Why some states fail: the role of culture.

Shikida, Claudio D. ; de Araujo, Ari Francisco, Jr. ; SantAnna, Pedro H.C. 等

There are many studies on the relationship between economic development and institutions. Institutions can be classified as formal or informal. This article emphasizes the importance of the relationship between culture (informal institutions) and the quality of public goods supplied by the government, using a measure of state failure: the Failed States Index. The results suggest that culture is more important than formal institutions in explaining differences in the degree to which states fail.

The Importance of Institutions

In 2008, the World Public Organization (WPO) carried out a poll indicating a decrease in people's optimism regarding the market and an increase in their belief in the need for stronger state regulation. That popular sentiment was widespread across countries, including Latin America, Germany, Italy, Nigeria, the Philippines, Turkey, and South Korea (WPO 2008). Do those personal beliefs and values influence development beyond the traditional variables of Solow's economic growth model? Are ideas and values important for economic development, or are they only consequences of it?

In this article, we study the relationship between institutions and governments. Our objective is to understand the impact of formal and informal institutions on the quality of states--and thus to better appreciate how institutions help shape public policies and affect economic development. We survey the most recent work on the role of institutions and then present and test our model.

Markets, Culture, and Development

In 1936, Leacock criticized the cultural values that Adam Smith admired. Having witnessed the Great Depression, Leacock clamored for a cultural change: Would it not be the time to convince people to alter their beliefs about the operation of free markets?

In less poetic terms, North (1989), among others, has emphasized the importance of understanding the influence of formal and informal institutions on economic development. One of the most important questions has concerned the role of property rights in the process of development (de Soto 2000).

Is it possible to distinguish between formal and informal institutions? The laws of a country can represent the codification of a significant part of its informal institutions, and, following the literature, one can label formal institutions as "laws" for operational purposes. This distinction, however, does not deny the existence of informal institutions, under the label of "culture," including people's values, behavioral norms, and traditions (Williamson and Kerekes 2009), all of which, in many cases, play an important role in determining market results. (1)

Guiso, Sapienza, and Zingales (2006: 23) define the informal institutions that embody culture as "those customary beliefs and values that ethnic, religious, and social groups transmit fairly unchanged from generation to generation." Boettke, Coyne, and Leeson (2008: 338) use the Greek term "metis" to convey culture:

   Metis is characterized by local knowledge resulting from practical
   experience. It includes skills, culture, norms, and conventions,
   which are shaped by the experiences of the individual. This concept
   applies to both interactions between people (e.g., interpreting the
   gestures and actions of others) and the physical environment (e.g.,
   learning to ride a bike).

Such definitions suggest that one proxy for informal institutions may be culture (or metis), translated into some type of value or initial belief that influences the choice among formal institutions. Empirically, the impact of formal and informal institutions on economic development is affected by government quality (La Porta, Lopez-de-Silanes, and Shleifer 1999, 2008). (2) Guiso, Sapienza, and Zingales (2006) show that the cultural factor should not be underestimated: They find that preferences regarding redistributive policies affect the policies themselves. In other words, public policies can result from cultural factors. (3)

The literature reviews by Guiso, Sapienza, and Zingales (2006) and La Porta, Lopez-de-Silanes, and Shleifer (2008) present evidence that institutions are important for the economy, although it is not possible to affirm that there is consensus regarding the mechanisms of transmission of such (formal and informal) institutions to economic development. This article does not seek to resolve that issue; rather, we revisit the problem of the influence of institutions on the quality of governance.

Model and Data

The importance of culture in determining state quality can be expressed as

(1) [Y.sub.i] = [alpha] + [[beta].sub.1][F.sub.i] + [[beta].sub.2][I.sub.i] + [summation over (i)] [[delta].sub.i] [X.sub.i] + [[epsilon].sub.i]

where Y is the index of state failure, F represents formal institutions and I informal institutions, X is a vector of control variables, and [epsilon] is an error term (i.e., "white noise").

Definition of Variables

We now give a detailed explanation of each variable.

The Failed States Index. To operationalize the concept of "socioeconomic arrangement," we use the 2007 edition of the Failed States Index, published by the Fund for Peace. The Failed States Index includes 177 states, and has been published since 2005. Using its own methodology (the Conflict Assessment System Tool or CAST), the Fund for Peace examines the social, economic, and political dimensions of a state by relying on news items. The social dimension is captured by four indicators, the economic dimension by two, and the social dimension by six. Values based on an ordinal scale of 0 to 10 are attributed to each indicator and then aggregated, creating a "tendency index." That index is used to estimate the risk of occurrence of conflicts in the country. In other words, the Failed States Index measures the deterioration (failure) of the capacity of a state or government to resolve peacefully the problems of a society--that is, to supply basic public goods such as security and stability for various social groups in their economic and noneconomic transactions. In the measurement of this index, states such as Iraq (113.7) and Sudan (111.4) were, in 2007, at the top of the ranking (most failed states), whereas Finland (18.5) and Norway (17.1) were, respectively, the 176th and 177th (last ones) of the list. (4)

Formal Institutions. According to Williamson and Kerekes (2009) and Glaeser et al. (2004), formal institutions can be defined as the political constraints to which governments are subject. We use the measurements collected by Glaeser et al. (2004):

* Judicial Independence is the sum of three variables, measured about 1995, with regard to the tenure of Supreme Court judges (or judges on the highest court) of each country, and, for administrative cases, judges of last instance courts which conterns the existence or not of ease law. The aggregated variable is normalized, assuming values between 0 and 1, with higher values meaning greater judicial independence. This variable was originally measured by La Porta et al. (2004).

* Proportional Representation is a dummy variable that takes a value of 1 when there is a proportional representation system. We use the average value for 1975-2000. This variable was originally measured by Beck et al. (2001).

* Constitutional Review is the sum of two variables measured about 1995. One of them concerns the power of judges to review the constitutionality of the laws of a country; the other concerns how hard it is to change the constitution in the same country. This sum is normalized between 0 and 1, with higher values meaning greater power of constitutional review by the courts. This variable was originally measured by La Porta et al. (2004).

* Plurality is a dummy variable that takes a value of i when there is an electoral system that follows a winner-take-all rule: the winner (by simple or relative majority) controls all the chairs disputed in the election. We use the average value for 1975-2000. This variable was originally measured by Beck et al. (2001).

Informal Institutions. The choice of the variables that represent informal institutions follows Tabellini (2008a) and Williamson and Kerekes (2009). Such information is obtained from the 1995-1997 and 1999-2000 reports of the World Values Survey. Four aspects of the culture (or the average belief) of a country are considered: control, respect, trust, and obedience. Obedience represents the percentage of people in the sample who mentioned obedience as an important factor in society. Trust and respect, in a similar fashion, represent positive answers ("most people can be trusted" and "[most people show] tolerance and respect towards others") to these values, in the sample. Finally, control is the unconditional average response to the question about how much freedom of choice and control people have over their lives. The variable "culture" is then obtained by summing the values of control, respect, and trust, and then subtracting the value of obedience.

In what sense can culture be considered a proxy for liberal individualism? Initially, let us look at the negative belief "obedience." Consider the following passage:

One might well ask, if an individual is born with the obligation to obey, who is born with the right to command? If one wants a coherent theory of obligations, there must be someone, whether an individual or a group, with the right to the fulfillment of the obligation. If I am constituted as a person by my obligation to obey, who is constituted as a person by the right to obedience? Such a theory of obligation may have been coherent in an age of God-kings, but it seems rather out of place in the modern world [Palmer 2008].

Note that it is not about questioning the value of obedience as it exists, for example, in the form of informal agreements (e.g., family hierarchy) or formal agreements (e.g., commercial transactions governed by laws). The anti-individualist aspect of obedience is related to a culture in which experimentation and criticism are tightly restricted. In this sense, obedience identifies a conservative characteristic, not a liberal one (see Hayek 1960: postscript; Postrel 1998: chap. 5).

In contrast, respect, confidence, and control (over one's own life) are qualities cited as important in the development of exchanges, be they monetary or otherwise. Normally, discussion of the impact of those values on society involves the distinction between "'bonding" (or exclusive) social capital and "'bridging" (or inclusive) social capital. The former is the social capital that exists in groups of people who share some type of similarity, such as families or members of the same religious community, while the latter occurs in more diverse groups of people, such as the ones that involve economic exchanges. In a complex society, respect and confidence are important elements for the effectiveness of the market mechanism:

The development of bridging social capital ... necessarily involves a thin set of morals, such as tolerance of others, the observance of contracts and respect for private property, that can be shared by actors with otherwise diverse and perhaps even conflicting moral codes [Meadowcroft and Pennington 2008].

Finally, the variable "control" is one of the most obvious dimensions of individual freedom. It expresses, according to Tabellini (2008b), the idea that individual effort is rewarding.

These variables, in the form postulated by Williamson and Kerekes (2009), would tend to have a negative impact on state failure, because they would improve the quality of governance in the provision of public goods and reduce discrimination.

Control Variables. We follow Williamson and Kerekes (2009) in using the following control variables:

* Gdpgrowth_9000 is real GDP per capita growth, adjusted for purchasing power parity and measured in U.S. dollars (base year = 2000), averaged over 1990-2000.

* Yearsschooling_60 is the logarithm of the number of years of schooling of the population above 25 years of age, measured in 1960. This variable was obtained from La Porta, Lopez-de-Silanes, and Shleifer (1999).

* Urbanpop_9000 is the percentage of urban population, averaged over 1990-2000. This variable was obtained from the World Bank's World Development Indicators (WDI) database.

* Govcons_902000 is real government consumption expenditures, measured as a percentage of the GDP, averaged over 1990-2000. This variable was obtained from the World Bank's WDI database.

The Endogeneity Problem

Although there are theoretical arguments for the supposition that the variables that represent formal and informal institutions are exogenous, they may be endogenous in our exercise. In this ease, the correlation between the variables that capture the effect of institutions on state failure and the errors ([epsilon]) is different from zero. Acknowledging the possible endogeneity problem (reverse causality, measurement error, and omitted variables), we applied the two-stage least squares (2SLS) method with instrumental variables. Thus, the first-stage equations are:

(2) [I.sub.i] = [[phi].sub.1] + [[phi].sub.2][W.sub.i] + [u.sub.i]

(3) [F.sub.i] = [[psi].sub.1] + [[psi].sub.2][Z.sub.i] + [v.sub.i]

Following Williamson and Kerekes (2009), the origin of the legal code is the instrument (Z) for formal institutions (F), whereas latitude is used as an instrument (W) for the informal part of institutions (I), i.e., culture. It is expected that such instruments are sufficiently adequate to isolate the channel through which both formal and informal institutions affect state failure. In other words, for Williamson and Kerekes (2009), Z and W are correlated with F and I (respectively), but are not correlated with [epsilon]. (5) Latitude and legal code can be described as follows:

* LAT_ABST is latitude of the country, measured in absolute value. The closer to 0 the value is, the closer the country is to the equator. This variable was obtained from La Porta, Lopez-de-Silanes, and Shleifer (1999).

* LEGOR_UK is a dummy variable that takes a value of 1 if the country's legal code is of English origin. This variable was obtained from La Porta, Lopez-de-Silanes, and Shleifer (1999).

Once (2) and (3) are estimated, it is possible to compute the following second-stage equation:

(4) [Y.sub.i] = [alpha] + [[beta].sub.1][[??].sub.i] + [[beta].sub.2][[??].sub.i] + [summation over (i)] [[delta].sub.i][X.sub.i] + [[epsilon].sub.i]

where the vector X includes the control variables cited earlier.

We performed the Durbin-Wu-Hausman endogeneity test in the regressions estimated with instrumental variables, as described in Hayashi (2000). The null hypothesis of the test states that the ordinary least squares (OLS) estimators of the same equation would generate consistent results--that is, the endogeneity of the regressors for formal and informal institutions would not have deleterious effects on OLS estimates. Therefore, the rejection of the null hypothesis suggests that the regressors are endogenous, and that instrumental variables would be the most adequate method. The data have a chi-squared distribution with m degrees of freedom, where m is the number of regressors specified as endogenous in the regression with instrumental variables.

Results

In Table 1 we present the results of the simple and multiple regressions estimated by OLS. We observe, in the simple regressions (1 through 5), that the only variable that is not significant (at the 10 percent level) is constitutional review (negative sign). The parameter estimated for culture suggests a negative impact, as expected. The coefficients for judicial independence, proportional representation, and constitutional review all have the expected negative sign.

In columns 6 through 9 of Table 1 we present the results of the regressions in which the variable culture is maintained, and the other variables that represent formal institutions are added one by one. We observe that, in all cases, only culture has a negative and significant impact on the Failed States Index--that is, informal institutions improve the quality of state governance. Our results suggest that the inclusion of variables that represent formal institutions does not help explain any additional variation of state failure.

Table 2 (regressions 10 through 13) replicates regressions 6 through 9 with the introduction of traditional control variables. The results are very robust qualitatively. The variable culture remains significant (at the 5 percent level) and has the expected negative sign, whereas the variables that capture the influence of formal institutions on the Failed States Index do not have levels of significance statistically higher than the conventional ones. The same result occurs with the estimated parameters of the control variables urban population and consumption. The variables schooling and growth are negative and significant, as expected. Due to the possible presence of endogenous regressors, such results must be observed with caution, as they may be biased.

As previously described, we used the 2SLS with the endogeneity test. The results are presented in Tables 3 and 4--univariate regressions (14 through 16), bivariate regressions (17 and 18), and regressions with additional controls (19 and 20). It is worth noting that in only three regressions (all of which lacked controls) the tests suggest that the regressors are endogenous. In other words, endogeneity of formal and informal institutions does not present, in a general way, deleterious effects on OLS estimates. In any case, in the univariate regressions, the estimated parameter for culture suggests, just as previously, a negative and significant impact (at the 1 percent level of significance). The estimated parameters for judicial independence and proportional representation are not significant.

In the bivariate results, culture (informal institutions) is highly significant and exerts negative influence on the Failed States Index. The same result does not occur with formal institutions. After the inclusion of control variables (Table 4), the results remain the same.

In sum, using either OLS or 2SLS estimates, only culture has a negative and significant effect on the Failed States Index--that is, informal institutions appear to prevail over formal institutions in improving the quality of state governance.

Conclusion

Economic history is the natural laboratory for those who try to understand the determinants of economic development. The impacts of such formal and informal incentives are different, and, therefore, the development policy recommendations, if necessary, are not the same. It may be more difficult to understand how the cultural aspects of a society change, and how they influence economic development, than to study the economic impacts of legal changes in the same society.

In this article, we have sought to investigate the role of formal and informal institutions on the quality of governance. The results suggest that informal institutions (culture) are more important than formal institutions. In other words, a more pro-market culture is one of the determinants of better governance. That fact can be regarded as evidence that more open societies tend to produce governments that are more efficient in the provision of public goods.

There are many aspects to be explored in the analysis of the role of institutions in human actions, as well as in how such actions are reflected in economic outcomes. In this sense, the evidence presented in this article shows that policymakers should pay attention to the impact of different institutions (mostly to those of informal ones) on state quality.

References

Acemoglu, D., and Johnson, S. (2005) "Unbundling Institutions." Journal of Political Economy 113 (5): 949-95.

Acemoglu, D.; Ticchi, D.; and Vindigni, A. (2007)"Emergence and Persistence of Inefficient States." Working Paper, Massachusetts Institute of Technology. Available at http://econ-www.mit.edu/ files/2215 (December).

Beck, T. G.; Clarke, A.; Groff, P.; Keefer, P.; and Walsh, P. (2001) "New Tools in Comparative Political Economy: The Database of Political Institutions." World Bank Economic Review 15 (1): 165-76.

Berkowitz, D.; Pistor, K.; and Richard, J. F. (2003) "Economic Development, Legality, and the Transplant Effect." European Economic Review 47 (1): 165-95.

Boettke, P. J.; Coyne, C. J.; and Leeson, P. T. (2008) "Institutional Stickness and the New Development Economics." American Journal of Economics and Sociology 67 (2): 331-58.

De Soto, H. (2000) The Mystery of Capital. New York: Basic Books.

Dixit, A. (2007) "Evaluating Recipes for Development Success." World Bank Research Observer 22 (2):131-57.

Easterly, W., and Levine, R. (2003) "Tropics, Germs, and Crops: How Endowments Influence Economic Development." Journal of Monetary Economics 50: 3-40.

Faria, J. R. (1999) "Is There an Optimal Constitution?" Constitutional Political Economy 10 (2): 177-84.

Fund for Peace (2007) The Failed State Index, 2007. Available at www.fundforpeace.org.

Glaeser, E.; La Porta, R.; Lopez-De-Silanes, F.; and Shleifer, A. (2004) "Do Institutions Cause Growth?" Journal of Economic Growth 9: 271-303.

Greif, A. (2005) Institutions: Theory and History. Comparative and Historical Institutional Analysis. Cambridge: Cambridge University Press.

Guiso, L.; Sapienza, P.; and Zingales, L. (2006) "Does Culture Affect Economic Outcomes?" Journal of Economic Perspectives 20 (2): 23-48.

Hayashi, F. (2000) Econometrics. Princeton, N.J.: Princeton University Press.

Hayek, F. A. (1960). "Why I Am Not a Conservative.'" In The Constitution of Liberty. Chicago: University of Chicago Press.

La Porta, R.; Lopez-de-Silanes, F.; and Shleifer, A. (1999) "The Quality of Government." Journal of Law, Economics and Organization 15 (1): 222-79.

La Porta, R.; Lopez-de-Silanes, F.; Pop-Eleches, C.; and Shleifer, A. (2004) "Judicial Checks and Balances." Journal of Political Economy 112: 445-69.

La Porta, R.; Lopez-de-Silanes, F.; Shleifer, A.; and Vishny, R. (2008) "The Economic Consequences of Legal Origins." Journal of Economic Literature 46 (2): 285-332.

Leacock, S. (1936) Hellements of Hickonomics. New York: Dodd, Mead.

Meadowcroft, J., and Pennington, M. (2008) "Bonding and Bridging: Social Capital and the Communitarian Critique of Liberal Markets." Review of Austrian Economics 21 (January): 119-33.

North, D. C. (1989) Structure and Change in Economic History. New York: W. W. Norton.

Palmner, T. G. (2008) "Myths of Individualism." Cato Policy Report (September/October). Available at www.cato.org/pubs/policy_report/cpr-18n5- 1.html.

Pejovich, S. (2003). "Understanding the Transaction Costs of Transition: It's the Culture, Stupid." Review of Austrian Economics 16 (4): 347-61.

Postrel, V. (1998) The Future and Its Enemies. New York: Simon and Schuster.

Tabellini, G. (2008a) "The Scope of Cooperation: Values and Incentives." Quarterly Journal of Economics 123 (3): 905-50.

--(2008b) "Institutions and Culture." Journal of the European Economic Association 6 (2-3): 255-94.

Williamson, C. R., and Kerekes, C. B. (2009) "Securing Private Property: The Relative Importance of Formal versus Informal Institutions." Available at www.claudiawilliamson.com.

World Public Organization (WPO) (2008) "Erosion of Support for Free Market System: Global Poll." Available at www.worldpublicopinion.org/pipa/articles/home_page/471. php ?lb=hmpg2&pnt=471&nid=&id=>.

(1) Berkowitz, Pistor, and Richard (2003) detail, empirically, the importance of the way a country adopts a law for its economic development. Faria (1999) shows that, in a model of constitutional choice, it is possible to obtain two equilibriums: one that seeks to restrict political power, the other that seeks to foster civic virtues. The importance of culture first appeared in the theoretical literature with game theory (see, for example, Greif 2005). Tabellini (2008b) offers a theoretical model to explain the importance of cultural values for economic development.

(2) For a discussion of the determinants of economic development, and the relevance of formal and informal institutions, see Easterly and Levine (2003), Pejovich (2003), Acemoglu and Johnson (2005), Dixit (2007), and Tabellini (2008b).

(3) It is not by chance that modern socialism gives so much importance to the informal aspect of institutions. Gramsci's ideas are a good example. See Guido, Sapienza, and Zingales (2006) for a survey of studies on the relationship between the economy and culture.

(4) Acemoglu, Tiechi, and Vindigini (2007) present a model that seeks to explain the appearance and persistence of "Inefficient States." Their operational definition is not the same as for the Failed States Index, but their theoretical idea is quite similar.

(5) Williamson and Kerekes (2009) cite pertinent literature (for example, Acemoglu and Johnson 2005 and Easterly and Levine 2003) to defend latitude and legal code as valid instruments.

Claudio D. Shikida is Associate Professor of Economics at Ibmec Minas Gerais, Brazil; Ari Francisco Araujo Jr. is Assistant Professor of Economics at Ibmec Minas Gerais; and Pedro H. C. Sant'Anna is a candidate for M.A. in Economics at the Universidad Carlos III in Madrid. The authors thank Peter Boettke and Christopher Coyne for comments on a previous version of this article. They also thank discussants at the 5th Seminar of Economics of Belo Horizonte and at the 2008 Economic Freedom of the World Network Conference.

TABLE 1
INITIAL REGRESSIONS

                      Dependent Variable: Failed States Index

                                 Univariate Results

                         (1)             (2)             (3)

Culture              -0.5055 ***
                      (0.0588)
Judicial                             -20.2176 **
  Independence                        (8.7730)
Proportional                                        -16.6443 ***
  Representation                                      (3.6591)
Constitutional
  Review
Plurality

Constant            121.6618 ***     77.0137 ***     78.2252 ***
                      (7.5891)        (6.1533)        (2.5471)

[R.sup.2]               0.40            0.05            0.12
Number of
  Observations           75              68              149

                          Dependent Variable: Failed States Index

                       Univariate Results          Bivariate Results

                        (4)            (5)             (6)

Culture                                             -0.5235 ***
                                                    (0.0704)
Judicial                                            -12.3570
  Independence                                      (11.2669)
Proportional
  Representation
Constitutional        -11.6974
  Review             (13.0903)
Plurality                          14.9974 ***
                                     (4.4780)
Constant            69.1447 ***    59.1608 ***    132.2264 ***
                      (8.0628)       (3.8091)       (10.2474)

[R.sup.2]               0.01           0.08           0.47
Number of
  Observations           70            156             47

                       Dependent Variable: Failed States Index

                                Bivariate Results

                         (7)             (8)             (9)

Culture              -0.5187 ***     -0.5488 ***     -0.5168 ***
                      (0.0538)        (0.0668)        (0.0580)
Judicial
  Independence
Proportional           -8.5620
  Representation      (6.6692)
Constitutional                        -15.5072
  Review                              (12.2704)
Plurality                                              5.3841
                                                      (5.0571)
Constant            128.5368 ***    135.1527 ***    118.8696 ***
                      (7.8196)        (12.4242)       (9.2537)

[R.sup.2]               0.47            0.47            0.45
Number of
  Observations           72              47              73

NOTES: Standard error in parentheses. Level of significance: *** at 1
percent, ** at 5 percent, * at 10 percent.

TABLE 2
ORDINARY LEAST SQUARES WITH CONTROLS

                                 Dependent Variable: Failed
                                       States Index

                                    (10)             (11)

Culture                          -0.2187 **       -0.2363 **
                                  (0.1052)         (0.1020)
Judicial Independence             10.3677
                                  (9.5657)
Proportional Representation                         1.0198
                                                   (5.1937)
Constitutional Review

Plurality

GDP Growth Log                 -510.4539 ***      -373.619 *
                                  (183.64)        (193.3465)
Schooling in 1960 Log           -20.7263  **     -17.6328 ***
                                  (4.3222)         (4.3116)
Urban Population                   7.8733          -1.4706
                                 (19.3486)        (18.9353)
Government Consumption            -16.1779         -27.9581
                                 (107.0260)       (97.7202)
Constant                        103.075 ***      115.6419 ***
                                 (18.9225)        (16.3739)

[R.sup.2]                           0.71             0.70
Number of Observations               39               43

                                 Dependent Variable: Failed
                                        States Index

                                    (12)             (13)

Culture                          -0.2168 **       -0.2435 **
                                  (0.1016)         (0.1004)
Judicial Independence

Proportional Representation

Constitutional Review             -8.5649
                                 (12.1122)
Plurality                                          -3.8941
                                                   (4.2747)
GDP Growth Log                   -433.20 **       -386.84 **
                                 (203.1916)       (186.0939)
Schooling in 1960 Log           -19.6902 ***     -16.8812 ***
                                  (4.7249)         (4.1599)
Urban Population                   2.6209          -1.9603
                                 (17.1561)        (17.6447)
Government Consumption            -15.5549         -41.7796
                                 (100.9759)       (92.0209)
Constant                        117.014 ***      121.423 ***
                                 (14.3204)        (14.2025)

[R.sup.2]                           0.71             0.70
Number of Observations               39               43

NOTES: Standard error in parentheses. Level of significance: *** at 1
percent, ** at 5 percent, * at 10 percent.

TABLE 3
TWO-STAGE LEAST SQUARES WITH INSTRUMENTAL VARIABLES

                                        Failed States Index
                                        Univariate Results

                                  (14)           (15)          (16)

Culture                        -0.7286 ***
                                (0.1452)
Judicial Independence                          17.1493
                                              (22.1389)
Proportional Representation                                  -12.4806
                                                             (9.9576)
Constant                      150.0997 ***   49.1434 ***   75.8858 ***
                                (18.8990)     (16.9189)      (6.0483)
Endogeneity Test                  3.209         4.315         0.225
P-value (Endogeneity Test)       0.0733         0.0378        0.6351

Centered [R.sup.2]                0.32          -0.12          0.11
Number of Observations             75             68           149

                                  Failed States Index

                                   Bivariate Results

                                  (17)           (18)

Culture                        -0.7705 ***    -0.7258 ***
                                (0.1501)       (0.1512)
Judicial Independence            -7.6022
                                (19.937)
Proportional Representation                     3.7337
                                               (14.9975)
Constant                      161.0058 ***    145.4589 **
                                (20.3076)      (19.2148)
Endogeneity Test                  4.622          2.461
P-value (Endogeneity Test)       0.0992         0.2922

Centered [R.sup.2]                0.37           0.38
Number of Observations             47             72

NOTES: Standard error in parentheses. Level of significance: *** at 1
percent, ** at 5 percent, * at 10 percent.
Instruments: Latitude and English Legal Code.

TABLE 4
Two-STAGE LEAST SQUARES WITH INSTRUMENTS AND
CONTROLS

                               Dependent Variable:
                               Failed States Index

                                   (19)            (20)

Culture                          -0.9359 *       -0.8663 *
                                 (0.5127)        (0.4632)
Judicial Independence            -21.9763
                                 (47.3318)
Proportional Representation                       9.3037
                                                 (18.3839)
GDP Growth Log                    -92.207        -202.042
                                (547.3461)      (300.6824)
Schooling in 1960 Log             2.4177          -2.0431
                                 (19.0493)       (12.9621)
Urban Population                 -24.2255         -5.3661
                                 (44.4452)       (29.4208)
Government Consumption           197.5567         128.531
                                (203.1253)      (149.3644)
Constant                       172.4875 ***    144.8437 ***
                                 (63.8489)       (33.8771)

Endogeneity Test                   3.640           3.228
P-value (Endogeneity Test)        0.1620          0.1990

Centered [R.sup.2]                 0.17            0.36
Number of Observations              39              43

NOTES: Standard ** error in parentheses. Level of significance: *** at
1 percent, ** at 5 percent, * at 10 percent. Instruments: Latitude and
English Legal Code.