WV 2ag I o POLICY RESEARCH WORKING PAPER 2910 Boondoggles and Expropriation Rent-seeking and Policy Distortion when Property Rights are Insecure Philip Keefer Stephen Knack The World Bank F Development Research Group L 3 Investment Climate Team and Public Services Team October 2002 I POLICY RESEARCH WORKING PAPER 2910 Abstract Most analyses of property rights and economic They may reduce the incentives of governments to development point to the negative influence of insecure use tax revenues for productive purposes, such as public property rights on private investment. Keefer and Knack investment. focus instead on the largely unexamined effects of - They do so whether one regards the principal insecure property rights on government policy choices. problem of insecure property rights as the maintenance They identify one significant anomaly-dramatically of law and order, which government spending can higher public investment in countries with insecure potentially remedy, or as the threat of expropriation by property rights-and use it to make the following broad government itself, and therefore not remediable by claims about insecure property rights: government spending. They increase rent-seeking. The authors present substantial empirical evidence to support these claims. This paper-a product of the Investment Climate and Public Services Teams, Development Research Group-is part of a larger effort in the group to investigate the institutional roots of effective government policy. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Paulina Sintim-Aboagye, room MC3-422, telephone 202-473-8526, fax 202-522-1155, email address psintimaboagye@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at pkeefer@worldbank.org or sknack@worldbank.org. October 2002. (30 pages) The Policy Research Working Paper Senes disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Research Advisory Staff Boondoggles and expropriation: Rent-seeking and policy distortion when property rights are insecure Philip Keefer Stephen Knack The World Bank The World Bank pkeefer&worldbank.org sknack@worldbank.org Boondoggles and expropriation: Rent-seeking and policy distortion when property rights are insecure' The security of property rights has a well-known and significant role in economic development. In explaining the links between the two, most analyses point to the direct effect that insecure property rights have on private investment. This focus does not explain significant anomalies in government policy making between countries with secure and insecure property rights. This paper uses one such anomaly, the fact that the ratio of public investment either to national income or to private investment is dramatically higher in countries with insecure property rights than other countries, to make the following broad claims: insecure property rights increase rent-seeking by government; they may reduce the incentives of governments to use tax revenues for productive purposes; and, finally, these effects emerge regardless of whether one regards the principal problem of insecure property rights as the maintenance of law and order among citizens, which government spending can potentially remedy, or as the threat of expropriation by government itself, and therefore not remediable by government spending. Empirical tests reported in the second half of this paper provide evidence for each of these claims. On the other hand, our analytical results reject the argument that governments have an incentive to increase public investment when property rights insecurity drives down private investment. On the contrary, as long as public investment is complementary to private investment, the opposite is true. However, when one allows rent-seeking and public investrnent decisions to interact, the analytical results generate predictions that are consistent with the empirical record reported in the second half of the paper: observed public spending rises with the insecurity of property rights. With respect specifically to public investment, the analysis provides an explanation for the ambiguous and often negligible growth effects that have been found for public investment in the literature. The analysis here suggests that the effects of observed public investment on growth are likely to vary with the security of property rights. However, the interaction between the security of property rights and public investment has not been considered in this large literature. The tests reported at the end of this paper support the notion that public investment can have a positive effect on growth, but that this effect is I We are grateful to Karla Hoff, Martin NMeurers and Paul Zak for thorough and very helpful comments. revealed only when one accounts for the differing incentives of governments, under different property rights environments, to use public investment for rent-seeking purposes. Public investment and growth The analysis in this paper begins with Barro's (1990) canonical analysis of public investment and growth. In the sections following, we expand on the conception of property rights that Barro hints at in his model and contrast it to a more institutional approach to property rights that is common in the literature. In the final stage of the analysis we introduce the potential for government rent-seeking, from which we derive our testable predictions. Barro (1990) begins with a representative, infinite-lived household in a closed economy, maximizing utility given by U = f u(c)e-Odt, where c is consumption per person, p>O is the constant discount rate and u(c) = a , where a > 0. Production per capita in the model is given byy=kO = Ak(k k) . A flat income tax finances public good spending, so the marginal private return to private investment is (1- r) = (I r(Xl - a){ k ) The growth rate of consumption is therefore (1) yV =-[(I - rl -a)A _k ) -P]. Both benevolent and rent-seeking governments maximize (1) with respect tog/y to find the rate of public investment that maximizes growth. Rent-seeking governments then go on to maximize their utility with respect to rents, the amount of taxes they collect above and beyond the amount needed to fund growth-maximizing public investment. Assuming that all tax revenues are spent on public investment (r = gly) and that the production function is Cobb-Douglass, and noting that (g / k) = (g / y) - q(g / k), differentiation gives Barro's equation (15): (2) dr(g (O'-l). d(Got k Growth is miaximlized when the expression in (2) is equal to zero, yielding?z a. Barro 3 argues that since all governments would setg/k such that growth is maximized, there should be no observed correlation between public investment and growth. Rates of public investment should only vary across countries because of variations in the relative productivity of private and public capital which are not themselves correlated with growth. The effect of property rights on public investment and growth There are at least two ways to introduce the security of property rights into the model above. One approach is suggested in Barro's (1990) public investment work and is common in the growth literature (e.g., Barro, Mankiw and Sala-i-Martin 1995, or Belletini and Ceroni). In this literature, property rights insecurity is a tax, implying that property rights security can be modified by government with the same facility that it modifies taxes more generally. Barro (1990) specifically assumes that governments can spend resources to curb attempts by some households to steal the output of other households: "An increase in spending, g, in areas that enhance property rights causes a reduction in the effective value of rrather than a direct effect on the production function... .[t]he relation of growth and saving rates to the amount of government expenditure devoted to the enforcement of property rights would resemble [those of other productive government expenditures]." (Barro 1990, p. S116).2 Although public spending consequences are implied by this "fiscal" model of property rights, they have not been formally modeled nor tested. An alternative approach to property rights is taken in the more institutionally- oriented literature. In this literature, the problem is not one of "law and order", where households fear expropriation by other private parties, but rather one of the ability of the state to credibly commit not to expropriate sunk investments. North and Weingast (1989), Acemoglu and Robinson (2001), Clague et al. (1996), Keefer and Stasavage (2000) and many others emphasize the institutional roots of insecure property rights, particularly the ability of governments to commit credibly not to expropriate privately-held assets. Institutions are not easily susceptible to change by government officials. As with the fiscal model of property rights, the institutional model has not been formally integrated into a dynamic 2 Zak (2000) also models the property rights problem as one of expropriation by some households of others, and allows government to expend resources on policing to reduce this risk. Unlike Barro (1990), he endogenizes the expropriation risk that households impose on each other, however, his model abstracts from the decision by government to invest in public infrastructure. 4 model of government fiscal decision making in a growth context.3 The two approaches embrace two distinct problems. The fiscal approach to property rights focuses on the law and order threat that citizens pose to each other, and the costs to government of dealing with that threat. It does not explicitly ask what conditions make these costs higher in some countries than others. The institutional model focuses on the threat of government expropriation and directly suggests institutional sources of variations in the severity of this threat across countries. Most of the cross-country evidence that sheds light on the impact of property rights is most easily interpreted in the context of the second model including the evidence presented below. One might imagine that governments would have an incentive to offset reduced private investment due to insecure property rights with greater public investment. Under either approach to property rights, the theoretical analysis below shows that governments do not have an incentive to do this, as long as public investment is a complement to private investrnent, as assumed in all of the public investment literature, rather than a substitute for it. An additional consideration, besides the insecurity of property rights in and of itself, is necessary to explain our empirical findings that public investment is highest in countries with the weakest property rights. Property rights as a fiscal problem If property rights security were easily influenced by government budget decisions, then the decision to spend resources to improve the security of property rights would be directly linked to the decision to spend resources on public investment. To explicitly model this, we assume that households confront a new tax rate given by THH = rp, + rPR + Kpr ), comprised of the taxes they pay to finance public investment (rpz and more secure property rights (rp,), and of the irnplicit tax they confront, represented by iK(-Q), which is the residual risk to property rights after government has spent resources (e.g., on courts and police) to mitigate that risk, dK/d-rpR < 0. The burden of insecure property rights can then be described as (K + rpR ), the amount.that households pay to secure property rights, and the insecurity 3 A third approach views the security of property rights as neither a fiscal nor an institutional problem, but rather a "pre-institutional" problem of struggle over resources and of investment by all actors in resources that can be used either to defend their own assets or to expropriate the assets of others (see, e.g., Skaperdas 1995 and Sonin 2002). The government provision of property rights in these models is, however, as in the macroeconomics literature, a simple policy decision flowing from the identity of the median voter rather than a consequence of institutional features of the state. 5 that remains after taxes have been converted into police and courts. As usual, households choose a consumption path that mnaximizes utility, taking as given government policy, so that growth is given by (3) y = - T[( HH X( - a)A(k) -] Government then chooses r,f and rpR to maximize growth. Noting that rp, = g/y =A-1 (g/k)l-1, substituting forg/k , and differentiating (3) with respect first to TP, and then to TPR to get the marginal growth impacts of each, yields the two first order 1 a I conditions given by (4) (after dividing out -_ra A l-a from each of the two conditions). Pi (4.1) K' =-1 (4.2) rp, = a(l -,rpR - K(rpR Note that when property rights are irrelevant, equation (4.1) disappears and equation (4.2) converges to the previous solution for public investment, rp, = a. From these conditions, one can deduce immediately that more insecure property rights reduces the incentives of governments to make productive public investments. The ratios of productive public investment to national income and to private investment are lower in countries with insecure property rights. From equation (4.2), the greater is the burden of insecure property rights (K + rp, ) (spending to secure property rights and the residual insecurity of property rights), the lower is rp,, which is simply the ratio of public investment to national income. With respect to the ratio of public to private investment, recall that = (A rp, )I. Substituting from equation (4.2), and differentiating both sides k with respect to (K + rp ), the net burden imposed by insecure property rights on households, it follows immediately that the predicted ratio of public to private capital also drops as property rights insecurity increases.4 The law and order approach to property rights does not predict, however, that the security of property rights influences the growth effect of productive public investment. At the growth-maxiniizing level of public investrnent, additional public investmrent has zero 4 Note that restricting attention to (4.2) is possible because the growth-maxkriizing level of expenditures on 6 marginal impact on growth, regardless of the secutity of property rights.5 An institutional approach to property rights and public investment The institutional approach to property rights focuses on the threat of government expropriation, a threat which is unlikely to be solved by increasing expenditures on courts and police. Two of the significant explanations for variations in this threat across countries are checks on government discretion and the time horizon of government officials. North and Weingast and Keefer and Stasavage (2000) argue that the ability of any government to credibly promise not to renege on its policy commmitments (e.g., not to expropriate) is enhanced by the presence of institutional checks and balances or multiple veto players inside government. These institutional features are at the constitutional core of the state. Similarly, countries that have elections, which offer a sanction for decision makers who violate promises, can also strengthen the institutional underpinnings of secure property rights. Others (e.g., Clague et al. 1996) argue that the time horizon of leaders is another significant determinant of their decision to expropriate. Clague, et al. (1996) that the longer a dictator expects to survive in office, or the longer a democracy has been established, the more secure are property rights. Leaders with long time horizons are reluctant to engage in expropriatory actions, since the reduced investment that results hurts their ability to extract rents in the future. Leaders with short time horizons naturally care less about future rents. Regime horizons, though influenced by the actions of government officials while in office, have a substantial exogenous component that is beyond the influence of government policy instruments. To embed these characteristics in the Barro (1990) model we first allow for the possibility that at some point, government decision makers realize that they have only one period left in office, that at the end of the period they must leave office, removing all incentive they might have to maximize future growth and leading them instead to expropriate as much private capital as possible. The amount that they can take in the last period, though, depends on the overarching institutional environment in which they operate (e.g., where checks and balances are significant, it is less likely that any individual government actors can, in their last periods, extract significant rents). property rights protection establshed in (4.1) depends only on K. S Differentiating (3) with respect to rpi, and differentiating again with respect to (K + init), and evaluating the result at the growth maximnzing level of public investment given by (4.2) reveals no effect of propert rights 7 Assume that neither households nor government know in which period all government veto players will have to leave office, and that the exogenous probability that the government will have to leave office in any given period is given by 5) F(t = 1 - e*. 6 This distribution function has the convenient property that the probability that the government will have to leave office at some time t, conditional on not having had to leave office earlier, is given by F'(4/(1- F()) = h, where h is invariant over time. Upon leaving office, governments can take a share 8 of the capital that households have invested, above and beyond whatever tax rate the government had previously established. The key here is simply that governments can neither control nor predict the events that would make expropriation an optimal strategy. However, they can control taxes and spending conditional on the probability of the events occurring and the maximum rents that they can extract should they be forced to leave office. The specific assumptions - that the government leaves office, for example - are not meant to capture the entire set of circumstances under which governments might expropriate. The key assumption is simply that at some point in a government's tenure, its incentives change unexpectedly, leading it to shift tax policy in a way that is adverse to existing investors. The particular specification accomplishes this, although others would, as well. Institutions affect both F and 0. For example, in countries with strong, long-lived political parties, the likelihood that a party will have to leave office, never to return, is lower, reducing F and therefore b. In countries with regular elections and institutional checks and balances, it is more difficult for political actors to expropriate in the event that they are forced to leave office (or, more generally, in the event that their political incentives change). Such institutions effectively lower 0. Households maximize U = f u(c)e-Pdt, where 6.1) U = f u[(1 - F(t))(t) + F'(tXl - O)k(t)k-dt, subject to security on the relationship between public investment and econornic growth. 6 Kamien and Schwartz (1978) introduced this expression as the probability that a research program would bear fruit by some period t. The analogy to research and development is not strained, since R&D, like government tenure, is affected by factors both exogenous to and within the control of key actors. 8 6.2) k=(1-r)AkI-aga-c, c20. Equation (6.1) assumes that households consume c(t until government expropriates, at which point they live off of their remaining capital, (1 - O)k.7 The maximization problem can be approached by using (6.2) to rewrite (6.1) in terms of k and dk/dt. If the integrand of (6.1) is G, the solution to the problem is then given by the Euler equation, dGdt = Gk, or - u [1- F(t)]+ u'F'(t) + u'p[1- F(t)] = uf(1- F(t)Xl - rX1- a)A(g) + F(t)(l -9)] Recalling that u'=r' and u'=-arc- and rearranging terms yields the following expression for growth C= a [( rXI - a)A( k )-hp As expected, growth falls as the security of property rights deteriorates. Once again, all governments, whether rent-seeking or not, attempt to choose the level of public investrnent that maximizes growth. As before, substituting g =(A r)l-' into (7) and maximizing (7) with respect to r yields: (8) dY = I A l-a r I-a (a~ 1)=0- dr or The predictions of this model, incorporating an institutional approach to property rights, are somewhat distinct from those of the fiscal model of property rights. None of the institutional parameters appear in (7), so from (7) we can immediately deduce that the share of public investment in national income and the ratio of public to private investment are predicted to be invariant to the security of property rights, rather than declining with the security of property rights, as in the fiscal model. The influence of public investment on growth is invariant to the security of property rights, as in the previous model. In sum, the institutional approach to property rights, unlike the fiscal approach, 7 A more accurate, but more complicated setup would allow households to carry (1 - O)k into the next period, beginning production again under a successor government. Since utility under the successor government is a function only of the capital that households possess when the successor government comes to power, however, the simplification has no effect on the conclusions. 9 predicts that there is no spillover from the property rights environment to government fiscal decision making. However, in common with the fiscal approach to property rights, the institutional approach fails to predict the relationship between the magnitude of public investment and property rights that is observed in the data. Rent-seeking offers a logical avenue to explore to reconcile the property rights predictions and the empirical findings. We turn to this in the next section. Rent-seeking, institutions and public investment There is a significant distinction between rent-seeking and insecure property rights, and analytical advantages to not conflating them by, for example, defining those governments as rent-seeking that undermine property rights. Rent-seeking is best seen as that portion of predictable government taxes that are retained by government officials for their own uses. Threats to property rights, on the other hand, as the foregoing arguments suggest, are best seen as the unpredictable threat that either government or citizens will expropriate capital - possibly for their own use (as in rent-seeking), but not necessarily. Much public investment takes the form of rent-seeking. From pork barrel spending in the United States to white elephant steel factories in Nigeria, public investment has been seen to serve purposes other than maximization of growth, ranging from securing political support to securing personal fortunes. Among countless stories that show the extent to which rent-seeking is embedded in public investment in countries with insecure property rights is one of public investment in Turkmenistan: in a country where roads were crumbling and water was unavailable for hours on end, the authorities an international airport with the capacity to receive 4.5 million visitors a year, in a country that received a few hundred thousand a year, and for which the authorities insisted on building the control tower on the wrong side of the terminal, blocking the controllers' view of the runway.8 The public choice literature, with its emphases on rent-seeking interest groups (Buchanan and Tullock, 1962; Olson, 1971), bureaucratic and legislative incentives (Niskanen, 1971; Noll and Fiorina, 1978; Fenno, 1973), and the predilections of confiscatory autocrats (Olson 1994), provide ample reasons to believe that public investment should not be immune to the non-economic incentives of government decision makers. Pritchett (2000) cites numerous examples of public investments with no demonstrable positive impact 8 "Palaces and Poverty in Central Asia," The Washington Post, November 11, 1994, p. A35. 10 on economic outputs at alL Tanzi and Davoodi (1997) thoroughly demonstrate an inverse correlation between corruption and the quality of public investments across a large number of countries. However, the simple observation that rent-seeking is associated with public investment is not enough to explain why public investment is higher in countries with insecure property rights. For that, one needs to show that rent-seeking does not displace public investment or that, if it does, rent-seeking increases at a faster rate than public investment declines. Similarly, in the fiscal model of property rights, productive public investment declines with property rights protection; if property rights insecurity increases observed public investment because of higher rent-seeking, it must be the case that rent- seeking rises faster than productive public investment falls. In both the fiscal and institutional models of property rights, the sum of rent-seeking and productive public investment increases as the security of property rights drops. Barro (1990) describes a plausible way in which a rent-seeking government might decide how much to extract from households for its own, non-public purposes. First, government decision makers set the rate of public investrnent that naximizes growth, as in the preceding models. Then they fix the final tax rate to maxirnize their own consumption possibilities, conditional on the effect of this decision on future growth, holding public investment constant. In Barro (1990), the utility of government officials in any period is given by U(cg) = ul(r - a). y(O)e) j, where a is the growth-maximizing ratio of public investment to income. The higher is growth, y, the more that government can consume; higher taxes, though, suppress growth. Maximizing government utility with respect to r, holding public investment constant, leads rent-seeking governments to set taxes such that they can extract rents given by (9)r a = - I a)q The middle term in (9) simply reflects the fact that rents are higher when taxes have a smaller effect on growth, since an increase in taxes with no offsetting increase in public investment must have, by (1), a negative effect on growth. When property rights are a law and order problem that can be addressed by fiscal policy, governments can be thought of as first setting taxes and spending to protect property rights according to (4.1), then setting taxes to set the growth-mnaximizing level of public investment, according to (4.2), and then identifying the optimal level of rent-extraction by maxmizing U(cg) = U[,(THH - a(l - rpR - K(rpR)) - rpR) y(O)ew ] with respect to rHH. This yields rents given by (10) rHH -a(l-rpR K(rpR))P rR dV idr where the first term in (10) corresponds to the tax rate in (9), and the remamiing terms on the -left-hand side of (10) correspond to a in (9), such that the left-hand side of (10) is simply the amount tax revenues that are diverted to rent-seeking rather than productivity-enhancing activites. From (10), and specifically the denominator of the right-hand side term, we can deduce both that property rights insecurity exacerbates rent-seeking, and that insecure property rights drives up rent-seeking more rapidly than it drives down productive public investment. From the maximization of (3) and first order conditions (4.1) and (4.2), the right-hand side of (10) can be rewritten as (11) /T (I - a)A -rM-K(rpR ))a By inspection, it is evident that rent-seeking increases as the security of property rights declines (the left-hand side of (11) increases with K and -rpR). Moreover, the increase in rent-seeking offsets the decline in productive public investment predicted by equation (4.2). By (4.2), as property rights become more insecure (as (K + rpR ) rises), productive public investment drops by a Differentiating the right-hand side of (11) by (K + rpR) yields az a -a+1 a the amount by which rent-seeking increases with a decline in (I-a) A'-a(l-r-K) l-a the security of property rights (or, equivalently, an increase in the costliness of protecting property rights). For all reasonable values of the technology parameter A and the marginal utility of consumption, a-, this expression is larger than a and the insecurity of property rights 12 increases rent-seeking (much) faster than it suppresses productive public investment.9 Intuitively, governments forego current rents if the resulting increased household investment and faster growth promise more than offsetting increased future rents. When property rights are insecure (or costly to secure), the promise of future rents falls for two reasons. Directly, insecure property rights deter private investmnent; however, they also reduce private investment indirectly, by reducing government incentives to undertake productive public investment. The direct effect is what leads rents to increase faster than productive public investment falls. The addition of rent-seeking to the objectives of government turns the previous predictions of the fiscal model of property rights on their head: as long as rent-seeking takes the form of unproductive public investmnent, observed public investment (the sum of productive and unproductive public investment) can rise both as a fraction of income and as a fraction of private investment when property rights become less secure. Furthermore, the non-productive public investnent that results from rent-seeking has a negative imnpact on growth (or, more precisely, the taxes that finance it have a negative impact on growth). Consequently, one would expect the effect of observed public investment on economic growth to increase with the security of property rights, rather than exhibit no relation, as predicted in the absence of rent-seeking. The addition of rent-seeking to the institutional approach to property rights yields similar conclusions. In this case, the utility function of government officials is given by (12), where government consumes all the capital that it can expropriate, and all of the tax revenue it retains after bringing public investment to the growth-maximizing level. (12) U(cg) = u[(I - F(t -r --gy(O)e$ + F'(t)fc(O)ex] Maximizing (14) with respect to r yields (13) g _ F'(t) * 6k(O) 1 y 1- F(t) y(O) 8r/ 9 This conclusion holds even if households have an outside investment option, not vulnerable to property rights problems and not taxable by the government, as long as the rate of return on the outside option is sufficiendy low. Obviously, if households have a very good investment option, even a slight decline in the property rights environment would lead to a complete abandonment of the more risky productive activity, and public investment would plummet, along widh rents. However, in most countries, most investors/households do not have such attractive exit options. 13 The question we would like to ask is whether rents from taxation, given by.the right hand side of (13), and rents from expropriated capital, given by the last term in (12), increase or decrease as property rights become less secure (as 0 rises). Summing the two and differentiating with respect to 0, yields -h (o) + he k(0)(1-Ohl recalling = h. This expression is greater than zero - that is, total consumption by government officials as a fraction of national income rises as property rights become less secure - when (14) - 1 +eh eA' (1- h9) > O. y(o) The right hand term of (14) is always positive and the left hand term is small. Expression (14) indicates that over the most plausible parameter ranges, rent-seeking does, indeed, generally rise as property rights become less secure.10 The intuition is broadly similar to that in the fiscal model, in that as property rights become less secure, the payoffs to government from deferring rents to the future decline. Since productive public investment, in this model, is unaffected by the security of property rights, the net effect of insecure property rights is to increase observed public investment as a fraction of national income and private investment. There are clear implications for the interpretation of public investmnent coefficients in growth regressions. Since they are based on observed public investment, which includes non-productive public investment created for rent-seeking purposes, the estimated growth effects of public investment are naturally likely to be low or negative.. Failing to control for the effects of the property rights environment on the effects of additional public investment conflates these two sets of countries, leading to no or negative associations between public investment and growth. Conditional on the security of property rights, however, we would predict public investment to have a positive impact on growth. The evidence presented below is consistent with all of these predictions. Different institutional variables and their expected impact In the empirical analysis that follows, we first establish the anomalies that motivate 10 If the expression in parentheses were.01, h were 99, growth were zero, and t one, the right hand side term would be .0037. Initial period income per capita could be lower than $350 (when the left hand term would equal -.0029), and (14) would still be positive. 14 this paper - that the ratios of public investment to national income and to private investment are significantly higher in countdies with insecure property rights. We then examine the implications of the foregoing analysis for economic growth, and look at the interaction of public investment and property rights on growth. Measures of the institutional environmnent are obviously fundamental to these tests. Unfortunately, there are no cross-country data that allow for a specific test of the fiscal model of property rights - the threats that citizens irnpose on each other, and the inherent differences across countries that make these threats more cosdy to combat in some places than others. Available variables measure a mix of both concepts. On the other hand, we are able to use two sets of variables related to the institutional model of property rights. One set refers direcdy to the threats posed by government action to investors. The second set captures some of the underlying institutional factors that are argued in the institutional literature to increase the threat of government expropriation. Risk of expropriation and contract repudiation Indicators of the risks of expropriation or the repudiation of contracts by government are representative of the capacity of the government to act arbitrarily more generally. In the work below, two indicators for these risks are employed, "Risk of Expropriation" and "Risk of Repudiation of Contracts by Governments." These variables are published as part of the International County Risk Guide (ICRG), provided monthly to subscribers, mainly multinational investors. "Expropriation" is a subjective indicator of the risk of confiscation and forced nationalization. "Repudiation of Contracts by Governments" is a subjective indicator of the risk that governments will repudiate or odterwise unilaterally change the terms of contracts with foreign businesses. Because these two indicators provide similar information, in the regressions below an additive index of them is used, labelled "Predictable Government" Bureaucratic quality and corruption Low quality bureaucracies are the second institutional dimension investigated below. Bureaucratic organization and incentives are important components in their own right of the institutional framework within which public investment decisions are made and implemented. They are also indicative of the state of other institutions. The decisions of bureaucrats are nearly always subject to the oversight of politicians. If politicians are 15 unconstrained in the demands that they can place on bureaucracies, or if they are subject to frequent replacement, bureaucratic quality is likely to be lower.11 Directly or indirectly, therefore, bureaucratic quality enters into the determination of the institutional parameter 9. Two variables from the ICRG are relevant The first, "Quality of the Bureaucracy", measures the degree to which the bureaucracy has the strength and expertise to govern without drastic changes in policy or interruption of governmental services. Such bureaucracies have established mechanisms for recruitment and training, and some autonomy from political pressure. For the second variable, "Corruption in Government," low ratings are assigned to countries in which high government officials are likely to demand special payments and where illegal payments are generally expected throughout lower levels of government. Again, because the two variables offer similar information, they are grouped in an additive index labelled "Bureaucratic Performance."12 Executive discretion and competitive elections At the center of institutional arguments related to the insecurity of property rights is the extent to which government hands are institutionally tied or not in questions of expropriation. Underlying the security of property rights and the performance of the bureaucracy is the structure of political institutions. The checks and balances arguments of North and Weingast and others are precisely in this vein: when multiple government actors must agree to expropriation or contract repudiation, the possibility of these actions being undertaken is reduced. Two indicators of executive discretion are used below. One, Executive Constraints, indicates whether the executive in a country is subject to restraints on unilateral decision making by other branches of government or parts of society (aggers, K and T.R. Gurr 1995). The Database of Political Institutions (DPI), version 3 (Beck, et al. 2001) characterizes the presence of checks and balances with the variable Checks3. This variable is built up from several other variables collected in the data set, induding the number of parties in the government coalition (for parliamentary systems), whether the president's party has a majority in the legislature (presidential systems), and whether elections 11 Noll and Fiorina (1978) is one of the first of a growing literature examining the interaction of political and bureaucratic agents. 12 Resuits using either of the two components are very similar to results obtained from using the index. The same is true for "Pretictable Govemment". 16 are governed by closed list or open list rules (the former granting more authority to the heads of parties). On the presumption that constitutional checks on executive behavior mean little if the relevant actors are not elected, the construction of Checks3 also takes into account legislative and executive indices of electoral competitiveness (EIEC and LTEC in DPI), scaled one to seven.13 However, for completeness, we also consider the competitiveness of elections separately. Where elections are more competitive, constraints on excessive rent- seeking and expropriation by leaders are likely to be tighter. Public investment data The other key variable in the analysis below is public investment. To measure this we follow Levine and Renelt (1992) and Devarajan, et al. (1996) by taking data on public investment from the Government Financial Statistics (GFS) of the International Monetary Fund. While GFS has some data on public investment by state and local governments, its most complete and reliable coverage is of central government expenditures, not including investments by state-owned enterprises.14 Property rights and the rate of public investment Both models of property rights predict that, in the presence of rent-seeking that takes the form of non-productive public investment, observed public investment ratios to national income and private investment should be higher in countries that exhibit less secure property rights. Table 1 focuses on the institutional model of property rights, and compares 13 Where there are no elections, countries receive a one; the scores rise to seven when there are multiple candidates and multiple parties, and no single-party or candidate receives more than 75 percent of the vote. If the legislative index of electoral competitiveness is less than five (where five indicates that multiple parties can legally be established, but where only one party wins any seats in the legislature), checks is one. This reflects the notion that legislatures that are not competitively elected are less likely to exercise decision making authority independent of the executive. Otherwise, coding of this variable depends on whether countries are presidential or parliamentary. In presidential systems, checks is the sum of one (if EIEC is greater than four), one (for the president), one for each legislative chamber, and one if the first government party is closer in political orientation (eft, right or center) to the first opposition party than to the party of the president. If the legislature is closed list (voters must vote for parties and cannot register candidate preferences) and the president's party has a majority in parliament, the legislature is not counted as a check. In parliamentary systems, checks is the sum of one (for the prime minister) and the number of parties in the governing coalition; the number of parties is reduced by one if there is a closed list and the prime minister's party is in the coalition. 14 Barro (1991) appears to use the general government public investment, including decentralized government expenditures from GFS where available. For a few dozen countries, Easterly and Rebelo (1993) supplement GFS data on central govemment public investment with World Bank country reports that most notably include data on investments by state-owned enterprise. 17 public investment across countries with different property rights and institutional environments. The institutional/property rights variables are measured at the beginning of the period to minimize endogeneity concerns. Table 1: Average Public Investment to GDP (1974 - 98) Under Different Institutional Arrangements Range Mean values of Significance of (see note) public difference in means investment/GDP (# of countries) Bureaucratic (0 - 6.99) 5.9 (54) .0001 Performance (7 - 12) 3.5 (39) Predictable (0 - 1.99) 6.0 (47) .0001 Government (12 - 20) 3.8 (46) Executive (5 - 7) 5.8 (60) .0001 Constraints (see note) (1 - 4.99) 4.1 (45) Checks3 (1) 6.3 (60) .003 (>1) 4.2 (54) EIEC (1 - 2) 6.5 (31) .054 (3 - 7) 4.9 (83) N.B. Higher values imnply better bureaucratic performance, more predictable govemment, more checks, more competitive elections, but fewer executive constraints. The first and second rows show that public investment as a share of GDP (averaged over 1974-98) is signifcantly lower in countries that exhibit higher scores on the two property rights variables, predictable government and bureaucratic performance.15 However, the institutional model of property rights fiurther suggests that specific political institutions underlie the security of property rights. Again, public investment is higher in countries that exhibit lower values for these institutional variables: executive constraints, checks3 and the competitiveness of executive elections. Table 1, therefore, supports both the predictions and the underlying institutional arguments of the institutional model of property rights. 15 Tanzi and Davoodi (1997) make arguments about the effects of corruption on public investment which lead them to consider the amount of public investment in corrupt and non-corrupt countries: they also find that 18 Table 2: Effect of institutions on public investment/GDP (1974-98) (OLS, robust standW errors in parntheses) Bureaucratic Predictable Checks3 Electoral Performance Government Competitiveness (EMEC) Institutional -0.327 -0.201 -0.945 -0.754 Variable (0.155) (0.124) (0.267) (0.218) Intercept 20.625 23.049 25.050 26.058 (6.677) (7.229) (6.849) (7.244) Log initial -0.757 -0.924 -1.282 -1.250 income/capita (0.720) (0.767) (0.715) (0.790) Area -0.190 -0.258 -0.306 -0.321 (0.202) (0.200) (0.229) (0.247) Population -0.533 -0.499 -0.389 -0.420 (0.291) (0.314) (0.321) (0.366) Secondary 0.022 0.014 0.013 -0.002 enrollment (0.028) (0.031) (0.028) (0.030) Primary 0.045 0.047 0.043 0.058 enrollment (0.025) (0.026) (0.024) (0.021) Initial price of -0.249 -0.234 0.018 0.024 investment goods (0.202) (0.197) (0.208) (0.213) N 77 77 85 85 R2_ _ .45 .42 .43 .47 Table 2 examines the logic underlying Table 1, but controls for other variables that might influence the government decision to supply public investment, including initial income per capita, initial population and area, the initial price of investment goods in the country relative to the US, and average human capital (educational achievement) over the period. Public investment is consistently higher across all measures except for Exewtive Constraints (not reported), for which the coefficient is of the right sign, but has a I-statistic of only 1.08. public investment is higher in corrupt countries. 19 Determinants of the ratio of public to private investment Strictly speaking, the two property rights models, with rent-seeking, predict that where property rights are more insecure, the ratio of public to private capital is higher. Data on the stock of private capital is necessary to test these predictions directly, but is unavailable. Instead, we use investment ratios, focusing as before on the threats to property rights created by the potential of government expropriation.16 Table 3: Average Public to Private Investment Ratios (1974 - 98) Under Different Institutional Arrangements Range (see note) Mean values of Significance of public/private difference in means investment (# of countries) Bureaucratic (0 - 6.99) 27.0 (53) .0001 Performance (7 - 12) 15.5 (38) Predictable (0- 11.99) 28.7 (46) .0001 Government (12 - 20) 15.5 (45) Executive (5 - 7) 32.6 (54) .036 Constraints (see note) (1 - 4.99) 17.9 (44) Checks3 (1) 34.5 (54) .0058 (>1) 18.0 (53) EIEC (1 - 2) 40.1 (29) .002 (3-7) 21.2 (78) N.B. Higher values imply better bureaucratic performance, more predictable government, more checks, more competitive elections, but fewer executive constraints. Table 3 shows the country means of the ratio of public to private investment under different institutional arrangements. In all cases, the ratio in countries where institutional scores are lower is significantly higher than in countries with higher institutional scores, as the "institutional" model of property rights and public investment predicts. 16 At any given point in time, the stock of private capital is determined by the initial capital stock and the change in capital stock over time, d/kdt, or koe4h/ As long as the rate of depreciation of capital is not too low and the time period over which average rates of investment are compared is not too short, though, initial capital stocks are likely to be small relative to capital stocks in most subsequent periods. 20 Table 4: Effect of institutions on public to private investment (OLS, robmst standard emrs in parentheses) Bureaucratic Predictable Checks3 Electoral Performance Government Competitiveness __________ ~~~~~~~~~~~~~~~(EIEC) Institutional -1.229 -1.155 -5.058 -4.176 Variable (0.777) (0.610) (1.527) (1.806) Intercept 74.305 81.880 153.76 159.836 (35.038) (33.905) (65.632) (65.309) Log initial -0.797 -1.010 -5.651 -5.545 income/capita (4.455) (4.193) (4.574) (4.832) Area -0.277 -0.624 1.576 1.465 (0.949) (0.961) (3.463) (3.314) Population -2.176 -1.800 -4.251 -4.361 (1.232) (1.267) (4.028) (3.767) Secondary 0.007 0.001 0.032 -0.034 enrollment (0.121) (0.122) (0.149) (0.156) Primary 0.067 0.062 -0.289 -0.206 enrollment (0.110) (0.107) (0.340) (0.317) Initial price of -1.049 -0.878 0.530 0.642 investment goods (0.889) (0.842) (1.187) (1.213) N 76 76 83 83 R2 .34 .35 .19 .20 As with Table 1, the results in Table 3 might be sensitive to omitted variables and to possible biases created by the use of private investment rather than the private capital stock. Table 4 presents a more controlled comparison of the determinants of the ratio of public to private investment across countries. The log of initial income per capita (from Summers and Heston, 1991) is used, and can be interpreted as a proxy for initial capital stocks. In addition, the regressions in the table include standard explanatory variables employed in cross-country investment equations, including primary and secondary school enrollment (averaged over the period), and the initial price level of investrnent goods relative to the U.S. (from Summers and Heston, 1991). 21 Although the effects of these on the ratio of public to private investment is unclear, education is expected to have a positive impact on private investment; investment goods prices and initial income should have a negative influence. The population and area of a country are also taken into account, as rough proxies of the demand for public goods and economies of scale in infrastructure provision. Both might be expected to increase the demand for public goods, although again their impact on the ratio of public to private investment is unclear. Table 4 reports results for predictable government, bureaucratic perfortnance, Checks3 and the Executive Index of Electoral Competitiveness. The signs of each are as predicted and statistically significant. The less secure are property rights and the less constrained by political institutions are government decision makers, the higher is the ratio of public to private investment. That is, even controlling fot a variety of other possible influences on the ratio of public to private investment, this ratio is still higher in countries with lower scores in the institutional indicators. Executive Constraints (not reported) is again of the correct sign, but not quite significant at conventional levels, with a t-statistic of 1.54. Rent-seeking, property rights and the quality of infrastructure If the analysis in the paper is right, then we would expect the connection between observed public investment and the quality of public infrastructure to be highest in countries where property rights are secure - and where, therefore, incentives to extract high rents in the form of "white elephants" would be lowest. To the extent that observed public investment reflects rent-seeking, however, any relationship between spending and infrastructure services would be weakened. Table 5 presents some preliminary evidence on these issues. The dependent variable is the change in quality of infrastructure over the 1900-2000 period, as measured by an index constructed from four variables: population with access to an improved water source (in percent); access to improved sanitation (in percent); electricity losses in distribution and transmission (as a percentage of output); and kilometers of paved roads.17 Control variables include initial (1990) infrastructure quality, growth in per capita income and population over 17 Changes over the decade are measured in percentage points for the first three variables; because paved roads are measured in kilometers rather than percent, the percentage change is constructed by dividing the 1990-2000 difference by initial (1990) kilometers of paved roads. The infrastructure index is constructed by standardizing each of the four variables to have mean 0 and standard deviation of 1, and taking the mean. The electricity losses variable is reversed, so that greater losses are reflected in lower values of the index. 22 the decade, price level of investment goods (from Summers and Heston), and log of land area (in square kilometers). Table 5: Public Investment, Property Rights, and Improving Infrastructure (m bust standard e x rs in pamntheses) Equation 1 2 3 Institutional Variable None Bureaucratic Executive Performance consttaints Public 0.015 0.014 0.028 Investment/GDP (0.016) (0.028) (0.016) Institutional Variable -0.060 -0.025 (0.033) (0.026) Public Investment* 0.014 0.014 Institutional (0.008) (0.008) Variable Initial infrastructure -0.146 -0.018 -0.006 quality (0.084) (0.090) (0.084) Per capita income -0.230 -1.735 -0.454 growth 1990-98 (1.562) (1.652) (1.581) Population growth 11.085 7.699 9.607 1990-98 (3.972) (4.336) (4.058) Price level of 0.002 0.002 -0.002 investment goods, (0.003) (0.004) (0.001) 1990-98 mean Log of land area 0.033 0.041 0.050 (0.021) (0.026) (0.024) Constant -0.708 -0.320 -0.543 (0.466) (0.548) (0.396) N 89 76 84 R2 .17 .28 .31 Equation 1 in the table shows the weak overall relationship between public investrnent over the decade and improvement in infrastructure. The coefficient on initial quality is significant at the .10 level, and negative as expected. The strongest predictor is population growth. Equation 2 adds burraucraicpeformance and its interaction with public investment. The coefficient on the interaction term is positive (and significant at the .10 23 level), indicating that public investment expenditure produces larger improvements in infrastructure where bureaucratic quality is lower and corruption is higher. This coefficient implies that the impact of public investment varies from a low of -.043 (with a standard error of .029) conditional on a bwreameamticpeformance value of 3 to a high of .079 (with a standard error of.06) when bsuraucraticpec*rmance is 12. Equation 3 uses an alternative measure of constraints on last period political decision makers to expropriate, executive constraints, which varies from a minimum value of one to a maximum of seven. Results are very similar, as the interaction term has a positive coefficient, significant at the .08 level. The impact of public investment on infrastructure, conditional on the minimum executive constraints value of one, is -.024 (a standard error of .025). Conditional on the maximum value of 7 - which characterizes 24 of the 84 countries in the sample - the marginal effect of public investment on infrastructure is .062 (a standard error of .031). In fact, spending is significantly associated with infrastructure improvement as well conditional on any executive constraints values of 5 or more.18 Interactions are not significant, however, usingpredictablegovernment, checks3 and the executive index of electoral competitiveness (EIEC). Public investment and growth Conflicting or inconclusive results permeate the existing, extensive empirical literature on government spending in general and public investment in particular. Barro (1991), De Long and Summners (1991), Levine and Renelt (1992) and Landau (1983) report that public investrnent has an insignificant impact on growth. Easterly and Rebelo (1993), on the other hand, conclude that some types of public investment (transport and communication) have a positive correlation with growth, as does government investrnent in general. Devarajan et al. (1996) show that the substitution of government investment for other types of government expenditures has a negative impact on growth, however. The foregoing analysis suggests that the neglect of property rights and rent-seeking considerations in this literature may have influenced its findings. When property rights are insecure, observed public investment is more likely to consist of rent-seeking, and therefore be unproductive. High levels of observed public investment due to rent-seeking, which would be associated with lower levels of growth (ower, because the taxes that finance higher 18 At 5, the coefficient is .033 and the standard error is .017; at 6, the coefficient is .048 and the standard error 24 rent-seeking suppress growth), are therefore conflated with high levels of productive public investment with potentially positive growth effects. This problem would not emerge if one could easily distinguish productive and non-productive public investment, but the data do not permit this. We test whether the property rights/rent-seeking nexus identified above in fact obscures the growth effect of productive public investment. We do this by including the interaction term (public investment x institutional variable) in the standard empirical growth equation that has been used in the past to examine the effects of public investment The control variables used are the school enrollment and initial income per capita variables from Tables 2 and 3, and private investment/GDP averaged over the period. Initial income is included to account for any "catch-up" effect that might exist, including the effect of initial capital stock.19 Changes in human capital are proxied by primary and secondary school enrollment. The growth effect of public investment is determined in the model jointly with incentives to undertake private investments, so private investment over GDP is included. Any growth specification is vulnerable to the claim that omitted variables account for observed relationships; to mitigate this problem, we follow common practice in including continent dummies in the base specification to capture omitted effects that have continent-wide effects. We also control for inflation. To the extent that governments finance public investment through unsustainable mechanisms, such as money creation, it is less likely to have positive growth effects. Inflation is a better measure of unsustainable financing than tax revenue or deficit measures, since the efficiency with which governments can collect taxes and the extent to which money demand is sensitive to government deficits vary widely. Table 6 displays the core results. Equation 1 is the base specification and includes no institutional variables or interactions. Consistent with Devarajan et al. (1993), the coefficient on public investment is negative and (marginally) significant. Other significant variables include inflation, initial income, and the Africa dummy (the OECD is the omitted category). is .024. 19 However, see Keefer and Knack (1997) for institutional detemiinants of the strength of this effect. 25 Table 6: Public Investment, Property Rights, and Growth (robust standard errors in parentbeses) |Equation l 2 3 Institutional Variable None Bureaucratic Predictable l ~~~~~~~~~~~Perforrnance Government Pulc Investrnent/GDP -0.112 -0.054 -0.021 (0.057) (0.087) (0.075) Institutional Variable 0.085 -0.054 (0.100) (0.110) Public Investment* 0.013 0.042 Institutional Variable (0.017) (0.020) Log of initial per capita -0.679 -0.853 -0.547 GDP (0.350) (0.436) (0.369) Secondary enrollment 0.002 0.002 0.007 (0.017) (0.019) (0.016) Primary enrollment 0.015 0.015 0.016 (0.010) (0.011) (0.010) Private Investment/GDP, 0.058 0.053 0.044 Mean 1974-89 (0.049) (0.049) (0.049) Latin America & -0.780 -0.679 -1.322 Caribbean (0.890) (0.829) (1.055) Sub-Saharan Africa -3.186 -3.280 -3.593 (1.133) (1.130) (1.243) East Asia& Pacific 1.371 1.231 1.062 (1.078) (1.098) (1.143) Middle East & North -0.572 -0.424 -0.638 Africa (0.902) (0.897) (1.103) Eastern Europe and -1.287 -1.207 -1.724 Central Asia (1.328) (1.437) (1.403) South Asia 0.235 0.146 -0.166 (1.053) (0.981) (1.124) Inflation Mean (log) -0.524 -0.488 -0.563 (0.202) (0.209) (0.188) Constant 7.005 7.566 6.625 (3.144) (3.491) (3.067) R2 .55 .56 .58 Note: Sample size is 80. Mean of dependent variable is 1.09. 26 Equations 2 and 3 display results for the two property rights variables, Bureaucratic Petformance and Predictabk Government, and their interactions with public investment20 As the "institutional" model of public investment and property rights predicts, the interaction term in equation 3, with Predictable Gouernment, is positive and statistically significant. This is the variable that most precisely relates to the threat of government expropriation. Although positive, the interaction in equation 2, Bureaucratic Perfojrance, less directly related to government expropriation, is not significant. Whenpredictablegovernment equals its sample mean (12.4 on the 20-point scale), the net effect of public investment on growth is near zero (-.021). Whenpredictablegovernmentis equal to 10 or less (as it is for 28 of the 80 countries in the sample), a one percentage point increase in public investment has a significant negative association with growth. Whenpredictabklgovernment is at its maximum level, a one percentage-point increase in public investment as a share of GDP increases growth by nearly 0.3 percentage points.21 Conclusion The analysis in this paper exposes and explains an important effect of insecure property rights that has not been previously noted: insecure property rights not only drive down private investment; they also distort important public policy decisions of government. We document significant variation in public investment across countries with secure and insecure property rights, showing that observed public investment as a fraction of national income or private investment is far higher in the latter than the former. The intuitive explanation for this finding is not correct: governments do not have an incentive to invest more in productive public investment in order to offset the negative effects of insecure property rights on private investmrient. On the contrary, we show that productive public investment is, at best, unaffected by insecure property rights (when insecure property rights are due to the threat of government expropriation), and at worst is significantly reduced, when property rights are a law and order problem that can be remedied by government spending. 20 Interactions are computed as the product of the deviations of public investment and the institutional variables from the ir sample means. Thus, the coefficient on public investment in equations 2 and 3 indicates the effect of a 1 percentage-point increase in that variable conditional on the institutional variable being equal to its mean value. 21 Only Luxembourg scores a perfect 20, however, followed by Switzerland (19.6), Netherlands (19), and the U.S. and U.K (18.8). Most developed nations score above 17. Therefore, the effect is significant at the .076 27 Instead, we show that insecure property rights increase government rent-seeking. Rent-seeking governments are willing to accept lower rents in the current period if the correspondingly lower tax rate triggers higher private investment and a correspondingly larger stream of future rents. Insecure property rights reduce the future stream of rents and therefore encourage rent-seeking. We show that the effect of insecure property rights on rent-seeking can be dramatic, and indeed is likely to be substantially greater than any reduction in productive public investment. At the same time, we note that observed public investmnent in many countries is associated with rent-seeking activity, thereby suggesting a solution to the puzzle with which we begin the paper. Significantly higher observed public investment in countries with insecure property rights - and the scant or negative effect of that investment on economic growth - is in fact a reflection of the enhanced rent-seeking incentives of governments in environments where property rights are more insecure. The results of this analysis go beyond the usual arguments underlining the importance of institutional issues in economic growth. For example, development policy often focuses on a two-pronged approach that emphasizes both non-distortionary and investment-friendly tax policies, but also secure property rights and protection of investors against unexpected expropriatory activities. The analysis here suggests that these two prongs are not independent: the very insecurity of property rights induces governments to distort fiscal policy in growth-suppressing ways. The results therefore draw attention to the complementarity of institutional quality and policy effectiveness and are likely generalizable to other policy settings: the success of policy interventions generally, like the effectiveness of public investment in boosting growth, is likely to depend on the institutional environment in which those policy changes are made. level for a one-tailed test. 28 References Acemoglu, Daron and James A. Robinson (2001). "A Theory of Political Transitions." American Economic Review 91:4, 938-63 (September). Barro, Robert (1990). "Government Spending in a Simple Model of Endogenous Growth." Journal of Political Economy, 98: S103-125. Barro, Robert (1991). "Economic Growth in a Cross Section of Countries." Quarterly Journal of Economics, 106: 407-444. N. Gregory Mankiw and Xavier Sala-I-Martin (1995). "Capital Mobility in Neoclassical Models of Growth." American Economic Review, 85:1, 103 - 115 (March). Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer and Patrick Walsh (2001). "New Tools in Comparative Political Economy: The Database of Political Institutions." The World Bank Economic Review, 15: 1, 165-176 (September). Belletini, Girogio and Carlotta Berti Ceroni (2000). "Financial Liberalization, Property Rights and Growth in an Overlapping-Generations Model." Retiew of International Economics, 8:2, 348-359. Buchanan, James and Gordon Tullock (1962). The Calculus of Consent. Ann Arbor: University of Michigan Press. Clague, Christopher, Philip Keefer, Stephen Knack and Mancur Olson (1996). "Property and Contract Rights under Democracy and Dictatorship." The Journal of Economic Growth 1:2, 243-276 dune). De Long Bradford and Lawrence Summers (1991). "Equipment Investment and Economic Growth." Quartery Journal of Economics, 106: 445-502. Devarajan, Shantayanan, Vinaya Swaroop and Heng-fu Zou (1996). "The composition of public expenditure and economic growth." Journal of Monetay Economics, 37: 313-44 (April). Easterly, William and Sergio Rebelo (1993). "Fiscal Policy and Economic Growth: An Empirical Investigation." Mimeo, The World Bank. Fenno, Richard (1973). Congressmen in Committees. Boston: Little, Brown and Co. Jaggers, K. and T.R. Gurr (1995). "Tracking Democracy's Third Wave with Polity III Data." Journal of Peace Research 32:XXX 469-482. Kamien, Morton I. and Nancy L Schwartz (1978). "Potential Rivalry, Monopoly Profits and the Pace of Inventive Activity." Review of Economic Studies 45:3, 547 - 557 (October). Keefer, Philip (1996). "Protection Against a Capricious State: French Investment and Spanish Railroads." Journal of Economic Histoy. 56:1, 170-192 (March). 29 ____ and Stephen Knack (1997). "Why don't poor countries catch up? A cross-national test of an institutional explanation." Economic Inquiy 35:3, 590-602 (uly). Keefer, Philip and David Stasavage (2000). "Bureaucratic delegation and political institutions: When are independent central banks irrelevant?" World Bank Policy Research Working Paper 2356. (une). Knack, Stephen and Philip Keefer (1995). "Institutions and Economic Perfortnance: Cross- Country Tests Using Altemative Institutional Measures." Economics and Politics 7:3, 207-228 (November). Kormendi, R. and P. Meguire (1985). "Macroeconomic Determinants of Growth: Cross Country Evidence." Journal of Monetay Economics, 16: 141-63. Landau, David (1983). "Govermment Expenditure and Economic Growth: A Cross- Country Study." Southern EconomicJournal, 49: 783-92. Levine, Ross and David Renelt (1992). "A Sensitivity Analysis of Cross-Country Growth Equations." American Economic Review, September. Niskanen, Willian (1971). Bureaucray and Representative Government. Chicago: Aldine- Atherton. Noll, Roger and Morris Fiorina (1978), Voters, legislators and bureaucrats: A rational choice perspective on the growth of bureaucracy." Journal of Public Economics. North, Douglass and Barry Weingast (1989). "Constitutions and Commitments: The Evolution of Institutions Governing Public Choice in 17th Century England." Journal of Economic Histoy 49 (December): 803-832. Olson, Mancur (1971). The Logc ofCollectiveAction: Public Goods and the Theogy of Groups. New York: Schocken Books. (1994). "Dictatorship, Democracy and Development." American Political Science Review, 87: 567-76. Pritchett, Lant (2000). "The tyranny of concepts: CUDIE (cumulated, depreciated investment effort) is not cabital."World Bank Policy Research Working Paper No. 1660 (May). Skaperdas, S. (1992). "Cooperation, Conflict and Power in the Absence of Property Rights." American Economic Retiew 82: 4, 720-39. Sonin, Konstantin (2002). "Why the Rich May Favor Poor Protection of Property Rights." Working Paper, New Economic School, Moscow. Stasavage, David (2000). "Credible Commitment in Early Modern Europe: North and Weingast Revisited." Mimeo, London School of Economics. 30 Tanzi Vito and Hamid Davoodi (1997). "Corruption, Public Investrnent and Growth." IMF Working Paper WP/97/139 (October). Zak, Paul (2000). "Larceny." Economics of Governance 1: 157-79. Policy Research Working Paper Series Contact Title Author Date for paper WPS2881 Returns to Investment in Education: George Psacharopoulos September 2002 N. Vergara A Further Update Harry Anthony Patrinos 30432 WPS2882 Politically Optimal Tariffs: Dorsati Madani September 2002 P. Flewitt An Application to Egypt Marcelo Olarreaga 32724 WPS2883 Assessing the Distributional Impact B. Essama-Nssah September 2002 0 Kootzemew of Public Policy 35075 WPS2884 Privatization and Labor Force Alberto Chong September 2002 H. Sladovich Restructuring around the World Florencio Lopez-de-Silanes 37698 WPS2885 Poverty, AIDS, and Children's Martha Ainsworth September 2002 H. Sladovich Schooling: A Targeting Dilemma Deon Filmer 37698 WPS2886 Examining the Feasibility of Jerry R. Skees September 2002 E. Laguidao Livestock Insurance in Mongolia Ayurzana Enkh-Amgalan 82450 WPS2887 The Demand for Commodity Alexander Sarris September 2002 M. Fernandez Insurance by Developing Country 33766 Agricultural Producers: Theory and an Application to Cocoa in Ghana WPS2888 A Poverty Analysis Macroeconomic Luiz A. Pereira da Silva September 2002 R. Yazigi Simulator (PAMS) Linking Household B. Essama-Nssah 37176 Surveys with Macro-Models Issouf Samake WPS2889 Environmental Performance Rating Hua Wang September 2002 Y. D'Souza and Disclosure: China's Green- Jun Bi 31449 Watch Program David Wheeler Jinnan Wang Dong Cao Genfa Lu Yuan Wang WPS2890 Sector Organization, Governance, Antonio Estache September 2002 G. Chenet-Smith and the Inefficiency of African Water Eugene Kouassi 36370 Utilities WPS2890 Sector Organization, Governance, Antonio Estache September 2002 G. Chenet-Smith and the Inefficiency of African Water Eugene Kouassi 36370 Utilities WPS2891 Trends in the Education Sector from Nga Nguyet Nguyen September 2002 E. Khine 1993-98 37471 WPS2892 Productivity or Endowments? Hiau Looi Kee September 2002 P. Flewitt Sectoral Evidence for Hong Kong's 32724 Aggregate Growth WPS2893 Banking on Foreigners: The Behavior Maria Soledad Martinez September 2002 A. Yaptenco of International Bank Lending to Peria 31823 Latin America, 1985-2000 Andrew Powell Ivanna Vladkova Hollar WPS2894 Telecommunications Sector Jean-Paul Azam September 2002 P. Sintim-Aboagye Reforms in Senegal Magueye Dia 38526 Tchetche N'Guessan Policy Research Working Paper Series Contact Title Author Date for paper WPS2895 Telecommunications Reform in Jean-Jacques Laffont September 2002 P. Sintim-Aboagye C6te d'lvoire Tch6tche N'Guessan 38526 WPS2896 The Wage Labor Market and John Luke Gallup September 2002 E. Khine Inequality in Vietnam in the 1990s 37471 WPS2897 Gender Dimensions of Child Labor Emily Gustafsson-Wright October 2002 M. Correia and Street Children in Brazil Hnin Hnin Pyne 39394 WPS2898 Relative Returns to Policy Reform: Alexandre Samy de Castro October 2002 R. Yazigi Evidence from Controlled Cross- Ian Goldin 37176 Country Regressions Luiz A. Pereira da Silva WPS2899 The Political Economy of Fiscal Benn Eifert October 2002 J. Schwartz Policy and Economic Management Alan Gelb 32250 in Oil-Exporting Countries Nils Borje Tallroth WPS2900 Economic Structure, Productivity, Uwe Deichmann October 2002 Y. D'Souza and Infrastructure Quality in Marianne Fay 31449 Southern Mexico Jun Koo Somik V. Lall WPS2901 Decentralized Creditor-Led Marinela E. Dado October 2002 R. Vo Corporate Restructuring: Cross- Daniela Klingebiel 33722 Country Experience WPS2902 Aid, Policy, and Growth in Paul Collier October 2002 A. Kitson-Walters Post-Conflict Societies Anke Hoeffler 33712 WPS2903 Financial Globalization: Unequal Augusto de la Torre October 2002 P. Soto Blessings Eduardo Levy Yeyati 37892 Sergio L. Schmukler WPS2904 Law and Finance: Why Does Legal Thorsten Beck October 2002 K. Labrie Origin Matter? Ashi Demirgu,-Kunt 31001 Ross Levine WPS2905 Financing Patterns Around the World: Thorsten Beck October 2002 K. Labrie The Role of Institutions Asl1 Demirgu,c-Kunt 31001 Vojislav Maksimovic WPS2906 Macroeconomic Effects of Private Lourdes Trujillo October 2002 G. Chenet-Smith Sector Participation in Latin Noelia Martin 36370 America's Infrastructure Antonio Estache Javier Campos WPS2907 The Case for International Antonio Estache October 2002 G. Chenet-Smith Coordination of Electricity Regulation: Martin A. Rossi 36370 Evidence from the Measurement of Christian A. Ruzzier Efficiency in South America WPS2908 The Africa Growth and Opportunity Aaditya Mattoo October 2002 P. Flewitt Act and its Rules of Origin: Devesh Roy 32724 Generosity Undermined? Arvind Subramanian WPS2909 An Assessment of Carsten Fink October 2002 P. Flewitt Telecommunications Reform in Aaditya Mattoo 32724 Developing Countries Randeep Rathindran