Growth and Political Institutions
31 March - 1 April, 1995
This conference was organized by CREI and Harvard University. The program was prepared by Professors Ramon Marimon (European University Institute, Florence and Universitat Pompeu Fabra) and Robert Barro (Harvard University).
FRIDAY 31 MARCH, 1995
Democracy and Growth
Robert J. Barro. (Harvard University, Bank of England, NBER)
Discussant: Robert Waldman (EUI, Florence)
Session Chair: Ramon Marimon (EUI, UPF, CREI, CEPR, NBER)
Politics and the Effectiveness of Foreign Aid
Peter Boone (LSE, Center for Economic Performance)
Discussant: William Easterly (The World Bank)
Growth, the Terms of Trade and Power Concentration
Philip R. Lane (Harvard University) and Aaron Tornell (Harvard University, NBER)
Discussant: Daniel Cohen (CEPREMAP, CEPR )
Session Chair: Joan M. Esteban (IAE-CSIC)
Per Krusell (University of Rochester) and José Victor Ríos Rull (University of Pennsylvania)
Discussant: Andreu Mas-Colell (Harvard University, UPF)
Session Chair: Josep M. Colomer (IESA-CSIC, UPF)
On The Number And Size Of Nations
Alberto Alesina (Harvard University, NBER, CEPR), Enrico Spolaore (ECARE – Université Libre de Bruxelles)
Discussant: Oded Galor (Brown University, Hebrew University)
Session Chair: Salvador Barberà (UAB)
SATURDAY 1 APRIL, 1995
Property and Contract Rights under Democracy and Dictatorship
Philip Keefer (The World Bank), Stephan Knack (IRIS – University of Maryland, American University), Christopher Clague and Mancur Olson (IRIS – University of Maryland)
Discussant: Edward Glaesser (Hoover Institution, Harvard University)
Session Chair: Teresa Garcia-Milà (UPF, CREI).
Abhijit V. Banerjee (MIT)
Discussant: Xavier Sala-i-Martín (Yale University, UPF, CEPR)
An Empirical Investigation on Income Distribution, Democratic Institutions and Growth
Roberto Perotti (Columbia University and NBER)
Discussant: Torsten Persson (IIES – U. of Stockholm, CEPR)
Session Chair: Albert Marcet (UPF)
Robert Barro and Ramon Marimon
Democracy and Growth
Robert Barro (Bank of England, NBER)
Growth and democracy (subjective indexes of political freedom) are analyzed for a panel of about 100 countries from 1960 to 1990. The favorable effects on growth include maintenance of the rule of law, free markets, small government consumption, and high human capital. Once these kinds of variables and the initial level of real per-capita GDP are held constant, the overall effect of democracy on growth is weakly negative. There is a suggestion of a nonlinear relationship in which democracy enhances growth at low levels of political freedom but depresses growth when a moderate level of freedom has already been attained. Improvements in the standard of living -measured by GDP, life expectancy, and education -substantially raise the probability that political freedom will grow. These results allow for predictions about which countries will become more or less democratic in the future.
Robert Waldmann’s (EUI) Discussion of “Democracy and Growth” by Robert Barro: Robert Waldmann emphasized that Robert Barro’s research improved on previous work on democratic institutions and economic growth along two lines. First, Barro estimates the effect of democratic institutions on growth by making use of instrumental variables to control for possible feedback effects and reverse causation and second, Barro uses the improved data set (developed by Barro with Jong-Wha Lee (1994)) on educational attainment and human capital. Waldmann argued, however, that Barro’s main finding that democratic institution do not significantly promote economic growth, is incorrect because Barro wrongly controls for endogenous variables like fertility, rule of law, and life expectancy. Waldmann argues that democratic institution tend to decrease fertility and increase the rule of law and life expectancy. Once the growth regression is estimated without those three endogenous variable, the effect of democratic institution on economic growth is positive and significant. To make his point, Waldmann presented regression results which suggested that past changes in the democracy index increased life expectancy, controlling for a variety of variables like GDP and the level of life expectancy. His estimates suggested that going from the most undemocratic to the most democratic system (in the sample) would increase life expectancy by 10 percent in the long run. He suggested that this may be because democratic institutions allow for broader participation in the making of public health policy. He noted that a similar effect appeared to be at work regarding education expenditures. Past changes in the democracy index increased education expenditures while they have no effect on government consumption.
Alberto Alesina suggested that while there may be no effect of democratic institutions on average growth rates, there may be an effect on the variance of average growth rates. In particular, a casual comparison of the growth performance of Asian dictatorships (“good dictators”) with dictatorships in Africa (“bad dictators”) indicates that, in terms of economic growth, dictators produce more variable outcomes. He also felt that it was odd that all regressions controlled for investment given that democratic institutions may affect the investment rates. Finally, he wondered how much of the variability in the sample was due to the differences between countries in Asia and countries in Africa and suggested that including dummies might help to address this question. Robert Barro replied that he did include dummies for Africa and Asia and that they were insignificant in explaining the results. Even the dummy variable for Latin America was insignificant, once the rate of inflation was controlled for. Mancur Olson argued that the lack of a theoretical framework in Barro’s analysis makes it difficult to understand which variables should be controlled for. For example, he argues that thinking in terms of “good dictators” and “bad dictators” is misguided. People behave according to the incentives they are given. The very definition of a dictator, for example, is that he is the law. Hence controlling for a rule of law index at the same time as controlling for democratic institutions fails to attribute economic growth to the proper sources. Stephen Knack added that the time variation in the democracy index is highly suspect. He pointed out that the changing trends in the index coincided with changes in the group of people judging the strength of democratic institutions across countries. Barro agreed but emphasized that the trend in the democracy index coincides with the trend in other indices like civil liberties and political rights. Other question addressed the econometric stability of the results across sub-samples and sub-periods. Barro stated that, generally speaking, the results were robust along this dimensions.
Politics and the Effectiveness of Foreign Aid
Peter Boone (London School of Economics, CEPR)
Critics of foreign aid programs have long argued that poverty reflects government failure. In this paper I analyze the effectiveness of foreign aid programs to gain insights into political regimes in aid recipient countries. My analytical framework shows how three stylized political /economic regimes labeled populist, rent seeking and laissez-faire would use foreign aid. I then test reduced from equation using data on nonmilitary aid flows to 96 countries. I find that models of rent seeking political regimes best predict the impact of foreign aid. Aid does not increase investment and growth, nor benefit the poor as measured by improvements in human development indicators, but it does increase the size of government. I also find that the impact of aid does not vary according to whether recipient governments are liberal democratic or highly repressive. But liberal political regimes and democracies, ceteris paribus, have on average 30 % lower infant mortality that the least free regimes. This may be due to greater empowerment of the poor under liberal regimes even though the political elite continues to receive the benefits of aid programs. An implication is that short term aid targeted to support new liberal regimes may be a more successful means of reducing poverty than current programs.
William Easterley’s (World Bank) Discussion of “Politics and the Effectiveness of Foreign Aid” by Peter Boone By way of introduction, William Easterley pointed out that Peter Boone’s paper was the first serious attempt to investigate the effect of aid on growth and standards of living. The main problem with the pre-existing literature being that the sample selection problem due to the endogeneity of who gets aid is not addressed. Easterley emphasized that, without addressing sample selection, analyzing the effect of aid on welfare is like analyzing the effect of hospitals on health by comparing people which are in hospitals with people which are not. This is why it is crucial to go to instrumental variable estimation as done by Boone. But the choice of instruments becomes critical. There Easterley was critical of the two main instruments used, “Friend of France” and “Friend of US.” Easterley argued that “Friend of US,” being defined as countries who get more than 1 percent of the US aid budget, is clearly not exogenous. Presumably, the state of the economy plays an important role when deciding on US aid. “Friend of France,” defined as member of the Franc zone, appears the better instrument. Being a member of the Franc zone does, however, also entail fixed exchange rates with France, trade agreements, and many other special deals which may affect growth etc. in their own right. Easterley was also critical regarding the evidence of the effect of aid on infant mortality. While it is true that the estimation suggests that aid does not reduce infant mortality, not much else appears to matter either. For example, income growth appears to have no effect on infant mortality despite other evidence suggesting the contrary. Also the fact that high infant mortality seems to imply higher future infant mortality is suspicious and appears to indicate a lack of stability may be as a result of misspecification. This might be due to measurement error which, already serious in level estimation, becomes potentially fatal once the left-hand variable is in differences. Finally Easterley pointed out that indirect taxes to GDP is not a good measure of distortions. Whenever data on revenue raised and tax rates is available, they suggest that there is no correlation because of tax-evasion. In any case, indirect taxes are not very distortionary and the black market premium would appear to be the better measure.
Alberto Alesina suggested that the times a country voted with the US might be a better instrument than “Friend of US.” Edward Glaeser suggested that the objective of aid was to improve the receiving countries attitude towards donors and that therefore this, not growth etc., should be the LHS variable. Mancur Olson suggested that one effect of aid may be to impose better institutions and that these effect might not be discernible in short sample. He cited the Marshall plan as an example.
Growth, Terms of Trade and Power Concentration
Philip Lane (Harvard University)
Aaron Tornell (Harvard University)
We develop an economic growth model that focuses on the role of property rights -defined by limits on discretionary redistribution- in determining growth performance. We show that a country’s response to a terms of trade disturbance is conditioned on the distribution of power. In a country in which power is concentrated, a positive shock will lead to an increase in redistribution, that is financed by distortionary taxation, and slower growth. The opposite is true in a country in which power is dissipated. The distortion is that the increase in taxation induces a reallocation of capital from the formal sector, that is taxed, to the less efficient shadow economy, that is more difficult to tax. Empirically we construct proxies for power concentration and show that, in growth regressions, conditioning the relationship between the terms of trade and growth on our proxies for economic concentration uncovers the negative effect mentioned above.
Daniel Cohen’s (CEPREMAP, CEPR) Discussion of “Growth, Terms of Trade and Power Concentration” Daniel Cohen began his discussion by pointing out that, theoretically, the effect of power concentration on economic efficiency is not monotonic as emphasized by the authors. If, for example, there was one only group making decisions on taxation, that group would effectively internalize the effect on capital flight and make an efficient decision. In fact there was a parallel regarding the efficiency of labor markets. Both the very competitive labor market in the US as well as the labor market in Australia, with centralized bargaining among few decision makers, result in little unemployment. The European organization of labor markets appears to be located between these polar incentives with the result of high and persistent unemployment. Cohen’s discussion turned next to the question of equilibrium selection. The authors proved that an equilibrium of the model had certain comparative static properties. In particular, a windfall may permanently harm prospects for growth. But why should that equilibrium materialize. In effect, it appeared as if the situation was unstable as power groups may have two types of incentives to deviate: First, individually they may lower appropriation in order to give incentives to repatriate capital (or to, more generally, invest it into the formal sector) and hence increase their profits by taking a smaller share of a larger pie. Second, there appeared to be no reason why power group could not solve the two different problems at hand separately. They could first set the revenue maximizing tax and second decide over how to split revenues among themselves. Cohen presented a simple static model of the division of a possibly shrinking pie next. He demonstrated that if an increase in the pie was correlated with the elasticity with switch the pie shrinks due to influence taking activity, then the authors results pertained. Still, Cohen argued that the theoretical model predicted three effects: (1) growth should decrease with a terms of trade shock, (2) growth should increase with the number of power groups, (3) the effect of a terms of trade shock on growth should be larger (and negative) the fewer the power groups. Cohen the argued that the authors empirical result demonstrated at most (1) but that they did not demonstrate (2) and (3).
Edward Glaeser emphasized a parallel between the interior equilibrium of the model and the mixed strategy equilibria in some games. Both appear to have unintuitive comparative static results. The rest of the discussion concentrated on the authors measurement of power concentration. There was agreement on the fact that diversification of exports was not a good measure. It was suggested to get a measure of union activity as a better proxy.
Per Krusell (University of Rochester)
José Victor Ríos-Rull (University of Pennsylvania)
We propose a political theory of transition from a non-market to a market economy. Economic success in the model is synonymous with the adoption of new production techniques and this process requires explicit investment in human capital. Economic policy consists of regulation of the process of technology adoption, and a key aspect of the economy is that there is inherent disagreement in the population over how this policy should be conducted. Our theory predicts: (i) that an economic transition is associated with a substantial drop in output; (ii) that it is in the interest of large groups in the population to resist the reforms; (iii) that democratization is not sufficient for an economically successful transition -indeed it will lead to “backwards” politics and slow growth; and (iv) that an interim period of dictatorship which imposes laissez-faire is necessary for long-run prosperity to be sustainable in the context of democratic policy determination.
Andreu Mas-Colell’s (Harvard University, UPF, and CREI) Discussion of “Politico-Economic Transitions” By way of introduction Andreu Mas-Colell emphasized that he was not going to spend much time commenting on the technical aspects of the paper, other than stating at the outset that it was a technical tour-de-force. He felt, however, that many of the points of the paper could be made in simpler game theoretic models. He went on to do so in a simple two period, three player model. He started out with a model where innovation took place in both periods because a majority of players (two out of three) were better of when innovation occurred in both periods. Then he proceeded to changing the payoffs of the model. If the second innovation did not benefit one of the player much, then the outcome could be that players voted to innovate in the first period but that this particular player then sided with those against innovation. Hence the outcome became innovate/not-innovate. Finally, innovation could be increasing the pie and still leave two out of three players worse off. Hence innovation could be prevented in both periods. The egalitarian solution, distributing the payoff from innovation equally, could, but need not, be a solution to the problem. Abolish majority-voting in the first period could, in this case, yield innovation and growth in the second period. Mas-Colell felt, however, that this conclusion was misguided. The benefits of democracy and majority voting must be decided on more grounds than this. In particular, simply allowing for side payment could improve the situation and the winners would compensate the losers so that they vote with them.
Ramon Marimon emphasized that not having election is just one of many possible commitment devices, while Daniel Cohen pointed out that the problem with the young bribing the old to vote for innovation appeared to be serious when the economy started from a relative backward state. Still, he was not convinced that irreversibilities in human capital were important enough to warrant the conclusions in the paper. Other comments well states that if one allows for voting one should also allow for competing platforms. Per Krusell replied that the main point of the model was to show how changes in one period, not vote, could propel the economy into a high growth equilibrium. Mancur Olson stated that the basic points of the paper, that there may be more gains than losses but more losers than winners and that even constitutions had to be self-enforcing, were important and a first step in the right direction.
On the number and Size of Nations
Alberto Alesina (Harvard University, NBER, CEPR)
Enrico Spolaore (ECARE)
This paper studies the equilibrium determination of the number of political jurisdictions in different political regimes, democratic or not, and different economic environments, with more or less integration and free trade. We focus on the trade off between the benefits of large jurisdiction in terms of economies of scale and the costs of heterogeneity of large and diverse populations. Our model implies that: i) democratization leads to secessions; ii) without an appropriate redistributive scheme (which we characterize) in equilibrium one observes an inefficiently large number of countries; iii) the equilibrium number of countries is increasing in the amount of economic integration. We also study the welfare effects of economic integration and free trade when the number of countries is endogenous.
Oded Galor’s (Brown University) Discussion of “On the Number and Size of Nations” by Alberto Alesina and Enrico Spolaore: Oded Galor emphasized that the basic trade-off present in the paper appears important in actuality. On the one hand, a larger population implies a decreasing average cost in the supply of public goods with fixed costs. On the other, an increasing population implies a larger variety of prefernces and hence, from the point of the individual, a possibly less desirable bundle of public goods. Galor emphasized the three main results of the paper: (1) an increase in the degree of economic integration increases the number of nations. (2) an increase in democracy implies an increase in the number of nations. (3) the equilibrium number of nations is smaller than socially efficient. Galor than proposed the following alternative assumptions. (A1) mobility increases as income increases. This may be due to better means of communications or because of technological progress. (A2) economic integration implies a narrower variety of preferences as people’s preferences become more equal. This is what appears to have happened, for example, with the European Union. (A3) there are public good spillovers across countries. Examples are environmental standards, health expenditures, and public TV and Radio. Galor showed that under this assumptions the main results are reversed. Economic integration lead to an increase in the size of nations and a decrease in the number and that the number of countries may be inefficienctly low. He also critisized the fact that people along the border are next to indifferent in which country to live. Counterexample abound, but most to the point were the Israeli residents of the Golan Heights and the Syrians just east.
Ramon Marimon emphasize that diversity of preferences may be a reason why government function should be split up regionally. Mancur Olson pointed out that similar arguments could be used to analyze the optimal number and size of jurisdictions, especially when different problems must be dealt with a different level. Environmental pollution, for example, must be dealt with at a global level. Daniel Cohen argued that in any case non-linear tax schedule could decentralize the efficient allocation as people further away from the political center could be made to pay less.
Institutions and Economic Performance: A New Look at Democracies and Growht
Christopher Clague (IRIS, University of Maryland)
Philip Keefer (IRIS, University of Maryland)
Stephen Knack (IRIS, University of Maryland)
Mancur Olson (IRIS, University of Maryland)
This paper examines some intermediate links between the type of political regime in a country and its economic growth. One of the ways in which regime type affects growth is through its effect on the character and security of property rights and contract enforcement. We introduce a new measure of these rights, called “contract-intensive money”, which is the ratio of noncurrency money to the total money supply. Using this measure, along with ratings by expert observers of such country characteristics as “rule of law” and risk of expropriation”, we find evidence that democracies tend to offer more credible guarantees of contract and property rights than autocracies, and that more secure rights in turn promote economic growth. We also examine the relationship between the duration of the political regime and the security of property and contract rights, finding that the longer the duration of the democracy and the longer the tenure of an autocrat are associated with more secure property rights.
Edward Glaeser’s (Hoover Institution and Harvard University) Discussion of “Institutions and Economic Performance: A New Look at Democracies and Growth” Edward Glaesser emphasized the importance of the basic question, the demand for different types of regimes, underlying the research. Why does it appear that the population demand “dictators” at certain historical moments. One answer may be because of their belief that dictators may be able to create effective protection for property rights. But they must, of course, trade-off the protection of their property rights relative to their neighbors with the protection of their property rights relative to the “dictator.” Regarding the empirical work Glaeser emphasized the importance of quantifying the protection of property rights with objective economic data and not subjective opinions of some observers. In this sense, the authors use of “contract intensive money” was an important first step. He also argued that dealing with the panel the authors have by assuming that consecutive years are draws from independent distribution, while it does not introduce a bias in the estimates, does bias the standard errors downwards. Hence the precision of the estimates will be overestimated. A better way should be found to deal with the serial correlation in consecutive observations. Instrumental variable would also be preferable to the OLS techniques used. Possible instruments are economic conditions in neighboring countries and ethno-linguistic fractionalization. Finally, he argued that more effort should be dedicated to reconcile the time-series and cross-section results. One possibility was that only in desperate situations do democracies become dictatorships, which biases the estimates of dictatorship on economic conditions downwards.
Ramon Marimon emphasized the difficulty of classifying regimes because political freedom and stability for doing business may be separate issues. China and Italy are two cases in point. One a seemingly very stable dictatorship with no political freedom but stability for doing business and the other with much political freedom but no stability for business. Mancur Olson also emphasized that secure dictators may be good for the protection of property rights while enforced democracy may be very ineffective. Daniel Cohen reiterated the pre-occupation with the fact that the time of regime change was endogenous. Robert Barro stated that contract intensive money appeared to be more related to financial development than effective protection of property rights. Mancur Olson replied that it is the effective property rights which makes financial development through contract intensive money possible. William Easterley emphasize credit-worthiness, another right-hand side variable, said more about macr-policy and macro-managment than well defined property rights.
Abhijit Banerjee (M.I.T.)
This paper starts with the theoretical observation that under some ostensibly rather weak conditions, both corruption and inefficiency due to corruption in things like tax collection, pollution control and the distribution of publicly provided private goods can be eliminated by partially legalizing profit making by government bureaucrats. We contrast this with the practical observation that while corruption is rampant in many developing countries and is widely perceived to be highly undesirable, there is no discernible tendency towards such legalization. We go on to discuss possible economic and political costs that might work against a move toward legalization.
Xavier Sala-i-Martin’s (Yale University and UPF) of “Eliminating Corruption” Xavier Sala-i-Martin started by pointing out that the analysis of cooruption was particular relevant in Spain in this moment because of the case of Roldan who is suspected to have received 500 million dollars to give rights to build housing for police. In fact, Banerjee’s first model suggested that the government should have sold the right to become chief officer of the police force and then given the chief officer the right to sell permits to build housing for the police. This would have eliminated corruption and implemented the efficient allocation. Sala argued that before discussing how to eliminate corruption the question is whether this is warranted. Often corruption is just one way to get around ineffective government legislation. Japan and Italy being two cases in point. In any case, Sala regarded as the main result of the paper the fact that it appeared very easy to eliminate corruption. Clearly, that depended on the particular assumption made in the paper and the important next step for research was to identify which of the assumption was not satisfied in actuality. In many ways, Sala argued, this result is reminiscent of Ricardian Equivalence, which also helped in identifying which assumption had to be relaxed for government bonds to matter. Finally, he argued that important clues for fighting corruption might be gotten from the behavior of firms. How does IBM, for example, fight corruption? Understanding firm behavior might shed light on possible government policy against corruption.
Christopher Clague emphasized that governments can always fight corruption by large fines. But this is prevented in actuality by notions of fairness. The best way to fight corruption is to create loyalty, which is used in large firm like IBM. Michele Boldrin asked why some courts should be honest? Presumably this is endogenous and related to incentives. Furthermore, why aren’t individuals capable to identify honest courts which leads to overworked courts? This cannot be, as the model suggest, because in equilibrium nobody goes to court. people must believe that courts are honest and verify this. In any case, the problem of tax evasion appears to be under reporting of rich people, not the fact that some people, the poor in the model, pay too much taxes. Regarding Banerjee’s point on tax collecting of the Mafia in Sicily, Boldrin pointed out that Sicily has the largest fraction of tax evasion to output.
An EmpriricalInvestigation on Income distribution, Democratic Institutions and Growth
Roberto Perotti (Columbia University)
Many theoretical models posit a positive relationship between income distribution and growth. However, the theoretical mechanisms that can lead to this relationship vary enormously. By its nature, a reduced form estimate cannot shed light on the underlying mechanisms. Hence the importance of the third issue -evaluating the specific channel (s) of operation of income distribution by estimating the structural models behind the reduced form. The paper explores the three most popular channels: endogenous fiscal policy, socio-political instability, and borrowing constraints. The main conclusion of the paper in this regard is that there is no evidence in support of the first explanation, god evidence in support of the second, while the results on the third are mixed.
Torsten Persson argued that the paper could be seen as testing four hypothesis. (P1) the robustness of the relationship from inequality to growth. (P2) the question whether this relationship differed between democracies and non-democracies. (P3) the fiscal policy approach which argues for a negative link between the income share of the middle class and taxes and for a negative link between taxes and growth. (P4) the political instablity approach which argues for a negative link between socio-political stability and the income share of the middle class and a negative link between socio-political stability and growth. Persson begun with commneting on (P2), In contrast to Persson and Tabellini (1994), Perotti does not fins any significant differences in the relationship between inequality and growth between democracies and non-democracies. This may be due to the different definitions used. Persson and Tabellinin used a democracy dummy, while Perotti uses a continuos index. Going to the test of the structural fiscal policy mode, in (P3, Persson proposed the following simple alternative: Suppose that growth=(1-p)(1-t)A-constant, where a smaller p stands for a better protection of property rights and t stands for taxes, then we get that growth=(1-p)A-tA+ptA-constant, in other words taxes and protection of property rights must be interacted. Taking this empirical approach, proxying for protection of property rights by the rule of law index, Persson finds that the results in Perotti are overturned. Regarding (P4) Persson wondered whether this actually corresponded to a structural form estimation, referring to apaper by Svensson (1994) which did take a structural approach. Svensson argues that instability lowers the incentive to reform institutions and therefore lowers investment, a hypothesis strongly supported in the data. In any case, the main problem, only unsatisfactorily dealt with, is reverse causation of growth on the distribution on income. Here Persson proposed going to instrumental variable estimation.
Daniel Cohen argued that regarding the unintuitive results on social security, it may be important to go away from income per capita to average labor productvity. Robert Barro stated that he did not find social security to be the correct measure of re-distribution. It measures transfers to old people not poor people.