spring 2017 senior distinction papers-class of 2013 · bitcoin. nearly ten years since the first...

1

Omicron Delta Epsilon International Honor Society in Economics

Beta Chapter: St. Olaf College

Executive Board 2013-2014 President: Kelly Tomera and Nick Evens

Vice President: Erik Springer ODE Journal Executive Editor: Rebecca Gobel

ODE Journal Associate Editor: William Lutterman

About Omicron Delta Epsilon

Omicron Delta Epsilon is one of the world’s largest academic honor societies. The objectives of Omicron Delta Epsilon are recognition of scholastic

attainment and the honoring of outstanding achievements in economics, the establishment of closer ties between students and faculty in economics within colleges and universities, the publication of its official journal, The American

Economist, and the sponsoring of panels at professional meetings as well as the Irving Fisher and Frank W. Taussig competitions.

Currently, Omicron Delta Epsilon has 578 chapters located in the United

States, Canada, Australia, the United Kingdom, Mexico, Puerto Rico, South Africa, Egypt, France, and the United Arab Emirates. With such a broad international base, chapter activities vary widely, ranging from invited

speakers, group discussions, dinners, and meetings, to special projects such as review sessions and tutoring for students in economics. Omicron Delta Epsilon plays a prominent role in the annual Honors Day celebrations at

many colleges and universities.

Senior Distinction Papers-Class of 2013 Spring 2017

2

Spring 2017

Omicron Delta Epsilon

International Honor Society in Economics Beta Chapter: St. Olaf College

Executive Board 2016-2017

President: Matthew Lasnier Vice President: Spencer Knack

ODE Journal Executive Editor: Kelsey Myers ODE Journal Associate Editor: Erik Davidson-Schwartz

About Omicron Delta Epsilon

Omicron Delta Epsilon is one of the world’s largest academic honor societies.

The objectives of Omicron Delta Epsilon are recognition of scholastic attainment and the honoring of outstanding achievements in economics, the

establishment of closer ties between students and faculty in economics within colleges and universities, the publication of its official journal, The American

Economist, and the sponsoring of panels at professional meetings as well as the Irving Fisher and Frank W. Taussig competitions.

Currently, Omicron Delta Epsilon has 578 chapters located in the United

States, Canada, Australia, the United Kingdom, Mexico, Puerto Rico, South Africa, Egypt, France, and the United Arab Emirates. With such a broad international base, chapter activities vary widely, ranging from invited

speakers, group discussions, dinners, and meetings, to special projects such as review sessions and tutoring for students in economics. Omicron Delta Epsilon plays a prominent role in the annual Honors Day celebrations at

many colleges and universities.

St. Olaf College’s Beta Chapter of Omicron Delta Epsilon aims to build a bridge between the economics faculty and students, actively providing input and

assistance as needed to improve departmental events; they also publish an in-house economics journal, encouraging, reviewing, selecting, and publishing

original work from economics students at the college.

3

The St. Olaf College Economics Department’s

Omicron Delta Epsilon Journal of Economic Research

__________________________________________________________________ Contents Spring 2017 __________________________________________________________________ Nicholas Golberg: Economic Analysis of Cryptocurrency.……............4 Rebecca Kunau: On Race and Envy: Racial Segregation and Voter Support for Redistributive Referenda …………………………..…...…22 Aaron F. Miller: Quality of Care in Hospitals after Vertical Integration with an HMO………………………………...…………………………44 Robert J. Leet: American Manufacturing: How Important is it to the U.S. Economy? ……………………………………………………….…..…63

4

“Economic Analysis of Cryptocurrency” by Nicholas Golberg

Economic Analysis of Cryptocurrency Nicholas Golberg Abstract: This paper is an analysis of cryptocurrency as a disruptor to traditional national fiat currencies. The revolution in money for the digital age has roots spawning from the financial crisis with the release of the first cryptocurrency Bitcoin. Nearly ten years since the first conception of Bitcoin and its many spin offs, what does the future hold for the movement? Will mainstream success become a reality, or will excessive and dire weaknesses prevent full implementation? Using the financial crisis as context and the Free Banking Era as a past example, this study aims to expand current literature by exploring cryptocurrency through alternative viewpoints. Hopefully, this analysis can provide guidance for future cryptocurrency economists. Regulators will continue to foster a mixed relationship with cryptocurrency because it poses a natural threat to centralized authority, but this system may be the sole savior of digital innovation.

I. Economic Analysis of Cryptocurrency

New technologies are paving the way for transformational change in

the global economy. Advancements in network computing and encryption are

supporting a transition in marketplace exchanges throughout the economy. A

relatively new development is the birth of cryptocurrencies. A cryptocurrency

is a form of virtual currency that facilitates peer-to-peer exchange without the

need for the conventional method involving central authorities and

clearinghouses. The benefits of cryptocurrencies are considerable in breadth

and depth, including greater speed and efficiency of transactions. The

underlying engine that allows cryptocurrencies to function is the distributed

ledger system that keeps track of all transactions in a large network using

separate but connected nodes instead of one central authority. While there are

5

St. Olaf College’s Omicron Delta Epsilon Journal of Economic Research

valuable benefits to cryptocurrency, there are also some drawbacks. There is

considerable risk for money laundering, terrorism funding, and tax fraud. In a

macro view, monetary policy may be impacted by the market share of

cryptocurrency in the long run, which may lead to higher risks of financial

instability and systemic risk.

This new technology has not gone unnoticed by authorities and

regulators. Though cryptocurrencies are still in an early stage of

development, regulators are beginning to devise their responses.

Cryptocurrencies pose a natural challenge to regulators, as their

decentralized, anonymous, and borderless nature makes them inherently

difficult to regulate. These qualities stem from the currency’s unique ability

to share cash-like properties. Regulators across the globe have responded in a

multitude of ways, including clarifying existing monetary laws to apply to

cryptocurrency, banning financial institutions from operations involving

cryptocurrency, and even totally prohibiting citizens from using all virtual

currency. This is a preliminary reaction by nations around the globe, and

developments in policy will need to address the potential challenges of

widespread circulation of cryptocurrency.

But what is currency? After all, it has been a long time since

economists have needed to seriously consider a new system. The three main

functions of money are: to provide a unit of account for such things as

agreeing on a fair price that two parties can understand, to serve as a means

6


of exchange so that two parties need not barter goods, and to store economic

value, allowing for savings and delayed consumption. These three main

functions are the groundwork for the present-day monetary system that

society relies upon. Money has continually evolved throughout human history

to support the accumulation of greater economic wealth. The evolution of

money continues into the digital age, with new innovations that are

challenging economists to question: what is qualified as money?

The internet created an opportunity for a new wave of innovation,

allowing economic activities to take place outside the physical world. Digital

currency is any form of money that is stored and transferred electronically,

such as through Paypal or online transactions; a virtual currency is a digital

representation of value, not issued by a central bank, which can sometimes be

used as an alternative to money, like Amazon Coin. A cryptocurrency is even

more granular and is a digital and virtual currency specifically protected by

cryptography and free from a centralized authority. Cryptocurrency is the

idea of exchanging value without the approval of an institution (Dibrova

2017). Cryptocurrency is not a foreign currency, not a tradeable commodity,

and not a payments network. The most popular and successful cryptocurrency

in existence is Bitcoin.

This paper analyzes the future of cryptocurrency and the probability

of its mainstream success in substituting for traditional fiat currency supplied

by central banks. Historically, financial crises have often served as the

7


catalyst for governments to institute change in financial systems, and the

most recent recession is no different. In addition, using an analogy of the Free

Banking Era in US history provides historical examples of a fragmented

currency market with many competing currencies that are not under a

centralized authority. The primary hypothesis of this paper is that

cryptocurrencies will fail as an alternative to conventional currency unless a

central authority provides some level of support through ensuring legal

protections, regulation in the marketplace, and establishing acceptance as

legal tender, because these development will not organically occur.

II. Literature Review

Current literature discussing cryptocurrency is enthusiastic about its

possible social benefits. The development and forward progress of the

internet has increased shared responsibility and social accountability through

the decentralization of the multicurrency monetary system, which will

eventually lead to structural changes in society. Kleineberg 2016 predicts that

technologies used to develop Bitcoin will create a paradigm shift that could

resolve many of the problems of modern society. Some of the more notable

benefits deriving from the original white paper by Satoshi Nakamoto 2009

are: the advent of micropayments to allow for efficient payments at fractions

of what is economically viable with conventional currency, an increase in

global access to commerce that will benefit international trade, and a

8


reduction in fraudulent activity like chargebacks. These assets are designed to

ultimately give users complete control over their money.

Academic literature discussing the topic also highlights the potential

weaknesses of cryptocurrency and the threats that change in the monetary

system can create. Boel 2016 analyzes the potential implications of financial

innovations for monetary policy and central banks. This study points out that

if society transitioned to no longer using cash currency in favor of a solely

digital system, central banks could theoretically set negative interest rates as

an added tool for stimulating the economy. This is because it is impossible to

charge interest on cash. Halaburda 2016 also examines the potential pitfalls

of cryptocurrencies, such as the waste of resources required to “mine” new

Bitcoins and update the Blockchain. The security of the underlying system is

threatened by a potential for a super miner or super pool of miners to accrue

too much power in the Bitcoin system. This arms race would give too much

power to the best computers. A traditional monetarist view of money gives

theoretical evidence of deflationary pressures, because Bitcoin is capped at 21

million. Kubat 2015 examines cryptocurrencies through one of the main

functions of money: its ability to store value. Bitcoin’s historic volatility

relative to gold and other asset types makes it a relatively poor replacement

for traditional currency in this regard. Bitcoin is widely called money, but

Kubat 2015 concludes it does not fit the criteria. Likewise, loss of useable

Bitcoin due to lost passwords will increase over time and hurt the payment

9


system. There are doubts by academics and professionals as to whether the

new technology will encourage investment in assets such as shares of a

company to increase production in the economy. An economy is measured by

its production and growth, and any threat to this production is a serious and

universal concern.

Given the strengths and weaknesses of cryptocurrency, the current

literature has positive views for its eventual mainstream success. Boel 2016

argues cryptocurrencies will survive as a technology only if the transaction

costs are sufficiently low. Weber 2014 analyzes Bitcoin’s efficacy in

achieving its proposed goals, as well as the public’s will in regards to this

issue according to various citizens and accountable representatives. Prineas

2016 believes that once corporations and consumers assemble a critical mass,

the leading cryptocurrencies will have refined their products, addressing any

shortcomings and creating more opportunities. These new transaction

technologies will evolve in parallel with the online and offline world, and

businesses and governments will adopt both to suit their needs.

In contrast to these positive outlooks provided by current literature,

negative predictions persist that certain flaws will undermine the ability of

cryptocurrency to achieve mainstream success. There are countless possible

ways that Bitcoin could fail: “The technology fails. Over speculation causes

an irrecoverable crash. The price never settles down. Deflationary pressure

annihilates liquidity. A government shuts it down. A new currency makes it

10


obsolete. Users abandon it for some other reason” (Rose 2015). Boel 2016

theorizes that the acceptance and use of a cryptocurrency, which is

intrinsically worthless and not backed by any authoritative body, will only

proliferate in economic conditions of high inflation and volatility. Rose 2015

concludes Bitcoin may fail because the odds are objectively against Bitcoin’s

long-term success. This is because it is extremely difficult to develop an

entirely new concept that depends upon popular acceptance and use for its

success. It is even more difficult when the new concept is a payment and

monetary system. Kubat 2015 proposes that the main competitive advantages

of Bitcoin may not even remain relevant, as traditional banks and payments

systems catch up in terms of speed, efficiency, and guarantees.

III. Financial Crisis as a Catalyst

The financial crisis of 2007 spurred a wave of mistrust and doubts

about the power dynamics of the current financial system. The economic

downfall also inspired experts to fix these conundrums. As part of the

repercussions of the crisis, technology became targeted as a possible solution

to the problem. The first and most storied study concerning technological

solutions is the white paper released by Satoshi Nakamoto titled “Bitcoin: A

Peer-to-Peer Electronic Cash System” in November 2008. The idea sparked a

wave of cryptocurrencies that offer an alternative to the central bank model of

the monetary system. This new form of currency is not under the control of a

11


centralized authority and gives power to the entire user base. Weber 2014

analyzes the legitimacy crisis of current money and payment systems in

reaction to the recent financial crisis and assesses the merits of Bitcoin as a

disruptor to traditional established currencies. Weber 2014 believes a severe

economic crisis, like the Great Recession of 2007, unsettles established

compromises and practices concerning the governance of money and

payment systems. It is under these circumstances that cryptocurrencies like

Bitcoin can sway the public to take interest in a completely new practice.

It has been almost eight years since the creation of the first

cryptocurrency, Bitcoin. Where is Bitcoin now in its development? An infant,

an adolescent, or somewhere in between? Necessary infrastructure and

systems are in place to legitimize the practical use of Bitcoin, but lack of

mainstream support has left Bitcoin much like a preteen who is able to begin

taking on some responsibility in the household, but not yet allowed access to

the keys for the family car. Trust in the Bitcoin payment system is imperative,

and ironically lack of trust in the prevailing monetary system is part of the

reason why it was created in the first place.

Given the advantages and disadvantages of popular cryptocurrencies,

primarily Bitcoin, cryptocurrencies will need to be clearly legislated and

included in the incumbent monetary system in order to survive. Upon further

analysis, various issues persist that cause a lack of trust in Bitcoin and

therefore lagging adoption of this cryptocurrency system.

12


First, what impacts the prominence and utility of a system?

Connection systems like Facebook, telephones, and Craig’s List rely on the

economic principle of network effects. In a network system, increased

membership universally enhances member benefits. When only a few people

belong to the network, its reach and effect is greatly diminished. Once a

network is established, the winner-take-all effect also begins to come into

play. If one person leaves the network, there are no equally comparable

alternatives, because the owner of the network has the most utility to offer,

despite any shortcomings. The greater the size of the Bitcoin network, the

more respect and trust it will popularly receive.

In addition to the network effect, there is a profound lack of clear

legislation and supervision of cryptocurrencies, which reinforces the popular

belief that these systems are primarily exploited by the black market for

money laundering. Without legislation formally acknowledging

cryptocurrencies, possible negative impacts of this disruptive innovation are

doubled by reducing any social protections. What happens when a theft

occurs? Can someone rightfully claim their Bitcoin wallet as valuables in a

police report? The now infamous Mt. Gox Bitcoin exchange hacker attack

and Silk Road marketplace shutdown are examples of these types of flaws in

security and transparency of cryptocurrencies. Mckinney 2013 analyzes the

concept that a digital currency must first be considered intangible personal

property similar to trademarks, copyrights, and patents. Without this

13


recognition, legal protections are not ensured, and consumer confidence is

lacking. Legal protections could decrease market volatility by reducing the

risk of asset losses. Second, ownership disputes must be subject to a judicial

system, such as a judicial proceeding or binding arbitration, to resolve

property conflicts. While personal property recognition is an important

barrier, without the means to resolve ownership rights, the risk of loss is not

reduced. Overall, digital currencies must be subject to similar regulation as

other financial instruments used in facilitating exchanges.

Furthermore, the trust in traditional banking institutions lost in the

financial crisis has not transformed into increased faith in cryptocurrency.

Banks brought the economy to its knees through moral hazard and misaligned

incentives, but society thus far has appeared to prefer the devil they know. It

may take another crisis to fully push cryptocurrency from innovators and

early adopters to early majority and late majority users. As shown in Figure 1,

the life cycle of Bitcoin is at a low acceptance rate and is still not transitioned

from an early adopter to an early majority stage.

14


IV. Free Banking Era as a Case Study

Different currencies have coexisted within the same country in the

past. For instance, during the early stages of American history, there were

virtually no barriers of entry into the banking industry. Banks could even

issue their own currency. In the first half of the 19th century, these methods

changed with President Andrew Jackson. President Jackson was a man of the

common people and rebelled against corporate oligarchs corrupting

establishment powers in government. In a watershed decision for the

American financial system, he declined to renew the charter of the Second

Bank of the United States in 1832. This event initiated what is now known as

the Free Banking Era. Five years later, the states enacted free banking laws

that allowed free entry into the banking industry. These laws granted banks

the ability to issue notes on the condition that designated securities backed

the notes. Such securities were placed on deposit with state regulators. During

15


this era, almost anyone could issue paper money. States, municipalities,

private banks, railroad and construction companies, stores, restaurants,

churches and individuals printed an estimated 8,000 different types of money

by1860. If an issuer went bankrupt then the note would become valueless. To

start a free bank, the owners would typically sell subscriptions or shares of

stock in the bank, and then use the proceeds to buy eligible government

bonds to deposit with the state authority. If the bonds were approved, the state

authorities would allow the bank to start issuing banknotes.

With so many different bank notes in circulation, it was important to

determine which currencies were valid and trusted, and which were risky

before a transaction occurred. Newspapers were the primary source of

updated information about the validity of various banknotes and the going

market exchange rates. Bank notes that were strong would trade at or near

their face value, while weaker notes traded at a discount due to the risk of the

issuer. The Free Banking Era notes were almost the same as the fiat currency

issued by the US government today, except for two significant differences.

Bank notes circulated without a central authority, and there was no deposit

insurance to back the notes. Despite these risky qualities, banks notes

functioned as cash for the economy during this period.

Cryptocurrencies are parallel to these Free Banking Era notes,

because both currencies do not have intrinsic value, but still function as a

medium of exchange. If consumers and merchants did not trust that these

16


currencies could be exchanged as legal tender, then their economic functions

would break down. If there are many types of currencies in circulation, like

during the Free Banking Era and within today’s cryptocurrency market, then

merchants and consumers will be more skeptical of and less willing to accept

lesser known currencies. Analysis of the Free Banking Era suggests that

privately issued currency can circulate at face value, while successfully

serving as a medium of exchange, when there is adequate regulation to assure

the public that the currency is backed sufficiently. If the public loses

confidence, the currencies will cease to circulate or lose value. During the

Free Banking Era, it was clear to regulators that numerous bank failures and

steep reductions in the value of their associated private currencies resulted

from collateral restrictions. These limitations left banks helpless against the

swings in state finances, which in turn caused counterproductive financial

volatility (Sanches 2016). Altogether, it is apparent that a balance is

necessary between unhemmed innovative growth and authoritative support.

The Free Banking Era ended in 1863, because the US government

needed to fund the Civil War and stabilize the financial system. The era was

plagued with many problems, including inflation, as a result of so many

separate entities printing bills. Bank runs were a common fear. Additionally,

a delicate balance persisted between the banks’ rush to conserve specie and

the note holders’ push to convert banknotes to coin. In the end, the Free

Banking Era concluded because the Union needed a uniform way to pay for

17


military supplies, and the power of currency was restored to the Federal

government. The Federal government ordered so called “greenbacks” to be

printed. As these bills became the most widely circulated currency in the

country, the notes from the Free Bank Era naturally phased out. However,

this period did encourage significant economic growth by making capital

more available, enabling transactions to be more efficient, and ensuring

formal intermediation between borrowers and savers.

V. Conclusion

Cryptocurrency has the potential to become an influential player in

the global monetary system. Given the complex nature of this system, it is

difficult to accurately and precisely predict the nature of cryptocurrency’s

eventual role in our financial and economic landscape. Currently, it appears

that cryptocurrencies will not achieve mainstream adoption unless another

watermark event, like the recent global financial crisis, encourages broader

acceptance and trust in cryptocurrencies. After thorough analysis, this study

has identified underlying deficiencies of cryptocurrencies, primarily Bitcoin,

that discredit the attractiveness and functionality of these currencies as true

money. Among the many definitive functions of money, cryptocurrencies will

have the greatest difficult in fulfilling the objective of serving as a store of

value. This is because money that is a reliable store of value is usually backed

by a government or some central authority, or else it has inherent value, such

18


as gold or silver. If a currency has neither backing nor intrinsic value, people

will not trust it over time.

Society is now approaching a tipping point, where the world’s

governments must decide to either suppress or support the development of

cryptocurrency. Ideally governments will act before a critical mass in

transactions and usership accumulates, at which point regulators will be

pressured to arbitrate this system to prevent a potential financial crisis or stem

illegal activities funded through the exploitation of these systems. In the long

run, it is uncertain whether cryptocurrencies will generally beneficial to

society. America’s Free Banking Era demonstrates that simultaneously

issuing hundreds of different private currencies can foster economic growth,

but that ultimately national governments rely on the power of centralized

authorities to control the monetary system. Will cryptocurrency systems,

which similarly issue many types of currencies at once, suffer the same fate

as the banknotes of the Free Banking Era? Most likely not, but governments

can learn from the past and create legislation that amplifies the benefits and

the mitigates the risks of this revolution in money.

19


VI. References Boel, Paola. "Thinking About the Future of Money and Potential Implications

for Central Banks". Sveriges Riksbank Economic Review 1 (2016): 147-157. Web. 8 Mar. 2017.

Dibrova, Alina. "Virtual Currency: New Step in Monetary Development". Procedia - Social and Behavioral Sciences 229 (2016): 42-49. Web. 8 Mar. 2017.

Kleineberg, Kaj-Kolja and Dirk Helbing. "A “Social Bitcoin” Could Sustain a Democratic Digital World". The European Physical Journal Special Topics 225.17-18 (2016): 32313241. Web. 8 Mar. 2017.

Kubát, Max. "Virtual Currency Bitcoin in The Scope of Money Definition and Store of Value". Procedia Economics and Finance 30 (2015): 409-416. Web. 8 Mar. 2017.

McKinney, R. E., Shao, L. P., Shao, D. H. & Rosenlieb, D. C. The reality of digital currency as a financial medium of exchange. Journal of International Finance Studies (2013): 13(3), 45-50. Web. 8 Mar. 2017.

Nakamoto, S. 2009. Bitcoin: A Peer-to-Peer Electronic Cash System, http://bitcoin.org/bitcoin.pdf

Rose, Chris. "The Evolution of Digital Currencies: Bitcoin, A Cryptocurrency Causing a Monetary Revolution". International Business & Economics Research Journal (IBER) 14.4 (2015): 617. Web. 15 Mar. 2017.

Sauer, Beate. "Virtual Currencies, The Money Market, and Monetary Policy". International Advances in Economic Research 22.2 (2016): 117-130. Web. 8 Mar. 2017.

Sanches, Daniel. "The Free Banking Era: A Lesson for Today?". Economic Insights 1.3 (2016): 9-14. Web. 7 May 2017.

Taran, Ekaterina Mikhailovna et al. "Features of Using the Cryptocurrency". Asian Social Science 11.14 (2015): 330-336. Web. 8 Mar. 2017.

Weber, Beat. "Bitcoin and the Legitimacy Crisis of Money". Cambridge Journal of Economics 40.1 (2014): 17-41. Web. 15 Mar. 2017.

Wolfson, Shael. "Bitcoin: The Early Market." Journal of Business & Economics Research (Online) 13.4 (2015): 201. Web.15 March 2017.

20


On Race and Envy: Racial Segregation and Voter

Support for Redistributive Referenda Rebecca Kunau

I. Introduction They came to feel that they were being overtaken. This sense that other countries were catching up contributed to the rise of Thatcherism and Reaganism. The catchup was largely a mechanical consequence, but the people of Britain and the United States nevertheless found it hard to accept. The wealth hierarchy is not just about money; it is also a matter of honor and moral values.

–Thomas Piketty1

Even though the Occupy Wall Street protests of 2011 exposed

America’s discomfort with growth in executive compensation and other

markers of dramatic income inequality, there has been no meaningful

increase in demand for redistributive policies. Most data easily demonstrates

that significantly more families live in poverty than in wealth in the United

States, and the ratio has only become more unbalanced in the last few

decades.2 Therefore, if pure fiscal self-interest was the only variable on which

voters based their choices, and if everyone who could vote chose to do so, the

growing population living at or below the poverty line would bombard U.S.

politicians with demands for increasingly redistributive political platforms.3

Low-income voters would vote for redistributive policies and high income

1 Edited for length. Thomas Piketty and L. J Ganser, Capital In The Twenty-First Century, 1st ed. (London, England: Belknap Harvard, 2014). 509. 2 US. Bureau of the Census. 2016. “Income Gini Ratio of Families by Race of Householder, All Races [GINIALLRF]” FRED Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/GINIALLRF. 3 Wolfgang Höchtl, "Inequality Aversion And Voting On Redistribution”, 2016, European Economic Review 56 (2012) 1406–1421.

21

“On Race and Envy: Racial Segregation and Voter Support for Redistributive Referenda” by Rebecca Kunau

voters would vote against such policies.4 Polling data from the 1970s-1990s

further corroborates the idea that “the vast majority of Americans agree …

that American society is structured in a way that benefits the rich and

penalizes the poor.”5 Despite the rise in inequality and class-conscious

sentiments, an intuitive rise in redistributive demand did not occur.

Instead, empirical evidence demonstrates that increases in actual and

perceived levels of national inequality do not result in more social programs,

progressive taxes, or other redistributive policies.6 Income taxes are nowhere

near as progressive as they were from 1932–1980, when the top marginal

income tax rate in the U.S. averaged 81 percent.7 Further, the average

American voter has demonstrable “difficulty linking broad concerns about

inequality or economic self-interest to congruent policy preferences.”8 In fact,

strict “personal economic motivations play little role in the formation of

[their] opinions” regarding national policies or personal voting decisions.9

4 This is also the basis of the Median Voter Hypothesis, which contends that a voter in an unequal society with the median level of income (“the median voter”) would be relatively less well off in relation to the mean income. When the taxes are more progressive and governmental cash transfers are higher, the median voter stands to fiscally benefit. Therefore, they would choose to vote for more redistributive policies. 5 Leslie McCall, “Do They Know and Do They Care? Americans’ Awareness of Rising Inequalities”, May 2005, Russel Sage Foundation Social Inequality Conference, University Of California, Berkeley, Conference Paper. 8. 6 Kenworthy, Lane and Leslie McCall. 2007. "Inequality, Public Opinion And Redistribution". Socio-Economic Review 6 (1): 35-68. doi:10.1093/ser/mwm006. 7 Piketty, Capital in the 21st Century. 507. 8 Franko, William, Caroline J. Tolbert, and Christopher Witko. 2013. “Inequality, Self-Interest, and Public Support for “Robin Hood” Tax Policies.” Political Research Quarterly. University of Utah. 924. 9 Ibid, 924.

22


It’s clear that there is more to learn about voter preferences, since general,

low-grade awareness of institutionalized barriers to class mobility has not

pushed American citizens to actually demand redistributive policies at the

ballot box.

Redistributive policies themselves can be broadly defined to include

everything from social programs for protected classes, such as food stamps

for women and children in poverty10 and health care for veterans, to

investments in public goods and infrastructure projects, to minimum wage

policies that directly impact the incomes of those living within striking

distance of the poverty line.11 But the focus here is not on styles of

redistribution; rather, the driving force behind this research is a desire to

understand the dearth of support for such policies. Thus the question in focus

is small, but important: what characterizes communities that are more likely

to support redistributive policies?

I argue that Americans vote for minimum wage increases not for

personal, pecuniary reasons, but because of envy and racism. Though these

are less traditional or economically rational concepts, the salience of voters’

social standing ultimately does have a meaningful impact on whether the

community ultimately supports the redistributive policy. Here, that

10 Commonly referred to as WIC, the Special Supplemental Nutrition Program for Women, Infants and Children is a federal program that provides grants to states to support the healthcare and nutrition of low-income pregnant women, breastfeeding women, and at-risk infants and children under the age of five. 11 Alexander Pacek and Benjamin Freeman, “The Welfare State and Quality of Life: A Cross-National Analysis”, 1995, Texas A&M University.

23


redistributive policy is defined as a proposed increase in the minimum wage,

and salient social standing is measured by the voter’s exposure to people with

a higher income and different racial background than their own. The results

demonstrate that counties are more likely to support redistributive policies

when they are racially segregated and exhibit unbalanced income

distributions.

This paper focuses on statewide redistributive ballot measures, relying

on counties as the unit of analysis. These ballot measures are somewhat

abstracted from the partisan commitments of politicians, while still allowing

my research to examine different communities’ sentiments regarding

redistributive policies. Though this research cannot definitively establish

causal mechanisms, it can look more closely at the characteristics of counties

that do (or do not) support redistributive policies, and use social theories to

understand why certain characteristics might be correlated with these

communities. These standalone ballot items are perfect tools of analysis since

they are disaggregated from other policy concerns and explicitly measure

public opinion on just one topic. After the tighter questions of

characterization are answered, further research can focus on the causal

influence of inequality salience or racial homogeneity on redistributive

preferences.


24


Explicit pecuniary self-interest plays little to no role in political

behavior.12 Research over the last 30 years has demonstrated that American

citizens do not link their actions in voting booths with the potential absolute

fiscal benefits, which means that they place less importance on utility exacted

from monetary gains than on socially-grounded utility.13 So, if monetary self-

interest doesn’t drive policy support, what does?

One theory, referred to as the social rivalry effect, centers on the idea

that some people derive utility from being surrounded by people who they

perceive as being in their same socioeconomic stratum.14 Redistributive

policies are designed to improve the lives of those with lower incomes; these

policies make it possible for low-income people to get their children into

better schools and to enter new neighborhoods, which improves their chance

of achieving intergenerational upward mobility and increases their exposure

to different communities.15 As such, American voters may oppose

government-organized redistribution even if they derive no monetary

disadvantage from it, since it may increase their likelihood of interacting with

someone of a lower socioeconomic background as ‘the poor’ are integrated

into the community. Though the literature does not make this leap, it may

12 Stanley Feldman, “Economic Self-Interest and Political Behavior”, 1982, American Journal of Political Science 26(3). 446–466. 13 Ibid. 14 Giacomo Corneo and Hans Peter Grüner, “Individual Preferences for Political Redistribution,” 2002, Journal of Public Economics 83. 83–107. 15 Raj Chetty, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez, "Where Is The Land Of Opportunity? The Geography Of Intergenerational Mobility In The United States" (2014).

25


also follow that racial segregation would operate in a similar manner: people

may be more likely to support policies that allow them to continue to interact

with people who are racially similar and prevent the intrusion of ‘others.’

Furthermore, the perception of racial bias in redistributive policies

could have a significant impact on support for these policies since means-

tested welfare programs are thought to be more heavily utilized by Blacks.

There is literature that bears this out; Americans disproportionately believe

that Black Americans rely on welfare and anti-poverty programs at a higher

rate than Whites– even when the policies themselves are explicitly race-

neutral.16 This means that “non-racial survey questions [recording

respondents’ attitudes] toward the poor may reflect their racial attitudes as

well,” and that their voting behavior for redistributive ballot measures could

potentially demonstrate this same implicit racial bias.17 Essentially, voters

could be less likely to support redistributive policies if they believe people of

a different racial background will be the primary beneficiaries.

The utility of perceived social status should be given considerable

weight in analyses of voters’ redistributive preferences as there are good

reasons to expect that redistributive sentiments might vary based on

proximity to wealth. People do not conduct assessments of their wellbeing in

16 Martin Gilens, “Chapter 7: Racial Attitudes and Race-Neutral Social Policies: White Opposition to Welfare and the Politics of Racial Inequality”, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy, University of Chicago Press, 2000. 17 Ibid, 5.

26


a vacuum but rather (unconsciously) make comparisons of their own utility

relative to a reference group. These groups can and should be primarily

defined geographically since it is that population, which is the most salient to

any given voter. When reference groups are experimentally defined in terms

of geographic proximity, support for redistributive politicians depends

primarily on exposure to higher-income neighbors.18

In a similar vein, last-place aversion theorizes that low-income voters

elect to “punish those slightly below themselves to keep someone [in their

community] beneath them,” undermining support for redistributive policies.19

This revelation is demonstrated by the preferences of workers earning just

above minimum wage, who have been demonstrated to be “most likely to

oppose minimum-wage increases, as they would no longer have a lower-wage

group beneath them.”20 This relationship explains why people who are

impoverished by national standards (and would therefore stand to benefit

from transfers and redistribution by the federal government) do not uniformly

support redistributive policies. If the voter has a low income, but makes a

wage exceeding the average in their small communities, they benefit from the

social utility they gain from feeling as though they are on top. This small fish,

18 Jacob Vigdor, “Fifty Million Voters Can’t Be Wrong: Economic Self-Interest and Redistributive Politics.” National Bureau of Economic Research. Working Paper No. 12371. (2006). 19 Ilyana Kuziemko, Ryan Buell, Taly Reich, and Michael I. Norton. “”Last-place Aversion": Evidence and Redistributive Implications.” National Bureau of Economic Research. Working Paper No. 17234. (2011). 20 Ibid.

27


small pond idea has a strong impact on individuated utility calculations and

helps us understand why redistribution is not demanded at higher rates even

as inequality increases. A concern that arises from this relationship is that

counties themselves are decreasing in ideological integration, making it less

likely that citizens’ exposure to different backgrounds and mindsets will

occur.21

III. Methods

I argue that some of the variability in votes for redistributive policies

can be described by the racial segregation present in the county, and rely on

the tools used to describe de facto segregation in Chetty’s descriptive

analysis.22 Another variable that has an impact on the redistributive

preferences of the county is the income concentration of the community. This

is demonstrated by the share of income that is attributable to 1% of the

population of that county.23

I chose to focus on state-level referenda to determine redistributive

policy preferences (rather than politicians or other methods of partisan-

influenced redistributive preference determination) because they are less

21 Jesse Sussell, "New Support For The Big Sort Hypothesis: An Assessment Of Partisan Geographic Sorting In California, 1992–2010", PS: Political Science & Politics 46, no. 04 (2013): 768-773. 22 Chetty et al, "Where Is The Land Of Opportunity? The Geography Of Intergenerational Mobility In The United States”. 23 Ibid.

28


likely to be influenced by other value preferences. The referenda and ballot

initiatives referred to in this paper were discovered and compiled through

Ballotpedia, a nonpartisan, nonprofit, online encyclopedia of American

politics and elections.

The raw county-level referenda and initiative voting data analyzed in

this paper were provided by the state websites of Arizona24, Colorado25, and

Maine.26 The ballot measures in these states are structured in similar ways; all

occurred in 2016 and center around the same topic: raising the minimum

wage. I also chose these referenda because they avoid inflammatory or biased

language in their titles and descriptions.27 The states (Arizona, Colorado, and

Maine) were chosen because their data was accessible online and because

they were sufficiently geographically different from one another, rather than

all coming from the same region of the United States.

The data compiled on inequality is thanks to Raj Chetty, who offers

simple tools and comprehensive county and commuting zone-level analyses

24 "Arizona Secretary Of State Election Night Reporting", Results.Arizona.Vote, 2017, http://results.arizona.vote/2016/General/n1591/Results-State.html#ballots. 25 “Official Certified Results: November 8, 2016 General Election” Colorado Election Results, 2017, http://results.enr.clarityelections.com/CO/63746/184388/Web01/en/summary.html#. 26 "Bureau Of Corporations, Elections & Commissions, Elections And Voting, Results, 2014 Tabulations", Maine State, 2017, http://www.state.me.us/sos/cec/elec/results/results16-17.html#tally. 27 These ballot measures are titled “Arizona Minimum Wage and Paid Time Off Initiative,” aka Proposition 206 (2016), “Colorado $12 Minimum Wage,” known as Amendment 70 (2016), and the “Maine Minimum Wage Increase” aka Question 4 (2016).

http://results.arizona.vote/2016/General/n1591/Results-State.html#ballots

http://results.arizona.vote/2016/General/n1591/Results-State.html#ballots

http://results.enr.clarityelections.com/CO/63746/184388/Web01/en/summary.html

http://results.enr.clarityelections.com/CO/63746/184388/Web01/en/summary.html

http://www.state.me.us/sos/cec/elec/results/results16-17.html#tally

29


of the fifty states through the Equality of Opportunity project. Each of the

variables described in the following list were constructed by Chetty for his

work on intergenerational mobility in America. In a quick summary:

• The “Racial Segregation” variable represents an index that calculates

the maximum possible randomness of the data less the demonstrated

randomness in order to discern the degree of actual, non-random

disparity.28 Here, it focuses at the census-tract level over four groups:

White alone, Black alone, Hispanic, and Other.

• The “Top 1% Income Share” variable represents the fraction of income

within a commuting zone that goes to just 1% of the population.

• The “Share Below the 25th Percentile/Above the 75th Percentile” is

estimated by computing household income for 16 income groups, as

defined by the 2000 census.

This paper primarily utilizes the following model,

V = 𝛽𝛽AZ ∗ AZ + 𝛽𝛽CO ∗ CO + 𝛽𝛽ME ∗ ME + 𝛽𝛽R ∗ R + 𝛽𝛽T ∗ T + 𝜀𝜀

in which a given county’s support of (or opposition to) minimum-wage-

raising ballot measures (V) varies with the state that they are in (Arizona,

Colorado, and Maine, represented by their appropriate abbreviations), the

racial segregation of the county (R), and the share of the nation’s top 1% of

income earners who live in that county (T). I use the states as fixed effects

28 Pedro Conceição and Pedro M. Ferreira, "The Young Person's Guide To The Theil Index: Suggesting Intuitive Interpretations And Exploring Analytical Applications", SSRN Electronic Journal, 2000.

30


here in order to ensure that any error due to potential lack of independence of

the data is identified. The dependent variable itself is set up in such a way

that a positive coefficient indicates support for the redistributive policy, while

a negative coefficient exhibits opposition to the policy. It is very simple, and

calculated as 𝑉𝑉 = (Yes Votes−No Votes)Total Votes

.

This analytical structure has the added benefit of accessibility to

academics without backgrounds in advanced statistical modeling. For a

presentation of an alternative analysis that utilizes a multi-level model, see

the appendix. The resulting analysis is effectively the same as this simple

linear model.

31


IV. Results

Heavily segregated counties are more likely to support minimum

wage referenda, as shown by Model 6 in Table 1. For every 1% increase in

the degree of racial segregation, there is an associated increase in the degree

of demonstrated support for the policy. This appears to validate the racial

extrapolation of the social rivalry effect, indicating that people are more

likely to support redistributive policies when they only interact with other

community members of their same race. The racial segregation of the

community is also more important than a simple increase in the presence of a

minority group, which is demonstrated by the introduction of ‘Fraction

Black’ into Model 2 and Model 5.

Contrasted with each other, wealth and poverty play very different

roles in a county’s support for redistributive policies. For every 1% increase

in the share of the county’s income that is taken up by a small subset (1%) of

the population, the likelihood of that community supporting minimum wage

referenda meaningfully increases. In contrast, the data show no predictive

relationship (in significance, magnitude, or even consistent direction)

between the fraction of citizens in the county that live below the federal

poverty line and the redistributive sentiment of the county. This lends further

credence to the argument that demonstration of support for redistribution by

voting is not a solely pecuniary act; otherwise, the share of the population

32


below the poverty line would be highly predictive of support for all

redistributive policies.

Racial segregation and top 1% income share are robust in different

variations of the model and hold roughly stable significance and magnitude

across different conditions. Unfortunately for this analysis, de facto

segregation of impoverished or wealthy population subsets are not robustly

associated with support for or opposition to redistributive ballot measures.

Table 2 demonstrates that they are significant when controlling for the racial

segregation of the county, but wholly dominated by the introduction of

wealth distribution indicators like ‘Top 1% Income Share’ (as shown in Table

1, Model 5).

33


V. Discussion

The relationship between racial segregation and redistributive

preferences fully aligns with Gilens’ empirical evidence, which indicates that

welfare policies are placed in an implicitly racial context for voters.29

Redistributive ballot measure voting behavior demonstrates the same implicit

racial bias exhibited in the national surveys Gilens cites in his book. When

voters in a given county are less exposed to people of different races, they are

more likely to support redistributive policies. This relationship demonstrates

that voters are less likely to support redistributive policies when people of

different racial backgrounds are more salient within their communities.

Voters are less likely to support redistributive policies when they interact

with people of a different racial heritage than their own, regardless of their

income equality or imbalance.

In contrast, Bartels’ work appears to contradict the economic voting

patterns I put forth. He empirically demonstrates that the choices of American

voters at all income levels are sensitive to the fortunes of the wealthy– in that

they become more likely to support the incumbent (and less likely to demand

change) when wealthy families experience income growth, regardless of their

own pecuniary successes.30 Bartels explains that this supports the Republican

Party in that even when income gains are not experienced by middle or low-

29 Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy, 5. 30 Bartels, Unequal Democracy, 88.

34


income households, high-income households’ income growth positively

influences broader voter support for conservative politicians. This

relationship may seem to exist in contradiction with my results, which

directly link the county’s top 1% income share to community support for

redistributive policies and imply a certain level of envy between people of

different socioeconomic statuses who live in the same county. But, Bartels’

work is not easily comparable to my data and subsequent analysis. He

examines growth experienced by different income groups at a national level,

which likely does not have an impact on the social utility experienced within

a given county; astronomical salaries achieved by executives in New York

City would hardly be salient for the daily social status utility experienced by

an American voter in Scottsdale, AZ. Furthermore, the methods I use look at

support for specific redistributive policies in 2016 rather than using

nationwide votes for incumbent candidates collected over the last sixty years

as a proxy for economic sentiment.

My research is more closely aligned with Jacob Vigdor, who asserts

that support for redistributive politics depends primarily on exposure to

higher-income neighbors.31 We have both found that when voters live in

communities with higher income disparity, they are more likely to support

redistributive policies. His analysis, however, relies on county and individual-

31 Jacob Vigdor, “Fifty Million Voters Can’t Be Wrong: Economic Self-Interest and Redistributive Politics.”

35


level party-voting choices from 1980 and 2000 rather than a less partisan

measurement like ballot measures.

My work also has several key limitations. It focuses exclusively on

minimum wage-raising ballot measures as a proxy for redistributive

sentiment rather than relying on a broader set of ballot measure concepts. The

analysis also centers on just three U.S. states; though I could not demonstrate

a robust relationship between redistributive preferences and class integration,

a larger sample size may demonstrate such an association. All of these

limitations are rooted in the challenge of using cross-sectional data to

establish causality. My research relies on three ballot measures from 2016,

which cannot prove a singular causal relationship with racism or envy.

VI. Conclusion

My ultimate goal is to understand the dearth of support for

redistributive policies like minimum wage hikes; a good place to start has

been to understand the economic and social dynamics of the communities that

vote to support or oppose these policies. The analysis of redistributive

preferences demonstrates that counties are more likely to support

redistributive policies when they are racially segregated and exhibit

unbalanced income distributions. Americans appear to vote with their pride

and their biases, not their wallets; people tend to care much more about their

status relative to their local community than they do about their absolute net

36


worth.32 Here, the status is measured in the distribution of wealth and in

implicitly racist social constructs. It appears that Americans vote largely to

identify as superior to their neighbors; they are trying to make America great

only for those who look just like them.33

Further Research

Much more investigative work is necessary in this field. Collection

encompassing county-level data from all states that offer ballot measure

procedures would further validate the research. Once there is a full dataset

based exclusively on minimum wage referenda and initiatives, it would be

possible to ask more questions: with inequality and the ratios of very-low- to

very-high-income earners held constant, how much effect does the U.S. state

itself have on the sentiment, if any? Is there wild variation across state lines,

or are people equally sensitive to their communities’ levels of inequality,

regardless of where they are in the Union? It would also be interesting to

know if this analysis holds true for other types of redistributive policies that

are found in ballot measures, and if these preferences have been stable over

time, or if they have fluctuated.

Linking this data with more explicitly partisan voting records would

be interesting, too: would the counties’ partisan tendencies be more predictive

32 This is in reference to the research presented in this paper, as well as the following: William Franko, Caroline J. Tolbert, and Christopher Witko. 2013. “Inequality, Self-Interest, and Public Support for “Robin Hood” Tax Policies.” Political Research Quarterly. University of Utah. 33 And are we alone in this? Do most other nations have similar preferences, or is the U.S. an outlier in demonstrating this sentiment?

37


(in magnitude and significance) of their sentiments towards redistributive

policies, or would the salient inequality and relevant latent racism win out

again? I am inclined to believe that the partisanship of the county would be

less indicative than it may seem, but again it would be useful to see that

demonstrated in the data.

Ethnographic research could also be conducted in the counties that lie

on the extremes: those that are most significantly segregated based on income

and race, and those that are the most integrated, economically and racially

diverse spaces. Such work would provide insight into these communities and

help researchers better understand whether the posited behavioral theories

play conscious roles in policy preferences.

38


VII. Appendix

Though this paper relies on a simple OLS regression, another option

functions just as well: a multi-level linear model (MLM). Pictured in Figure 1

below, it offers similar assessments to those offered by the simple regression

and again highlights the significance of the direct correlations between

support for the redistributive ballot measure at hand, and racial segregation

and a large share of income attained by a small portion of those in the

community as being characteristics of the counties supporting these policies.

Use of the MLM ensures that the research does not commit the atomistic

fallacy, which would violate the assumption of independence and bias our

results.

The utility of an MLM is further highlighted by the graphs of Figures

2 and 3, which demonstrate the difference between the analysis that is

performed on the aggregate and faceted, multi-level analytics. The dependent

variable in each is the support for minimum wage ballot measures: above

zero, the citizens in the county voted in favor of the referendum, and below,

they voted in opposition. Figure 3 applies a different slope and intercept to

each, allowing analysts to easily appreciate the value of the nested data.

These graphs ultimately depict the importance of introducing the states into

the model (in any regard, but with particular focus on MLM) so that the

regression performed is not dictated by the position of the different states on

the plot.

39


The only issue in this particular instance is that there are not many

groups to be separated out for each level; there are only two levels in this

MLM, and there are only three groups (the states) that could be used on one

of the two levels (the other being all of the counties that exist in each state).

There does not appear to be much to be gained from framing the problem like

this.

That said, a multi-level model would be critically important for

understanding larger datasets that involve a larger sample of state-level

referenda. As it is, there is not much to glean from the differentiated, split-

level model that cannot be understood from the simple linear regression

presented in the results.

40


41

“Quality of Care in Hospitals after Vertical Integration with an HMO” by Aaron F. Miller

Quality of Care in Hospitals After Vertical Integration with an HMO Aaron F. Miller

Abstract: Objective: Hospital mergers and acquisitions continue to rise due to government and market pressures to join accountable care organizations and lower costs. Our study examines the changes in a quality of care measure when hospitals were acquired by HMOs to analyze the effect that vertical integration has on coordination of care. Methods: Using California patient discharge data from 2000-2011, our analysis used differences-in-differences and logistic models to test for a change in a quality of care measure before and after hospitals merged with an HMO. We utilized Patient Safety Indicator #2, Death Rate in Low-Mortality Diagnosis Related Groups, to measure quality of care. Results: Hospitals experienced decreases in low-mortality death rates after being acquired by an HMO. This group of hospitals had increasing measures of the quality indicator prior to the merge as well, suggesting selection of well-performing hospitals by HMOs. Hospitals acquired by HMOs also faced increased Type-2 diabetes rates post-merge. Conclusions: Our results suggest that hospitals merging with a vertically integrated health care system may lead to increases in quality of care. It appears that this could be due to either HMOs providing more coordinated care, or that HMOs acquire hospitals already trending towards better care. I. Introduction

United States hospital consolidations doubled from 2009 to 2012,

topping 100 total deals [1]. The majority of such acquisitions have been

hospital-to-hospital mergers, or horizontal integrations. However, a growing

subset of the consolidations is mergers of hospitals with health systems that

provide their own health plans, or vertical integration with HMOs [1).

According to the Henry J. Kaiser Family Foundation, over 25% of all

Americans receive their health insurance from HMOs [2]. Further, from 2013

to 2015, enrollment in HMO insurance plans has increased 11%. This

42


increase largely coincides with the expansion of Medicaid in 31 states under

the Affordable Care Act, and is expected to continue to rise as more states do

the same.

Increases in hospital acquisitions are consistent with federal

healthcare reform that focuses on incentivizing managed care. There is

pressure on hospitals to join accountable care organizations and the

government is increasing Medicare reimbursement. These programs, coupled

with an overall decrease in patient volume following the recession, make it

cost-effective for many hospitals to join larger healthcare systems [3].

Comparably, it is economically favorable for large health systems to acquire

hospitals in an effort to lower costs, raise capital, better coordinate care, and

more efficiently manage a population’s healthcare needs [3].

Economic theory suggests that horizontal and vertical integrated

health care systems can either make consumers better or worse off. However,

HMOs assume economies of scope, lower transaction costs associated with

outsourcing, and have incentive to invest in preventative care. Despite such

economic advantages, do hospital mergers with HMOs translate to improved

quality of care for patients? At a time when enrollment in HMOs continues to

rise, this question becomes considerably more important for patient outcomes

across the United States.

We used California inpatient hospital data from The Healthcare Cost

and Utilization Project (HCUP) from 2000-2011 of 47,442,140 patient

43


discharges from 506 total hospitals. We used logistic regression to measure

the low-mortality death rate at hospitals before and after a merge with an

HMO, and adjusted for patient demographic and health factors. We used this

data to consider the affect that a rise in HMO enrollment has on patient

outcomes in the United States.

Conceptual Framework

Vertical consolidations are often thought to be a much better option

for payers and providers, because one company owns several aspects of the

supply chain. Economic theory proposes that vertically integrated healthcare

systems should display multiple internal efficiency gains. These efficiencies

include: economies of scope, lower transaction costs associated with

outsourcing, and incentive to invest in health maintenance efforts which

reduce costs in the long run. The financial efficiencies that hospitals glean

when consolidating free up resources to devote to improved quality of care.

Hospitals have low operating margins despite wide variation in

reimbursement rates, and there is low variation in operating margin across

hospitals. Together, these facts hint that hospitals likely find places to spend

additional profits, and it is reasonable to believe that an investment in quality

improvement will rank at the top of the list for discretionary spending.

Economies of scope are cost advantages that result when a firm

provides a variety of products, rather than specializing in the production of

just one. Economic theory suggests that unit costs increase as output

44


increases. These cost improvements include productivity gains and waste

reduction from the ability to eliminate extraneous costs by operating multiple

businesses under a centralized management. In 2014, the operating margin

for an average hospital was 8.3% [1], whereas operating margins for five of

the largest health insurers was 4.3%34 [4, 5, 6, 7, 8]. In both healthcare industries,

administrative costs constitute a substantial portion of the expenses, reaching

over 25% in hospitals [9]. As such, vertically integrated health care systems

are able to increase economies of scope and operate under a centralized

management thereby cutting total administrative costs for both businesses by

eliminating duplication. The capability for vertically integrated health care

systems to have more available services could lead to increases in

coordination of care and cost reductions, both benefitting the consumer and

the business.

Vertically integrated firms are expected to have lower transaction

costs associated with outsourcing. Transaction costs are the expenses that

incur in the purchase of goods and services. In the 1930’s, Ronald Coase

developed transaction cost economics, which thoroughly underlines the

numerous inefficiencies that occur in the existence of transactions. These

transactional inefficiencies include arguments that contracts include

substantial transaction frictions, such as: interest rates, commission fees, and

opportunity costs [10]. In the 1970’s, Oliver Williamson expanded on Coase’s

34 Calculated the mean operating margin using 2014 annual reports from the five largest insurance companies in terms of revenue

45


theory and developed transaction cost theory of integration. Williamson

argues that integration can be an effective response when transactional

frictions are present [11]. Currently, the healthcare system is filled with

transaction frictions due to its fragmented nature between insurance

companies, hospitals, physicians and patients. As such, vertical integration

addresses these difficulties by consolidating production within a single

organization, thereby eliminating the need for contracting among separate

firms [11]. Therefore, all production is managed internally by one healthcare

system and such economic efficiency gains may get passed down to the

consumer.

HMOs are incentivized to invest in preventative care to obtain cost-

savings; vertically integrated healthcare systems manage the insurance rates

for their customers. As such, firms seek healthier patient populations in order

to have competitive health insurance prices. Thus, firms have economic

incentive to perform care in a way that makes the patient as healthy as

possible, in order to obtain cost savings in the future. The National

Commission on Prevention Priorities (NCPP) and the National Business

Group on Health (NBGH) found multiple preventative interventions to be

cost-saving. These interventions include childhood immunization, adult

Influenza immunization, counseling on use of low-dose aspirin, and screening

and counseling for alcohol and tobacco [12]. Therefore, healthcare systems

have incentive to invest in short-run preventative care efforts to experience

46


long run cost savings. As a result, patients experience an emphasis on

preventative care, which leads to better quality outcomes in the long run [13].

In contrast, horizontally integrated healthcare systems bill directly to private

insurance companies not affiliated with their practice. As a result,

horizontally integrated healthcare systems may have economic incentive to

perform unnecessary diagnostic tests and increase patient admissions to drive

up their reimbursement. Patients are more likely to pay higher medical bills

and suffer from less focused care because of horizontal integration.

There are generally two sizeable obstacles for vertical partnerships to

overcome, which contribute to higher prices and lower quality of care for the

patient. First, high acquisition costs for two large entities bears a substantial

amount of risk, leading to an ability to raise prices and increase admission

rates. Second, consolidation could lead to monopolization of the market,

subsequently leading to higher prices through reduced competition.

Vertical mergers of all sorts assume expensive acquisition costs. For

example, in 2015 Kaiser Permanente had an operating income of $1.9 billion,

compared to Seattle-based Group Health Cooperative’s $740 million.

However, Kaiser Permanente paid a total of $1.8 billion in 2015 to acquire

Group Health Cooperative to expand its geographic reach. Kaiser Permanente

is bearing a substantial amount of risk when putting forth costs over two

times greater than the operating income of Group Health Cooperative the

preceding year, without a guaranteed return on investment. As a result,

47


economic pressures for Kaiser Permanente to recoup its investment could

lead to increases in price, admissions, and diagnostic testing. These would in

turn lead to higher prices for consumers and possibly less focused care.

Monopolization of the market is also an important factor worthy of

consideration. Antitrust laws regulate the conduct and organization of

business corporations to promote fair competition for the benefit of

consumers. These laws exist to contain costs, improve quality, expand choice,

and incentivize innovation [14]. All companies must abide by these antitrust

laws before proceeding through any consolidation to ensure that it will not

harm the consumer. However, in 2015 the Federal Trade Commission

narrowed its definition of what constitutes an anti-competitive merger [15].

This will likely drive more large, vertical consolidations in the healthcare

industry. This raises concerns for consumers, because as healthcare systems

increase market share in geographical locations with little competition, firms

can drive up prices for an inelastic good like healthcare.

In total, this paper intends to empirically investigate whether such

operational efficiencies lead to improved health outcomes for patients.

II. Methodology

48


Data and Group Selection Process

In order to examine quality of care in hospitals affiliated with HMOs

versus those that are not, we used a differences-in-differences model to look

at low-mortality death rates in hospitals before and after merging with an

HMO. Furthermore, we used logistic regression on California inpatient

hospital data from The Healthcare Cost and Utilization Project (HCUP). The

data set in total contained 47,442,140 patient discharges from 506 total

hospitals, with 151 measured variables from 2000-2011.

We sorted hospitals into three groups: hospitals that had merged with

an HMO between 2000-2011, hospitals that had always been affiliated with

an HMO, and hospitals that had never been associated with an HMO

(control). Figure 1 depicts a representation of our selection process. We

excluded 50 hospitals that reported zero discharges in a given year between

2000-2011.

Figure 1. Methodological Flowchart

49


First, we created a control group that consisted of hospitals that had

sustained operations from 2000-2011. We studied the parent name of each

hospital from 2000-2011 using utilization data provided by HCUP, to identify

hospitals that never had a parent name affiliated with an HMO. This group

was made up of 328 hospitals and included 32,756,343 discharges.

Next, we sorted all hospitals that had sustained operations with an

HMO from 2000-2011. To thoroughly classify this subset of hospitals, we

identified all hospitals that maintained a given HMO parent name throughout

this timeframe. This group consisted of 59 hospitals and 8,561,390

discharges.

Finally, we separated one group to include all hospitals that had

vertically merged to an HMO between the years 2000-2011. To

systematically determine which hospitals were merged with an HMO, we

studied the parent name of each hospital from 2000-2011 to identify changes

that had taken place. A change suggested that a hospital merged with another

healthcare system. We further examined all hospitals that had a parent name

change in this timeframe by reviewing insurance plan information from the

Department of Managed Healthcare for the state of California to differentiate

between HMO consolidations versus non-HMO integrated consolidations.

This HMO-merged group consisted of eight hospitals and included 827,992

discharges.

50


Figure 2. Regressing low-mortality death rates in freestanding hospitals

versus HMO hospitals from 2000-2011 using simple linear regression

The eight hospitals that were acquired by an HMO from 2000-2011

merged in different years. Figure 2 regresses freestanding and HMO-

affiliated hospitals over time on low-mortality death rate (per 10,000). These

data show a reduction in death rate in HMO but not freestanding hospitals.

This observation suggested that we should control for variances in the

healthcare system over time. To do so, we added a variable to assign a value

corresponding to number of years pre/post-merge to depict when the hospital

0

0.5

1

1.5

2

2.5

2000 2002 2004 2006 2008 2010 2012

Low

Mor

talit

y D

eath

Rat

e (P

er 1

0,00

0 D

isch

arge

s)

Year

Low Mortality Death Rate - Freestanding vs. HMO Hospitals

Freestanding

HMO

51


had merged (Figure 3). For example, if a hospital merged in 2006, its 2008

data is assigned a 2 (years post-merge).

Measuring Quality of Care

The response variable we selected as a proxy for quality of care was

the AHRQ Patient Safety Indicator #2, “Death Rate in Low-Mortality

Diagnosis Related Groups”. We chose this measure because the underlying

assumption of the Patient Safety Indicator is that when patients admitted for

an extremely low-mortality condition or procedure die, a healthcare error

across multiple components of the healthcare system is likely to be

responsible. Thus, it attempts to assess coordination of care, which is what

vertically integrated systems claim to improve. This Patient Safety Indicator

describes in-hospital deaths per 1,000 discharges for low mortality (< 0.5%)

Diagnosis Related Groups among patients ages 18 years and older or obstetric

patients. In accordance with the Agency for Health Related Quality, the

quality measure excludes cases with trauma, cancer, an immunocompromised

state, and transfers to an acute-care facility [16]. This indicator intends to

identify in-hospital deaths in patients unlikely to die during hospitalization.

To study the health of the patient populations in both the experimental

and control hospitals, we studied the prevalence of Type-2 diabetes in each

group using ICD-9-CM codes. We analyzed this variable to study the health

of the patient populations in all respective hospital groups.

52


Econometric Models

Each observation of our quality indicator is a binary variable, where

Y is 1 if the patient died from a low-mortality condition. We used a

differences-in-differences approach to study the differential effect of vertical

integration with an HMO on quality. We used logistic regression analysis

modeled by the equation shown below, where X is a vector of patient

characteristics to study low mortality death rate in each respective hospital

group:

Low Mortality Death Rate = β0 + β1Post-merge*HMO +

β2Year + β3X + ε (1)

We conducted logistic regressions to control for demographic variables such

as: age, race, gender, diabetes rate, and length of stay.

Characteristics of the Sample

The characteristics of the different hospital groups that we focus on

are important to understand. Table 1 reports summary statistics for the main

control variables for the three categories of hospitals: freestanding hospitals,

HMO hospitals, and HMO-acquired hospitals.

Freestanding HMO HMO-Acquired

53


Table 1. Summary statistics for three groups of hospitals: freestanding,

HMO-affiliated, and HMO-acquired hospitals from 2000-2011.

Most hospital characteristics and patient demographics do not differ

substantially across the different hospital types. The HMO hospitals

experienced higher mean patient discharges per hospital than the freestanding

hospitals from 2000-2011, and this same trend is followed by the HMO-

acquired hospitals that increased mean patient discharges per hospital after

merging. Further, the mean length of stay was higher in the freestanding

hospitals than both the HMO and HMO-acquired patient populations. Shorter

lengths of stay are typically associated with better quality of care.

Pre-Acquisition Post-Acquisition

Total Hospitals Included 328 59 8 8 Total Patient Discharges 32,756,343 8,561,390 429,523 398,469 Mean Patient Discharges/Hospital/Year 8,322 12,092 8,516 9,738 Change in Patient Discharges (2000-2011) 7.0% 11.7%

Mean Low Mortality Death Rate/10,000 Discharges 1.47% 0.97% 1.94% 0.66% Change in Low Mortality Death Rate (2000-2011) -5.7% -55.9%

Mean Diabetes Rate 14.3% 15.5% 16.6% 22.4% Change in Diabetes Rate (2000-2011) 72.0% 65.3%

Male 32.8% 31.9% 33.5% 33.7% Female 48.9% 49.7% 53.3% 53.7% White 55.1% 52.7% 75.8% 69.9% Black 4.7% 6.6% 2.9% 4.8% Asian 4.8% 4.8% 1.4% 2.4% Mean Age (years) 42.7 44.5 53.4 53.1 Mean Length of Stay (days) 5.4 4.1 4.4 4.1

54


III. Results

***=Statistically significant at the 0.01 level; ** = Statistically significant at

the 0.05 level; * = Statistically significant at the 0.10 level

Table 2. Regressing on low-mortality death rate using a logistic model, with

and without diabetes

Our main finding best supports the notion that hospitals experience

better coordination of care within a couple years of vertically integrating into

an HMO. The AHRQ patient safety indicator “low-mortality death rate” is a

measure of hospital quality. As such, we conducted our analysis regressing an

interaction between hospital type and the post-merge time period on low-

mortality death rate in Table 2, to identify which forms of administrative

structures are linked to better care. Prior to being acquired by an HMO, these

(Excluding diabetes)

HMO hospitals -0.4009*** -0.395***

Acquired HMO hospitals 0.2094* 0.242**

Acquired HMO hospital-post-merge interact -1.0097*** -1.000***

Diabetes 0.6429*** --

Length of Stay 0.0013*** 0.001***

Year -0.0129*** -0.008**

Constant 17.1621** 7.79

N 35,611,548 35,611,548

55


hospitals had higher low-mortality death rates than freestanding hospitals.

However, the acquired HMO hospital-post-merge interaction term has a

negative and statistically significant coefficient, suggesting that the hospitals

experienced reductions in low-mortality death rates following a merge with

an HMO.

According to Table 2, it appears that HMOs have improved low-

mortality death rates compared to freestanding hospitals. Given the fact that

total patient discharges are increasing faster in HMOs than in freestanding

hospitals, patient selection may be responsible. Further, the negative

coefficient on our time variable indicates that overall quality in all hospitals

may be increasing over time. We obtained similar results for each variable

with different model specifications, meaning that our estimates are robust.

Omitting any control variable from our model does not change the results.We

regressed freestanding hospitals from 2000-2011 on low-mortality death rates

to identify how quality changed over time. Figure 3 shows that the mean low-

mortality death rate in freestanding hospitals was fairly consistent over time.

In general, there was a slight decrease in low-mortality death rates from

2000-2011. This suggests that the hospitals in our analysis likely experienced

rather modest increases in quality over time.

56


Figure 3. Regressing on low-mortality death rate for hospitals not affiliated

with an HMO from 2000-2011 using simple linear regression

On the other hand, Figure 4 shows regression of hospitals over time,

relative to the year they merged with an HMO, on low-mortality death rate.

This demonstrates that on average, merging with an HMO was correlated

with decreases in a hospital’s overall low-mortality death rate. The data are

broken out into two different time series groups, pre-merge and post-merge.

Both groups had relatively the same, downward sloping trend. The greater

magnitude of the pre-merge slope, in comparison to the slope of the

freestanding hospitals from Figure 3, suggests that hospitals that were bought

up in acquisition deals already had trends towards decreasing low-mortality

death rates. However, merging with an HMO was associated with an overall

0

0.5

1

1.5

2

2.5

3

2000 2002 2004 2006 2008 2010 2012

Low

Mor

talit

y D

eath

Rat

e (p

er 1

0,00

0 D

isch

arge

s)

Year

Low Mortality Death Rate - Freestanding Hospitals

Freestanding

57


reduction in the low-mortality death rate even after accounting for the trend,

indicative from the decreased y-intercept value in the post-merge time series.

Figure 4. Regressing on low-mortality death rate for hospitals six years

before and six years after a merge with an HMO using simple linear

regression

We regressed hospital groups and patient demographics on Type-2

diabetes rates to determine if HMO hospitals cream-skimmed to obtain

healthier patient populations. According to Table 3, the pool of hospitals that

HMOs chose to acquire had higher rates of diabetes before the acquisition,

and that rate increased a bit once the merge happened. The continued

increase, coupled with the result that always-existing HMO hospitals

experienced increased rates of diabetes as well, may be due to diabetes

management programs attracting diabetes patients to the HMOs that offered

0

0.5

1

1.5

2

2.5

3

-6 -4 -2 0 2 4 6

Low

Mor

talit

y D

eath

Rat

e (P

er 1

0,00

0 D

isch

arge

s)

Time Since Merge

Low Mortality Death Rate - HMO-Acquired Hospitals

Pre-Merge

Post-Merge

58


them. This increase may suggest that HMO hospitals did not cream-skim the

healthiest patients.

Diabetes

Rate

HMO hospitals 0.082***

Acquired HMO hospitals 0.350***

Acquired HMO hospitals post-merge

interaction 0.082***

Female -0.279***

Black 0.507***

Asian 0.251***

Time 0.054***

Constant -110.7***

N 42,145,725

*** = Statistically significant at the 0.01 level; ** = Statistically significant at

the 0.05 level; * = Statistically significant at the 0.10 level

Table 3. Regressing on diabetes rate using a logistic model

IV. Discussion

We analyzed the difference in quality of care delivered at hospitals

before and after being acquired by an HMO. We found that hospitals

experienced a drop in low-mortality death rates associated with the date of

59


their merger. We do not believe this change is due to selection of patient

populations, because there was an increase in the diabetes rate in the post-

merge hospital patient population consistent with the rising diabetes rates that

freestanding hospitals experienced over time. Acquired hospitals had a

greater proportion of white patients than that of the HMO and freestanding

hospitals. However, all other races were associated with lower low-mortality

death rates relative to whites. Thus, both findings suggest that the result of

higher quality is not attributable to healthier patient populations, but perhaps

rather to improved care following a merge with an HMO.

One possible explanation for increased quality in post-merge HMO-

affiliated hospitals is that the new management may have instituted clinical

protocol specific to the HMO to achieve system-wide standardization. In line

with our findings, on average HMOs have approximately 50 percent more

discharges per hospital than freestanding hospitals. Comparably, hospital

discharges increased 14 percent for hospitals per year following an HMO

merge. In response to excess patient flow, HMOs likely used clinical protocol

at a higher rate than other hospital types to maintain consistent “best

practices” throughout its hospital system. Literature supports the

effectiveness of protocol on positive patient outcomes on a wide range of

different workflow processes and procedures [17, 18, 19, 20]. Standardized

protocols decrease unwarranted variability, which could result in lower low-

mortality death rates following a merge with an HMO, consistent with our

60


findings [18].

Further, it is possible that merging with an HMO allows a hospital

access to a wider array of resources and more specialized care for their

patients. HMO and post-merge patient populations were associated with

shorter lengths of stay in comparison to freestanding and pre-merge hospitals.

Patient access to specialized clinics due to a larger health network system in

the HMO hospitals could be responsible for more focused and efficient care.

Past findings have revealed that HMO-affiliated physicians and more

specialized physicians were both more efficient, irrespective of patient illness

characteristics [21]. Both results suggest that merging with an HMO could

possibly lead to more specialized and efficient care, in turn increasing quality.

An alternative interpretation is that HMOs acquired higher-

performing hospitals that were trending towards increased quality,

irrespective to the merge. This explanation is consistent with the steeper

downward trend of the low-mortality death rate in the pre-merge hospital

group. This hypothesis would assert that the acquired hospitals are not a

representative sample of the effects of an HMO merge on quality of care.

However, we found that hospitals saw large reductions in low-mortality death

rates the third year and beyond after merging with an HMO. These decreases

are likely associated with the HMO, rather than the hospitals themselves. This

finding is consistent with past literature that showed initial decreases in

quality after converting to a for-profit hospital, but regained its initial quality

61


performance the third year following conversion [22]. Having made this

distinction, we feel that the evidence may point to a switch rather than a

continuation of a trend.

Previous literature analyzing the relationship between managed care

and quality show a relatively equal number of significantly better and worse

HMO results for healthy patients, compared with non-HMO plans [23, 24].

Studies used multivariate regression in HMO versus non-HMO hospitals over

time to study quality of care indicators, rather than pre- and post-HMO merge

data like we did. Within differing HMO plans, the literature suggests that

investor-owned HMOs deliver lower quality of care than not-for-profit plans

[25]. Nonetheless, all of the aforementioned research was conducted prior to

the implementation of the Affordable Care Act, warranting further studies.

Prior research has examined the impact that hospital mergers have on

quality of care. The majority of studies we reviewed found that increasing

hospital market concentration has relatively no effect on quality [26, 27, 28]. Of

the studies reviewed, Carlin, Dowd and Feldman most closely resembled our

study [29]. They used administrative data for health plan enrollees in the

clinics that were acquired by integrated delivery systems in the Minneapolis-

St. Paul area for two years prior to and four years after the acquisitions. They

used a differences-in-differences model and estimated Probit models, similar

to ours, with errors clustered within enrollees to compare changes in quality

measures such as breast cancer, colorectal cancer, and cervical cancer. The

62


study found that the clinics that vertically integrated had slightly better

quality outcomes. This suggests that IDS clinical acquisitions have the

potential to improve patient outcomes, consistent with our findings.

Our study had several limitations due to the fact that we used patient

outcome data. First, we are incapable of knowing the level of accuracy of the

patient outcome data. The construct of the Patient Safety Indicator that we

used relies on the assumption that every patient diagnosis was identified,

correct, and documented by the hospital staff. Our study is dependent on

hospital precision, and our merged hospital dataset only contained eight

distinct hospitals. As a result, poor documentation practices at one or more of

these hospitals could possibly influence the validity of the outcome of this

study. Despite this drawback, the Patient Safety Indicators are still widely

used in the literature [30, 31, 32].

Second, the demographic data for each hospital group had missing

values due to the fact that this statistic was self-reported. As a result, it is

possible that gender and race proportions may have varied more than

expected. This could have led to misinterpretations of the patient populations

in our analysis.

Finally, we did not have access to comprehensive financial data for

each hospital. Thus, we were unable to investigate further into the financials

of the merged hospital group. If we had access to said data, we could have

assessed hospital performance pre- and post-merge with an HMO, as well.

63


V. Conclusion

We examined changes in the low-mortality death rate in three groups:

non-HMOs, HMOs, and hospitals that vertically integrated into an HMO. We

used a differences-in-differences model to study the effects on the HCUP data

set that included inpatient hospital data for patients in California from 2000-

2011. Our findings suggest that merging with an HMO was associated with

lower low-mortality death rates. We believe that merging with HMO

hospitals could potentially result in better coordination in care, and increased

quality of care as a result.

There are multiple ways in which this research can be meaningfully

extended. First, there is a need to explore the prevalence of protocol in

different hospital types. Evidence from that research can explain whether the

decrease in low-mortality death rate observed from this analysis was simply

due to increased use of protocol, or whether the cause is due to an alternative

aspect of the HMO model.

Second, it is increasingly important to study those hospital

characteristics that dictate the success of a merge in the healthcare sector at a

time of frequent consolidations. Hospitals foreclosures result in hundreds

unemployed and thousands seeking care. As previously noted, acquisitions

are costly and risky, thus it is important to understand aspects that contribute

to the viability of hospitals following consolidation.

64


VI. References

[1] American Hospital Association: trends affecting hospitals and health systems [Internet]. Trendwatch Chartbook 2012; [cited 2017 January 23]. Available from: http://www.aha.org/research/reports/tw/chartbook/index.shtml.

[2] The Henry J. Kaiser Family Foundation: total HMO enrollment [Internet]; [cited 2017 January 28] Available from: http://kff.org/other/state-indicator/total-hmo-enrollment/.

[3] Saxena S, Sharma A, Wong A. Succeeding in Hospital & Health System M&A. 2013. [4] Aetna: 2014 Annual Report [Internet]; [cited 2017 January 21]. Available

from: https://materials.proxyvote.com/Approved/00817Y/20150313/AR_2 39789/

[5] Anthem: 2014 Annual Report [Internet]; [cited 2017 January 21]. Available from: http://media.corporate-ir.net/media_files/IROL/13/130104/2014AR/export7/pdfs/Anthem_2014AR.pdf

[6] Humana: 2014 Annual Report [Internet]; [cited 2017 January 21]. Available from: http://phx.corporate-ir.net/phoenix.zhtml?c=92913&p=irol-reportsannual

[7] Kaiser Permanente: 2014 Annual Report [Internet]; [cited 2017 January 21]. Available from: https://share.kaiserpermanente.org/static/kp_annualreport_2014/

[8] UnitedHealth Group: 2014 Annual Report [Internet]; [cited 2017 January 21]. Available from: http://www.unitedhealthgroup.com/2014-annual/Default.aspx

[9] Himmelstein, D. U., Jun, M., Busse, R, et al. A Comparison of Hospital Administrative Costs in Eight Nations: U.S. Costs Exceed All Others by Far. Health Affairs. 2014;33(9);1586-94. doi: http://dx.doi.org/10.1377/hlthaff.2013.1327.

[10] Coase RH. The Nature of the Firm. Economica. 1937;4(16):386-405. [11] Williamson OE. The Economics of Organization: The Transaction Cost

Approach. American Journal of Sociology. 1981;87(3):548-77. doi: http://dx.doi.org/10.1086/227496.

[12] Goodell S, Cohen J, Neumann P. Cost Savings and Cost-Effectiveness of Clinical Preventive Care. Princeton (NJ): Robert Wood Johnson Foundation; 2009 Sep. Report No.: 18.

[13] Maciosek M, Coffield A, Flottemesch T, Edwards N, Solberg L. Greater Use of Preventive Services In U.S. Health Care Could Save Lives At Little Or No Cost. Health Affairs. 2010;29(9):1656-60. doi: http://dx.doi.org/10.1377/hlthaff.2008.0701.

[14] Gaynor M, Vogt WB. Antitrust and competition in health care markets. Handbook of Health Economics. Volume 1, Part B: Elsevier; 2000. p.

http://www.aha.org/research/reports/tw/chartbook/index.shtml

http://dx.doi.org/10.1086/227496.

http://dx.doi.org/10.1377/hlthaff.2008.0701.

65


1405-87. [15] Bendix J. MACRA Doesn't Spell End of Independent Medical Practices.

Medical Economics. 2016. [16] AHRQ: Patient Safety Tools and Resources [Internet]; [cited 2016 Dec

12]. Available from: http://www.hcup-us.ahrq.gov/reports/statbriefs/sb180-Hospitalizations-United-States-2012.pdf.

[17] Hijazi M, Al-Ansari M. Protocol-Driven vs. Physician-Driven Electrolyte Replacement in Adult Critically Ill Patients. Ann Saudi Med. 2005;25(2):105-10.

[18] Friedman SM, Mendelson DA, Kates SL, McCann RM. Geriatric Co-Management of Proximal Femur Fractures: Total Quality Management and Protocol-Driven Care Result in Better Outcomes for a Frail Patient Population. J Am Geriatr Soc. 2008;56(7):1349-56. doi: http://dx.doi.org/10.1111/j.1532-5415.2008.01770.x.

[19] McCallie KR, Lee HC, Mayer O, Cohen RS, Hintz SR, Rhine WD. Improved outcomes with a standardized feeding protocol for very low birth weight infants. J Perinatol. 2011;31(S1):S61-S7. doi: http://dx.doi.org/10.1038/jp.2010.185.

[20] Ogilvie-Harris DJ, Botsford DJ, Hawker RW. Elderly Patients with Hip Fractures: Improved Outcome with the Use of Care Maps with High-Quality Medical and Nursing Protocols. Journal of Orthopaedic Trauma. 1993;7(5):428-37. PubMed Central PMCID: PMC8229379.

[21] Chilingerian J. Evaluating Physician Efficiency in Hospitals: A Multivariate Analysis of Best Practices. European Journal of Operational Research. 1995;80(3):548-74.

[22] Picone G, Chou S-Y, Sloan F. Are For-Profit Hospital Conversions Harmful to Patients and to Medicare? RAND Journal of Economics. 2002;33(3):507-23.

[23] Miller RH, Luft HS. Does Managed Care Lead to Better or Worse Quality of Care. Health Affairs. 1997;16(5):7-25. doi: http://dx.doi.org/10.1377/hlthaff.16.5.7.

[24] Hellinger FJ. The Effect of Managed Care on Quality: A Review of Recent Evidence. Arch Intern Med. 1998;158(8):833-84.

[25] Himmelstein D, Woolhandler S, Hellander I, et al. Quality of Care in Investor-Owned vs. Not-For-Profit HMOs. JAMA. 1999;282(2):159-63. doi: http://dx.doi.org/10.1001/jama.282.2.159.

[26] Volpp K, Ketcham J, Epstein A, Williams S. The Effects of Price Competition and Reduced Subsides for Uncompensated Care on Hospital Mortality. Health Services Research, 2005;40(4).

[27] Mutter R, Wong H. The Effects of Hospital Competition on Inpatient Quality of Care. Agency for Health Care Research and Quality. 2004.

[28] Shen YC. The Effect of Financial Pressure on the Quality of Care in Hospitals. Journal of Health Economics. 2003;22(2).

http://dx.doi.org/10.1111/j.1532-5415.2008.01770.x.

http://dx.doi.org/10.1038/jp.2010.185.

http://dx.doi.org/10.1377/hlthaff.16.5.7.

http://dx.doi.org/10.1001/jama.282.2.159.

66


[29] Carlin C, Dowd B, Feldman R. Changes in Quality of Health Care Delivery After Vertical Integration. Health Services Research. 2015;50(4):1043-68. doi: http://dx.doi.org/10.1111/1475-6773.12274.

[30] Cook A, Gaynor M, Stephens Jr M, Taylor L. The effect of a hospital nurse staffing mandate on patient health outcomes: Evidence from California's minimum staffing regulation. Journal of Health Economics. 2012;31(2):340-8. doi: http://dx.doi.org/10.1016/j.jhealeco.2012.01.005.

[31] Cuellar AE, Gertler PJ. Strategic integration of hospitals and physicians. Journal of Health Economics. 2006;25(1):1-28. doi: http://dx.doi.org/10.1016/j.jhealeco.2005.04.009.

[32] Rose J, Evans C, Barleben A, et al. Comparative Safety of Endovascular Aortic Aneurysm Repair Over Open Repair Using Patient Safety Indicators During Adoption. JAMA. 2014;149(9):926-32. doi: http://dx.doi.org/10.1001/jamasurg.2014.1018.

http://dx.doi.org/10.1111/1475-6773.12274.

http://dx.doi.org/10.1016/j.jhealeco.2012.01.005.

http://dx.doi.org/10.1016/j.jhealeco.2005.04.009.

http://dx.doi.org/10.1001/jamasurg.2014.1018.

67

“American Manufacturing: How Important is it to the U.S. Economy?” by Robert J. Leet

American Manufacturing: How Important is it to the U.S. Economy? Robert J. Leet

Abstract: Employment in American manufacturing has been in consistent decline for roughly thirty years. The purpose of this paper is to gain an understanding of why this decline is taking place by examining various micro and macroeconomic factors at play. Multiple sources of literature attribute a lack of federal funding, a lack of adequate workplace skills, China’s world market presence, and the US current account deficit to be the main causes of persistent domestic manufacturing unemployment. The Heckscher-Ohlin (H-O) and Stolper-Samuelson (S-S) theorems bring greater clarity to these issues in the analysis section by explaining why American manufacturing unemployment is high in labor-intensive goods and why capital owners are unambiguously better off than ordinary workers. China’s devalued yuan is also analyzed and is shown to negatively contribute to America’s preexisting current account deficit and cause more manufacturing unemployment. The conclusion discusses several downsides of the current president’s tariff plans and presents more sustainable empirical-based policies. Economic theory is finally applied to create a production subsidy and consumption tax policy, which, as opposed to a tariff, sustains strong trade relationships, stabilizes government finances, and improves domestic manufacturing employment. I. Introduction

When the word “manufacturing” comes to mind, many individuals

define this term as an economic activity that combines human effort (labor)

with physical plant and equipment (capital) to transform raw materials into

products.35 This definition of manufacturing is indeed true, but it does not

accurately characterize domestic manufacturing today. According to a

government report in 2013, manufacturing in the US “generally [refers] to the

35. Jeremy Atack and Fred Bateman, "Manufacturing," in Economic Sectors, ed. Susan B.

Carter, et al., Millennial ed., vol. 4, Historical Statistics of the United States (New York, NY: Cambridge University Press, 2006), 573.

68


production of scientifically and technologically-intensive products, in which

the economic value derives from the inputs of knowledge and design more

than it reflects traditional inputs such as labor and materials.”36 American

manufacturing, in other words, is a constantly changing term, and its

contextual malleability largely reflects its compositional changes in the US

over the past several decades.

Domestic manufacturing employment has reached staggering lows

since peaking in 1979.37 According to data collected by the Bureau of Labor

Statistics, employment in US manufacturing declined from 30 percent in

1950 to roughly 10 percent in 2016 as a share of total non-farm employment

(see appendix A for a graphical representation of this decline).38

These employment figures may lead one to ask: is this decline

something I should be worried about? To answer this question, one must first

weigh the importance of manufacturing to the US economy. According to

several Congressional hearings that took place from 2009 to 2013, domestic

manufacturing is a key driver of innovation to the nation’s economy—it

accounts for 70 percent of US private-sector research and development

36. United States Congress Government Accountability Office (GAO), Global Manufacturing:

Foreign Government Programs Differ in Some Key Respects from Those in the United States, by Andrew Sherrill and Lawrence Evans, report no. 13-365 (Washington, DC: Government Accountability Office, 2013), 3, accessed May 9, 2017, http://www.gao.gov/assets/660/656239.pdf.

3. U.S. Bureau of Labor Statistics, "Employment, Hours, and Earnings from the Current Employment Statistics survey (National)," U.S. Bureau of Labor Statistics, last modified 2017, accessed April 5, 2017, https://data.bls.gov/pdq/SurveyOutputServlet.

4. Marlene A. Lee and Mark Mather, "U.S. Labor Force Trends," Population Reference Bureau

63, no. 2 (June 2008), 7, accessed May 9, 2017, http://www.prb.org/pdf08/63.2uslabor.pdf.

69


(R&D) spending,39 and generates more than 60 percent of all the nation’s

patents.40 Domestic manufacturing also creates many positive externalities

for the nation; while the industry composes only 11 percent of the country’s

GDP,41 it creates $1.40 of additional economic activity for every dollar that

the government spends toward this sector.42 In addition, every job created in

manufacturing creates two-and-a-half jobs in other sectors.43 Lastly,

manufacturing jobs offer comfortable salaries and benefits—workers earn

wages that are 17 percent higher on average than their counterparts in other

industries.44 From this evidence, one can infer that domestic manufacturing

has significant benefits to the success and sustainability of the nation’s

economy.

This essay will review several pieces of literature that have examined

the current manufacturing situation in America. The literature attributes (1) a

39. Committee on Commerce, Science, and Transportation (United States Senate), The Role

of Manufacturing Hubs in a 21st Century Innovation Economy, S. Rep. No. 113-664, 1st Sess., at 1 (2013) (Conf. Rep.), 11, accessed May 9, 2017, https://www.gpo.gov/fdsys/pkg/CHRG-113shrg95215/pdf/CHRG-113shrg95215.pdf.

40. Committee on Commerce, Science, and Transportation (United States Senate), Manufacturing Our Way to a Stronger Economy, S. Rep. No. 112-293, 1st Sess., at 1 (2011) (Conf. Rep.), 18, accessed May 9, 2017, https://www.gpo.gov/fdsys/pkg/CHRG-112shrg73231/pdf/CHRG-112shrg73231.pdf.

41. Ibid., 2. 42. Committee on Banking, Housing, and Urban Affairs (United States Senate), The United

States as a Global Competitor: What are the Elements of a National Manufacturing Strategy?, S. Rep. No. 111-281, 1st Sess., at 1 (2009) (Conf. Rep.), 8, accessed May 9, 2017, https://www.gpo.gov/fdsys/pkg/CHRG-111shrg55018/pdf/CHRG-111shrg55018.pdf.

43. Ibid. 44. Committee on Commerce, Science, and Transportation (United States Senate), The Role

of Manufacturing Hubs in a 21st Century Innovation Economy, 11.

70


lack of federal funding, (2) a lack of adequate workplace skills, (3) China’s

world market presence, and (4) the US current account deficit to be the main

causes of the nation’s persistent manufacturing unemployment. These issues

are further examined in the analysis section, which applies the Heckscher-

Ohlin (H-O) and Stolper-Samuelson (S-S) theorems to address points (1) and

(2). The analysis proceeds to address points (3) and (4) by explaining how

China’s devalued currency is linked to America’s current account deficit,

which causes further unemployment in American manufacturing. Finally, the

conclusion explains how President Trump’s tariff plans are a risky,

unsustainable trading strategy. Instead of tariffs, several empirical-based

studies encourage increased federal funding toward manufacturing assistance

programs. Implementing production subsidies and consumption taxes is also

a reasonable policy that could grow employment in US manufacturing while

providing a stable source of revenue for the government.


The US manufacturing industry was described earlier as an essential

component to the nation’s private-sector innovation spending market. Despite

this recognition, most domestic manufacturing companies misallocate their

research and development (R&D) spending toward existing technology.45

45. Matthew Phillips and Peter Coy, "Look Who’s Driving R&D Now," Bloomberg, last

modified June 4, 2015, accessed May 9, 2017, https://www.bloomberg.com/news/articles/2015-06-04/look-who-s-driving-r-d-now.

71


This allocation occurs because many manufacturing companies (particularly

small- to medium-sized enterprises) lack adequate funding and face

disincentives to invest in innovative R&D; R&D is expensive, time

consuming, and it often fails. Many investors are therefore hesitant to make

an investment in these firms without assurance that their investment will yield

a return. Consequently, many manufacturing firms lack sufficient finances to

fund innovative R&D.46 This cycle of events is known as the “valley of

death,” and it summarizes the struggle manufacturing firms face when they

attempt to commercialize their products.47

Critics believe the “valley of death” occurs because private sector

R&D spending receives little-to-no federal support.48 According to

Bloomberg, federal R&D funding represented only 0.8 percent of the GDP in

2015 (appendix B contains further information of this data).49 There is also

not a federal program devoted to identifying emerging technologies with

broad potential impact in America’s manufacturing sector.50 What is more

surprising is that according to a recent poll done by the American Institute of

46. United States Congress Government Accountability Office (GAO), Global Manufacturing,

7. 47. Committee on Commerce, Science, and Transportation (United States Senate), The Role of

Manufacturing Hubs in a 21st Century Innovation Economy, 46. 48. Phillips and Coy, "Look Who’s," Bloomberg. 49. Ibid. 50. Committee on Commerce, Science, and Transportation (United States Senate), The Role of

Manufacturing Hubs in a 21st Century Innovation Economy, 15.

72


Physics in 2014, citizens ranked the scientific community (many of whom are

involved in manufacturing R&D) as the second most trusted community in

the US, ahead of medicine, education, and the US Supreme Court, all of

which receive substantial federal funding.51 Given the facts stated above, the

US manufacturing industry is struggling relative to the world because of

lackluster R&D funding by the federal government.

Another issue frequently brought up by literature is the lack of

adequate workplace skills of American manufacturing workers. In a 2012

podcast episode of NPR’s This American Life, reporter Jack Davidson stated

that technological changes in manufacturing caused a greater demand for

high-skilled workers and a smaller demand for low-skilled workers.52

Davidson also mentioned that hiring qualified workers is an extremely risky

process since manufacturing firms invest significantly in vocational training

without certainty that their hires will remain at the firm for a reasonable

amount of time.53 The opportunity cost of hiring, in other words, is extremely

high for manufacturing companies, and to reduce this dilemma,

manufacturing firms need outside public support. The Government

51. Mike Henry, U.S. Global Lead in R&D at Risk as China Rises, report no. 10 (Washington,

DC: American Institute of Physics (AIP), 2016), 7, accessed May 9, 2017, https://www.aip.org/fyi/2016/report-us-global-lead-rd-risk-china-rises.

52. Adam Davidson, "The Past and Future of American Manufacturing," This American Life,

podcast audio, January 10, 2012, accessed May 9, 2017, http://www.npr.org/sections/money/2012/01/10/144978487/the-tuesday-podcast-the-past-and-future-of-american-manufacturing.

53. Ibid.

73


Accountability Office (GAO) reveals that the US devolves vocational

(manufacturing) training to states and localities, but these local governments

often do not have sufficient funding.54 The GAO adds that the US does not

have a national system to issue industry-recognized credentials, which can

lead to adverse selection problems for companies since they have a harder

time distinguishing qualified workers from disqualified workers.55 Due to

hiring inefficiencies, high investment in workers, and lack of public support

for vocational training, American manufacturing companies waste time,

money, and energy that should be devoted toward more productive purposes,

like product development.

China’s presence in the world market is one of the most common

arguments for the loss of American manufacturing jobs. Critics argue that

China’s trading practices, such as currency manipulation, suppression of

labor rights, export subsidies, and import tariffs are causing a merchandise

trade deficit in the US current account, which contributes to the nation’s

manufacturing unemployment.56 According to a speech given to Congress by

Leo W. Gerard (President of United Steel Company), the US import

penetration rate (which measures the importance of imports in a domestic

54. United States Congress Government Accountability Office (GAO), Global Manufacturing,

37. 55. Ibid., 2. 56. Committee on Commerce, Science, and Transportation (United States Senate),

Manufacturing Our Way to a Stronger Economy, 17.

74


economy) has increased over the past decade. Gerard attributes much of this

increase to China’s presence in the world market.57

International economist Robert E. Scott takes a more thorough

approach to this subject as his analysis focuses on how America’s

manufactured goods trade deficit is associated with changes in output,

productivity, and employment in domestic manufacturing. He finds that

changes to the manufacturing trade deficit are the best determinant of changes

to the nation’s manufacturing output.58 Scott furthermore discovered that

these variables have an inverse relationship; so, the more rapidly America’s

manufacturing trade deficit grows, the slower domestic manufacturing output

gets produced. His study also shows a negative relationship between the

nation’s manufactured goods trade deficit and manufacturing employment,

and a positive relationship between the nation’s manufactured goods trade

deficit and manufacturing productivity.59 Appendix C provides a graphical

representation of these findings.

III. Analysis

57. Ibid., 15, 17. 58. Robert E. Scott, Manufacturing Job Loss: Trade, Not Productivity, Is the Culprit, issue brief

no. 402 (Washington, DC: Economic Policy Institute, 2015), 3, accessed May 9, 2017, http://www.epi.org/publication/manufacturing-job-loss-trade-not-productivity-is-the-culprit/.

59. Ibid., 3-4.

75


The Heckscher Ohlin (H-O) and Stolper-Samuelson (S-S) theorems

help explain why domestic manufacturing firms lack vocational programs and

federal funding. China’s pegged exchange rate is also analyzed and is shown

to contribute to the US current account deficit. These factors ultimately

contribute to unemployment in America’s manufacturing industry.

The H-O theorem states that countries export products that use their

abundant factor intensively and import products that use their scarce factor

intensively. According to The World Economy: Trade and Finance, the US

has historically been, and continues to be, a capital-abundant nation, while

countries, like China and India, have been predominantly labor-abundant

nations (appendix D contains this data).60 An empirical study by Devashish

Mitra, a Professor of Economics at Syracuse University, revealed that

countries with more stringent labor standards use higher capital-intensities in

their manufacturing production.61 Because of America’s capital abundance

and multitude of labor regulations, America is assumed to produce and export

capital-intensive goods while importing labor-intensive goods from more

labor-abundant nations. This concept intuitively makes sense when

considering the literature mentioned earlier—the US specializes in more

60. Beth V. Yarbrough and Robert M. Yarbrough, The World Economy: Trade and Finance, 7th

ed. (Mason, OH: Thomson South-Western, 2006), 65.

61. Rana Hasan, Devashish Mitra, and Asha Sundaram, "The Determinants of Capital Intensity in Manufacturing: The Role of Factor Endowments and Factor Market Imperfections," World Development 51 (November 2013): 91, doi:10.1016/j.worlddev.2013.05.012.

76


advanced, capital-intensive manufacturing and is abandoning more basic,

labor-intensive manufacturing.

The H-O theorem also applies to the occupational inequalities in

manufacturing described earlier. Since the US is a capital-abundant nation, it

will export capital-intensive (i.e., technology-oriented) goods, which demand

high-skilled, knowledgeable workers. Conversely, the US will import labor-

intensive goods that do not require much technical knowledge; so, low-skilled

workers are in lesser demand and can be characterized as America’s “scarce

factor” with regards to the H-O theorem. This lack of demand for low-skilled

(i.e. scarce factor) workers could explain why capital-intensive firms lack

motivation to create vocational programs for low-skilled workers. Even with

increased demand for high-skilled workers, domestic manufacturing firms

have limited hiring space, and they do not need many high-skilled workers to

manage and operate capital (due to the nation’s capital abundance). This lack

of job availability could explain why the nation’s manufacturing employment

has been in consistent decline for over 50 years. Domestic manufacturing

companies that are labor-intensive also suffer from this equation; since labor

is America’s scarce factor, America will import labor-intensive products,

which exposes domestic labor-intensive firms to greater international

competition.

The S-S theorem brings more clarity to the manufacturing

unemployment situation in the US as its definition expands on occupational

77


inequalities in American manufacturing, which could be linked to the

industry’s job loss and lack of vocational training and federal funding. The S-

S theorem states that if the price of a good increases (or decreases), then the

factor (labor L or capital K) used intensively in that good will be

unambiguously better off (or worse off). The increase (or decrease) in the

return will be higher than the increase (or decrease) in price. The other factor

is unambiguously worse off (or better off). Given the H-O theorem stated

earlier, the US is assumed to specialize in and export capital-intensive goods,

like robots. If world demand for US robots increases, then the price of robots

rises, which causes capital-owners to be unambiguously better off since their

return will be higher than the increase in price of robots. Meanwhile,

workers’ wages will be unambiguously worse off, and they earn a smaller

income as a result. This magnification effect could also explain why the US

has faced consistent unemployment in its manufacturing sector for over 50

years. As workers’ real wages continuously decline (because of increased

demand for capital-intensive manufactured products), their wages eventually

become unsustainable, and they are forced to find another industry that pays

better. Appendix E models this development.

The term “capital-owners” in the S-S theorem equates to a company’s

management team today since they are considered modern “owners” of

capital. If a rising price of a capital-intensive good makes managers (i.e.,

capital-owners) unambiguously better off (and workers worse off), then

78


company managers would be more inclined to stick with their current

production system and avoid creating new programs that might suck away

their growing incomes. Managers could therefore perceive vocational training

as costly to their incomes. This perception could explain why management

teams of manufacturing companies choose not to create new vocational

training programs, but rather, stick with their existing production schemes.

This perception could also explain why manufacturing firms receive little-to-

no federal support; perhaps, management teams do not what the government

to get involved and learn about their surging incomes.

The S-S theorem has strong empirical support; Paul S. Adler, an

international research economist at the Marshall School of Business at USC,

found that managerial employment in US manufacturing surged from 1975 to

1990 while more menial manufacturing jobs (like clerical and operative

work) declined.62 This finding proves the validity of the S-S theorem—that as

the prices of capital-intensive goods increase, capital owners (i.e., managers)

are unambiguously better off, while workers (i.e., operative and clerical

employees) are unambiguously worse off.

Much of the reviewed literature blamed China for the nation’s

manufacturing job-loss. China’s artificially low exchange rate has indeed

negatively impacted the US merchandise trade deficit, but their presence only

62. Paul S. Adler, Technology and the Future of Work (Oxford, U.K.: Oxford University Press,

1992), 61, 63-64, https://ebookcentral.proquest.com/lib/stolaf-ebooks/reader.action?docID=241618&ppg=61.

79


accounts for a small sliver of this deficit. Economists Galina Hale and Bart

Hobijn at the Federal Reserve Bank of San Francisco did a study on this

matter and found that imports from China amounted to only 2.5 percent of

US imports in 2010. They also found that 88.5 percent of domestic

consumption in the US goes toward “made in USA” products, while only 2.7

percent of domestic consumption is spent on “made in China” products.63

This study refutes the commonly heard claim that China is the primary reason

for America’s merchandise trade deficit and manufacturing job loss.

Yet, this study is not to say that China has no impact on American

manufacturing unemployment. Economic theory proves that China’s

devalued currency puts the US current account into greater deficit, which

ultimately contributes to America’s manufacturing job-loss. China’s central

bank maintains a pegged exchange rate above the flexible equilibrium rate to

make the yuan cheaper relative to the dollar. From the Chinese perspective,

the cost of converting yuan into dollar-denominated deposits is high, while

the cost of converting dollars into yuan-denominated deposits is low from the

US perspective. The exchange rate is therefore cheaper for US citizens, and

the price of US goods rises relative to Chinese goods. As the relative price of

US goods increases, US exports decrease (since US goods cost more to

import for foreigners), and US imports increase (since US citizens have more

63. Galina Hale and Bart Hobijn, "The U.S. Content of “Made in China”," Federal Reserve

Bank of San Francisco, last modified August 8, 2011, accessed May 9, 2017, http://www.frbsf.org/economic-research/publications/economic-letter/2011/august/us-made-in-china/.

80


purchasing power from a higher real exchange rate). As US exports decrease

and imports increase, the US loses GDP and falls into a current accounts

deficit. Shrinking GDP causes America’s manufacturing industry to be more

negatively affected; manufacturing firms will have to cut jobs and output to

adjust to the economic circumstances. Appendix F models these changes.

This pattern of events is supported by an empirical study from the

Government Accountability Office (GAO). The GAO analyzed the Trade

Adjustment Assistance (TAA) program, whose goal was to improve sales and

productivity of struggling domestic manufacturing firms. The GAO found

that as domestic imports rose, sales of TAA clients declined. For every 1

percentage point increase in the nation’s manufacturing import penetration

ratio (which measures the importance of imports to a nation or industry),

sales of manufacturing firms declined by about 16 percent on average.64 This

study closely resembles the series of events described in the paragraph

above—as US imports rise, domestic output decreases—evidenced by

declining sales of manufacturing firms from GAO’s analysis.

IV. Conclusion

What type of policies can reduce job loss and improve employment in

US manufacturing? The current president has proposed implementing tariffs

64. United States Congress Government Accountability Office (GAO), Trade Adjustment Assistance: Commerce Program Has Helped Manufacturing and Services Firms, but Measures, Data, and Funding Formula Could Improve, by J. Alfredo Gomez, report no. 12-930 (Washington, DC: Government Accountability Office, 2012), 23, accessed May 9, 2017, http://www.gao.gov/assets/650/648213.pdf

81


to solve this dilemma. This proposal, however, is an unreliable, risky plan,

and can hurt manufacturing employment in the long run. Instead of

implementing tariffs, much of the literature has proposed increasing

commercialization funding, providing outside vocational training, and

developing manufacturing “institutes” as a means of improving this situation.

Implementing production subsidies and consumption taxes also present a

realistic solution to improve manufacturing employment because they have

similar insulating effects of a tariff but do not put the nation’s trade

relationships in jeopardy.

President Trump has proposed implementing import tariffs against

China, Mexico, and several other countries to protect American

manufacturing firms from foreign competitors. Since the US economy

constitutes a large share of the world market, implementing tariffs do have

the potential to create short-term monetary gains for the nation (due to the

terms-of-trade effect shown in appendix G). This strategy, however, is far too

risky if the goal is to preserve and improve domestic manufacturing

employment. Imposing large import tariffs can lead to beggar-thy-neighbor

policies or risk retaliation by other countries, which would limit trade access

and increase trading costs. The closed-off, protectionist nature of tariffs is

also not a sustainable trading strategy since tariffs weaken trade relationships.

By attempting to specialize in comparative-disadvantage goods through

tariffs, the domestic economy and international economy is usually worse off

82


overall. The world has also learned from history that tariffs are not the most

effective trading strategy. In The World Economy: Trade and Finance, Beth

and Robert Yarbrough show that tariffs rates around the world have declined

over the past 18 years (from 1980 to 1998), and by 1990, the US tariffed only

13 products, all of which had a negative net effect on US welfare.65 Given

this data, President Trump’s tariff plans are not only an unrealistic long term

solution; they also have very slim chances of creating a positive net welfare

effect for the nation.

As opposed to implementing tariffs, empirical literature has called for

greater federal funding for new, practical manufacturing assistance programs.

The Government Accountability Office (GAO) did a fascinating study in

2013, analyzing how several foreign governments support their

manufacturing sectors. The study focused on Canada, Germany, South Korea,

and Japan, and learned that these governments 1) devote significantly more

funds to innovation and R&D, 2) offer localized government programs that

provide technical support and product development, 3) provide public

infrastructure and information systems, 4) support innovation “clusters”

(which are regional concentrations of large and small manufacturing firms),

and 5) have a sustained commitment to vocational training.66 The GAO

mentioned that the US has similar programs, but these programs were

65. Yarbrough and Yarbrough, The World, 149, 155-156. 66. United States Congress Government Accountability Office (GAO), Global Manufacturing,

8-9, 26-27, 29-30, 36.

83


characterized as too small, disorganized, underfunded, and limited in

focus.67 68

Another proposal brought up by literature is to create manufacturing

“institutes,” which are government-sponsored centers that allow

manufacturing firms, universities, national labs, and public organizations to

participate in collaborative research projects.69 US Secretary of Commerce

Penny Pritzker believes that “by using shared facilities, manufacturers can

pool their risk and drive down the cost of commercialization.”70 This idea

would inevitably expedite innovation and help manufacturing firms avoid

R&D pitfalls described earlier.

With benefits come costs, however, and implementing these changes

would require significant government spending. Government spending would

contribute to the nation’s existing fiscal deficit and perhaps affect the US

current accounts deficit, which would cause a twin deficit. The twin deficit is

further analyzed in appendix H.

Implementing production subsidies and consumption taxes is another

policy idea that could grow the nation’s manufacturing sector. While a tariff

raises consumer prices of foreign-produced goods, production subsidies and

67. United States Congress Government Accountability Office (GAO), Trade Adjustment, 36. 68. United States Congress Government Accountability Office (GAO), Global Manufacturing,

28-30, 37, 39. 69. Committee on Commerce, Science, and Transportation (United States Senate), The Role

of Manufacturing Hubs in a 21st Century Innovation Economy, 13-14. 70. Ibid., 13.

84


consumption taxes maintain low free trade prices and protect domestic

manufacturers via subsidies. Deadweight loss from this policy would be the

same amount as a tariff, but strong foreign trade relationships would prevail.

The government would also have a more stable stream of tax revenue to

finance its subsidies and would avoid plunging the nation into further fiscal

debt and current accounts deficit. This proposal largely reflects many of the

empirical-based solutions presented earlier—government funding toward

R&D, for example, largely embodies a production subsidy. Consumption

taxes were not mentioned in the literature, but implementing them would

inevitably improve the government’s financial stability. Appendix I models

this proposal.

Many of today’s news media outlets associate trade with

competition—who can produce the most output, who can export the most,

and who has the strongest currency. International economic theory, however,

reveals that trade does not necessarily resemble competition, but rather

cooperation. Yet engaging in trade can produce industry inequalities and job

losses within a country, such as America’s current manufacturing sector. The

H-O and S-S theorems shed light on this idea by explaining how America’s

capital-abundance causes domestic labor-intensive firms to be more exposed

to international competition and how capital-owners are unambiguously

better off than workers when the price of a capital-intensive good rises.

Perhaps, these theorems have greater implications for the longer-run

85


outcomes of US manufacturing; this hypothesis was not addressed in this

paper and needs to be further examined. In the short run, however, a

deteriorating domestic manufacturing industry calls for greater support by the

federal government to strengthen the sector, stabilize the nation’s economy,

and improve domestic and international welfare overall.

86


V. Appendices Appendix A

Fig. 1: Shows decline in US manufacturing employment over 56 years. Manufacturing composed roughly 30 percent of nonfarm nationwide

employment in 1956, dropping to less than 10 percent in 2006. Marlene A. Lee and Mark Mather, Share of Nonfarm Employment by Major Industrial

Sector, 1950-2007. 2008, graphical chart. Source: Bureau of Labor Statistics, Current Employment Statistics. Available from: Population Bulletin,

http://www.prb.org/pdf08/63.2uslabor.pdf

87


Appendix B

Fig. 2: Shows changes in federal R&D spending as a share of domestic GDP. Matthew Phillips and Peter Coy, Look Who’s Driving R&D Now, timeline chart. Source: Bloomberg, https://www.bloomberg.com/news/articles/2015-

06-04/look-who-s-driving-r-d-now Fig. 3: Examines business and other nonfederal R&D spending versus federal government R&D spending. Mike Henry, Federal R&D Support—Ratio of U.S. R&D to GDP, by funders: 1953-2013, graphical

chart. Source: American Institute of Physics (AIP), https://www.aip.org/fyi/2016/report-us-global-lead-rd-risk-china-rises

88


Fig. 4: Compares US R&D spending as a percentage of GDP with other developed nations. Mike Henry, R&D: Intensity (R&D/GDP), 1996-2013, graphical chart. Source: American Institute of Physics (AIP), https://www.aip.org/fyi/2016/report-us-global-lead-rd-risk-china-rises

89


Appendix C

Fig. 5: Graphs changes of US manufacturing goods trade deficit over 25 years. Provides absolute and relative values (relative values are expressed as a percentage of GDP). Robert E. Scott, U.S. manufacturing goods trade deficit, 1989-2014, graphical chart. Source: Bureau of Economic Analysis, Available from: Economic Policy Institute (EPI), http://www.epi.org/publication/manufacturing-job-loss-trade-not-productivity-is-the-culprit/

90


Fig. 6: Exhibits relationships between changes in US manufactured goods trade deficit (shown in figure 5 above) and US manufacturing output, productivity, and employment over time. The nation’s manufacturing trade deficit has a negative, positive, and negative relationship with domestic output, productivity, and employment respectively. Robert E. Scott, Average annual change in output, productivity, and employment growth in US manufacturing, 1989-2014, graphical chart. Source: Bureau of Labor Statistics, Available from: Economic Policy Institute (EPI), http://www.epi.org/publication/manufacturing-job-loss-trade-not-productivity-is-the-culprit/

Appendix D

Fig. 7: Illustrates factor endowment differences (i.e., labor abundance versus capital abundance) around the world. Countries with higher capital-per-worker endowments are more capital-abundant; countries with lower capital-per-worker endowments are more labor-abundant. Beth V. Yarbrough and Robert M. Yarbrough, Capital-Per-Worker Endowments, 1990, data table. Source: American Economic Review, Available from: The World Economy: Trade and Finance. Mason, OH, 2006. 65.

91


Appendix E

Fig. 8: Given the fact that US manufacturing is capital-abundant, the manufacturing industry will specialize in producing and exporting capital-intensive goods and will import labor-intensive goods (because of the H-O theorem). When the price of a capital-intensive good rises, there is an initial surge in demand for capital, and the demand for capital increases from DK

0 to DK

1. There is also decreased production of other goods (that use a smaller amount of capital) to compensate for the increased demand of capital for this specific good. The capital demand curve thus shifts down slightly from DK

1 to DK

2. The change in capital demand reflects the rental rate of capital owners, who experience a net increase in their rental rate income (r). Capital owners are unambiguously better off than the increase in price of this good.

92


Fig. 9: When the price of a capital-intensive good rises for an American manufacturing firm, there is decreased production of labor-intensive goods. This decrease is shown as a shift of the labor demand curve from DL

0 to DL1.

Increased production of capital-intensive goods would require some labor to operate the equipment and machinery; so, there is a slight upward shift in the labor demand curve, explaining the movement from DL

1 to DL2. These

changes in labor demand reflect the wages of workers, who experience a net decrease in wages (w) and are unambiguously worse off than the increase in price of this good.

Appendix F

Fig. 10: Shows China’s foreign exchange market for the dollar. China’s central bank maintains a pegged exchange rate above the equilibrium exchange rate by absorbing excess dollars and selling yuan to the public. This fixed exchange rate allows the yuan to remain artificially depreciated (relative to the dollar) by remaining above the equilibrium exchange rate of a flexible regime.

93


Fig. 11: Shows US foreign exchange market for the yuan. As the yuan depreciates (increasing the denominator value of the US exchange rate, e), the US exchange rate decreases overall, and the dollar appreciates. This dollar appreciation is characterized by a downward shift of US demand for yuan (from D¥

0 to D¥1). The US operates under a flexible exchange rate regime.

94


Fig. 12: As the US nominal exchange rate (e) decreases, its real exchange rate (R) increases, allowing the US to have more purchasing power over other nations’ goods. US goods also become more expensive to other nations. The increase in R causes US imports to rise (due to increased purchasing power), and US exports to fall (due to more expensive US goods). Higher imports and lower exports decrease the nation’s GDP by shifting the aggregate expenditure curve down in the US expenditure market from Y0 to Y1. The decrease in GDP is directly related to the US current account balance (shown in graph below the US expenditure market)—the import curve shifts up, and the export curve shifts down. Lower GDP from increased imports and decreased exports translates into a current account deficit since US imports exceed exports. Domestic manufacturing firms will have to cut output and

95


jobs since they will be exporting less (earning less revenue) and are more exposed to international competition (because of the US’ importing more foreign-produced goods).

Appendix G

Fig. 13: Illustrates the effects of an import tariff policy on the US economy. Since the US economy constitutes a large share of the world market for manufactured goods, the shape of its supply curve influences the shape of the total (world) supply curve. If the US government is considering implementing a tariff for a specific manufactured good Y, then its domestic producers must be at a comparative disadvantage in producing good Y. This comparative disadvantage explains why good Y is consumed at a greater quantity (Y0) than it is supplied by domestic producers (Y1). The space between point C and I (Y0 – Y1) represents the amount of US imports for good Y. Domestic producers must compete with foreign producers, which explains why domestic firms sell good Y for the same low price as foreign firms (at PY

0). If the US imposes a tariff, domestic consumers must pay a higher price for good Y and the total (world) supply curve shifts up (to Sd+w+t). The new price of good Y is PY

1, and a higher price decreases the amount of good Y consumed to Y2 (from Y0). Domestic firms, on the other hand, are more protected under the tariff law and enjoy greater profits from the increased price of good Y. Quantity supplied thus increases to Y3, and total US imports shrink to Y2 – Y3. Area “a” represents a direct transfer of consumer surplus to producer surplus, as domestic firms enjoy greater profits and protection from the tariff law, while consumers suffer from an increased price of good Y. Area “b” represents a deadweight loss due to domestic production inefficiencies—

96


domestic manufacturers try to increase production in a comparative-disadvantage good. Area “c” represents a transfer from consumer surplus to domestic government—a tariff on good Y forces consumers to pay an extra tax for this good, and the tax goes directly to the federal government. Area “d” represents another deadweight loss due to consumption inefficiencies—consumers can no longer obtain Y0 units due to increased price from the tariff, PY

1. Lastly, area e represents the terms-of-trade effect—since the US market of good Y composes a large share of the world economy; a tariff decreases the domestic quantity demanded of good Y, which could force world suppliers to accept a lower minimum price of selling this good (represented by PY

2). If the domestic deadweight loss (areas b+d) is smaller than the profits lost by foreign producers (represented by area e), then the US would experience a positive net welfare effect, while the world would suffer a negative net welfare effect. The chances of the US experiencing a positive welfare effect are very rare, however, and the world is always worse off when a tariff is implemented.71

Appendix H

Fig. 14: For the federal government to increase its spending toward domestic manufacturing programs, it would have to issue bonds to finance its spending, so the domestic supply of bonds increases. Increased bond supply is illustrated by shifting the bond supply curve to the right. This shift lowers the price of bonds, which increases demand for them.

71. Yarbrough and Yarbrough, The World, 159.

97


Fig. 15: The federal government spends money toward domestic manufacturing programs, which increases domestic output but increases the fiscal deficit. After the government issues bonds to finance its spending, bond supply expands, which increases demand for bonds. Increased demand for bonds correspond to increased interest rates, which is illustrated by a rightward shift of the IS curve on the IS-LM market above. Increased government spending also influences the rightward shift of the IS curve.

98


Fig. 16: Due to increased interest rates (modelled in figure 15), domestic demand for foreign currency decreases. The demand curve for foreign currency (such as British pounds), shifts down. This downward shift lowers the US nominal exchange rate, and the dollar appreciates.

Fig. 17: Nominal exchange rates decrease for US citizens (shown in figure 16), which increases the US real exchange rate, R. A higher real exchange rate means more imports and less exports for the US, which lowers the nation’s GDP and causes a current account deficit (these changes are explained more thoroughly in figure 12). Increased government spending (from figure 15) contributes to the nation’s fiscal deficit and can further impact the nation’s current account deficit (shown above), which ultimately leads to a twin deficit.

99


Appendix I

Fig. 18: Due to the complexity of modeling a production subsidy and consumption tax on a large-economy welfare model, the small-economy welfare model is used. By not implementing a tariff, world prices remain low and constant for producers and consumers (at PY

0). Consumers initially consume Y0 units of good Y for a price of PY

0 (at point C), and US manufacturers produce Y1 units of good Y and sell this good at a price of PY

0 (at point P). Y0 – Y1 is the amount of good Y the US imports before a production subsidy and consumption tax is implemented. When a production subsidy implemented, the government provides more funding to suppliers, which allows them to produce more of good Y; production increases to point M (Y1 to Y2). When a consumption tax is implemented, consumers obtain less of good Y, and domestic consumption drops to point N; consumers now consume Y3 units of good Y. Area “b” represents production inefficiencies that arise from manufacturers attempting to produce more of good Y, which is assumed to be a comparative-disadvantage good. Area “d” represents consumption inefficiencies that arises from consumers not being able to obtain Y0 units of good Y because of a consumption tax. Area “b+d” represents an overlap of production and consumption inefficiencies. Implementing this policy does decrease US imports (to Y3-Y2), but world prices remain the same. A production subsidy and consumption tax equates to the same deadweight loss as a tariff, but this policy does not put trade relationships in jeopardy. Low world prices are maintained and domestic producers (i.e., US manufacturing firms) receive consistent funding from the

100


government. The government also enjoys a stable source of financing from consumption taxes. VI. References Adler, Paul S. Technology and the Future of Work. Oxford, U.K.: Oxford

University Press, 1992. https://ebookcentral.proquest.com/lib/stolaf-ebooks/reader.action?docID=241618&ppg=61.

Atack, Jeremy, and Fred Bateman. "Manufacturing." In Economic Sectors, edited by Susan B. Carter, Scott Sigmund Gartner, Michael R. Haines, Alan L. Olmstead, Richard Sutch, and Gavin Wright, 573-592. Millennial ed. Vol. 4 of Historical Statistics of the United States. New York, NY: Cambridge University Press, 2006.

Committee on Banking, Housing, and Urban Affairs (United States Senate), The United States as a Global Competitor: What are the Elements of a National Manufacturing Strategy?, S. Rep. No. 111-281, 1st Sess., at 1 (2009) (Conf. Rep.). Accessed May 9, 2017. https://www.gpo.gov/fdsys/pkg/CHRG-111shrg55018/pdf/CHRG-111shrg55018.pdf.

Committee on Commerce, Science, and Transportation (United States Senate), Manufacturing Our Way to a Stronger Economy, S. Rep. No. 112-293, 1st Sess., at 1 (2011) (Conf. Rep.). Accessed May 9, 2017. https://www.gpo.gov/fdsys/pkg/CHRG-112shrg73231/pdf/CHRG-112shrg73231.pdf.

Committee on Commerce, Science, and Transportation (United States Senate), The Role of Manufacturing Hubs in a 21st Century Innovation Economy, S. Rep. No. 113-664, 1st Sess., at 1 (2013) (Conf. Rep.). Accessed May 9, 2017. https://www.gpo.gov/fdsys/pkg/CHRG-113shrg95215/pdf/CHRG-113shrg95215.pdf.

Davidson, Adam. "The Past and Future of American Manufacturing." This American Life. Podcast audio. January 10, 2012. Accessed May 9, 2017. http://www.npr.org/sections/money/2012/01/10/144978487/the-tuesday-podcast-the-past-and-future-of-american-manufacturing.

Hale, Galina, and Bart Hobijn. "The U.S. Content of “Made in China”." Federal Reserve Bank of San Francisco. Last modified August 8, 2011. Accessed May 9, 2017. http://www.frbsf.org/economic-research/publications/economic-letter/2011/august/us-made-in-china/.

Hasan, Rana, Devashish Mitra, and Asha Sundaram. "The Determinants of Capital Intensity in Manufacturing: The Role of Factor Endowments and Factor Market Imperfections." World Development 51 (November 2013): 91-103. doi:10.1016/j.worlddev.2013.05.012.

Henry, Mike. U.S. Global Lead in R&D at Risk as China Rises. Report no. 10. Washington, DC: American Institute of Physics (AIP), 2016.

101


Accessed May 9, 2017. https://www.aip.org/fyi/2016/report-us-global-lead-rd-risk-china-rises.

Lee, Marlene A., and Mark Mather. "U.S. Labor Force Trends." Population Bulletin 63, no. 2 (June 2008): 1-15. Accessed May 9, 2017. http://www.prb.org/pdf08/63.2uslabor.pdf.

Phillips, Matthew, and Peter Coy. "Look Who’s Driving R&D Now." Bloomberg. Last modified June 4, 2015. Accessed May 9, 2017. https://www.bloomberg.com/news/articles/2015-06-04/look-who-s-driving-r-d-now.

Scott, Robert E. Manufacturing Job Loss: Trade, Not Productivity, Is the Culprit. Issue brief no. 402. Washington, DC: Economic Policy Institute, 2015. Accessed May 9, 2017. http://www.epi.org/publication/manufacturing-job-loss-trade-not-productivity-is-the-culprit/.

United States Congress Government Accountability Office (GAO). Global Manufacturing: Foreign Government Programs Differ in Some Key Respects from Those in the United States. By Andrew Sherrill and Lawrence Evans. Report no. 13-365. Washington, DC: Government Accountability Office, 2013. Accessed May 9, 2017. http://www.gao.gov/assets/660/656239.pdf.

United States Congress Government Accountability Office (GAO). Trade Adjustment Assistance: Commerce Program Has Helped Manufacturing and Services Firms, but Measures, Data, and Funding Formula Could Improve. By J. Alfredo Gomez. Report no. 12-930. Washington, DC: Government Accountability Office, 2012. Accessed May 9, 2017. http://www.gao.gov/assets/650/648213.pdf.

U.S. Bureau of Labor Statistics. "Employment, Hours, and Earnings from the Current Employment Statistics survey (National)." U.S. Bureau of Labor Statistics. Last modified 2017. Accessed April 5, 2017. https://data.bls.gov/pdq/SurveyOutputServlet.

Yarbrough, Beth V., and Robert M. Yarbrough. The World Economy: Trade and Finance. 7th ed. Mason, OH: Thomson South-Western, 2006.

102


Omicron Delta Epsilon St. Olaf College: Beta Chapter

Class of 2017 Alexander Aristides

Erik Berthelsen Holly Jansen Elliot Knuths

Katherine Kroening Matthew Lasnier

Evan Lebo Kelsey Myers

Class of 2018 Matthew Damhof

Erik Davidson-Schwartz Spencer Knack

Robert Leet Lindsay Mattei

Alexandra Rosati Maren Weaver

103

Omicron Delta Epsilon

Journal of Economic Research St. Olaf College: Beta Chapter

Executive Editor Kelsey Myers

Associate Editor Erik Davidson-Schwartz

Spring 2016 Papers Maija Johansen Chloe Mitchell

Abigail Schnaith Aaron Stets

104

.

spring 2017 senior distinction papers-class of 2013 · bitcoin. nearly ten years since the first...

Documents