The impact of economic inequality on economic freedom.
Murphy, Ryan H.
Contemporary economic policy debates are dominated by concerns
regarding the rise in inequality (Stiglitz 2012, Piketty 2014).
Primarily, this has led to a focus in re-invigorating redistribution.
For instance, Robert Shiller (2014) has recently argued for indexing top
marginal tax rates to inequality and using the revenues to fund transfer
payments. Secondarily, there are the longstanding objections to
"neoliberalism in general, which has encouraged globalization and
the liberalization of markets. To the extent that liberal reforms have
improved economic institutions, might today s inequality subsequently
derail them?
It is often difficult to find firm evidence finking negative
outcomes to inequality (Deaton 2003, Porter 2014). However, some
economists have argued that inequality may harm the quality of
institutions. For example, Acemoglu et al. (2013) have argued drat
concentrations of wealth may subvert democracy. This argument is also
present in political science (Bartels 2008), and Easterly (2001) has
made similar points. Such arguments offer a more rigorous conception of
the popular notion of inequality subverting politics, a concern that is
especially salient following Citizens United v. Federal Election
Commission, and more recently, McCutcheon v. Federal Election
Commission. Acemoglu has made this point explicitly regarding Citizens
United, saying, "Instead of trying to stem that tide, we've
done the opposite and we've now opened the sluice gate and said you
can use that money with no restrictions whatsoever" (Garofalo
2012). Generally, the debate has centered on the notion that inequality
will weaken institutions by swinging policies toward favoring the
economic interests of the rich.
The approach here will differ, looking at the effect of inequality
on free economic institutions. The measure used will be the Economic
Freedom of the World (EFW) index, published by the Fraser Institute
(Gwartney, Lawson, and Hall 2013). The index runs from 0 to 10 using
five components of economic freedom, with higher index values
corresponding to greater economic freedom. This index has been used in a
variety of academic journals to investigate a broad range of issues
(Hall and Lawson 2014). Numerous studies using this index have
investigated whether economic freedom worsens inequality (Berggren 1999,
Scully 2002, Carter 2007, Clark and Lawson 2008), finding mixed results,
but not whether inequality may worsen economic freedom. Also relatedly,
recent research by Young and Lawson (2014) finds that economic freedom
is associated with a higher share of labor income.
Using a similar index for the United States, Apergis, Dincer, and
Payne (2014) argue that there is a bidirectional relationship between
inequality and economic freedom, with the possibility that policies that
are meant to reduce inequality will reduce economic freedom, which will
then only make inequality worse. Bennett and Vedder (2013) investigate
the relationship between the two variables, also using U.S. data, and
find similar results. In this article, I do not seek to identify
bidirectional effects; rather, I wish to investigate the long-run
effects of inequality on economic freedom in an international context.
This article also fits with the growing literature that uses the
EFW index as the dependent variable. While the index has been used a
large number of times as an independent variable, far less work has gone
into explaining economic freedom. Recent scholarly work has examined the
impact of foreign aid (Bearce and Tirone 2010), personal characteristics
of politicians (Dreher et al. 2009), and culture (Jing and Graham 2008)
on economic freedom as measured by the EFW index. Inequality too may
play a role in determining economic freedom.
The primary method this article employs is to control for economic
freedom at the beginning period, in effect differencing the data, and
then determine the impact of the Gini coefficient, (1) a common measure
of income inequality, in the first period on the EFW index in the future
period. We find that a one standard deviation increase in the Gini
coefficient reduces (worsens) the EFW index by 0.18-0.26 standard
deviations, depending on the specification. This magnitude persists
across the other three specifications of the baseline model, though it
loses significance upon the inclusion of fixed effects. (2)
In addition, the same procedure was applied to each of the five
subcomponents of the EFW index. Of the five subcomponents, inequality
has the largest impact in the later period on the size of government.
The most counterintuitive result is the mixed results regarding the
impact of inequality on regulation. Upon inclusion of fixed effects, a
one standard deviation increase in the Gini coefficient improves the
regulation score by 0.46 standard deviations. While the effect is only
significant with 90 percent confidence, the magnitude is very large.
Besides the impact of inequality on regulation, its impact on the other
components of the EFW index is generally intuitive.
Data and Method
Differencing (or controlling for levels in the first period)
alleviates many concerns regarding endogeneity, but the tradeoff that
arises is that there is often little variation from year to year. The
approach used in this article avoids that problem by comparing periods
10 years apart. The most parsimonious specification employed is to use
the Gini coefficient in year t to predict economic freedom in year t +
10 while controlling for economic freedom in year t. This specification
can be found in Equation 1. Despite its simplicity, this specification
is reasonably robust. Any proposed variable attacking this result must
be correlated with the change in EFW and the Gini coefficient at the
beginning period, and it is not immediately obvious what would do so,
especially upon the inclusion of fixed effects. (3)
(1) [EFW.sub.t+10] = [[beta].sub.0] + [[beta].sub.1][EFW.sub.t] +
[[beta].sub.2][gini.sub.t] + [epsilon].
In addition to this estimation, an analogous method was used to
measure the effect of inequality on each of the subcomponents of the EFW
index. Area 1 (i.e., the first subcomponent) measures the size of
government in the economy, with higher scores corresponding to smaller
governments. Equation 2 provides the parsimonious specification for
predicting Area 1 as an example. Area 2 measures the integrity of the
legal system and the enforcement of property rights. Area 3 measures the
soundness of money. Area 4 measures the freedom of trade
internationally, and Area 5 measures the regulatory environment. Of
these components, the most obvious conduit by which governments may
respond to inequality is Area 1, by means of increasing transfer
payments. However, it is easy to imagine ways in which inequality may
affect other Areas, for instance inequality leading to a backlash
against trade liberalizations.
(2) Area[1.sub.t + 10] = [[beta].sub.0] + [[beta].sub.1]
Area[1.sub.t] + [[beta].sub.2][gini.sub.t] + [epsilon].
Table 1 provides summary statistics for each of these variables.
(4) In addition to those already mentioned, data on ethnic, linguistic,
and religious fractionalization from Alesina et al. (2003) are included.
Unfortunately, only cross-sectional data are available for
fractionalization, but it is hoped that these variables help to capture
the cohesiveness of the observed countries that is unrelated to, but may
be correlated with, inequality. Data on the Gini coefficient are from
the World Banks online databank, which contains observations beginning
in 1978.
The sample size these data yield may be smaller than expected.
Until 2000, the EFW index was available only once every five years going
back until 1975, and only for a much smaller number of countries.
Additionally, the most recent EFW index ranks countries based on 2011
data. I include only observations for which the World Bank reports the
Gini coefficient in the same year t for which there is an EFW score both
in year t and year t+10. This means that t may only take the value of
the years 1980, 1985, 1990, 1995, 2000, and 2001. Ultimately, this means
that no regression has more than 114 observations. A full list of the
country-years in the sample appears in Table 2.
Results
Table 3 provides the baseline results. The Gini coefficient is
negatively associated with lower scores for the EFW index in the future.
Regression (2) provides the headline result. A one standard deviation
increase in the Gini coefficient decreases the EFW index by 0.15 points
10 years later, about 0.18 standard deviations. This is a modest effect,
but tangible and important considering the host of other variables that
may change the quality of economic institutions. The effect is
statistically significant at the 95 percent level. The effect is
reasonably robust across specifications.
Time and country fixed effects were both also attempted. The result
remains statistically significant in all specifications except that
which includes both country and time fixed effects, which is
unsurprising given that the data are already effectively differenced and
the data points are relatively few in comparison to similar models. In
the model with country fixed effects, for instance, the model consumes
75 degrees of freedom when only 112 observations are available. Despite
this, the point estimate of the effect of inequality is virtually
identical to those of the other models.
Tables 4-8 replicate these regressions for each Area of economic
freedom. The empirical results in Table 4 for Area 1 (size of
government) are surprising. The first three specification all show the
Gini coefficient having virtually zero impact on the size of government,
but when country fixed effects are included, a one standard deviation
increase in the Gini coefficient decreases the country's score in
Area 1 by 0.71, about 0.55 standard deviations. If the results of
Regression 8 are believed over Regressions 5-7, this is significant
evidence that inequality drives demands for increases in the size of the
welfare state, contrary to the hypothesis that inequality will lead to
lower taxes and social spending.
Table 5 reports the results of the impact of inequality on the
legal system. These results are similar to, but weaker than, the results
found for the overall EFW index. Like the overall index, the Gini
coefficient loses statistical significance (but keeps its sign) upon the
inclusion of country fixed effects. Using the results from Regression
10, we find that a one standard deviation increase in the Gini
coefficient decreases the score in Area 2 by 0.30 points, or about 0.22
standard deviations.
The results for Area 3 found in Table 6 are weak. This is not
surprising given the public's lack of familiarity with monetary
policy in comparison to the other components of the EFW index. While the
coefficient on the Gini coefficient in Regression 13 is statistically
significant and negative, the result immediately disappears in all other
specifications. And, as shown in Table 7, there are no discernable
effects of the Gini coefficient on Area 4 of the EFW index.
Results for Area 5, regulation, are perhaps the most surprising. In
Table 8, the Gini coefficient has negative effects on economic freedom
in the first three specifications, all of which are statistically
significant, and these effects are of similar magnitude to others found
here. However, upon inclusion of fixed effects, the sign flips and the
result is statistically significant at the 90 percent level. Though such
significance is weak evidence, it is worth noting that, if the point
estimate is accurate, the magnitude is fairly large. A one standard
deviation increase in the Gini coefficient, in this model, would
increase the score in Area 5 by 0.44 points, or 0.46 standard
deviations. However, the most we can say is that the evidence regarding
the effect of the Gini coefficient on regulation is mixed. (5)
Conclusion
Overall, inequality appears to have a negative impact on economic
freedom. While some of the evidence is mixed and at times
counterintuitive, a one-point increase in the Gini coefficient decreases
economic freedom (as measured by the Fraser Institute's Economic
Freedom of the World index) by 0.013-0.019 points. Equivalently, a one
standard deviation increase in the Gini coefficient reduces economic
freedom by 0.18-0.26 standard deviations. Inequality appears to increase
the size of government and to have a negative effect on the rule of law,
little effect on the soundness of money or trade, and ambiguous effects
on regulation.
Taken as a whole, this is not a cheery outcome. Those like Shiller
who call for higher taxes and more transfers in response to the growth
in inequality may be prophetic in the sense that policy is likely to
move in that direction, regardless of whether or not the rationales for
such policies hold water. Ironically, while those favoring more
interventionist policies in response to greater economic inequality will
likely win out, the predictions that inequality will allow the economic
interests of the rich to capture more of the political process will be
shown to have been wrong--that is, taxes will rise, not fall.
One implication is that those who wish to promote economic freedom
as measured by the EFW index should enthusiastically promote
liberalizations that also promise to reduce inequality. Reforms that do
both include educational reform, ending corporate welfare, and
intellectual property reform. Prioritizing those liberalizations over
others promises to improve the political climate for other
liberalizations. Liberalizations of the past that likely increased
inequality in the developed world, (6) like globalization, though
entirely justifiable on their own merits, may hinder the market-oriented
policy proposals of the present.
Proponents of free markets, from Hayek (1976) to Nozick (1974), are
often skeptical of the very philosophical meaningfulness of inequality.
Regardless of how inequality should be thought ol from a normative point
of view, in a positive sense we may say that it inhibits the development
of free economic institutions. Therefore, proponents of free markets
should be opponents of inequality.
References
Acemoglu, D.; Naidu, S.; Restrepo, P.; and Robinson, J. A. (2013)
"Democracy, Redistribution and Inequality." NBER Working Paper
No. 19746.
Alesina, A.; Devleeschauwer, A.; Easterly, W.; Kurlat, S.; and
Wacziarg, R. (2003) "Fractionalization." Journal of Economic
Growth 8 (2): 15.5-94.
Aspergis, N.; Dincer, O.; and Payne, J. E. (2014) "Economic
Freedom and Income Inequality Revisited: Evidence from a Panel
Correction Model." Contemporary Economic Policy 32 (1): 67-74.
Bartels, L. (2008) Unequal Democracy: The Political Economy of the
New Gilded Age. Princeton, N.J.: Princeton University Press.
Bearce, D. H., and Tirone, D. C. (2010) "Foreign Aid
Effectiveness and the Strategic Goals of Donor Governments."
Journal of Politics 72 (3): 837-51.
Bennett, D. L., and Vedder, R. K. (2013) "A Dynamic Analysis
of Economic Freedom and Income Inequality in the 50 U.S. States:
Empirical Evidence of a Parabolic Relationship." Journal of
Regional Analysis & Policy 43 (1): 42-55.
Berggren, N. (1999) "Economic Freedom and Equality: Friends or
Foes?" Public Choice 100 (3-4): 203-23.
Carter, J. R. (2007) "An Empirical Note on Economic Freedom
and Income Inequality." Public Choice 130 (1-2): 163-77.
Clark, J. R., and Lawson, R. (2008) "The Impact of Economic
Growth, Tax Policy, and Economic Freedom on Income Inequality."
Journal of Private Enterprise 23 (3): 23-31.
Deaton, A. (2003) "Health, Income, and Inequality." NBER
Research Reporter: Research Summary. Available at www.nber.org
/reporter/spring03/health.html.
Dreher, A.; Lamia, M. J.; Lein, S. M.; and Somogyi, F. (2009)
"The Impact of Political Leaders' Profession and Education on
Reforms." Journal of Comparative Economics 37 (1): 169-93.
Easterly, W. (2001) "The Middle Class Consensus and Economic
Development." Journal of Economic Growth 6 (4): 317-35.
Garofalo, P. (2012) "MIT Economist: Income Inequality In the
U.S. Is Crushing the Middle Class's Political Power." Think
Progress. Available at http://thinkprogress.org/economy/
2012/03/23/451166/acemoglu-income-inequality-political-power.
Gwartney, J.; Lawson, R.; and Hall, J. (2013) Economic Freedom of
the World: 2013 Annual Report. Vancouver, BC: Fraser Institute.
Hall, J., and Lawson, R. (2014) "Economic Freedom of the
World: An Accounting of the Literature." Contemporary Economic
Policy 32 (1): 1-19.
Hayek, F. A. (1976) Law, Legislation, and Liberty, Volume 2: The
Mirage of Social Justice. Chicago: University of Chicago Press.
Jing, R., and Graham, J. L. (2008) "Values versus Regulations:
How Culture Plays Its Role." Journal of Business Ethics 80 (4):
791-806.
Milanovic, B. (2012) "Global Inequality by the Numbers: In
History and Now." World Bank Policy Research Paper No. 6259.
Nozick, R. (1974) Anarchy, State, and Utopia. New York: Basic
Books.
Piketty, T. (2014) Capital in the Twenty-First Century. Cambridge,
Mass: Harvard University Press.
Porter, E. (2014) "Income Equality: A Search for
Consequences." New York Times (26 March).
Scully, G. W. (2002) "Economic Freedom, Government Policy, and
the Trade-off between Equity and Economic Growth." Public Choice
113(1-2): 77-96.
Shiller, R. (2014) "Better Insurance against Inequality."
New York Times (13 April).
Stiglitz, J. (2012) The Price of Inequality: How Today's
Future Divided Society Endangers Our Future. New York: W.W. Norton.
Young, A., and Lawson, R. (2014) "Capitalism and Labor Shares:
A Cross-Country Panel Study." European Journal of Political Economy
33: 20-36.
(1) The Gini coefficient is bounded by zero and one. Zero
corresponds to perfect income equality (the income of everyone in
society is identical) and one corresponds to perfect income inequality
(one person in society has all the income).
(2) While it loses significance, the magnitude of the coefficient
actually grows.
(3) Consider: the fixed effect captures variables related to the
country-specific trajectory, not just the country-specific levels, of
EFW.
(4) The dataset was constructed such that country-years with Gini
coefficient data available were first identified, and subsequently EFW
data were matched to it. This explains why the Gini coefficient has more
data points than the EFW index.
(5) One robustness check on these results was attempted. The
results were essentially unchanged when the sample was split into OECD
versus non-OECD countries. While replicating (when possible) each of the
24 regressions using restricted samples did not uniformly conform to the
estimated ranges found above, qualitatively it gives no reason to doubt
the conclusions reached.
(6) This is not to say that globalization promoted global
inequality, which has actually fallen (see Milanovic 2012).
Ryan H. Murphy is a Research Associate at the O'Neil Center
for Global Markets and Freedom at SMU Cox School of Business. He thanks
Robert Lawson, Doug Murdoch, and Yu-Hsi Liu for their useful comments.
TABLE 1
DESCRIPTIVE STATISTICS
Variable Obs Mean Std. Dev. Min Max
Gini Coefficient 465 42.364 11.0329 19.400 74.330
EFW, Year t 112 6.302 1.152 3.030 8.650
EFW, Year t+10 347 6.716 0.812 2.940 9.100
Area 1 of EFW, 114 6.115 1,596 2.773 9.305
Year t
Area 1 of EFW, 348 6.654 1.293 2,363 9.262
Year t+10
Area 2 of EFW, 110 5.389 1.777 1.884 9.491
Year t
Area 2 of EFW, 347 5.221 1.366 1.600 9.005
Year t+10
Area 3 of EFW, 113 7.243 2.326 0 9.838
Year t
Area 3 of EFW, 347 7.9322 1.437 0 9.698
Year t+10
Area 4 of EFW, 111 6.859 1.948 0.941 9.485
Year t
Area 4 of EFW, 347 7.131 1.102 2.376 9.708
Year t+10
Area 5 of EFW, 113 5.906 1.161 1,579 8.433
Year t
Area 5 of EFW, 354 6.629 0.955 3.764 9.338
Year t+10
Ethnic 455 0.441 0.228 0.002 0.930
Fractionalization
Linguistic 455 0.333 0.285 0 0.923
Fractionalization
Religious 455 0.397 0.219 0.004 0.860
Fractionalization
TABLE 2
LIST OF COUNTRY-YEARS IN SAMPLE
Algeria 1995 Italy 2000
Argentina 1995, 2000, 2001 Jamaica 1990, 2001
Austria 2000 Latvia 1995
Bangladesh 2000 Lithuania 2000, 2001
Belgium 2000 Luxembourg 2000
Belize 1995 Madagascar 1980, 2001
Bolivia 2000, 2001 Malaysia 1995
Brazil 1985, 1990, Mali 2001
1995, 2001
Bulgaria 1995, 2001 Mexico 2000
Cameroon 2001 Morocco 1985, 2001
Canada 2000 Nepal 1985
Chile 1990, 2000 Nicaragua 2001
China 1990 Norway 2000
Colombia 1980, 2000, 2001 Panama 1995, 2001
Costa Rica 1990, 1995, Paraguay 1990, 1995, 2001
2000, 2001
Cote d'Ivoire 1985, 1995 Peru 2000, 2001
Croatia 2000, 2001 Philippines 1985, 2000
Dominican Rep. 2000, 2001 Poland 1985, 2000, 2001
Ecuador 1995, 2000 Romania 2000, 2001
Egypt 2000 Russia 2001
El Salvador 1995, 2001 Rwanda 1985, 2000
Estonia 1995, 2000, 2001 Senegal 2001
Finland 2000 South Africa 1995, 2000
France 1995 Spain 2000
Georgia 2000, 2001 Sri Lanka 1985
Germany 2000 Sweden 2000
Greece 2000 Switzerland 2000
Guatemala 2000 Tanzania 2000
Haiti 2001 Thailand 1990, 2000
Honduras 1990, 1995, 2001 Tunisia 1985, 1990,
1995, 2000
Hungary 2000, 2001 Ukraine 1995
Indonesia 1990 United States 2000
Iran 1990 Uruguay 1995, 2000, 2001
Ireland 2000 Venezuela 1995, 2001
Israel 2001 Zimbabwe 1995
TABLE 3
BASELINE REGRESSIONS
(1) (2) (3) (4)
LHS EFW, Year EFW, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
EFW, Year t 0,518 *** 0.498 *** 0.424 *** 0.179
(0.055) (0.056) (0.072) (0.107)
Gini Coefficient -0.017 *** -0.014 ** -0.013 ** -0.019
(0.006) (0.006) (0.007) (0.013)
Ethnic -0.652 * -0.581
Fractionalization (0.366) (0.389)
Linguistic -0.029 -0.017
Fractionalization (0.288) (0.311)
Religious 0.547 * 0.530 *
Fractionalization (0.314) (0.318)
Constant 4.197 *** 4.253 *** 3.925 *** 4.149 ***
(0.440) (0.449) (0.667) (0.612)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 112 112 112 112
Adjusted [R.sup.2] 0.459 0.474 0.474 0.906
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
TABLE 4
REGRESSION RESULTS FOR AREA 1
(5) (6) (7) (8)
LHS Area 1 Area 1 Area 1 Area 1
EFW, Year EFW, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
Area 1 of EFW, 0.493 *** 0.529 *** 0.551 *** 0.124
Year t (0.077) (0.079) (0.084) (0.172)
Gini Coefficient 0.009 0.006 0.002 -0.064 **
(0.011) (0.013) (0.013) (0.026)
Ethnic -0.263 -0.606
Fractionalization (0.553) (0.587)
Linguistic -0.359 -0.146
Fractionalization (0.448) (0.468)
Religious 0.882 * 0.910 *
Fractionalization (0.484) 0.493
Constant 2.990 *** 2.898 *** 4.226 *** 5.552 ***
(0.429) (0.475) (0.920) (1.529)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 114 114 114 114
Adjusted [R.sup.2] 0.399 0.404 0.400 0.748
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
TABLE 5
REGRESSION RESULTS FOR AREA 2
(9) (10) (11) (12)
LHS Area 2 Area 2 Area 2 Area 2
EFW7, Year EFW7, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
Area 2 of EFW, 0.532 *** 0.553 *** 0.582 *** 0.116
Year t (0.057) (0.060) (0.066) (0.088)
Gini Coefficient -0.029 *** -0.024 ** -0.18 * -0.007
(0.009) (0.010) (0.010) (0.016)
Ethnic -0.541 -0.390
Fractionalization (0.525) (0.528)
Linguistic 0.169 0.042
Fractionalization (0.411) (0.41)
Religious -0.534 -0.524
Fractionalization (0.465) (0.468)
Constant 3.884 *** 3.890 *** 2.468 ** 4.327 ***
(0.621) (0.633) (1.165) (0.866)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 110 110 110 110
Adjusted [R.sup.2] 0.599 0.601 0.615 0.938
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
TABLE 6
REGRESSION RESULTS FOR AREA 3
(13) (14) (15) (16)
LHS Area 3 Area 3 Area 3 Area 3
EFW, Year EFW, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
Area 3 of EFW, 0.362 *** 0.350 *** 0.211 *** 0.051
Year t (0.063) (0.065) (0.071) (0.123)
Gini Coefficient -0.029 ** -0.020 -0.022 0.052
(0.014) (0.015) (0.014) (0.039)
Ethnic -1.610 * -1.701 **
Fractionalization (0.851) (0.835)
Linguistic 0.210 0.584
Fractionalization (0.669) (0.658)
Religious 0.910 0.585
Fractionalization (0.736) (0.681)
Constant 6.609 *** 6.459 *** 6.429 ***
(0.774) (0.843) (1.352)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 113 113 113 113
Adjusted [R.sup.2] 0.242 0.255 0.375 0.674
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
TABLE 7
REGRESSION RESULTS FOR AREA 4
(17) (18) (19) (20)
LHS Area 4 Area 4 Area 4 Area 4
EFW, Year EFW, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
Area 3 of EFW, 0.383 *** 0.370 *** 0.356 *** -0.181 ***
Year t (0.044) (0.045) (0.062) (0.066)
Gini Coefficient -0.009 -0.005 -0.006 -0.000
(0.008) (0.008) (0.009) (0.015)
Ethnic -0.616 -0.156
Fractionalization (0.492) (0.501)
Linguistic -0.008 -0.317
Fractionalization (0.388) (0,395)
Religious 0.184 0.220
Fractionalization (0.428) (0.416)
Constant 4.826 *** 4.957 *** 3.241 *** 2.554 ***
(0.478) (0.512) (0.797) (0.811)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 111 111 111 111
Adjusted [R.sup.2] 0.412 0.408 0.453 0.915
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
TABLE 8
REGRESSION RESULTS FOR AREA 5
(21) (22) (23) (24)
LHS Area 5 Area 5 Area 5 Area 5
EFW, Year EFW, Year EFW, Year EFW, Year
t+10 t+10 t+10 t+10
Area 3 of EFW, 0.567 *** 0.536 *** 0.446 *** 0.124
Year t (0.064) (0.065) (0.069) (0.153)
Gini Coefficient -0.019 *** -0.014 * -0.016 ** 0.040 *
(0.007) (0.007) (0.007) (0.020)
Ethnic -0.751 * -0.658
Fractionalization (0.421) (0.415)
Linguistic 0.440 0.497
Fractionalization (0.332) (0.326)
Religious 0.673 * 0.670 *
Fractionalization (0.371) (0.351)
Constant 4.083 *** 3.966 *** 3.673 *** 1.592
(0.485) (0.488) (0.703) (1.138)
Time Fixed Effects N N Y Y
Country Fixed N N N Y
Effects
n 113 113 113 113
Adjusted [R.sup.2] 0.428 0.443 0.517 0.736
* Denotes significance at 90 percent level.
** Denotes significance at 95 percent level.
*** Denotes significance at 99 percent level.
