Digitized by the Internet Archive in 2011 with funding from Boston Library Consortium IVIember Libraries http://www.archive.org/details/unbundlinginstitOOacem L5 c3- Massachusetts Institute of Technology Department of Economics Working Paper Series Unbundling Institutions Daron Acemoglu Simon Johnson Working Paper 03-29 July 2003 Room E52-251 50 Memorial Drive Cambridge, This MA 021 42 paper can be downloaded without charge frpm the Network Paper Collection at Social Science Research http;//papers.ssm.com/abstract_id=XXXX5o(X t^L0O^ INSTITUTE S£P 192003 I Unbundling Institutions* Daron Acemoglu Simon Johnson MIT MIT July 2003. Abstract TMs paper evaluates the importance of "property rights institutions", which protect by the government and powerful ehtes, and "contracting citizens against expropriation institutions", wliich enable private contracts between citizens. We exploit exogenous by colonial liistory, and dociunent strong firstbetween property rights institutions and the determinants of European stage relationships colonization strategy (settler mortality and population density before colonization), and between contracting institutions and the identity of the colonizing power. Using tliis instrumental variables approach, we find that property rights institutions have a first-order effect on long-run economic growth, investment, and financial development. Contracting institutions appear to matter only for the form of financial intermediation. A possible explanation for this pattern is that individuals often find ways of altering the terms of variation in both types of institutions driven their formal and informal contracts to avoid the adverse effects of contracting institutions, but axe imable to do so against the risk of expropriation. Keywords: Contracts, Economic Grovrth, Financial Development, Institutions, Law and Finance, Legal Formalism, Legal Origin, Pohtical Economy, Politics, Property Rights. JEL Numbers: E44, G18, KOO, N20, P16, P17. * We thank Simeon Djankov, Rafael La Porta, Amir Licht, Florencio Lopez-de-Silanes, Andrei Shleifer, and seminar participants at the Canadian Institute for Advanced Research, MIT, and the University of Illinois for helpful comments and discussions. We also thank Simeon Djankov, Ross Levine, Florencio Lopez-de-Silanes, and Todd Mitton for generously providing data. Introduction 1 Douglass North opens Structure and Change in Economic History by distinguishing be- tween a "contract theory" of the state and a "predatory theory" of the state (1981, pp. According to the 20-27). legal first theory, the state and associated framework that enables private contracts to facilitate institutions provide the economic transactions According to the second, the state "reduce transaction costs"). transferring resources from one group to another. Throughout is (i.e., an instrument for book, North develops liis a story combining the two theories, and argues that good institutions will simultaneously support private contracts and provide checks against expropriation by the government or other pohtically powerful groups. There is a growing consensus among economists and poUtical scientists that the broad outlines of North's story are correct: the social, economic, legal, of a society, i.e., its "institutions," is and political organization a primary determinant of economic performance. However, Hke North, the contemporary hterature has not attempted to determine the relative roles of institutions supporthig private contracts institutions constraining tions").^ Instead, it ( "contractmg institutions" ) and government and ehte expropriation ("property rights has documented the importance of a "cluster" of institutions that include both contracting and private property protection elements. This is weU-estabhshed theoretical arguments emphasizing each set of institutions. ple, the contract Wilhamson (1975, 1985), links the efficiency of organizations For exam- and societies to what type and enforced, and thus underscores the importance of con- tracting institutions (see also Grossman and Hart, 1986, Hart and Moore, 1990, and Hart, In contrast, other authors emphasize the importance of private property rights, especially their protection against goverrunent expropriation (see, 1981, in spite of theory hterature, starting with Coase (1937, and especially 1960) and of contracts can be written 1995). institu- De Long and This paper is Slileifer, among others, Jones, 1993, or Olson, 2000). an attempt to mibmidle the broad cluster of institutions, and learn more we could refer to institutions constraining government expropriation as "political inwe did in a previous version of this paper. We decided not to use this term, since some ^Alternatively, stitutions" as readers interpreted it as referring to the type of constitution or to the ideological leanings of politicians We opted for the former term, since certain aspects of the legal institutions, such as the independence of the judiciary from politics, may have an important effect on the security of property rights against expropriation by the or society. In addition, "contracting institutions" could be called "legal institutions" government or other powerful groups. . about the relative importance of contracting versus property rights institutions at the macro level. Such an attempt has to We will tions. start with some proxies two for the sets of institu- proxy property rights institutions using two alternative measures: Pohtical Risk Services' assessment of protection against government expropriation in a country, and Polity IV's constraint on the executive measure.^ For contracting would like ties are ordinary citizens), and able measures the of collecting we proxy number this with the data on legal formahsm developed Shleifer (2002, 2003). TMs on an unpaid check or evicting a non-paying tenant. These authors show formahsm have higher and lower perceived longer delays in courts, fairness costs of enforcing simple contracts, and efficiency of the judiciary system. then natural to presiuxie that greater legal formahsm is a proxy for worse contracting In Ordinary Least Squares (OLS) regressions, long-run economic growth, institutions."' investment rates, and financial development are correlated with both contracting and property causal effect. rights institutions. To make further progress, of variation in both legal formahsm and However, we need to OLS correlations isolate potentially and Porta, Lopez-de-Silanes, Lopez-de-Silanes, (1960) and Slileifer, Merryman exogenous sources private property rights. on the degree of effect Shleifer (2002, 2003), building and Vishny (1997, 1998) and by (1985), legal insti- do not estabhsh a Fortunately, the hterature offers potential instrrmients for both variables. La vari- of formal legal procedures necessary to resolve a simple case that coimtries with greater legal tutions we a measure of costs of enforcing private contracts (contracts where both par- and used by Djankov, La Porta, Lopez-de-Silanes, and It is institutions, show that the on work by La Porta, legal scholars such as Dawson "legal origin" of a coimtry has formahsm, and most relevant Djankov, an important for our sample, coimtries with French legal origin have substantially higher degrees of legal formalism than Enghsh legal origin countries. Moreover, as they point out, at least for former legal ^ system can be thought of as "exogenous" because , Constraint on the executive this measure as a proxy for is it was European largely colonies, the imposed by colonial clearly related to the "political institutions" of a society. Our use of property rights institutions reflects the fact that political constraints on the executive are closely interwoven with the security of property rights. al. do not relate legal formalism to long-run economic growth. But this seems a natural formalism affects the costs of contract enforcement (as Djankov et al. argue) and if contracting institutions are important for growth (as North and many others argue), then there should be •^Djankov et step: an if legal formalism on economic growth. formalism data are available from the World Bank's Doing Business website, under the heading Contract Enforcement: http://rru.worldbank.org/DoingBusiness/. effect of legal The legal powers.'* Our previous work Acemoglu, Johnson, and Robinson (2001, 2002), on the in other hand, shows the importance of the mortahty rate facing potential European settlers and population density before colonization on the colonization strategy of Europeans. Via this channel, these variables have influenced the liistorical and the degree of private property society relations development of the state- rights enforcement in the former colonies today. Our approach in this paper is to use a multiple instrumental variables (IV) strategy, exploiting these sources of variation. on the two The success of the multiple sets of instruments to isolate the contracting TV strategy depends and property rights charmels. In this respect, colonial history offers an ideal setup. In the sample of former European colonies, the legal legal formalism, system imposed by colonial powers has a strong and almost no effect on measures of property At the same time, both mortality rates for potential effect on the degree of rights institutions today. European settlers and population density in 1500 have a large effect on current property rights institutions, and no impact on measures of The ing: we legal formalism. results of our empirical investigation using this multiple find strong support for the IV strategy axe interest- importance of property rights institutions on current economic outcomes. Countries with greater constraints on politicians and eUtes, and more protection against expropriation by these powerful groups, have substantially higher in- come per capita (i.e., liigher long-rmi to the private sector relative to growth rates), greater investment rates, GDP, and more developed stock markets. our findings indicate that the role of contracting institutions with greater legal fonnalism have for the effects of less is more credit In contrast, more Umited. Countries developed stock markets. However, once we control property rights institutions, legal formahsm seems to have no impact on income per capita, the investment to GDP ratio, and the private credit to These results suggest that contracting institutions affect the GDP ratio. form of financial inter- ^The La Porta et al. papers suggest a number of channels through which legal origin could affect economic outcomes. In addition, Glaeser and Shleifer (2002) argue that the origin of the legal system affects not only legal transactions, but also regulates the power of politically powerful groups. Mahoney (2001) also argues that legal origin has a positive effect on economic growth through channels other than legal formalism. If these views are correct, our IV estimates for the importance of legal formalism, and therefore for contracting institutions, will be biased upwards. Consequently, our estimates of the importance of contractual institutions can be interpreted as upper bounds. The reduced-form evidence in the Appendix Table A3 does not show any evidence of a significant (direct or indirect) effect of legal origin on economic growth, investment, bank credit, or stock market development once we control and instrument for property rights. mediation, but have less effect on economic growth, investment, and the overall level of financial development. It seems that society can function in the face of weak contracting institutions without first-order a significant economic risk of expropriation costs, but has a much harder time deahng with from the govermnent or other powerful groups.^ Our terpretation, consistent with the simple model we use to highhght the distinction between contracting and property rights institutions, is that contracting institutions affect pri- when vate transactions and create ex post transfers between parties (for example, face large costs of collecting on in- their loans lenders from borrowers). Private contracts or other reputation-based mechanisms can, at least in part, alleviate these problems.*' For example, when it is more difficult for lenders to collect banks that can monitor effectively will play a credit relationsliips will develop. ments therefore Umit the on their loans, interest rates increase, or more important role, or reputation-based Private contracting and alternative financial arrange- effects of contracting institutions and legal formalism. In contrast, protection of private property rights relates to the relationship between the state and the citizens. elites, private citizens When there are no checks on the state, on pohticians, and on do not have the security of property rights necessary for investment. In this case, they are also unable to enter into private arrangements to circumvent these problems; it is impossible to write credible contracts with the state to prevent future expropriation, since the state, with arbiter of contracts (see At this point its monopoly of legitimate violence, is the ultimate Acemoglu, 2003a). we have to emphasize the limitations of our analysis. First, to the extent that contracting and property rights institutions interact in regulating relations between the state and citizens, or even between citizens, the interpretation of our results 'This pattern La is also consistent with the results We barriers data more from Djankov, number of new business is legal formalism greater legal formalism, naturally, means more when we use their measures for the total costs of opening new businesses, the main Porta, Lopez-de-Silanes, procedures to open a procedures. However, we obtain using the entry is and Shleifer (2002). find that the key determinant of the — determinant appears to be property rights institutions, not legal formalism. Therefore, it again appears that contracting institutions aiTect the form of economic transactions, but have less impact on the overall economic outcome (here, the total cost of opening a business). ®See the emphasis of, among others, Ellickson (1991) and Greif (1989) on the ability of individual agents to substitute reputation-based arrangements for legal contracts. The World Bank's Doing Business study has found that in countries such as Bulgaria, Egypt, Mozambique, and Tunisia, creditors structure contracts so as to be able to seize collateral when a borrower defaults without using standard slow court procedures (Djankov, 2003). Naturally, as also highlighted by the theoretical model below, there may exist a certain threshold beyond which contracting institutions may matter more, and consequently, our results here are consistent with the notion that a substantial worsening in contracting institutions could have significant economic implications. difficult/ Second, our "property rights institutions" are and could reflect the eff'ect In addition to work by still somewhat of a black box, of other political or non-pohtical institutional features.^ La Porta et al., Djaakov et al., and Acemoglu et al., the papers closest to our work are Beck, Demirgiig-Kuiit, and Levine (2003a, 2003b), and Rajan and Zingales development. origin (2003), wliich critically evaluate the eff'ect of legal origm on financial Beck, Demirgiig-Kunt, and Levine (2003a) find evidence that both legal and potential settler mortality matter for financial development.^ However, they only estimate reduced-form relationsliips and do not specify the mechanisms tlirough wliich legal origin may and Levine (2003b) test affect economic and financial outcomes. Beck, Demirgtig-Kunt, whether legal origin matters because it affects state control over the judiciary, or because some legal systems are more "adaptable" than others. They do not address the nature or origin of property rights institutions. ^° Rajan and Zingales (2003) off'er of investor protection in Europe. an "interest group" explanation They argue that changes for the in financial development arrangements at the turn of the twentieth century are evidence against "time invariant" explanations, such as the legal origin approach, and instead support their theory in which incumbent producers oppose financial development to prevent entry from newcomers. In contrast, in our model and empirical work, we focus on the effect of legal origin on legal formahsm, and show that contracting institutions and legal formalism generally matter less than property rights institutions for credit, investment, and long-rim economic growth.^^ ^But in our defense, we find no evidence in the data politicians and political elites are in for a significant interaction effect between formalism also determines how constrained practice, this would create an upward bias in our estimates for the property rights and contracting institutions. Moreover, if legal importance of contracting institutions. '*In Acemoglu, Johnson, and Robinson (2001, 2002), we provide detaOed evidence that the effects of mortality rates for potential European settlers and population density in 1500 are working through institutions, and not through geographic, religious, or some other omitted factors. '^Levine (2002), Beck and Levine (2002), and Demirgug-Kunt and Maksimovic (2002) all find a link between legal origin and both the level of financial development and the extent to which external finance is market- rather than bank-based. Levine (2003) reports results where legal origin explains the level of financial development across countries and these in turn account for differences in long-run growth. ^''See also Berkowitz, Pistor, and Richard (2003), who argue that countries that developed their own adapted any "transplanted" law, have a more effective legal system, and Johnson, McMillan, and Woodruff (2002), who present micro evidence from five post-communist countries showing that effective property rights matter more for firm investment than financial constraints. Finally, there is also some recent work investigating which type of institutions matter more for economic outcomes. For example, Persson and Tabellini (1999) find that within the set of democracies, presidential regimes have smaller governments and majoritarian (non-proportional representation) electoral systems are correlated with less government spending and less welfare spending. Barro (1997), on the other hand, investigates the relative importance of rule of law and democracy in stimulating economic legal systems, or that substantially ' ^ Section 2 develops a simple model to liighlight how contracting and property rights different effects on economic outcomes. Section 3 discusses our empirical strategy and the basic data. Section 4 provides details on the sample and institutions might have descriptive statistics. Section 5 shows some basic univariate results. Section 6 provides our main results, contrasting the impact of contracting and property rights institutions on a range of economic outcomes. It also contains robustness checks. Section 7 concludes. Theory 2 We now outhne a simple reduced-form model to higlihght how differences in contract- and property ing rights institutions affect financial We and economic outcomes. think of contracting institutions as mainly affecting the costs of enforcing private contracts. Property rights institutions, on the other hand, determine the interaction between individuals and the government (and the pohtical ehtes who property rights institutions do not constrain these eUtes are elites, control the government). more Ukely When to violate the property rights of individual producers and expropriate their incomes or assets. The pm-pose of this model is not to develop a micro-founded analysis of financial intermediation or the role of property rights institutions, but simply to higlihght issues that will help with interpretation of our empirical results. In particular, the model trates how, under certain circumstances, contracting institutions of financial transactions but have relatively Umited effects on the may illus- influence the form overall level of financial intermediation, investment, and output, because individuals can vary the terms of their contracts ex ante to deal with the ex post costs of contract enforcement. The Environment 2.1 The model and the for lasts for elite. one period and consists of three groups of agents: producers, lenders, All agents are risk neutral. investment or consumption. have sufficient We There is a imique good that can be used normalize the number of producers to 1. Lenders funds to lend to producers, but no investment opportunities. Producers have productive investment opportimities but no funds to undertake these investments. These studies typically do not isolate an exogenous source of variation in institutions, so the reflect omitted factors or be driven by differential measurement error (hence attenuation bias) various measures of institutions. growth. results in may Finally, elites do not have a directly productive role, but control the state apparatus, and can use their poUtical power to expropriate the incomes of other groups in We will ^^ think of contracting institutions as regulating the relationsMp between lenders and producers, and property rights uistitutions as affecting the relationship ducers and lenders on the one hand and the Each producer can produce a > elite distribution of Cj in the population is e^ when them needs this using either a debt contract or other. good by investing unit of loan to a producer with a debt contract to borrow 1 -I- m'^ administrative costs, or costs of collecting funds from savers. R that The G {e). unit from the lenders. We an equity contract. is 1 Each miit. 1 she undertakes production. discussion of the debt contract, returning to equity contracts later. a gross interest rate between pro- given by a continuous distribution function Since producers have no funds, each of They can do on the units of the final 1 producer j also incurs a non-pecuniary cost 1 society. The cost of providing where m'^ < a A debt start with a — 1 represents contract will specify the lender has to repay. However, the producer can renege on her payment promise. In this case, the lender can take the producer to court. The cost of taking producer j to court, i.e., filing a complaint, c^ + where C^ is degree of "legal formalism". how costly example, if e, a feature of the legal system that imphes that enforcing debt contracts it will be is more The parameter is is (1) specific to debt contracts. costly, wliich may be is a C^ liigh which determines malfeasance by the producer. For the project turns out to be very complex (or difficult to monitor or adjudicate), the producer may find ways of not repaying vdthout the court easily detecting this, e.g., diverting the proceeds wliile pretending to be bankrupt. If the lender he incurs the cost in (1) files a complaint, and always wins and receives the promised payment, while the producer incurs some positive cost '^A central assumption here is e. If the lender does not file a complaint after the that producers are distinct both from the ehte and from the lenders (we do not need these groups to be disjoint, but simply to be sufficiently distinct). is higher because there 9j is project complexity, for the lender to prove that there is A reasonable given our focus on financial relations and growth in This Eissumption modern economies. Although there have been societies such as the plantation economies of the Caribbean between 17th and 19th centuries where the elites were also the producers, in most modern societies there are important producer groups outside the elite. Moreover, in many societies expropriation by government is a major concern, and in this case there is a natural distinction between producers and groups involved in expropriation. In addition, the distinction between lenders and producers is a key element of any modern economy without this distinction, a discussion of financial intermediation would not be meaningful. — producer reneges, there We project. is assume that Finally, the ehte no repayment and the producer receives the 6j is distributed full output of the uniformly in the population between can decide to expropriate the returns from the project. and 1. However, they can only do so when the checks and balances that the society imposes on them are We sufficiently lax. model by assuming that "property rights institutions" constrain tliis the ehte in their expropriation. More the aggregate state of nature a is this formulation, P specifically, after and expropriation realized, measures the degree of investment decisions take place, political checks avoided is if a < P. In and balances on the ehte. For example, a stronger, more independent legislature (assuming this represents non-elites) would correspond to a liigh value of the eUte expropriate producers. P, meaning that only in special circumstances can We assume that the distribution of a continuous distribution function, F(cr); What modelling exphcitly here. to emphasize that imposes on the The timing elite tliis represents political events that tliis given by a we are not matters for the economic decisions of producers and lenders will be the risk of expropriation we want is by the risk is in elite, and with tliis reduced-form modehng tm:n related to pohtical constraints that society and the government. of events in this economy is as follows: 1. Producers observe their non-pecuniary cost of production, {cj}. 2. Lenders compete to provide funds to producers. 3. The aggregate 4. Producers and lenders observe the reahzation of the project complexity, {9j}. 5. If state a is reahzed, and the ehte take the expropriation decision. there has been no expropriation, producers decide whether to renege on their payments. 6. Lenders decide whether to take producers 7. Returns are reahzed and consrmied. We who have reneged to court or not. next characterize the subgame perfect equilibrium of this game in the standard way by backward induction. This equihbrimn will turn out to be imique. Equilibrium with Only Debt Contracts 2.2 Let us start with a subgame where there has been no expropriation and producer who has borrowed Then at the rate R, reneges. the payoff from the strategy of fihng a complaint for the lender, conditional on the realization of project complexity V (file = R — l — m'^ — 9j — C'', \R,9j) incurred the investment cost, when he does not but legal costs, file 1+ a complaint will also is 9j, F (no the legal costs will incur file | = —1 — ^j) /?, rri'^, + C^. The Oj payoff as he wiU not incirr not receive the payment R. Therefore, the lender wiU file if ej<R- C^. when Since reneging on the debt contract no benefits, the producer will is: since in this case he will get back R, he has already and he m*^, j, (2) the lender renege only when files a complaint does not hold. (2) assumption, the probability of repayment before the reaUzation of is costly and creates Given the miiform 9j is therefore min(max(i?-C"',0),l), where the and tween < that max and min 1. We — C"^ < i? > whenever a are realized) is sure that the probability remains assume tliroughout that 1 (see footnote 14). C"^ > (l + m'') F {P)~ — 1, Then, using the fact that the ehtes boimded be- which ensures will expropriate P,^^ the expected return of a lender at the lending stage (before a and 9j is: V (lend where the make operators first \R) = -l-m'^ + F{P){R- C'^)R, two terms are the costs of lending and monitoring, and the tlaird term the probability that there will be no expropriation, F{P), times the probability that there no reneging, {R is lender will receive cient funds among 0. — C^), times the repayment amoimt, R. Thus, we need lenders to cover V (lend all the | i?) > demand 0. By not lending, the Moreover, since there are for funds suffi- from producers, whenever V (lend R) > 0, competition between lenders wiU reduce R, so in equiUbrium we must have V (lend i?) = 0. Straightforward algebra shows that the miique positive solution \ | '^With a reasoning similar to that for the Laffer curve, for very high values of P the ruling elite may want to reduce P in order to encourage greater investment and increase their revenues. The simplifying assumption here is that they are unable to do so (see Acemoglu, 2003b, for or the government more discussion). to this equation is:^^ R= However, for the ^ (3) R< producer to be able to repay, we also need a, or C^<C^F).a-(l±^4^^. (4) In other words, for the credit market not to collapse, the legal costs of enforcing contracts need to be less than a critical threshold. Moreover, C'^ (P) is increasing in P, so that better property rights institutions increase the range of legal costs over which a credit market can be supported. It dR/dC^ > can also be verified that institutions and stronger constraints on and dR/dC^ < elites thus better contracting 0, reduce the required repayments (the interest rate) in the credit market. Now going to the first stage of the game, we can write the expected utihty of a producer with effort cost Ej as: U,{e,\R) = F{P)[a-{R-C')R]-e^, which takes into accomit that with probabihty ehte. Otherwise, the producer receives the output back to lenders depends on the reahzation of < R— 0j we C'^i \ so she will have to make the — F (P), a, 9j. there is expropriation by the and whether she makes the payment With probability R— C^, we have payinent. Using (3) to substitute out for R, obtain: Uj {cj) Notice an important feature of on C^. This is = F{P)a-l-m'^- (5): as long as (4) Cj. is satisfied, (5) Uj (cj) does not depend because lenders and producers write ex ante contracts, and they can change the terais of these contracts to deal with the fact that there is a low probabihty R— C^ > is always true. Straightforward algebra using (3) establishes that as long C^ > (l + to'*) F{P)~^ — 1, we also have i? — C* < 1, as claimed in the text. Alternatively, if C* < (1 + m'^) F {P)~^ - 1, then the probabihty that 6j < R - C^ is equal to 1, and as a result, R = {1 + to'') F(P)"\ and the credit market collapses only if (l + m'') F (Py^ > a. Thus, if C* < ^"•Note that as (l a a + m'') F {P)~ — 1, the interest rate is R — (1 + to'')F(P)~ (as long - (1 + m'^)F{Py'^ /a > C' > {1 + m'^) F {Py'^ - 1, the interest rate is — (l + to'') F {P)~ /a, the credit market collapses. 10 as this is less than given by (3), and if a); if C^ > of repayment when C^ high. In other words, is R adjusts to keep the expected payments from the producer to the lender constant irrespective of the value of C^. Therefore, when C^ is liigh, the producer more likely to be able to avoid payments, have to promise to pay a liigher interest will To so when calculate total investment not collapse, notice that 0, is we need Uj distribution of Cj for is given by G (e), | i?) all investment is (4) is satisfied, i.e., when the credit market does Thus using the producer j to invest. total investment financially intermediated, this diation) in the economy. This expression and rate.^^ is fact that the given by: = G[F{P)a-l-7n% I Since she the producer does not midertake the investment, she receives if > (e^ l^ut in retirrn, credit are independent of C^. is (6) also total credit (finaiicial interme- shows that, as long as (4) is satisfied, Contracting institutions therefore investment may have limited effects on investment, because ex ante contracting enables the parties to circumvent po- tential enforcement problems. However, there are limits to this argument: contract enforcement, C^, is very if the cost of the credit market will collapse and there will be liigh, no investment. when Notice also that (4) pohtical constraints on ehtes, have a "more important" is holds, investment dl /dP > i.e., effect is always increasing in the degree of The reason why property rights institutions 0. on investment than contracting institutions in this model between lenders and producers, there are no ex ante that, in contrast to contracts contracts between producers (or lenders) and the elite that can be used to circumvent the ex post holdup and expropriation problem. Here we take possibilities are absent, but it is clear why tliis it would be as given that such contracting so: it is impossible to write credible contracts with the state to prevent future expropriation, since the state, with its monopoly We of legitimate violence, can now summarize '''The fact that these two tliis is the ultimate arbiter of contracts.^'' discussion as follows: effects exactly cancel never go to court along the equilibrium path. If in this case, is a special feature of this model, where parties R would have to increase further to compensate C^ would have an effect on Uj {cj). Nevertheless, it ex ante contracting possibilities would reduce the effect of C^ on lenders for the expected court costs as well, and can be verified that even out they did, investment. "'One way of "writing" such contracts through trigger strategies in a repeated game, whereby if Acemoglu (2003a,b) shows that there are generally limits to how useful these trigger strategies will be in a political context. is the state expropriates too much, agents will stop investing. 11 Result 1 In the model outlined above with only debt C^ < for legal costs C"' (P) such that if on equilibrium investment. and investment falls If C^ & rises Qd ^ Qd ^p-j^ g^ higher Figiu-e 1 is a simple {P), changes in legal costs have no effect above C^{P), the credit market collapses to zero. Greater constraint the likeUhood of credit market collapse, P contracts, there exists a threshold i.e., on expropriation, P, reduces increases C'^ (P). Moreover, as long as raises investment. diagrammatic representation of the equilibrium, with the costs of debt contract enforcement, i.e., C* on the legal formalism, on eUte expropriation, P, on the horizontal constraint elite vertical axis and Above the hne &{P), the axis. market collapses because the costs of contract enforcement are too high. In credit legal this P does not affect investment, since investment is already 0. Below line the credit market exists and a liigher value of P raises investment directly. region a higher value of the C''- (P) In this region, private contracting undoes the effects of higher legal costs, so effect on investment. to cross 2.3 A greater C^ reduces investment only when from above to below the threshold hne C^ has no induces the economy it C"^ (P). Equity Contracts To introduce equity we assume that contracts in the simplest possible way, in stage 2 of the above timing of events, lenders can decide between a debt contract as described above, or an equity contract where they returns. With will lend 1 unit in return for equity contracts, raising impUes that, everything 1 unit of funds costs 1 a fraction s of the project + m^, else equal, equity finance is cheaper, for where The different makes better use costs of court action from the of market information (e.g., when the producer < vn!^. This example because money can be raised from a broader group of savers, or because equity finance less costly or rrf is administratively Holmstrom and Tirole, 1993). reneges on her equity payments are legal costs involved in enforcing debt contracts, and are equal to C^ > C^, so that setthng disputes related to equity are more costly. This might be, for example, because when the producer fails to repay debt, the lender her assets. In contrast, shareholders do not have access to may be tliis able to foreclose on option when the producer does not pay dividends, making the enforcement of equity contracts more costly for lenders (see, e.g., The Hart, 1995). analysis in the pure equity case is parallel to the debt case, 12 and we can easily see when the producer reneges on that 6j < and sa — C^. The equiUbrium naturally, we need market to function, = s < a complaint only file ^ P) 2a wliich gives a condition similar to before for the credit 1, i.e., C < C (P) Once Uj again, as long as this condition (ej) = F {P)a — 1 — — Tjf Cj, (8) is satisfied, we have the utility of and consequently, aggregate investment producer j as is I^G[F{P)a-l-m^], which, Hke (6), Result 2 In the model outlined above with only equity for legal costs C"^ (P) such that on equilibrium investment. (i.e., (9) does not depend on legal rules C^, as long as investment falls increases If if C^ to zero. Greater C^ (P)), and C^ < C^ (P), changes rises P above (8) holds. Tliis gives us: contracts, there exists a threshold in legal costs have no effect (P), the equity market collapses C"^ and reduces the likeUhood of equity market coUapse as long as C^ < C"^ (P), it straightforward to characterize the equihbrimn It is also if: equity share given to lenders will be: — ^ the payment, the lender wiU raises investment. when both debt and equity contracts are available. In this case, competition between lenders ensiues that the contract form maximizing the utilities of utility of producers producers with debt and equity contracts. As long as credit markets will function wiU prevail as long as m*^ < when (4) holds, contracts even Result 3 If if m^ < both contracts (4) but m'', i.e., as long as equity may be is the more and when m^ < and (8) are satisfied, We summarize form of financial the equilibrium will featiue debt this discussion as follows: (8) are satisfied, m'^. If (4) efficient greater than C^, an interesting configiuation (8) fails to hold. In this case, m'^. (4) both with debt and equity contracts, and equity contracts intermediation. However, since C^ arises Thus, we need to calculate the will prevail. the unique equifibrimu involves only eqvuty holds and (8) involves only debt contracts. 13 fails to hold, the unique equiUbrimn Now in light of these results, let us compare two economies that have the same P, but one has a higher legal cost of enforcing both debt and equity contracts, corresponding to high values of C and C"^, and the other has low values of C^ and C^. enforcement-cost economy, but not in the high-cost one, and will if (4) If (8) holds in the low- holds in both, then we observe debt contracts in the high-cost economy and equity contracts in the low-cost economy. Despite this sharp difference in the form of financial intermediation, differences in total credit and investment may be in the high-cost differ a potential interpretation for wiU total investment G [F (P) a — 1 — economy and investment levels wiU not rules small: much in the rrf] in is G [F (F) a the low-cost one. — 1 If rrf — m^] « m'^, two economies. This configuration provides our empirical results where economies with different legal exliibit large differences in the form of financial intermediation, particularly the use of equity, but only small differences in levels of long-run income, investment, and total credit (debt plus equity). Therefore, in this model, legal costs of private contract enforcement can influence the form of financial intermediation at the same time as having a relatively small effect on investment and output. In contrast, the by the state and pohtical ehtes typically has a major effect risk of expropriation on investment output. Empirical Strategy and Data 3 Basic Specification 3.1 We are interested in investigating the separate effects of contracting institutions Ignoring nonhnearities, the economic relationship property rights institutions. we and are interested in identifying can be written as: Fc where Y^ is =« • F, + /? • 7e -F the outcome of interest for coimtry Z^ 7o + e, c, for example, per capita income, the investment rate, or the level of financial development. Fc which captures the institutions, 7o is and Zc is a measure of legal formahsm, legal costs of contract enforcement, Ic is a is a set of other controls, a and j3 (10) measure of property rights are the parameters of interest, and a vector capturing effects of the control variables in Zc-^^ ^^In addition, we have also investigated whether there is contracting institutions by adding interaction terms such an interaction between property rights and and whether there are significant jFc We did not find any evidence for significant £is -'^ci by adding higher-order terms in Fc and Icinteractions or nonlinearities, so we do not report these results to save space. nonlinearities 14 The four outcomes we focus on GDP the level of are: per capita, wliich is a good measure of long-rmi growth since around 1750 there were only ininor differences come per capita GDP, investment to money as a across comitries (Acemoglu, Johnson, wliich and Robinson, 2002); the the best measure of whether a society is into productive investments; the amoimt ratio of able to chamiel is of private credit as a percent of measure of finance provided tlirough the banking sector and trade in in- credit; GDP, and stock market capitalization as a percent of GDP, which provides a measure of equity finance. In our baseline regressions, we choose outcomes from by data availabihty and our desire to start the analysis have institutions data are aU independent For Ic — when the tliis choice is dictated comitries for wliich we states. we use two measures. Our base measure by government, averaged over 1985-95, from first the 1990s "protection against expropriation" is Political Risk Services. These data were used in economics by Knack and Keefer (1995), and are also the main measure used in Acemoglu, Johnson, and Robuison (2001). Pohtical Risk Services reports a value between and 10 each comitry and year, with for The second measure expropriation. is indicating the lowest protection against "constraint on the executive" from the Pohty dataset, capturing the degree of constraints on pohticians (Giirr, 1997).^*' Tliis constraints. In our measure ranges from main regressions, 1 to 7, IV and pohtically powerful ehtes where a higher score indicates greater we use the average of the values between 1990 and 2000 inclusive.^^ As already noted, we proxy for contracting institutions ism, Fc, using measures constructed (2003). 3.2 The The with the degree of legal formal- by Djankov, La Porta, Lopez-de-Silanes, and Slileifer details of these variables are discussed in Section 3.4 below. Empirical Strategy simplest strategy is to estimate the model in equation (10) using ordinary least squares (OLS) regression. There are two distinct problems with this strategy. First, both contractmg and property rights institutions are endogenous, so we verse causahty, or the effect of some omitted may be ^'^The latest version of Polity is IV is available ignore that year for the purposes of constructing the average. on the executive may be a down- on the web at http://www.cidcm.umd.edu/inscr/polity/ missing or the coding indicates an interregnum of some kind results using constraint re- characteristics (e.g., geography, reUgion, or other variables). Second, both variables are measured with error, so there ^^Where a year capturing in 1990, in 1970, 15 and We its (e.g., civil war), we also checked the robustness of our average value in 1950, 1960, and 1970. ward attenuation More important, bias. if contracting and property rights institutions are correlated, the effect of the institution that on to the other respond to the causal regressions will give results that —upward or dovimward bias is cor- So we would Uke to estimate equation possible. Two-Stage Least Squares (2SLS) with formahsm and property do not contracting and property rights institutions on economic effect of outcomes legal will load variable. Both of these concerns imply that OLS (10) using measured with greater error is rights. distinct and plausible instruments for These instruments should be correlated with the endogenous regressors but orthogonal to any other omitted characteristics (i.e., uncorre- lated with the outcomes of interest through any chaimel other than their effect via the endogenous regressors). ^° In this paper, two we pursue a multiple IV strategy The to identify the effects of interest. first-stages are: where Mc Fc = (5i-Lc + Ic = 62- Lc + r}2- Mc + Zl,-'y2 + U2c 77i-Mc + Z;-7i either the log potential mortality rate of is + iiie Eiuopean (11) settlers or log of the indigenous population density in 1500, and conceptually corresponds to the instrument for property rights institutions. dummy We explain these measures in Section 3.3 below. Lc for English legal origin (or equivalently, colony) and is is whether or not the comitry was a British the instrmnent for legal formahsm (i.e., contracting institutions). This discussed further in Section 3.4. If these instrmuents are vahd, the IV strategy wiU f3 is solve the endogeneity, the omitted variables bias, and the measiurement error problems, and can estimate the a and a we parameters consistently. is that legal origin may affect economic outcomes through channels other than example, La Porta et al., 1998, Glaeser and Shleifer, 2002, and Mahoney, 2001). In terms of our framework, in particular, equations (10) and (11), this would amount to Cov {Sc, Lc) 5^ 0, ^"A potential concern legal formalism (see, for whereas the 2SLS identifying assumption is that this covariance should be 0. Since the existing literature suggests that English legal origin should have a positive effect on the economic outcomes studied here, we expect that, if anything, Cov{ec,Lc) > 0, and in this case, the estimate of the impact of legal formalism on economic outcomes, a, will be biased upwards, and our results can be interpreted as potential upper bounds on the importance of legal formalism and contracting institutions. The results reported in Appendix Table A3 do not show any evidence of a major effect of legal origin on the outcome variables here. 16 Our and Population Density Settler Mortality 3.3 first instriiment for property rights institutions is settler mortality in coimtries that were colonized by European nations between 1500 and 1900.^^ Acemoglu, Johnson, and Robinson (2001) docmnented that European colonization strategies had radically different impUcations for economic development. Places prospered when Europeans set up institutions that protected private property rights and powerful established elites. first —or took over existing—extractive on politicians when Europeans institutions. the Europeans' colonization strategy? There were two key factors. was the disease environment facing Europeans. Wliere the disease environment was favorable political effective constraints In contrast, areas stagnated or grew only slowly What determined The and placed Eirropean settlements, they migrated in large mmibers and developed for and economic institutions very similar to, or even substantially better than, the contemporary mstitutions in Europe. These Canada, Austraha, or New settler colonies, such as the United States, Zealand, rapidly developed and maintained good institutions, with tight constraints on pohticians and elites, and secure property rights. In many other colonies, for example in sub-Saharan Africa, South Asia, and Central America, Europeans faced in some liigh or very liigh mortality rates (up to 50 percent mortality per year and settlement was not places) much more hkely feasible. In these areas, the colonizers were to develop extractive institutions, used mostly to exploit the native population for the benefit of European colonizers. After independence the beneficiaries of extraction changed, that had rapacious today. Based on an instrmnent more and the form of extraction has evolved over time, but countries rule mider colonialism typically have worse property rights institutions tliis reasoning, we will use potential European settler mortahty rates as for current institutions (see discussion). Acemoglu, Johnson, and Robinson, 2001, for ^^ The second determinant of Em-opean colonization strategy was initial indigenous pop- ^'We use the series constructed by Acemoglu, Johnson, and Robinson (2001) based primarily on Curtin (1989, 1998) and Gutierrez (1986). ^^ Malaria and yellow fever caused the majority of European deaths during the early colonization period. Although these diseeises were fatal to Europeans, they had much less effect on indigenous adults with aquired or inherited immunity. These diseases are therefore unlikely to be the reason why many countries in Africa and Asia are poor today. More generally, when we measure the effect of institutions correctly, there is no evidence that the large income differences between former colonies are due to geography, religion, or culture (for more details of this analysis, see Acemoglu, Johnson and Robinson, 2001, 2002). These results are robust to alternative variables (Acemoglu, Johnson measures of outcomes, institutions, and control and Robinson, 2001, and Easterly and Levine, 2003). 17 Illation density. Where population and put it this to was high, Europeans were more work in some form there is and pop- less likely Acemoglu, Johnson, did not settle. that, for coimtries colonized initial by European powers, a strong negative relationship between population density in 1500 and income per capita today. Tliis relationship is population density in 1500 had, and due to the fact that former colonies with greater have, worse property rights institutions. still density of indigenous population per square kilometer in 1500 alternative instrmnent. Because settler mortality but should have similar effects on property is in 1500 capture between the two measures rights, using these The therefore an appeahng and population density different sources of variation in practice (the correlation is Wliere likely settle themselves, when they to develop extractive institutions even and Robinson (2002) provide evidence of forced labor system. more ulation density was low, Europeans were likely to "capture" the local is 0.4), two instruments separately a good check on our results. Legal Origin and Legal Formalism 3.4 The fundamental idea in the Une of research of Vishny (1997, 1998) legal, is La that countries have distinct "legal origins" and these matter for economic, and financial outcomes.^"^ La Porta distinction between the two great legal traditions: part of the Porta, Lopez-de-Silanes, Shleifer, and UK or the British Empire, and "Civil (1998) draw the strongest "Common Law" countries that were et al. Law" countries where a French, German, or Scandinavian legal system has prevailed. Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003) offer two measures for the operation of contract enforcement through the legal system. These quantify the "formal" procedures associated with collecting on a boimced check and evicting a tenant for non- payment of rent. In both cases, the defendant has no justification payment. The underlying idea is and avoids volvmtary that a pure "neighbors" model, in which disputes are resolved informally by disinterested local third parties based on fairness criteria, would quickly rule in favor of the plaintiff (see Shapiro, 1981, and also EUickson, 1991, for real commimity works under such a model) More . legal how a formalism creates additional costs of enforcing the contract implied by the check or the tenancy agreement. Djankov et — (2003) measure the extent of these costs 23 legal —by surveying expert opinions formahsm See Glaeser and Shleifer (2002) on the origins of these distinct "legal famiUes". 18 al. of lawyers in an international network of law firms in 109 countries. They then construct indices that are comparable across comitries. A legal system more formal, is in their metric, if it involves professional judges lawyers, written rather than oral argtmients, the legal justification of claims decisions, the regulation of evidence, superior review of first-instance and creates delay formahsm of the Djankov collected legal formalism data on more it gives 9 or 10 Whether a country has a common law or civil al. good reasons to regard this as law systems on the country they colonized, law systems. civil We Enghsh legal origin law system other Eiiropean powers and exogenous an important determinant may be a choice, but for former —the British imposed common As we wiU by other European formahsm with legal origin see in greater detail below, formahsm than those that were colonized by less now have some version of civil law. The Samples and Descriptive Statistics Samples 4.1 Our is therefore instrument legal have former also confirm all our results wliile countries colonized sample of former European colonies. ^^ for We (2003). of legal formahsm. In general, the legal origin of a country colonies there are Bank subsequently more observations We data.^^ using the eviction measure directly from Djankov et 4 (2003) countries, but only for the "check" measure. we have aU the other necessary colonies for which colonies with al. raises the cost of adjudication (2003) study, the World et al. use this measure in our base regressions, as in the et in the resolution of disputes. As an extension powers have and judges' judgment, other "engagement formalities", and more required independent actions. Djankov also present evidence that a greater degree of legal and basic sample is that of former European colonies. For aU these comitries we have information on their legal origin and an estimate of their population density in 1500 (from Acemoglu et al., 2002). From smaUer subsets measures of institutions, and for measures of legal for potential settler of countries, we also have data mortahty rates (from Acemoglu formahsm (from Djankov et al., for various et al., 2001) 2003). ^''The extended check measure data were kindly provided by Simeon Djankov in a private communica- tion. ^^Djankov et al (2003) relate legal formalism to legal origin in the whole world sample and show that legal origin explains about 40 percent of the variation 19 in legal formalism. common In total, there are 42 Djankov et al. law "English legal origin" countries in the extended (2003) sample, of which almost base sample, we have 30 EngUsh legal origin when we all 1 In our lose additional observations civil law "French legal origin" (2003) sample, of wliich 35 are former al. summarizes our data. Column we have countries on which standard deviation or all European less) of down into former data. 1 reports Colmnn mean shows 2 values oiu* and standard deviations for data for the former colonies of cases, the values for former colonies are quite close (within half a former British colonies these and we colonies.^'' ^^ Eiuopean powers. In for coxintries, European Descriptive Statistics 4.2 Table are former use the settler mortality instrument. There are 47 countries in the extended Djarrkov et colonies. all the values for (i.e., common all countries. Column 3 reports mean values law countries), and columns 4 through 6 break British colonies with low, medium, and high settler mortahty (with the break points given by values that divide aU excolonies roughly into thirds). Colimms 7 through 10 provide parallel data for former colonies with French legal origin. The first two rows in Table 1 report our measures of property rights institutions: protection against expropriation and constraint on the executive. Rows 3 and 4 report the two measures of legal formahsm: the "check measiire" and the "eviction measure" Former colonies with lower on average, better property Rows GDP 5, 6, 7, settler mortahty and those with an Enghsh rights institutions and lower and 8 describe our four main dependent per capita (in PPP terms) in 1995.^* legal origin have, legal formalism. variables. Row 5 reports log of We can see a clear relationship between settler ^^The exceptions are: Britain and Ireland in Europe; Thailand, which was never colonized; and IsBahrain, and the United Arab Emirates, which were formerly part of the Ottoman Empire and administered by Britain under League of Nations mandates after the end of World War I. We do not treat former League of Nations mandate countries as excolonies for three reasons. First, European control was relatively short-lived and did not generally have major transformative effects on political institutions. Second, because these mandates were granted in the twentieth centiuy, European powers were already moving towards decolonization or at least minimal control, rather than the previous forms of colonial control for either settlement or extraction. Third, by the early twentieth century, advances in medicine meant that settler mortality was much more even across countries, so our data on this from earlier centuries do not allow construction of a valid instrument for the League of Nations rael, mandate countries. ^^The exceptions, in addition to France, are Belgium, Greece, Italy, Luxembourg, Malta, the Netherlands, Portugal, and Spain in Europe, and Jordan, Kuwait, and Turkey in the former Ottoman Empire. ^^These data are from the World Bank (2003). The results are robust to using GDP per capita data from other years or from the Summers-Heston data set. 20 mortality and income per capita: former colonies with higher settler mortality rates have substantially lower income per capita today. In addition, a comparison of colmnns 3 and 7 shows that English legal origin colonies have liigher average income per capita than French legal origin colonies. Row 6 prices reports om* data on the ratio of private investment to and averaged over the 1990s (from Heston in former colonies amount shown in row Investment ratios are higher et al. 2002). we use two standard measures of credit to the private sector in the 7; in current with lower settler mortahty rates, and liigher in former British colonies. In our basehne regressions total GDP, measured and stock market capitalization, economy shown in of financial development: the as a percent of row S.'^^ GDP in 1998, Former colonies with lower settler mortality rates and an English legal origin have, on average, liigher levels of credit to the private sector. For the size of the stock market, we use average stock market capitaUzation as a percent of GDP, provided by Beck et al. (2003a) ."'° much mortality rates and Enghsh legal origin have Below we official will also look at the barriers to entry for The and first is second measure Shleifer (2002). the is number businesses, particularly the La Porta, Lopezin Table of formal procedures needed to register a business. GDP on potential entrepreneurs. settler new There are two main measures (rows 9 and 10 the cost of registration. Tliis as a percentage of annual settler higher stock market capitaUzations. regulations required for registration as measured by Djankov, de-Silanes, 1). Former colonies with lower per capita. A is calculated by Djarikov et higher cost is al. The (2002) obviously a greater burden French legal origin former colonies and those with higher mortahty rates have more procedures required the overall cost of registering a business is for registering higher as a percent of a new business and GDP per capita. ^^Our base measure for banking system development is credit to the private sector, from the World (2003). This measure refers to financial resources provided to the private sector, such as through loans, purchases of nonequity securities, and trade credits and other accounts receivable, that establish a Bank claim for repayment. For some countries these claims include credit to public enterprises. In the choice and other mesisures, we are following the financial development literature; see, for example, Levine and King (1993), Rajan and Zingales (1998), and Levine (2003). •*"This measure corresponds to the total value of outstanding shares as a percent of GDP. The series averaged over 1990-95, which is appealing since it excludes the large run-up in global stock markets in of this (1997), Levine is the late 1990s (as well as the large, but mostly temporary, effects of the Asian financial 21 crisis). Results 5 First-Stage Results 5.1 Our regression analysis disentangles the effects of legal origin acting through legal formal- and colonization strategy acting through property ism, showing the Figure rights institutions. We start by first-stage relationships in Figures 2 tlirough 4. upper 2, left corner, plots on the y-axis the residuals from regressing average protection against expropriation on Enghsh legal origm, against on the x-axis the residuals from regressing log settler mortality on English legal origin. This is a visual representation of the strong first-stage relationship between the settler mortahty instrument rights institutions today. The upper right corner of Figure 2 out the effects of log settler mortahty, there is shows that, after partialing a substantially weaker but stiU positive between protection against expropriation and English relationsliip and property legal origin. As our regression analysis below will illustrate, however, the relationship between protection against expropriation Figure this on log legal origin is generally not robust. lower right, plots the residuals from regressing the check measure of legal 2, formahsm on origin and Enghsh Enghsh log settler mortahty, agauist the residuals from regressing settler mortahty. This shows that there measure of contracting institutions and is also a strong relationship Figure legal origin. 2, lower left, legal between depicts a weaker (and statisticaUy insignificant) relationship between the check measure of legal formahsm and log settler mortahty. Figure 3 shows a similar pattern using constraint on the executive and the eviction measure of legal formahsm as the dependent more variables. In this case the first stages are even separable: log settler mortahty has a strong effect institutions and no effect on this on measure of contracting this measure of property rights institutions, while Enghsh legal origin here affects contracting institutions but not property rights institutions. Figiire 4 presents the first-stage results for protection against expropriation and the check measxire of legal formalism with log population density in 1500 as the instrmnent for property rights institutions; these 5.2 show a 2. Univariate Regressions To provide a benchmark. Table hand similar pattern to that in Figure side variable. In Panel A 2 reports results just using legal the dependent variable 22 is log formahsm GDP as the right- per capita in 1995. Columns Column and 2 report OLS regressions using the check measure of 1 uses data from 1 109 comitries for wliich we have all measure of legal formaUsm; column 2 Umits by European powers and There OLS for GDP legal formalism. data and the check to just the 65 countries that were colonized tliis which we have data on formahsm. legal a significant coefficient on the check measure of legal formahsm in the basic is regression —a one standard deviation increase over a 30 percent dechne in malism with GDP legal origin, there is per capita today. a strong associated with is Wlien we instrument for legal for- stage (R^=0.58), but in the second stage fii'st the coefficient becomes msignificant, though formaUsm in legal remains quantitatively large; the point it estimate of -0.18 in colmnn 3 implies that a one standard deviation increase in legal mahsm 5, and will GDP reduce log per capita by 0.20 of a standard deviation. '^^ In colmnns the eviction measure of legal 6, specifications, and the 2, the first In Panel B, the dependent variable is columns a significant negative and 1 level (but 2. formahsm size of the effect is the IV specifications of Table There The is coefficient is not significant in either OLS about 1/4 of that imphed by colmnn stage R^ is in the 3. 4, or IV In all range 0.56-0.6L"^^ the average investment- GDP ratio in the 1990s. on formahsm legal column 3 coefficient of -1.77 in is in the OLS specifications of significant at the 10 percent not at 5 percent) and imphes that a one standard deviation increase in legal formaUsm would reduce the investment-GDP columns for- 4, smaU. The 5, and 6, ratio by 0.3 of a standard deviation. the eviction measure of legal formaUsm size of the effect implied the effect suggested by colmnn 3. by the IV Since the first coefficient in is both insignificant column 6 is In and about 1/2 of stages in this and subsequent panels are almost identical to those in Panel A, we do not report them to save space. Panels C C and D show significant the dependent variable coefficient is -0.14 in it is results for both measures of legal formaUsm. In Panel credit to the private sector as a percent of colmnn 2 (OLS) and -0.13 in column 3 implies that a one standard deviation increase in legal standard deviation faU in the credit to one standard deviation change GDP ratio. The GDP m (IV). Tliis is 1998. a large effect formaUsm causes about a results for the eviction The half measure measure of legal formalism in the former colonies change of 0.22, which is about 20 percent of the standard deviation of log GDP, which is approximately 1 for former colonies. ^^The difference between the results using the check and the eviction measures is not due to differences in the samples. When we run regressions with the check measure restricting the sample only to those countries that have data on the eviction measure, we get very similar results to those using the full check measure sample. •^'A sample is in the check 1.24, thus the estimate implies a 23 and 6 of Panel B, are similar (though of legal formalism, in coliumis 4, 5, slightly less significant). Panel D a percent of reports the GDP same set of specifications now with stock market capitaUzation as (averaged over 1990-95) as the dependent variable. Both legal formahsm measures are significant determinants of stock market capitalization, and the measure (columns 2 and results are quite similar for the check is about twice the A 6). OLS size of the 3) coefficient for the eviction , OLS and IV while the IV coefficient measvue (columns 5 and one standard deviation increase in either legal formahsm measure imphes a half standard deviation dechne in stock market capitahzation. Table 3 reports parallel univariate specifications using our measures for property rights and constraint on the executive, as the institutions, protection against expropriation dependent variables, and with the instrimients. In if example columns TMs mortahty and log population density we only panels of Table 3 we have data on both regressions in all settler 2, 3, and 4, is instriunents in in- 1500 as include a country in the excolonies — this ensures that the sample size, for the same, so the coefficients are directly comparable. and restriction does not affect the results, in subsequent tables, when we use log population density as instrument, we will report results with the fuU set of coimtries for which we have data on that instrmnent. Panel A shows a strong first stage from log settler mortality to property rights insti- an R^ of 0.23 in column tutions, with specifications, so to save space The 3. we do not repeat them the IV coefficient using log settler mortahty This is similar to the results in that the is OLS almost identical in aU in the other panels. In approximately double the is OLS aU panels coefficient. Acemoglu, Johnson, and Robinson (2001), and suggests estimates are significantly biased downwards because of attenuation. This not surprising: since the available measmes of property rights institutions correspond quite poorly to the concepts ror, creating tliis we would downward attenuation like to measme, there The bias. is a form of measmement er- instrimiental variables strategy corrects for type of attenuation bias. The first-stage relationship between log population density in 1500 4. first-stage results are and protection against expropriation For this instrmnent, the IV coefficient a bit weaker, with an R^ of 0.17 in column also typically twice the is Property rights institutions have a large ification of is effect OLS coefficient. on income per capita. In the IV spec- Panel A, the coefficients of 1.05 and 1.07 in colmnns 3 and 4 imply that a one standard deviation improvement in this measiue of institutions leads to about a 24 1.5 standard deviation increase in cohimn executive, in Property ratio, The estimated per capita. effect of constraint on the 7, is similar. riglits also GDP credit to GDP have a large effect on the investment to and on stock market development. GDP ratio, on the private For example, in colimui 3 of Panel B, the coefficient of 5.5 impUes that a one standard deviation improvement in property rights causes a 1.1 standard deviation increase in the investment-GDP ratio. The estimates in colmnn 3 of Panels improvement GDP credit to The in property rights ratio and a 1.2 results using constraint C and D, imply that a one standard deviation would cause a 1.5 standard deviation increase in stock market capitalization. on the executive are qualitatively and quantitatively Taken together. Tables 2 and 3 show strong on GDP standard deviation increase in the effects of legal formalism) 6 of contracting institutions (proxied on credit and stock market development, with more limited GDP per capita and the investment-GDP results property rights institutions per capita, investment, credit, and stock market development both in univariate OLS and IV regressions. They also show strong effects by sunilar. ratio. We next tiurn to investigating effects how on these change when the two sets of variables are included simultaneously. Contracting vs. Property Rights Institutions 6.1 Main Results Tables 4, 5, 6 and 7 report our main results. In aU fovu tables, colmuns 1 tlirough 4 use protection against expropriation as the measiue of property rights institutions and columns 5 through 8 use constraint on the executive. Columns measiue of legal formalism, measure of legal and columns 3, 4, 7, 1, 2, 5, and 8 include and 6 use the check results with the eviction formaUsm. For each specification we present a pair of odd-numbered colimms and IV in even-numbered columns. results: OLS in For the IV specifications, the top panel reports the second stage and the other panels report the first stages for the measure of property rights institutions and the measiue of legal formahsm in that order. All four tables use log settler mortality rates as the instrument for property rights institutions, and Tables A4, A5, A6 and A7 in the Appendix have an identical structure, but use log population density in 1500 as the instrument. ^'^ ^^In addition, Appendix Table A2 reports regressions in which the dependent variable is economic growth between 1970 and 1995. The results are similar to Table 4, where log GDP per capita is the 25 The first-stage regressions in all eight tables are quite similar patterns shown in Figures 2 tluough using either measiue, there is between -0.65 and -0.92, legal origin is positive and is 4. In the first and consistent with the stage for property rights institutions, a strong robust effect of log settler mortality and —the coefficient typically four times larger than the standard error. English significant in the specification of coliimn 2, but in colunm 4 it When we use constraint on the executive as the measure of property rights institutions (columns 6 and is significant only 8), in when we use log population density as the instrument. the sign on English legal origin is insignificant and usually negative. Overall, the which coimtries were colonized, but not who colonized them, is way a robust determinant of property rights institutions. In the first legal origin Enghsh stage for legal formalism, using either the check or eviction measure, has a strong robust negative effect. In contrast, log settler mortahty and log population density in 1500 are not significant in any specification. Wlio colonized, but not the way in which countries were colonized, appears to affect the extent of formahsm in legal procedures. These strong first-stage results enable us to disentangle the effect of property rights and contracting institutions. In expropriation 0.95. rights is wiU lead to a results in column 2 of Table 4, the coefficient on protection against This impUes that a one standard deviation improvement in property 1.4 standard deviation increase in per capita. column 6 imply that a one standard deviation increase executive causes a two standard deviation increase in on coefficient legal columns 2 and higher GDP GDP 4, it formalism is often not significant, has the "wrong" sign significant, 2 of that table). This change in the implies that in OLS in constraint on the per capita. In contrast, the and when it is significant, e.g., —coimtries with more formalism actually have per capita. Note the change from Table formahsm, while not always GDP Similarly, the 2, where the was negative and quite imphed effect of legal coefficient on large (e.g., -0.18 in legal column formahsm on long-run growth or in regressions that do not control for property rights institutions, the importance of contracting institutions is exaggerated because they capture some of the differences in protection of property rights. Appendix Table A4, which uses mortality, has in Table 4. outcome of log population density in 1500 in place of log settler between 6 and 9 more observations, and shows results consistent with those For example, the estimated coefficient for protection against expropriation in interest. 26 column it is 2 0.97 1.09 is also similar in the compared with 0.95 compared with 0.18) (s.e. 0.22), (s.e. two GDP 0.25) in Table 4. (s.e. effect of The colmnn 4 first-stage pattern is property rights institutions on the investment According to the estimate in colmim ratio. 0.16) in Table 4, while in tables. Table 5 shows a robust positive to 1.53 (s.e. 2, a one standard deviation strength- ening of property rights causes a 0.95 standard deviation increase in the investment to GDP ratio. In contrast, there is no significant efTect of legal formalism on the investmentGDP ratio in any specification. The coefficient on legal formahsm is negative and small in colmnns 5 tlirough 8, Appendix Table A5 are and actually positive in colmnns 1 through The 4. results in siinilar. Table 6 shows a strong effect of property rights institutions on credit to the private The sector. coefficient of 0.28 in column 2 implies that a one standard deviation increase in protection against expropriation causes a 1.2 standard deviation increase in the credit to GDP ratio. improvement The coefficient of 0.27 in in constraint cohunn 6 imphes that a one standard deviation on the executive causes a 1.61 standard deviation increase in credit. Once we formahsm control for the effect of property rights institutions with om' is no longer a specifications of Table 6. 2 and 4. -0.08 in When significant determinant of credit The coefficient negative, it is it is formalism m column is strategy, legal insignificant in aU the IV actually positive in columns e.g., the coefficient of -0.13 in Table 2). parallel results using log population density in 1500 as The shghtly lower, but the effects are coefficient of 0.25 it is highly insignificant and small, instrvunent in place of settler mortahty. now legal column 6 (down from the univariate estimate of Appendix Table A6 reports are on — IV coefficients still large on the property and Mghly rights an measures significant (e.g., the 2 impfies a one standard deviation increase in protection against expropriation causes about a one standard deviation increase in credit to the private sector). and The coefficients on the legal formahsm measures are consistently small insignificant. The pattern ferent. The in Table 7, effect of where we look at stock market capitahzation, is slightly dif- a one standard deviation increase in protection agauist expropriation (using the coefficient of 0.21 in in stock market capitahzation. column The 2) is about 0.8 of a standard deviation increase coefficient of 0.2 in column 6 implies that a one standard deviation unprovement in constraint on the executive would raise stock market 27 capitalization by one standard deviation. With protection against expropriation regression, the coefficient on legal formalism is negative but not significant. In contrast, with constraint on the executive as the measure of property rights, there significant negative effect of legal the estimate in colrmin ple, increase in legal 6, (s.e. a statistically For exam- 0.07) implies that a one standard deviation formahsm would reduce stock market capitalization deviation. So the effect of contracting institutions, even smaller than that of property rights institutions. The when by 0.2 of a standard significant, is substantially Appendix Table A7, with results in log population density in 1500 as the instrmnent, are similar. There is property rights institutions, and a smaller effect of legal formahsm that when competing is formalism on stock market development. -0.16 in the a strong effect of is significant only against constraint on the executive. Overall, the results in this section suggest that property rights institutions have a first- order effect on income per capita, the ratio of investment to stock market development. In contrast, legal GDP, formahsm appears stock market development; for the other outcomes, the effect from zero. Moreover, for all variables, and attempts to have The an effect only on not significantly different the effect of legal formahsm smaller than the effect of property rights institutions. this pattern is robust is the level of credit, and is quantitatively rest of this section much shows that to develop a potential explanation."^^ Alternative Samples 6.2 Table 8 reruns owe basic regressions using alternative samples. To save space, the first-stage regressions in identical. All full just in Panel A; the first stages for Panel B it reports are almost columns use protection against expropriation, instrumented with log settler mortaUty. Results using constraint on the executive as the measure of property rights or log population density in 1500 as the instrument are very similar, but not reported to Appendix Table A3 reports which we instrument through 6 and log population density in columns 7 through 12), but legal origin is entered directly. The specifications are useful since they do not restrict the effect of legal origin to work solely through legal formalism, and are informative on whether legal origin has a direct effect on the economic outcomes studied here. The results are similar to our baseline results ^''in addition, for property rights (using settler mortality in in Tables 4 on GDP through 7, but now English partial reduced- form specifications in columns legal origin 1 has a negative (i.e., wrong signed) and significant effect per capita and does not appear to be a significant determinant of stock market capitalization. We have also experimented with other measures of credit and financial development, including the same measures as in Tables 6 and 7 but calculated over different periods, measures of overall size of financial system (sum of bank credit and stock market capitalization), M2 over GDP, and liquid liabilities over GDP. The results are similar to those in Tables 6 and 28 7, and are available upon request. save space. 1 and British colonies) and In columns 5, we columns 2 and 6 to just French in drop legal formalism, and examine whether there within each legal "family" . There a strong is is first 1 and and a much higher R^ (e.g., common effect of much property rights institutions on GDP among common law Columns 3 is vs. 0.07, for French comparing colmnns is a large significant effect of effect of approximately the same within "legal families" as The results for the investment to and stock market development there is GDP a stronger effect coimtries and 7 drop the fom: "neo-Europes," closest geographic conditions to Zealand, and the law countries than French legal origin countries). The across aU excolonies (compare with Table 4). ratio are similar, while for credit property rights institutions per capita for both legal systems (with a larger, less precisely estimated, coefficient for it is Here we law comitries can be explained by the colonization In the second stage there property rights on income per capita former of the considerable variation in strategy of European powers. but (i.e., stage for log settler mortality in both R^ of 0.30 2 of Panel A). This result suggests that institutions witliin the set of an countries legal origin countries. common specifications, with a slightly larger coefficient for legal origin comitries common law sample to just restrict the USA. the richest former colonies with the i.e., Western Europe (Crosby, 1972): Austraha, Canada, Tliis is useful to show that the New results are not driven simply by the contrast between these four comitries and other former European colonies. Without these countries, the coefficient for protection against expropriation goes GDP up slightly in the per capita, investment, and private credit regressions and increases by about 50% market capitahzation. Legal formalism shows the same pattern as before; it is for stock significant only for stock for log GDP market capitalization. Legal formahsm is also almost significant per capita, but with the wrong sign. Columns 4 and 8 in Panel A report the regressions just for countries above GDP per capita and investment-GDP ratio median world income, to estabUsh that the results are not driven simply by the comparison of rich and poor countries. The results are close to those in Table 4. Although there are no GDP significant outhers in the ratio data used here, there are outcomes. Columns 4 and 8 in Panel GDP some major outhers B per capita and investment to in the financial development drop outhers from the credit and stock market capitalization regressions respectively. For credit, the outhers are Malaysia, South vVfrica, 29 and the United Africa. The and States, market they are Malaysia, Singapore, and South for the stock change results for property rights institutions little; the coefficient 0.28 (column 2 in Table 6) to 0.24 (column 4, Panel B, Table 8); 2 in Table 7) to 0.14 (column declines, so the effect remains highly legal formalism is no longer a 8, Panel B, Table 8). from and from 0.21 (colunm In both cases the standard error also significant. Notably, significant falls once these outliers are dropped, determinant of stock market development. Additional Control Variables 6.3 Table 9 includes with three important control variables from the fiterature on long-rrm groAvth 2, 5, 8 and financial development.'^^ and 11 control for latitude, inflation over 1970-97. In Columns 1, 4, and coltmms 7 and 10 control 3, 6, for refigion; colmims 9 and 12 control for log average aU columns, we include these control variables in the first-stage regressions but do not report their coefficients to save space; the first-stage results are very similar to those in earher tables. To save space, Table 9 only reports results using protection against expropriation to measrue property rights and log settler mortahty as the instrument. However, the results are similar if we use constraint on the executive to measure property rights or log population density as the instrmnent. A number development al. of recent papers have taken religion seriously as a determinant of financial (e.g., Stulz and Williamson, 2003). We (1999) with the percent of the population that add to "other"; these shares is 100. In our specification, base category. Including rehgion in this form hardly against expropriation. insignificant in most notable change formalism is The check measure column is no longer 1, use the measrue from La Porta et in the results for stock In we treat aff'ects MusUm and CathoUc as the omitted the coefficient on protection of legal formalism remains positive, small, and negative, smaU, and significant. Catholic, Protestant, insignificant in columns 4 and market capitaUzation: all cases, in column 7. and The 10, legal the p- values of the F-test for rehgion variables indicate that these variables theinselves are not significant.'^'^ '^'^A fourth potential control is the level of human with our measures of property rights institutions capital. Unfortunately, this —only countries with relatively is highly correlated good institutions have encouraged the majority of the population to accumulate human capital. It is therefore difficult to disentangle the separate effects of human capital, property rights, and the legal system without additional instruments. When we include this measure as an exogenous regressor and use the protection against expropriation measure, we find similar results to those in our baseline estimates. However, when we use the constraint on the executive measure, there is too much multi-collinearity. ^^ Stulz and Williamson (2003) argue that religion should be coded differently, with a "1" for the most popular religion of a country and a zero for all other religions, no matter how large. Using this alternative 30 It might also be useful to control for latitude, because coimtries that are closer to the equator are often argued to be poorer, perhaps because of the hotter cUmate or because they are exposed to more virulent diseases. The results in Table 9 confirm findings we have reported in other work (Acemoglu, Johnson, and Robinson, 2001, 2002) control for institutions there is These tables also show there no significant effect for geography on income per capita. significant effect of geography on investment and on no is measures of financial development. The pattern of results and is legal formahsm no longer a Log essentially is is for property rights institutions michanged, except the check measiue of legal formahsm significant determinant of stock inflation —once we market capitahzation (see column 11). included as the most plausible macroeconomic determinant of financial development. Countries with high inflation are often thought to have less developed finan- systems. However, the addition of this variable does not reduce the effects of property cial rights institutions, or increase the effects of legal formahsm, on growth, investment, and financial development. Controlling for 6.4 GDP Per Capita We have shown robust effects of on investment, and stock market development. These credit, GDP per capita, property rights institutions on as well as results are not informative, however, on the question of whether the effect of institutions on investment, credit and stock market development that is is, is direct or indirect. One possibihty is that the effect present-day institutions affect these outcomes directly. The polar is direct, alternative a purely indirect effect whereby current and past institutions have determined per capita, and GDP per capita today is the main influence on investment and GDP financial development. In practice, both direct and indirect effects are hkely to be present. Table 10 briefly investigates this issue. right Since log hand (2001), include log side of regressions for the investment to we do not have GDP We plausible instruments for per capita as exogenous. tliis As explained GDP GDP ratio per capita in 1995 on the and financial development. GDP, we have no in choice but to treat Acemoglu, Johnson, and Robinson procedure generally leads to an upward bias in the estimated coefficient for the variable treated as exogenous, and a corresponding downward bias for the effect of property rights institutions. Here the problem coding does not significantly affect our main results. 31 is much worse, since log GDP per capita and property rights institutions are very strongly correlated precisely because these institutions have a causal effect 4) show that they are a on economic development major determinant of Not surprisingly given these concerns, the collinearity, and log GDP GDP — in fact, our results Table (e.g., per capita. results in Table 10 show a high degree of per capita, property rights institutions, and legal formalism are not individually significant. However, joint significance tests reported in the table show that GDP per capita and property rights institutions are almost always jointly highly significant, while GDP per capita and legal formahsm are never jointly significant. These results therefore confirm the overall patterns shown above but not contracting institutions have significant effects —property rights institutions on investment and financial devel- opment. Unfortunately, however, they do not enable us to ascertain whether the property rights institutions on investment and financial development through affecting GDP is effect of direct or indirect per capita. This remains an interesting area for future research. Corroborating Evidence 6.5 In terms of long-rim economic and financial outcomes, our results suggest that the effects of legal case. formahsm Both are quite hmited. Tables 11 tables use data from the Djankov and 12 further explore why (with an extended sample provided by Djankov, 2003). measures of legal formahsm; the fuU first may be the (2002) "Regulation of Entry" study et al. in pairs, for the protection against expropriation this We show OLS and IV results, measure of property rights and both IV stage results are provided for specifications. Table 11 reports regressions in wlrich the log of the number of procedures for registering a new business is the dependent variable. This is a good measure of the formal steps needed to legally start a new firm which, as shown by Djankov et (2002), al. correlated with other measiues of formal requirements imposed on firms. results in this table indicates that have significant effects both legal formahsm and property is highly The pattern of rights institutions on the nmnber of procedures. However, in Table 12 the results are quite actual cost of registering a new different. Here the dependent variable business, as a percent of Property rights institutions again have a large effect GDP is the per capita (per annum). on these costs: a one standard deviation increase in protection against expropriation reduces the cost of registering by 0.65 of a standard deviation, which is quantitatively similar to 32 its effect on the number of procedures. in the IV But no robust in this table there is The specifications. coefficient of 0.03 in coefficient column on legal significant effect for legal formaUsm is formaUsm consistently small; e.g., the 2 implies that a one standard deviation increase in legal formalism causes only a 0.03 of a standard deviation increase in the cost of registration. These fuidings seem to corroborate our main a significant effect have more limited Contracting institutions have residts. on the formal procedures that regulate business. But they appear to on the major economic outcomes effects —long-rrm growth, investment, the overall size of the financial system, and the real costs of regulation that businesses face. Interpretation 6.6 Our empirical investigation reveals an interesting pattern: contracting institutions and approximated by legal formafism, have some legal rules, as (the use of equity versus debt contracts) lar, effect on the form of finance and the form of business regulation (in particu- the nimaber of procedures necessary to open a business). But they have limited or no effects on major economic outcomes, including long-rim growth (cmrent income the investment to economy (also GDP ratio, and the overall amoimt of on medium-rrm growth). Property rights degree to which the government, pohticians, and levels), financial intermediation in the institutions, wliich determine the constrained in their relation- efites are ships with the rest of the society, on the other hand, matter significantly for all these outcomes. Although a precise explanation for this pattern of results is not possible with our current level of knowledge, our simple reduced-form model suggests a potential explanation: legal rules and procedures primarily affect the contracting relationship private individuals, in particular between lenders and producers. As long institutions are not extremely dysfunctional, individuals can avoid effects of bad legal rules arrangements. One way we may of contracting around these rules see are cheaper to enforce. institutions as these legal most of the adverse by changing the terms of their contracts or entering into informal is to change the form of financial intermediation, so in places with a high degree of legal formafism institutions), between more debt rather than equity, (i.e., perhaps because debt contracts Once these adjustments have been made, the on investments and growth is relatively Umited. 33 worse contracting effect of contracting When it comes to arrangements regulating property tween the state and individuals, i.e., rights and the relationsliip be- property rights institutions, the option to engage in ex ante contracts to avoid ex post distortions is not available. Individuals caimot write contracts with the state to constrain futme actions by the state and ehtes controlling the more important Therefore, property rights institutions have a state. outcomes than do contracting An alternative, and effect on economic institutions. related, explanation would be based on the recent paper by Djankov, Glaeser, La Porta, Lopez-de-SUanes, and Shleifer (2003), where they introduce the concept of an "institutions possibihties frontier," representmg the trade-off between and centrahzed disorder as moving an economy along a given less disorder). Changes on the other hand, would induce with worse property rights institutions, a society has to both disorder and centrahzed property rights institutions of legal rules formahsm may correspond to frontier (e.g., greater in property rights institutions, shifts of the frontier, so that suffer higher costs of we can think control, hi terms of their framework, may As a control. have much more pronounced result, differences in effects on investment and output than do differences in contractmg institutions. At present, these interpretations are merely conjectmes, however, and more detailed work is necessary to investigate them in greater detail and also to determine the precise channels through which property rights institutions matter so much for finance, invest- ment, and long-run economic performance. Conclusion 7 There is now considerable evidence that "institutions" are of first-order importance for economic and financial outcomes. Douglass North, sure the importance of "contracting institutions" izens, rulers. and "property rights institutions" , , for example, emphasized in equal mea- enabling private contracts between cit- protecting the property rights of citizens against Despite the importance of these questions for the study of long-run economic per- formance, there has been relatively httle work investigating which types of institutions matter more and for which economic outcomes. This paper offers a step in that direction. We proxied contracting institutions with the legal formalism measure of Djaiikov et and property rights institutions with measures of protection ment expropriation and constraints on govermnent power. 34 We al., for citizens against govern- used a multiple uistrumen- tal variables strategy in the institutions legal rules (i.e., institutions are sample of former European colonies, where both contracting and procedures shaped by the experience of colonialism. In first-stage relationships between and property for private contracts) and legal origin rights this sample, there are strong legal formalism on the one hand, and between colonization strategy and property rights institutions on the other. Using multiple instrmnental variables strategy, stitutions have a development, and the wliile contracting institutions amomit total Our conjectme appear to regulation, but have more affect the form of financial interme- limited effects on growth, investment, that individuals can structure contracts to reduce the adverse is institutions, for Because of these adjustments, the usual effect of rules governing contracts on investment and growth limited; exceptions occur only when ef- example, by changing the form of intermediation to reduce the cost of providing outside finance to firms. trast, financial of credit in the economy.'^" from contractmg fects that property rights in- major influence on long-rmi economic growth, investment, and and the form of diation we found robust evidence tliis may be relatively contracting institutions are extremely bad. In con- because enforceable contracts between the state and individuals are not possible, property rights institutions constraining arbitrary behavior and expropriation by the state and elites controlling We view paper as a tliis first step, and much more empirical and theoretical work Oiu* explanation for the patterns in the data needed. is the state have more important effects on economic outcomes. and detailed work using both macro and micro data is is no more than a conjecture, necessary to investigate whether individuals are indeed avoiding the costs of distortionary legal rules by changing the terms of their contracts and the form rights institutions black box growth? —how Is it of their financial relations. Moreover, the effects of property on econoinic outcomes, though highly robust, axe something of a stiU exactly do property rights institutions affect investment, credit, and because the government and poUtically powerful groups are expropriatmg the incomes of other groups? Or is it because they are blocking entry by producers? Or because they are creating a non-level playing field new groups and and a liigh degree of inequality? ^ ' Our results also suggest that there may be a mispecification in growth regressions that use legal origin as an instrument for financial development (see Levine, 2003, for a discussion of this Legal origin is control for property rights institutions. in these common approach). correlated with total financial intermediation or credit, but this effect disappears once we Consequently, the instrumented financial development variable growth regressions could be capturing the effect of 35 property rights institutions on growth. References 8 Acemoglu, Daron (2003a) "Why Not a Commitment and Politics," Political Coase Theorem? Social Conflict, Journal of Comparative Economics, forthcoming. Acemoglu, Daron (2003b) "State Power and Economic Development" MIT mimeo. Acemoglu, Daron, Simon Johnson, and James A. Robinson (2001) "The An Colonial Origins of Comparative Development: Economic Review, December, Empirical Investigation," American 91, 5, 1369-1401. Acemoglu, Daron, Simon Johnson, and James A. Robinson (2002) of Fortmie: Geography and Institutions in the Making of the "Reversal Modern World Income Distribution," Quarterly Journal of Economics, 117, November, 1231-1294. Acemoglu, Daron, Simon Johnson, Jcunes A. Robinson, and Yunyong Thaicharoen (2003) "Institutional Causes, Macroeconomic Symptoms: VolatiUty, and Growth," Journal of Monetary Economics, Crises, 50(1), January, 49-123. Barro, Robert (1997) Determinants of Economic Growth, MIT University Press, Cambridge. Beck, Thorsten, Asli DemirgiiQ-Kunt, and Ross Levine (2001) "The Financial Structure Database," chapter 2 in Financial Structure and Economic Growth, edited by Asli Demirgiig-Kimt and Ross Levine, MIT Press, Cambridge, MA, 17-80. Beck, Thorsten and Ross Levine (2002) "Industry Growth and Capital tion: Alloca- Does Having a Market-Based or Bank-Based System Matter?" Journal of Financial Economics, 64, 147-180. Beck, Thorsten, Asli Demirgiig-Kunt, and Ross Levine (2003a) "Law, Endowments, and Finance," Journal of Financial Economics, forthcoming. Beck, Thorsten, Asli Demirgiig-Kunt, and Ross Levine (2003b) "Law and Finance: Why Does Legal Origin Matter?" Journal of Comparative Economics, forth- coming. Berkowitz, Daniel, Katharina Pistor, Jean-Francois Richard (2003) "Economic Development, Legality, and the Transplant Effect," European Economic Review, 47(1), 165-195. Coase, Ronald (1937) "The Nature of the Firm," Economica, Coase (1960) "The Problem of Social Cost," Journal of 1-44. 36 4(16), 386-405. Law and Economics, III, Crosby, Alfred (1972) The Columbian Exchange: Biological and Cultural Conse- quences of 1492, Greenwood Press, Westport, Connecticut. Curtin, Philip D. (1989) Death by Migration: Europe's Encounter with the Tropical World m the 19th Century, Ciirtin, Philip Cambridge University Cambridge University Dawson, John (1960) A De New York. D. (1998) Disease and Empire: The Health the Conquest of Africa, bridge, Press, Press, New of European Troops in York. History of Lay Judges, Harvard University Press, Cam- MA. Long, J. Bradford and Andrei Shleifer (1993) "Princes and Merchants: European City Growtli before the Industrial Revolution," Journal of Law and Economics, 36, October, 671-702. Demirgiig-Kunt, Asli ajid Vojislav Maskimovic (2002) "Funding Growth Bank-Based and Market-Based Financial Systems: in Evidence from Firm-Level Data," Journal of Financial Economics, 65(3), 337-363. Djankov, Simeon (2003) Personal Communication. Djankov, Simeon, Edwcird Glaeser, Rafael La Porta, Florencio Lopez-deSilanes, and Andrei Shleifer (2003) "The New Comparative Economics," Journal of Comparative Economics, forthcoming. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer (2002) "The Regulation of Entry," Quarterly Journal of Economics, 117(1), 1-37. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silcines, and Andrei Shleifer (2003) "Courts," QuaHerly Journal of Economics, 118, William and Ross Levine (2003) Easterly, Endowments hifluence 2, May, 453-517. "Tropics, Germs, and Crops: Economic Development," Journal of Monetary Economics, How 50(1), January, 3-39. EUickson, Robert C. (1991) Order Without Law: Harvard University Press, Cambridge, Glaeser, Edw^ard and Andrei How Neighbors Settle Disputes, MA. Shleifer (2002). "Legal Origins," Quarterly Journal of Economics, November, 117, 1193-1230. Grossman, Seinford and Oliver Hart (1986) "The Costs and ership: A Theory of Vertical and Lateral Ownership," Journal of 691-719. 37 Benefits of Political Own- Economy, 94, Greif, Avner (1989) "Reputation, and Coalitions in Medieval Trade: Evidence from the Maglxribi Traders," Journal of Economic History, XLIX, Gurr, Ted Robert (1997) "PoUty 1986," UnpubUshed paper. University II: 4, Political Structures December, 857-882. and Regime Change, 1800- of Colorado, Boulder. Gutierrez, Hector (1986) "La Mortalite des Eveques Latino- Americains aux XVIIe et XVIII Siecles," Hcirt, Oliver Annales de Demographic Historique, 29-39. and John Moore (1990) "Property Rights and the Nature of the Firm," Journal of Political Economy, 98, 1119-58. Hcirt, Oliver (1995) Firms, Contracts, and Financial Structure, Oxford University Press, Oxford. Heston, AlEin, Robert Summers, and Bettina Aten (2002) "Perm World Table , Version 6.1," Center for International Comparisons at the University of Pennsylvania (CICUP), October, downloaded in March 2003 from http://webhost.bridgew.edu/baten/ Johnson, Simon, John McMillan, and Christopher Woodruff (2002) "Property Rights and Finance," American Economic Review, December, 92, 1335-1356. Jones, Eric L. (1981) The European Miracle: Environments, Economies, and Geopolitics in the History of Europe and Asia, Cambridge University Press, Knack, Stephen and Philip Keefer (1995) "Institutions New York. and Economic Perfor- mance: Cross-Country Tests using Alternative Measures," Economics and Politics, 7, 207-227. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, Vishny (1997) and Robert "Legal Determinants of External Finance," Journal of Finance 52, 1131- 1150. La Porta, W. Vishny and Robert (1998) "Law and Finance," Journal of Political Economy, 106, 1113-1155. La Porta, W. Vishny Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, Rafael, Florencio Lopez-de-SilcUies, Andrei Shleifer, and Robert (1999) "The QuaUty of Government," Journal of Law, Economics and Organization, V15, Nl, 222-279. Levine, Ross (1997) "Financial Development and Economic Growth: Views and Agenda," Journal of Economic Literature, June, 35, 688-726. Levine, Ross (2003) "Finance and Growth: Theory, Evidence, and Mechanisms," Handbook of Economic Growth, forthcoming, Marcli draft. Levine, Ross cuid Robert G. King (1993) "Finance and Growth: Schumpeter 38 Might Be Right," Quarterly Journal of Economics, 153, August, 717-738. Levine, Ross, Norman Loayza, and Thorsten Beck (2000) "Financial Intemie- diation and Growth: CausaUty and Causes," Journal of Moneta.ry Economics, 46, 31-77. McEvedy, Colin and Richard Jones (1978) Facts on File, New Atlas of World Population History, York. Mahoney, Paul G. (2001) "The Common Law and Economic Growth: Hayek be Right," Journal of Legal Studies, MerrymEin, John (1985) The Miglit 30, 503-525. Civil Law Tradition, Stanford University Press, Stan- CA. ford, North, Douglass C. (1981) Structure and Change Norton & Co., Olson, New in Economic History, W.W. York. Mancur (2000) Power and Prosperity: Outgrowing Communist and Capi- Dictatorships, Oxford University Press, Oxford. talist Persson, Torsten and Guido Tabellini (1999) "Tlie Size and Scope of Government: Comparative Pohtics with Rational Pohticians, 1998 Alfred Marshall Lecture," European Economic Review, 43, 699-735. Rajan, Raghurani and Luigi Zingales (1998) "Financial Dependence and Growth," American Economic Review 88, 3, 559-586. Rajan, Raghurani cind Luigi Zingales (2003) "The Great Reversals: The Pohtics of Financial Development in the 20th Century," Journal of Financial Economics, 69(1), 5-50. Shapiro, Mcirtin (1981) Courts, The University of Cloicago Press, Cliicago. Stulz, Rene and Rohan Willicunson (2003) "Culture, Opermess, and Finance," Journal of Financial Economics, forthcoming. Willicunson, Oliver (1975) Markets and Hierarchies: Analysis and Antitrust Implications, New York, NY, Free Press. Willicunson, Oliver (1985) The Economic Institutions of Capitalism, NY, New York, Free Press. World Bank (2003) World Development Indicators, 39 at publications.worldbank.org/WDI/. Table 1 Descriptive Statistics (1) (2) Whole World Sample (3) (4) (5) (6) English English English excolonies with excolonies with excolonies with Excolonies English low settler Sample excolonies mortality medium of Expropnation Average Investment-GDP ratio Credit to the Private Sector Stock Market Capital^ation Log Number Procedures a New Business Cost of Registering a New Business as Percent of GDP per capita Log Settler Mortality of for Registering Mean in Log 6.34 5,63 (1.81) (1.47) (1,70) (1.63) (1,42) (1,43) (1,16) (1,05) (1,07) (120) 4.48 4.15 4.48 5.65 4,54 3,57 3,99 4,32 4,87 3,31 (2.09) (2.09) (2.06) (1-55) (2,60) (1,88) (1.78) (2,05) (172) (1,50) 3.66 378 2-77 2-38 2,57 3,15 4 65 427 479 4.60 (1.06) (1.24) (0.88) (0,59) (0,79) (0.75) (0,74) (0-89) (0.54) Average protection against which 13 with low 96) 375 2 99 2 98 261 3,15 4.60 4 44 4,65 4.58 (1.09) (0.68) (0,82) (0,49) (0,58) (0.79) (0,86) (0,77) (0.91) 8.36 7.91 8-17 9,04 8,21 7,65 7,71 8,20 8,23 7.22 (1.11) (1,00) (1-14) (1.03) (1,12) (091) (0,83) (095) (057) (0,65) 14.73 12.30 1486 19 42 10,13 1367 1053 10,98 1331 8,59 (7.89) (7.21) (8.37) (9,71) (4,97) (7,34) (5,49) (6,63) (436) (5.22) 0.42 0.36 048 083 036 0,29 0,26 0,45 027 0.19 (0.40) (0.35) (0,44) (0,55) (0,22) (025) (0,21) (0 19) (017) (0,22) 0.22 0.17 0,30 072 0,13 004 0,07 0,18 0,07 0.03 (0.37) (0.37) (051) 1 88 (056) (068) 1 64 (0,15) (0,06) (0,14) (0,28) (0,08) 209 2.19 2,61 2,38 2,68 2,62 (0,59) (043) (0.36) (0.29) (0,19) (0,29) (0,32) 2.25 2.27 (0 51) (0.57) 048 0,70 049 (0.81) (1.05) (088) 4.25 (149) 4.68 6 with medium 1 035 1.05 0.89 0.34 1,02 1,00 (023) (1,33) (1.17) (0.42) (1,42) (1.04) 3,01 4,56 6,14 4,93 376 4 44 5.75 (0,64) (0,92) (046) (025) (0.64) 15 0,25 -0,54 (026) 1 35 (0,94) 0.52 0,37 075 0.83 036 0.99 (1.53) (1,77) (253) (1,15) (1.13) (1,29) (2.61) (120) (0,95) medium 500 defined only ranges from 1 (0.09) (018) parentheses, "High", "medium", and "low" settler mortality are defined as roughly one in log population density in risk of expropriation settler mortality. (0 3.73 (0.92) n.a. and mortality 5,31 1500 settler mortality high settler (with standard deviations in parenttieses) 6,02 n.a. values, with standard deviations settler 6,21 former colonies' 29 English legal origin, o( which 14 with low mortality, 7 with mortality. mortality 6,21 (1.20) Log Population Density excolonies with medium mortality 7.93 1995 (PPP Measure) in excolonies with low settler excolonies 6.93 Legal Fonnalism (Eviction Measure) per capita excolonies with French mortality 6.39 Legal Formalism (Check Measure) GDP (10) French 7.07 Constraint on the Executive Log (9) French high settler mortality Mean Values Average Protection Against Risk settler (8) French (7) mortality, for excolonies. and 8 with high mortality; 51 third each French legal all excolonies We have log settler mortality data for 80 which 8 with low mortality. 20 with medium mortality and 23 with high of the distribution for origin, of For detailed sources and dermitions see Appendix Table A1 where a higher score indicates more protection The number of observations is 120 (whole world); 75 excolonies: 31 English legal origin excolonies of and 12 with high settler mortality; 44 French legal origin excolonies, of which 6 with low settler mortality, 18 with medium settler mortality, and 20 with htgh to 10, settler mortality, settler mortality. Constraint on the Executive ranges from settler mortality. 5 with medium where a higher score indicates more constraint. The numt>er of observations is 159 (whole worid); 92 excolonies: 35 English and 1 7 with high settler mortality: 56 French legal origin excolonies. of which 7 v/ith low settler mortality, 20 with medium settler to 7, 1 settler mortality Legal Formalism (Check Measure) ranges from lo 7, where a higher score indicates a higher English legal ongin excolonies of which 11 with low settler mortality, 5 with settler mortality, and 1 settler mortality, and 14 with high settler mortality; 35 legal origin Legal Fomialism {Eviction Measure) ranges from and 6 with high ratio is in with high settler mortality; settler mortality; current prices is to 7, where a higher score indicates a higher medium medium Log procedures mortality, 2 are is Cost of registering a are medium mortality is GDP level of control or intervention in the judicial process. settler mortality, and 13 with high settler mortality; 26 French settler mortality- The number of observations legal origin excolonies. of is 167 (whole worid): 98 excolonies; 42 English legal ongin excolonies of which 13 with low which 7 with low setUer mortality, 19 with medium settler mortality and 29 with high settler mortality. of observations The number of observations the market value of settler mortality the log of medium The number legal origin excolonies, of percent of is 165 (whole worid); 97 excolonies; 43 English 54 French legal origin excolonies, of which 8 with low Stock Market Capitalization mortality, 7 with and 29 with high is is 1 03 {whole world); 55 excolonies; 29 which 4 with low settler mortality, 16 with medium settler mortality. 55 French Credit to the private sector 22 with high mortality, observations Log GDP per capita in 1995 is PPP adjusted- The number of observations is 168 (whole worki); 97 excolonies; 42 English legal origin excolonies of which 13 with low 22 with high settler mortality; 55 French legal origin excolonies, of which 8 with low settler mortality, 19 wfith medium settler mortality and 28 with high settler mortality, lnvestment*GDP which 13 with low legal origin excotonies of 113 (whole worid); 65 excolonies; 30 French excolonies, of which 5 with low settler mortality, 17 with medium of 3 with high settler mortality. English legal ongin excolonies of which 11 with low settler mortality, 5 with settler mortality, medium The number level of control or intervention in the judicial process. and 1 all settler mortality, traded stocks as a percent of 19 with GDP, The number of medium legal origin excolonies of settler mortality, observations is 1 1 and 27 with high settler mortality, 7 with settler mortality, 7 with which 14 with low settler mortality, medium medium settler mortality 7 with medium 1 and and 22 settler mortality, and settler mortality. 5 (whole world); 84 excolonies; 37 English legal origin excolonies of which 7 with high settler mortality: 47 French legal origin excolonies, of which 7 with low settler mortality, settler mortality 7 with medium settler mortality number of prcicedures needed to open a new business. We have data on 82 countries in our sample, of which 43 are former colonies: 20 English and 7 are high mortality: 23 French legal origin excolonies, of which 4 are low mortality, 1 1 are medium mortality and 8 are high mortality- and 23 with high 1 3 with low settler settler mortality, legal ongin excolonies. of which 1 1 are low mortality new business as and a percent of 7 are high mortality: GDP 22 French per capita per annum. origin excolonies, of We have data which 4 are low Log Population density in 1 500 is available for 98 former colonies, of which: 42 are English are low mortality. 20 are medium mortality, and 30 are high mortality. on 81 countries mortality, origin, of 1 1 are which 1 in our sample, of which 42 are former colonies: 20 English origin excolonies, of which 11 are low mortality, 2 medium mortality and 7 are high 3 are low mortality, 7 are medium mortality. mortality, and 22 are high mortality; and 56 are French legal origin, of which 6 Table 2 GDP per capita. Contracting Institutions: Credit, and Stock (2) (3) Excolonies Excolonies Sample Sample (1) Whole World Market Capitalization (4) (6) (5) Excolonies Whole World Sample Excolonies Sample 2SLS, with Check OLS 2SLS, with OLS Measure Panel A: Dependent variable Legal Formalism (Check Measure) -0.28 -0.21 -0.18 (0.10) (0,10) (0.14) log is Legal Formalism (Eviction Measure) R-Squared in OLS 0.07 0.07 First Stage for in First Panel B. (0.12) (0.12) (0.17) 0.008 0.005 0.56 99 65 Dependent var table is -1.85 -1.19 -\ni (0.71) (0.94) OLS 0,07 0.04 110 65 Panel Legal Formalism (Check Measure) C. • 65 Dependent variable is credit to -0.16 -0.14 -0.13 (0.04) (0,03) (0.05) 55 average ratio of investment -0.69 Number of Observations tin? 55 to -1.15 (1.22) (0.82) (0.91) 0.007 0.004 100 55 -0.08 -0.12 (0.05) (0.07) 0.21 0.03 0.05 104 65 94 55 65 is \ ofGDPin 1998 -0.08 0.16 Panel D: Dependent variable 55 (0.05) OLS in 1990s -0.43 of Observations R-Squared GDP in private sector as a percent Legal Formalism (Eviction Measure) Number -0.04 0.58 -' in -0.06 -1.61 Legal Formalism (Eviction Measure) R-Squared -0.10 (0.20) (0.67) Legal Formalism (Check Measure) 1995 -1.87 65 109 in (0.20) Stage Number of Observations GDP per capita Measure of Legal Formalism (Check or Eviction Measure) English Legal Origin R-Squared Eviction Measure 55 stock market capitalization as a percent of GDP, average over 1990-95 Legal Formalism (Check Measure) -0.17 -0.17 -0.16 (0.03) (0.04) (0.05) Legal Formalism (Eviction Measure) R-Squared in OLS 0,24 0.26 90 62 Number of Observations Standard errors are is: in Panel A, log 62 in parentheses. All regressions are cross-sectional GDP per capita (in PPP terms) ratio, in current prices, in 1995, World Bank's of GDP, 1990-95, from Beck (2003a). et al from the World Bank's WDI on-line; The independent and in variable in WDI 0.05 80 52 52 is: in columns 1, 2 and 3, legal formalism, using the "check formalism, using the "eviction measure" from Djankov, La Porta, four panels the measure of legal formalism is instrumented using a dummy variable for whether a country has an English legal origin, as classified by Djankov, La Porta, Lopez-de-Silanes, and Shieifer (2003) and Djankov (2003). GDP Panel D, the level of stock market capitalization as a percent 4, 5, 6, legal 0.06 on-line; in Panel B, the investment to Lopez-de-Silanes, and Shieifer (2003). In and -0.17 (0.07) Panel C, level of credit to the private sector as a measure", from Djankov (2003); in columns all -0.09 (0.06) with one observation per country. The dependent variable average over 1990s, from Penn World Tables 6.1; percent of GDP in 1998, from the -0.11 (0.05) Table 3 GDP per capita. Property Rights Institutions: (3) (4) Excolonies Sample Excolonies Sample Excobnies Sample (5) Against Risk of Exprop. in OLS 2SLS, with 2SLS, with 2SLS, with Constraint on Constraint on Exprop. Exprop. Executive Executive 0.56 0.52 1.05 1.07 (0.06) (0.19) (0.22) 0.54 First Log OLS is log GDP per capita • Settler Mortality R- Squared Stage in First 65 Against Risk of Exprop. R-Squared in OLS Against Risk of Exprop. 0.17 65 65 average ratio of investment is 5.50 5.48 (1.12) (1.33) 65 65 is R-Squared in OLS 037 0.31 (0.02) (0.03) (0.08) (0.08) 0.33 109 66 66 66 Pane! D: Dependent variable Average Protection Against Risk of Exprop. is 0.11 0.14 0.30 0.26 (0,03) (0.08) (0.09) Constraint on Executive R-Squared in OLS 0,24 0.25 95 63 63 63 WDI on-line; through 4, Polity IV. and GDP in 1.61 1.40 4.16 3.76 (0.42) (1.04) (1.08) 0.16 0.14 ratio, in current prices, average over 1 990s, from Penn World Tables average protection against the risk of expropriation as used The measure of institutions is in 6,1; in 8, of GDP in 1998 0.08 0.08 0.25 0.17 (0.02) (0.02) (0.06) (0.05) 0.15 0.19 140 70 70 1 990-95 0.06 0.06 0.21 0.14 (0.02) (0.02) (0.06) (0.05) 0.12 0.07 is: 66 in Panel A, log et al (2003a). Acemoglu. Johnson and Robinson (200 1 1 ); in per capita < in The measure of institutions used columns 5 through 850 (where mortality using log population density in 1500 from Acemoglu, Johnson and Robinson (2002). is per 8, constraint 1 66 66 GDP Panel C, the level of credit to the private sector as a percent of GDP instrumented: in columns 3 and 7, using log settler mortality before and Robinson (2001); and in columns 4 and 70 105 Panel D, the level of stock market capitalization as a percent of GDP, 1990-95, from Beck 69 (0.30) Standard errors are in parentheses. All regressions are cross-sectional with one observation per country. The dependent variable the investment to 0.24 69 1990s stock market capitalization as a percent of GDP. average over (0.02) Number of Observations GDP in credit to the private sector as a percent 0.16 0.32 to 149 0.15 0.26 69 146 Constraint on Executive Number of Observations ) -0.56 0.23 Panel C: Dependent variable Average Protection 0.34 (0.12) 3.00 65 0.35 •0.36 (0.48) 0.38 0.70 (0.15) (0.10) 3.17 113 0.76 (0.15) -0.80 (0.33) 0.45 0.32 (0.05) (0.16) Constraint on Executive Number of Observations 0.34 (0.04) (0.13) Panel B: Dependent variable Average Protection 1995 -0.57 1500 Number of Observations in Stage for Measure of Institutions (Protection Against Expropriation or Constraint on Executive Log Population Density in (8) Excolonies Sample Protect. Against (0.04) 0.63 (7) Excolonies Sample 2SLS.with Constraint on Executive R-Squared (6) Excolonies Sample Protect, Against Panel A: Dependent variable Average Protection and Stock Market Capitalization Credit, (2) (1) PPP terms) in 1 in 1995; in Panel B, 998, from the as an independent variable World Bank's in columns is: on the executive, averaged over the 1 990s, from 000 per annum with replacement) from Acemoglu, Johnson 1 Table 4 GDP Contracting vs. Property Rights Institutions: (2) (1) (4) (3) Depe,ndeni variable per capita (5) is log (6) GDP per capita in 2SLS, with 2SLS, with Protect Protect Against Exprop. OLS & Check Agamst Exprop. OLS Measure OLS Average Protection Against Risk of Expropriation 0.63 0.95 (0.08) (0.16) & or Second 0.64 1.53 (0.08) (0.25) Constraint on Executive Legal Formalism (Checls Measure) in OLS 2SLS, with Constraint on & Check & Eviction Measure Regression 0.31 0.99 0.33 0.75 (0.07) (0.29) (0.07) (0.18) 0.11 0.35 -0.16 0.05 (0.15) (0.10) (0.24) 0.62 Exec. OLS Measure (0.09) Legal Formalism (Eviction Measure) R-Squared 2SLS, with Exec. Sta]ge (8) Constraint on OLS Eviction Measure (7) 1995 0.28 0.45 -0.04 0.08 (0.09) (0.16) (0.11) (0.20) 0.34 0.33 0.61 First Stage for Measure of Institutions (Protect. Against Expropriation or Constraint on Executive) English Legal Origin Log Settler Mortality R-Squared in First 0.60 0.44 -0.002 -0.22 (0.31) (0.33) (0.48) (0.46) -0.71 -0.70 -0.65 -0.92 (0.12) (0.15) (0.19) (0.21) 0.21 0.33 0.42 0.50 Stage First Stage for English Legal Origin Log Settler Mortality R-Squared in First Stage Standard errors are 51 in parentheses. All per capita (PPP adjusted) in -1.69 -1.88 -1.7 (0.24) (0.23) (0.24) 0.15 -0.02 0.11 0.01 (0.10) (O.ll) (0.09) (0.11) 0.57 1 43 51 0.62 43 51 regressions are cross-sectional with one observation per country; 995, from the World Bank's protection against the risk of expropriation as used in 1990s, from Polity IV. -1.96 (0.24) 0.63 Number of Observations GDP Measure of Legal Formalism (Check or Eviction Measure) WDI on-line. we The measure of institutions used use only the excolonies sample. as an independent variable is: 42 The dependent in columns 1 variable through 4, is log average Acemoglu. Johnson and Robinson (2001); in columns 5 through 8. constraint on the executive, averaged over the as an independent variable is: in columns 1-2 and 5-6, the "check measure", from Djankov (2003); in columns 3- The measure of legal formalism used 4 and 7-8, the "eviction measure" from Djankov, La Porta, Lopez-de-Silanes. and Shleifer (2003). In columns the instruments are: a 0.56 42 51 dummy Djankov (2003); and log variable for whether a country has an English legal origin, as classified settler mortality before 1850 (where mortality is per 1000 per annum with 1, 3, 5, by Djankov, La replacement). and 7 we report OLS results. In Porta, Lopez-de-Silanes, columns 2, 4, 6, and 8 and Shleifer (2003) and Table 5 Contracting vs. Property Rights Institutions: Investment-GDP ratio (1) Dependent variable is 2SLS, with 2SLS, with Constraint on Constraint on & Check Exprop. OLS Measure OLS 3,88 4.68 4.23 5.28 (0.65) (1.11) (0.72) (1.18) & Exec. OLS Eviction Measure OLS -1.05 -0.80 (1.55) 0.47 -0.46 -0.85 (1.01) (1.64) 0.10 -0.002 -0.22 (0.48) (0.46) -0.71 -0.70 -0.65 -0.92 (0.12) (0.15) (0,19) (0.21) 0.21 0.33 0.42 0.50 Stage for Measure of Legal Formalism (Check or Eviction Measure) -1.96 -1.69 -1.88 -1.70 (0.24) (0.24) (0.23) (0,24) 0.15 -0.02 0.11 0.01 (0.10) (0.11) (0.09) (0.11) Stage 0.63 51 columns 1 GDP ratio, 43 through 4, average protection against the risk of expropriation as used executive, averaged over the 1990s, from Polity IV. Djankov (2003); results. In Silanes, in columns over 1990s, from Penn World Tables columns 3-4 and 2, 4, 6, 51 are: a we use only the excolonies sample. The dependent variable The measure of institutions used as an independent variable as an independent variable is: in columns 1-2 and dummy variable for settler mortality whether a country has an English before 1850 (where mortality is legal origin, as classified per 1 8, is: 1, 3, 5, constraint and 7 by Djankov, La 000 per annum with replacement). we is in on the 5-6, the "check measure", measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003). In columns and Shleifer (2003) and Djankov (2003); and log 42 42 51 Aceraoglu, Johnson and Robinson (2001); in columns 5 through The measure of legal formalism used 7-8, the "eviction and 8 the instruments in 6.1. 0.56 0.62 0.57 43 51 in current prices, average on Executive) 0.44 Standard errors are in parentheses. All regressions are cross-sectional with one observation per country; the investment to 0.06 (Protect. Against Expropriation or Constraint (0.33) Settler Mortality in First 0.91 0.60 English Legal Origin R-Squared 3.91 (1.40) (0.31) Stage Number of Observations & Eviction Measure (1.20) 0,47 Measure of Institutions First Log 1,04 (0.66) (0.83) Settler Mortality in First 4.70 (1.87) 0.57 English Legal Origin R-Squared 1.08 (0.57) (1.08) 1.57 Exec. OLS or Second St age Regression 0.67 First Stage for Log & Check Measure (0.71) (0.85) in 1990s 2SLS, with Legal Formalism (Eviction Measure) R-squared (8) (7) GDP in Protect Against Constraint on Executive Legal Formalism (Check Measure) to 2SLS, with OLS Against Risk of Expropriation (6) average ratio of investment Protect Against Exprop. Average Protection (5) (4) (3) (2) report from OLS Porta, Lopez-de- Table 6 Contracting vs. Property Rights Institutions: Private Credit (2) (1) (3) (4) Dependent variable is (fi) (5) credit to the private sector in 1 99S as a percent oj 2SLS, with Protect & Exprop. Against Exprop. OLS Check Measure OLS 0.13 0.28 0.19 0.36 (0.04) (0.07) (0.05) (0.09) OLS Average Protection Against Risk of Expropriation & 2SLS, with Constraint on Constraint on Exec. OLS Eviction Measure or in OLS in First -0.13 -0.08 (0.08) 0.001 -0.08 -0.13 (0.09) (0.05) (0.08) 0.34 Stage Settler Mortality in First Stage 0.44 -0.002 -0.22 (0.33) (0.48) (0.46) -0.71 -0.70 -0.66 -0.92 (0.12) (0.15) (0.19) (0.21) 0.21 0.33 0.42 Measure of Legal Formalism (Check or Eviction Measure) -1.96 -1.69 -1.88 -1.70 (0.24) (0.24) (0.23) (0.24) 0.15 -0.02 0.11 0.01 (0.10) (0.11) (0.09) (0.11) 0.57 43 51 51 average protection against the the 1990s, from Polity IV. columns 3-4 and 4, 6, risk 51 in 1998, we 42 51 use only the excolonies sample. The dependent variable from the World Bank. The measure of institutions used as an independent variable of expropriation as used in 42 is; in columns 1 is through 4, Acemoglu, Johnson and Robinson (2001); in columns 4 through 8, constraint on the executive, averaged over as an independent variable is: in columns 1-2 and 5-6, the "check measure", from Djankov (2003); in The measure of legal formalism used 7-8, the "eviction and 8 the instruments of GDP 0.56 0.62 43 Standard errors are in parentheses. All regressions are cross-sectional with one observation per country; credit to the private sector as a percent 0.13 0.60 0.63 Number of Observations 0.28 (0.31) 0.50 English Legal Origin R-Squared 0.01 (0.05) Stage for Measure of Institutions (Protect. Against Expropriation or Constraint on Executive) First Stage for Log 0.19 (0.06) (0.04) Settler Mortality R-Squared 0.06 (0.03) 0.01 English Legal Origin Log 0.27 (0.10) (0.07) First & Eviction Measure 0.06 -0.09 0.42 Exec. OLS (0.03) (0.04) Legal Formalism (Eviction Measure) R-squared & Check Measure Second Stage Regression Constraint on Executive Legal Formalism (Check Measure) GDP 2SLS,with 2SLS. with Protect Against (7) are; and Djankov (2003); and log measure" from Djankov, La Porta, Lopez-de-Silanes. and Shleifer (2003). In columns 1,3,5, and 7 dummy we report OLS results. In columns 2, vanable for whether a country has an English legal origin, as classified by Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003) settler mortality before 1850 (where mortality is per 1000 per annum with replacement). Table 7 Contracting vs. Property Rights Institutions: Stock Maricet Capitalization (1) (2) Dependent variable is Against Risk of Expropriation 2SLS, with 2SLS, with 2SLS, with Protect Against Constraint on Constraint on & 0.10 0.21 (0.04) (0.07) & Exprop. OLS Check Measure Exec. OLS Eviction Measure or in OLS 0.32 (0.06) (0.10) R-Squared -0.17 -0.16 (0.07) -0.03 -0.10 -0.10 -0.20 (0.07) (0.10) (0.06) (0.08) 0.23 Stage English Legal Origin Settler Mortality R-Squared in First Stage (Protect. Against Expropriation or Constraint on Executive) 0.61 0.45 -0.05 -0.28 (0.31) (0.33) (0.49) (0.46) -0.71 -0.70 -0.65 -0.91 (0.12) (0.15) (0.19) (0.21) 0.42 0.20 0.33 Measure of Legal Formalism (Check or Eviction Measure) -2.00 -1.76 -1.93 -1.77 (0.24) (0.23) (0.23) (0.23) 0.15 -0.004 0.11 0.004 (0.09) (0.10) (0.09) (0.11) 0.66 50 42 50 Standard errors are in parentheses. All regressions are cross-sectional with one observation per country; 1 GDP in 1990-95, from Beck et al (2003). 0.62 0.65 0.62 42 50 the level of stock market capitalization as a percent of 0.11 0.31 Measure of Institutions 0.50 Number of Observations 0.17 (0.06) (0.04) First Stage for Log 0.05 (0.04) -0.10 0.43 in First 0.20 (0.09) (0.07) Settler Mortality we 50 41 Djankov (2003); results. In Silanes, in columns 1 990s, from Polity IV. columns 3-4 and 2, 4, 6, The measure of legal formalism used 7-8, the "eviction and 8 the instruments are; a as an independent variable as an independent variable is: in columns 1 -2 and and Shleifer (2003) and Djankov (2003); and log variable for whether a country has an English legal origin, as classified settler mortality before 1850 (where mortality is per 1000 per annum with 8, in is columns on the "check measure", from 1, 3, 5, and 7 by Djankov, La replacement). is; constraint 5-6, the measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003). In columns dummy 41 use only the excolonies sample. The dependent variable The measure of institutions used through 4, average protection against the risk of expropriation as used in Acemoglu, Johnson, and Robinson (2001); and in columns 5 through executive, averaged over the & Eviction Measure 0.04 -0.15 English Legal Origin Exec. OLS (0.03) (0.05) First Stage for Log & Check Measure Second Stage Regression 0.17 Legal Formalism (Eviction Measure) R-squared (8) 2SLS, with Constraint on Executive Legal Formalism (Check Measure) (7) (6) stock market capitali. alion. average 1990-95. as a percent of GDP OLS Average Protection (5) Protect Against Exprop. OLS (4) (3) we report OLS Porta, Lopez-de- Table 8 Contracting vs. Property Rights Institutions: Alternative Samples (1) (2) (3) French Just Just Common legal origin Law countries countries Dependent variable Panel A: is log (4) (5) (7) (6) Without Just countries neo- above median Just Common legal origin neo- Europes world income Law countries countries Europes GDP per capita Just 1995 in French Dependent variable Without investment-GDP is (8) Just countnes above median world income ratio in 1990s Second Stage Regression Average Protection Against Risk of Expropriation 1.12 1.45 1.10 1.15 6.30 5.37 5.72 6.65 (0.27) (0.67) (0.23) (0.29) (1.78) (2.46) (1.57) (2.18) Legal Formalism (Check Measure) 0.31 0.40 0.26 1.52 (0.16) (0.22) (1.12) (1.63) First Stage for English Legal Origin Log in First 0.61 0.33 0.61 (0.30) (0.39) (0.30) (0.39) -0.59 -0.32 -0.55 -0.65 -0.59 -0.45 -0.55 -0.65 (0.19) (0.19) (0.13) (0.19) (0.19) (0.20) (0.13) (0.19) 0.30 0.07 0.32 0.41 0.30 0.12 0.32 0.41 Settler Mortality R-Squared Stage First English Legal Origin Log Settler Mortality R-Squared in First Average Protection Against Risk of Expropriation 0.33 Stage Number of Observations 40 25 Stage for Check Measure of Legal Formalism -1.89 -2.16 -1.89 -2.16 (0.26) (0.32) (0.26) (0.32) 0.09 0.12 0.09 0.12 (0.11) (0.16) (0.11) (0.16) 0.56 0.65 0.56 0.65 47 42 47 42 40 25 Just French Without Just Common legal origin neo- Without Just Common legal origin neo- Without Law countries countries Europes Outliers Law countries countries Europes Outliers Dependent variable is Just Dependent variable credit to the private sector as a percent of GDP Panel B: Average Protection Against Risk of Expropriation 0.44 0.24 0.35 0.24 0.39 (0.17) (0.11) (0.06) (0.15) (Check Measure) Number of Observations 40 26 is stock market capitalization as a percent of GDP (0.13) Legal Formalism Without French 0.09 0.36 0.14 (0.05) (0.11) (0.05) -0.03 -0.02 -0.14 -0.06 (0.08) (0.06) (0.07) (0.05) 47 48 25 46 38 47 Standard errors are in parentheses. All regressions are two stage least squares; cross-sectional data with one observation per country; the excolonies sample. line; in columns sector as share of Beck et al. The dependent 5-8, Panel GDP (2003a). variables are: in A, investment-GDP in 1998, ratio, from the World Bank's The measure of institutions used colunms 1-4, Panel A, log GDP per capita, PPP adjusted, fi^om the average over 1990s, from Penn World Tables 6.1; WDI on-line; in columns as an independent variable 5-8, Panel B, stock is in columns 1-4, we World Bank's use only WDI on- Panel B, credit to the private market capitalization as share of GDP, from average protection against risk of expropriation, as used in Acemoglu, Johnson, and Robinson (2001). The measure of legal formalism used as an independent variable is the "check measure", from Djankov (2003). The instruments are: in columns 3-4 and 7-8 a dummy variable for whether a country has an English legal origin, as classified by Djankov, La Porta, Lopez-de-Silanes, per annum with Columns 3 and Shieifer (2003) and Djankov (2003); and 7 drop the four "neo-Europes": Australia, Canada, below median world income (using all in all columns, log settler mortality before 1850 (where mortality is per 1000 replacement). GDP per capita, former colonies of European Powers for which PPP, New Zealand, in 1995). In we have data. and USA. columns 4 and 8, In Panel B, columns 4 and we 8, Panel A, we drop all countries drop outliers. The Excolonies Sample includes For more detailed data definitions and sources see Appendix Table 1 Table 9 Contracting vs. Property Rights Institutions: Other Control Variables (I) (4) (2) Dependent variable log is GDP per capita in (5) 1995 Religion is (S) (7) (6) Dependent Variable investment-GDP Inflation Religion Latitude Dependent variable private credit as a percent of GDP ratio Log Latitude is Log Inflation (12) is stock market capitalization as percent of Log Religion (II) (10) (9) Dependent Variable Inflation Religion Latitude GDP Log Inflation Second Stage Average Protection Against Risk of Expropriation 0.88 0.96 0.98 3.65 5.60 4.68 0.28 0.30 0.28 0.21 0.28 0.22 (0.16) (0.18) (0.16) (1.36) (1.32) (1.13) (0.09) (0.09) (0.07) (0.09) (0.09) (0.07) Legal Formalism 0.09 0.34 0.32 -2.56 0.93 0.55 -0.05 0.001 0.01 -0.14 -0.08 -0.10 (Check Measure) (0.28) (0.15) (0.16) (2.31) (1.14) (1.14) (0.15) (0.08) (0.07) (0.15) (0.08) (0.08) p-value for religion [0,50] Latitude Log 037 -12.70 0,31 -0.86 (1.00) (7.42) (0.49) (0.51) 0.08 0.11 -0.03 -0.02 (0.09) (0.65) (0.70) (0.04) Inflation First Stage for English Legal Origin Log Settler Mortality R-Squared in First Stage Average Protection Against Expropriation 0.09 0.61 0.58 0.09 0.61 0.58 0.09 0.61 0.58 0.08 0.62 0.58 (0.42) (0.31) (0.33) (0.42) (0.31) (0.33) (0.42) (0.31) (0.33) (0.44) (0.31) (0.33) -0.70 -0.65 -0.71 -0.70 -0.65 -0.71 -0.70 -0.65 -0.71 -0.70 -0.65 -0.71 (0.12) (0.13) (0.13) (0.12) (0.13) (0.13) (0.12) (0.13) (0.13) (0.13) (0.14) (0.13) 0.55 0.52 0.50 0.55 0.52 0.50 0.55 0.52 0.50 0.55 0.52 0.50 First Stage for English Legal Origin Log Settler Mortality R-Squared in First Stage -1.10 -1.97 -1.94 -1.10 -1.97 -1.94 -1.10 -1.97 -1.94 -1.13 -2.01 -1.94 (0.24) (0.25) (0.28) (0.24) (0.25) (0.28) (0.24) (0.25) (0.29) (0.24) (0.25) 0.19 0.12 0.15 0.19 0.12 0.15 0.19 0.12 0.15 0.19 0.12 0.15 (0.08) (0.11) (0.09) (0.08) (0.11) (0.09) (0.08) (0.11) (0.09) (0.08) (0.10) (0.09) 0.76 0.64 0.66 0.76 0.64 0.66 0.76 0.64 0.66 0.76 0.66 0.66 51 51 50 51 51 50 51 51 50 50 50 50 Standard errors are in parentheses. All regressions are two stage log GDP per capita independent variables in Check Measure of Legal Formalism (0.28) Number of Observations 3, least squares; cross-sectional data columns 1, 4 and 7. dummy absolute normalized value of distance from the equator variables for majority religions, Protestant, (i.e.. a standardized measure of latitude), in we use only the excolonies sample. The dependent variables are: in columns of GDP; in columns 10-12, stock market capitalization as share of GDP. The with one observation per country; 1995; in columns 4-6, invcstment-GDP ratio, average over 1990s; in columns 7-9, private credit as share are: in [0.4S] [0.82] [0.24] Muslim, and Other, with Catholic as the omitted category, from La Porta which a higher value indicates that a country's capital average annual inflation in the Consumer Price Index from 1970 to 1997, from the World Bank. The other control variables are included The measure of legal formalism used as an independent variable over the 1990s, from Polity IV. The instruments are; a (2003); and log settler mortality before 1 detailed data definitions and sources see Is the "check measure", dummy variable 850 (where mortality Appendix Table 1 is per 1 for in the first is further origin, as classified (1999); in columns 2, 5 and 8, the from the equator; columns in 3, 6 and 9. the log of stage but not reported to save space. from Djankov (2003). The measure of institutions used as an independent variable whether a country has an English legal et al is constraint on the executive, averaged by Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003) and Djankov 000 per annum with replacement). The Excolonies Sample includes all 1 - former colonies of European Powers for which we have data. For more Investment-GDP Depend e III V 2SLS, wiih lahlc r nera^e ratio imeslmi'nt lo 2SLS. wiih Protect Against Protect Against Exprop, & Exprop ratio, Credit, & Check Measure Check Measure Table 10 and Market Capitalization: Controlling for GDP in IVWs 2SLS,with 2SLS,wUh Prolecl Against Protect Against I:\prap & & Exprop tviclion Eviction Measure Measure Dfpendet iahle is 2SLS, with 2SLS. with Protect Against Protect Against Exprop. & Check Measure Exprop & Check Measure credit to GDP per capita pn\ale wiior Dependent variable 2SLS.with 2SLS.wiih Protect Against Protect Against Exprop & & Exprop Eviction Eviction Measure Measure is 2SLS, wilh 2SLS, wiih Protect Against Protect Against Exprop & Exprop & Check Measure Check Measure market capitalization 2SLS. with :SLS. uith Protect Against Prolecl Against Exprop & & Exprop Eviction Eviction Measure Measure Seconii Stage Regression Average Protection -0 Against Risk of Expropnalion 59 (672) Legal Formaiism (Ciieck Measure) 1 28 (2 66) 5.11 271 J 14 051 -O04 (9.27) (20.17) (404) (0.64) (042) GDP per capita p-value for joint Check Measure GDP 0.07 -Oil 003 -0 14 (0.25) (0.18) (0.25) (0.22) -0.18 0-02 0.62 -0.18 0.60 -0-22 (2.03) (2.29) (012) (2.53) (0 15) GDP per capita in 2.57 2 42 fl.24 0.27 -1.47 0.37 -0.21 0.15 (19 49) (3.38) (058) (034) (5.20) (0.20) (0.57) (041) (5 76) (026) |0 00) 10.00] (000) 10.00] 1001] 10.00] 10.43] |000) 10.05] 10.00) 10.59] 10.00) 10-46) (0.42) 10.65) 10.36] 1090] (0,62) 1096] 1020) 10,42] 10.49) 10.89) 10.32) 0,67 0.80 (025) (024) 70 (028) 65 in 1995 080 070 0.71 0.70 082 0.75 0.73 (0-24) (0.28) (0.25) (0.25) (0.24) (0.28) (0-26) -0,17 0.80 (0.13) (022) -1 Settler Moruility GDP per capita in 1 (0.08) 1995 Number of Observations Standard errors are (0-20) -0.08 (009) (0.08) 0.70 084 0.80 0.95 0.70 0.86 (0.16) (0.17) (0 13) (0.22) (0.16) (0.17) 0.68 062 061 0.53 -0 12 (0,09) 0.81 099 071 (0.13) (023) (0,16) 056 0.68 0.62 0-58 97 -1.94 -1-72 -1.66 -1.97 -1.94 -1.72 -1.66 -2.02 -1.98 -1.81 -173 (022) (0,25) (0-22) (024) (0.22) (0.25) (0.22) (0.24) (0.21) (0.23) (0.20) 50O SUlge (0.15) -0.12 062 61 053 Suigc for Legal Formalism (Check or Eviction Measure) 005 -0-08 0.05 in (0.20) -0.07 68 (018) (0 14) Log Population Density -001 -0.16 -0.03 (015) (0 09) 0.84 Stage 0.67 (0.25) -012 (0.17) Stage for Prolcclion Against Expropnation 0-71 -0.03 095 (024) in First -1 (0.25) (0.20) -007 1500 English Legal Origin R-Squared 026 064 (751) First Log -0 04 (0 31) (858) (008) Log 95) 5 35 (0 15) in First 1.96 (5 (6 14) -017 Settler Mortality Log Population Density R-Squared O04 (0.51) 099 Firei Log 0-42 (0.62) per capita English Legal Origin Log -012 (0.24) GDP per capita test of Formalism and 1995 Protection against test Lxpropriation and p-value for joint in 83 (4.02) Legal FonTialism (Eviction Measure) Log 1 (5.38) 003 -0.08 -0.13 (0 14) (0.18) (0.14) (0.17) 0.03 -0.04 0.03 -0.04 0.04 -0.03 (008) (0.08) (0.08) (008) (0.08) (0.07) -0.14 -0 12 (0.16) (0.12) 0.62 0.63 51 57 in parentheses, All regressions are cross-sectional 09 -0.09 -0-14 -0.12 -0.09 -0.09 0.18 0.13 -0.18 -Oil (0-20) (0.14) (016) (0,12) (0.20) (014) (0.16) (0-12) (0 19) (013) 0.57 062 063 0.57 0.57 0.65 0.60 0.60 51 57 56 42 48 -0 57 43 49 with one obserNation per countr>': we use only the excolonics sample The dependent 43 variable 67 49 is: in columns 50 1 through 4, the investment to GDP ratio, in current pnces, average over 1990s, World Bank, and m columns 9 through 12, stock market capiializaiion as a percent of GDP in 1990-95, from Beck el al (2003). The measure of institutions used as on independent variable is average protection against the risk of expropriation as used in Accmoglu, Johnson and Robinson (2001). The measure of legal formalism used as an independent variable is; in columns 1-2. 5-^, and 9-10 the "check measure", from Djankov (2003), m columns 3-4. 7-8. and 1-12 the "eviction measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer{2003). The instruments are; mail columns, log GDP per capita in 1995 and a dummy variable for whether a country has an English legal origin, as classified by Djankov. La Porta, Lopcz-de>Si lanes, and Shieifer (2003) and Djankov (2003); in odd numbered columns, log settler mortality before I8S0 (where mortality is per 1000 per annum with from Penn World Tables 6.1; in columns 5 through 8. credit tolhe pnvale sector as a percent of GDP 1 replacement), and in even numbered columns, log population density in 1500. in 1998. from the Table 1 Contracting vs. Property Rights Institutions: Procedures for Entry (2) (1) (3) Dependent variable is log Legal Formalism (Check Measure) 2SLS, with 2SLS, with 2SLS, with 2SLS, with Protect Against Protect Against Protect Against & in OLS & Exprop. Exprop. & Exprop. OLS Check Measure OLS -0.17 -0.20 -0.21 -0.31 -0.17 -0.24 -0.21 -0.27 (0.05) (0.08) (0.05) (0.09) (0.04) (0.10) (0.05) (0.08) OLS Eviction Measure or & Eviction Measure Second Stage Regression 0.15 0.18 0.16 0.16 (0.08) (0.05) (0.08) 0.17 0.21 0.17 0.22 (0.06) (0.09) (0.06) (0.08) 0.50 0.51 0.50 OLS Check Measure (0.05) Legal Formalism (Eviction Measure) R-squared (8) number of procedures for registering a new business OLS Against Risk of Expropriation (V) (6) Protect Against Exprop. Average Protection (5) (4) 0.52 First Stage for Protection Against Expropriation English Legal Origin Log Settler Mortality 0.53 0.35 0.83 0.51 (0.32) (0.34) (0.32) (0.31) -0.72 -0.71 (0.12) (0.15) Log Population Density R-Squared in First Stage 0.43 0.51 First Stage for English Legal Origin Log Settler Mortality in First Stage 48 the log of the number of procedures needed 0.37 -1.86 -1.64 -1.87 -1.60 (0.25) (0.22) (0.22) 0.16 -0.01 (0.10) (0.11) 48 40 to register a business, expropriation as used in Acemoglu, Johnson and Robinson (2001). 0.08 0.003 (0.06) (0.06) 0.54 0.61 40 54 54 Standard errors arc in parentheses. All regressions arc cross-sectional with one observation per country; is 0.35 (0.25) 0.62 Number of Observations -0.38 (0.08) Measure of Legal Formalism (Check or Eviction Measure) Log Population Density R-Squared -0.36 (0.09) we use only the excolonies sample. from Djankov (2003). The measure of institutions The measure of legal formalism used 0.55 is 46 46 The dependent variable average protection against the risk of as an independent variable is: in columns 1-2 and 5-6, the "check measure", from Djankov (2003); in columns 3-4 and 7-8, the "eviction measure" from Djankov, La Porta, Lopez-de-Silanes, and Shieifer (2003). In columns 3, 5, and 7 we report classified OLS results. In columns 2, 4, 6, and 8 the instruments are: in all colurrms, a by Djankov, La Porta, Lopez-de-Silanes, and Shieifer (2003) and Djankov (2003); per 1000 per annum with replacement); and in colunms 6 and 8, in dummy columns 2 and log population density in 1500. I, variable for whether a country has an English legal origin, as 4, log settler mortality before 1850 (where mortality is Table 12 Contracting vs. Property Rights Institutions: Cost of Entry (11 (3) (2) is cost of registering 3 < & Exprop. & 2SLS. with 2SLS, with Protect Against Protect Against Eviction -0.44 -0.23 -0.38 -0.40 -0.28 -0.22 -0.25 (0.11) (0.18) (0.05) (0.09) (0,10) (0.22) (0.05) (0.09) Measure Check Measure -0,07 0.03 0.06 0.11 (0.18) (0.11) (0.19) -0.13 -0.17 -0.12 0.10 (0.10) (0.06) (0.09) 0.28 Settler Mortality 0.29 Stage for Protection Against Expropriation First English Legal Origin 0.53 0.37 0.83 0.51 (0.32) (0.32) (0.32) (0.31) -0.72 -0.71 (0.12) (0.15) Log Population Density Stage Log 0.43 0.51 Settler Mortality R-Squarcd in First Stage is -1.64 -1.87 -1.6 (0.25) (0.22) (0.22) 0.16 -0.01 (0,10) (O.II) 0.54 0.62 48 the cost of registenng a business as a percent of average protection against the risk of expropriation as used columns 1-2 40 and Shleifer (2003). In columns 1, 3, 5. 1 in 850 (where mortality is per 1 La OLS 0.55 wc 46 46 54 use only the cxcolonies sample. The dependent from Djankov (2003). The measure of institutions used as an independent variable results. In in columns 3-4 and columns 2, 4, 6, in columns 6 and 8, 1 ). The measure of legal formalism used 7-8, the "eviction is as an independent measure" from Djankov, La Porta, Lopez-de- and 8 the instruments Porta, Lopcz-dc-Silanes. and Shleifer (2003) 000 per annum with replacement); and 0.003 (0.06) 54 Acemoglu, Johnson, and Robinson (200 and 7 we report has an English legal origin, as classified by Djankov, before GDP per capita, 0.08 (0.06) 0.61 40 and 5-6. the "check measure", from Djankov (2003); is: in Silanes, 0.37 -1.86 48 variable 0.35 (0.25) Standard errors arc in parentheses. All regressions arc cross-sectional with one obscr\'ation per country; vanablc -0.38 (0.08) i Log Population Density Number of Observations -0.36 (0.09) Stage for Measure of Legal Formalism (Check or Eviction Measure) First English Legal Origin Eviction Measure (0.07) 0.33 0.30 OLS or Second Stage Regress ion (0.12) OLS in First & Exprop. -0.43 Legal Formalism (Eviction Measure) R-Squared & Exprop. OLS OLS Legal Formalism (Check Measure) Log (8) Measure OLS in (7) ofGDPpt'r capita OLS Average Protection R-squared (6) business as percent Protect Against Check Against Risk of Expropriation new 2SLS. with Protect Against Exprop. (5) (4) Dependent variable 2SLS. with arc: a and Djankov (2003); log population density in 1 in dummy variable for whether a country columns 2 and 500. 4, log settler mortality Appendix Tables: Not for Publication Appendix Table Al Variable Definitions and Sources Variable Source Description Average Protection against Expropriation Risk Risk of expropriation of private foreign investment by government, from calculated ihe mean value for the scores in to 10, where a higher score means years from 1985 to 1995. This variable all is less risk. We Acemoglu, as previously used in Johnson, and Robinson (2001). Dataset obtained directly from Political Risk Services, September 1999. These data were previously used by Knack and Keefer(l995) and were organized in electionic form by the IRIS Center (University of Maryland). The original compilers of these data are Political Risk Services. A Constraint on Executive seven category scale, from authority; score I of 3 indicates to 7, with a higher score indicating 1990 Legal Formalism (Check Measure) to more constraint Score of I indicates unlimited IV Polity dataset, downloaded from Intcr-Unjversity Consortium moderate limitations; score of 5 indicates substantial limitations, score of 7 indicates Social Research. Variable described slight to executive parity or subordination. Scores of 2, 4, and 6 indicate intermediate values. in for Political and Gurr (1997). We calculated average values from 2000. inclusive, treating interregnums as missing values. Index of formality in legal procedures for collecting on a bounced check, from 1 Djankov (2003), an extension of the data to 7. Djankov, La Porta, Lopez-de-Silanes, in and Shieifcr (2003). Legal Formalism (Eviction Measure) Log GDP per capita in 1995 (PPP Measure) Average Investment -GDP ratio in 1990s Index of formality in legal procedures for evicting a tenant fornonpayment of rent, from Logarithm of GDP per Ratio of investment to year Credit to the Private Sector (or private credit) As is capita, GDP, on Purchasing Power in currenl prices, I Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003) to 7. World Bank, World Development Parity Basis, in 1995. average from 1990 to 1999, If data are missing for a country in any year, this Penn World Tables version web ignored in calculating the average. a percent of GDP in 1998: financial resources provided to the private sector, such as through loans, purchases of at 6.I., Indicators, on-line version, February 2003. Heston, Suimners. and Aten (2002), d ht^:/Avebhost.bridgcw.cdu/baterv' World Bank (2003), Worid Development Indicators, on-line version, Febniaiy 2003. nonequity securities, and trade credits and other accounts receivable, that establish a claim for repayment For some countries these claims include credit to public enterprises. Stock Market Capitalization Market value of all traded stocks as a percent of GDP, average over 1990-95 web Beck, Demirgii^-Kunt, and Levine(2003a); data from the at http://www.woridbank.org/rcsearch/bios/tbeck/LPF.xls. Procedures for Registering a New Bi Number of separate procedures needed to register a new Djankov (2003), an extension of the data business in Djankov, La Porta, Lopez-de-Silanes, in Djankov. La Porta, Lopez-de-Silanes, and Shleifer (2002). Cost of Registering a New Business Alternative Measure of Credit to the Private Sector Cost of registering a Credit by deposit new business, as a percent of money banks and GDP per capita Djankov (2003), an extension of the data and Shleifer (2002). other financial institutions to the private sector as share of GDP, average over 1980-95. web Beck, Demirguij-Kunt, and Levine (2003a); data from the at http://w\vw worldbank.org/research/bios/tbeck/LPF xls M2 M2 monetary aggregate, as percent of GDP, average over 1 Beck, Demirgui;-Kunt, and Levine (2001 980-95. h ttp Liquid Liabilities Ratio of cunency plus demand and interest-bearing liabilities of banks and nonbank financial intermdiaries to GDP, average over 1980-95. Alternative Measure of Stock Market Capitalization : // ), data from the www w ori dbank .org/research/proj ec ts/ . fi web at nstruc ture/datab ase web Beck, Demirgu^-Kunt, and Levine (2003a); data from the . htm at http ://wwT.v.wori dbank.org/reseaTch/bios/tbeck/LPF xls , Ratio of total value of outstanding traded shares to GDP, average over 1980-95. web at Beck, Demirguij-Kunt, and Levine (2003a); data from the http://www.worldbank.org/research/bios/tbeck/LPF.xls Overall Size of Financial System Credit lo the private sector, as percent of GDP, in 1998 plus stock market capitalization, average over 1990-95 Credit from World Bank (2003), World Development Indicators, on-line version, February 2003. Stock market capitalization from Beck et web Overall Size of Financial System, Alternative Deposit money bank assets plus stock market capitalization as share of GDP, averages over 1980-95. Measure Religion Variables at al (2003a); data from the http://\vww. worldbank.org/research/bios/lbeck/LPF.xls Beck, Demirgu^-Kunt, and Levine (2001). Data from Ihe web at http://www.woridbank.org/research/projects/finstiuctute/database.htm Percentage of the population that belonged religions; Roman Catholic, Protestant, in 1 980 (or for 1 990-95 for countries formed more recently) to the following Muslim, and "other". La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1999) Description Variable Absolute value of the latitude of the country, scaled to lake values between Oand Latitude Log Log of average annual Inflation inHation in the Consumer Source 1. where is La the equator Porta. Lopez-de-Si lanes, Shleifer, World Bank. World Development Price Index from 1970 to 1998, and Vishny {1999} Indicators, CD-Rom, 1999, as used in Acemoglu. Johnson. Robinson, and Thaicharoen (200?) linglish Legal Origin (or Common Law countries) Coded zero or one One indicates that country was colonized by Bhiain and English legal code was transferred La Porta, Lopez-de-Si lanes, Shleifer, and Vishny (1999), and Djankov, La Porta, Lopcz-de-Si lanes, and Shleifer (2003). French Legal Origin (or Civil Log Sclllei Mortality Law countries) Coded zero or one. One indicates legal code was transferred. Log of estimated monalily mortality is for that country European was colonized by France, Spain, Belgium, Portugal or Germany and French La Porta, Lopez -de-Si lanes. Shleifer. and Vishny (1999), and Djankov, La Porta, Lopcz-de-Silanes, and Shleifer (2003) settlers during the early period of European colonization (before 1850). Settler calculated from the mortality rates of European-bom soldiers, sudors and bishops when stationed in colonies. Acemoglu, Johnson, and Robinson (2001), based on Cunin (1989) and other sources. It measures the effects of local diseases on people without inhenled or acquired immunities. Log Population Density Log of population density in 1500. population density is inhabitants per square kilometer, Acemoglu. Johnson, and Robinson (2002), based on McEvedy and Jones (1978) Table A2 Contracting vs. Property Rights Institutions: Econoinic Growth, 1970-95 (1) (3) (2) (4) Dependent variable 2SLS, with 2SLS, with 2SLS, with Against Risk of Risk of Constraint on Constraint on & Expropriation Check Measure 0.64 1.00 0.59 0.99 (0.20) (0.32) (0.22) (0.38) OLS OLS or & Exec. OLS Eviction Measure Second Stage Regression, with log Constraint on Executive in 0.11 0.25 -0.21 -0.05 (0.25) (0.36) (0.22) (0.38) 0.07 0.19 47 40 OLS 47 40 0.79 0.25 0.75 0.37 (0.21) (0.53) (0.23) (0.42) 1 results. In in -0.41 (0.29) columns 0.29 -0.10 -0.10 -0.22 (0.27) (0.39) (0.25) (0.33) 45 45 1970 to 1995, from the World Bank (2003) 2, 4, 6, and 8 the instruments The measure of legal WDl CD-Rom. in 1 500. we 45 54 use only the excolonies sample. The measure of institutions used both Panels, a The dependent variable is as an independent variable measure" from Djankov, La Porta, Lopez-de-Silanes, and Shieifer (2003). In columns are: in 45 is: in in dummy 1, 3, 5, and 7 variable for whether a coimtry has an English legal origin, as classified Lopez-de-Silanes, and Shieifer (2003) and Djankov (2003); in Panel A, log settler mortality before Panel B, log population density 54 Acemoglu, Johnson, and Robinson (2001); in columns 5 through 8, constraint on the formalism used as an independent variable is: in columns 1-2 and 5-6, the "check measure", from average protection against the risk of expropriation as used 7-8, the "eviction 0.01 0.02 0.21 52 regressions are cross-sectional with one observation per country; GDP per capita from columns 3-4 and 0.21 (0.42) -0.23 executive, averaged over the 1990s, from Polity IV. Djankov (2003); 0.06 (0.16) (0.21) 0.27 4, 0.09 (0.44) -0.30 in parentheses. All 39 0.04 (0.42) 52 39 (0.15) 0.11 OLS through 47 (0.22) Number of Observations columns 0.03 or Second Stage Regression, with log population dens ity as instrument Legal Formalism (Eviction Measure) Standard errors are 0.68 (0.38) -0.10 Legal Formalism (Check Measure) average annual growth in -0.05 (0.15) (0.41) Constraint on Executive in 1.00 -0.38 Panel B: instrument (0.52) (0.20) 47 & Eviction Measure -0.07 0.13 0.29 Exec. OLS (0.15) (0.31) Number of Observations R-squared settler mortality as -0.08 OLS Average Protection Against Risk of Expropriation & Check Measure (0.21) Legal Formalism (Eviction Measure) R-squared (8) 1970-95 Protection Against OLS Legal Formalism (Check Measure) capita. Protection Panel A; Risk of Expropriation (7) (6) 2SLS, with Expropriation Average Protection Against (5) annual average growth rate of GDP per is 1 850 (where mortality is per 1000 per armum with we report OLS by Djankov, La Porta, replacement); and in A3 Table Regressions with Reduced (1) 12) (3) Form (4) for Origin and IV for Property Rights Institutions 161 (5) (71 Dependent vanablc indicated Log QDP per capita in Investmenl- GDP ratio in Credit to the Market Log number pnvate sector capitalization procedures for as pertent of as percent of GDP Cost at the if registering a I GDP new business (percerntofGDPper per capita in business capita) 1995 new registering a m (8) Log (10) (II) (12) Cost of registering a top of"each column Investment- GDP ratio in Credit to the Market Log number pnvate sector capitalization procedures for as percent of as percent of GDP GDP regisienng a new new business (percent of GDP per 1995 1990s GDP 1,20 6,02 039 0,3 -024 -0.45 1.08 5.55 0.28 0.24 -0.27 -0 31 (010) (009) (0,06) (0.15) (0.23) (1-41) (0.08) (0.09) (0.08) (0.19) 1990s business capita) Second Stage Regression Average Protection Against Risic of Expropriation Englisli Legal Origin (024) (1 39) -0 91 -3,16 -0,13 -0,32 0.05 -083 -2.67 -0.06 0.03 -0.27 -0 19 (0.38) (2,36) (0 16) (0 15) (0,14) (0.32) (0.34) (2.18) (0.12) (0.13) (0.14) (0.31) 0.76 0,73 0,70 0,81 053 0.53 090 0.90 0.87 098 083 0.83 (0.32) (0,33) (0,32) (033) (0,32) (0.32) (029) (0.30) (0.29) (0.31) (032) (0.32) 01 First Stage for Proteclior1 English Legal Origin Log Settler Mortality -0.49 -0,54 •0,48 -0,54 072 -0.72 (0 13) (0,13) (0 13) (0,13) (0,12) (0.12) Against Expropriation Log Population Density R-Sqnared in First Stage Number of Observations Standard errors are in 0.27 0,32 0,28 030 0,51 65 65 66 63 48 parentheses All regressions are two stage least squares, cross -sectional data with one obser\'aiion per country; 1995; in columns 2 and 8. investmenl-GDP ratio; in columns 3 and register a business; in 51 48 columns 6 and 12. the cost 9, credit to the pnvate sector as share of GDP. Djankov, La Porta. Lopez -de-Si lanes, and Shleifer (2003) and Djankov (2003) The instruments 12, log population density in 1500. in columns 4 and of registering a business as a percent of GDP per capita. Sources are as arc: in columns 1 in through we -0 34 -0,36 -0.34 -0.34 -036 -0.36 (0.09) (0.09) (0.09) (0.09) (0.09) (0.09) 0.28 0.27 0.28 0.28 0.32 0.35 71 71 72 69 54 use only the excolonies sample 10. stock The dependent market capitalization as share of GDP, previous tables. All columns have a 6. log settler mortality dummy in columns variables are in columns 5 and 1 1 , the log I and 54 7, log GDP per capita numberof procedures required in to variable for whether a country has an linglish legal origin, as classified by before 1850 {where morlalily is per 1000 per annum with replacement), and in columns 7 through Table Contracting vs. Property Rights Institutions: (2) (1) A4 GDP per capita, with alternative instrument (3) (5) (4) Dependent variable is log GDP per capita 2SLS, with 2SLS, with Protect Protect Against Exprop. & Check Against Exprop. OLS Measure OLS 0.64 1.09 0.64 0.97 (0.08) (0.22) (0.09) (0.18) OLS Average Protection Against Risk of Expropriation & in OLS R-Squared 1 (0.21) 500 in First Stage GDP per capita (PPP (0.24) 0.87 0.59 (0.30) 7-8, the "eviction are: a 0.05 -0.26 (0.43) (0.47) -0.36 -0.38 -0.40 -0.40 (0.08) (0.13) (0.13) 0.37 0.15 0.16 Stage for Measure of Legal Formalism (Check or Eviction Measure) -1.95 -1.64 -1.74 -1.59 (0.22) (0.21) (0.21) (0.21) 0.07 -0.01 0.04 -0.02 (0.06) (0.06) (0.06) (0.06) 0.57 49 57 in parentheses. All regressions are cross-sectional and 8 the instruments 0.31 (0.09) 57 adjusted) in 1995, from the 0.27 (0.30) World Bank's in WDI 60 we 50 50 60 use only the excolonies sample. The dependent variable on-hne. The measure of institutions used as an independent variable is: in columns 1 through is log 4, in columns 5 through 8, constraint on the executive, averaged over columns 1-2 and 5-6, the "check measure", from Djankov (2003); in Acemoglu, Johnson, and Robinson (2001); is: in measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003). In columns 1,3,5, and 7 dummy 0.54 0.57 49 with one observation per country; from Polity IV. The measure of legal formalism used as an independent variable columns 3-4 and 4, 6, -0.004 Stage for Measure of Institutions (Protect. Against Expropriation or Constraint on Executive) average protection against the risk of expropriation as used the 1990s, -0.05 (0.11) 0.62 Number of Observations Standard errors are 0.45 (0.16) 0.52 English Legal Origin R-Squared 0.28 (0.11) 0.35 1 0.88 (0.27) -0.002 First in 0.31 (0.07) -0.13 in First Stage Log Population Density 0.88 (0.27) (0.10) 500 Eviction Measure 0.29 0.44 English Legal Origin & Exec. OLS (0.07) (0.18) First in & Check Measure 0.15 0.55 Log Population Density 2SLS, with Constraint on (0.09) Legal Formalism (Eviction Measure) R-Squared 2SLS, with Constraint on or Second Sta ge Regression Constraint on Executive Legal Formalism (Check Measure) (8) 1995 Exec. OLS Eviction Measure (7) (6) in we report OLS results. In columns variable for whether a country has an English legal origin, as classified by Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003) and Djankov (2003); and log population density in 1500 from Acemoglu, Johnson, and Robinson (2002). 2, Table A5 Contracting vs. Property Rights Institutions: Investment-GDP ratio, with alternative instrument (1) (2) (4) (3) Dependent variable 2SLS, with 2SLS, with 2SLS, with Constraint on Constraint on & Exprop. OLS 3.81 5.80 4.07 4.48 (0.63) (1.50) (0.70) (1.26) & Eviction Measure or OLS R-Squarcd in First in 1 -1.34 (1.37) R-Squared in First the investment to 500 columns 1 through 4, -0.54 -1.42 (0.93) (1.78) 0.09 1500 0.59 0.05 -0.26 (0.30) (0.43) (0.47) -0.36 -0.38 -0.40 -0.40 (0.09) (0.08) (0.13) (0.13) 0.37 0.15 0.16 Measure of Legal Formalism (Check or Eviction Measure) -1.95 -1.64 -1.74 -1.59 (0.22) (0.21) (0.21) (0.21) 0.07 -0.01 0.04 -0.02 (0.06) (0.06) (0.06) (0.06) 0.57 49 57 49 regressions arc cross-sectional with one observation per country; ratio, in current prices, on Executive) (0.30) 0.62 57 ( 0.06 Protect. Against Expropriation or Constraint 0.87 0.35 in parentheses. All GDP 1.10 (1.18) 0.43 Stage Standard errors are 1.38 (0.81) Stage for Measure of Institutions English Legal Origin Number of Observations 4.82 (2.02) -0.94 First Stage for in 0.94 (0.57) (0.76) Stage Log Population Density 4.24 (1.77) 1.27 English Legal Origin & Eviction Measure 1.00 (1.21) First Exec. OLS (0.51) 0.71 0.43 Log Population Density & Check Measure (0.68) Legal Formalism (Eviction Measure) in Exec. OLS Second S tagc Regression Constraint on Executive R-squared 1990s Protect Against OLS Legal Formalism (Check Measure) (8) (7) GDP in 2SLS, with Exprop. Against Risk of Expropriation (6) to Protect Against Check Measure OLS Average Protection (5) average ratio of investment is average over 1 0,54 0.57 60 we 60 50 50 use only the excolonies sample. The dependent variable World Tables 6.1. The measure of institutions used as an independent variable is: in used in Accmoglu, Johnson, and Robinson (2001); in columns 5 through 8, constraint on is 990s, from Penn average protection against the risk of expropriation as the The measure of legal formalism used as an independent variable is: in columns 1-2 and 5-6, the "check measure", from Djankov (2003); in columns 3-4 and 7-8, the "eviction measure" from Djankov, La Porta, Lopez-de-Silancs, and Shieifer (2003). In columns 1, 3, 5, and 7 we report OLS results. In columns 2, 4, 6, and 8 the instruments are: dummy variable for whether a country has an English legal origin, as classified by Djankov, La Porta, executive, averaged over the 1990s, from Polity IV. Lopez-de-Silanes, and Shieifer (2003) and Djankov (2003); and log population density in 1500. Table A6 Contracting vs. Property Rights Institutions: Private Credit, with ahemative instrument (3) (2) (1) (4) Dependent variable is (5) credit to the private sector in 2SLS, with Protect & Exprop. Against Exprop. OLS Check Measure OLS 0.13 0.25 0.18 0.24 (0.04) (0.09) (0.04) (0.08) OLS Average Protection Against Risk of Expropriation & or Second in OLS R-Squared 1 -0.08 in First 0.002 -0.02 -0.08 -0.10 (0.05) (0.07) (0.05) (0.07) 0.30 500 500 Stage 0.87 0.59 0.05 -0.26 (0.30) (0.43) (0.47) -0.36 -0.38 -0.40 -0.40 (0.09) (0.08) (0.13) (0.13) 0.15 0.16 0.37 Measure of Legal Formalism (Check or Eviction Measure) -1.95 -1.64 -1.74 -1.59 (0.22) (0.21) (0.21) (0.21) 0.07 -0.01 0.04 -0.02 (0.06) (0.06) (0.06) (0.06) 0.57 49 57 57 of GDP in 1998, from the 60 we 60 as an independent variable average protection against the risk of expropriation as used in Acemoglu, Johnson, and Robinson (2001); and in columns 5 through columns 3-4 and 4, 6, 7-8, the "eviction and 8 the instruments are: a is: in columns 1-2 and Djankov (2003); and log population density for in 1 whether a country has 8, is: The dependent in constraint variable columns 4 through is 8, on the executive, averaged and 5-6, the "check measure", from Djankov (2003); in measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003). In columns dummy variable 50 50 use only the excolonies sample. World Bank. The measure of institutions used over the 1990s, from Polity IV. The measure of legal formalism used as an independent variable 0.54 0.57 49 Standard errors are in parentheses. All regressions are cross-sectional with one observation per country; credit to the private sector as a percent 0.13 (0.30) 0.62 Number of Observations 0.23 Stage for Measure of Institutions (Protect. Against Expropriation or Constraint on Executive) English Legal Origin R-Squared 0.18 (0.08) (0.06) First Stage for 1 0.06 (0.03) -0.11 0.35 in 0.17 (0.07) (0.04) in First Stage Log Population Density 0.06 0.01 English Legal Origin in & Eviction Measure (0.02) (0.07) First Exec. OLS Regression Stiage -0.08 0.36 Log Population Density & Check Measure (0.04) Legal Formalism (Eviction Measure) R-squared 2SLS, with Constraint on Exec. Constraint on Executive Legal Formalism (Check Measure) 2SLS, with Constraint on OLS Eviction Measure (8) 1998 as a percent of GDP 2SLS, with Protect Against (V) (6) 1, 3, 5, and 7 we report OLS results. In columns 2, an English legal origin, as classified by Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003 500 from Acemoglu, Johnson, and Robinson (2002). Table A7 Contracting vs. Property Rights Institutions: Stock Market Capitalization, with alternative instrument (2) (1) (3) Dependent variable is Against Risk of Expropriation 2SLS, with 2SLS,with 2SLS,with Protect Against Constraint on Constraint on & Exprop. OLS Check Measure 0.10 0.21 (0.04) (0.09) & Eviction Measure in OLS 0.22 (0.05) (0.09) R-Squared in First in First 1500 Stage Number of Observations -0.03 -0.11 -0.09 -0.17 (0.08) (0.05) (0.08) 0.26 Measure of Institutions (Protect. Against Expropriation or Constraint the level of stock market capitalization as a percent of 1 through 4, 0.02 -0.31 (0.31) (0.30) (0.44) (0.47) -0.37 -0.38 -0.39 -0.38 (0.09) (0.08) (0.13) (0.13) 0.37 0.14 0.16 Stage for Measure of Legal Formalism Check or Evict ion Measure) -1.99 -1.71 -1.78 -1.65 (0.21) (0.20) (0.20) (0.20) 0.08 0.008 0.05 -0.006 (0.06) (0.06) (0.06) (0.06) 0.64 0.62 0.59 48 48 59 GDP in 1990-95, from Beck et al (2003). we The measure of institutions used report OLS results. Porta, Lopez-de-Silanes, In in columns 3-4 and columns 2, 4, 6, 7-8, the "eviction and 8 the instruments 49 as an independent variable average protection against the risk of expropriation as used in .^cemoglu, Johnson, and Robinson (2001); and measure", from Djankov (2003); we 0.59 49 59 use only the excolonies sample. The dependent variable on the executive, averaged over the 1990s, from Polity IV. The measure of legal formalism used as an independent variable 7 on Executive) 0.60 56 56 0.10 0.89 Standard errors are in parentheses. All regressions are cross-sectional with one observation per country; columns 0.15 (0.09) (0.06) 0.24 English Legal Origin R- Squared 0.05 (0.03) -0.14 0.36 in 0.13 (0.07) (0.05) First Log Population Density 0.04 -0.15 Stage Eviction Measure (0.03) (0.04) 500 & Exec. OLS or Second Stage Regression 0.17 -0.08 First Stage for 1 Check (0.07) English Legal Origin in & Measure -0.13 0.38 Log Population Density Exec. OLS (0.04) Legal Formalism (Eviction Measure) R-squared ofGDP 2SLS, with Constraint on Executive Legal Formalism (Check Measure) (8) (7) slock market capitali -Mtion, average 1990-95. as a percent OLS Average Protection (6) Protect Against Exprop. OLS (5) (4) is: in in columns 5 through columns 1-2 is: 8, constraint and 5-6, the "check measure" from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2003). In columns are; a dummy 1, 3, 5, variable for whether a country has an English legal origin, as classified by Djankov, and Shleifer (2003) and Djankov (2003); and log population density in is in 1500 from Acemoglu, Johnson, and Robinson (2002). and La Figure 1 Credit market collapses; C^ Legal costs of higher P does not affect .C\P) investment contract C^ enforcement investment only affects when cross from one region to the other Credit market exists; higher P increases investment Political constraints, P Figure 2 GMB NZL USA CAN USA CAN NZL SGP SGP AUS GAB GAB IDN «* TO AUS iND GMB CMR JAM <="* HttOc PNG KEN g6f AGO NIC cffl" o^" miR SUR PNG •»^gAM TZA CUR 'K&g GIN NER KEN LKA COG T2A -—^ ^- MU GHA IS ^_ HND —-.^NGA MDG BFA BCD NIC LKA ZAP PAK ^TO3 CDa BGO "t^ MJ HTI SUR ZAR LM3A ^, Z*R SDN SDN -r 2 Residuals from Regressing Log Settler Mort. on Eng. Leg. Orig. 5 Residuals from Regressing Eng, Leg, Ohg on Log Settler Mort. - ^ ARg"' TZA ^AN ^ '^° BGD NGA ETH KEN USA SGP __^L-- —-^ NIC E«D GHA ^__SEN_„ BRB gQY DOM " MU ew NER MDG CAN TifiRV DOM AUS ZAF NZL EGY CIV BLZ ETH E e VNM BRA HKG 2 Residuals from Regressing Log, Sett. Mort on Eng. Leg. ,5 Ong Residuals from Regressing Eng. Leg. Ong on Log. Sett. Mort NGA Figure 3 4PNG MDG 2 - JAM TTO MLt IND MUS WPOM BEN *^^HTI CAF ^^ "^^ 0!<fl!i um P^ GMB GHA °"^ LKA RGQ rjl C" NGA TZA seR»»»g,n -2 - DZA ETM E OJI SGP ZAR AFG SDN -4 - 2 Residuals from Regressing Log Settler Mort- on Eng. Leg. Orig. .5 Residuals from Regressing Eng. Leg. Orig. on Log Settler Mod, 2 - PWJ Gt*N LKA ZAF PAWfJ^ IND BGD USA NGA MS*R EttfO KEN - SHA DOM UGA SLV BRB CAN TZA JAM AUS 1 lEN-rzA CPV EGY - CAN BRaW UGA AUS ^ TTO NZL BL2 VNM 2 1 2 Residuals from Regressing Log. Sett. Mort. on Eng. Leg. Orig 1 .5 Residuals from Regressing Eng. Leg. Orig on Log. Sett. Mort. Figure 4 USA CAN AUS NZL SGP 8RA GAB CHL ION ~~~~^~~' m ^ GMB MKG BXiA^ ^^--^ OIHS ^t'ftv G'^ viftfA M-et,,^ ARG fND njN "^OL^^^^^^^—.....ETH GHA~—---...^j. 'ms- ^=*'-bnA_PAK KEN r^^lW GUY SLE COG MOG SUR MMR BGD MU HTI UGA ZAR a: "%J, SDN E S SOM "1 -2 Residuals from Reg. Log Pop Density in 2 1500 on Eng. Leg. Ong. 1 1 -.5 .5 Residuals from Regressing Eng. Leg. Ong. on Log Pop. Density in O a ABG SWAN BWA "^V PAN B0(. swz m ™ CRI VCT o o E -2 Residuals from Regressing Log. Pop. Density 2 In 1500 on Eng. Leg. Ong .5 Res. from Reg. Eng. Leg. Orig on Log. Pop. Density in 1500 1500 u - '^ J MOV 1 1 2003 Date Due MIT LIBHARIES i'" I !' •"n\ ..ii 3 9080 02613 1323