eerc oilcurse _fin_alexeev

The “Oil Curse” and Institutions: A Brief Survey

by

Michael Alexeev

2

NATURAL RESOURCE CURSE CLAIM: LARGE ENDOWMENT OF OIL AND MINREALS REDUCES LONG-TERM GROWTH RATES MAIN TRANSMISSION MECHANISM STRESSED IN RECENT LITERATURE: DELETERIOUS EFFECT OF NATURAL RESOURCES ON INSTITUTIONS, INCL. FORM OF GOVERNMENT MAIN CULPRIT: “POINT-SOURCE” RESOURCES

SUCH AS OIL

3

BASIC LOGIC OF THE NATURAL RESOURCE CURSE: NATURAL RESOURCES GENERATE RENTS MORE RENT MORE EFFORTS TO OBTAIN THAT RENT (I.E., RENT-SEEKING EFFORTS) LESS EFFORT IS DEVOTED TO NON-RENT-SEEKING ACTIVITIES IF THE LATTER PRODUCE POSITIVE EXTERNALITIES, ECONOMIC GROWTH DECLINES.

4

AN EXAMPLE OF A FORMAL MODEL: MEHLUM ET AL. (EJ, 2006) Mehlum, Halvor, Karl Moene, and Ragnar Torvik, “Institutions and the Resource Curse,” Economic Journal 116:508:1–20. Mehlum et al.’s model sketch: There is a fixed number of entrepreneurs in the economy, 𝑁𝑁 = 𝑛𝑛𝑃𝑃 + 𝑛𝑛𝐺𝐺, 𝑛𝑛𝑃𝑃 = 𝛼𝛼𝑁𝑁, with each choosing between being a “producer” or a “grabber.” Total rent = 𝑅𝑅. Grabber’s payoff = 𝜋𝜋𝐺𝐺 = 𝑠𝑠𝑅𝑅/𝑁𝑁; producer gets 𝜆𝜆𝑠𝑠𝑅𝑅/𝑁𝑁 of rent, where 𝜆𝜆ϵ[0,1] is “institutional quality” (i.e., grabbers get ≥ rent than producers) and 𝜕𝜕𝜕𝜕𝜕𝜕𝑛𝑛𝐺𝐺

< 0. Two production technologies: “modern” (IRS with fixed costs, low MC, positive profit) and “fringe” (no fixed costs, high MC, zero profit). Labor = 𝐿𝐿,𝑤𝑤 = 1. Producer’s production profit, 𝜋𝜋(𝑛𝑛𝑃𝑃), increases in 𝑛𝑛𝑃𝑃 (b/c total income and demand ↑). Producer’s total profit is 𝜋𝜋𝑃𝑃 = 𝜋𝜋(𝑛𝑛𝑃𝑃) + 𝜆𝜆𝜋𝜋𝐺𝐺. If 𝜆𝜆 is low and 𝑅𝑅 is high, there are grabbers in equilibrium where 𝜋𝜋𝑃𝑃 = 𝜋𝜋𝐺𝐺 implying that 𝜋𝜋𝑃𝑃(1 − 𝜆𝜆) = 𝜋𝜋(𝑛𝑛𝑃𝑃), and total income, 𝑌𝑌 = 𝑁𝑁𝑁𝑁(𝑛𝑛𝑃𝑃)

1−𝜆𝜆+ 𝐿𝐿 → 𝑅𝑅 ↑ implies that 𝑛𝑛𝐺𝐺 ↑ and 𝑛𝑛𝑃𝑃 ↓

→ 𝑌𝑌 ↓ , i.e., there is “resource curse.” It occurs because 𝑅𝑅 ↑ shifts some producers into rent grabbing and reduces the positive externality generated by producers. Here the conclusion is that only countries with weak institutions (low 𝜆𝜆) suffer from a resource curse.

5

ANOTHER STANDARD ARGUMENT FOR THE DELETERIOUS EFFECT OF NATURAL RESOURCES ON INSTITUTIONS: NATURAL RESOURCE WEALTH MEANS THAT THE GOVERNMENT DOES NOT NEED MUCH TAX REVENUE AND HAS NO INTEREST IN MAINTAINING GOOD PRODUCTIVE INSTITUTIONS.

6

THE “ANTI-CURSE” ARGUMENT IS THAT NATURAL RESOURCES BRING IN THE POSSIBILITY FOR THE GOVERNMENT TO IMPROVE ITS FUNCTIONS WITHOUT TAXING PEOPLE AND BUSINESSES TOO MUCH. THIS IS PARTICULARLY IMPORTANT IN POOR COUNTRIES THAT MAY NOT EVEN HAVE THE ADMINISTRATIVE CAPACITY TO TAX A COUNTRY, PARTICULARLY A RELATIVELY POOR COUNTRY, CAN IMPROVE ITS INSTITUTIONS WITHOUT HAMPERING THE GROWTH OF THE PRIVATE SECTOR.

7

OTHER STORIES: DUTCH DISEASE CIVIL CONFLICT AUTOCRACY

8

MAIN EARLY EMPIRICAL PAPERS ON THE OIL CURSE: Sachs and Warner (1995, 2001) claims: “… a statistically significant, inverse, and robust association between natural

resource intensity and growth over the past twenty years.” “[e]mpirical support for the curse of natural resources is not bulletproof, but it is

quite strong” Sala-i-Martin and Subramanian (2003) “[s]ome natural resources – oil and minerals in particular – exert a negative and non-linear impact on growth via their deleterious impact on institutional quality.”

9

Gylfason, T., and G. Zoega (2002a, b): “[e]conomic growth is inversely related to natural resource dependence…” M. Ross (2001): “…oil and mineral wealth tends to make states less democratic” Isham, et al. (2003) Leite and Weidmann (1999) and many others

10

REGRESSIONS OF INSTITUTIONAL QUALITY (RULE OF LAW, CORRUPTION, ETC.) ON NATURAL RESOURCE ENDOWMENT

𝑰𝑰𝑰𝑰𝒊𝒊,𝒕𝒕 = 𝝁𝝁𝟎𝟎 + �𝝁𝝁𝒊𝒊

𝒊𝒊

𝑿𝑿𝒊𝒊 + 𝝑𝝑𝟏𝟏𝒀𝒀𝒊𝒊,𝒕𝒕−𝟐𝟐𝟎𝟎 + 𝝑𝝑𝟐𝟐𝑵𝑵𝒊𝒊 + 𝝎𝝎𝒊𝒊

WHERE 𝑰𝑰𝑰𝑰𝒊𝒊,𝒕𝒕 -- INSTITUTIONAL QUALITY; 𝑵𝑵𝒊𝒊 -- NATURAL RESOURCE ABUNDANCE OR A MEASURE OF RENTS; 𝒀𝒀𝒊𝒊,𝒕𝒕−𝟐𝟐𝟎𝟎 -- “INITIAL” GDP PER CAPITA; 𝑿𝑿𝒊𝒊 -- CONTROLS

11

PROBLEMATIC ISSUES: - “INITIAL GDP” AS A CONTROL VARIABLE - MEASURES OF NATURAL RESOURCES (GDP SHARES

VS. PER CAPITA; EXPORTS/GDP; INCLUSION OF RENTS )

- ENDOGENEITIES (BOTH DUE TO POTENTIAL

REVERSE CAUSALITY AND OMITTED VARIABLES)

12

ALEXEEV AND CONRAD FIRST SHOW THAT PER CAPITA GDP IS STRONGLY POSITIVELY CORRELATED WITH NATURAL RESPURCE ENDOWMENT:

𝒀𝒀𝒊𝒊,𝟐𝟐𝟎𝟎𝟎𝟎𝟎𝟎 = β𝟎𝟎 + �𝜷𝜷𝒊𝒊

𝒊𝒊

𝑿𝑿𝒊𝒊 + γ𝑵𝑵𝒊𝒊 + ε𝒊𝒊

13

Table 1. Effect of Oil Wealth on PC GDP (Dep. Var.: Log of PC PPP GDP in 2000; Large sample)

Variable 1 2 3 4 5 6

Hydrocarbon deposits, per capita

.059*** (.016)

.051*** (.010)

Value of oil output, per capita

.096*** (.023)

.086*** (.015)

Oil/GDP ratio 1.51** (.693)

1.26*** (.313)

Absolute latitude .037*** (.005)

.038*** (.005)

.038*** (.005)

Ethnolinguistic fractionalization

-170 (.232)

-.455* (.250)

-.436* (.262)

European population 1.34*** (.202)

1.30*** (.202)

1.43*** (.208)

-.054 (.308)

.097 (.286)

.066 (.322)

Latin America 1.02*** (.155)

.926*** (.154)

1.06*** (.171)

.814*** (.134)

.662*** (.135)

.774*** (.151)

East Asia 1.70*** (.334)

1.67*** (.290)

1.77*** (.269)

.572*** (.195)

.594*** (.172)

.618*** (.213)

Rule of law (Instrumented)

1.14*** (.150)

1.02*** (.150)

1.09*** (.165)

Observations 111 118 118 111 117 117 Adj. R-squared .739 .725 .708 .862 .869 .844

Notes: *** - significant at 1%; ** - at 2%; * - at 10%; Instruments for Rule of law: absolute latitude, English language speakers; European language speakers

14

Contrast the above results with the following statement

by Sachs and Warner (2001):

“casual observation suggests that there is virtually no overlap in the set of countries that have large natural resource endowments – and the set of countries that have high levels of GDP.”

15

Table 2. The Effect of Interaction between Institutions and Natural Resources on PC GDP

(large sample; dependent variable: Logarithm of per capita PPP GDP in 2000) Variable 1 2 3 4 5


.056*** (.014)


.101*** (.021)

Oil/GDP ratio 1.14* (.678)

Mining/GDP ratio 1.64 (1.14)

Mining output, per capita

.072** (.028)

Rule of law (fitted values)

1.14*** (.189)

1.08*** (.202)

1.15*** (.218)

1.13*** (.238)

1.19*** (.236)

(Rule of law)* Natural resources

-.041*** (.012)

-.065*** (.017)

-.834 (.913)

-1.53 (2.20)

-.053*** (.019)

Observations 111 117 117 117 117 Adj. R-squared .783 .770 .754 .745 .756

Notes: *** - significant at 1% level; ** - significant at 2% level; * - significant at 10% level.

16

THE EFFECT OF MINERAL WEALTH ON INSTITUTIONS

EFFECT OF “INITIAL GDP” AS A CONTROL VARIABLE: OIL RAISES GDP IN 1970 KUWAIT HAD MUCH GREATER

PER CAPITA GDP THAN FRANCE, BUT HAD CONSIDERABLY

WORSE INSTITUTIONS THAN FRANCE CONTROLLING

FOR INITIAL GDP LEADS TO A NEGATIVE COEFFICIENT

FOR OIL

17

COMPARISON OF FRANCE AND KUWAIT IN TERMS OF PER CAPITA GDP AND INSTITUTIONAL QUALITY

GDP PER

CAPITA 1970

RULE OF LAW

1998

CONTROL OF CORRUPTION

1998

GOVERNMENT EFFECTIVENESS

1998 FRANCE 12,115 1.44

(0.18) 1.75

(0.18) 1.64

(0.23) KUWAIT 41,357 1.16

(0.24) 1.07

(0.18) −0.01 (0.31)

Note: standard errors are in parentheses

18

ALEXEEV AND CONRAD’S APPROACH: FIRST, ESTIMATE 𝒀𝒀�:

𝒀𝒀𝒊𝒊,𝟏𝟏𝟏𝟏𝟏𝟏𝟎𝟎 = β𝟎𝟎 + �𝜷𝜷𝒊𝒊𝒊𝒊

𝑿𝑿𝒊𝒊 + 𝒖𝒖𝒊𝒊 → 𝒀𝒀� = 𝒃𝒃𝟎𝟎 + �𝒃𝒃𝒊𝒊𝒊𝒊

𝑿𝑿𝒊𝒊

WHERE 𝑿𝑿𝒊𝒊 ARE ABSOLUTE LATITUDE, EUROPE, LATIN AMERICA, EAST

ASIA

𝒀𝒀� = 𝟔𝟔.𝟖𝟖 + 𝟎𝟎.𝟎𝟎𝟎𝟎 × 𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨 + 𝟏𝟏.𝟏𝟏𝟏𝟏 × 𝑬𝑬𝑬𝑬𝑬𝑬𝑬𝑬𝑬𝑬𝑬𝑬 + 𝟎𝟎.𝟖𝟖𝟎𝟎 × 𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑳𝑳 + 𝟎𝟎.𝟒𝟒𝟏𝟏 × 𝑬𝑬𝑨𝑨𝑨𝑨𝑨𝑨

Adj. 𝑅𝑅2 = 0.55; No. obs. = 118

NOTE THAT 𝐶𝐶𝐶𝐶𝑅𝑅𝑅𝑅(𝑁𝑁𝑖𝑖 ,𝑌𝑌𝑖𝑖 ) >> 𝐶𝐶𝐶𝐶𝑅𝑅𝑅𝑅(𝑁𝑁𝑖𝑖 ,𝑌𝑌�) ≈ 0

19

THEN, CONTRAST

𝑰𝑰𝑰𝑰𝒊𝒊,𝟐𝟐𝟎𝟎𝟎𝟎𝟎𝟎 = 𝝁𝝁𝟎𝟎 + �𝝁𝝁𝒊𝒊𝒊𝒊

𝑿𝑿𝒊𝒊 + 𝝑𝝑𝟏𝟏𝒀𝒀𝒊𝒊,𝟏𝟏𝟏𝟏𝟏𝟏𝟎𝟎 + 𝝑𝝑𝟐𝟐𝑵𝑵𝒊𝒊 + 𝝎𝝎𝒊𝒊

AND

𝑰𝑰𝑰𝑰𝒊𝒊,𝟐𝟐𝟎𝟎𝟎𝟎𝟎𝟎 = 𝝁𝝁𝟎𝟎 + �𝝁𝝁𝒊𝒊𝒊𝒊

𝑿𝑿𝒊𝒊 + 𝝑𝝑𝟏𝟏𝒀𝒀� + 𝝑𝝑𝟐𝟐𝑵𝑵𝒊𝒊 + 𝝎𝝎𝒊𝒊

20

Table 3. The Effect of Oil Wealth on the Rule of Law (ethnic and linguistic controls only; large sample) (Dependent variable: Rule of law index for year 2000)

Variable 1 2 3 4 5 6


-.042** (.017)

-.005 (.017)


-.068** (.026)

-.012 (.027)

Oil/GDP ratio -1.33** (.536)

.003 (.666)

GDP 1970, per capita

.570*** (.090)

.552*** (.091)

.569*** (.103)

GDP 1970, per capita, fitted

.508*** (.173)

.466*** (.169)

.473*** (.160)

Observations 112 112 118 118 118 118 Adj. R-squared .717 .623 .710 .617 .712 .616

Notes: robust standard errors are in parentheses; *** - significant at 1% level; ** - significant at 2% level; * - significant at 10% level

21

Table 4. The Effect of Oil and Mineral Wealth on Corruption (Dep. var.: Index of control of corruption, 2000)

Variable 1 2 3 4 5 6 7 8 9 10

Hydrocarbons, PC

-.014 (.015)

.014 (.013)

Value of oil output, PC

-.060** (.023)

-.008 (.019)

Oil/GDP ratio -1.12** (.442)

.111 (.363)

Mining/GDP ratio

-.554 (1.25)

.814 (.829)

Mining output, per capita

-.022 (.027)

.019 (.025)

PC GDP 1970 .498**

* (.100)

.578***

(.088)

.579***

(.093)

.517***

(.103)

.528***

(.099)

PC GDP 1970, fitted

.765***

(.167)

.800***

(.153)

.804***

(.154)

.814***

(.155)

.803***

(.157) Observations 110 110 113 113 113 113 113 113 113 113 Adj. 𝑅𝑅2 .732 .756 .738 .740 .734 .740 .720 .743 .721 .742

Notes: *** - significant at 1% level; ** - significant at 2% level; * - significant at 10% level.

COMPARISON OF BELARUS, RUSSIA, AND UKRAINE IN TERMS OF PER CAPITA GDP AND INSTITUTIONAL QUALITY

PPP GDP

per capita, 2012 US$

Rule of law

2012

Control of corruption

2012

Government effectiveness

2012 BELARUS 15,327 −0.92

(0.13) −0.52 (0.13)

−0.94 (0.22)

RUSSIA 23,589 −0.82 (0.12)

−1.01 (0.12)

−0.43 (0.19)

UKRAINE 7,298 −0.79 (0.12)

−1.03 (0.12)

−0.58 (0.19)

23

A&C CONCLUSIONS: GIVEN THE DATA AVAILABLE SO FAR, NATURAL RESOURCES, INCLUDING OIL: - DO NOT REDUCE LONG-TERM GROWTH RATES - DO NOT MAKE INSTITUTIONAL QUALITY SIGNIFICANTLY WORSE HOWEVER, A&C DID NOT LOOK AT VOICE AND ACCOUNTABILITY INDEX

24

THE OIL CURSE AND THE ECONOMIES IN TRANSITION

MOST ECONOMIES IN TRANSITION EXPERIENCE A DRASTIC CHANGE IN

INSTITUTIONS AND THUS PRESENT AN OPPORTUNITY TO GET AROUND

THE PROBLEM OF INSTITUTIONAL PERSISTENCE

SIMILAR APPROACH AS DESCRIBED EARLIER BUT WITH A DUMMY

VARIABLE FOR ECONOMIES IN TRANSITION SHOWS THAT IT IS MOSTLY

NOT TRUE.

25

Table 5. Institutional quality regressions (reduced form; OLS) Dependent variables: Measures of institutional quality for 2005

Variable name Corr. control

Corr. control

Rule of law

Rule of law

Voice & account.

Voice & account.

(1) (2) (3) (4) (5) (6) Transition economy -1.11***

(.159) -1.13***

(.168) -.968***

(.166) -.991***

(.176) -.443** (.145)

-.449*** (.150)

CIS -.301 (.232)

-.246 (.233)

-.341 (.228)

-.329 (.221)

-.584* (.319)

-.619* (.330)

Per capita oil output .013 (.021) - .016

(.020) -.038** (.017) -

Transition*PC oil output

-.035 (.038) - -.037

(.040) -.020 (.046) -

Resource depletion/GNI - -.036

(.040) - -.051 (.042) - -.140***

(.040) Transition*Resource depletion/GNI - .002

(.068) - .005 (.067) - .032

(.082) R-squared .604 .665 .573 .647 .579 .626 Observations 162 148 162 148 162 148

26

AS THE ABOVE TABLE SHOWS, OIL ABUNDANCE IS ASSOCIATED WITH WORSE VALUES OF VOICE AND ACCOUNTABILITY INDEX.

THIS RESULT IS CONSISTENT WITH EGOROV ET AL. (2009) WHO CONSIDER AUTOCRT’S INCENTIVES WITH RESPECT TO FREE MEDIA.

INTUITION FOR EGOROV ET AL.’S MODEL: FREE MEDIA PROVIDE INFO ABOUT BUREAUCRAT’S PERFORMANCE, BUT INCREASE THE PROBABILITY OF AUTOCRAT’S DEMISE.

WHEN TAX REVENUES FROM NON-MINING IS IMPORTANT, THIS TRADEOFF FAVORS RELATIVELY FREE MEDIA.

HOWEVER, WHEN THE AUTOCRAT COULD GET REVENUES FROM OIL, PERFORMANCE OF THE REST OF THE ECONOMY IS LESS IMPORTANT AND THE AUTOCRAT STIFLES THE MEDIA.

EMPIRICAL APPROACH: FIXED EFFETS REGRESSIONS

EMPIRICAL PROBLEMS WITH THE PAPER: “ABSOLUTE” NATURAL RESOURCE MEASURE (I.E., NOT RELATED TO EITHER POPULATION OR GDP); CONTROLS FOR PER CAPITA GDP; POTENTIAL ENDOGENEITIES.

27

OIL AND DEMOCRACY M. ROSS (2014): “The core finding that more oil wealth is associated with less

democracy has been replicated many times, using better data and more sophisticated methods”

THIS STATEMENT IS CORRECT. HOWEVER, THERE HAVE BEEN SEVERAL RECENT STUDIES THAT FOUND NO ASSOCIATION (OR POSITIVE ASSOCIATION) BETWEEN OIL WEALTH AND DEMOCRACY (E.G., HABER AND MENALDO 2010).

28

CITING WRIGHT ET AL. (2014), ROSS CLAIMS THAT THERE MIGHT HAVE BEEN NO ASSOCIATION PRIOR TO MID-1970s, BUT THERE HAS BEEN A FAIRLY STRONG RELATIONSHIP SINCE THEN.

HOWEVER, WRIGHT ET AL. SHOWED THAT WHILE HIGH OIL WEALTH EXTENDS AUTOCRATIC REGIME’S LIFE, DECREASING OIL WEALTH PROMOTES A TRASITION TO ANOTHER AUTOCRACY RATHER THAN TO DEMOCRACY.

ROSS ALSO EMPHASIZES THE DIFFERENCE BETWEEN SHORT-TERM EFFECTS OF OIL WEALTH CHANGES AND LONG-TERM EFFECTS, BUT IT IS HARD TO SEE WHAT LONG-TERM EFFECTS COULD BE FOUND “SINCE MID-1970s”.

29

A NICE PAPER IN THIS LITERATURE IS BRUCKNER ET AL. (2012) WHO SHOW THAT OIL PRICE SHOCKS (𝜃𝜃 × 𝐶𝐶𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂, where 𝜃𝜃 =𝑂𝑂𝑒𝑒𝑒𝑒𝑒𝑒𝑂𝑂𝑒𝑒 − 𝑂𝑂𝑖𝑖𝑒𝑒𝑒𝑒𝑂𝑂𝑒𝑒𝑠𝑠) IMPROVE DEMOCRACY SCORES OF OIL EXPORTERS (IN A SENSE CONTRADICTING AREZKI AND BRUCKNER 2009) BUT WORSEN DEMOCRACY IN OIL IMPORTERS.

ALTHOUGH THE ECONOMETRICS IN THIS PAPER APPEARS TO BE SOUND, THIS RESULT IS SUSPECT, BECAUSE ACCORDING TO THE AUTHORS, DEMOCRACY SCORES REACT TO OIL SHOCKS ONLY WITHIN A YEAR.

INTERESTINGLY, OIL PRICE SHOCKS HAVE A ROBUST POSITIVE (NEGATIVE) EFFECT ON GDP OF EXPORTERS (IMPORTERS).

30

USE OF SUB-NATIONAL DATA

ADVANTAGES VS. DISADVANTAGES OF SUB-NATIONAL DATA

RESEARCH HAS BEEN MAINLY ON THE US AND MAINLY ON CORRUPTION PROBLEMS WITH MEASURING CORRUPTION (CONVICTIONS VS. SURVEY OF REPORTERS VS. AP INDEX).

31

PAPYRAKIS AND GERLACH (2007) FIND THAT NATURAL RESOURCES ARE ASSOCIATED WITH GREATER CORRUPTION. PROBLEMS: FEW OBSERVATIONS (49); NATURAL RESOURCES ARE DEFINED TO INCLUDE AGRICULTURE, FORESTRY, FISHING, AND MINING AS A SHARE OF GSP ON THE SCALE OF 1-10; RESULT HOLDS ONLY WITH NO CONTROL VARIABLES. CONTROLLING FOR INCOME, STATISTICAL SIGNIFICANCE DISAPPEARS.

32

Table 6. Correlations among corruption measures (US states) Convictions

per 1000 people

Legal corruption (reporters)

Illegal corruption (reporters)

AP corruption index

Convictions per 1000 people

1.0


0.1919 1.0


0.1032 0.4800 1.0

AP corruption index

0.1218 0.6967 0.6736 1.0

33

ACCORDING TO MY OWN ESTIMATES FOR THE US, MINING MEASURES (BOTH SHARE AND PER CAPITA) ARE INDEED ASSOCIATED WITH GREATER CORRUPTION IN THE US AS MEASURED BY CONVICTIONS, BUT NOT AS MEASURED BY SURVEY OF REPORTERS. MOREOVER, EVEN FOR CONVICTIONS, STATISTICAL SIGNIFICANCE IS REDUCED WHEN WE ADD SOME CONTROLS. THE RESULTS FOR RUSSIA ARE AMBIGUOUS PRESUMABLY BECAUSE CORRUPTION MEASURES FOR RUSSIA’S REGIONS ARE ALL OVER THE PLACE

34

Table 7. Natural resource abundance and corruption in the US states Convictions

per 1000 people



AP corruption index

Log PC mining

.009* (.005)

-.097 (.121)

.101 (.103)

-.009 (.017)

Log of population

.0002 (.0029)

.208** (.094)

.330*** (.109)

.036*** (.012)

Urban share (%, 2000)

-.0004* (.0002)

-.012 (.009)

.011 (.011)

.003** (.001)

Racial diversity

.0006 (.0169)

1.97** (.794)

.309 (.784)

.041 (.086)

BA Degree (%, 2000)

.0003 (.0007)

-.038 (.037)

-.050 (.033)

-.008*** (.003)

“Moralistic” nature

-.016*** (.005)

-.380 (.328)

-.481* (.292)

-.033 (.027)

R-sq .295 .331 .496 .570 Notes: Log per capita GSP doesn’t affect estimates and is typically insignificant; all regressions have a constant; very similar results obtain for share of mining in GSP

35

RUSSIAN CORRUPTION MEASURES CARNEGIE CENTER (2004) – HIGHER VALUES REFLECT LESS CORRUPTION INDEM FOUNDATION (2002; “INTEGRAL ESTIMATE” AND “VOLUME”) – HIGHER VALUES IMPLY MORE CORRUPTION FOM SURVEY (2011) – HIGHER VALUES MEAN MORE CORRUPTION BRIBE TAX (BEEPS; 2011) – HIGHER VALUES CORRESPOND TO GREATER CORRUPTION BRIBE FEQUENCY (BEEPS; 2011) – HIGHER VALUES CORRESPOND TO BRIBES BEING MORE COMMON

36

Table 8. Correlations among corruption measures (Russia’s regions) Carnegie

index (2004)

INDEM “estimate”

(2002)

INDEM “volume”

FOM survey (2011)

Bribe tax

(BEEPS 2011)

Bribe frequency (BEEPS)

Carnegie index (2004) 1.0 INDEM “estimate” (2002)

0.048 1.0

INDEM “volume” -0.362 0.296 1.0 FOM survey (2011) -.494 0.430 0.461 1.0 Bribe tax (BEEPS; 2011)

-.044 -0.424 -0.376 -0.187 1.0


-0.100 0.223 0.332 0.220 0.425 1.0

37

Table 8. Natural resource abundance and corruption in Russia’s regions

Carnegie

index (2004)


(2002)

INDEM “volume”

(2002)

FOM survey (2011)

Bribe tax (BEEPS;

2011)


Oil / GRP (%; 2001) ×

Oil Price

.001 (.002)

-.000 (.001)

-.003** (.001)

-.010** (.005)

.003** (.001)

-.001 (.001)

Log of population

-.134 (.109)

.022 (.070)

.191** (.078)

.934 (1.14)

.300 (.225)

.400*** (.100)

Urban share (%)

.006 (.008)

-.001 (.005)

.001 (.010)

.072 (.084)

-.008 (.015)

.004 (.007)

Student / population

2.38 (7.84)

3.75 (3.16)

4.17 (4.69)

62.60 (77.27)

-57.03*** (11.50)

-11.02 (.6.78)

Mean Jan temperature

-0.043*** (.013)

.017* (.037)

.006 (.010)

.770*** (.140)

-.021 (.030)

-.000 (.011)

Log of dist. to Moscow

-.314*** (.085)

.033 (.052)

-.036 (.050)

2.64*** (.766)

.148 (.146)

.067 (.080)

R-sq .211 .175 .311 .358 .437 .396 No. Obs. 73 36 36 68 35 35

Notes: Log of per capita GRP is sometimes statistically significant but doesn’t qualitatively affect the estimates of oil abundance coefficients

38

Table 8. Natural resource abundance and corruption in Russia’s regions

Carnegie

index (2004)


(2002)

INDEM “volume”

(2002)

FOM survey (2011)

Bribe tax (BEEPS;

2011)


Log of PC Oil (2001) ×

Oil Price

.004 (.026)

.002 (.015)

-.033* (.017)

-.281 (.208)

.054 (.040)

-.045** (.018)

Log of population

-.133 (.106)

-.024 (.070)

.214*** (.076)

1.10 (1.19)

.337 (.241)

.426*** (.098)

Urban share (%)

.006 (.008)

.001 (.005)

.001 (.009)

.078 (.085)

-.005 (.017)

.001 (.007)

Student / population

3.11 (7.65)

3.52 (2.72)

3.14 (4.17)

52.55 (81.13)

-48.09*** (12.78)

-12.1** (5.40)

Mean Jan temperature

-0.043*** (.013)

.017* (.009)

.010 (.009)

.797*** (.145)

-.023 (.033)

.000 (.011)

Log of dist. to Moscow

-.314*** (.085)

.034 (.051)

-.017 (.049)

2.77*** (.812)

.094 (.155)

.099 (.079)

R-sq .209 .175 .310 .347 .379 .453 No. Obs. 73 36 36 68 35 35

Notes: Log of per capita GRP is sometimes statistically significant but doesn’t qualitatively affect the estimates of oil abundance coefficients

39

CONCLUSIONS

- LARGE ENDOWMENTS OF OIL/MINERALS DO NOT LOWER LONG-TERM ECONOMIC GROWTH, ALTHOUGH THEY MAY DISTORT THE RELATIONSHIP BETWEEN GDP AND INSTITUTIONAL QUALITY

- BECAUSE OF THIS DISTORTION, THE NEGATIVE EFFECT OF LARGE

ENDOWMENTS OF “POINT-SOURCE” RESOURCES ON INSTITUTIONS CLAIMED IN THE LITERATURE IS MOSTLY DUE TO THE USE OF INITIAL GDP VALUES AS CONTROL VARIABLES

- THERE MIGHT BE A NEGATIVE EFFECT OF OIL ON GOVERNMENT

ACCOUNTABILITY, BUT THIS REQUIRES FURTHER RESEARCH - EVIDENCE ON THE EFFECT OF OIL ON DEMOCRACY IS

CONTRADICTORY - REGIONAL DATA FOR RUSSIA AND THE US PRESENT CONTRADICTORY

EVIDENCE ON THE EFFECT OF OIL ON CORRUPTION

eerc oilcurse _fin_alexeev

Science