania Assistant Professor of Business Lawand Business Ethics

Aneel G. Karnani Ph.D., Harvard Associate Professor of Strategy;Chair of Strategy

Scott A. Moore Ph.D., University of Pennsylvania Arthur F. Thurnau Professor;BBA Program Director; Associate Professorof Business Information Technology

Andrea Morrow M.A., University of Michigan Director of Writing Programs; Co-Coordinator BBA Communication Program

Dana M. Muir J.D., University of Michigan Arthur F. Thurnau Professor of Business Law

Anu Nagarajan Ph.D., University of Michigan Lecturer of Strategy

Lisa A. Pawlik M.B.A., University of Michigan

Adjunct Lecturer of Business Communication; Co-Coordinator of

BBA Communication Program

Josh Pierce Ph.D., Michigan State University

Assistant Professor of Financeat University of South Carolina

Lloyd Sandelands Ph.D., Northwestern

Professor of Management and Organizations& Professor of Psychology

Dennis G. Severance Ph.D., University of Michigan

Accenture Professor of Business Info. Tech.

Amitabh Sinha Ph.D., Carnegie Mellon University

Assistant Professor of Operationsand Management Science

Valerie Y. Suslow Ph.D., Stanford

Associate Dean for Degree Programs;Louis and Myrtle Moskowitz Research

Professor of Business and Law;Associate Professor of

Business Economics and Public Policy

David B. Wooten Ph.D., University of Michigan

Associate Professor of Marketing

Lynn P. Wooten Ph.D., University of Michigan

Clinical Assistant Professor of Strategyand Management & Organizations

5

CONTRIBUTING INSTITUTIONSManuscripts for this edition were submitted from the following institutions:

Arizona State UniversityBabson CollegeBrandeis UniversityCollege of William and MaryColumbia UniversityDuke UniversityGeorgia Institute of TechnologyHartwick CollegeMohammad Ali Jinnah UniversitySkyline Institute of Engineering and TechnologyTexas A&M UniversityUniversity of California, Los AngelesUniversity of PennsylvaniaUniversity of the Punjab


(ISSN# 1941-5745)

Copyright © 2010 Stephen M. Ross School of Business at the University of Michigan. This work may be reproduced and redistributed, in whole or in part, without alteration

and without prior written permission, solely by educational institutions for nonprofit ad-ministrative or educational purposes provided all copies contain the following statement: “Copyright © 2010 Stephen M. Ross School of Business at the University of Michigan.

This work is reproduced and distributed with the permission of the Stephen M. Ross School of Business at the University of Michigan. No other use is permitted without

the express prior written permission of the Stephen M. Ross School of Business at the University of Michigan. For permission, contact: The Michigan Journal of Business, 701

Tappan Street, Ann Arbor, MI 48109, e-mail: [email protected].”

7

CONTENTS

Editor’s Note 8

The Feasibility of a Monetary Union in MERCOSURMOISES NUMABabson College

11

Law of Yuan Price: Estimating Equilibrium of the RenminbiJOSHUA KLEIN LIPMANUniversity of Pennsylvania

61

A Multi-Criterion Model for Evaluating the Efficiency of Solar Energy IncentivesCHRISTOS MAKRIDISArizona State University

91

Real Option Analysis of a Large-scale Space Solar Power VentureYAEL GILBOA and XIAOYU GUOCollege of William and Mary

121

VOLUME 4: ISSUE 2 APRIL 2011


EDITOR’S NOTEAt the conclusion of the 2011 academic school year, the Michigan Journal

of Business is pleased to offer its 8th issue of undergraduate theses. This issue brings the Journal’s fourth volume to a close, and in doing so, continues the realization of the objective that the founding editors had in mind: to provide a common platform for undergrads to share their research with a broader aca-demic community. As noted in previous issues, the Journal is still accessible from over 200 libraries including the United States Library of Congress, Har-vard Business School’s Baker Library and Princeton’s Main Library as well as the Directory of Open Access Journals.

As usual, the editorial board had the honor to consider a wide variety of topics ranging from monetary unions to sports marketing. In spite of a relatively more modest batch of nineteen submissions, the quality of work has not diminished, making our decisions no easier. After careful refereeing, four articles were selected to be presented in this issue, which vary both in method as well as content. In light of this variance, the articles can be considered with respect to two distinct topics: currency and solar energy.

This issue begins with a discussion of the possibility of implementing a monetary system similar to that of the EU within South America. In “The Feasibility of a Monetary Union in MERCOSUR” Moises Numa of Babson College systematically profiles each potential country and tests for current lev-els of integration with various econometric models. Due to the current state of economic performance and integration, the author largely finds that such a system would be imprudent today. We follow this broad consideration of cur-rency unionization with an analysis of one particular currency, China’s Ren-minbi. In the second article, “Law of The Yuan Price: Estimating Equilibrium of the Renminbi”, Joshua Lipman of the Wharton School of the University of Pennsylvania contributes to the ongoing discussion regarding the equilibrium value of the Renminbi. While his estimate is primarily derived from several Purchasing Power Parity studies, he further substantiates his claims by consid-ering those found in previous literature by way of more complex macroeco-nomic approaches.

The remaining two articles are primarily concerned with renewable en-ergy dilemmas facing the world today. In “A Multi-Criterion Model for Evalu-ating the Efficiency of Solar Energy Incentives” Christos Makridis of Arizona State University looks to help facilitate the adoption of renewable resources by developing both a theoretical model and a list of qualitative criteria for the purposes of assessing current public policy incentives. The author also offers a discussion of the effectiveness of one particular incentive: feed-in-tariffs. Finally, our fourth article, “Real Option Analysis of a Large-scale Space Solar

9

Power Venture”, concludes this theme by exploring the financial possibility of investing in a space solar power program. In light of the many static financial projections of this kind of venture, authors Yael Gilboa and Xiaoyu Guo of the College of William and Mary augment these analyses by accounting for the series of decisions made in response to project’s updated state and current en-ergy prices. In observing this overlooked flexibility, the authors find that such a venture could indeed be feasible to undertake.

The Journal would like to extend its gratitude to the many contributors and faculty supervisors that make this publication possible. In particular, we would like to thank Professor Tammy Feldman for her advice in the methods and direction of the Journal along with Professor Scott Moore for his contin-ued support in the Journal’s activities. We would also like to thank all of our other advisors at the Stephen. M. Ross School of Business for their help reach-ing out to other schools for submissions. Ms. Erika Busch was, as always, instrumental in our logistical operations as well. Mr. Thomas C. Jones and the Alumni Association at Ross have consistently shown their support through their financial contributions. Thank you to the editorial board for another se-mester’s worth of diligent work. And finally, thank you to the authors who not only provide the content through their insightful work but also endure several rounds of revisions.

Dominic SpadaceneEditor-In-Chief

11

The Feasibility of a Monetary Union in MERCOSUR

Moises Numa1

Babson College

Abstract

Based on the progress of the European Union (EU) and the development of the European Monetary Union (EMU), many economists have studied the pos-sibility of implementing a single monetary unit system in other regions of the world. When considering the countries in the region of Latin America, experts suggest they are currently not ready for a monetary union. This paper analyzes the feasibility of a currency union within MERCOSUR – a trade agreement between Argentina, Brazil, Paraguay and Uruguay – with a long-term per-spective. In order to form a conclusion, the concept of a currency union in MERCOSUR was analyzed qualitatively and quantitatively. The existing Euro model was studied and frequently used as an example. In addition, the country profiles of all four members of MERCOSUR were conceptually compared and also analyzed with econometric models to test the current level of integration. The results suggest that these four countries are not ready for a currency union because their level of integration is not strong enough and there is too much volatility in their economies. To develop into an optimum currency area, sev-eral actions must first occur: Argentina, Brazil, Paraguay and Uruguay must increase their level of integration amongst each other, improve their economic performance, and lastly, all four countries must believe and act in such a way that shows they are unified under a single goal. Lastly, the Eurozone has prov-en that to have a sustainable currency union, there must be fiscal integration to absorb and smooth economic shocks. If the ultimate goal is a monetary union, then an economic union should also be taken into consideration.

1 Moises Numa graduated from Babson College in December 2010 with a B.S. in Business Management. This paper was originally written as a senior honors thesis under the guidance of his thesis adviser Kent Jones.


I. IntroductionFundamentally every nation has the freedom to decide how they will

manage their currency. Among the array of available options is the election of a fixed exchange rate as a means to prevent value fluctuation, and the election of a flexible exchange rate, under which the value of the currency is essentially determined by the markets. These alternatives forego the potential for inte-gration and collaboration with other countries to create a competitive advan-tage. Unlike the aforementioned options, monetary unions allow neighboring countries to share the same currency. However, governments participating in a monetary union have a lack of autonomy regarding their monetary policies. Despite the disadvantages, monetary unions are appealing to countries who seek to increase trade, eliminate transaction costs, and increase foreign direct investments, among others.

The purpose of this paper is to assess the feasibility of a currency union in Latin America. It is divided into seven sections to critically analyze the po-tential implementation and success of the respective currency union. The first section serves as an introduction to monetary unions and outlines the rest of the paper. Section two provides an overview of a literature review written on this particular topic. Section three compares the advantages and disadvantages of monetary unions. Section four is a review of the monetary union in Europe from which general recommendations are drawn for policy makers who hope to pursue a monetary union. Section five presents a study of the MERCO-SUR countries in order to conceptualize the structures of these four countries and the political and economic driving forces of their economy. Section six contains quantitative analysis which evaluates the degree to which Argentina, Brazil, Paraguay and Uruguay are integrated using four statistical models. The seventh and final section is the conclusion and offers recommendations that MERCOSUR could follow.

By analyzing, both quantitatively and qualitatively, the feasibility of a monetary union in Latin America, this paper concludes that MERCOSUR is not ready for a currency union. It also suggests steps that will increase the prospect of a single currency for the MERCOSUR countries which are drawn from all seven sections of the paper.

II. Literature ReviewIn the article, “A Theory of Optimum Currency Areas” Robert Mundell

introduced the theory of a single currency. He also explained situations where groups of nations would benefit from such a currency union; hence, what he called an “Optimum Currency Area”. Mundell suggested that if labor and capi-tal “are mobile across national boundaries then a flexible exchange rate system

13The Feasibility of a Monetary Union in MERCOSUR

becomes unnecessary”2. He implies that only under these conditions would a monetary union make economic sense. Mundell’s article and suggestions are important for this paper as they correspond to fiscal integration, which seems to be a necessity for a monetary union.

Also, due to the success of the EMU, many debates have come about of the potential benefits of a similar union in other “Optimal Currency Areas”. This search has been a very popular one, especially trying to identify whether or not a monetary union would make sense in Latin America.

Similarly, Sebastian Edward analyzed the economic performance and shocks of Latin American countries in his article, “Monetary Unions, External Shocks and Economic Performance: A Latin American Perspective”. The first part of Edward’s results suggests that “belonging to a currency union has not lowered the probability of experiencing either a sudden stop or a current ac-count reversal.3” The second half of his study emphasizes on some of the costs and potential risks of a monetary union by explaining “that the effects of ex-ternal crises on GDP growth have tended to be more severe in currency union countries than in countries with a currency of their own.”

Alternatively, Fernando Ferrari-Filho focuses on MERCOSUR in the ar-ticle “Why does it not make Sense to Create a Monetary Union in Mercosur? A Keynesian Alternative Proposal”. Ferrari-Filho analyzes the existing data to determine whether a currency union would be feasible among the countries of MERCOSUR. Ferrari-Filho concludes that a monetary union is still premature because “(i) the factor markets are not sufficiently unified, (ii) the volume of intra-regional trade among the MERCOSUR countries is still low and (iii) the macroeconomic variables of the MERCOSUR failed in addressing the conver-gence criteria.”4

The purpose of this paper is to study the feasibility of a potential mone-tary union in Latin America – specifically in MERCOSUR. This paper contrib-utes to the existing literature because it assesses the feasibility of a currency union in MERCOSUR quantitatively and qualitatively. By applying several econometric and statistic methodologies, examining the country profiles of the MERCOSUR members and studying the Euro model, this paper analyzes the proposed question in the long run.

III. Advantages and Disadvantages of a Currency UnionA currency union has both advantages and disadvantages. The advantages

2 Robert Mundell A Theory of Optimum Currency Areas, 1961.3 Sebastian Edwards, Monetary Unions, External Shocks and Economic Performance: A Latin American

Perspective,20064 Fernando Ferrari-Filho, Why does it not make sense to create a monetary union in MERCOSUR? A

Keynesian Alternative Proposal, 2001


are numerous. One of the biggest benefits of a monetary union is the increase in trade. Reuven Glick and Andrew K. Rose estimate that “bilateral trade ap-proximately doubles/halves as a pair of countries forms/dissolves a currency union” in their paper titled: “Does a Currency Union affect Trade? The Time Series Evidence”5. A currency union causes trade to increase for several rea-sons: transaction costs and tariffs are eliminated, price transparency increases, and free labor mobility encourages specialization and competition. Also, a suc-cessful currency union will lead to a more stabilized currency which increases the flow of investments into the bloc. Lastly, since the country risk is reduced, the value of the securities in that area would increase because of the reduction in investment risk. This increase in the markets would have a positive effect on the overall wealth of the nations.

The disadvantages of a currency union are not as plentiful, but can be very risky. The first cost of a monetary union is the introduction of a new currency into the markets. “The European Commission (1990) estimated that the elimination of currency conversion costs would amount to one-quarter to one-half of 1 percent of the Community GDP”6. The main risk to an individual country in a monetary union is the ceding of their monetary policy powers. By foregoing this mechanism, governments depend solely on the new established central bank. This bank pursues the best interest of the bloc as a whole, so policies that benefit the group can be detrimental to one of the countries. Also, another risk of a monetary union is the effect a crisis in one country can have on the entire union. For example, as the markets feared a possible default from Greece, contagion spread to other countries. As a matter of fact, the acronym PIIGS was created to reference Portugal, Ireland, Italy, Greece and Spain.

For these reasons, it is imperative that a currency union should be strictly composed of countries with very similar economies. By doing so, the members of the union will share business cycles, thus reducing the risks of asymmetric shocks. If the central bank policies help all member countries, then the risk of foregoing monetary policy is significantly reduced.

IV. European Monetary UnionIn 1999 the European Monetary Union and the Euro currency were offi-

cially introduced in eleven countries. In 1951, the first step towards integration was taken with the commencement of the European Coal and Steel Commu-nity (ECSC)7. A key moment that led to the creation of the Euro was the Maas-tricht Treaty, which was officially signed on February 7, 1992. This treaty set 5 Reuven Glick and Andrew K. Rose, Does a Currency Union Affect Trade? The Time Series Evidence,

20016 George Tavlas, Benefits and Costs of Entering the Eurozone. Cato Journal, 89-106, 2004.7 http://www.german-way.com/euro-timeline.html


the timetable for the implementation of the Euro and clarified the criteria to evaluate the admission of countries into the EMU as follows:

1. Inflation cannot be 1.5% higher than the inflation of the top three EU countries within the last twelve months

2. Countries cannot have government budget deficit higher than 3% of GDP

3. Countries must not have a government debt ratio of more than 60% of GDP

4. Average nominal long-term interest rates that do not exceed 2% points of the top three member states

5. Exchange rate has to be normal compared to that of the Exchange Rate Mechanism (ERM) and without severe tensions for two years8

The criterion set forth in the Maastricht Treaty was supposed to create a very exclusive group of nations based on what appeared to be very selective admission criterion. However, Mundell’s suggestions of what constitutes an optimum currency area are circumstances that advance fiscal integration, yet none of the Maastricht Treaty criteria does the same. While the criterion set in 1999 is a guideline that encourages member countries to have a certain level of discipline, the suggestions offered by Mundell would help smooth out shocks caused by differences in business cycles among the member countries.

Although the Maastricht Treaty omitted Mundell’s suggestions, the Euro seemed to be holding up very strongly. Until the recent global financial crisis, the Euro was perceived by many to be a great success. The currency was very stable and consistently traded at a higher value than the American dollar. Also, the paper “Has the Euro Increased Trade” by Maurice Bun and Franc Klaassen shows the significant impact the Euro has had on trade by estimating “a total cumulative increase in intra-EMU exports of 3.9% in 1999, 6.9% in 2000, 9.6% in 2001 and 37.8% in the long run”.9

However, as a result of the economic crisis, many of the issues of the currency union started to show. Greece, for example, was unable to contain their debts to the levels specified by the Maastricht Treaty and unemployed Spaniards found it difficult to migrate around the Eurozone to balance their staggering high unemployment rates. 8 Eduard Hochreiter and Helmut Wagner, The Road to the Euro: Exchange Rate Arrangements in

European Transition Economies, Sage Publications.9 Maurice Bun and Franc Klaassen, Has the Euro Increased Trade, 2002. (Bun and Klaassen 2002)


In the article “Currency Substitution and European Monetary Union” Ni-kiforos Laopodis, analyzed whether the net benefits for Greece, Spain and Portugal of substituting their currency for the Euro made sense. Laopodis used an “error-correction model through the application of the two-step estimator of co-integrated systems”. He concluded that due to the high inflation rates and the low degree of financial market liberalization in these countries, the net benefit of joining the EMU did not outweigh the costs of substituting their currency. The results question the impulse of Greece, Spain and Portugal for joining the Euro and of the motivation for stronger countries to accept weaker members.10

In the essay “Greek Monetary Economics in Retrospect: The Adventures of the Drachma”, Sophia Lazaretou describes, chronologically, important mo-ments of the Greek economy and the steps Greece needed to take to join the Euro. According to Lazaretou, in 1990, the Greek economy was suffering from instability. However, during the 1990’s the Greek government took actions to improve its economy and their image to meet the Maastricht criteria and gain admittance into the EMU. For example, inflation was more than halved level-ing at around 3.5%. Greece also joined the Exchange Rate Mechanism (ERM) of the European Monetary System to demonstrate the country’s stability.

Lazaretou shows Greece was able to implement strict strategies to earn a spot in the EMU. However, she does not prove that their acceptance was the best route for either party. Before admitting a country into an almost irrevers-ible union, a certain time frame should be in place to gauge their long-term stability. This might have given the EMU a clearer picture of how Greece’s policies would shape out in later years, and potentially prevent the contagion created from the Greek debt crisis. This goes without mentioning the rumor that the Greek government manipulated their books in order to present them-selves as worthy of joining the EMU.

Given the fragility of some of the European countries, the EMU faced a difficult decision: whether to bail out these struggling countries or allow them to ask for external lending. Greece needed a package of around €40-€50 billion11. However, as German Finance Minister, Wolfgang Schaeuble said, “Monetary Union is unprepared for extremely severe situations of the type we are now seeing”12.

The contagion effects caused by the Greek crisis have had negative con-sequences on some of the other EMU countries. Spain and Portugal were the first victims after their ratings were downgraded as Standard & Poor’s cut 10 Nikiforos Laopodis, Currency Substitution and European Monetary Union, Ekonomia, 2001.11 Elisa Parisi-Capone and Arun Motianey, Is a European Monetary Fund a Useful Complement to the

IMF?, 2010.12 http://www.roubini.com/critical-issues/48728.php


Greek debt to junk status13. The downgrade of Spain’s rating triggered a sell-off of Italian government bonds. In addition, Spain’s high unemployment rate rose above 20%. Santiago Carbo, an economics professor at the University of Granada stated, “Most of the 20 percent is structural, it’s here to stay”.14

The consequences of reversing the Euro suggest it is best for the union to stay together. Reversing the Euro to the Drachma, for example, would prob-ably cause Greece to continue suffering. A devaluation of their hypotheti-cal independent currency would increase their debt and the costs of exports. Simultaneously, Greece would lose the credibility the EMU once provided. Concurrently, the markets would expect currency depreciations, which would increase currency-risk and reduce FDI. In addition, the cost of currency trans-actions with the EMU would be a factor again, thus further reducing trade and investments15. Lastly, reintroducing the Drachma would be expensive and local banks would suffer, as most people would withdraw their money further impacting the already weak banking system16.

In conclusion, given the recent volatility of the EMU, the structure of any future currency union, should be diligently analyzed. The European model can provide insight into finding ways to select the appropriate countries, set the best criteria, and create a back-up plan.

Country selection should account that admittance into a currency union is almost irreversible. Initially, the EMU seemed to be very selective; however, this was not the case. During the decade of the 1990’s many countries did not meet the requirements set by the Maastricht Treaty. For example, the debt conditions in 1997 in “Belgium, Greece, and Italy were well above the 60% limit… Denmark, Austria, Germany, Ireland, Portugal, Spain, Sweden and the Netherlands were closer to meeting the goals but still above the limit”. As of 1996, only Luxembourg qualified to be admitted into the EMU. Also, between 1997 and 1998, Greece was well over the inflation rate, deficit as a percentage of GDP, debt as percentage of GDP and long-term interest rate criteria.17

Hindsight pinpoints the European crisis to some struggling countries who were, in all probability, not the best fit for the union. To learn from the EMU, policy makers should carefully consider whether a smaller group of countries, with very similar traits, would be the most beneficial. In addition, the criteria set to determine eligibility of a nation should be tested for a longer period of time and perhaps be audited. This would force countries who want to join to

13 Wall Street Journal: Spain Downgrade Spurs Contagion Fear14 The New York Times: Europe Now Moving Swiftly on Greek Rescue Plan15 Financial Times, Greece will fix itself from inside the eurozone, 2010.16 The Economist, The Future of the Euro: Don’t do it, 2010.17 John Edmunds and John Marthinsen, Wealth by Association: Global Prosperity Through Market

Unification, 2003.


find sustainable ways to integrate for the long-run rather than focusing on the short-term fix.

In addition, the admissions criteria should not be based on one coun-try’s economy (i.e. low inflation in Germany and the EMU), but rather on a mixture of similarities amongst the member countries. The optimum currency area criteria, outlined by Mundell, should also be a factor to determine which countries should join. This is important because since the adjustment mecha-nism of monetary policy will be ceded, there should be other alternatives for countries to smooth the shocks of the different economies. George Tavlas, the Director General of the Bank of Greece, suggests that a good method to do this is to have fiscal integration, such as labor mobility. This would “provide fis-cal transfers from low-unemployment regions to high unemployment regions” which would help mitigate some of the shocks various countries may face.18

Lastly, the EMU was in a sense of urgency as Greece suffered from the high possibility of default. This had very negative consequences on the Euro currency and the Eurozone. This problem could have been resolved before it escalated to this degree by an emergency fund. To solve this problem, a “safety net based on centralized fiscal transfers” should be in place19. This concept is controversial in the EU as some countries do not want the burden of support-ing the weaker countries. However, only after a group of countries are willing to set aside their short term interests to pursue the long term goals of the group will a currency union succeed.

The Federal Minister of Finance in Germany, Wolfgang Schaeuble, pro-posed a way to fund the potential European Monetary Fund. He suggests, “Monetary penalties imposed on Stability Pact breaches.” He explained that by “imposing a 1% fine on excessive debt and excessive deficit to all mem-bers since the start of the EMU, the [European Monetary Fund] would have accumulated 120 billion Euros [by] 2009”20. Having a similar fund for future monetary unions will provide member countries easier access to capital in case of an emergency. This fund would threaten members from diverging outside the criteria while also discouraging unfitted countries to attempt to join.

V. MERCOSURMERCOSUR, or Southern Common Market, is a trade agreement be-

tween Argentina, Brazil, Paraguay and Uruguay. Although it was officially introduced in 1991 under the Treaty of Asuncion, its origins go back to the 1980s when Argentina and Brazil came together to form the “Argentina-Brazil 18 George Tavlas19 Kent Jones20 Elisa Parisi-Capone and Arun Motianey, Is a European Monetary Fund a Useful Complement to the

IMF?, 2010.


Integration and Economics Cooperation Program or PICE”21. This original program also proposed a common currency called “Gaucho” as an end goal. MERCOSUR was created to augment internal trade and the bloc’s negotiat-ing power. Advocates praise its success. In 2006 Brazil’s deputy UN envoy, Piragibe Tarragô, suggested that “poverty reduction within MERCOSUR countries ‘has been remarkable’”22. However, others argue that MERCOSUR is becoming too politicized. For example, Riordan Roett told the Council on Hemispheric Affairs that “‘Mercosur is no longer about trade… The new join-ers don’t have much to trade; they are opposed to free trade it seems. The organization is more and more political.’”23

Currently, all four countries are leaning towards a leftist political party with elected presidents. This can prove to be beneficial if they pursue a mon-etary union because the political figures seem to have the same general ideas. Normally, though, these countries have a corrupt political and judicial system which can create obstacles and great delays in the progress towards a currency union. In addition, disagreements amongst the four countries can be a signal of the poor level of integration, which are becoming prevalent. For example, Bolivia provides natural gas for both Argentina and Brazil, but it does not have enough natural gas to support the demand of both nations. After discussions between the three countries, Brazil “refuse[d] to share Bolivian natural gas”24. In addition, Uruguay’s interest in the continuity of MERCOSUR is debatable as they are discussing trade agreements with other countries without includ-ing its counterparts, which goes against the MERCOSUR agreement. Both of these examples raise a red flag in the level of integration of these four countries to pursue the next steps toward a monetary union.

The economies of the MERCOSUR countries seem to be strongly driven by their exports and agriculture. In addition, Brazil and Argentina both have strong tourism. The fact that exports and tourism are drivers for the MERCO-SUR economies should give confidence to policy makers of the possibility of a currency union for this group. If monetary policy is ceded, all four economies would benefit from currency depreciation in order to promote exports and en-courage tourism by increasing foreign purchasing power, for example.

Although the late 1990’s and early 2000’s was a very difficult time period for Latin American countries, it proved the level of influence one MERCO-SUR country can have on the others. When Brazil went through its currency crisis, both Argentina and Uruguay were also affected. During the 1980’s, the Argentinean currency was pegged to the dollar to stabilize hyperinflation. 21 http://www.allied-co.com/ri/ri/003030125.html22 http://www.cfr.org/publication/12762/mercosur.html23 Joanna Klonsky and Stephanie Hanson, Mercosur: South America’s Fractious Trade Bloc, 2009.24 http://www.nytimes.com/2008/02/25/business/worldbusiness/25iht-energy.1.10354521.html?_r=2


However, the dollar appreciated in the 1990’s, which forced the exchange rate in Argentina to increase as well. Concurrently, Brazil faced a currency crisis, so their exchange rate was depreciated, making Brazilian exports more afford-able. As a result, the Argentinean exports became less attractive, driving the already struggling economy to an economic crisis. As the Brazilian Real and the Argentinean Peso were depreciated, the exports of Paraguay and Uruguay decreased which eventually led to a financial crisis in Uruguay.

The goal of a currency union is such that when one country is going through difficult times, the shocks can be absorbed and balanced by its mem-ber countries. However, MERCOSUR is not a monetary union yet, so it would be unfair to expect them to perform like such at this early stage. Nevertheless, the volatility experienced by these countries during the last two decades and the current vulnerability of Argentina, Paraguay and Uruguay causes one to question the sustainability of a sound monetary union in MERCOSUR.

VI. Methodologies and ResultsTo quantitatively test the level of integration between the MERCOSUR

countries, three econometric models were used: Ordinary Least Square regres-sion, Granger Causality and Vector Autoregression (VAR). These econometric models will help gauge the relation amongst the four MERCOSUR countries. The last model is based on descriptive statistics to measure their level of trade amongst each other.

The variables used for the first three models are macroeconomic indica-tors that affect monetary policy: inflation, exchange rate and budget surplus/deficit as a percentage of GDP. The last model is based on a potential benefit of a currency union: increase in trade. The selected variables used are: imports, exports and GDP.

Ordinary Least Square Regression and resultsThe OLS model can estimate the dependent variable based on other in-

dependent variables. It helps approximate the value of the dependent variable based on changes of one of the independent variables, while the other inde-pendent variables are held fixed. The strength of the model can be confirmed through different tests. For example, the R-Square is a representation of what percentage of the dependent variable is explained by the independent vari-ables. There is no specific requirement of how high an r-square must be, but generally the higher the better. However, if irrelevant variables are added or omitted from the model it may bias the results and/or create multicollinearity. Models with low r-squared can be helpful at determining the relationship be-tween an independent variable and the dependent variable.


Ideally, the model should be as simple as possible, so only relevant vari-ables are included, which is determined through a hypothesis test. After drop-ping a variable, the strength of the model can be verified by comparing the full-model with the simple model by the probability of the variables, the adjusted r-squared and AIC and SIC numbers. The OLS equation is the following:

iii eXbbY ++= 10 25

The OLS is used in this paper to better understand how closely related the economy of all these countries are to each other. For example, the first regres-sion of exhibit 2 is an equation where the dependent variable is the inflation of Argentina and the independent variables are the inflation of Brazil, Para-guay and Uruguay plus a constant variable. This regression has an adjusted r-squared of .365, and AIC value of 15.47 and SIC value of 15.65. The equation is the following:

InfArg = (-219.430)C + (0.358)InfBrz + (14.263)InfPar + (3.524)InfUruThe only variable that is significant at 10% is the inflation of Brazil be-

cause it has a p-value of .0357. The inflations of Paraguay and Uruguay have a p-value of .546 and .590, respectively, so they are insignificant. Given the high p-value (higher than .1), these two variables are dropped from the equation to test if a simpler model would yield better results.

The simpler model has an adjusted r-squared of .328 and AIC and SIC values of 15.47 and 15.56, respectively. While these numbers did not improve (except for SIC), the significance of the inflation of Brazil for the model did. The p-value of the inflation of Brazil decreased to .0003 which suggests a cer-tain level of relationship between the inflation of Argentina and the inflation of Brazil. The simplified equation is:

InfArg = (67.632)C + (0.546)InfBrzBased on the results from the OLS Regression model, the following conclu-sions can be made which can be further analyzed in Exhibit 2:

Inflation: There seems to be a relationship between the inflation of Argentina and Brazil as well as the inflation of Paraguay and Uru-guay. All four full-models were simplified to only have one indepen-dent variable.

Exchange Rate: There appears to be correlations between the ex-change rates of all four countries, so all full-models were kept. How-

25 Yunwei Gai


ever, these results need to be analyzed carefully because there could be multicollinearity between the variables given the nature of the macroeconomic indicator used. All exchange rates were based on the dollar, so outside factors such as the strength of the dollar could have influenced the model.

Budget Surplus/Deficit as a % of GDP: The results when Argentina and Brazil were the dependent variables imply that their respective budgets are correlated by the budget of the other three countries. However, when Uruguay was the dependent variable, Paraguay was insignificant, so a simpler model was used after dropping Paraguay. When Paraguay was the dependent variable, Uruguay was dropped, and then Brazil became insignificant in the second model so it was also dropped, leaving the budget of Paraguay to be explained only by the budget of Argentina.

Granger CausalityOne of the major limitations of the OLS model is that it can only deter-

mine whether there is correlation between two variables. It cannot pinpoint which variable is causing the relationship. As a result, other econometric mod-els are also used to complement the OLS findings.

Correlation does not imply causation. Like previously explained, the OLS model only helps determine correlations between variables, but Granger Cau-sality determines the direction of these relationships. The Granger causality test is a statistical hypothesis test to determine the cause and effect of two variables. The information the Granger test yields helps determine which vari-able is the effect of a change in the other variable. This is important for this paper because it is not only imperative to understand that there is a relationship between two variables, but also what the direction of that relationship is. The results are in Exhibit 3, and the conclusions are the following26:

Inflation: Based on the Granger Test using Lag 1 and Lag 2, done on inflation of the four countries, it appears that the inflation of Ar-gentina “Granger Causes” the inflation of the other three countries, but not the other way around. It can be concluded from this that the inflation of Argentina influences the inflation of Brazil, Paraguay and Uruguay. These results raise a red flag given the volatility of Argentina’s inflation in the past three decades.

26 Eviews Help: Granger Causality


Exchange Rate: Based on the Granger Test with 2 lags, the Brazil-ian Real and the Paraguayan Guarani seem to be the two currencies causing the changes in the other two exchange rates. Brazil affects Paraguay and Uruguay while Paraguay affects Argentina and Uru-guay. When using 1 lag, there are a greater number of “Granger Cau-salities” where all the currencies are influencing at least one other exchange rate. These findings are questionable given Paraguay’s small size and economy. However, the limitations of the Granger Causality methodology will be clarified in more detailed below, which can help explain these findings.

Budget Surplus/Deficit as a % of GDP: When running the test with 2 lags, Uruguay seems to be causing Argentina to change and Para-guay causing Uruguay to change. Similar results come up when run-ning the test with one lag, except that the Argentinean budget causes the Paraguayan budget to change. So there seems to be a triangle between Argentina, Paraguay and Uruguay. Brazil’s budget appears to be unaffected by any of these three variables according to the Granger Causality Test.

Similar to the OLS Model, the Granger Causality Test model also has limitations. The causations implied by the model might not be perfectly ac-curate. There are instances where there is a bidirectional Granger Causality between two variables which makes it difficult to interpret which one is caus-ing the other. “The Granger test is designed to handle pairs of variables, and may produce misleading results when the true relationship involves three or more variables”27.

For example, when analyzing the exchange rate of Uruguay and Brazil with 1 lag, both hypotheses can be rejected, thus concluding that the exchange rate of Uruguay affects the exchange rate of Brazil to change and vice versa. Based on the Granger model a relationship exists, but there could be external factors other than the Uruguayan Peso and the Brazilian Real that could be affecting the exchange rates. Since both variables are based on the value of the US Dollar, the results may be biased as the model does not take the dollar into account. Another example is the link between inflation and exchange rate through the theory of Purchasing Power of Parity. Although both of these in-dicators are studied in this paper, they are analyzed separately, so the Granger model does not capture the effects inflation has on exchange rate.

27 Eviews Help: Granger Causality


Vector Autoregression: Impulse Response Function and Variance Decomposition

Vector Autoregression (VAR) is an econometric model that includes all variables based on a lag. “All the variables in a VAR are treated symmetrically by including for each variable an equation explaining its evolution based on its own lags and the lags of all the other variables in the model”28. The impact of a change in one variable over the others can be measured through the structural analysis allowed by a VAR model. For this paper, the VAR model is mainly used to interpret the impulse response functions, and forecast error variance decompositions. These two components of the VAR model help gauge the change one variable has over the others.

The impulse response function within the VAR model helps estimate the path that a one time-shock to the independent variable creates on the depen-dent variable. This tool yields the change all the independent variables have over the dependent variables for the selected period of time. For example, the results on Exhibit 5, “Response of InfBrz to InfArg” suggest that if the infla-tion of Argentina changes, the Inflation of Brazil will also see a significant effect on their inflation. These results provide an additional means to better gauge the relationship between MERCOSUR29.

The results of inflation seem to be consistent with the Granger Causality test. According to the results from the aforementioned test, the inflation of Argentina “Granger causes” the inflations of Brazil, Paraguay and Uruguay based on 1 and 2 lags. The Impulse Response of inflation also yields similar results. The first column of Exhibit 5 depicts the shocks a change in the Argen-tinean inflation would have on the other respective inflations. So for example, the graph titled “Response of INBRZ to INFARG” illustrates the changes the inflation of Brazil would go through if the inflation of Argentina changes. Based on these results, a change to the inflation of Argentina would cause the inflation of Brazil to experience great volatility as well. The inflation of Brazil would first sharply increase and then return to its base value at around time period 4, but then increase again and slowly decrease. These results suggest that a shock to the inflation of Argentina would cause the inflations of Brazil, Paraguay and Uruguay to change. The results are in Exhibit 5 and the conclu-sions are the following:

28 Eviews Help: Vector Autoregression29 Eviews Help: Impulse Response Function


Impulse Response Function Results - Inflation:

Overall, a change in the Argentinean inflation would seem to have the greatest impact on the inflation of the other MERCOSUR coun-tries. However a change in the inflation of the other three MERCO-SUR countries has little to no effect on the inflation of Argentina.

Impulse Response Function Results - Exchange Rate:

Overall, a change in the Brazilian Real seems to have the greatest impact on the exchange rate of the other MERCOSUR countries. A shock to the Paraguayan Guarani also affects the exchange rates of the other three countries, but not as significantly as those shocks to the Brazilian Real.


Impulse Response Function Results - Budget Surplus/Deficit as a Percentage of GDP:

Overall, a change in the Argentinean surplus/deficit has the greatest impact on the other MERCOSUR countries, but Brazil also seems to be influencing this metric.

Even though the impulse response functions estimates the consequences of a change of a variable to another variable in the VAR, variance decomposi-tion is an additional tool used because it “separates the variation of the variable into the component shocks to the VAR.” The difference is that the variance de-composition yields information about “the relative importance of each random innovation in affecting the variables in the VAR.”30 In other words, variance decomposition explains how much of a change in a variable is due to its own shock and how much is due to the shocks of the other variables.

Similar results are seen for the Variance Decomposition in Exhibit 6 as they were in Exhibit 5 for Impulse Response. This model shows that a change in their own country’s economic indicator causes a large percentage of the fu-ture changes of that country’s indicator. For instance, Exhibit 6 under “Percent InfBrz variance due to InfBRz” implies that a large portion of Brazil’s inflation is caused by a change in their inflation, especially in the first periods. Similar results are evident in the inflation of Argentina, Paraguay and Uruguay. The country whose inflation is most affected by their own inflation is Argentina, as the percentage of variance remains very high throughout the whole graph. However, after further analyzing the graphs of the other three countries, the percentage variance caused by their own inflation decreases after only a few years, which implies their inflation is affected by other factors, unlike the Ar-gentinean inflation. Interestingly, the percentage variance of inflation for these

30 Eviews Help: Variance Decomposition/


three countries caused by the inflation in Argentina sharply rises around the same time. This suggests that the inflation of these countries is strongly influ-enced by the inflation of Argentina.

The results can be seen in Exhibit 6, and the conclusions are the follow-ing:

Variance Decomposition Results - Inflation:

Overall, a change in the Argentinean inflation seems to have the greatest impact on the inflation of the other MERCOSUR countries. Again, this is not a good sign for the performance of the MERCO-SUR countries as the Argentinean inflation has not performed very well in the past few decades

Variance Decomposition Results - Exchange Rate:

Overall, a change in the Brazilian Real appears to have the greatest impact on the exchange rate of the other MERCOSUR countries fol-lowed by the Argentinean Peso.


Variance Decomposition Results - Budget Surplus/Deficit as a Percentage of GDP:

Overall, a change in the Argentinean surplus/deficit as a percentage of its GDP has the greatest impact on the other MERCOSUR coun-tries, but Brazil also seems to be influencing this metric.

Descriptive StatisticsThis model will help analyze the current level of trade between the MER-

COSUR countries. The percentage a country exports and imports from and to each MERCOSUR country will determine the amount of total trade that is done between the four countries. In addition, the GDP as a percentage of ex-ports and as a percentage of imports will help compare the trade levels of each country with the overall economy. The results are divided into four sections, each representing one of the MERCOSUR countries.

ArgentinaThe majority of exports and imports of Argentina either go to or come

from Brazil. Over the past 30 years the percentage of Argentina’s total exports and imports to Brazil is 16% and 22%, respectively. The total percentage of trade with Paraguay and Uruguay is not as significant as a total of Argentina’s total trade. Similarly, the trade as a percentage of GDP is the largest with Bra-zil at an average of 2%. Imports as a percentage of GDP appears to be larger than the exports as a percentage of GDP as seen in Exhibit 7.

BrazilIn MERCOSUR, the majority of Brazil’s trade comes from and goes to

Argentina; however, it is not as significant as Argentina’s percentage of total trade to Brazil. In addition, the amount of trade in Brazil with MERCOSUR as a percentage of GDP is also minimal at no more than 1% with any of the countries as can be seen in Exhibit 8. As a whole, the exports as a percentage


of GDP are greater than the imports as a percentage of GDP for Brazil.

ParaguayIn the 1980’s and 1990’s the majority of Paraguay’s total exports went

mainly to Brazil, but also to Argentina. However, in the last decade the per-centage of exports to Argentina and Brazil slightly decreased, but the percent-age that goes to Uruguay spiked up from 2% to 19% as can be seen in Exhibit 9. As a whole, the percentage of Paraguay’s exports going to MERCOSUR increased in the past decade due to the increase in exports to Uruguay. How-ever, the percentage of total Paraguayan imports from MERCOSUR countries remained stable over the past three decades: the majority coming from Brazil, then Argentina while Uruguay remained at a low 2%.

As a result, exports as a percentage of GDP are very similar. The major-ity of exports go to Argentina and Brazil, but the last decade saw a significant increase in exports to Uruguay. Overall, imports as a percentage of GDP was significantly higher than exports as a percentage of GDP for the past three de-cades with an average of 29% and 14%, respectively.

UruguayAs can be expected from the previous results, the majority of exports

from Uruguay to MERCOSUR countries went to Argentina and Brazil. Al-though exports increase from the 1980’s to the1990’s, they declined in the last decade. Similarly, Argentina and Brazil make up the majority of the imports to Uruguay, but as was seen in the previous results, the trade from Paraguay to Uruguay increased from 1% to 4% in the last decade, and as a matter of fact they went up to 7% in 2009. Overall, Uruguay imports more than they export measured by a percentage of GDP, but not by much as can be seen in Exhibit 10. The average for imports as a percentage of GDP for the past thirty years was 17% vs. 15% for that of exports.

In conclusion, the majority imports and exports go to or come from Ar-gentina and Brazil. However, the level of trade of Argentina and Brazil with MERCOSUR is significantly less than that of Paraguay and Uruguay. This can be explained from the size difference of Argentina and Brazil with that of Paraguay and Uruguay.

VII. ConclusionSince the inception of MERCOSUR, Argentina, Brazil, Paraguay and

Uruguay, collectively, have gone through tremendous changes. Three of the countries recently underwent an economic crisis. Brazil has also completely turned around to become one of the most promising countries in the world.


It is important to point out that when the currency crisis struck Brazil, both Argentina and Uruguay followed suit. For better or for worse, this is in-dicative of the influence one economy can have on others. Also, Uruguay and Paraguay, the smaller economies in the group, heavily depend on exports from both of their main trading partners - Argentina and Brazil. This is another sign of the similarities amongst the MERCOSUR members.

However, similarly to previous studies this paper also suggests that Latin America is not economically and politically ready for a currency union. Also, based on the limited level of integration between MERCOSUR countries, and the struggles experienced in the EMU, the feasibility of a monetary union in MERCOSUR is unrealistic. If MERCOSUR were to introduce the “Gaucho” (the name Brazil and Argentina originally gave to their potential currency) today, the union would be made up of four completely different countries. For example, Argentina is still trying to improve its global relationships from the economic crisis it suffered earlier in the early 2000’s when the country de-faulted on its loans. Brazil, on the other hand, has recuperated very well from its currency crisis and is positioned for tremendous growth. While Paraguay and Uruguay seem to depend on Argentina and Brazil, the results indicate Ar-gentina and Brazil are not influenced much by these two.

Germany is one of the most powerful nations of the European Union, so after the debt crisis in Greece exploded, they saw the future of the Euro partially on their shoulders. Given the strong growth in Brazil, if MERCO-SUR introduced a single currency, there would probably be great parallelism between Germany and Brazil. When the economy is prospering, the media would applaud the successful union, similarly to how the Euro was praised. However, as the cycles of the economies change, MERCOSUR would not be able to withstand itself as one bloc. Instead, each country would most likely pursue its own interests instead of the well-being of the “Gaucho”. A country like Argentina would probably play the role of Greece and hold large amounts of debts that they cannot sustain. History would then repeat itself and Brazil would be forced to save the struggling “brothers”.

As a result, MERCOSUR should not introduce a monetary union right now. If MERCOSUR were to implement this union, Brazil would probably be the economically dominant member. Consequently, the Brazilian government should cautiously consider whether a unification of currencies with the other three MERCOSUR countries can be mutually beneficial. Brazil’s broad econ-omy and strong current global position limits the potential benefits of a cur-rency union, and as recently experienced in Europe, the costs can be tremen-dous. Likewise, Paraguay and Uruguay should also carefully analyze whether the costs involved are in fact worthwhile. As of right now, both countries have


strong trading relationships with Argentina and Brazil, so the potential increase in trade would probably not suffice the loss of independent monetary policy. Argentina should also slowly consider the possibility of a monetary union. They are in the process of overcoming the financial crisis that left the country deeply indebted. Argentina should first solve their financial difficulties before embarking into a complex and limiting agreement.

In “Wealth by Association: Global Prosperity through Market Unifica-tion”, Edmunds and Marthinsen outline the steps a group of countries should follow to attain a successful currency union, economic union and/or political union. The initial steps are as follows: Preferential Trading Agreement, Free Trade Area, Customs Union, and Common Market. Only after a group has a successful common market should it pursue further integration. Edmunds and Marthinsen do not suggest any particular order for the next step, so the group should only pursue what is more aligned with their goals31.

Since MERCOSUR is in the process of becoming a customs union, it should first focus on strengthening that union. The customs union between these four countries is a restricted one and it is not at the same stage as the cus-toms union in Europe. As Dr. Jones, economics professor at Babson College explains, a customs union has a common external tariff which is controlled centrally and should only be considered complete after they have a voice in the World Trade Organization (WTO). MERCOSUR’s customs union has limita-tions, so it does not apply equally to all products and services and a common external tariff is not in place yet. In addition, the vision for a currency union in MERCOSUR has gradually declined. Uruguay has pursued trade agree-ments with other countries in order to reduce their reliance on the rest of MER-COSUR. This goes against a customs union since all members must hold the same level of protectionism with nonmember countries. Until the goals of all members are not clear and similar, MERCOSUR cannot continue pursuing a monetary union.

Following the customs union, MERCOSUR should create a common market. The common market for MERCOSUR would be the equivalent to the European Common Market in Europe. The difference between this step and a customs union is that the common market opens the borders to labor and capi-tal mobility. This would increase the competition to produce products/services with the highest level of efficiencies available for the group as a whole.

All these steps towards a monetary union are essential not only because they set a strong foundation, but also because they encourage integration which would facilitate the ultimate goal. Edmunds and Marthinsen explain that the

31 John Edmunds and John Marthinsen, Wealth by Association: Global Prosperity Through Market Unification, 2003.


next step can be any of three choices: monetary union, economic union, and political union. The difference between a common market and a monetary union is that the group would have to assign an authoritative body who would have total control of all monetary policy. Even though every country would cede its monetary policy tools, this central bank would be made up of repre-sentatives from each country. If the ultimate goal of this potential union would be to have a “gaucho,” or in other words, a monetary union, careful consider-ation should be given to some of the deficiencies from the European model.

A monetary union should be accompanied by an economic union, where a supranational authority responsible of fiscal policy is included, similar to the system in the US. As a matter of fact, the United States had an economic union many years before having a currency union. The Federal Reserve Bank was created at the end of 1913, thus creating a central bank and a monetary union in the US. Europe, on the other hand, did it the other way around. They created a currency union without an economic union. As previously explained there is no formula for the best way to attain an optimum currency area, but the European crisis has made it clear that a monetary union is not feasible without a systematic way to promote fiscal union.

If a central fiscal authority is not in place to promote this level of integra-tion, then strict guidelines should be set to attain disciplined fiscal policies between the member countries. Fiscal integration can provide the group with the necessary adjustments needed to overcome certain crisis such as the recent one seen in Greece.

In conclusion, the findings from this paper support the results found from the previous literature reviews: a currency union in Latin America would not make sense right now. There are numerous reasons for why this would be a disadvantageous union for MERCOSUR to pursue right now. For example, the economies in MERCOSUR are too different, so the policies each of them takes might be completely diverse. Also, the level of unity is not quite there yet, while some countries influence others in some aspects, they are not inte-grated enough for such a drastic transformation. Uniting two countries that are not completely aligned with each other economically under one currency can prove to cause problems as those seen in Europe.

Until a set of countries can prove that they are in it together, then a cur-rency union should be avoided. This should go beyond one country bailing out the other – merely on the idea of avoiding contagion throughout the rest of the group. Instead, this level of “brotherhood” should be demonstrated before the unification of currencies. Although the EMU has taken many steps to encour-age such integration, there is still a great deal of work ahead of them, which should be a cautionary flag to policy makers in Latin America and the rest of


world. The EMU confirmed they are loyal to each other. However, signs suggest

otherwise. For example, how is it possible that France can trade with Ireland without transactions costs, while someone in Italy has to pay roaming charges for calling Greece? “Each country has its own service providers that enjoy monopoly types of rents in sectors such as telecommunications, electricity and banking”32. These examples of deficiencies and lack of “brotherhood” in Eu-rope could be resolved with the proper measures of fiscal integration.

George Tavlas, the Director General of the Bank of Greece and the Al-ternate to the Bank of Greece’s Governor on the Governing Council of the European Central Bank identifies nine criteria for selecting members of a cur-rency union: “similarity of inflation rates, degree of factor mobility; openness and size of the economy; degree of commodity diversification; price and wage flexibility; degree of goods market integration; fiscal integration; the need for real exchange rate variability; and political factors”33. Based on Dr. Tavlas’ criteria, MERCOSUR is not remotely ready. The level of light integration in Latin America and the idea of sharing a currency is a formula for disaster.

MERCOSUR will not be ready for such a radical move until several things happen. Firstly, these countries need to become more integrated. Based on the results from this paper, members of MERCOSUR are not economically integrated enough to mitigate some of the risks of ceding monetary indepen-dence. Secondly, all MERCOSUR members need to increase their economic performance in the long-run because convergence alone will not suffice. If all four countries become economically healthier, then the benefits of integrating into a currency union would become more apparent. Lastly, all MERCOSUR members must believe and act in such a way that proves they are all in it together. This point is an important one because there are no formulas to de-termine when countries form an optimum currency area, so there is no data that can indicate when MERCOSUR will be ready. As a result, MERCOSUR should analyze the Euro and other currency unions. Lastly, once all member countries are economically and politically suited for a monetary union, they should determine if a political union would make more sense before a currency union, and develop strategies to avoid a crisis similar to the one in Europe.

Despite the fact that the results suggest that a currency union in Latin America would not make sense, if the right measures are taken, in the long term it is a possibility. The benefits of a currency union can be very reward-ing, as long as the risks are carefully mitigated. If MERCOSUR follows the right steps similar to the aforementioned, then one day they will be ready for a

32 Walter Molano, Europe: Forging a Union, 2010.33 Dwight Venner, Challenges to Central Banking From Globalized Financial Systems, 2002.


monetary union. After all, the development of the European Monetary Union lasted about half of a century. In addition, data shows Europe was not ready for a currency union, but they decided to pursue the Euro due to the political ben-efits, which has never been analyzed for Latin American countries. Also, by incorporating a “gaucho”, MERCOSUR could stabilize their currencies and avoid similar currency crises as experienced in the late 1990’s.


Exhibits

Exhibit 1: General Statistics of MERCOSUR countries (2009)34

Exhibit 2: Ordinary Least Square Regression (OLS) Results:

InfArg = C + InfBrz + InfPar + InfUru

InfArg = C + InfBrz

34 https://www.cia.gov/library/publications/the-world-factbook/


InfBrz = C + InfArg + InfPar + InfUru

InfBrz = C + InfArg

InfPar = C + InfArg + InfBrz + InfUru


InfPar = C + InfUru

InfUru = C + InfArg + InfBrz + InfPar

InfUru = C + InfPar


ExcArg = C + ExcBrz + ExcPar + ExcUru

ExcArg = C + ExcBrz + ExcUru

ExcBrz = C + ExcArg + ExcPar + ExcUru


ExcPar = C + ExcArg + ExcBrz + ExcUru

ExcUru = C + ExcArg + ExcBrz + ExcPar

BudArg = C + BudBrz + BudPar + BudUru


BudBrz = C + BudArg + BudPar + BudUru

BudPar = C + BudArg + BudBrz + BudUru

BudPar = C + BudArg + BudBrz


BudPar = C + BudArg

BudUru = C + BudArg + BudBrz + BudPar

BudUru = C + BudArg + BudBrz


Exhibit 3: Granger Causality Results:

Inflation; 2 Lags:

Inflation; 1 Lag:


Exchange Rate; 2 Lags:

Exchange Rate; 1 Lag:


Budget Surplus/Deficit as a % of GDP; 2 Lags:

Budget Surplus/Deficit as a % of GDP; 1 Lag:


Exhibit 4: Vector AutoRegression (VAR) Results:

Inflation:


Exchange Rate:


Budget Surplus/Deficit as a % of GDP:


Exhibit 5: Impulse Response Results

Inflation:


Exchange Rate:


Budget Surplus/Deficit as a % of GDP


Exhibit 6: Variance Decomposition

Inflation:


Exchange Rate:


Budget Surplus/Deficit as a % of GDP:


Exhibit 7: Average of Trade between Argentina and MERCOSUR35

Exhibit 8: Average of Trade between Brazil and MERCOSUR

35 Euromonitor International


Exhibit 9: Average of Trade between Paraguay and MERCOSUR

Exhibit 10: Average of Trade between Uruguay and MERCOSUR


References

Allegret, Jean-Pierre, and Alain Sand-Zantman. “Monetary Integration Issues in Latin America: A Multivariate Assessment.” Panoeconomicus, 2008: 279-308.

Arun, Motianey, and Elisa Parisi-Capone. Is a European Monetary Fund a Useful Complement to the IMF? March 22, 2010.

Back, Aaron. “The Wall Street Journal.” EU Barroso: Greek Rescue Package to be Finalized Very Soon, April 30, 2010: 1.

Barrionuevo, Alexei. The New York Times: Argentina seeks to avoid energy crisis, as Brazil refuses to share Bolivian natural gas. February 25, 2008. http://www.nytimes.com/2008/02/25/business/worldbusiness/25iht-energy.1.10354521.html?_r=2 (accessed November 2010).

BBC News: Single Currency Pros and Cons. November 21, 1997. http://news.bbc.co.uk/2/hi/special_report/single_currency/25081.stm (accessed 03 15, 2010).

BBC: Argentina’s Economic Crisis. February 12, 2003. http://news.bbc.co.uk/1/hi/business/1721061.stm (accessed November 2010).

Berg, Andrew, Eduardo Borensztein, and Paolo Mauro. “An evaluation of monetary regime options for Latin America.” North American Journal of, 2002: 213-235.

Bruno, Kaitlin. Global Envision: Brazil’s Boom. August 4, 2008. http://www.globalenvision.org/2008/08/04/brazils-boom (accessed November 20, 2010).

Castle, Landom, and Stephen Castle. New York Times. February 28, 2008. http://www.nytimes.com/2010/03/01/business/global/01union.html (accessed March 15, 2010).

Edmunds, John C, and John E Marthinsen. Wealth by Association: Global Prosperity Through Market Unification. Babson Park: Praeger, 2003.

Edwards, Sebastian. Monetary Unions, External Shocks and Economic Performance: A Latin American Perspective. National Bureau of Economic Research, 2006.

EIU. Greece. February 2010.“EIU: Greece March Report.” European Intelligence Unit. March 2010. (accessed

March 2010).Emsden, Christopher. “Euronomics: Spain Downgrade Spurs Contagion Fear, Italy

Hit.” Wall Street Journal, April 28, 2010: 1.Euromonito: Country Profile: Uruguay. October 15, 2010. http://www.portal.

euromonitor.com.ezproxy.babson.edu/Portal/DocumentView.aspx (accessed November 15, 2010).

“Euromonitor: Country Profile Argentina.” Euromonitor. November 7, 2010. http://www.portal.euromonitor.com.ezproxy.babson.edu/Portal/ResultsList.aspx (accessed November 20, 2010).


“Euromonitor: Country Profile: Brazil.” Euromonitor. November 5, 2010. http://www.portal.euromonitor.com.ezproxy.babson.edu/Portal/ResultsList.aspx (accessed November 20, 2010).

Euromonitor: Country Profile: Paraguay. October 15, 2010. http://www.portal.euromonitor.com.ezproxy.babson.edu/Portal/ResultsList.aspx (accessed November 20, 2010).

Euromonitor: Paraguay and Uruguay experience muted slowdowns. June 17, 2009. http://www.portal.euromonitor.com.ezproxy.babson.edu/Portal/DocumentView.aspx (accessed November 20, 2010).

Evangelist, Mike, and Valerie Sathe. “BRAZIL’S 1998-1999 CURRENCY CRISIS.” 2006.

Eviews Help: Granger Causality. 2010. C:\Program Files\EViews6SV\Help Files\wwhelp\wwhimpl\js\html\wwhelp.htm (accessed November 2010).

Eviews Help: Impulse Response Function. C:\Program Files\EViews6SV\Help Files\wwhelp\wwhimpl\js\html\wwhelp.htm (accessed November 20, 2010).

Eviews Help: Variance Decomposition. C:\Program Files\EViews6SV\Help Files\wwhelp\wwhimpl\js\html\wwhelp.htm (accessed November 25, 2010).

Eviews Help: Vector AutoRegression. 2010. C:\Program Files\EViews6SV\Help Files\wwhelp\wwhimpl\js\html\wwhelp.htm (accessed November 20, 2010).

Ferrari-Filho, Fernando. “Why Does it Not Make Sense to Create a Monetary Union in Mercosur? A Keynesian Alternative Proposal.” Journal of Post Keynesian Economics, 2002: 235-252.

Flippo, Hyde. German-Way: Euro-Timeline. 2010. http://www.german-way.com/euro-timeline.html (accessed April 27, 2010).

Gai, Yunwei. “Simple Regression Models and Diagnostics Part I.” Babson Park, MA, February 2009.

Google Finance. April 30, 2010. http://www.google.com/finance?q=EURUSD (accessed April 30, 2010).

Hallwood, Paul, Ian Marsh, and Jörg Scheibe. “An assessment of the case for monetary union or official dollarization in five Latin American countries.” Emerging Markets Review, 2006: 52–66.

Jones, Kent, interview by Moises Numa. Honors Project Thesis (April 2010). Babson College. Babson Park, MA.

Kiser, Sherry, and William C Gruben. Brazil: The First Financial Crisis of 1999. Southwest Economy, 1999.

Klonsky, Joanna, and Stephanie Hanson. Council on Foreign Relations: Mercosur: South America’s Fractious Trade Bloc. August 2009. http://www.cfr.org/publication/12762/mercosur.html (accessed November 2010).


Kopits, George. “Central European EU Accession and Latin American Integration: Mutual Lessons in Macroeconomic Policy Design.” North American Journal of Economics and Finance, 2002.

Laopodis, Nikiforos. “Currency Substitution and European Monetary Union.” Ekonomia, 2001: 190-207.

Lazaretou, Sophia. “Greek Monetary Economics in Retrospect: The Adventures of the Drachma.” Economic Notes, 2005: 331-370.

Mallet, Victor. “Roubini Global Economics: Overview of Solutions: Is a European Monetary Fund a Useful Complement to the IMF?” Roubini Global Economics. March 2010. http://www.roubini.com/critical-issues/48728.php (accessed April 2010).

Miles, William. “One Continent, One Currency? Varieties of Common Currency Experience in Europe and Latin America.” 2006.

Molano, Walter. “Europe: Forging a Union.” Roubini Global Economics, June 1, 2010.

Mundell, Robert. A Theory of Optimum Currency Areas. September 1961.Parisi-Capone, Elisa. “Roubini Global Economics: Is a European Monetary Fund a

Useful Complement to the IMF?” Roubini Global Economics. March 22, 2010. http://www.roubini.com/analysis/109795.php (accessed March 25, 2010).

Provopoulos, George. “Greece will fix itself from inside the eurozone.” Financial Times, January 21, 2010: 1.

Ross, Andrew. “The New York Times: Europe Now Moving Swiftly on Greek Rescue Plan.” The New York Times. April 29, 2010. http://dealbook.blogs.nytimes.com/2010/04/29/europe-now-moving-swiftly-on-greek-rescue-plan/?src=busln (accessed April 30, 2010).

Tavlas, George, email correspondance with Moises Numa. (April 6, 2010).Tavlas, George S. “Benefits and Costs of Entering the Eurozone.” Cato Journal

(Cato Journal), 2004: 89-106.The Economist. “The Future of the Euro: Don’t do it.” December 2, 2010.The World Fact Book. 2010. https://www.cia.gov/library/publications/the-world-

factbook/ (accessed December 2010).Trading Economics: Euro Area Unemployment Rate. February 2010. http://www.

tradingeconomics.com/Economics/Unemployment-rate.aspx?Symbol=ESP (accessed April 27, 2010).

Venner, Dwight. “Challenges to Central Banking From Globalized Financial Systems.” 2002.

Wagner, Helmut, and Eduard Hochreiter. The Road to the Euro: Exchange Rate Arrangements in European Transition Economies. Vol. 579. Sage Publications, Inc.


Willis, Andrew. “EU Observer: Plans emerge for European monetary fund .” EU Observer. March 08, 2010. http://euobserver.com/9/29623 (accessed March 25, 2010).

61

Law of Yuan Price: Estimating Equilibrium of the Renminbi

Joshua Klein Lipman1

The Wharton School of the University of Pennsylvania

Abstract

This study seeks to estimate an equilibrium value for the Chinese currency, the Renminbi. While existing literature puts forth various measures of equilibrium value, there exists a gap in reconciling historical trends with current estimates, which vary from 60% undervaluation to 5% overvaluation. I perform several Purchasing Power Parity (PPP) studies of the Renminbi vis-à-vis the U.S. Dol-lar, the Euro, and the Yen. The tests presented represent an estimate of the market equilibrium value given the relevant macroeconomic data available. There are different methods to interpret the available data, which result in dif-ferent estimations of equilibrium. In order to present a complete picture using all relevant data, I perform 14 empirical tests. They show the effect of changes in: base year (1978 through 2005), Absolute versus Relative PPP and the use of CPI versus PPI. I compare my findings to three macroeconomic estimates of equilibrium: Fundamental Equilibrium Exchange Rate (Williamson 1994), Behavior Equilibrium Exchange Rate and Permanent Equilibrium Exchange Rate (Funke and Rahn 2006). This comparison suggests the macroeconomic studies are generally in line with PPP comparisons from the early 1990s. The results also show a considerable effect on the exchange rate from the 1994 appreciation of the Renminbi by Chinese monetary authorities. The summa-tion of evidence suggests an equilibrium exchange rate around 4.0 RMB/USD, which implies a 37.5% undervaluation. For all measures of undervaluation, I use the exchange rate from the time of publication (April 2011), 6.5 RMB/USD. Lastly, I examine how China has recently allowed Renminbi apprecia-tion. Assuming similar actions going forward, I propose it could take at least five years of steady appreciation for the Renminbi to reach equilibrium.

1 Joshua Klein Lipman graduated cum laude from the Wharton School of the University of Pennsylvania in December 2010 with concentrations in Finance and Management. He would like to thank Gordon Bodnar, Morris W. Offit Professor of International Finance and Director of the International Economics Program at the School of Advanced International Studies of The Johns Hopkins University for his generous insight, guidance, and mentorship throughout writing this paper. Comments can be directed to the author at [email protected]. All errors are his.


I. TheoryBecause the Chinese Renminbi is not traded in a free market, there exists

no clear market equilibrium value of the exchange rate. The objective of this research is to estimate what an equilibrium rate might be using other mac-roeconomic data, such as inflation. Generally, economists use two methods to estimate equilibrium exchange rates. First is the Purchasing Power Parity (PPP) method, which assumes that exchange rates will come to equilibrium at the rate which aligns prices between countries. I use various empirical tests using two types of PPP methods: Relative Purchasing Power Parity (RPPP) and Absolute Purchasing Power Parity (APPP). With RPPP, equilibrium is as-sumed to hold in a certain base year, and then inflation differentials between two countries drive the difference in the estimate for equilibrium going for-ward. APPP, however, goes further in assuming that the price of a given good will reach equilibrium through a change in the exchange rate.

The second method of calculating exchange rates involves finding mac-roeconomic balance, which seeks an exchange rate that would allow for an equilibrium balance of payments. The macroeconomic balance approach is more complex; it draws on more data and requires more subjective assump-tions (Dunaway and Li 2005). While PPP uses less data as input, it also has the advantage of theoretical consistency in making fewer assumptions. While I present empirical evidence with only the PPP framework, I explain in the third section some of the leading macroeconomic balance estimates for the equilibrium value of the Renminbi.

Evidence under the PPP framework show how currencies change relative to the inflation rates in the countries. The theory is based on arbitrage and the Law of One Price. If there is free access to trade markets (no trade barriers, transaction or transportation costs) and exchange rates are determined by sup-ply and demand, then rational agents will buy and sell goods and currency until the price of goods across the world reaches equilibrium. The price cannot depart from equilibrium because that would provide an arbitrage opportunity to buy the cheaper good, sell the more expensive good and earn a risk-free profit. Therefore, there will be one, and only one, price of any given good. Practically, there are barriers to this type of arbitrage, such as price stickiness, transportation costs and transaction costs. However, the theory still holds, albeit with variance in the data (Froot, Kim and Rogoff 1995).

There are two ways to apply PPP. First, in its “strongest” form, PPP is an affirmation of the Law of One Price. Also called Absolute Purchasing Power Parity (APPP), this interpretation of PPP suggests that arbitrage will drive the price levels to equal each other. This assumption is difficult to prove, espe-

63Law of Yuan Price: Estimating Equilibrium of the Renminbi

cially in the short term, because it assumes easy trading (no transportation or transaction costs) and free markets. The “weaker” form of PPP is Relative Purchasing Power Parity (RPPP), which assumes equilibrium in a base year and then uses the difference in inflation from one country and another to pre-dict the expected change. RPPP suggests that “the proportionate change in the home country’s price level is equal to the proportionate change in the product of the foreign price level and the exchange rate.” Thus, RPPP does not claim that price levels equal each other (which APPP claims), but rather uses infla-tion differentials to show relative change between currencies. (Krueger 1983)

In general, PPP has been shown empirically as a weak predictor of future currency changes, and it holds especially weak in the short run. There is a significant effect, however, of mean reversion in the long run (MacDonald and Stein 1999). This is the theoretical basis behind the empirical tests in the paper. The reason that PPP holds in the long run is that exchange rates are the macroeconomic tool by which prices among countries adjust to allow for the Law of One Price to hold. If prices are not at equilibrium between two coun-tries, the action of buying in one country and selling in another requires similar trades in the currencies of the two countries. That market activity drives the exchange rate to equilibrium in the long run. For an economy with a fixed exchange rate, such as China, PPP estimates suggest an estimate of the equi-librium exchange rate based on the inflation differentials and an assumption of equilibrium in the base year. Because it is so difficult to know when (if ever) there was equilibrium, I show tests using a variety of the base years to provide a range of equilibrium points as the start of RPPP comparisons.

(A) Relative Purchasing Power ParityThis study uses inflation data and RPPP to find estimates of the RMB/

USD exchange rate. For each test, a base year is selected, and the exchange rate is assumed to be at equilibrium in that year. Assuming the base year repre-sents equilibrium, the difference of inflation rates between the two currencies is used to estimate what the exchange rate should be. If the base year in fact represents equilibrium, then the inflation-adjusted exchange rate represents equilibrium at every given date. Most importantly, the inflation-adjusted ex-change rate today gives the estimate of equilibrium today, which can be used to calculate the current under- or overvaluation of the Renminbi.

For example, suppose the U.S. Dollar and the Renminbi are at an equilib-rium exchange rate of 5.0 RMB/USD in the base year. Then suppose inflation for a year in the U.S. is 5% while inflation in China is 10%. The new inflation-adjusted exchange rate implied by RPPP is 5 * 1.10/1.05 = 5.24 RMB/USD. This implies that the U.S. Dollar will appreciate. As inflation in China out-


paces U.S. inflation, goods become relatively more expensive in China. The exchange rate adjusts for this so that one U.S. Dollar can buy the same goods in China at the end of the year as it could at the start of the year. An “inflation-adjusted exchange rate” (IAXR) line is included in each figure below to illus-trate the exchange rate that PPP would predict for any given year.

To provide a full picture of the current Renminbi exchange rate, this study compares Renminbi inflation with that of three other currencies: the U.S. Dol-lar, the Japanese Yen, and the Euro. These three countries comprise a majority of Chinese trade, and show how the Renminbi has changed vis-à-vis individ-ual trade partners. The U.S. and Japan are China’s two largest trade partners. Countries from the European Union constitute an additional 13% of China’s trade. Taken together, the U.S., Japan and Europe account for more than 40% of China’s exports and imports (The U.S.-China Business Council 2011) (Eu-ropean Commission 2011).

This study attempts to use as many base years as possible to show the changing effect of inflation differentials over time. In each figure, the ex-change rate is assumed to be at equilibrium at the base year. By using vari-ous base years, the study aims to provide a complete picture of exchange rate changes with different starting point assumptions. However, the lack of data on China makes it difficult to compare inflation rates starting from different base years. Producer Price Index (PPI) data is only available from late 1998 for China. Therefore, this study uses 2000 and 2005 for base years of RPPP comparison with PPI data between China and the U.S., Japan, and Europe.

Additionally, consumer price index (CPI) rates between the U.S. and China are compared. For CPI, there is data from China dating back to 1978, 1980, 1985, and annually from 1989. CPI data is normally less optimal to use in PPP than PPI data because PPI more accurately represents costs associated with tradable goods. In this case, however, CPI is used when it is the only inflation data available.

(B) Absolute Purchasing Power ParityAs compared with RPPP, APPP makes stronger assumptions. APPP does

not assume equilibrium in the base year. Instead, it estimates the equilibrium exchange rate in each year by applying the Law of One Price. APPP dictates that the equilibrium exchange rate will make one price for goods between cur-rencies. The exchange rate is the rate at which the price of a good or service is equal in the two currencies.

To estimate APPP equilibrium exchange rates, this study uses two “bas-kets” of goods. One is the World Bank GDP PPP rates. This is using GDP as the basket of goods, to which the World Bank applied a factor to adjust


for PPP. Secondly, this study uses the Big Mac Index from The Economist as a basket of goods to apply APPP. The Economist collects data on the price of one McDonald’s Big Mac hamburger in countries around the world. By treating the Big Mac as a basket of goods, this study finds the exchange rate at which a Big Mac would carry the same price between the U.S. and China. While the Big Mac may not seem to be a useful basket, it is attractive because of the availability of data and uniformity of the basket. The Big Mac Index as an indicator of PPP is discussed further with the corresponding data.

(C) Estimates of the Balassa-Samuelson EffectWhen calculating PPP equilibrium exchange rates, it is important to note

the effect of productivity changes and their effect on the real exchange rate equilibrium. This concept – known since its discovery in 1964 as the Balassa-Samuelson effect – argues that relative productivity discrepancies between countries affect the way that prices must be interpreted. Countries experience productivity gains through varies means: increased infrastructure, reduced barriers to international trade and advances in the technology of production.

Productivity gains, however, progress at different rates between trad-able and non-tradable products, made up of goods and services. The tradable products are able to capture more productivity increases, and cause wages in tradable product sectors to increase because of labor scarcity and workers’ increased power. These wage increases affect all labor markets, and wage in-creases extend to non-tradable products. In tradable product sectors, wage increases are balanced by productivity gains. In non-tradable sectors, there are also wage increases (the labor is from the same pool). However, there are no corresponding productivity gains along with the increase in prices in non-tradable sectors. This causes prices in the non-tradable sector to increase at a faster rate, which leads to inflation. When a country experiences inflation more than the rest of the world, its currency appreciates (Dunaway and Li 2005). This study does not calculate productivity differences as part of the equilibrium estimates. There already exist several studies that estimate the Balassa-Samuelson effect in China over the past decade. After presenting PPP estimates of the Renminbi equilibrium, this study summarizes some of the articles that have focused on the Balassa-Samuelson effect on the Renminbi exchange rate.

While the pure PPP estimates presented in this study show no significant shift in the value of the Renminbi between 2000 and 2010, many economists feel productivity changes have caused the Renminbi to be significantly under-valued, and that this undervaluation has grown over this time period. The un-dervaluation is consistent with the PPP estimate that uses 1978 as an assump-


tion of equilibrium, but the change from 2000 to 2010 cannot be explained by prices alone. Several studies have found a Balassa-Samuelson effect in China, which implies that the Renminbi is more undervalued than a simple PPP esti-mate would predict (Dunaway and Li 2005).

However, the magnitude of the Balassa-Samuelson effect in China is not clear. Specifically, the size of the effect varies based on the number of coun-tries included in the comparison and whether the country set is made up of developed or developing countries. In the case of China, rapid growth and high inflation make comparisons with other developing countries more applicable (Dunaway and Li 2005). There are various reasons why the Balassa-Samuel-son effect may not hold in modern China. First, the effect relies on accurate price data, and there is a lack of quality price surveys in China (which also presents a limitation of PPP estimates). Prices in China are very low, which can cause models to underestimate the degree of undervaluation of the Ren-minbi. (Cline and Williamson, Estimates of the Equilibrium Exchange Rate of the Renminbi 2007)

Additionally, the Balassa-Samuelson effect assumes that the economy is at or close to full employment. That is a necessary condition for increases in wages in tradable sectors to cause increases in wages in non-tradable sectors. In China, this may not be the case, as the past decade has seen new workers migrate from the rural areas to new jobs in the cities. While many of these workers may have worked in rural areas, they were not part of the formal work force in the cities. Their abundance and recent entrance into urban work forces has had the effect of adding new workers to the cities without requiring employers to increase wages. Non-tradable sectors have been able to maintain low wages, and have not caused significant price increases or the resulting ap-preciation of the currency. Therefore, the Balassa-Samuelson effect may be minimal in China over this time period (Dunaway and Li 2005).


II. Findings

(A) Relative Purchasing Power Paritya. RMB/USD Comparisons

Figure 1: RMB/USD CPI (1978)

This figure demonstrates how the historical exchange rate has depreci-ated significantly more than the inflation adjust exchange rate would have predicted. Starting with an assumption of an equilibrium exchange rate of 2.46 RMB/USD in 1978, the inflation-adjusted exchange rate predicts a stron-ger Renminbi than how the exchange rate actually changed. The exchange rate predicted by inflation would be 4.13 RMB/USD (assuming equilibrium in 1978. This translates to a 36% Renminbi undervaluation off the current value of 6.5 RMB/USD.

Using 1978 as a base year for China is not arbitrary. On a global scale, it was only a few years after the fall of the Bretton Woods system. The price of gold was no longer fixed, and currencies were trading on floating markets. In China, Deng Xiaoping came to power in 1978 and was a strong proponent in opening up the country economically to the rest of the world. Part of this economic liberalization was allowing for greater flexibility in currency con-version. Previously, from 1949 until 1978, China’s exchange rate had been pegged at 2.46 RMB/USD (Cheng 1998).



In 1980, the government began to allow 10 “swap centers” around the country, and China began a trend of a long-term, slow depreciation of the cur-rency (Cheng 1998). From 1978 to 1980, the Renminbi had been depreciating vis-à-vis the U.S. Dollar, and the nominal exchange rate was 1.74 RMB/USD in 1980.

Figure 2 appears similar to the one that uses 1978 as a base year, but Figure 2 uses an even lower assumption of equilibrium in 1980. Using this baseline assumption of 1.74 RMB/USD from 1980, the inflation differentials suggest a current inflation-adjusted exchange rate of 3.41 RMB/USD, which implies a 48% undervaluation. This is greater than the 36% implied by using 1978 as starting equilibrium. It is not clear which, if either, is a good estimate of the fair equilibrium exchange rate. It is important to note, however, that both show an estimate of equilibrium significantly lower than the nominal exchange rate today. These both imply that if 1978 and 1980 represent equi-librium, inflation differentials suggest Renminbi undervaluation today.

When 1985 is used as equilibrium, the model predicts an inflation-ad-justed exchange rate relatively close to where the exchange rate is trading cur-rently. Assuming the 1985 equilibrium, the inflation-adjusted exchange rate today is 6.63 RMB/USD, which implies a 2% overvaluation. The rise in the inflation-adjusted rate between 1985 and 2008 was due to high inflation rates in China. This inflation came along with strong economic growth of China through the 1990s and 2000s, driven in large part by economic liberalization


and strong Chinese exports.The fact that this figure arrives at an estimate close to the current rate

shows that the exchange rate since 1985 has tracked fairly closely with the change in inflation. In Figure 2, 1985 is used as equilibrium, and then equilib-rium is shown again today. But this is just one figure, and there may not have been equilibrium in 1985. Therefore, the relative under- or overvaluation of the Renminbi with the U.S. Dollar in 1985 is consistent with the current valu-ation of the Renminbi.

Using 1989 as a base year produces an estimate of equilibrium consistent with the estimate that uses 1985 as a base year. They both end with estimates of an exchange rate close to where current rate. In this case, using 1989 as equilibrium, inflation differential would suggest a current inflation-adjusted exchange rate of 7.05 RMB/USD. This implies an 8% overvaluation today. In the figure, the sharp depreciation of the Renminbi and ensuing inflation in China between 1993 and 1994 is apparent. In fact, on January 1, 1994, China officially changed its exchange rate to 8.68 RMB/USD, effectively entering the “Unified Rate Phase” (Cheng 1998).

In Figure 10, which uses the end of 1994 as equilibrium, there is no change nearly as dramatic as the devaluation in January 1994 that produces such strong inflation in China. In this figure, the increase in Chinese inflation can be noticed through the rise in the inflation-adjusted exchange rate through 1998. Assuming the exchange rate was at equilibrium in 1994, the figure sug-gests a current inflation-adjusted exchange rate of 9.24 RMB/USD, which translates to a 42% overvaluation.

This demonstrates why 1994 probably does not provide a useful base year to assume equilibrium. After the shocking devaluation of January 1994, inflation began to take hold in China, and took a few years to increase Chinese prices. The devaluation itself is evidence that equilibrium is not likely at this time. Assuming equilibrium directly after a sharp devaluation, as this figure does, can lead to misleading analysis that should not form the basis of sound conclusions about equilibrium.

Using 2000 as the base year for equilibrium shows an even more concen-trated view of the exchange rate and inflation changes. Since then, there have not been significant differences in inflation, which lead to a flat inflation-ad-justed exchange rate line ending at an equilibrium estimate of 8.24 RMB/USD, above the current exchange rate. That translates to a 27% overvaluation. The historical exchange rate was steady until July 2005, when China unpegged the Renminbi, and began to allow it to appreciate. That it now sits above the cur-rent exchange rate implies Renminbi overvaluation, but this is likely a cause of the choice of 2000 as equilibrium. It is unlikely that the currency was effec-


tively at equilibrium in 2000. Therefore, Figure 11 – along with the following figures that use 2000 as equilibrium – serves merely to demonstrate how infla-tion rates between the U.S. and China have differed since 2000, but are likely inaccurate as predictors of equilibrium exchange rates today.

Figure 3: RMB/USD PPI (2000)

Figure 3, which uses producer price indices (PPI) and a base year of 2000, is included here to establish a similarity between CPI and PPI data, because all figures comparing the Euro and Yen use PPI data. Additionally, Figure 12 in the appendix shows the PPI test with a base year of 2005. The PPI charts are in fact similar to their CPI counterparts. Inflation in China slightly outpaced U.S. inflation in this time period, but the inflation-adjusted exchange rate line remains relatively flat. The previous figures showed the same trend with CPI data. The CPI data showed less of a difference in inflation rates between the U.S. and China using 2000 as equilibrium. The historical exchange rate ap-pears to slope downwards more than it does in the CPI figure because this figure uses monthly rather than annual data, and this figure extends to 2010.

The PPI estimate that uses 2000 as a base year shows the inflation-adjust-ed exchange rate equal to the historical exchange rate around 2004, whereas the rates equal one another around 2006 in the CPI estimate. This is to be expected, as PPI is a leading indicator of CPI: Consumer goods tend to have sticky price and vary based on producer costs. The PPI estimate that assumes 2000 as equilibrium suggests an inflation-adjusted exchange rate today of 9.66 RMB/USD, which implies a 49% overvaluation. The PPI figure that as-


sumes 2005 as equilibrium shows less of an overvaluation of the Renminbi, and suggests a current equilibrium of 9.02 RMB/USD, which implies a 39% overvaluation.

b. RMB/EUR ComparisonsWhile the RMB/EUR estimate, using 2005 as a base year (Figure 14),

appears similar to the RMB/USD graph from 2005 (Figure 12), the Euro com-parison from 2000 (Figure 15) shows a marked difference from its USD coun-terpart, Figure 3. The Renminbi depreciated vis-à-vis the Euro from 2000 to 2005, which is consistent with larger currency moves at the time. The Euro was appreciating vis-à-vis the U.S. Dollar at that time, and the Renminbi was pegged to the U.S. Dollar. With an assumption that the Renminbi and Euro were at equilibrium in these figures, the tests produce estimates of 9.80 RMB/EUR using the 2000 base year, which implies 51% overvaluation. Using 2005 as a base year produces an estimates of 12.14 RMB/EUR which trans-lates to 87% overvaluation.

c. RMB/JPY ComparisonsFigures 15 and 16 in the appendix show the Renminbi against the Jap-

anese Yen in the same format as the previous figures. While the standard convention is to quote currencies against the Yen as JPY/FC, the currency is quoted in this paper as RMB/JPY to be able to visually compare the Figures 15 and 16 with previous figures. Otherwise, the movements would be inverted.

In comparing the Yen analysis with the others, it is clear that the infla-tion-adjusted exchange rate line appears different than it did in the Euro and U.S. Dollar figures. This is consistent with expectations because Japan has experienced such low levels of inflation. Therefore, it appears to have gained purchasing power against the Yuan. In addition, the historical exchange rate remains flat across both time periods. While the Renminbi has appreciated vis-à-vis the U.S. Dollar, the Yen has also appreciated, so the two currencies have maintained a fairly stable cross-rate. With an assumption of 2000 as equi-librium, the figure suggests an inflation-adjusted exchange rate of 0.13 RMB/JPY (7.50 JPY/RMB). This translates to 15% overvaluation. While that shows a significant discrepancy from the current exchange rate, the exchange rate suggested by using 2005 as equilibrium is 0.11 RMB/JPY and 45% over-valuation (9.42 JPY/RMB), which closer to the current exchange rate of 0.08 RMB/JPY (12.37 JPY/RMB).


(B) Absolute Purchasing Power Paritya. World Bank GDP Implied PPP Exchange Rates

Figure 4: RMB/USD World Bank Implied Rate (2000)

This rate is based on GDP figures published by the World Bank. The World Bank publishes PPP-adjusted GDP estimates, which are the World Bank’s estimate of what a country’s GDP would be if it had the purchasing power of the U.S. Dollar. The number is published as GDP/capita in U.S. Dollars. To find the RMB/USD exchange rate, I multiply the PPP GDP rate by population, and divide the actual Renminbi GDP by that number. As a check on the population number, I multiply the World Bank’s GDP/capita by popula-tion, and it in fact is in line with total GDP figures. The APPP estimates from these rates imply significant undervaluation of the Renminbi. In 2008, the implied exchange rate is 3.80 RMB/USD, which implies 42% undervalua-tion of the Renminbi.


b. The Economist Big Mac Index

Figure 5: RMB/USD Big Mac Index (2000)

The Economist has compiled the Big Mac index since 1986, reporting the price of one Big Mac hamburger in many of the 110 countries where it sold around the world. While it is not an actual basket of goods, the Big Mac Index has been gaining attention among academics and media outlets over the last few decades. Economists agree that the Big Mac Index produces a number of problems. First, it is only one product, and a food product, which is exposed to highly cyclical price shifts that are not indicative of prices in the rest of the economy. Second, the Big Mac itself is marketed as a different product in dif-ferent countries. In some countries, such as the U.S., the Big Mac hamburger is inexpensive and could be an inferior good compared with other food items. In other economies, including China, the Big Mac can be expensive for much of the population, and may even be seen as a higher-end food product. This discrepancy limits the effectiveness of the Big Mac Index.

Nonetheless, there are certain undeniable advantages to the Big Mac In-dex. Practically, it is cheap and easy to administer, and the data has been reli-ably gathered annually in many countries for almost 25 years. Also, the idea of the Big Mac “basket” encompasses all of the inputs into a Big Mac, including raw food costs, transportation costs, marketing and labor production costs. Regardless of the reason, several academic studies “generally find the Big Mac is surprising accurate in tracking exchange rates over the long run.” (Funke and Rahn 2006). Using the Big Mac Index, an equilibrium exchange rate for October 2010 is estimated to be 3.48 RMB/USD, or 46% undervaluation.


(C) Summary of FindingsThis section presented five figures and referred to nine more from the ap-

pendix. They all demonstrate various measures of the equilibrium Renminbi exchange rate. First, Relative Purchasing Power Parity presents evidence sup-porting the claim that the Renminbi is undervalued. Specifically, the devalua-tion in 1994 appeared to have significantly shifted the effective exchange rate. The increase in Chinese inflation relative to U.S. inflation suggests that the 1994 devaluation resulted in an exchange rate that undervalued the Renminbi.

The use of several years as the equilibrium starting points demonstrates the level of inflation differentials over time. From the data, it’s clear that infla-tion since 2000 has been relatively similar in the U.S. and China. Therefore, the level of over- or undervaluation of the Renminbi is likely similar today to its level in 2000. Before that, from 1985 to 2000, inflation in China outpaced inflation in the U.S (see Figure 1). While the two countries experienced simi-lar levels of inflation from 1990 to 1994, the significant disparity in inflation rates occurred from 1985 to 1990 and 1994 to 2000.

The comparison of the Renminbi with the Euro and Yen show that the trends vis-à-vis the U.S. Dollar are not specific to the Renminbi-Dollar rela-tionship. The graphs are not the same because they reflect the unique nature of the Euro and Yen during these time periods. The Euro experiences significant appreciation against other currencies in during the first part of the decade, and that accounted for much of the changes in Figures 13 and 14. The Yen has historically experienced very low levels of inflation, which accounted for the large inflation differentials in Figures 15 and 16.

With Absolute Purchasing Power Parity, the demonstrated effects of un-dervaluation are stronger. Both the implied World Bank rate and the Big Mac Index would predict an exchange rate below 4 RMB/USD, significantly below the exchange rate in April 2011 of 6.5 RMB/USD. These predictions assume that their basket of goods is an accurate representation of prices within the country. While both the Big Mac Index and the World Bank PPP rate are not perfect baskets, they are still relevant and provide a baseline of comparison between the U.S. Dollar and Renminbi.

III. Other MethodologyIn addition to the PPP estimates presented in the previous section, this

study reviews some of the leading estimates of Renminbi equilibrium found in recent academic studies. The most relevant to compare to the previous section are those which use the more complex macroeconomic balance approach. Be-sides for macroeconomic balance, almost all methods are variants of PPP and


draw on similar data sets. There are three such methods described in this study: the Fundamental Equilibrium Exchange Rate (FEER), the Behavioral Equilib-rium Exchange Rate (BEER), and the Permanent Equilibrium Exchange Rate (PEER). John Williamson of the Institute for International Economics devel-oped the FEER approach, and has updated his models as of January 2010. A 2006 paper by Michael Funk and Jörg Rahn of Hamburg University evaluated the Renminbi with the BEER and PEER methods.

In addition to the methods described here, there has been other significant research on the value of the Renminbi. One recent article analyzed 130 empir-ical estimates of the value of the Renminbi. It found that estimates were wide-ranging, from -66.70% to 161.70% misalignment of the Renminbi (Bineau 2010). The average is currently around 13%, but has changed significantly over time, as illustrated in Figure 6.

Figure 6: Estimates of Renminbi Misalignment over Time (Bineau 2010)

a. Fundamental Equilibrium Exchange RateThis method aims to find an equilibrium exchange rate so that a country

simultaneously achieves internal and external balance. For the purposes of this method, Williamson describes internal balance as “acceptance of the his-torically determined wage rate and achievement of a level of effective demand such as to sustain the highest level of activity consistent with the control of inflation.” External balance is a target current account large enough that it is sustainable and does not rely on capital inflows attracted by high interest rates. The target current account balance should also be small enough to avoid the domestic political pressure of a postponement of consumption and foreign po-litical pressures (such as the political pressures the U.S. is currently applying to China in hopes of reducing China’s large current account balance) (Wil-


liamson 1994).The FEER is a medium term equilibrium concept (Bineau 2010). To

arrive at a final value, the method uses values the exchange rate based on a number of real factors, the “fundamentals.” These variables include taxes, international terms of trade, commercial policy, capital flows, aid flows, and technology. Because this method is a function of so many factors, its value depends not only on the current values but also on the expectations of policy changes and economic shifts in the future (Elbadawi 1994).

In January 2010, Williamson published updated Fundamental Equilib-rium Exchange Rates, and included an estimate for the Renminbi at 4.9 RMB/USD, implying a 28% undervaluation from the rate at which it was trading, 6.8 RMB/USD. This estimate, however, relied on the IMF’s expectations for China’s current account surplus through 2012, which was around 10%. This figure appeared to be too high, as recent data suggest the current account sur-plus may be closer to 5-6%. A lower current account surplus has a dramatic effect on the FEER. For example, if the current account surplus is 7%, the FEER implies a 25% (rather than 28%) undervaluation vis-à-vis the U.S. Dol-lar (Cline and Williamson, Notes on Equilibrium Exchange Rates 2010).

b. Behavioral and Permanent Equilibrium Exchange RatesThe Behavioral Equilibrium Exchange Rate (BEER) is based on funda-

mentals of an economy surrounding internal and external balance, similar to the FEER. The difference, however, is that the BEER does not require equilib-rium. It avoids the subjective assumptions involved with selecting a sustain-able current account rate. Instead, the BEER is a “data-determined systematic component of the exchange rate in the medium and long run” (Funke and Rahn 2006).

The Permanent Equilibrium Exchange Rate (PEER) attempts to address the problem with BEER and FEER that they rely on fundamentals in the rest of the macroeconomy that are not at equilibrium. In order to arrive at values for these fundamentals, the PEER starts with a Hodrick-Prescott or bandpass filter to account for the effect of the business cycle in the data. Then, the PEER extends this notion to create a time series for fundamentals that breaks the vari-ables into permanent and transient components. The PEER uses the perma-nent components in a series of equations to find an equilibrium exchange rate.

The results of the evidence suggest that there is only modest undervalu-ation of the Renminbi. The BEER implies that the Renminbi has been under-valued by 6-15% at different points between 1999 and 2006. The PEER sug-gests that the Renminbi is undervalued by about 12%. Considered together, these estimates show that the undervaluation may be less than other estimates


suggest. The study also assumes a Balassa-Samuelson effect, which may not be significant in modern China (see Findings section).

IV. Significance of StudyThe Chinese exchange rate has become increasingly important in past

years as it has affected trade balances, the Sino-U.S. diplomatic agenda and currency policies in other countries. In each of these areas, the value at which the Renminbi trades has been important in determining how other countries set policy and interact with each other. An exchange rate has ripple effects throughout the world economy and global political environment.

(A) Current Account BalanceThere is no doubt that China has experienced massive economic growth

in the past decade. In fact, a recent working paper from the World Bank esti-mates that Chinese GDP grew by 10% or greater every year from 2003-2008 (Kuijs 2009). A strong driver of this growth has been Chinese exports, which approximately doubled between 2005 and 2009. These exports mostly com-prise manufactured goods, and their export price depends on the Renminbi exchange rate (World Bank 2010). With a weak Renminbi, exports are made relatively inexpensive, and foreign consumers demand more Chinese goods.

Simple macroeconomic theory dictates that a weak currency encourages exports (and decreases imports). The question is to what extent the current ac-count balance depends on the exchange rate. To test this link, William R. Cline from the Institute of International Economics conducted analysis on exchange rate and trade data to determine the statistical significance of the effect of the Chinese exchange rate vis-à-vis the U.S. Dollar on China’s global trade bal-ance and its bilateral trade balance with the U.S.

The empirical evidence at first glance seems to suggest a weak link. From July 2005 to August 2008, China allowed its currency to appreciate 21%, yet the U.S. bilateral trade deficit with China remained largely unchanged during the period. Yet Cline shows that China’s current account surplus is greatly influenced by the real exchange rate and the growth differential. The relation-ship shows an adjusted R2 correlation of 0.89. The coefficient in the equation, which is statistically significant, shows that a 1% rise in the real exchange rate leads to a 0.30%-0.45% fall in the current account balance as a percent of GDP (Cline, Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit 2010).

Then why didn’t the current account balance fall during the 21% appre-ciation of the Renminbi? The historical evidence shows a significant effect of the real exchange rate of the Renminbi on China’s global current account


balance and the bilateral trade balance with the U.S. The effect is significantly lagged, however. This can be attributed to the time it takes for manufacturers to interpret an exchange rate change as a long-term realignment rather than a short-term fluctuation. Also, assuming companies often hedge much of their foreign currency exposure, they may not feel the effects of immediate fluctua-tions in exchange rates. These factors contribute to a lag of about two years in the effect of exchange rates on a trade balance (Cline, Renminbi Undervalua-tion, China’s Surplus, and the US Trade Deficit 2010).

The data start in 2000, at a time when the Renminbi was pegged to the U.S. Dollar. Therefore, until the Renminbi was unpegged in mid-2005, the U.S. Dollar vis-à-vis other world currencies was the sole determinant of the value of the Renminbi against non-U.S. trading partners.

In 2003, China had a modest current account surplus of 2.8% of GDP. In the period of 2002-2005, the U.S. Dollar depreciated against other major world currencies. With that the Renminbi linked to that depreciation, its ex-ports benefited from the relative weak currency. The effect was shown in 2007 (consistent with the two year lag), when China experienced a current account surplus of 11% of GDP (Cline, Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit 2010).

In the following years, the U.S. Dollar saw a sharp appreciation. This occurred during the financial crisis, when investors fled from risky emerging markets to assets in safer countries such as the U.S. and Japan. The result was an appreciation of the U.S. Dollar during the crisis. Concurrently during the crisis, China effectively re-pegged its currency to the dollar and stopped its recent trend of appreciation to minimize the negative impact on exports. Therefore, the U.S. Dollar appreciation in this period affected the Renminbi in a way similar to the U.S. Dollar appreciation during 2002-2005. Following the crisis, and a two-year lag, China’s current account surplus has declined to around 6% of GDP in 2010 (Cline, Renminbi Undervaluation, China’s Sur-plus, and the US Trade Deficit 2010).

From this data, we can conclude that the Renminbi exchange rate has a significant effect on the current account balance. The effect is about 30-45%. This means that, using 2010 data, a 10% appreciation of the Renminbi would reduce China’s global current account surplus by $170 billion to $250 billion (Cline, Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit 2010).

(B) Sino-U.S. Political and Diplomatic AgendaSince the strong growth China exhibited in the 1990s, the Renminbi ex-

change rate has been a closely watched and highly contentious issue in politics


and diplomacy between the U.S. and China. From 1994 to 2002, the Chinese economy doubled while the trade-weighted value of the Renminbi increased by 30%. Then, from 2002 to 2005, China experienced even stronger growth, but not as much appreciation because the U.S. Dollar, to which the Renminbi was pegged, had been depreciating. China’s rapid growth could be largely at-tributed to exports; its export-GDP ratio was close to 40%. Over half of these exports were sold to consumers in the U.S. Politically, the issue arose as a major point of contention between the U.S. and China. There are two tools that the U.S. has used to pressure China on this issue: tariffs on imports and the characterization of currency manipulation in an annual report by the Treasury Department (Goldstein, Renminbi Controversies 2005).

On July 21, 2005, China announced that it would end its decade long peg to the U.S. Dollar. This followed a bill that was then in the U.S. to impose a 27.5% tariff on all goods imported from China. The Senate Finance Commit-tee had scheduled hearings on the Chinese currency regime for July 23, 2005 (The Economist 2005). Shortly after the announcement, in November 2005, the U.S. Treasury acknowledged that “a rigid exchange rate hinders China’s ability to move away from its current dependence on exports.” It went on to re-port, “President Hu told President Bush that China would unswervingly press ahead with reform of its exchange rate mechanism. The Chinese authorities should do so by the time this report is next issued.” The report labels certain countries as “currency manipulators,” but left China off the list, at least in part because of its steps taken in July 2005 (Department of the Treasury 2005). Since then, the Treasury Department has still not labeled China as a currency manipulator.

After the initial 2.1% appreciation, China met this expectation and al-lowed gradual appreciation of the Renminbi each year until 2008 (Goldstein and Lardy, The Future of China’s Exchange Rate Policy 2009). Then, in light of the global economic crisis, the Chinese authorities began to scale back the rate of appreciation. From September 2008 to early June 2010, the Renminbi was essentially fixed at 6.83RMB/USD (Cline, Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit 2010). Yet the political pressure did not immediately return. Other countries had been taking protectionist mea-sures, and there was no strong pressure on China to appreciate its currency during a global recession.

In 2010, however, the issue re-emerged. As the world began to recover from the recession, politicians in the U.S. and around the world once again felt the negative effects of Chinese exports. The U.S. began to consider measures to label China as a “currency manipulator” in the annual report from Treasury to Congress. A G-20 meeting in Toronto in July of 2010 provided a forum


where many world leaders were expected to pressure China to allow the Ren-minbi to appreciate.

On July 21 2010, exactly 5 years after the initial un-pegging, China an-nounced that it would once again begin to allow the Renminbi to appreciate. China followed the same strategy of trying to preempt political pressures by appreciating its currency. Yet, in the months following the announcement, the Renminbi appreciated less than 2%. The appreciation has been disappointing for U.S. lawmakers (Beattie 2010).

In response, two U.S. Congressmen introduced a bill treating currency manipulation as an illegal export subsidy. This bill would have given the U.S. more authority to impose tariffs on Chinese goods, but it did not pass. Re-sponding to U.S. pressures, a spokesperson for the Chinese Foreign Ministry, Jiang Yu, said, “I would point out that appreciation of the Renminbi will not solve the US deficit and unemployment problems.” In addition, the Senate Committee on Banking, Housing, and Urban Affairs and the House Ways and Means Committee called Treasury Secretary Timothy Geithner to testify on September 16, 2010. The committees spoke about the annual Treasury report to Congress, and asked Secretary Geithner whether China may be labeled as a currency manipulator in the report that is due to Congress on October 15th. Although he generally agreed that action was needed, he questioned the use-fulness of labeling China as a currency manipulator in the Treasury’s report: “Wishing something does not make it so – and issuing a report that requires me to go and consult – changes nothing” (Beattie 2010).

The effect of the Chinese exchange rate on the U.S.-Chinese bilateral relations is clear. In 2006, the U.S. Treasury appointed a full-time Financial Attaché to represent U.S. interests in this area and other (Department of the Treasury 2005). The Renminbi exchange rate will continue to be a significant part of the bilateral diplomatic agenda between the U.S. and China.

(C) Domino Effect on Other CountriesIf the Renminbi is in fact undervalued, that would create a competitive ad-

vantage for China’s exports. In a global economy desperate for growth, many other countries have begun to consider currency interventions to depreciate the value of their own currencies and help their export economies. In many ways, this resembles the devaluations of fixed currencies under Bretton Woods that led to the eventual collapse of the gold standard in 1971. However, most of the major world currencies are floating, not fixed as they were under Bretton Woods. Even though they are floating, there are other ways that countries can influence the value of their currency. Central banks can take steps to intervene in currency markets to buy foreign currency. This increases demand, and thus


price, of foreign currencies, effectively depreciating the home currency. The perceived undervaluation of the Renminbi and corresponding competitive ad-vantage have contributed to decisions by Japan and South Korea to lower the value of their own currencies.

On September 15, 2010, Japan spent roughly two billion Yen in the larg-est currency intervention in history. It came at a time the Yen was at a 15 year high against the dollar, and the Japanese government was facing intense pres-sures from its domestic producers to take action against the high value of the Yen (Garnham 2010). The intervention had the intended effect of depreciating the Yen and – more importantly – signaling to investors the government’s will-ingness to intervene in currency markets.

In addition, other countries, such as South Korea, have attempted to de-preciate the value of their currency. Even the recent U.S. program of quantita-tive easing has to depreciation of the Dollar. In light of all of these shifts, some countries are left with strong currencies. One such country is Brazil, which ap-preciated about 25% vis-à-vis the U.S. Dollar from January 2009 to September 2010. Brazil’s Finance Minister, Guido Mantega, recognizes the challenges faced by these recent interventions, “We’re in the midst of an international cur-rency war, a general weakening of currency. This threatens us because it takes away our competitiveness” (Wheatley and Garnham 2010).

However, not all interventions are meant to depreciate exchange rates for an advantage in trade. For years, Japan bought U.S. dollars with yen, and accumulated reserves exceeding $893 billion in June of 2007. Japan largely stopped its currency interventions in 2004 (Nanto 2007). Yet while they were actively intervening, the U.S. did not object to these Japanese actions in the same way it objected to China’s intervention in currency markets. This is due to the motivations behind Japan’s actions. An April 2005 Government Ac-countability Office (GAO) report found that although Japan had made signifi-cant interventions in currency markets in 2003 and 2004, “Japan did not meet the criteria [of a currency manipulator] in 2003 and 2004 in part because its exchange rate interventions were considered to be part of a macroeconomic policy to combat deflation” (Government Accountability Office 2005). In ad-dition, the interventions had little lasting effect on the exchange rate (Nanto 2007).

Even so, most currency interventions in today’s economy carry with them the intent to depreciate the value of a country’s currency. China has led the way in preventing the appreciation of its currency, and other countries have followed suit to protect their export industries. The key question going for-ward will be how likely China is to appreciate its currency. The question is political in the way that China makes its decisions, but clearly has important


ramifications on the current account, Sino-U.S. relations and on the likeliness of other countries to devalue their own currency.

V. ConclusionAlmost all the relevant data show an undervaluation of the Renminbi.

The magnitude of the undervaluation is unclear, but some RPPP studies and both APPP studies suggest an inflation-adjusted exchange rate of 4 RMB/USD and below. When compared with other methodology, 4 RMB/USD appears to be a lower bound within the academic estimates of the value of the Renminbi.

First, the relevant RPPP estimates in this paper seem to be from before the 1994 Renminbi devaluation. After that devaluation, Chinese inflation was rel-atively high (possibly itself a result of the devaluation and effect on exports), and it’s unlikely that the exchange rate has been at equilibrium since 1995. Using a base year of 1978 and 1980, the inflation-adjusted exchange rate is around 4RMB/USD. This is significantly lower that the current rate of 6.50, and implies 38% undervaluation. With later base years of 1985 and 1989, the inflation-adjusted exchange rates are 6.63 RMB/USD and 7.05 RMB/USD, much closer to the current value of 6.50 RMB/USD.

The APPP estimates of from the World Bank implied GDP calculation and The Economist Big Mac Index are generally in line with the RPPP esti-mates from 1994 and earlier. The World Bank implied rates show a current value of the Renminbi at 3.80 RMB/USD, with an implied undervaluation of 42%. The Big Mac Index suggests a fair value for the Renminbi even lower than the World Bank, 3.48 RMB/USD, which corresponds to 46% undervalu-ation. These appear to be high estimates of undervaluation when compared with other academic studies.

While it is difficult to say conclusively what a fair value is for the Ren-minbi, it is useful to consider possible thresholds and how long it would take to reach those levels. China maintains control over its currency, and it is unclear what rate China will choose to appreciate the Renminbi, if it allows appre-ciation at all. A relevant benchmark comes from the period of 2005 to 2008, when China allowed steady appreciation of the Renminbi after strong U.S. po-litical pressures. During that period, China averaged 6.4% annual appreciation of the Renminbi. It also did so fairly evenly throughout the year. After 2008, China stopped Renminbi appreciation, and effectively repegged the Renminbi to the U.S. Dollar. In August 2010, however, China resumed appreciation. The figure below shows 6.4% appreciation starting in November 2010.


Figure 7: 6.4% RMB/USD Appreciation

This graph shows how the Renminbi would change assuming China al-lows 6.4% appreciation. In this scenario, it would take until April 2015 (ap-proximately 4 years) to reach 5.0 RMB/USD. The exchange rate would then reach 4.5 RMB/USD in December 2016, which would about 5.5 years. Fi-nally, if China chooses to continue appreciation, the Renminbi will get to 4.0 RMB/USD in October 2018, almost eight years after the start point of appre-ciation. It is impossible to predict the level to which China will allow the Ren-minbi to appreciate and the rate at which it will take place. But if China cedes to international pressure, it may be a long time until the Renminbi reaches a value close to equilibrium.


Appendix




Figure 12: RMB/USD PPI (2005)

Figure 13: RMB/EUR PPI (2000)


Figure 14: RMB/EUR PPI (2005)

Figure 15: RMB/JPY PPI (2000)


Figure 16: RMB/JPY PPI (2005)


References

Beattie, Alan. “Geithner Calls for Action on Renminbi.” Financial Times, 2010 17-September.

Bineau, Yannick. “Renminbi’s misalignment: a meta-analysis.” Economic Systems, 2010 йил 6-February: 259-269.

Cheng, Joseph Y.S., ed. China in the post-Deng era. Hong Kong: Chinese University Press, 1998.

Cline, William R. Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit. Washington, D.C.: Institute for International Economics, 2010.

Cline, William R., and John Williamson. “Estimates of the Equilibrium Exchange Rate of the Renminbi.” Conference on China’s Exchange Rate Policy. Washington, D.C.: Peterson Institute for International Economics, 2007.

—. “Estimates of the Equilibrium Exchange Rate of the Renminbi: Is There a Consensus and, If Not, Why Not?” Conference on China’s Exchange Rate Policy. Washington, D.C.: Peterson Institute for International Economics, 2007.

Cline, William R., and John Williamson. Notes on Equilibrium Exchange Rates. Policy Brief, Washington, D.C.: Peterson Institute of International Economics, 2010.

Department of the Treasury. “Report to Congress on International Economic and Exchange Rate Policies.” 2005.

Dunaway, Steven, and Xiangming Li. “Estimating China’s “Equilibrium” Real Exchange Rate.” IMF Working Paper, 2005.

Elbadawi, Ibrahim A. “Estimating Long-Run Equilibrium Real Exchange Rates.” In Estimating Fundamental Equilibrium Exchange Rates, by John Williamson, 93-132. Washington, D.C.: Institute for International Economics, 1994.

European Commission. European Commission: Trade: China (Bilateral relations). 2011 25-March. http://ec.europa.eu/trade/creating-opportunities/bilateral-relations/countries/china/ (accessed 2011 21-April).

Froot, Kenneth A., Michael Kim, and Kenneth Rogoff. “NBER Working Paper Series.” Law of One Price over 700 Years. May 1995. http://www.economics.harvard.edu/files/faculty/51_NBER1995.pdf (accessed April 24, 2011).

Funke, Michael, and Jörg Rahn. “Just How Undervalued is the Chinese Renminbi.” Asia-Pacific Economic Association Conference. Seattle: Blackwell Publishing, 2006.

Garnham, Peter. “Intervention: The genie has escaped from the bottle.” Financial Times, 2010 27-September .

Goldstein, Morris. “Renminbi Controversies.” Conference on Monetary Institutions and Economic Development. Washington, D.C., 2005.


Goldstein, Morris, and Nicholas R. Lardy. The Future of China’s Exchange Rate Policy. Washington, D.C.: Institute of International Economics, 2009.

Government Accountability Office. Treasury Assessments Have Not Found Currency Manipulation, but Concerns about Exchange Rates Continue. Report to Congressional Committees, Washington, D.C.: Government Accountability Office, 2005.

Krueger, Anne O. Exchange-Rate Determination. Cambridge: Cambridge University Press, 1983.

Kuijs, Louis. China through 2020 - A Macroeconomic Scenario. Beijing, China: World Bank China Office, 2009.

MacDonald, Ronald, and Jerome Stein. Equilibrium Exchange Rates. Boston: Kluwer Academic Publishers, 1999.

Nanto, Dick K. Japan’s Currency Intervention: Policy Issues. CRS Report for Congress, Washington, D.C.: Congressional Research Service, 2007.

The Economist. “Precisely Wrong.” The Economist, 2005 йил 26-July: 76.The U.S.-China Business Council. US-China Trade Statistics and China’s World

Trade Statistics. 2011. http://www.uschina.org/statistics/tradetable.html (accessed 2011 22-April).

Wheatley, Jonathan, and Peter Garnham. “Brazil Warns of ‘Currency War’.” Financial Times, 2010 27-September.

Williamson, John. “Estimates of FEERs.” In Estimating Equilibrium Exchange Rates, by John Williamson, 179. Washington, D.C.: Institute for International Economics, 1994.

World Bank. “Quarterly Update.” Beijing, 2010.

91

A Multi-Criterion Model for Evaluating the Efficiency of Solar Energy Incentives

Christos Makridis1

Arizona State University

Abstract

Regardless of any scientific uncertainty about climate change, unequivocal evidence illustrates the emergence of various renewable energies in the inter-national economy. One promising and prevailing technology is solar energy. Given the heightened demand for renewable energy, the current debate centers on how to foster an environment conducive to investment and efficiency in order to increase total welfare. This article provides a survey of the major incentives that exist in the international community to promote the adoption of solar power, introduces a unique working model that compiles and modi-fies core conclusions in the literature, and provides a methodology to assess the effectiveness of these incentives. This article’s underlying objective is to create a reasonable theoretical model and qualitative list of criteria that can be applied to various incentives, like feed-in-tariffs, in order to determine optimal incentives for use in United States energy policy.

1 Special thanks and appreciation is given to Professors Dr. Nicholai Kuminoff and Dr. Matthew Croucher for their time and energy, both of whom provided helpful and valuable input for the review of the article.


I. IntroductionEconomists have spent a significant amount of time analyzing the poten-

tial economic ramifications of climate change and the effects of various policy proposals to mitigate those costs. However, uncertain climactic models have not precluded rapid renewable energy development. Trends towards sustain-ability have stimulated renewable energy development and growing public support. Moreover, if fossil fuel prices rise in the coming decades and con-cerns over greenhouse gasses (GHG) continue to mount, there will be further pressure on renewable energies to serve as the world’s power source. Because current renewable energy technology (RET) is not yet cost-competitive with fossil fuels, there is an opportunity for the implementation of policies that incentivize renewable energy development and increase economic growth. While the goal for policymakers is to optimize social welfare by creating a climate conducive to supply and demand forces, interest groups will inevi-tably clamor for subsidies to artificially boost their profits and international competitiveness.

Even though solar power, a prevailing renewable energy, is currently not cost competitive with fossil fuels because of storage challenges, higher re-search and development costs, and energy intermittency2, there is still sig-nificant opportunity for gains in profitability. One of the prevailing renewable energies is solar power. Yet, while governmental agencies are clamoring to provide the industry with various types of incentives, there must be a critical discussion of incentive efficiency and cost-effectiveness. The article has three connected purposes: (1) compile the major incentives used for solar power in the United States and abroad; (2) create a model for evaluating the effec-tiveness, or value added, of an incentive; (3) develop central questions and answers regarding the attributes of a successful incentive package for the solar industry.

The research concludes with an application of the developed gener-al quantitative model and qualitative criteria for incentive valuation. Using growth in the share of capital, contributions to welfare, and measured techno-logical innovation, in addition to other criteria such as analyses of forecasted business confidence, efficiency, and unintended consequences, feed-in-tariffs are shown to be an ineffective incentive instrument employed by many na-tional governments. Building on existing research, the article shows that an optimal incentive must enhance the long-term profit maximizing points in the market. Without a sustained commitment to innovation and cost-reduction, incentives merely induce short-term production without real economic gains.

2 This list is not meant to be a comprehensive explanation of why solar is not as expansive as fossil fuels.

93A Multi-Criterion Model for Evaluating the Efficiency of Solar Energy Incentives

Therefore, to promote long-term progress, agencies must implement incen-tives that are structured on the basis of value-enhancing criteria for renewable energy development.

II. A Brief Background on Solar Market InformationThe renewable energy market as a whole is projected to achieve a net

worth of $349.2 billion by 2020. Specifically, solar power, one of the most lucrative sectors positioned for profitability, is expected to achieve $113.6 bil-lion in net worth by 2020, and was estimated to be worth $71.2 billion in 2010 (Pernick et al; 2011). Despite the recent economic contraction, in 2008 to 2009, the solar photovoltaic market value grew by 15% and 45% in GW capacity (EPIA; 2010). Market forecasts by the European Photovoltaic Indus-try Association (EPIA) estimate that solar energy could satisfy roughly 12% of total European Union energy demand by 2020, absent any further external price support mechanisms.

While developing predictions for market expansion, it is important to dif-ferentiate between the two major forecasting models: (1) the business-as-usual scenario where there are no additional external incentives provided, and (2) the policy-driven scenario where additional support mechanisms, such as feed-in-tariffs (FiT) are used more extensively. The EPIA estimates that under scenario (1) solar power will provide approximately 15,000 MW by 2014, whereas un-der scenario (2), solar power is estimated to provide 30,000 MW by 2014, in the European market. While the discrepancy between the policy-driven and business-as-usual scenarios is large, it should not justify the implementation of inefficient and ineffective mechanisms that create perverse incentives for energy development.

These realities demonstrate the importance of providing optimal incen-tives for the solar industry; there is a unique opportunity to increase econom-ic growth, but it requires strategic implementation of sound public policies. However, potential for growth within the market does not alone serve as a justification for providing subsidies for the industry for two reasons. First, incentives that are directed to the solar industry could potentially be more efficiently allocated elsewhere; policymakers will need to persuasively illus-trate that incentives for a given industry and/or companies will be spent com-paratively better than allocation in other areas. Second, there is a real risk for protectionist behavior in a scenario where each country operates under a policy-driven scenario, namely because countries will perceive gains as zero-sum activities. Policymakers must make a conscious effort to avoid the per-ception of artificially inflating the domestic industry at the expense of foreign competition. However, the two aforementioned concerns need not prevent


meaningful implementation of sound economic incentives in the solar sector. If policymakers develop a valid rationale for slowly phasing strategic and cost beneficial incentives for the solar industry, the two concerns can be managed, if not completely obviated.

III. Review of the Literature and Analytic CompilationsThis section will survey current market research on the solar industry,

probe pressing questions facing its future growth, and lay a foundation for analyzing the role of incentives in the alternative energy market.

A. Existing Market Based Incentives in the United States and AbroadThe international community has implemented a wide array of federal

and local incentives over the past decades. In order to accurately assess the costs and benefits of incentives, it is necessary to be aware of those that exist. This allows for empirical testing and more precise forecasting of the effects of various incentives. The subsequent section summarizes the most impor-tant incentives that have been cited in the literature, laying a foundation for the model and qualitative analysis that will be developed (Price, et al; 2010):


The European and United States markets have excelled in the production and distribution of solar photovoltaics. Understanding the basic accomplish-ments achieved by major renewable energy producers provides empirical sup-port that renewable energy can result in profitable investments. This provides a foundation for further analysis to determine the most effective method to increase profitability, since not all incentives are created equal when it comes to increasing total welfare. The following provides an overview of the various accomplishments of key players in the solar industry (EPIA; 2010):


In addition to the advancements seen in the European and United States PV markets, other countries such as China and India have also exhibited in-tense economic growth in their solar energy sectors. Alternative energy subsi-dies, particularly for solar power, tend to be higher in China and India. Howev-er, this is likely due to the structure of their governments, which allow leaders more authoritative decision making power (Ritger and Vidican; 2010; Markets and Markets; 2010). Nonetheless, it is clear that many countries are imple-menting incentives for catalyzing greater adoption of solar energy.

B. Regulatory and Market ApproachesWhile market based approaches have proven to successfully encourage

alternative energy adoption and reduce supplier production costs, it is impor-tant to acknowledge that there are many companies that still do not transition to cleaner fuels because of the high cost of clean energy, even when there are incentives. This is a public policy dilemma, since it has already been estab-lished that fossil fuels create negative externalities, such as carbon dioxide and sulfur oxide emission, which are detrimental to human health and eco-systems (Rivers and Jaccard; 2006). As such, the social costs are not currently reflected in the market. Therefore, the regulatory approach claims that policy intervention must be taken to return social welfare to its optimal level. More-


over, proponents of such an approach claim that costs for alternative energy will decrease because firms will gain experience with producing a technology.

A thorough body of empirical evidence indicates that a firm’s costs will decrease and productivity will increase as experience is gained (Arrow; 1962). For example, the cost of solar energy will decrease for a company that has learned to scale effectively and build and distribute its product efficiently by using absorbent photovoltaics and cost-effective inputs. While learning-by-doing may increase productivity, firms typically do not have incentives to quantify the social benefits that accrue, skewing their perceived estimation of return on investment. Therefore, many firms choose to invest less in renewable energies than what is optimal from a social perspective (Rivers and Jaccard; 2006).

However, the reality that there is a socially optimal level that has not yet been reached does not necessitate regulatory approaches. Rather, it should ac-centuate the value of conservatively targeting effective incentives in the mar-ketplace. In this sense, learning-by-doing is best achieved by creating incen-tives to develop efficiencies that are recognized by the market. Rivers and Jaccard provide a base analysis that is used in the article’s assumption that a market based approach is socially optimal.

In order to avoid unnecessary complications, we accept the results ac-quired by Rivers and Jaccard, which are elaborated on in part A of the Ap-pendix, pertaining to their specific model on heterogeneous firms with imper-fect foresight regarding the socially optimal level and benefits of clean energy technology. The existence of heterogeneity in firms provides further credence to the argument that market based approaches, rather than regulatory, are the most efficient and cost-productive policies. Market based instruments allow firms to allocate emissions reductions between technologies in the most ef-ficient method, as opposed to a one-size-fits-all approach in a command and control policy. Rivers and Jaccard find that a market based approach is almost always more cost-effective than a regulatory instrument. The presence of het-erogeneity only further decreases the benefits of a command and control policy by increasing the average cost for firms. Thus, Rivers and Jaccard provide an excellent introductory analysis, concluding that cost savings increase through the use of market mechanisms.

C. Relevant Qualitative LiteratureThe National Renewable Energy Laboratory created a seminal study, led

by Gouchoe (2002), which provides a broad qualitative analysis of the effec-tiveness of alternative energy incentives. Although there are numerous ways to define effective renewable energy incentives, including, but not limited to, a


reduction in the cost of technology over a given time period, installed capacity, and the amount of electricity produced, the policymaking community needs more precise criterions for evaluation. While Gouchoe does not provide quan-titative guidelines for evaluation, seven notable suggestions are offered from the literature regarding criteria that renewable energy incentives must take into account in order to create economic value:

a) Funding stability and duration: Incentives should create certainty for recipients and be operational for optimal duration. Timeframes that are too short will be ineffective, while those that are too long will be inefficient.

b) Incentive amount: The magnitude of the incentive must be large enough to catalyze investment and result in an eventual correction of the negative externality.

c) Quality assurance: Incentive programs must be designed to monitor performance and provide incentives for the company to use financial stimuli effectively.

d) Application process: Incentives should be easy to apply for and in clude relevant assistance from program administrators.

e) Consumer education and awareness: Marketing is necessary to edu cate the public about renewable energy technologies and existing in centives.

f) Institutional barriers: Foundational issues must not inhibit the effec tiveness of incentive programs by creating unnecessary inefficiency.

g) Complementary financial incentives: Incentives will be most effective when combined; packages of incentives are the most effective way to stimulate wide-reaching demand3.

While some of the aforementioned criteria that characterize an incentive may appear to be intuitive, it is important to acknowledge core contributions to the literature and consensuses that have been achieved. These overriding criteria for policy evaluation of incentive packages are integrated throughout the model and policy discussion in the paper.

D. Introduction of the Basic Cobb-Douglas FunctionBarro, Sala-i-Martin, and Mankiw (1995) created a model that can be

applied to analyze endogenous growth in capital, as it pertains to enhancing firm productivity. They use a Cobb-Douglas production function that uses

3 This does not imply that every recipient must receive a variety of incentives. However, on the aggregate, a mix of incentives should be used.


the physical stock of capital, the quantity of raw labor, and a fixed technology parameter as inputs. Their model is constructed to analyze capital formation across countries and the direction that it flows. It is clear that the concern here is not with capital formation and its direction, but rather the emphasis is on developing a model for predicting growth in capital, as it pertains to contribu-tions to the multiplier effect for energy incentives. As such, the Barro, Sala-i-Martin, and Mankiw model is modified (illustrated in 2.11) so that it may be applied to the discussion of energy economics for incentive valuation.

E. A Base Model for Welfare AnalysisLi and Lofgren (2008) develop a sound model for welfare analysis. It

is well known that policymakers aim to effectively optimize social welfare. However, measuring the effectiveness of public policies can be challenging because the effects of specific policies must be isolated amidst a dynamic and multi-dimensional international economy. A problem with many earlier wel-fare models is that they apply unjustifiable and inaccurate discount rates for future social profits. In this sense, if there is a low risk of environmental harm resulting from climate change and/or pollution, many historical models as-cribe an unrealistically low economic value and therefore a high discount rate. Li and Lofgren change the emphasis from ‘second-best’ constraints to envi-ronmental concerns and the effects of projects’ lifetime consequences. Their approach to welfare analysis is an essential element in the model because of the interaction that welfare contributions have to the multiplier effect for in-centives.

F. The Implications of Technological ChangeClarke, Weyant, and Edmonds (2008) have made a significant contribu-

tion to the literature on the sources of technological change4, namely the role that incentives play in fostering innovation and greater efficiency measures. Technological change is one of the most efficient methods for driving eco-nomic growth, originally concluded by Solow, but specifically analyzed by Clarke et al (2006). Technological innovation in the solar industry can help correct the negative externalities regarding atmospheric carbon dioxide, pollu-tion, and the potential for climate change. This accentuates the opportunity for policies to positively influence the growth of industries and lead to improve-ments in technology and long-run efficiency. For example, when coupled with effective conditions, grants for research and development (R&D) can induce technological change. For policymakers, an important result is the spillover 4 Technological change refers to a given knowledge base, intangible or tangible (physical stock or

intellectual capital). Induced technological change refers to a change in the rate or direction of technological change in response to a particular set of policies, for example, targeted incentives.


that occurs. Extensive studies illustrate that the social rates of return for R&D spillovers are comparatively higher than those for private rates (Griliches; 1992). Therefore, the challenge becomes optimizing incentive packages to create the highest rate of return possible, by providing private firms with nec-essary resources for research and development, product development, or any other relevant project associated with launching renewable energy technology into the market.

This model is developed purposely by Clarke, Weyant, and Edmonds (2006) to remain abstract in order to accentuate the value of influences on tech-nological change, namely indirect spillovers through intra-industry efforts and innovation driven by external incentives. External incentives serve as catalysts for intra-industry innovation. In this sense, fostering conditions for knowledge spillovers drives technological change. The explanation of the Clarke et al (2006) model is explained in more detail in part B of the appendix.

IV. ModelAlthough definitions of incentives vary, the general consensus in eco-

nomic literature is that incentives are discrete offers that elicit an expected response. Incentives are created for the purposes of motivating an individual / entity (or, individuals / entities) to choose an alternative to the status quo (Sweeney; 2004, Grant; 2002). For the purposes of the article, incentives can be defined as financial transfers that promote a change in behavior.

While the definition of an incentive is established, it is more challenging to develop criteria for evaluating its success or failure. There are multiple ways in which to interpret the outcome of an incentive for solar power. These inter-pretations include, but are not limited to: the number of jobs that are created, the economic growth that ensues, and the amount of energy that is generated.

While the above interpretations of success are valid, they lack consoli-dation and precision. To formulate a generalized principle for evaluating the effectiveness of incentive mechanisms for solar power, I develop an equation that recognizes the necessary parameters and variables. The following are the variables that will be used and their corresponding meaning:


5

Infrastructure, mindset, and depreciation are held constant as exogenous variables, whereas the multiplier and total expenditures are endogenous vari-ables. The terms are treated as such to capture changes in value when total ex-penditures and the multiplier are varied. Using comparative statics to measure changes in relevant endogenous variables, the following model is deduced:

With constraints of:

where pT is the price of the incentive (quantity of dollars spent), r is the interest rate, and n is the number of years.5 Objectively expressing these attitudes is challenging because of the subjectivity latent in the qualitative

data. However, we assume that a similar approach will be taken to the explanation in West, et al (2010).

(2.1)

(2.2)

(2.3)

(2.4)

(2.5)


The total value added by an incentive is subject to the constraints of the total incentive expenditure in present value terms, which is the amount that a given policymaker intends to spend on a project. Converting the expenditures into present value terms is important because policymakers are often forecast-ing incentive mechanisms to take place at a particular date in the future – not in the current n0 time period.

The results from (2.3) and (2.4) also illustrate an important consideration, namely that total expenditures for a given incentive will increase the value added up until point δ; following δ, diminishing marginal productivity sets in and the value added will decrease. The purpose of this article is not to ascertain the exact point δ, not only because it is path dependent on the type of incentive and its respective targeted community, but also because it would sacrifice other important findings. The cardinal argument is merely that pure expenditure, without thoughtful formulation in context of the specific energy, will fail to re-sult in the adoption, production, and profitable implementation of solar energy.

One can also use comparative statics to derive an optimization problem and its corresponding first order conditions:

First Order Conditions:

We set condition 2 equal to 3 and arrive at max V: T* = K*(G*, W*, Q*)

We assume that 2 and 3 are maximized following:

Second Order Conditions:

(2.6)

(2.7)

(2.8)

(2.9)

(2.10)


GrowthWe first look to define G, which is part of K(G, W, Q) because of its sig-

nificance in the model. A neoclassical Cobb-Douglas closed-economy growth model is used to simplify the argument (Barro, et al; 1995):

This Cobb-Douglas function uses G, total growth with inputs, A, a fixed technology parameter6, P, international positioning achieved through growth in output and capital, α, an exogenous growth factor for P, K, stock of physical capital, η, an exogenous growth factor for K, L, a given rate that labor and hu-man capital, with growth g and time n, contribute to total growth.Growth is maximized when:

where α > 0, η > 0, α + η < 1.

WelfareLi and Lofgren’s model (2008) is used because it best describes the so-

cial cost and benefit that result from a policy project; many other models that conduct welfare analysis purely focus on changes in tax rates or regulation. Instead, this model allows the policymaker to input different scenarios, such as a tax or regulation, into the model to observe the effects. There are a few adjustments to the Li and Lofgren model, namely that capital goods are not counted since they are included in (2.11). This also allows the use of a more simplified model without employing Hamiltonian dynamics.

W can be defined by amending the Li and Lofgren welfare analysis mod-el. This describes a given incentive’s contribution to welfare (Li and Lofgren; 2008):Let C = (C1, C2...,Cn) be an n-dimensional vector of consumption flow at a given time t, which includes all possible goods and services that implicate so-cial welfare through consumer and/or producer surplus. C also includes long-term horizon goods, such as environmental services and ecosystem functions. Therefore, the prices of these services are rental prices. I use a utilitarian mea-sure of welfare at time t = 0 with the expression:

where U(C) is an arbitrary concave, non-decreasing, instantaneous utility

6 Technology is held constant in this portion of the model for purposes of simplicity.

(2.11)

(2.11)

(2.12)


function with continuous second order derivatives C ≥ 0 and Θ is the utility rate of discount.At each point in time t, consumption C(t) is allocated within the n-dimensional possibility set S(C(t):α) where α is a collection of parameters such that it re-mains strictly convex.

Technological ChangeFinally, we define Q as the production that changes as a direct result of

technological change. Innovation is a driving force behind efficiency and prof-it-maximization. As such, it is important to consider its composition. Consider the following modification to Clarke (2008):

where Q is an n-dimensional vector of technologies, like solar cells, knowl-edge pools, and human capital, which all exist at a given time t. Consider a time, T, that represents a time in the future. T is used to represent the techno-logical change in time periods for vector Q. A utilitarian measure is used to quantify technological change.

where Q(A) is an arbitrary convex exponential function that illustrates the relationship between gains in technology and production

where Q*(A) is the optimal amount of technological change that occurs given the constraints

(2.13)

(2.14)

(2.15)

(2.16)

(2.17)


We use (2.17) to demonstrate a general form for maximum technological productivity. The interaction between technological change and productivity inevitably implicates the effect of the incentive multiplier; an incentive that spurs technological innovation will create knowledge pools and spillovers greater than what previously existed. There are three different activities that drive technological change: intra-industry initiatives, direct spillovers, and in-direct spillovers (Clarke; 2008). Of particular importance are direct spillovers, as well as their contribution to subsequent indirect spillovers. Policies that result in technological change and spillover within an industry will offer not only immediate benefits of a new technology, but also an amplification of the rate at which existing pools of knowledge can be used. As such, the general form in (2.17) illustrates one of the objectives of policymakers designing in-centive packages.

Model RemarksThe above work has clearly failed to leverage the exogenous variables.

This is because focus is given to the factors that are most relevant to policy-makers working on a specific project. Factors, such as I(t) and M(C, H), can-not be changed as easily, as they require top-down and bottom-up structural decisions in federal tax policy and social mindset, respectively. As such, the endogenous variables are highlighted and given context to accentuate the sig-nificance of the value-added model established above. However, a description of t, C, and H is provided to establish a working knowledge for the general factors that influence the acquisition of values for the exogenous variables, in case future analysis on them is pursued.

More importantly, G, W, and Q – which determine K – are scrutinized to develop a broader model for incentive based evaluation. Growth, welfare, and technological productivity are important criteria because they are the funda-mental drivers of incentive effectiveness. To improve upon the criteria estab-lished by Gouchoe (2002), an incentive that enhances international position-ing and capital growth, contributions to aggregate welfare and technological productivity will inevitably increase total energy generation, profitability, and cost-effectiveness. While there are many models that analyze the cost-effec-tiveness of solar energy in depth using a certain region as a case study, few have compiled holistic criteria for incentive valuation. This work serves as a


synthesis of core models to develop a more comprehensive method for evalu-ating policy decisions, both quantitatively and qualitatively.

V. Discussion

A. Solar Energy’s CompetitivenessGenerating a model that attempts to evaluate the effectiveness of pub-

lic policy incentives is a prerequisite to solar energy’s expansion. In order to make solar power more cost competitive, the true social costs and benefits must be reflected in the market, and effective incentives must be applied to the energy industry to resolve the existence of fossil fuels’ negative externalities. Although solar energy is not yet cost competitive with fossil fuels and nuclear power because of storage challenges, higher research and development costs, and energy intermittency, it is making significant strides (Borenstein; 2008). According to some estimates, primary costs for installation including parts and labor have fallen significantly in recent years. However, another primary cost, replacement of inverters, has not been falling as quickly, resulting in a lag for consumer wide adoption of solar energy as a source for electricity.

Because solar power is expensive relative to other energy production methods, the industry needs incentives to innovate if it is to quickly establish itself as a dominant energy source in the market. Incentives must reconcile with the reality that the net present cost of installing solar photovoltaic tech-nology exceeds the net present benefits, assuming current real interest rates and current real costs for electricity (Borenstein; 2008). At its most basic level, the question concerns how to make the marginal benefits outweigh the mar-ginal costs for solar adoption.

In the past, the majority of U.S. solar development has been almost en-tirely focused on generating electricity from rooftops or ground-mounted pho-tovoltaics on residential and commercial communities. However, there has been a recent spike in interest for utility scale plants, implementing newer technologies and efficiency standards (Nimmons; 2008).

Technological advancements have impacted the levelized cost of energy (LCOE7) for photovoltaic panels and are fostering increasing adoption of util-ity scale solar power. For the past decade, the cost of PV has declined by 2-4% per year on average, with the current price under $4/watt (DeJongh; 2010). Moreover, as more utility scale projects materialize, the investment costs will continue to decline due to cost reductions in equipment components. In this sense, the two effects – reductions in the cost of producing solar photovoltaics

7 The LCOE is accepted as the net present value of the total life cycle costs of the anticipated solar energy project divided by the quantity of energy produced over the system life.


and the cost of equipment in utility scale plants – will interactively enhance each other. Utility companies realize that they will not be able to meet demand solely through the use of residential scale projects because of their inherent limitations, namely that the magnitude of energy remains limited unless ad-opted on a widespread scenario. Without significant incentives for consumer adoption of residential solar power, it remains unlikely that it would expand enough to make a significant impact (DeJongh; 2010).

While studies support the use of both residential and distribution based solar energy, a vast majority of economic literature indicates that, all else equal, scaling and utility based solar power will result in the very substantial cost savings. Even though the cost of producing photovoltaics has been decreasing steadily per year, it has not decreased enough to make residential solar power a cost-effective option for a large portion of consumers. Without a technological breakthrough, studies indicate that one of the best options that utilities have to reduce their cost per kilowatt hour is to purchase plots of land and place vast quantities of solar panels to produce electricity. Supporting evidence comes from a Berkeley case study on California’s solar economy. Findings, based on the analysis of two systems (CPUC, California Public Utilities Commission, and CEC, California Energy Commission), illustrate that the average system cost falls substantially for larger systems. The largest systems in the CEC data-set are roughly $2.5/WAC cheaper than 1 kW installations; the largest CPUC-funded systems are roughly $1.5/WAC less expensive than the smaller systems funded by that program (Wiser, et al; 2006).

While utility scale solar power offers concrete benefits and efficiency improvements for an expanding solar market, the emphasis is not purely on comparative residential and utility scale solar projects. Instead, utility and residential scale solar power is used to accentuate the complexity latent in


the solar industry, namely that each renewable energy technology has many subgroups that warrant thoughtful policy and economic analysis. This is an important conclusion because it necessitates that incentives for the solar indus-try be specific and targeted. Because, for example, solar and residential scale projects are unique in their financing and power generational abilities, they require distinct incentives for economic growth. As such, the value of having a predictive model that gauges the expected success of incentives will be pivotal in order to distinguish between profitable investments in various renewable energy technologies.

B. Surveying Feed-in-Tariffs and their RamificationsThe above discussion has been useful in describing incentives from the

perspective of utility recipients. On the other side is the provider of incentives – local, state, and federal governments. The question of value maximization and cost minimization necessitates a conclusion regarding effective incentive outcomes. One incentive that has received significant attention by the econom-ic and reporting community alike, touted to resemble the aforementioned char-acteristics the most, is the feed-in-tariff (FiT) (Couture and Gagnon; 2010). The majority of feed-in-tariffs guarantee a minimum price per kilowatt hour that an electrical utility has to pay to a private, independent, producer of re-newable power. While FiTs have succeeded in promoting greater use of solar energy in many European countries, they have been criticized because of high costs, inefficiencies, and distortions between regional electrical utilities (Sijm; 2002).

Although the reviewed models provide an effective way for quantifying gains from various incentives, it is important to also qualitatively assess the implications of policy action in application of the already employed quantita-tive techniques. Therefore, a few criteria can be developed to represent central features of an incentive. These criteria include, but are not limited to: business confidence & certainty, effectiveness at achieving the desired goal, efficiency, and extent to which significant unintended consequences result. This section will begin to reconcile core issues developed in studies from the literature.

Business Confidence & CertaintyFeed-in-tariffs were created to catalyze renewable energy production,

with a salient feature of instilling certainty within the market. However, FiTs can be interpreted as being tantamount to regulation because an electrical util-ity is forced to pay an independent producer a set fee for electrical generation. Therefore, often times, feed-in-tariffs can work at odds with a given commu-nity’s economic structure by distorting true supply and demand.


While some proponents of feed-in-tariffs characterize FiTs as stable and predictable incentives for renewable energy generation, an economic consen-sus on the comparative superiority of FiTs has yet to be reached (Ragwitz; 2007, Sijm; 2002). In this sense, accepting the premise that incentives and incentive packages must vary based on the region and/or targeted energy, other incentive mechanisms, such as tax credits and rebates, may provide a more ef-fective catalyst for value creation in the energy sector.

One issue, however, that has not been covered as extensively in the lit-erature is the ability of local and federal governments to implement long-term growth objectives through feed-in-tariffs. Because FiTs only set a price be-tween a buyer and a seller, competition does not constantly monitor the extent to which the price is overvalued or undervalued. Although this can be remedied to a degree with price adjustments by regulatory officials, the long-term total costs can still be unsustainable. Therefore, even though FiTs unequivocally promote short-term energy generation, it is questionable whether long-term welfare can increase enough to justify the costs. Under a Ricardian approach, individuals will perceive feed-in-tariffs as short-term stimuli. However, in the long term, feed-in-tariffs and their regulatory counterparts do not induce companies to pursue new and more efficient profit maximizing points. This is empirically illustrated; feed-in-tariffs have been ineffective at lowering the levelized cost of electricity for renewable energy, specifically wind and solar energies, because they offer little long-term incentives to innovate given that they guarantee a long-run price (del Rio and Gaul; 2006). Nonetheless, there is room for FiTs to be more effective if their pricing schemes can be modified, not only to fit the economic environment, but also to attach decisive and opti-mal timeframes for prices to avoid long-term market distortion.

Effectiveness & EfficiencyFeed-in-tariffs have effectively created greater renewable energy produc-

tion. However, this is no surprise because of the majority of FiTs’ regulatory nature, which involves a set price below the market level for independent producers. As such, the question shifts to quantifying the costs of renewable energy production and whether the benefits of increased RET development outweigh the costs of subsidizing RET.

Analyzed by the Energy Research Center between 1994 and 1998, the minimum bid prices for contracts between Britain’s auction model and Germa-ny & Spain’s feed-in-tariffs fell 40% and 1%, respectively. While FiTs resulted in stable electrical rates, Britain’s auction market exhibited significant relative success at lowering energy prices. This small endeavor in turn helped create a sustainable system for long-term investment in RET.


Feed-in-tariffs are commended for their simplistic administrative nature because they typically require little supervision due to an upfront establish-ment of a long-term price scheme (Del Rio and Gaul; 2006). While this is true in certain respects, a regulatory approach creates indirect inefficiencies in the market. For example, by stifling innovation and charging a price that oper-ates externally from the market8, inefficiency arises. There are methods for circumventing this type of inefficiency, for example, phasing feed-in-tariffs out of a given economy over time (Sijm; 2002). However, it is important to recognize that many studies do not evaluate how realistic these remedies are, namely that once regulatory and/or full subsidies are instituted they are rarely eliminated.

Another flaw in the feed-in-tariff model is that they do not permit utility-scale solar power. Because regulatory agencies create pricing schemes that induce solar power usage primarily in residential communities, there are no incentives for scaling. Residents are unable to act as uniform bodies that pool resources to invest in larger solar production locations. In this sense, residents typically produce small amounts of energy by placing solar panels on rooftops. For these reasons, FiTs have generally been ineffective at achieving the ulti-mate goal: catalyzing innovation for the purposes of lower energy costs and gains in efficiency.

Unintended Consequences Feed-in-tariffs have resulted in many material unintended consequences.

Because FiTs are, by definition, at odds with competitive electrical pricing, they inevitably heighten the costs of electricity production by artificially creat-ing a demand that would otherwise not exist. In addition, the incentive to pro-duce at lower cost and innovate is not as strong because a set fee is determined by an external regulatory authority. Although there are methods to mitigate moral hazard and enhance the drive for innovation, FiTs are generally at a comparative disadvantage when paired against other market based incentive mechanisms.

Analyzing the effect that feed-in-tariffs had on Germany is particularly useful in understanding whether models predicting FiT success accurately represent reality. Many German utilities and consumers now oppose feed-in-tariffs because of the high costs that resulted to support renewable energy pro-ducers. For example, in 2000, a new wind turbine project would be paid 11 cents per kilowatt hour for the first five years with a falling rate thereafter. The Renewable Energy Law (EEG) that was instituted created a buyback tariff rate

8 Although every incentive alters the market, positively or negatively, certain incentives allow for more market-based interaction, such as tradable permit schemes and tax reductions for minimum purchase.


that increased 5% annually. Increased costs reached a point that resulted in net disutility and consumer dissatisfaction (Lesser and Su; 2007). This demon-strates that constant modification on FiTs is not always a feasible option, even if it is theoretically possible.

Denmark is another example of a country that has relied heavily on feed-in-tariffs, since their initial use in 1992. Utilities were obligated to purchase renewable energy from private producers at a fixed price between 70 and 85 percent of the market price for the given electricity (non-fossil fuels). Al-though there was a sharp increase in renewable energy production – which was expected – Denmark abandoned its guaranteed pricing in 2000, due to simi-lar concerns about the high cost of supporting renewable energy producers. Departing from FiTs, Denmark began to institute tradable green certificates (TGC) to create a market for renewable energy (Lesser and Su; 2007). While there is inconclusive evidence as to whether TGCs create maximum incentives for market based competition, Denmark realized that feed-in-tariffs had satu-rated their market and failed to generate as much success for their renewable energy industry as they initially anticipated. It must be noted that the challenge for Denmark was not electrical production, but the lack of both reduced kilo-watt/hour costs and long-term efficiency gains.

In addition to imposing hidden costs onto utilities and the general popu-lation, national feed-in-tariffs are eligible only to domestic generators (Sijm; 2002). Unless FiTs are structured to be received by all relevant entities – both domestic and foreign – within a given location, the incentive heightens trade protectionism. The National Foreign Trade Council has set guidelines that environmental trade legislation must follow to be compatible with interna-tional law; targeting FiTs to specific domestic producers would violate trade law (Syunkova; 2007). Although the purpose of this article is not to quantify the impact that such a violation would have, it is important to note the various unintended consequences that could possible result, such as disruptions to in-ternational trade. The negative ramifications of undermining free trade would reverberate throughout the international economy.

Conclusions on Feed-in-TariffsThe mathematical model and qualitative analysis demonstrates that in-

centives must encourage the maximum utilization of the free market, rather than implementing an inherently inefficient regulatory framework. Couture and Gagnon (2010) provide an explanation detailing the distinction between market- independent and dependent tariff policies. While market independent FiTs establish a fixed payment rate for a set duration, market dependent FiTs have more flexible price structures capable of adapting to demand and reflect-


ing environmental and social attributes. This versatile approach offers signifi-cant benefits over market independent feed-in-tariffs, primarily the gains in efficiency for utilities. Under a fixed price scheme, utilities are forced to price energy demand throughout the day at a fixed rate, which has resulted in lower marginal cost generators being scaled back (Langniss, et al; 2009, Couture Gagnon; 2010). Tamas, Shrestha, and Zhou (2010) also provide substantial evidence for this argument. Fixed price feed-in-tariff schemes are tantamount to government subsidies because they create an artificially low price to induce renewable energy development. Social welfare under this type of policy is lower than it is for tradable green certificates because the latter creates a mar-ket that compensates electrical producers for transitioning to renewable energy technologies. These findings are aligned with the argument espoused in the literature and model reviews, namely that approaches to engage markets are comparatively more efficient than regulatory schemes.

Although this analysis simplifies the issue at hand to some extent, it is meant to compile a breadth of information and accentuate the value of assess-ing incentives based on total benefits versus total costs – not using a single criterion for evaluation. While the discussed arguments are not intended to be fully comprehensive, they demonstrate some of the shortcomings that inhibit market independent feed-in-tariff approaches in comparison to other incentive mechanisms.

VI. ConclusionSolar power is being subsidized in most developed countries. However, in

order to grow and avoid becoming dependent on subsidies, the solar industry needs to proactively search for ways to use its received incentives more effi-ciently. Ultimately, solar power should become a cost-competitive technology that will displace conventional energies.

Economic incentive policy is equally as important; policymakers that pro-vide poorly constructed incentives will heighten the tendency for solar firms to misallocate cash flow and invest in inefficient projects. Therefore, it is es-sential to develop criteria for evaluating the effectiveness of a given incentive, incorporating central aspects of various incentive valuation interpretations. Without objective and uniform valuation methods, policymakers will merely perpetuate perverse incentives in the renewable energy sector.

The mathematical model and qualitative criteria that have been discussed thus far introduce useful mechanisms for valuing incentives and their expected success. Seminal literature in the field of environmental economics has been surveyed and digested in this article to probe key questions and develop intro-ductory conclusions. One such conclusion is that market based approaches are


most likely encourage optimal and sustainable success in the renewable energy market. While regulatory and short-term incentive mechanisms, like feed-in-tariffs, offer compelling short-term benefits for renewable energy adoption, policymakers must acknowledge that the criteria for success encompasses more than just electrical production. Other vital objectives for renewable en-ergy must also include converging upon and eventually surpassing the com-petitiveness of conventional energies in terms of cost-effectiveness, efficiency, and equity. Only by becoming profitable in the long-term will renewable en-ergy technologies, especially solar power, become popular among consumers and utilities alike.

Additional analysis that acquires data for the model’s endogenous vari-ables will enable critical insights to be reached. In particular, estimates con-cerning incentives’ contribution to the physical stock of capital, contributions to welfare, and technological change ought to be acquired. While developing accurate estimates of these values may prove challenging, I suggest that the wind energy production tax credit be analyzed as a base case for comparison, given that it serves as a strong model for an incentive’s relative success at ex-panding renewable energy development. Based on existing economic theory and models, I contend that tax credits and a version of tradable permits should first be evaluated because they are the incentives that empirically capitalize upon market mechanisms the most. Upon confirmation of their effectiveness, additional research ought to be conducted to determine the most efficient method of implementation, namely whether the federal government or state and local governments are expected to achieve the most efficient outcome. While subsequent data must be subject to quantitative econometric techniques, it is also important to develop conclusions in context of the qualitative analysis established in this article, including Gouchoe (2002).

The solar energy sector has unprecedented potential for growth and in-novation. However, this sector’s ability to enhance total welfare is a function of both policymaker incentive construction and recipient incentive use. Deter-mining and implementing optimal incentives for the solar industry is a chal-lenge that can be overcome with careful analysis.


Appendix

A. Rivers and Jaccard on Firm HeterogeneityThe model has three periods: past (period 0), present (period 1), and future (period 2). No discounting takes place within periods, but costs are discounted from period 2 to period 1 by discount rate δ. Firms are also assumed to be electricity generators with three available technologies:

• A standard technology: a relatively cheap technology that emits a high amount of pollution. This is referred to, in the Rivers and Jac card model, as subscript b.

• An alternative technology: this technology is relatively more expensive, but emits less pollution than baseline technology b. Natural gas generation is an example. This technology is referred to as subscript f.

• A clean energy technology: this technology is much more expensive than the baseline technology, but emits very little pollution. This technology is referred to as subscript r.

The alternative and clean energy technologies experience decreasing costs, explained via the learning-by-doing assumption (Arrow; 1962):

Where cj,t is the cost of producing energy from the first unit of technology j installed in period t, Nj,t is the total cumulative production of energy from technology j up to but not including period t, and bj is a parameter defining how fast the cost of producing energy from technology j falls as cumulative production increases.

Without policy action that increases the relative costs of producing or forces a regulatory action, firms will use the cheapest technology, the baseline option. The marginal cost of pollution abatement, a, is measured in dollars per ton of pollution abated, calculated for clean energy and alternative technolo-gies:

(A1)


Where cj,t is the marginal cost of technology j in period t in $/kWh and ej and cr are increasing and concave values. The supply curve for technology j in period t is:

Where cj,t is the cost in $/kWh for the first unit of energy produced by technol-ogy j in period t, kj is a constant that reflects the steepness of the supply curve for technology j, denoting how fast per unit costs increase in a given period. If (A3) and (A2) are substituted for each other, one will find the marginal emis-sions abatement cost curve for each technology in time period t.

B. Clarke et al and Technological ChangeOf particular importance in the Clarke model are the factors underlying tech-nological change. They create a given production function for industry i as follows:

Where qi is output in industry I, Ai is the sum total of technology parameters for industry i, and xi is the sum total of inputs for industry i, including, but not limited to capital and labor; f(z) represents an activity, or set of activities, taken by actors within the model to help induce technological change (intra-industry efforts). Yet, technological change can also arise from external inducement, or incentives, through direct spillover. This is manifested in (B3) where tech-nological change occurs from potential policy interventionists. This is repre-sented with –i to refer to outside industry activities. The preceding equations

(A2)

(A3)

(B1)

(B2)

(B3)

(B4)


are used to speculate the existence of (B4) which refers to the interaction of outside industry activities and their impact on technology parameters driven by intra-industry activities (Clarke, et al; 2006).


References

Arrow, K., The economic implications of learning-by-doing,” Review of Economic Studies 29 (1962): 155-173.

Barro, R., Mankiw, G, & Sala-Martin, X, “Capital Mobility in Neoclassical Models of Growth,” American Economic Review 85:1 (1995): 103-115.

Borenstein, S, “The Market Value and Cost of Solar Photovoltaic Electricity Production,” The University of California Energy Institute (2008): http://www.ucei.berkeley.edu/PDF/csemwp176.pdf

Chuan-Zhong, L., & Lofgren, K.-G, “Evaluating Projects in a Dynamic Economy: Some New Envelope Results,” German Economic Review 9:1 (2008): 1-16.

Pernick, Ron., Wilder, Clint., Winnie, Trevor., Sosnovec, Sean., “Clean Energy Trends 2011,” (2011):http://www.cleanedge.com/reports/pdf/Trends2011.pdf

Couture, T., & Cory, K, “State Clean Energy Policies Analysis (SCEPA) Project: An Analysis of Renewable Energy Feed-in Tariffs in the United States,” The National Renewable Energy Laboratory (2009): http://www1.eere.energy.gov/wip/pdfs/tap_webinar_20091028_45551.pdf

Couture, T., Gagnon, Y, “An analysis of feed-in tariff remuneration models: Implications for renewable energy investment,” Energy Policy 28:2 (2010): 955–965.

Clarke, L., Weyant, J., Edmonds, J, “On the sources of technological change: What do the models assume?” Energy Economics 30:2 (2006): 409-424.

DeJongh, T, “Financing Utility-Scale Solar Projects in the United States,” (2010):http://taylor-dejongh.com/wp-content/uploads/2010/07/Financing-Utility-Scale-

Solar-in-the-US.pdfFrondel, M., Ritter, N., & Schmidt, C, “Germany’s solar cell promotion: Dark clouds

on the horizon,” Energy Policy 36:11 (2008): 4198-4204. “Global market outlook for photovoltaics until 2014,” The European Photovoltaic

Industry Association, EPIA (2010): http://www.epia.org/fileadmin/EPIA_docs/public/ Global_Market_Outlook_for_Photovoltaics_until_2014.pdf

“Global Renewable Energy Market Worth $614.92 Billion by 2015,” Marketsandmarkets (2010): http://www.marketsandmarkets.com/PressReleases/renewable-energy-market-worth-$615-billion-by-2015.asp

Gouchoe, S., Everette,, V., & Haynes, E, “Case studies on the effectiveness of state financial incentives for renewable energy,” The National Renewable Energy Laboratory (2002): http://www.nrel.gov/docs/fy02osti/ 32819.pdf

Grant, R, “Ethics in Human Subjects Research: Do Incentives Matter?” Journal of Medicine and Philosophy 29:6 (2004): 717-738.


Griliches, Z, “The search for R&D spillovers,” Scandinavian Journal of Economics 94:29-47 (1992).

Klein, A.; Pfluger, B.; Held, A.; Ragwitz, M.; Resch, G.; Faber, T, “Evaluation of Different Feed-in Tariff Design Options - Best Practice Paper for the International Feed-in Cooperation: 2nd edition.” Karlsruhe, Germany, The Fraunhofer Institute Systems and Innovations Research (2008): http://www.feed-in-cooperation.org/wDefault_7/content/research/research.php (main)

Langniss, O., Diekmann, J., Lehr, U, “Advanced mechanisms for the promotion of renewable energy: models for the future evolution of the German Renewable Energy Act,” Energy Policy 37 (2009):1289–1297.

Lesser, J., & Su, X, “Design of an economically efficient feed-in tariff structure for renewable energy development,” Energy Policy 36:3 (2007): 981-990.

Nimmons, J, “Utility Solar Business Models: Emerging Utility Strategies & Innovation,”

The Solar Electric Power Association (2008): http://www.solarelectricpower.org/media/84333/sepa%20usbm%201.pdfPrice, S., & Margolis, R, “2008 Solar Technologies Market Report. Retrieved from

U.S,” Department of Energy and Efficiency (2010): http://www1.eere.energy.gov/solar/pdfs/46025.pdf

del Rıo, P., & Gual, M, “An integrated assessment of the feed-in tariff system in Spain,” Energy Policy 35:2 (2006): 994-1012.

Rigter, J., Vidican, G, “Cost and optimal feed-in tariff for small scale photovoltaic systems in China,” Energy Policy 38:11 (2010): 6989-7000.

Rivers, N., Jaccard, M, “Choice of environmental policy in the presence of learning-by-doing,” Energy Economics 28:2 (2006): 223-242.

Sijm, J. P, “The Performance of Feed-in Tariffs to Promote Renewable Electricity in European Countries,” The Energy Research Center (2002): http://www.ecn.nl/docs/library/report/2002/c02083.pdf

Sweeney, M, “The Challenge of Business Incentives for State Policymakers: A Practitioner’s Perspective,” (2004): http://www.mccallumsweeney.com/uploads/ ARTICLE-23-10%20%20Challenge%20of%20Business%20Incentives%20for%20State%20Policymakers%20-%20Spectrum%20-%2001-04.pdf

Syunkova, A, “WTO – Compatibility of Four Categories of U.S. Climate Change Policy,” The National Foreign Trade Council (2007): http://www.nftc.org/default/Trade%20Policy/Climate_Change/Climate%20Change%20P aper.pdf

Tamas, M., Shrestha, S.O., Zhou, H, “Feed-in tariff and tradable green certificate in oligopoly,” Energy Policy 38:8 (2010): 4040-4047.


West, J., Bailey, I., Winter, M, “Renewable energy policy and public perceptions of renewable energy: A cultural theory approach,” Energy Policy 38:10 (2010): 5739-5748.

Wiser, R, “Letting the Sun Shine on Solar Costs: An Empirical Investigation of Photovoltaic Cost Trends in California,” National Renewable Berkeley Lawrence Laboratory (2006): http://eetd.lbl.gov/ea/ems/reports/59282.pdf

121

Real Option Analysis of a Large-scale Space Solar Power Venture

Yael Gilboa and Xiaoyu Guo1

College of William and Mary

Abstract

This study values a large-scale Space Solar Power (SSP) venture by using Real Option Analysis (ROA), an advanced financial technique that takes manage-rial flexibility into account. We develop a model to represent the entire project as a series of decisions in a binomial tree. We calibrate the tree using data from energy markets as well as Monte Carlo simulations of the distribution of project value under expected conditions. In the end, specific sources of flexibility are formulated as Real Options within the binomial tree and the cal-culated value of these options is added to the Net Present Value (NPV) of the project. Unlike previous studies that argue SSP is economically unfeasible, we conclude that large-scale SSP is a viable business venture as long as the project is implemented in stages and real options are exercised optimally. Previous studies on the economic feasibility of SSP have used only static Discounted Cash Flow (DCF) models, which assume that a project will be carried out until completion under any circumstances. Our study, on the other hand, incorpo-rates the flexibility to enter or abandon the project in response to energy prices, launch costs, and other technological parameters that affect the viability of the project. Modern decision analysis tools demonstrate that such flexibility adds tremendous value to a venture facing large uncertainties, as is the case with SSP.

1 We would like to acknowledge the guidance of Vladimir Atanasov, Professor of Finance at The College of William and Mary, in his capacity as thesis adviser and his insights and assistance during this process. We are grateful to Leonard Weinstein, DSc (NASA Langley Research Center) for his advice and/or assistance in the direct or indirect provision of data for this study.


I. IntroductionSpace Solar Power (SSP) is a technology that collects solar power on

satellites in space and wirelessly transfers this power for use on Earth. The initial idea, known as “Satellite-Solar Power System (SSPS)” was first de-scribed in November 1968. At the time, however, SSP faced major technologi-cal challenges that prevented it from becoming a feasible source of energy. In 1973, Peter Glaser carried the concept of SSP closer to reality by inventing the method of transmitting power over long distances using microwaves (Glaser, 1973). He proposed the idea of transmitting solar power from a large antenna in orbit to a rectifying antenna on Earth (referred to as a “rectenna”), which would convert the waves to electricity.

Compared to other renewable energies, which are available only inter-mittently (i.e. hydro and wind energies), SSP is continuously available and reliable. A solar power satellite deployed into geostationary orbit can produce energy at peak power 99% of the time because it is directly lit by the Sun. The average daily energy reaching Earth’s surface is only one tenth of that in orbit. In addition, ground solar power requires a large amount of land. The recten-nas used for SSP are 90% transparent and can be built over farmland, water or strategic remote locations using conventional construction methods. Crops and farm animals may be raised underneath a rectenna, as the thin wire used for support only slightly reduces the sunlight.

Besides the advantages mentioned above, SSP is one of the few options that will meet the power demand of the fast-growing world economy. The 2007 National Security Space Office (NSSO, now called the DoD Executive Agent for Space) report states “A single kilometer-wide band of Geosynchro-nous Earth Orbit experiences enough solar flux in one year to nearly equal the amount of energy contained within all known recoverable conventional oil reserves on Earth today.” (NSSO, 2007). Adapt from the benefit of collecting huge amount of energy in a short period of time, SSP may also be the answer to the long-term energy generation and climate control issues facing humanity. The report summarizes the attractive features of SSP relative to conventional energy sources and argues that SSP is the only option that is clean, safe, and reliable and can be used for base-load power, the amount of power required to meet minimum demands based on reasonable expectations of customer re-quirements. As a byproduct of SSP, carbon-neutral synthetic fuel could ad-ditionally be captured through three major steps (further explored in Section IV.C Using SSP for Synthetic Fuel) to help reduce the carbon footprint of the world economy.

Our study values a large-scale SSP venture by using Real Option Analy-

123Real Option Analysis of a Large-scale Space Solar Power Venture

sis (ROA), an advanced financial technique that takes flexibility into account. We extend the work of Atanasov and Lenard (2010), who apply ROA to a small-scale SSP system designed for high-end users, and focus on large-scale gigawatt-power SSP systems. We develop a model to represent the entire proj-ect as a series of decisions in a binomial tree, which is a graphical representa-tion of the possible intrinsic values that an option may take at different nodes or time periods, and calibrate the tree using data from financial, energy, and other markets. In the end, specific sources of flexibility are formulated as real options within the binomial tree and the calculated value of these options is added to the Net Present Value (NPV) of the project.

After evaluating the numerous real options, we conclude that large-scale SSP is a viable business venture as long as the project is implemented in stages and real options are exercised optimally. Furthermore, we propose potential opportunities related to SSP ventures such as the valuation of obtaining con-struction materials from the Moon or the satellites of Mars as a non-conven-tional alternative to launching all components from Earth. The energy savings of obtaining material from extraterrestrial sources are compared with the esti-mated costs of constructing automated facilities to harvest this material. Our analysis calculates the critical scale of SSP systems, above which harvesting material from extraterrestrial sources dominates the traditional launch-from-Earth model. We also suggest possible future research using large-scale SSP to manufacture carbon-neutral synthetic fuels, which will undoubtedly lead to a reduction of our carbon footprint.

The remainder of the paper is structured as follows. The next section de-scribes the background of SSP analysis, including previous economic analysis based on DCF models, followed by a background description of ROA and our research objectives. Section III first follows the traditional approach by calculating NPV using a DCF model, then adds back the hidden values of real options discovered from ROA to the project value. Section IV provides sev-eral thoughts on the financing aspect of such a venture, and explores the possi-bilities of reducing the large overhead costs by adopting alternative launching methods to reduce launch costs and harvesting synthetic fuel as a byproduct to claim more returns from such a venture. Section V concludes our paper. All details about our key assumptions are presented in the Appendix.

II. Background

II.A Previous Economic Analyses of SSPSince the 1970s, a number of studies have analyzed the economic fea-

sibility of SSP business ventures (Mankins, 1997; Macaulay et al, 2000; Xin


et al, 2009). The consensus among these studies has been pointing at the large uncertainty about revenues and costs, the current lack of technology that makes SSP more expensive than current energy sources. Risk factors in these analyses give SSP an unattractive profile. However, there have been several major issues with previous economic analysis of SSP ventures.

These studies have used only traditional financial tools to value the proj-ects, primarily static Discounted Cash Flow (DCF) models. Such valuation methodologies create problems because they assume that a project will be car-ried out until completion and excludes the uncertainties that may occur during the implementation (Charania & Olds, 2000). For example, a five-year project could involve the deployment of ten satellites overall, but not all ten of them have to be deployed at once. If, at the end of stage one, the first satellite en-countered problems or more investment in R&D is required, a company has the flexibility to choose whether or not to continue with the project at that mo-ment. DCF analysis does not capture such flexibility in valuing the feasibility of the project.

Due to the nature of the SSP, most people who value space business ven-tures have advanced technical knowledge, but use antiquated valuation meth-ods. Most previous studies use static DCF methods to value the projects. The most basic of these is NPV (Net Present Value), which discounts future cash flows to the present. A positive NPV means that a project is worth pursu-ing. Similarly, IRR (Internal Rate of Return) looks at the future cash flows of a given project and determines the internal rate of return that is necessary for the project cash flows to break even. Unsurprisingly, most NPV and IRR valuations of SSP have resulted in largely negative valuations. Considering the large upfront costs of investing in such a project and the fact that potential payoffs would occur in the distant future, the risk profile of SSP in an all-or-nothing way would certainly be unattractive.

Modern decision analysis tools tell us that the flexibility to expand or abandon adds tremendous value to a venture, especially when it depends on large uncertainties, as in the case with space business ventures.

II.B Real Options Analysis Real Option Analysis (ROA) is a technique that was originally developed

for financial options and values the flexibility of having a right, rather than an obligation, to make a certain decision. ROA allows for the flexibility of making decisions in the future and potentially abandoning, expanding or wait-ing on a project (Joseph, 2005). If a project is not as successful as initially expected, the project can be abandoned, or production capacity can be reduced or suspended. In this case, the loss of capital is limited to initial investments


and any incurred costs in the initial stages of the project (Copeland & Anti-karov, 2003). Alternatively, if a project seems more promising, then growth strategies can be implemented to expand the venture or enter into new markets. Additionally, certain decisions that can be put off until uncertainty is resolved in the future can add value to a project. These potential values are not included in a static DCF model.

Real Option Analysis has particular applications in industries where there is a great deal of uncertainty. For example, the success of a pharmaceutical company’s R&D project can highly depend on a number of factors that appear at various stages of the project. Although there may be a number of unknowns at the beginning of a process, using ROA, one can model reasonable ranges for the uncertainties and determine whether, all things considered, a project should be undertaken at anytime during the project. ROA also pervades in the energy sector, particularly oil and gas exploration. Since the profitability of an oil or gas venture is entirely dependent on the highly volatile prices of oil and gas in the market, DCF would be an unsatisfactory valuation method due to the difficulty in forecasting future cash flows. It also ignores the ability of compa-nies to defer exploration or development decisions to the future (Lund, 1999). Though given the similar risk characteristics in oil, gas sector and SSP venture, only a small number of studies have used ROA in the analysis of space appli-cations. The most prominent study of this type is Weck, Neufville & Chaize (2004), which analyzes the staged deployment of communication satellites.

Classic ROA models decisions to respond to a single source of uncer-tainty. When two or more sources of uncertainty are relevant to the project, the real options are called Rainbow Options (Kodukula & Papudesu, 2006). Rainbow Options are widely used to value natural resource deposits, which are usually exposed to two uncertainties – price and quantity (Smith & McCardle, 1999).

The two major sources of uncertainty that we take into consideration in this particular research project are 1) the present value of future revenues, which depends on future energy prices, and 2) the total cost of installation, which includes system cost, assembly expenses and launch expenses. We choose a five-years horizon for the model. This is approximately the time peri-od after which power solar will become commercially available. For example, Solaren Corporation signed a contract, allowing PG&E to deliver energy from a new solar power project (“Project”), which is expected to complete in 2016 (PE&G, 2009). A longer horizon for the model will only increase the value of the option, so five years is a rather conservative time frame. In Year 5–the final period of our model – each terminal node of the tree for the SSP venture is evaluated and the project is only launched if the difference between the present


value of revenues and total system costs is positive.

II.C Research ObjectivesOur research overcomes several shortcomings of previous economic

analyses. We develop a Real Options Analysis of a large-scale SSP system that explicitly accounts for flexibility to respond to resolution of uncertainty in the future. This is the first study to analyze a large- scale SSP venture for base-load power using ROA. We are also the first to analyze multiple sources of uncertainty in a large-scale SSP venture using the Rainbow Option frame-work. In addition, we provide a preliminary analysis of non-conventional al-ternatives, such as retrieving mass from the moon or the satellites of Mars as construction materials for the satellite.

III. MethodologyWe use five basic steps in developing our ROA model. First, we fol-

low the traditional approach and calculate the Net Present Value (NPV) of the project under no flexibility and baseline values of all parameters. Next, we use Crystal Ball to run Monte Carlo simulations on the variation in baseline NPV when each critical parameter is varied according to a given distribution. We run 20,000 trials of the Monte Carlo simulation in order to best represent the extent of various uncertainty situations. Although there are other methods available, we follow the Cox, Ross, and Rubinstein’s option-pricing model (1979) because it is the most well established in dealing with binomial trees. The Binomial Options Pricing Model requires specific factors to determine the Up and Down movements, which are calculated by using an underlying volatility. For our Real Options Analysis, we use the standard deviation of the NPV from the simulations to build the binomial tree model. Step four involves using the binomial tree model to evaluate the option of expanding or abandoning the project in Year 5. Our final step involves transforming the simple option into a more complex Rainbow Option by taking into account an-other source of uncertainty—the total project cost. The nesting of the steps in the model allows us to directly compare the value of the project as calculated from three alternative approaches– 1) The classic NPV; 2) ROA with a single source of uncertainty; and 3) ROA of a Rainbow Option based on two sources of uncertainty.

III.A Baseline NPV Analysis We first research the various inputs that affect costs and revenues for the

project and build a model with the most logical assumptions for the inputs. Table I shows inputs representing the most likely or current scenarios that are


built into the model to come up with an initial NPV. The variables highlighted in yellow (launch cost from GEO per kg; specific power density kg/kilowatt installed power; price per watt installed power; assembly cost per kg; price per kilowatt-hour; annual growth in price for electricity) are those that we later simulate in our NPV model. The annual growth in price for electricity follows a normal distribution, while we assume the other variables follow uniform distribution. For the purpose of this paper, it is important that we simulate the variables with large ranges of data to compensate for the lack of significant research on the proper ranges for the inputs. Research on most of the inputs cited inconsistent ranges that would not permit the use of normal distributions; thus, we used uniform distributions to conservatively account for the equally likely instances in each of the ranges. We did not simulate certain inputs that were found to be relatively standard in valuations of this nature, such as An-nual Loss of System Output, etc., because of the consensus on these figures in recent literature. For example, a combined efficiency of 65% was found by taking the product of the efficiencies of electricity transfer over various stages from space to the rectennas, a 1% annual loss of system output, a useful life of 25 years, and a Balance of System (BOS) multiple of 2. The BOS multiple measures the combined weight of the system relative to the power generation system. In other words, the system mass (in kg) equals the power output mul-tiplied by the power density and the BOS multiple of 2.

We set the cost of capital used to discount all future cash flow to 10%. Cash flows are projected for the useful life of 25 years, with annual revenues equal to the product of power output, 24 hours/day, 365 days/year, (1 - the an-nual loss in the system of 1%), the price per kilowatt-hour (which, due to the variance, we model different ranges for), (1 + the growth in price of electric-ity) to the power of (time), percent of time active and the combined efficiency.

The output in Table II shows the breakdown of the total initial expenses, as well as the present value of future cash flows and the NPV. Under the base-line methodology, the NPV for this project is -$2.26 billion.

III.B Monte Carlo Analysis of the Variation in NPVIn order to come up with a more robust valuation, we manipulate the vari-

ables mentioned above and highlighted in the table in order to obtain a range of values for NPV. The most important output from our Monte Carlo simula-tions is the annualized volatility of NPV. In order to simulate a realistic range of possibilities for the NPV scenario, each of the variable inputs to the model is set to a distribution, primarily uniform, lognormal or normal. These ranges are then used as inputs in a Monte Carlo simulation, which varies the NPV cal-culation according to the different cash flows. We run 20,000 simulations and


extract two parameters of the resulting distributions of NPV – the annualized standard deviation of the Present Value of all revenues, which approximately equals 30%, and the annualized standard deviation of expected total costs of the system, which approximately equals 20%. These two standard deviations are expected to best estimate the variations in both present value of revenues and expected total costs from all possible scenarios. They are used in the cali-bration of the ROA models below.

III.C Binomial Option Valuation Model – Simple Option to EnterBased on the annualized volatility from the NPV distribution above, we

build a binomial tree. The stock price of $8.7 billion is the present value of all the future cash flows generated by selling the energy, which is primarily deter-mined by price per kilowatt-hour and is varied in the sensitivity analysis. The exercise price of about $11 billion reflects all of the current costs enumerated above, including launch costs, assembly expenses and the system cost, all of which are primarily dependent on the system mass. The interest rate is set ap-proximately to the risk-free rate, which is 4%. We use an annualized volatility of 30%, which corresponds to the annualized standard deviation of the present value of revenues from the Monte Carlo simulations as described in II.B. For simplicity, we use a three-periods (t = 0,1,2) model with present time being 0 and each time period equaling 2.5 years, which results to the five-year model.

Two possible directions for the revenue are identified: the Up movement indicates that the revenue of the venture increases while the Down movement simply indicates that the revenue decreases. Based on the simple binomial approach (Cox, Ross & Rubinstein, 1979), the Up movement equals the expo-nent of annualized standard deviation (in our case 30%) times the square root of the length of time period (in our case 2.5). The Down movement equals one over the Up movement. The probability of Up movement pup is calculated by the following equation:

The probability of Down movement equals 1 – pup. The summary of these parameters can be found in Table III.

We determine that, without loss of generality, the binomial tree can be recombining, which means that an Up movement followed by a Down move-ment results in the same output as a Down movement followed by an Up move-ment. The tree is expanded based on the probability of each movement. The option to expand or abandon is a traditional and frequently used option. For a capital-intensive project such as this, one may have to spend $50 billion over


the lifetime of the project, but not all of it today (Tan, 2009). One can wait to see whether, for example, energy prices go up later on so that revenues in-crease, rather than funnel money into a negative NPV project or abandon it at the very beginning. In the case of SSP, there is a minimum threshold of price that makes the option of expanding feasible. This threshold is determined to be a price of 19 cents per kilowatt hour, and the project can be continued or abandoned in five years based on whether or not it meets that threshold. In this case, we decide to make the option for five years from now. The option has value when the present value of revenues, after a series of events, is greater than the total system cost. Figure I illustrates the results of the binomial tree based on 30% volatility. As seen in Figure I, the value of the project, once we account for the option to enter, equals $2.26 billion.

III.D Binomial Option Valuation Model – Rainbow OptionThe ROA model in III.C assumes that total system costs, which serve as

the exercise price for the option in Table III, are constant. This estimate of costs five years from now shows substantial variation based on variations in the parameters in our Monte Carlo simulation. This variation is represented by an annualized standard deviation of approximately 20%. To better capture multiple sources of uncertainty, we decide to incorporate the impact of this variation in total costs into the valuation model. We build a second binomial tree for total costs including system cost, assembly expenses, and launch ex-penses. Assigning Up and Down movements to the costs and using a 20% annualized standard deviation that we got from the Monte Carlo simulation of costs, we calculate a range of costs from $5.8 to 20 billion within two time periods of 2.5 years each.

The combined Rainbow Option tree consists of four branches for the first period and sixteen branches in the second period, incorporating the different combinations from adding the possibilities in the two major variables. These results are presented in Figure II. Compared to the $2.26 billion project value given by the simple binomial tree model, the project value in the Rainbow Op-tion model increases by another $300 Million to $2.56 billion.

III.E Discussion of ResultsWe derive several key conclusions in comparing the project values calcu-

lated from these three different methodologies: 1) -$2.26 billion from NPV; 2) $2.26 billion from the analysis of a simple option to enter; and 3) $2.56 billion from the analysis of a Rainbow Option, which depends on two sources of uncertainty. First, the difference between the value calculated using classic NPV and the project value calculated using Rainbow Option Analysis equals


$4.82 billion, which suggests that the flexibility to make decisions in the future adds a tremendous amount of value to capital-intensive projects, especially in highly uncertain emerging ventures like SSP. Second, the value of flexibility is not recognized without the aid of advanced financial modeling tools such as ROA. Third, by using advanced financial models to appreciate the values of flexibility, corporations will have incentives to structure investment projects in stages or other ways that maximize the flexibility to respond to future uncer-tainty. In our case, the decision to launch the system is done five years in the future rather than today, which allows a company to capture the upside when energy prices go up with limited downside when energy prices go down.

IV. Further Research Opportunities

IV.A Project FundingOur research mainly focuses on improving the effectiveness of the fi-

nancial valuation process of SSP ventures, which will be beneficial in the venture funding process. The ability to abandon the project at early state to eliminate future high investment costs is attractive to investors. Traditional funding methods involve steps like warm referrals from the investor’s trusted sources, or summits that enable investors meet companies face-to-face. Due to the special characteristics of an SSP venture (high up-front investment with huge amount of intangible value hidden under managerial decisions) we think that the funding for SSP requires a distinctive financing strategy (Xin, et al., 2009). Other alternative funding schemes such as emphasizing the environ-mental benefits could be further explored to obtain political supports on initial investment funding. (Jenkins, 2009)

IV.B Extraterrestrial Sources of Building MaterialsIn the case of SSP, launch materials and installation costs play signifi-

cant roles in determining the feasibility and profitability of SSP. Other than the conventional approach we choose to adopt in the scope of this research, there are other non-conventional launching methods that should be considered to reduce the launching cost, namely harvesting materials from the Moon or Phobos (Weinstein, 2003). We are able to broadly compare the conventional launching from Earth and the opportunity to build solar panels by borrow-ing certain materials from the Moon and Mars or launch from Moon or Mars (Mankins, 1997).

During conventional launching, construction materials are carried first to Low Earth Orbit (LEO) by using reusable launching vehicles, and then slowly transferred from LEO to GEO over the course of several years. For this launch


method, a 4 gigawatt capacity station would weigh about 80,000 metric tons assuming a solar panel mass of 20 kg/kilowatt, which means that between 40 and 150 heavy-lift launching vehicles are needed to send materials up to LEO. From there, Ion Propulsion Vehicles-first used as the main engines in NASA’s Deep Space 1 that launched in 1998-will push the arrays up to GEO. This leads to approximately a total launch cost of around $11 billion for low weight panels only.

Alternatively to the conventional launch approaches, by harvesting mate-rials from space, particularly the Moon or Phobos, launch costs are potentially much lower. The use of lunar resources would be significantly cheaper than Earth-based material for a system with as few as thirty systems of 10 gigawatt power each.

Since this concept relies less on human presence in space and more on self-replicating systems on the lunar surface under remote control of workers stationed on Earth, this proposal suffers from the current lack of such auto-mated systems and should be explored in the future. We expect to further valu-ate this venture based on alternative launching methods and with automated systems once the technology becomes promising. On the other hand, potential risks resulting from the unresolved issue over the ownership of the moon or extraterrestrial objects needs to be addressed in the future research.

IV.C Using SSP for Synthetic FuelAnother potential future research topic is harvesting synthetic fuel as a

byproduct of SSP to reduce the carbon footprint resulting from increased trans-portation (Weinstein, 2008). Given the currently expensive synthetic fuel pro-duction process, SSP could play a significant role in obtaining synthetic fuel from a more efficient process. In the first step, CO2 would be converted into carbon monoxide by using the “cerium-oxide-based” system. This two-cham-bered machine would use a rotating cerium-oxide ring and a parabolic mirror that employs solar energy to get oxygen from cerium-oxide and then pump out the oxygen. Then the de-oxygenated ring could be used in the other chamber to generate carbon monoxide. Second, a similar process could be used to con-vert water into hydrogen with the help of solar power. Lastly, mirrors could be used to heat chemical arrays to 400 degrees Celsius to form calcium carbonate by reacting CO2 and calcium oxide. After a few more reactions involving zinc oxide and calcium oxide, CO2 and solar power produce a synthetic fuel called Syngas and zinc oxide (World in 21st Century, 2010). With SSP assistance in synthetic fuel production, we could leverage the potential of Synthetic Fuel and make it more available for future energy consumption.


V. SummaryThis research project used Real Option Analysis to value a prospective

business venture that would build and operate a large-scale SSP system. The entire project was represented as a series of decisions in a binomial tree cali-brated using observable data from relevant markets and Monte Carlo simula-tions. The identification and modeling of market uncertainties by these meth-ods allowed us to develop a more accurate valuation of this business venture. By adopting advanced and flexible financial tools, we believe that it will be-come evident that large-scale SSP is a viable business venture. In addition to monetary uncertainties (i.e. the prices of oil and steel), non-monetary factors such as the global importance of sustainability and strategic energy security also add intangible value to this venture.


References

Atanasov, V., & Gianluigi, B., (2009). “Analysis of Two Space Business Opportunities.” Space-Based Technology and Commercialized Development: Economic Implications and Benefits.

Atanasov, V., & Lenard, R., (2010). “Real Option Analysis on a Privately Funded Space Based Solar Power Venture.” Proceedings of the 61st International Astronautical Congress.

Charania, C., & Olds, R., (October 2000). “A Unified Economics View of Space Solar Power.” 51st International Astronautical Congress. Rio de Janeiro, Brazil. IAF-00-R.1.06

Copeland, T., & Antikarov V., (2003). “Real Options: A Practitioner’s Guide.” 2nd ed. Thomson- Texere.

Cox, John, Stephen Ross, and Mark Rubinstein. (1979). “Option pricing: A simplified approach.” Journal of Financial Economics. 7, 229–263.

de Weck, O., Neufville, R., & Chaize, M., (2004). “Enhancing the Economics of Communications Satellites Via Orbital Reconfigurations and Staged Deployment.” Journal of Aerospace Computing, Information and Communication, v. 1. No.3: 119-136.

Glaser, P., (December, 1973). “Method And Apparatus For Converting Solar Radiation To Electrical Power.” United States Patent 3,781,647.

Joseph, A.D’Urso., (2005). “Valuing Employee Stock Options: A Binomial Approach Using Microsoft Excel.” The CPA Journal.

Kodukula, P., & Papudesu, C., (July 2006). “Project Valuation Using Real Options – A Practitioner’s Guide.” J. Ross Publishing, Incorporated.

Lund, M., (1999). “Real options in offshore oil field development projects. Natural Gas Marketing & Supply, Statoil.

Lyle, J., (December 2009), “Development of Space-Based Solar Power.” InTech.Macauley, et al., (March 2000). “Can Power from Space Compete? The Future of

Electricity: Markets and the Competitive Challenge to Satellite Solar Power.” Discussion Paper 00-16.

http://www.rff.org/documents/rff-dp-00-16.pdfMankins, J., (1997). “A Fresh Look at Space Solar Power: New Architectures,

Concepts and Technologies”. IAF-97-R.2.03. NASA. (1979). “Lunar Resources Utilization For Space Construction.” v.1.Pacific Gas and Electric Company., (April 2009). “Contract for Procurement

of Renewable Energy Resources Resulting from PE&G’s Power Purchase Agreement with Solaren Corporation”


Ruhnka, C., and Young, E., (Spring 1987). “A Venture Capital Model of the Development Process For New Ventures.” Journal of Business Venturing. v. 2, Issue. 2: 167-184

Smith, J., & McCardle, K., (1999). “Options in the Real World: Lessons Learned in Evaluating Oil and Gas Investments.” Operations Research 47, No. 1. 1-15.

Tufano, P., & Copeland, T., (March 2004). “A Real-World Way to Manage Real Option.” Harvard Business Review.

http://hbr.org/2004/03/a-real-world-way-to-manage-real-options/ar/1Tan, B., (2009). “Using Binomial Decision Trees and Real Options Theory to

Evaluate System Dynamics Models of Risky Projects.” System Dynamics Review.

The National Security Space Office (NSSO). (October 10, 2007). “Space-Based Solar Power: An Opportunity for Strategic Security”. Interim Assessment.

http://www.acq.osd.mil/nsso/solar/solar.htmThe World in the 21st Century, (2010). “Synthetic Fuel from CO2 and Solar Energy.” http://dgswilson.wordpress.com/2010/06/21/synthetic-fuel-from-co2-and-solar-

energy/Weinstein, L., (February 11, 2008). The Solution to Energy Crisis and Global

Warming.Weinstein, L., (September 2003). Why Phobos, Deimos and Many Asteroids

Probably Have Significant Amount of H2O Ice.Xin, S., Panier, E., Zund, C., & Gutiérrez Gómez, R., (May 18, 2009). “Financial and

Organizational Analysis for a Space Solar Power System: A Business Plan to Make Space Solar Power a Reality.) Master’s Thesis, Toulouse Business School.


Appendix

Assumptions

Launch cost to GEO per kg: Currently, the most expensive source of uncer-tainty for the model is launching costs. The cheapest options available are in Russia. The cheapest rocket is Dnepr, with an estimated launch price of $15 million ($2,000 US), which, considering the 4,400 kg capacity of the rocket leads to an estimated payload cost of $3,409/kg. As more and more options become readily available, launch costs will decrease, as they are expected to over the next 5-10 years. A major source of value for our Real Options Analy-sis is that the launch costs will be significantly lower in future years than they are today. For our simulation, we used a range of $500-$3,409/kg, which is in line with industry expectations. The static NPV model used a cost of $3,409.

Specific power density kg/kW: Based on a chapter on photovoltaic from Georgia Tech, we chose the lowest cost thin film with a specific power density of 1256 watts per kilogram. In order to get kg/kW, we took 1000/1256 = 0.796. Since this is already feasibly and currently the best efficiency for photovoltaic, we set the range at 0.5 to 0.79, since technology should help specific power density to decrease in the future. The static NPV model used 0.7961

Price per Watt of installed power: Previous studies have shown this could be anywhere from as high as $20 to as low as $1. The distribution for this was therefore uniform, because there was no reason to believe otherwise. For the NPV models, this was $10.

Assembly costs per kg: Assembly costs per kg is a uniform distribution with parameters of $50-$1,000. In the static NPV, this was set at $500.

Price per kWh: Currently, the break-even price per kWh that would make the NPV equal to zero is 19 cents. The price per kWh is currently around 10 cents. A number of states, however, have issued legislation that a certain percentage of their power come from renewable energy source including SSP, and have shown that they are willing to pay a steep premium, up to 30 cents in some states, for the renewable energy. Because this is the primary determinant for revenue in the future, a number of different scenarios were run with various ranges in order to get a range of NPV. The static NPV model used 15 cents to get an NPV of negative $2.26 billion. If the price were set at 30 cents, the NPV would be positive $6.44 billion.


Average growth in price for electricity: This caused some issues for us. In our preliminary research, we found that the majority of base load power is currently produced by coal, while the majority of peak load power is produced by natural gas. The ten-year average volatility of coal, based on the returns of the front-month futures contract QZ1 was 25.28%. This volatility represents a basis for off-peak volatility of power prices. The twenty-year average volatil-ity of natural gas, based on the ticker NG1, was found to be 57.43%, which represents a basis for the on-peak volatility of power prices. While the average volatility for power prices could be averaged to about 40%, this is far too high of a volatility to take as an aggregate amount. Since we use a single number for our average growth in the price for electricity, rather than a different randomly generated number for each year, the standard deviation of the average growth should be significantly less. This is based on some general principals of the standard error of the geometric mean.


Table IAssumptions for Initial NPV CalculationThe variables launch cost from GEO per kg, specific power density kg/kilo-watt installed power, price per watt installed power, assembly cost per kg, price per kilowatt-hour (kWh) and annual growth in price for electricity are those that were later modeled in the simulations of NPV. The inputs that were not simulated were those found to be relatively standard in valuations of this nature, including a combined efficiency of 65%, which is the product of the efficiencies of electricity transfer over various stages from space to the recten-nas, a 1% annual loss of system output, a useful life of 25 years, and a Balance of System multiple (BOS) of 2. BOS measures the combined weight of the system relative to the power generation system.


Table II Initial Expense Breakdown and Simple NPV

Cash flows were projected for the useful life of 25 years, with annual revenues equal to the output * 24 * 365 * (1-the annual loss in the system of 1%) * the price per kilowatt-hour (which, due to the variance, we modeled different ranges for) * (1 + the growth in price of electricity)^the year * percent of time active * combined efficiency. We set the cost of capital used to discount all future cash flow to 10%. The output below shows the breakdown of the total initial expenses, as well as the present value of future cash flows and the NPV.


Table IIIParameters for Simple Binomial Tree

The stock price of $8.7 billion is the present value of all the future cash flows generated by selling the energy, whose main input is price per watt of installed power, which was varied in the sensitivity analysis. The exercise price of about $11 billion includes all the current costs enumerated above, including launch costs, assembly expenses and the system cost, all primarily dependent on the system mass. The interest rate was set at approximately the risk free rate, at 4%, and the average volatility of 20% was used. The summary of the result was presented in the table below.


Table IV Rainbow Options Assumptions and Inputs

The basic assumptions are based on the results we derived from the earlier DCF models. Such as baseline NPV of $8.7 billion, annualized volatilities, and risk-free rate, etc,. For the parameters used in calculation that monitors the future projection, factors along with probabilities are assumed as we did for the simple binomial models. Based upon the assumptions mentioned in the tables above, the binomial tree was built around the total costs of the project and the NPV. All the assumptions and NPV are listed below.


Figure IBinomial Valuation Lattice of Simple Option to Enter

The figure presents the results from the evaluation of an option to launch the project that depends on one source of uncertainty – the present value of future revenues, with an annualized volatility of 30%. Each time period equals 2.5 years. All parameters required for calibrating the tree are presented in Table III. The terminal payoffs of the option in Period 2 equal Max(PV(FCF) – Total Costs, 0).


Figure IIBinomial Valuation Lattice of Rainbow Option

The figure presents the results from the evaluation of a Rainbow Option to launch the project that depends on two sources of uncertainty: 1) the present value of future revenues, with an annualized volatility of 30%, and 2) the total system costs, with an annualized volatility of 20%. Each time period equals 2.5 years. All parameters required for calibrating the tree are presented in Ta-ble IV. The terminal payoffs of the option in Period 2 equal Max(PV(FCF) – Total Costs, 0). The 16 terminal nodes correspond to the 16 combinations of four possible movements for PV(FCF) – UpUp, UpDown, DownUp, and DownDown, and four possible movements for Total Costs – UpUp, UpDown, DownUp, and DownDown.

volume 4: issue 2 april 2011 - the michigan journal …michiganjb.org/issues/42/text42.pdfkush patel...

Documents