Working Paper Series Department of Economics Alfred Lerner College of Business & Economics University of Delaware

Transcription

1 Working Paper Series Department of Economics Alfred Lerner College of Business & Economics University of Delaware Working Paper No Institutional Quality and Economic Growth: Maintenance of the Rule of Law or Democratic Institutions, or Both? James L. Butkiewicz and Halit Yanikkaya February c 2004 by author(s). All rights reserved.

2 INSTITUTIONAL QUALITY AND ECONOMIC GROWTH: MAINTENANCE OF THE RULE OF LAW OR DEMOCRATIC INSTITUTIONS, OR BOTH? James L. Butkiewicz Department of Economics University of Delaware Newark, DE Halit Yanikkaya College of Business and Administrative Sciences Department of Economics Celal Bayar University Manisa, Turkey

3 Abstract Analysis of the factors determining rates of economic growth has found that country-specific characteristics have important effects on growth performance. Empirical evidence to date suggests that maintenance of the rule of law promotes growth, while adopting democratic institutions does not appear to improve growth performance. We find that these conclusions are very sensitive to sample selection and to estimation technique. When an identical sample of countries is used, we find that countries with democratic institutions do enjoy superior growth performance. The relationship between growth and democratic institutions is also sensitive to the estimation technique used. Estimates using instrumental variable techniques suggest that democratic institutions do experience better growth performance. These results are especially relevant for developing nations. JEL Classifications: O40, P16 Keywords: Rule of Law, Democracy, Economic Growth 2

4 I. Introduction Governmental institutions establish the framework for economic activity within a country. Good institutions create an environment that promotes economic activity, inventiveness, and growth and development. Bad institutions typically result in economic stagnation. Two characteristics that typify most developed economies are democracy and maintenance of the rule of law. While related, these two institutional characteristics are not identical. Maintenance of the rule of law is not necessarily unique to democratic societies. However, these two institutional characteristics are high correlated. Evaluation of these two institutional frameworks for their effects on economic growth requires separate assessment of each. Empirical analysis of the impacts of each institutional type on growth has found that maintenance of the rule of law enhances growth, while establishing democratic political systems has no significant effect on growth. In this paper, we report findings suggesting that these conclusions are not robust to sample selection or estimation technique. Specifically, we find evidence indicating that both maintenance of the rule of law and democratic institutions increase real economic growth. The evidence also indicates that the effects of democracy are greatest for developing countries, a finding that is especially germane for development policy in these countries. The next section briefly reviews the literature investigating the impact of social institutions on growth. Section III presents the empirical model and discusses the data used in this study. Results are presented in Section IV. In Section V, the sensitivity of the results to estimation technique is analyzed. Concluding comments are in Section VI.

5 II. Literature Review Whether two of the most important institutional characteristics of countries the overall maintenance of rule of law and the practice of democracy affect economic growth differently, especially in developing countries, has been extensively debated in the growth and development literature. An apparent consensus of the empirical growth literature is that the maintenance of the rule of law is important for economic growth, whereas democracy has no strong effect on growth, after controlling for other important growth determinants. For example, Knack and Keefer (1995) conclude that the rule of law is a better measure of institutional quality than Gastil indices (measures of democracy) and political violence measures. Recently, Dollar and Kraay (2000) report that while better rule of law raises real per capita income and income of the poor, democracy has no strong effect on real per capita income and income of the poor. However, as emphasized in many other studies, it is important to recognize that democracy provides non-material benefits in addition to its potential economic impacts. Adherence to the rule of law is manifested by maintenance of property rights and absence of corruption. These characteristics can serve as alternative measures of the establishment and practice of the rule of law. In the theoretical literature, maintenance of property rights is considered one of the most important pillars of a free market economy. Well-defined property rights are an important determinant of economic growth through their effects on the level of investment (North, 1990; North and Weingast, 1989). A number of studies (Shleifer and Vishny 1993, Ehrlich and Lui 1999) argue that corruption is detrimental for economic growth. Although Bardhan (1997) agrees with the primary arguments made by Shleifer and Vishny, he argues that if pervasive and cumbersome regulations exist in a country, corruption may actually improve efficiency and increase growth (see Tanzi (1998) for a more complete review of the definitions and the economic impacts of corruption). Using the International Country Risk Guide (ICRG) data and/or Kaufmann, Kraay, and Zoido-Lobaton (1999) 2

6 data, a number of studies (such as Knack and Keefer 1995; Mauro 1995; Sala-i-Martin 1997; Dollar and Kraay 2000) provide evidence supporting the idea that some measure of the rule of law or property rights or corruption is significantly correlated with growth of per capita real income. Theory suggests that democratic institutions may be growth enhancing or growth retarding. The negative effect of democracy on growth is generally attributed to one of two factors. The first argument against democracy is such that democracy undermines investment because democracy leads to pressures for immediate consumption, resulting in reduced investment and thus reduced steady-state income and possibly growth. The second potential problem attributed to democracy is related to the idea of developmental states, that is, how well can states resist the pressures from special interest groups such as large firms and unions (Przeworski and Limongi, 1993)? In other words, economic development requires a strong hand from above. Necessary structural and economic reforms promoting long-run growth cannot be undertaken in the messy push and pull of democratic politics. The experience of a handful of economies in East and Southeast Asia under authoritarian regimes and Chile under Pinochet are widely cited examples for this latter argument. However, the many slowly growing countries with autocratic regimes contradict the assertion that democracy retards growth. Evidence that democracy promotes growth is found in a number of studies (Durham 1999; Przeworski and Limongi 1993; Bardhan 1997; Rodrik 1997, 2000). Rodrik ( 2000, 3) argues that participatory and decentralized political systems enable higher-quality growth: they allow greater predictability and stability, are more resilient to shocks, and deliver superior distributional outcomes. Bardhan (1997) suggests that compared to democracies, authoritarian regimes are actually less stable under deteriorating economic performance because they lack popular legitimacy, even 3

7 though authoritarian regimes might survive during periods of sustained economic growth. Similarly, Przeworski and Limongi (1993) emphasize the idea that authoritarian regimes are likely to be predatory states that eventually become a source of inefficiency. Empirical studies of the growth effects of democracy and political regime types suggest that a weak relationship exists between democracy and growth. Some studies (Knack and Keefer 1995; Sala-i-Martin 1997; Barro 1997, 1999; Rodrik 1997, 2000; Chowdhurie-Aziz 1997) find a positive impact; while other studies (Levine and Renelt 1992; Helliwell 1994; Dollar and Kraay 2000; Tavares and Wacziarg 2001) report a weak, negative relationship between democracy and growth (see Brunetti 1997 for a more complete literature review of this issue). Barro finds a non-linear relationship between democracy and growth, with growth increasing in democracy at low levels of democracy and decreasing in democracy at higher levels of democracy. The turning point appears to be at roughly the level of democracy existing in Malaysia and Mexico (in 1994), and somewhat above South Africa s level prior to its transition. Chowdhurie-Aziz (1997) finds a positive association between the degree of non-elite participation in politics and economic growth. Tavares and Wacziarg (2001), who estimate a system of simultaneous equations, find a positive effect of democracy on growth through the channels of enhanced education, reduced inequality, and lower government consumption. However, they report that democracy has an overall negative and significant, albeit small, effect on growth. Helliwell (1994) finds that democracy spurs education, measured by secondary schooling enrollment, and gross domestic investment, but has a negative (and insignificant) effect on growth holding constant investment and education. On balance, he finds no systematic net effects of democracy on subsequent economic growth. 4

8 (1994, p225) Dollar and Kraay (2000) find that there is a modest negative relationship between democracy and growth in growth regressions restricted to a sample of poor countries. Several recent papers show that either democracy or the rule of law can help to mitigate social or ethnic conflict (Collier 2000, Easterly 2001, Rodrik 1999). Countries having neither one of these institutional strengths are prone to conflict and crisis, and consequently suffer poor growth performance. III. Model and Data An empirical growth model is used to investigate the relationship between country institutional characteristics and long-run growth. In general form, this model can be characterized as yt ( y k, h ; Z ) γ = F, (1) t, t t ( t) where γ yt is country y s per capita annualized growth rate of real output in period t, y t is initial real GDP per capita, k t is the physical capital stock per person, and h t is initial human capital per person. Telephone mainlines per worker and life expectancy rates serve as proxies for the stock of physical and human capital, respectively. Although the initial GDP per capita level is employed to estimate the rate of conditional convergence, it is possible to interpret this variable as a proxy for the stock of capital for a country. The variable Z represents a vector of control and environmental variables, the majority of which are determined by the decisions of governments or individuals. These variables include a large number of the rule of law and democracy measures, trade intensity ratios, and four regional dummies. Real GDP growth is calculated using the national accounts data from the World Development Indicators 1999 CDROM (WDI 1999). Initial GDP per capita levels are 5

9 from Penn-World Table Data for telephone mainlines per employee and life expectancy rates are from Easterly and Sewadeh (2002)2. Data on trade shares as a percent of GDP are also from WDI This study uses five different measures of democracy. The first two measures of democracy, political rights and civil liberties (commonly known as Gastil indices), are used extensively in the literature. The data have been normalized from zero to one, with one denoting the greatest degree of freedom. We then use three measures of democracy from the Polity III database3. Democracy (autocracy), which measures the level of democracy (autocracy) in countries, takes values from 0 to 1. Higher (low) values indicate a higher level of democracy (autocracy) in a country. The last measure, political regime type, indicates whether a country has a civilian regime, military regime or a combination of these two regimes. This variable ranges from 1 to 4, with 1 indicating a civilian regime; 2 indicating a civilian-military regime; 3 indicating a military regime, and 4 indicating others (such as Iran and Afghanistan). The rule of law variable is a composite index based on measures of the quality of the bureaucracy, political corruption, the likelihood of government repudiation of contracts, the risk of government expropriation, and the overall maintenance of the rule of law. Each component of the rule of law index (the quality of the bureaucracy, corruption in government, the likelihood of government repudiation of contracts, the risk of government expropriation) is also used as a separate independent variable in the growth regressions. All five components of the rule of law index have been normalized from zero to one, with one denoting the greatest degree of rule of law. Data for all of 1Nuxoll (1994) and Summers and Heston (1991) discuss the reasons why the Summers and Heston data should be used for initial income levels while World Bank data should be used for income growth rates. 2 They maintain the Global Development Network Growth Database on the World Bank web site. 3 Polity III web site: 6

10 the component variables are from Easterly (1999)4. The mean values for each variable by decade are listed in Table 1. Cross-country regressions are estimated for a panel of one hundred countries5 for the sample period 1970 to The number of countries is limited by the availability of data. The system is a three-equation system estimated using the seemingly unrelated regression technique (SUR) and/or three stages least squares (3SLS) as in Barro (1997)6. The dependent variables are the average annual growth rates of real per capita GDP over three periods: , , and IV. Results Mean values of endogenous and exogenous variables by decade are listed in Table 1. Mean values of variables for the restricted sample are reported in the bottom portion of Table 1. Table 2 reports the correlation coefficients between the rule of law and democracy measures. For every pair of institutional variables there is a statistically significant correlation with the expected sign. The significant correlations suggest that multicollinearity would be a significant problem, precluding using these variables together in the same regressions. Given the high correlation between the rule of law and democracy measures, it is hard to understand the claim that the overall maintenance of rule of law is more important than democratic institutions in promoting economic growth, especially developing countries. The estimates of the basic growth model are reported in Tables 3 and 4. The model is estimated as a two or three-equation panel using the seemingly unrelated 4 See Knack and Keefer (1995), Gastil (1988), and the Polity III web site for detailed information about the institutional variables. 5 The list of countries is available from the authors. 7

11 regression (SUR) technique. Other than the constant terms, all coefficients are constrained to have the same value in each equation. The reported R-squares correspond to each equation in the panel of two or three equations. As is the practice in the literature, the effects of the rule of law and democracy are estimated separately. As expected, the proxies for human and physical capital have significant, positive effects on growth, while the initial level of real GDP per capita has a significant negative effect. The estimated rate of convergence is consistent with that found in several other studies. Trade as a percentage of GDP (a measure of openness) has an insignificant positive effect on growth in Table 3 but it has a significant positive, albeit small, effect on growth in Table 47. While the area-specific dummies indicate below average growth in Africa, Latin America, and the OECD countries, the area effect is positive but frequently insignificant for East Asian nations in most of the estimates. Equation 2 in Table 3 reports the estimated effects of the rule of law. The significant and positive estimated coefficient supports the hypothesis that overall maintenance of the rule of law promotes growth. The one possible mechanism by which the rule of law promotes growth is that overall maintenance of the rule of law creates a better investment environment, attracting more investment, especially foreign direct investment. The following four columns of Table 3 report the estimated coefficients for the individual components of the rule of law. Government repudiation of contracts and risk of expropriation indices, which indicate a commitment to the maintenance of the property rights in a country, have statistically significant and positive coefficients, indicating that maintenance of property rights is especially important for growth. Alternatively, the estimated coefficients for the corruption and 6 See Greene (1997) for a detailed discussion of these techniques. 7 The results in Table 3 for trade shares support the results of Rodrik et al. (2002). They report that trade shares lose their significance in growth regressions when used in combination with the several rule of law variables. It is important to note that in a restricted model, inclusion of all 8

12 bureaucratic quality indices are not statistically significant. Thus, these results suggest that political corruption does not hamper growth, which is not inconsistent with the theoretical literature reviewed in section II. Note that we focus first on the rule of law data because rule of law data are available for a larger sample of countries and combine all four measures of institutional quality. Gastil indices are used as measures of the level of democracy in each country. Results for the estimated specifications are reported in Table 4. Columns 2 and 3 in Table 4 contain estimated results for the growth model including political rights and civil liberties as measures of democracy. For both specifications, the estimates imply that the level of democracy has no effect on real GDP growth for the countries in the sample. As an alternative, three measures of democracy from the Polity III database are used in the growth regressions. The estimated coefficients for all three measures of democracy, reported in columns 4 through 6 in Table 4, indicate a positive but insignificant relationship exists between democracy and growth. While some previous studies find a positive, significant relationship between Gastil indices and growth, the results reported here fail to find any significant, positive relationship between democracy and long-run growth. A possible explanation is that the number of countries used in this study is considerably greater than the number of countries used in most other studies. For example, while Barro (1997) uses 87 countries, Knack and Keefer (1995) use 46 or 97 countries depending on the specification. An alternative approach is to divide the sample by income level using the World Bank country classification. (Our data include 29 developed countries and 85 developing nations according to this classification scheme.) The rationale for this approach is that the theoretical literature suggests that the developmental impact of institutional variables on growth depends upon the level of development in each of the democracy variables also obtains insignificant and positive coefficients for trade shares. 9

13 country. The estimated results for specifications segregating countries by income level are reported in Table 5. As shown by the results reported in the first Panel of Table 5, dividing the sample according to income levels obtains different implications for the rule of law variables. Now, while the estimated coefficient for overall maintenance of the rule of law is significant for developed countries, the estimated coefficients for government repudiation of contracts and risk of expropriation are significant for developing countries. This suggests that rule of law measures have different implications for countries at different stages of development. Estimates of the effect of democracy variables when countries are divided by income level are reported in the second Panel of Table 5. The results are similar to those previously reported. In all cases, for both developing and developed countries, the estimated coefficients indicate insignificant positive effects. Our results show that while some measures of institutional quality have a strong relationship with growth, democracy measures have no apparent relationship with growth, even though there is a very high correlation between these institutional measures that supposedly measure similar factors. The results reported so far are standard in the literature. For example, Knack and Keefer (1995) report similar results for these measures and conclude that the rule of law is a better measure of institutional quality than the Gastil indices. Although this may be true, we consider the possibility that the considerably different country and time coverage between the democracy measures and the rule of law measures may account for the weak correlation between democracy and growth. To investigate this possibility, we use the same specification as in column 2 of Table 3 with the same sample of countries and with the same time period (87 countries in the 1980s and 101 countries in the 1990s) for both institutional measures, and reestimate the regressions in Table 4. A simple examination of the See also Rodrik et al. (2002) for further discussion of the issue and additional references. 10

14 data reveals that all of the missing observations in the 1980s and 1990s (with the exception of Bahamas in the 1980s) are developing countries. For developed countries both growth rates and democracy measures are almost identical in both the unrestricted and restricted samples. However, growth rates are slightly higher in the 1990s for developing countries in the restricted sample. Developing countries also have slightly higher democracy levels in the restricted sample for two of the decades. Averages for the restricted sample are in the bottom panel of Table 1. The last panel of Table 5 (the restricted sample) reports these estimates. Compared to the unrestricted sample, in the restricted sample the estimated coefficients are substantially larger in absolute value for all countries and for the developing countries. Also, the all-country estimated coefficients for all of the democracy measures except for civil liberties are now positive and statistically significant. These estimates are reported in the first column of Table 5, Panel 3. It is significant that these results are driven by the experiences of the developing countries. These results demonstrate that the estimated effects of democracy measures are extremely sensitive to time periods and country coverage. This can explain the contradictory results reported throughout the literature. More importantly, these results challenge the idea that overall maintenance of rule of law is more important than democratic institutions for growth, especially for developing countries. The results presented here imply that both institutional types are important (or probably not important if we consider the unrestricted model) for economic growth. On the whole, our estimated results support the idea that the rule of law and democracy are equally important conditions for growth. These results also shed light on the widely debated topic as to whether any systematic difference exists between the growth performance of democratic regimes and undemocratic regimes. Our estimates 11

15 indicate that democratic countries grow faster, as suggested by the signs and magnitudes of the estimated coefficients for the various measures of democracy. V. Sensitivity Analysis A number of studies emphasize the possibility of simultaneity problems for estimates using institutional variables as explanatory variables. One possible source of bias is that rich countries have more resources to expend on building strong institutions. Another possible source of bias is that the populace demands more political freedom and civil liberty as income and wealth increase. To test whether simultaneity affects the estimated results, one widely used approach is to estimate the model using an instrumental variables technique. However, a major limitation with this approach is that it is difficult to find good instruments that are correlated with the exogenous variables but are not correlated with the error terms. Two studies in the literature suggest possible instruments. Mauro (1995) argues that the index of ethnic fractionalization is a valid instrument for institutional variables and Rodrik (1997) argues that secondary enrollment ratios are a valid instrument for democracy measures8. The correlations between these two variables and the institutional variables are reported in Table 2. Except for the correlation coefficient between regime type and ethnic fractionalization, all of the other coefficients are statistically significant, showing that countries with good institutions have less ethnic fractionalization and higher secondary enrollment ratios. In addition to the current or lagged values of other exogenous variables used in the specifications in Table 3 (or Table 4), these two variables are used as instruments in the three-stage least-squares (3SLS) estimates and the results are reported in Table 6. Comparing the SUR estimates in Table 5 and 3SLS estimates in Table 6 shows that 8 The ethnic fractionalization index is from Easterly and Sewadeh and secondary enrollment ratios are from Easterly (1999). 12

16 there are some differences between these two sets of estimates. Substantially larger, statistically significant coefficients are estimated using the 3SLS technique for all measures of rule of law with the exception of the corruption index. These results suggest a much higher positive impact of the rule of law on growth for all countries and for developing countries. In the unrestricted sample, controlling for the simultaneity obtains larger estimated coefficients for the democracy variables in the developing country sample. For the instrumental variable estimates, statistically significance results for civil liberties and for autocracy indicate that democratic developing countries grow faster than undemocratic developing countries. In the restricted sample, the SUR and 3SLS estimates are similar for the democracy variables, but while eight of the ten SUR estimates are significant, only five 3SLS estimated coefficients are significant for all countries and for developed countries. Overall, these results demonstrate that while controlling for the existence of endogeneity does not change the significance levels of most of the estimated coefficients, the coefficients estimated by 3SLS are, in most cases, larger than the SUR estimates, suggesting a much larger effect of institutional quality on growth. V. Conclusion Previous research has established the importance of maintenance of the rule of law for promoting economic growth, but found no evidence that democracy provided any growth benefit to developing economies. Here, using a larger sample than those used in previous studies, we find comparable results. The rule of law, but not democracy, enhances economic growth. However, we also find that these results are very sensitive to sample selection and estimation methodology. When an identical sample is used to assess the impact 13

17 of the rule of law and democracy, we find that both institutions promote growth. Democratic institutions appear to be particularly important for developing nations. Similarly, controlling for simultaneity is very important. Estimates using threestage least squares find large, significant effects of both the rule of law and democracy on growth. Again, the importance of democratic institutions for developing nations is particularly notable. Contrary to previous findings, we find that democratic institutions promote growth. Developing nations in particular benefit from the establishment of democratic political systems. Given the many nations that remain in early stages of development, the contribution of a democratic society to economic success cannot be overemphasized. 14

18 Table 1 Basic Statistics Developing Countries Developed Countries Variables 1970s 1980s 1990s 1970s 1980s 1990s GDP per capita growth rates (%) Initial GDP per capita (log) Life expectancy rates (log) Telephone mainlines per worker Trade (% of GDP) Rule of law (0-1) Gov't repudiation of contracts (0-1) Risk of expropriation (0-1) Corruption (0-1) Bureaucratic quality (0-1) Political Rights (0-1) Civil liberties (0-1) Democracy (0-1) Autocracy (0-1) Political regime (1-4) Restricted Sample GDP per capita growth rates (%) Political Rights (0-1) Civil liberties (0-1) Democracy (0-1) Autocracy (0-1) Political regime (1-4) For a discussion of these variables, see section 3. 15

19 Table 2 Pearson Correlation Coefficients for Various Institutional Measures a Gov't Repudiation of Contracts Risk of Expropriation Corruption Bureaucratic Quality Political Rights Variable Rule of Law Gov't Repudiation of Contracts Risk of Expropriation Corruption Bureaucratic Quality Political Rights Civil Liberties Democracy Autocracy Political Regime Type Ethnolinguistic Fractionalization Secondary Enrollment Ratios 1.00 a The top number is the simple correlation coeffcient for each decade averages; the bottom number are the P-level for each correlation coefficient. Civil Liberties Democracy Autocracy Political Regime Type Ethnolinguistic Fractionalizatio Secondary Enrollment Ratios 16

20 Table 3 SUR Estimates of the Effects of Rule of Law Measures on Growth Variable Log (Initial GDPSH) -1.92*** -2.79*** -3.11*** -2.99*** -2.32*** -2.61*** (2.43) (3.18) (3.26) (3.24) (2.74) (3.00) Log (Life Expactancy) 14.96*** 17.48*** 14.09*** 13.83*** 18.34*** 16.88*** (3.80) (3.89) (2.66) (2.65) (3.63) (3.39) Telephone mainlines per worker 0.011*** 0.010** 0.009** 0.009* 0.010** 0.010*** (2.64) (2.23) (1.87) (1.84) (2.33) (2.38) Trade (% of GDP) 0.009*** (2.42) (1.30) (1.03) (0.84) (1.37) (1.32) Rule of Law 2.06*** (2.43) Government Repudiation 3.52** of Contracts (2.28) Risk of Expropriation 3.89** (2.84) Corruption (0.50) Bureaucratic Quality 0.79 (1.31) AFRICA -1.11** -1.45** -2.33*** -2.37*** -1.59*** -1.78** (1.92) (2.15) (3.28) (3.45) (2.53) (2.87) LATIN -2.26*** -2.34*** -2.23*** -2.33*** -2.42*** -2.32*** (4.61) (4.72) (4.50) (4.74) (5.10) (4.90) EASIA ** 1.01** (1.41) (1.41) (1.16) (1.06) (2.20) (2.06) OECD -1.32** -1.94*** -2.05*** -2.16*** -1.12* -1.30** (2.09) (2.78) (2.74) (2.91) (1.84) (2.14) For each equation, R 2.45,.27.59,.24.63, ,.30.59,.29.63,.26 (N) Each specification is estimated using the SUR technique. The system has 2 equations, where the dependent variables are the per capita growth rate for each decade. Each equation has a different constant term (not reported here). Other coefficients are restricted to be the same for all periods. t-statistics are in parentheses. *** Significant at the 1 percent-level ** Significant at the 5 percent-level 17

21 Table 4 SUR Estimates of Democracy Measures on Growth Variable Log (Initial GDPSH) -2.49*** -2.60*** -2.57*** -2.37*** -2.33*** -2.37*** (3.66) (3.76) (3.71) (3.28) (3.25) (3.28) Log (Life Expactancy) 17.51*** 17.15*** 17.38*** 16.52*** 16.79*** 16.58*** 5.38) (5.23) (5.32) (4.77) (4.91) (4.78) Telephone mainlines per worker 0.010*** 0.010*** 0.010*** 0.009*** 0.009*** 0.009*** (3.01) (2.96) (3.26) (2.58) (2.64) (2.61) Trade (% of GDP) 0.011*** 0.011*** 0.010*** 0.012*** 0.011*** 0.011*** (3.37) (3.29) (3.26) (3.32) (3.26) (3.17) Political Rights 0.46 (0.90) Civil Liberties 0.36 (0.60) Democracy 0.56 (1.15) Autocracy (1.33) Political Regime Type (1.52) AFRICA -1.41*** -1.41*** -1.41*** -1.44*** -1.32*** -1.37*** (2.73) (2.72) (2.72) (2.66) (2.48) (2.56) LATIN -2.16*** -2.23*** -2.23*** -2.13*** -2.18*** -1.91*** (4.81) (4.90) (4.80) (4.51) (4.60) (3.99) EASIA (0.86) (0.87) (0.84) (1.24) (1.28) (1.38) OECD -1.30** -1.42*** -1.41** -1.57*** -1.58*** -1.29** (2.29) (2.44) (2.35) (2.60) (2.63) (2.21) For each equation, R 2.34,.42.33,.43.34,.42.32,.42.31,.43.32,.41 (N).28,(342).29,(342).28,(342).30,(321).29,(321).30,(321) Each specification is estimated using the SUR technique. The system has 3 equations, where the dependent variables are the per capita growth rates during each decade. Each equation has a different constant term (not reported here). Other coefficients are restricted to be the same for all periods. t-statistics are in parentheses. *** Significant at the 1 percent-level ** Significant at the 5 percent-level 18

22 Table 5 SUR Estimates of the Effects of Rule of Law and Democracy Measures on Growth Countries Divided by Income Level All Countries Developing Countries Developed Countries Variable R 2 R 2 R 2 SUR # of obs SUR # of obs SUR # of obs Rule of Law 2.07***.59, , **.82,.65 (2.43) 188 (1.45) 131 (2.19) 57 Gov't Repudiation 3.52**.63, ***.67, of Contracts (2.28) 155 (3.26) 120 (0.15) 35 Risk of Expropriation 3.89**.65, ***.65, (2.84) 155 (2.90) 120 (1.18) 35 Corruption , , ,.71 (0.50) 175 (0.74) 120 (0.52) 55 Bureaucratic Quality , , ,.69 (1.31) 175 (0.39) 120 (0.64) 55 Unrestricted Sample Political Rights ,.43, ,.32, ,.66,.67 (0.90) 342 (0.65) 255 (1.18) 87 Civil Liberties ,.42, ,.31, ,.66,.66 (0.60) 342 (0.67) 255 (0.90) 87 Democracy ,.42, ,.31, ,.68,.58 (1.15) 321 (0.83) 240 (0.18) 81 Autocracy ,.43, ,.32, ,.67, 58 (1.33) 321 (0.97) 240 (0.17) 81 Political Regime Type ,.41, ,.30, ,.68,.59 (1.52) 321 (1.08) 240 (0.65) 81 Restricted Sample Political Rights 1.35**.55, **.47, ,.71 (2.08) 188 (2.13) 131 (0.37) 57 Civil Liberties , **.47, ,.72 (1.54) 188 (1.99) 131 (0.74) 57 Democracy 0.99*.51, , ,.64 (1.70) 180 (1.49) 126 (0.50) 54 Autocracy -1.50**.53, *.45, ,.63 (2.15) 180 (1.78) 126 (0.01) 54 Political Regime Type -0.74*.50, *.41,.30 no variation (1.90) 180 (1.84) 126 See notes for Tables 3 and 4. 19

23 Table 6 3SLS Estimates of the Effects of Rule of Law and Democracy Measures on Growth: Countries Divided by Income Level All Countries Developing Countries Developed Countries Variable R 2 R 2 R 2 3SLS # of obs 3SLS # of obs 3SLS # of obs Rule of Law 7.99***.55, ***.53, ***.84,.45 (4.25) 155 (2.61) 103 (3.27) 52 Gov't Repudiation 10.92***.58, ***.69, of Contracts (2.60) 129 (3.61) 97 (0.09) 32 Risk of Expropriation 7.71***.60, ***.64, (2.51) 129 (2.67) 97 (0.73) 32 Corruption , , ,.66 (0.93) 147 (1.01) 97 (0.96) 50 Bureaucratic Quality 7.11***.66, ***.64, ,.63 (4.10) 147 (2.65) 97 (0.27) 50 Unrestricted Sample Political Rights ,.52, ,.46, ,.70,.56 (0.13) 252 (1.22) 177 (0.05) 75 Civil Liberties ,.52, **.21,.45, ,.70,.56 (1.01) 252 (2.21) 177 (0.04) 75 Democracy ,.49, ,.45, ,.70,.52 (0.24) 246 (1.47) 174 (0.83) 72 Autocracy ,.52, **.13,.49, ,.71,.53 (1.42) 246 (2.10) 174 (0.90) 72 Political Regime Type ,.48, ,.40, ,.70,.55 (0.15) 246 (0.02) 174 (1.58) 72 Restricted Sample Political Rights , **.46, ,.65 (1.10) 155 (2.03) 103 (0.82) 52 Civil Liberties , , ,.67 (0.23) 155 (1.56) 103 (1.43) 52 Democracy 2.92*.48, ***.43, ,.65 (1.70) 151 (2.75) 101 (0.70) 50 Autocracy -3.68**.52, ***.51, ,.63 (2.00) 151 (2.74) 101 (0.23) 50 Political Regime Type , ,.37 no variation (0.98) 151 (1.25) 101 The instruments are the five-year earlier log of (GDPSH) (for example, 1965 in the equation); five-year lagged values of log (LIFE) (for example, in the equation); actual values of TELPW, and previous five-year values of OPEN are used. For example, the equation uses averages for OPEN for the period. Finally, secondary enrollment ratios and an index of ethnolinguistic fractionaliaziton in 1960 are used in all specifications. 20

24 References Bardhan, Pranab. (1997). Corruption and Development: A Review of Issues, Journal of Economic Literature, Vol. 35 (September), Barro, Robert J. (1997). Determinants of Economic Growth: A Cross-Country Empirical Study. Cambridge and London: MIT Press. Barro, Robert J. (1999). Determinants of Democracy, The Journal of Political Economy, 107:6, S158-S183. Brunetti, A. (1997). Political Variables in Cross-Country Growth Analysis Journal of Economic Surveys, 11, Chowdhurie-Aziz, Monali. (1997). Political Openness and Economic Performance, unpublished paper, University of Minnesota, January Collier, Paul. (2000). Ethnicity, Politics, and Economic Performance, Economics and Politics, 12(3), Dollar, David, Aart Kraay. (2000). Property Rights, Political Rights, and the Development of Poor Countries in the Post-Colonial Period, World Bank Working Papers. Durham, Ebneson, J., (1999). Economic Growth and Political Regimes, Journal of Economic Growth, 4: Easterly, William. (1999). Life during Growth, Journal of Economic Growth, 4, Easterly, William. (2001). Can Institutions Resolve Ethnic Conflict?" Economic Development and Cultural Change, 49(4), Easterly, William, and Mirvat Sewadeh. (2002) Global Development Network Growth Database on the World Bank web site. Ehrlich, Isaac, Francis T. Lui. (1999). Bureaucratic Corruption and Endogenous Economic Growth, Journal of Political Economy, 107:6, S270-S293. Gastil, Raymond D. (1988). Freedom in the World: Political Rights and Civil Liberties New York: Freedom House. Greene, William H. (1997). Econometric Analysis. New Jersey: Prentice Hall. Helliwell, John. (1994). Empirical Linkages Between Democracy and Economic Growth, British Journal of Political Science, 24, Kaufmann, Daniel, Aart Kraay, and Pablo Zoido-Lobatón (1999). Aggregating Governance Indicators, World Bank Policy Research Department Working Paper No Knack, Stephen, Philip Keefer. (1995). Institutions and Economic Performance: Cross-Country Tests Using Alternative Institutional Measures, Economics and Politics, 7, Mauro, Paolo. (1995). Corruption and Growth, Quarterly Journal of Economics, Vol. 110 (August), North, Douglas. (1990). Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. North, Douglas, B. Weingast. (1989). Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England, Journal of Economic History, 49,

25 Nuxoll, Daniel A. (1994). Differences in Relative Prices and International Differences in Growth Rates, American Economic Review, 84, Polity III web site: Przeworski, A., F. Limongi. (1993). Political Regimes and Economic Growth, Journal of Economic Perspectives, 7, Rodrik, Dani. (1997). Democracy and Economic Development, Harvard University, Unpublished paper. Rodrik, Dani. (1999). "Where Did All Growth Go? External Shocks, Social Conflict, and Growth Collapses", Journal of Economic Growth, 4 (4), Rodrik, Dani. (2000). Institutions For High-Quality Growth: What They Are And How To Acquire Them, NBER Working Papers No Rodrik, Dani, Arvind Subramanian, Francesco Trebbi. (2002). Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development, Harvard University, Unpublished paper. Sala-i-Martin, Xavier X. (1997). I Just Ran Two Million Regressions, American Economic Review, 87, Shleifer, Andrei, Robert W. Vishny. (1993). Corruption, Quarterly Journal of Economics, Vol. 108 (August), pp Summers, Robert, Alan Heston. (1991). The Penn World Table (Mark 5): An Expanded Set of International Comparisons, , Quarterly Journal of Economics, 106, Tanzi, Vito. (1998). Corruption around the World: Causes, Consequences, Scope, and Cures, IMF Staff Papers, 45:4, Tavares, Jose, Romain Wacziarg. (2001). How Democracy Fosters Growth, European Economic Review, 45,