JAMES S. GUSEH Shaw University Raleigh, North Carolina

Government Size and Economic Growth in Developing Countries: A Political-Economy Framework * Many authors have examined the impact of growth in government size on economic growth. These studies, however, have not distinguished the impact among political and economic systems. This paper presents a model that differentiates the effects of government on economic growth across political and economic systems in developing countries. The results show that growth in government size has negative effects on economic growth, but the negative effects are three times as great in nondemocratic socialist systems as in democratic market systems.

1. Introduction Does a large government size promote or hinder economic growth? Although the issue is an important one, a variety of conflicting theoretical explanations has been advanced that can only be resolved through empirical investigations. Yet the results of the investigations conducted to date have been diverse and contradictory. Ram (1986, 1989) and Rubinson (1977) concluded that a large government size promotes economic growth, while Landau (1983, 1986) and Barro (1991) concluded that it depresses growth of per capita income. Conte and Darrat (1988) reported that changes in economic growth are not affected by public sector expansion, but Gemmell (1983) found that nonmarket sector growth has adverse macroeconomic effects that vary strongly from country to country. Bairam (1990) found positive effects for some countries and negative effects for others. Grossman (1988, 1990) concluded that government contributes both positively and negatively to economic growth, but the net effect appears to be marginally negative. The diversity of results, combined with the fundamental importance of the subject, necessitates not only further research but also the use of alternative methodologies. *An earlier version of this paper was presented at the Southwestern Economic Association Conference. I would like to thank Brian J. Berry, Irving J. Hoch, and Wim Vijverberg for their helpful comments.

Journal of Macroeconomics, Winter 1997, Vol. 19, No. 1, pp. 175–192 Copyright 䉷 1997 by Louisiana State University Press 0164-0704/97/$1.50

175

James S. Guseh This paper proposes an alternative theoretical framework based on the conventional neoclassical production function and tests it using a fixed effects model on fifty-nine middle-income developing countries over the period 1960–85. Structural differences, such as natural resource and human capital endowments, among countries can cause their levels of economic performance to differ significantly. These differences usually warrant the use of a separate regression equation for each country. However, the fixed effects model improves on this technique by controlling for the intertemporal dynamics and individuality of the entities being investigated. None of the previous studies controlled for both factors.1 To further extend the perspective of prior work, this study assesses the differential effects of government on growth among various political and economic systems. Although other studies have examined the relationship between politico-economic systems and economic performance (Barro 1991; Pourgerami 1988; Scully 1988), none has considered the dynamics of these systems over time. Instead, they have implicitly assumed that the political and economic systems of each country are constant. This study challenges this assumption, because over the years some countries have changed from one system to the other. To address this concern, the political and economic systems of each country in each year over the period 1960–85 are included in the analysis. Thus, a beginning is made of the study of economic growth that considers the intertemporal dynamics of political and economic systems. Finally, one debate in the literature centers on determining the appropriate specification of government size in assessing its impact on economic growth. Several studies have specified government size as G/Y, the share of government consumption expenditure (G) in gross domestic product (Y) (Landau 1983, 1986; Rubinson 1977). The impact of this specification on growth has generally been negative. Ram (1986), however, has argued that the appropriate specification of government size is the growth rate of government consumption expenditure, dG/G (or growth in the relative size of government expenditure, dG/Y). Its impact on growth has been found to be positive. Conte and Darrat (1988) have shown that both specifications are appropriate, with Ram’s specification measuring the short-run impact of government and the other specification measuring the long-run impact. Following the distinction made by Conte and Darrat, this study will focus on assessing the long-run impact of government, but will also test the impact of Ram’s specification. Such a richer model, it is hoped, will provide further insights into the role of government in economic growth, as well as provide information on 1 Landau (1985) attempted to control for the intertemporal dynamics, but only limited it to years that were either the contraction or recovery phase of a business cycle.

176

Government Size and Economic Growth optimal development strategies. The study concludes that growth in government size is negatively associated with economic growth, but the negative effects are greater in nondemocratic socialist systems than in democratic market systems. Section 2 outlines the theoretical framework. Section 3 presents the empirical results. Some specification issues are discussed in Section 4. The summary and conclusions are presented in Section 5.

2. The Theoretical Framework The basic model is an adaptation of the neoclassical production function and is derived as follows. Let an economy be described by the aggregate production function which is homogeneous of degree one, with only two factors of production, so that Y ⳱ f (K, L) ,

(1)

where Y is national economic output measured as gross domestic product (GDP); K is the stock of capital; and L is the labor force. Dividing Equation (1) by population (P), the production function becomes ´ L) ´ , Y´ ⳱ f (K,

(2)

where Y´ is GDP per capita, K´ is the capital-labor ratio, and L´ is the labor force participation rate. Since data on the labor force in developing countries are difficult to obtain, by convention population is used as a proxy for the labor force. Adapting the function in (2) to include government-size (G), measured as the share of government consumption expenditure in GDP2, yields ´ L, ´ G) . Y´ ⳱ f (K,

(3)

Differentiating the function in (3), dividing by Y´, manipulating the 2 Government consumption expenditure is defined as the sum of (i) purchases, less sales, of consumer goods and services by central, regional and local governments, reduced by the value of the own-account production of fixed assets, (ii) compensation of employees, (iii) compensation of fixed assets, and (iv) any payments of indirect taxes (World Bank 1989).

177

James S. Guseh results, and letting the results reflect pooled cross-sectional and time-series data yields the growth equation for the basic model as3 yit ⳱ b1kit Ⳮ b2git ,

(4)

where the lower case letters denote the growth rates of the relevant variables for country i in year t; and b1 and b2 are the elasticities of GDP per capita with respect to capital-labor ratio and government size, respectively. The lag of per capita income growth is also included in the model to address any partial adjustment in the economy, such as inertia. The variable used to measure capital in the capital-labor ratio is gross domestic investment since data on the growth rate of capital stock are more difficult to find for developing countries. Moreover, this variable is used only as a control variable since the focus of the study is to assess the impact of government on economic growth.4 To test for the differential effects of government across political and economic systems, this study adapts Gastil’s (1973, 1986) country rankings of political and economic systems. Gastil annually publishes country rankings of political and civil liberties, types of economic systems, and other measures of freedom. He derives his classification of political institutions from political rights and civil liberties. Political rights are rights to participate meaningfully in the political process. In a democracy, this means the right of all adults to vote and compete for public office, and for elected representatives to have a decisive vote on public policies. Civil liberties are rights to free expression, to organize or demonstrate, as well as to a degree of autonomy such as provided by freedom of religion, education, travel, and other personal rights. Political rights and civil liberties are both rated on a sevenpoint scale with 7 being the least free or least democratic and 1 being the most free or most democratic. From these two ratings, Gastil derives the annual status of political freedom for each country as either free, partially free, or not free. With respect to economic institutions, Gastil ranks countries from capitalist to socialist, with various mixtures of capitalism and socialism in be3 This specification can also be obtained by writing the neoclassical production function (Y ⳱ K, L) in the intensive form: Y´ ⳱ f (K´), where Y´ ⳱ Y/L, and K´ ⳱ K/L. Adapting the function to include government size (G) yields: Y´ ⳱ f (K´, G). Differentiating the function and manipulating the results yields: yit ⳱ b1kit Ⳮ b2git, where lower case letters denote the growth rates of the relevant variables. 4 The difficulty in obtaining data on the growth rate of capital stock is sometimes dealt with by specifying the growth model to include the share of gross domestic investment in GDP (i.e., I/Y). This specification is also estimated to compare the results with those obtained from the model employed. See note 8.

178

Government Size and Economic Growth tween, such as mixed capitalist and mixed socialist economies. For this study, the categories at the extremes are classified as either capitalist or socialist economies, and those in between are classified as mixed economies. This study transforms Gastil’s classifications of political freedoms and economic systems into dummy variables and applies these indices to each country over the entire sampling period. By way of summary, the variables for political systems are free, partially free, and unfree societies; the variables for economic systems are market, mixed, and socialist economies; k is the growth of the capital-labor ratio; g is the growth rate of government size; y is the growth rate of per capita income; and ytⳮ1 is the lagged value of the growth rate of per capita income. The classifications of the political and economic systems of the countries over the sampling period are reported in Appendix A.

3. Empirical Evidence The model is estimated by ordinary least-squares regression method using annual time-series data for the period 1960–85 for fifty-nine middleincome developing countries, as classified by the World Bank (1984: xxxiii– xxxv).5 All annual rates of growth are approximated by first differencing the logarithms of the variables for successive years. Unless stated otherwise, data are derived from the World Bank (1989). Monetary variables are in constant 1980 prices. To determine the extent to which the parameter estimates of the variables, especially the government-size variable, are influenced by the time and cross-sectional units of the fixed effects model, the model is estimated first without, and then with, these unit effects. The results are reported in Table 1. The coefficient of the linear term of the government-size variable in each equation is negative. Its absolute value decreases from 0.180% without any unit effects to 0.175% with only the time effects, to 0.148 with only the cross-sectional effects, and to 0.143% with both unit effects. Each coefficient is statistically significant at the 1% level. Clearly, the coefficient of government size is influenced by the effects of the year and country dummies.6 This is to be expected, because inclusion of these time and cross5

The World Bank (1984: xxxiii–xxxv) has classified developing countries into two categories, based on GNP per capita in 1981 U.S. dollars. Low-income countries have a GNP per capita of $405 or less, and middle-income countries have a GNP per capita between $405 and $6,900. The sample of countries was determined by the availability of data over the sample period. 6 The hypotheses of whether the time effects, country effects, or both effects taken together, are jointly equal to zero were tested. F-tests yielded F values of 2.27, 5.00, and 2.92, respectively, which are significant at the 1% level. Thus, we reject the hypotheses that the unit effects are jointly equal to zero.

179

James S. Guseh TABLE 1. Regression Results of the Impact of Government Size on Economic Growth Dependent Variable: Growth Rate of Per Capita GDP ( y)

Variable

With No Unit Effects

With Time Effects

With Country Effects

With Both Effects

1.6356c (11.598) k 0.1027c (16.794) 0.2514c ytⳮ1 (11.499) g ⳮ0.1797c (7.246) First order interaction with g1: Free 0.0899b (2.543) Not Free ⳮ0.0075 (0.247) Market ⳮ0.0146 (0.485) Socialist ⳮ0.1121c (3.124)

0.0663 (0.109) 0.0955c (15.412) 0.2349c (10.629) ⳮ0.1754c (7.185)

0.8172 (0.847) 0.1005c (16.609) 0.1741c (7.641) ⳮ0.1481c (5.863)

ⳮ1.1592 (1.055) 0.0913c (14.945) 0.1459c (6.341) ⳮ0.1434c (5.809)

0.0855b (2.466) ⳮ0.0183 (0.613) ⳮ0.0181 (0.608) ⳮ0.1099c (3.133)

0.0759b (2.131) ⳮ0.0386 (1.254) ⳮ0.0337 (1.116) ⳮ0.1372c (3.738)

0.0696b (2.007) ⳮ0.1518a (1.725) ⳮ0.0368 (1.243) ⳮ0.1336c (3.749)

R-SQUARE ADJ R-SQ DW Durbin h

0.4195 0.4060 2.03 ⳮ0.97

0.4265 0.3968 2.05 ⳮ1.06

0.4755 0.4380 2.03 ⳮ1.06

Constant

0.3748 0.3714 2.05 ⳮ0.97

NOTES: Estimates of the year and country effects are reported in Figure 1 and the Appendix, respectively, to save space. The t-statistics are in parentheses. 1 Partially free and mixed economy are the base categories. a ⳱ Significant at the 0.10 level. b ⳱ Significant at the 0.05 level. c ⳱ Significant at the 0.01 level.

sectional units affects the other variables in the X vector (Hoch 1962; Sayrs 1989), as seen in the reduced coefficients of the government-size variable. These results suggest that growth in government size has a significantly negative impact on the underlying rate of economic growth. Using the estimated model with both the time and cross-section effects in Table 1, the 180

Government Size and Economic Growth coefficient of the additive term of the government-size variable indicates that a 1% increase in government size decreases the rate of economic growth by 0.143%, ceteris paribus. We now investigate whether the impact of government size on growth differs among the various political and economic systems. Interaction terms for the government-size variable and the Gastil indices for political and economic systems allow us to test these hypotheses. Using estimates in the model with both the time and country effects in Table 1,7 the interaction variable for government size and free status is positive and statistically significant. This suggests that the negative impact of government on growth is lower in free societies than in partially free societies. On the other hand, the interaction term for government size and unfree status is negative and statistically significant. This suggests that the negative impact of government on growth is greater in unfree societies than in partially free societies. The interaction term for government size and market economy is negative but not significant, while the interaction term for government and socialist economy is negative and highly significant. These results indicate that the impact of government does not differ significantly between market and mixed economies, but differs significantly between mixed and socialist economies, with the negative impact being greater in socialist economies. The differential effects of government across political and economic systems are combined to arrive at a matrix of the impact of government size on growth across the spectrum of political economies. The results are reported in Table 2. The matrix allows us to compare the impact of government size among various combinations of political and economic systems. The results show a clear and consistent pattern as a country moves from one system to another. As a society moves from a politically free to an unfree status, the negative impact of government increases. Similarly, as a society moves from a market economy to a socialist economy, the negative impact increases. Thus, countries with political and economic freedoms have the lowest adverse effects of government, while countries without these freedoms have the highest adverse effects. Moreover, among the three economic systems, the mixed economic system appears to consistently perform well. As shown in Table 2, mixed economies that are politically free have the least adverse effects of government. Overall, the negative impact of government in countries with nondemocratic socialist institutions is three times that of countries with democratic market institutions. For example, a 10% increase 7 It is hypothesized that the parameter estimates of the additive terms of the political-economy variables are simultaneously equal to zero. An F-test of this hypothesis yielded F(4, 1286) ⳱ 1.89, which is insignificant. We therefore failed to reject the null hypothesis and proceeded to estimate the model without the additive terms.

181

James S. Guseh TABLE 2. Political-Economy Matrix of the Impact of Government Size on Economic Growth (Percent) Status of Freedom Economic System Market Mixed Socialist

Free

Partially Free

Not Free

ⳮ0.111 ⳮ0.074 ⳮ0.207

ⳮ0.180 ⳮ0.143 ⳮ0.277

ⳮ0.232 ⳮ0.195 ⳮ0.329

NOTE: These effects are obtained from the estimated model in the last column of Table 1 by differentiating the growth rate of per capita GDP ( y) with respect to government size (g), ⳵y/⳵g, and substituting therein the values of the relevant political-economy variables.

in government size yields a 0.74% decline in economic growth in democratic mixed economic systems, a 1.11% decline in democratic market systems, and a 3.29% decline in nondemocratic socialist systems, ceteris paribus. Thus, greater government size takes not only a toll on economic growth, but the type of political and economic systems present in a country affects the magnitude of the toll. These results indicate that growth of government retards economic growth. According to economic theory, government is inefficient in the provision of Pigovian goods and services and creates distortions in the economy through its taxing and spending mechanisms and unproductive rent-seeking activities. These inefficiencies and distortions retard economic growth in any political economic system, but the greatest slowdown is in socialist and nondemocratic systems. Thus, we see the current quest in Eastern Europe for political and economic liberties and the benefits that are likely to be achieved, as well as the drive toward privatization of state-owned enterprises in many developing countries. These conclusions are consistent with Scully’s (1988) finding that politically open societies with market economies grow at three times the rate experienced by countries in which these freedoms are circumscribed. Scully compares how efficiently capital and labor are employed among various political and economic systems. Differences in the rates of economic growth across politico-economic systems found in Scully’s work seem to reflect the differences in the adverse effects of government on growth found in this study. Growth in government size seems to interfere with the efficient allocation of resources leading to a slowdown in economic growth. Both studies, however, differ in several respects. This study compares the impact of government size on growth among political and economic 182

Government Size and Economic Growth systems. It also employs the political and economic systems of each country in each year over the period 1960–85. Scully, on the other hand, compares how efficiently capital and labor are employed among these systems. His variables for political and economic systems are the averages of Gastil’s rankings for the period 1973–80. The fixed effects model yields additional information in the form of estimates of the time- and country-effect parameters. Appendix B presents these estimates. The parameter estimates for the 1960s and 1970s are greater than those for the 1980s, indicating that growth in per capita income was higher in the 1960s and 1970s than in the 1980s. With respect to the cross-sectional units, about a third proved significant coefficients, meaning that the model fits quite well for most of the countries. The countries for which the model does not fit include Egypt, Jordan, South Korea, Kuwait, Libya, Malta, Portugal, Singapore, Thailand, and Yemen. They appear to be countries that have prospering manufacturing sectors or are endowed with natural resources, such as oil. Major oil producing countries, such as Kuwait and Venezuela, seem not to have done as well as one would expect over the period of our study based on common perception, as indicated by their negative coefficients relative to the base category, Liberia.

4. Specifications Tests Preliminary econometric tests of the assumptions of no autocorrelation and heteroscedasticity reveal that the estimates are robust. With a lagged dependent variable as an independent variable, the Durbin-Watson statistic is biased toward accepting the null hypothesis of no autocorrelation. Using the appropriate Durbin-h test, we also fail to reject the hypothesis. Additionally, an examination of the residuals did not reveal heteroscedasticity of the error term. Should this assumption be violated in a pooled data set, one of the best ways to address it is by introducing a fixed value that represents the variance unique to the cross-section and conditional on the sample. One way to condition the variance in least squares is by using a dummy variable (Sayrs 1989), as in the model employed in this study. Finally, a debate in the literature centers on the appropriate specification of government size. Besides the share of government consumption expenditure in GDP, the measure of government size employed in this study, another specification is the growth rate of government spending, dG/G, proposed by Ram (1986). We test this specification to determine its impact on growth across political and economic systems. All variables in the model employed in Table 1 are retained, except the government-size variable g, 183

James S. Guseh TABLE 3. Regression Results of the Impact of Government Size on Economic Growth Using Ram’s Specification Dependent Variable: Growth Rate of Per Capita GDP ( y)

Variable

No Unit Effects

With Time Effects

1.0036c 0.0574 (5.948) (0.088) k 0.1040c 0.1018c (15.702) (15.192) dG/G 0.0300 0.0234 (1.218) (0.960) ytⳮ1 0.2398c 0.2339c (10.148) (9.762) First-order interaction with dG/G1: Free 0.0650c 0.0619c (1.946) (1.880) Not Free 0.0698c 0.0606c (2.367) (2.082) Market 0.0072 0.0064 (0.251) (0.227) Socialist 0.0196 0.0283 (0.566) (0.830)

Constant

R-SQUARE ADJ R-SQ

0.2812 0.2774

0.3236 0.3079

With Country Effects

With Both Effects

ⳮ0.1865 (0.179) 0.1007c (15.336) 0.0365 (1.328) 0.1557c (6.328)

ⳮ1.4441 (1.209) 0.0968c (14.589) 0.0276 (1.018) 0.1398c (5.588)

0.0582 (1.571) 0.0730c (2.226) ⳮ0.0058 (0.175) 0.0358 (0.867)

0.0538 (1.476) 0.0596c (1.846) ⳮ0.0057 (0.175) 0.0465 (1.151)

0.3403 0.3062

0.3849 0.3410

NOTE: The estimates for the year and country effects are not reported to save space. 1 Partially free and mixed economy are the base categories. a ⳱ Significant at the 0.10 level. b ⳱ Significant at the 0.05 level. c ⳱ Significant at the 0.01 level.

which is replaced by Ram’s specification, dG/G. The results are presented in Table 3. Using the results of the model with both time- and country-effect parameters, as done in the previous analysis, the coefficient of the linear term of government size, dG/G , is positive, but is not statistically significant. The effect of government also does not differ significantly among political and economic systems, except in the case of unfree societies where the effect is marginally higher (b ⳱ 0.06, t ⳱ 1.8) than in partially-free societies. The implications of these results are discussed next. 184

Government Size and Economic Growth It has been reported that Ram’s specification is suitable for testing short-run growth effects, while the specification employed in this study assesses the effects of public sector expansion on the underlying growth rate (Conte and Darrat 1988). Growth and development is a long-run concept, while management of aggregate demand, a Keynesian prescription, is basically a short-term concept and thus is not economic growth. Therefore, while short-term measures of government spending through management of aggregate demand have a positive impact on an economy, the impact of government on the underlying growth rate is negative and differs significantly among political and economic systems as found in this study.8

5. Summary and Conclusions Many authors have investigated the relationship between growth in government size and economic growth. These studies, however, have not distinguished the differential effects of government size on growth across political and economic systems. This paper therefore examines this issue in the case of middle-income developing countries. A model which differentiates the effects of government on growth across political and economic institutions is developed and tested using annual time-series data for fifty-nine middle-income developing countries over the period 1960–85. The results suggest that growth in government size has adverse effects on economic growth in developing countries, but the adverse effects are three times as great in countries with nondemocratic socialist systems as in countries with democratic market systems. 8 Ram’s model has been challenged for being biased in the specification of government (Carr 1989) and for omitted variables (Rao 1989). According to Ram’s (1989, 284) reply, “richer models or additional evidence could certainly lead to different results.” This is confirmed by our study. Another specification issue involves the use of gross domestic investment instead of capital in the capital-labor ratio. As stated in note 4, this issue is often dealt with by specifying the growth model to include I/Y, the share of gross domestic investment in GDP. We estimate this specification to compare the results with those obtained from the model employed in this study. The estimation results, with the corresponding absolute values of the t-statistics in parentheses, are as follows:

y ⳱ ⳮ2.8315 Ⳮ 0.1037(I/Y) ⳮ 0.5475p Ⳮ 0.1408ytⳮ1 ⳮ 0.1582g Ⳮ 0.0530g*free (1.939) (5.123) (2.136) (5.596) (5.938) (1.399) ⳮ 0.0604g*not free ⳮ 0.0448g*market ⳮ 0.1280g*soc. (1.852) (1.391) (3.300) ADJ R-SQ: 0.3468 F Value: 8.887 The effects of government on growth from this model and the one employed in this study are comparable. Thus, the results of the model employed in our study are robust.

185

James S. Guseh Thus, greater government size takes not only a toll on economic growth, but the type of political and economic systems present in a country affects the magnitude of the toll. In light of what we have discovered, it appears that an appropriate policy prescription for economic growth and development should include a reduction in government size and the promotion of economic and political liberties. Received: January 1993 Final version: January 1996

References Bairam, Erkin. “Government Size and Economic Growth: The African Experience, 1960–85.” Applied Economics 22 (1990): 1427–35. Barro, Robert. “Economic Growth in a Cross Section of Countries.” The Quarterly Journal of Economics 106 (1991): 407–43. Carr, Jack. “Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data: Comment.” The American Economic Review 79 (1989): 267–71. Conte, Michael, and Ali Darrat. “Economic Growth and the Expanding Public Sector.” Review of Economics and Statistics 70 (1988): 322–30. Gastil, Raymond. Freedom in the World. Westport: Greenwood, 1986. ———. “The New Criteria of Freedom.” Freedom at Issue 17 (1973): 2–23. Gemmell, Norman. “International Comparison of the Effects of Non-Market Sector Growth.” Journal of Comparative Economics 7 (1983): 368–81. Grossman, Philip. “Government and Economic Growth: A Non-Linear Relationship.” Public Choice 56 (1988): 193–200. ———. “Government and Growth: Cross-Sectional Evidence.” Public Choice 65 (1990): 217–27. Hoch, Irving. “Estimation of Production Function Parameters Combining Time-Series and Cross-Section Data.” Econometrica 30 (1962): 34–53. Landau, Daniel. “Government Expenditure and Economic Growth: A CrossCountry Study.” Southern Economic Journal 49 (1983): 783–92. ———. “Government Expenditure and Economic Growth in the Developed Countries: 1952–76.” Public Choice 47 (1985): 459–77. ———. “Government and Economic Growth in the Less Developed Countries: An Empirical Study for 1960–80.” Economic Development and Cultural Change 35 (1986): 35–75. Pourgerami, Abbas. “The Political Economy of Development: A Cross-National Causality Test of Development-Democracy-Growth Hypothesis.” Public Choice 58 (1988): 123–41. 186

Government Size and Economic Growth Ram, Rati. “Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data.” American Economic Review 76 (1986): 191–203. ———. “Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data: Reply.” The American Economic Review 79 (1989): 281–84. Rao, V. Bhanoji. “Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data: Comment.” The American Economic Review 79 (1989): 272–80. Rubinson, Richard. “Dependency, Government Revenue, and Economic Growth, 1955–70.” Studies in Comparative International Development 12 (1977): 3–28. Sayrs, Lois. Pooled Time Series Analysis. London: Sage Publications, 1989. Scully, Gerald. “The Institutional Framework and Economic Development.” Journal of Political Economy 96 (1988): 652–62. World Bank. World Tables, 3d. ed. Baltimore: John Hopkins, 1984. ———. World Tables. Baltimore: John Hopkins, 1989.

187

Appendix A Country Ratings of Political and Economic Systems: 1961–85 Code

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

ARG BOL BRA BWA CHL CIV CMR COG COL CRI DOM DZA ECU EGY FJI GAB GRC GTM HND HUN IDN ISR JAM JOR KEN KOR KWT LBR LBY

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

P/X P/X P/S P/M F/M N/M P/M N/S F/M F/M F/M N/S P/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M N/M P/X N/M N/S

F/X P/X P/S F/M N/M N/M P/M P/S F/M F/M F/M N/S N/M N/S F/M N/M N/X F/M P/M N/S P/X F/X F/X N/M P/M P/M P/X N/M N/S

P/X N/X P/S F/M N/M N/M P/M P/S F/M F/M P/M N/S N/M P/S F/M N/M F/X P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

P/X N/X P/S F/M N/M N/M P/M P/S F/M F/M P/M N/S N/M P/S F/M N/M F/X P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

N/X P/X P/S F/M N/M N/M N/M P/S F/M F/M P/M N/S P/M P/S F/M N/M F/X P/M P/M N/S P/X F/X F/X N/M P/M N/M N/X P/M N/S

N/X P/X P/S F/M N/M N/M N/M N/S F/M F/M P/M N/S P/M P/S F/M N/M F/M P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

N/X P/X P/S F/M N/M N/M N/M N/S F/M F/M F/M N/S P/M P/S F/M N/M F/M P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

N/X P/X P/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/M P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

N/X N/X P/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/M P/M P/M N/S P/X F/X F/X N/M P/M P/M P/X N/M N/S

N/X N/X P/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/X N/M P/M N/S P/X F/X F/X N/M P/M P/M P/X N/M N/S

P/X F/X P/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/X N/M P/M N/S P/X F/X F/X N/M P/M P/M P/X P/M N/S

F/X F/X P/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/X P/M F/M P/S P/X F/X F/X P/M P/M P/M P/X P/M N/S

85 F/X F/X F/S F/M P/M P/M N/M N/S F/M F/M F/M N/S F/M P/S F/M N/M F/X P/M F/M P/S P/X F/X F/X P/M P/M P/M P/X P/M N/S

MAR MEX MLT MRT MUS MYS NGA NIC PAN PER PHL PNG PRT PRY SEN SGP SLV SWZ SYR THA TTO TUN TUR URY VEN YEM YUG ZAF ZMB ZWE

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S N/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M F/M P/X P/X N/X N/X P/X P/M N/X P/S N/S P/X F/M P/M N/S P/M F/X N/X P/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M P/M P/X P/X N/X N/X P/X P/M P/X P/S N/S P/X F/M P/M N/S P/M F/X N/X F/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M P/M P/X P/X N/X P/X P/X P/M P/X P/S P/S P/X F/M P/M N/S F/M F/X N/X F/X P/X F/X N/S N/S P/X P/S N/X

P/X P/X F/X N/X F/M P/M P/X P/X N/X P/X P/X F/M F/X N/S P/S P/X P/M P/M N/S N/M F/X N/X F/X N/X F/X N/S N/S P/X P/S N/X

P/M P/X F/X N/X F/M P/M P/X P/X N/X P/X P/M F/M F/X N/S P/S P/X P/M P/M P/S N/M F/M N/X F/X N/X F/X N/S N/S P/X P/S N/X

P/M P/X F/X N/X P/M P/M P/X P/X N/X P/X P/M F/M F/X P/S P/S P/X P/M P/M P/S P/M F/M N/X F/X N/X F/X N/S N/S P/X P/S P/X

P/M P/X F/X N/X P/M P/M F/X P/X P/X P/X P/M F/M F/X P/S P/S P/X P/M P/M P/S P/M F/M P/X F/X N/X F/X N/S N/S P/X P/S P/X

P/M P/X F/X N/X P/M P/M F/X P/X P/X F/X P/M F/M F/X P/S P/S P/X P/M P/M N/S P/M F/M P/X P/X P/X F/X N/S N/S P/X P/S P/X

P/X P/X F/X N/X F/M P/M F/X P/X P/X F/X P/X F/M F/X P/S P/S P/X P/M P/M N/S P/M F/X P/X P/X P/X F/X N/S N/S N/X P/S P/X

P/X P/X P/X N/X F/M P/M F/X P/X P/X F/X P/X F/M F/X P/S P/S P/X P/M P/M N/S P/M F/X P/X P/X P/X F/X N/S P/S P/X P/S P/X

P/X P/X P/X N/X F/M P/M N/X P/X P/X F/X P/X F/M F/X P/S P/S P/X P/M P/M N/S P/M F/X P/X P/X P/X F/X N/S P/S P/X P/S P/X

Sources: Adapted from Gastil (1973, 1986). NOTES: Country codes are defined in Appendix 8. For political systems, F ⳱ free; N ⳱ not free; and P ⳱ partially free. For economic systems, M ⳱ market; S ⳱ socialist; and X ⳱ mixed economy.

P/X P/X P/X N/X F/M P/M N/X P/X P/X F/X P/X F/M F/X P/S P/S P/X P/M P/M N/S P/M F/X P/X P/X F/X F/X N/S P/S P/X P/S P/X

Appendix B TABLE 1. Time-Effect Parameters Variable 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

Parameter Estimate

Absolute t-value

0.9424 2.9376 3.3958 1.0888 2.1600 0.9952 3.8704 2.9350 3.1678 2.8391 3.9386 1.8589 2.4851 1.5792 3.9291 0.6894 2.5894 2.3196 0.4197 0.3735 ⳮ0.3828 0.0623 0.4975

0.977 3.386 3.926 1.257 2.512 1.163 4.532 3.417 3.672 3.336 4.686 2.200 2.917 1.890 4.734 0.827 3.126 2.762 0.503 0.451 0.462 0.075 0.599

NOTE: The year 1985 is the base category.

190

Appendix B TABLE 2. Cross-Sectional Unit Effects Country Code Country DZA ARG BOL BRA BWA CHL CMR COG COL CRI CIV DOM ECU EGY SLV FJI GAB GRC GTM HND HUN IDN ISR JAM JOR KEN KWT LBY MAR MEX MLT MRT MUS MYS NGA NIC PAN

ALGERIA ARGENTINA BOLIVIA BRAZIL BOTSWANA CHILE CAMEROON CONGO COLOMBIA COSTA RICA COTE d’IVOIRE DOMINICAN REP. ECUADOR EGYPT EL SALVADOR FIJI GABON GREECE GUATEMALA HONDURAS HUNGARY INDONESIA ISRAEL JAMAICA JORDAN KENYA KUWAIT LIBYA MOROCCO MEXICO MALTA MAURITANIA MAURITIUS MALAYSIA NIGERIA NICARAGUA PANAMA

Parameter Estimate Absolute t-value 1.5745 ⳮ0.1080 0.3002 1.5575 4.5619 ⳮ0.1602 1.1188 2.2624 1.2927 0.8625 1.1885 0.3402 2.4335 3.4569 0.0975 0.9554 1.7094 2.4521 0.4390 0.1459 2.1544 1.6197 1.9308 0.3323 2.9043 1.3110 ⳮ6.1223 4.0456 1.0877 1.6897 3.9434 ⳮ1.0158 1.0351 2.2645 0.1587 0.0355 1.8529

1.209 0.083 0.231 1.181 3.461 0.123 0.847 1.733 0.992 0.662 0.913 0.260 1.757 2.096 0.075 0.726 1.312 1.877 0.337 0.112 1.449 1.239 1.479 0.252 1.907 0.997 4.050 2.957 0.834 1.294 3.008 0.779 0.786 1.715 0.121 0.027 1.421 191

Appendix B TABLE 2. Continued Country Code Country PNG PRY PER PHL PRT SEN SGP ZAF KOR SWZ SYR THA TTO TUN TUR URY VEN YEM YUG ZMB ZWE

Parameter Estimate Absolute t-value

PAPUA NEW GUINEA PARAGUAY PERU PHILIPPINES PORTUGAL SENEGAL SINGAPORE SOUTH AFRICA SOUTH KOREA SWAZILAND SYRIA THAILAND TRINIDAD & TOBAGO TUNISIA TURKEY URUGUAY VENEZUELA YEMEN YUGOSLAVIA ZAMBIA ZIMBABWE

NOTE: LBR LIBERIA is the base category.

192

0.2019 0.8134 0.4178 0.4835 2.6376 ⳮ0.6757 4.0952 0.5513 2.8328 0.9468 2.2094 2.6458 0.5371 2.3242 1.3535 0.1683 ⳮ1.3750 4.4678 1.2595 ⳮ0.7751 1.1431

0.153 0.617 0.321 0.371 1.998 0.519 3.112 0.419 2.152 0.718 1.650 2.026 0.407 1.763 1.037 0.129 0.838 2.927 0.962 0.587 0.836