BUREAUCRACY AND GROWTH:

A Cross-National Analysis

of the Effects of "Weberian" State Structures

on Economic Growth*

Peter Evans

Department of Sociology

University of California, Berkeley

and

James Rauch

Department of Economics

University of California, San Diego

Word Count: 7,082

_______________

* Please address all correspondence to Peter Evans, Department of Sociology, 410 Barrows Hall, University of California, Berkeley, CA 94720. These results grow out of the project on Bureaucratic Structure and Economic Performance directed by Peter Evans and James Rauch and funded in various stages by the Russell Sage Foundation, the Center for Institutional Reform, and the Informal Sector and Policy Research Department of the World Bank. We would also like to thank the 126 country experts for their generosity in sharing their knowledge and professional expertise. The findings reported here were produced by Linus Huang and John Talbot, working under the direction of Peter Evans and James Rauch. Patrick Heller and Mark Ritchie provided invaluable assistance in earlier stages of the data analysis. For comments on earlier drafts we are grateful to Ken Bollen, Neil Fligstein, Trond Petersen, Erik Olin Wright. Christy Getz played an invaluable role in the revision of the manuscript . We also are grateful to two anonymous reviewers and the ASR editor for their very useful comments.

Abstract

Bureaucracy and Growth:A Cross-National Analysis

of the Effects of "Weberian" State Structures on Economic Growth

The role of bureaucratic authority structures in facilitating economic growth has been a sociological concern since Max Weber's classic contributions of almost 100 years ago. Using a recently collected, original data set, we examine the characteristics of core state economic agencies and the growth records of a sample of 35 developing countries for the period 1970-1990. Our "Weberian State Scale" offers a simple measure of the degree to which these agencies employ meritocratic recruitment and offer predictable, rewarding long-term careers. We find that these "Weberian" characteristics significantly enhance prospects for growth, even when we control for initial levels of GDP per capita and human capital. Our results imply that "Weberianness" should be included as a factor in general models of economic growth. They also suggest the need for more attention by policy makers to building better bureaucracies and more research by social scientists on variations in the way state bureaucracies are organized.

Bureaucracy and Growth:A Cross-National Analysis

of the Effects of "Weberian" State Structures on Economic Growth

Explaining economic transformation at the national level is a classic sociological preoccupation as well as a central concern of economic analysis. There are many ways of approaching this task, but one of the most challenging involves trying to analyze the role of public institutions in fostering (and impeding) economic growth.

Growth depends on governance. Deciphering the relation between administrative structures and changing levels of economic output is, therefore, a perennial preoccupation of theorists and practitioners alike. In 1997, the World Bank took up the task again in a World Development Report called "The State in a Changing World." Both the choice of topic and the content of the report itself signified an important shift in thinking about the role of the state within the "development establishment." Explaining why some state bureaucracies are more effective than others finally seems to be taking precedence over simply condemning the excesses of state intervention. Pursuing this agenda requires re-exploring classic arguments on the comparative effectiveness of different forms administrative organizations. It is an obvious opportunity for sociological analysis to make a contribution to the understanding of cross-national differences in rates of economic growth.

Among the classic arguments that need to be brought together with some systematic comparative evidence, Weber's analysis of bureaucracy is perhaps the most obvious candidate. At the beginning of the century, Weber's (1968 [1904-1911]) monumental essays, Economy and Society, argued for the fundamental value of bureaucracy as one of the institutional foundations of capitalist growth. Subsequent comparative historical analysis (e.g. Polanyi 1944) echoed Weber's assertions, but the "bureaucracy as a tool of growth" thesis always had to contend with the historically prior and ideologically powerful "Smithian" view[1] that government, regardless of its organizational form, was the enemy of growth as soon as it went beyond protecting property rights.

In the seventies and eighties, neo-classical political economy and rational choice analysis provided new analytical reinforcement for the Smithian perspective (cf. Buchanan, Tollison, and Tullock 1980; Colclough and Manor 1991; Collander 1984; Krueger 1974). Case studies of "rent-seeking" and "predatory" states complemented these analytical arguments with equally powerful empirical reinforcement (e.g. Bates 1981; Klitgaard 1988). Unfortunately, in the rush to avoid the dangers of state intervention, the question of what kinds of state structures were most likely to promote economic growth was easily lost.

By the nineties, however, economists (but surprisingly not sociologists) began to focus on cross-national data that demonstrated the importance of looking more closely at how states were organized. Their results (Mauro 1995; Knack and Keefer 1995) showed various measures of "quality of government" to be powerfully connected to economic growth. This rapidly growing literature suggested that earlier neo-classical visions of government performance were too simplistic. Nonetheless, perhaps because of the absence of sociologists from the discussion, a way of describing what "good government" would look like was still lacking.

Contemporary empirical analyses of rent-seeking and corruption often use the term "bureaucracy" in its everyday pejorative sense rather than in the Weberian sense of a set of administrative organizations with specific structural features. Weber viewed bureaucracy, not as a generic collection of state officials, but as a particular kind of organizational structure, set in contrast to earlier patrimonial and prebendal forms of government administration.

The Weberian perspective does not negate the positive effects of strengthening market institutions, but it does postulate that bureaucratically structured public organizations, using their own distinct set of decision-making procedures, are a necessary complement to market-based institutional arrangements.[2] More precisely, Weber argued that public administrative organizations characterized by meritocratic recruitment and predicable, long term career rewards will be more effective at facilitating capitalist growth than other forms of state organization. This hypothesis cannot be dismissed simply by the discovery that people who call themselves bureaucrats have engaged in rent-seeking or that corrupt governments have undermined economic growth. Addressing the "Weberian state hypothesis" means answering the question, "Will countries whose administrative apparatuses more closely approximate bureaucratic forms of organization be characterized by higher rates of economic growth?" For some reason, students of economic development have lacked the incentive required to generate a systematic empirical response to this apparently simple question. The research reported here represents an initial effort to fill the lacuna.

Using an original data set, this analysis examines the effect on economic growth of certain structural features that were key elements in Weber's original characterization of bureaucracy. Our "Weberian State Scale" offers a simple measure of the degree to which core state agencies are characterized by meritocratic recruitment and offer predictable, rewarding long-term careers.

Our findings can be previewed succinctly. Examining the growth records of a sample of 35 developing countries for the period 1970-1990, we find that "Weberian" characteristics significantly enhance prospects for growth. This is true even controlling for initial levels of GDP per capita and initial levels of human capital. In addition, "Weberianness" provides a succinct, analytically satisfying explanation for observed regional differences in growth performance. Both the exceptionally good performance of the East Asian NICs and the exceptionally poor performance of the cases from sub-saharan Africa can be predicted on the basis of regional differences in the degree to which state bureaucracies are Weberian. These findings do not contradict the existing literature analyzing cross-national differences in growth rates, but they do offer an important complement.

Recent Literature: The literature on the role of state bureaucracies in promoting or impeding economic growth ranges from detailed cases studies of particular agencies in particular countries to cross-national analyses using statistical proxies drawn from many countries. Traditionally, political scientists dominated the production of case studies (e.g. Stepan 1978; Waterbury 1983) while sociologists focused more on cross-national analyses (e.g. Delacroix and Ragin 1981; Rubinson 1977; Snyder and Kick 1979). In recent years, awakened interest among economists in cross-national analysis has substantially expanded the production of cross-national statistical studies. The principal stimulus to renewed interest among economists has been the emergence of "endogenous growth theory," which offered formal theoretical support for the proposition that institutional factors could have a fundamental effect on rates of growth (cf. Lucas 1988; Romer 1994, 1990, 1986, etc.).

The endogenous growth perspective legitimated a variegated set of cross-national analyses by economists examining the impact of a variety of non-economic variables on national growth rates (see Crowley et al. forthcoming). One of the earliest and most influential of these studies, Barro (1991), emphasized the negative role of government by stressing the negative impact of government consumption (as a share of GDP) on growth rates.

What both the earlier sociological literature and the new cross-national literature in economics lacked was the possibility of seriously addressing the question of how variations in the form of state organization might affect economic dynamism. Arguments in the earlier sociological literature were formulated in terms of the dubious concept of "state strength" with state revenues and expenditures offered as proxies for "state strength" (e.g. Rubinson 1977). The cross-national regressions of the new endogenous growth models included even more unsatisfying proxies, such as using the annualized number of assassinations or revolutions as a proxy for political stability (e.g. Barro, 1991).

Neither economists nor (more surprisingly) sociologists made efforts to focus their cross-national analyses on organizational differences. Nonetheless, a literature had developed which paid serious attention to bureaucratic structures and was directly relevant to the Weberian hypothesis. It was built around detailed case studies of individual countries and focused primarily on a single region. Chalmers Johnson's (1982) classic study of Japan's MITI during the golden years of Japanese industrialization led the way. Johnson's portrayal was surprisingly consistent with a Weberian perspective. Subsequent studies of Korea (Amsden 1989) and Taiwan (Wade 1990) reinforced the same picture.[3] Since the "East Asian Tigers" described in these studies were also the most economically successful nations in the world during the 1970's and 1980's, they created a strong prima facie case in favor of the Weberian hypothesis. By the beginning of the 1990's even the World Bank seemed to join in the chorus with its East Asian Miracle (1993) report which emphasized the positive role played by East Asian bureaucracies in the region's spectacular industrialization.[4] Nonetheless, the case continued to rest primarily on case studies. The question remained as to whether the Weberian state hypothesis could be substantiated with a broader set of systematic evidence.

In the 1990's the case study literature began to be complemented by efforts to connect variations in the character of state bureaucracies to economic performance by means of quantitative cross-national analysis. Economists, rather than sociologists, took the lead in developing a more organizational focus. A new set of studies began to utilize the commercially available assessments of variations across national governments that businessmen had been using for some time.[5] The methodological validity and reliability of these measures were open to question. The rating services that provided them offered little explanation of how their data were derived or why they should be considered reliable. They seemed to be based primarily on the assessments of consultants, but the basis on which these consultants were selected is not usually specified and methodological issues were clearly not a primary concern.[6] The validity of these ratings as independent determinants of economic growth was also somewhat suspect. Most of them had clear "good" and "bad" poles -- e.g. more or less corruption, more or less red tape, higher or lower 'quality' of the bureaucracy, etc. Since raters were obviously aware of the economic performance of the countries they were rating there was likely to be a tendency to give "good" ratings to high performing countries and "bad" ratings to low performing ones, thus building in an automatic correlation between the ratings and economic growth.

Despite their flaws, these ratings did provide a way of comparing bureaucracies across a wider set of countries, and they did seem to produce results that confirmed the importance of variations across national bureaucracies in explaining variations in economic growth. Paolo Mauro (1995), using ratings on "corruption" and "red tape" from Business International, found that variation in these ratings was significantly associated with increased levels of investment, which were in turn one of the most obvious and powerful predictors of economic growth. Knack and Keefer (1995) used International Country Risk Guide (ICRG) ratings and Business and Environmental Risk Intelligence (BERI) ratings and again discovered that these ratings are directly related to variations in the growth of per capita income.[7]

Use of data on variations in state bureaucracies gives recent studies a clear advantage over earlier work that had to rely on gross measures like aggregate government expenditures or distant proxies like number of assassinations. The fact that these studies consistently find relationships between bureaucractic performance and economic growth provides new incentive for trying to refine our understanding of the roots of "bureaucratic performance." Even this new generation of studies remains, however, prisoner to the available measures. Convincing efforts to adjudicate the empirical validity of the Weberian state hypothesis must begin with information on how the structure of state bureaucracies varies across countries, which is what we have tried to do in the study that follows here.

Connecting Bureaucratic Structures and Growth: Contemporary analysis of comparative bureaucratic structures needs to move beyond Weber , but Weber's characterizations do provide a simple, accessible starting point for comparative research. In setting up the contrast between bureaucracies and prior organizational forms, Weber stressed a number of points that lend themselves to relatively objective empirical assessment. In the work that follows, we have chosen to emphasize two of them. The first is the importance of meritocratic recruitment, ideally based on some combination of education and examination (Parsons 1964:333,339; Gerth and Mills 1958:241). Second, a predictable career ladder provides long term tangible and intangible rewards for those who have been recruited into the bureaucracy (Gerth and Mills 1958:200-203; Parsons 1964:334-335; Stinchcombe 1974).

Other Weberian organizational features could have been selected.[8] One advantage of meritocratic recruitment and rewarding, predictable career ladders is that they are relatively easy to translate into simple measures that can be evaluated across countries. Focusing on them facilitates empirical testing. Equally, important plausible theoretical connections can be constructed between these structures and improved organizational ability to deliver the collective goods that constitute the state's potential contribution to economic growth.

Meritocratic recruitment not only increases the likelihood of at least minimal competence, but also helps generate corporate coherence and esprit de corps, which in turn can be plausibly argued to have substantive effects on the motivations of individual office holders. Bureaucrats who see themselves as having joined their confreres in office by virtue of sharing similar abilities are more likely to internalize shared norms and goals than those who know they owe their office to the favor of a particular kinsman or patron. Identification with colleagues and the organization should also create internalized intangible costs for corrupt activities that subvert organizational goals and increase the effectiveness of monitoring.

Rewarding long-term careers might also increase competence in the long run, but, regardless of their effects on competence, they will increase corporate coherence. Likewise, the predictable prospect of long-term career rewards reduces the relative attractiveness of the quick returns available from corrupt individual practices. This is obvious insofar as one of the aspects of long-term career rewards is competitive salaries. It is equally clear that careers which include the expectation of a series of promotions related to performance and conformity to organizational norms create disincentives to corrupt behavior, especially if such behavior undermines organizational goals. The costs of breaking organizational norms are also directly proportional to the expected longevity of membership in the organization and the expected rewards to longevity.[9] Overall, meritocratic recruitment and predictable career ladders should help structure the incentives of individual bureaucrats in a way that enhances the ability of the organizations they manage to effectively pursue long-term goals.[10]

If the argument that these structural features contribute to a more competent, purposive and cohesive bureaucracy is accepted, myriad specific causal paths leading to higher rates of economic growth are plausible. The longer time horizons associated with predictable, rewarding careers will increase the bureaucracy's propensity to advocate public sector infrastructure investment rather than consumptive expenditures. Since the returns from public infrastructure investments depend essentially on their "systemness," the coherence of the bureaucracy should enhance their effectiveness. Likewise, the reduction in individual maximizing (i.e. corrupt) practices should reduce the implicit tax on the private sector that such practices represent.

Diffuse links may be equally or more important. Most of the case study literature on "developmental states" (e.g. Johnson 1982; Amsden 1989; Wade 1991; World Bank 1993; Evans 1995) focuses primarily on the role of state bureaucracies in eliciting higher rates of private investment. Obviously, rational, risk averse entrepreneurs will avoid making long-term investments in plant and equipment if they face a corrupt, unpredictable bureaucracy unlikely to provide complementary public investments. By the same token, shared perceptions of the state bureaucracy as dependable, predictable, minimally competent and itself committed to long-term growth makes investment appear less risky.

Competent bureaucracies can help individual entrepreneurs overcome coordination problems that may be especially crucial in instigating new activities. They can also turn informational resources into public goods in ways that increase the likelihood and effectiveness of investment.[11] For example, when entrepreneurs in small countries are trying to upgrade into world markets, collective action to gather data on external markets and enforce standards among local producers may confer important advantages. Respected bureaucracies could act as "honest brokers" in overcoming collective action problems among exporters. A stronger version of this argument would see the bureaucracy itself as gathering information and providing advice and incentives that help local firms to better thread their way through the labyrinth of rapidly changing world markets.[12]

Adjudicating among the various paths that might account for a connection between competent, coherent state bureaucracies and economic growth would be a challenging task in itself, but it is not the aim here. The aim here is to establish the basic connection between bureaucratic structures and growth, thereby providing additional incentive to explore alternative mechanisms that might account for the connection.[13]

This study is then an assessment of a particular set of bureaucratic structures, not a comprehensive appraisal of all the features of bureaucratic structure that might enhance economic performance. Meritocratic recruitment and career ladders have been selected because of the strong tradition in the literature of claims on their behalf and because they constitute an empirically identifiable, theoretically plausible set of structural characteristics that makes a good starting point for getting research on the economic consequences of variations in bureaucratic structures off the ground.

Our strategy for connecting bureaucratic structures and growth is, of course, quite different from Weber's. Weber was interested in long-term historical changes in organizational forms. We are interested in cross-sectional comparisons in the contemporary period, essentially within the last thirty years. Our proposition is a very simple one. We predict that countries whose bureaucratic structures incorporated Weberian features will have experienced more rapid economic growth over the 20 years between 1970 and 1990.

Data:[14] The absence of comparable measures of bureaucratic structure for a substantial set of countries is one of the principal impediments to assessing the effects of variation in bureaucracy on economic growth. We decided that only the collection of a new, original set of data could surmount this obstacle. The "Weberian State data set" that we collected is built on comparable expert evaluations of bureaucratic structures in 35 countries, laboriously collected over a period of almost three years (1993-1996).

Our sample of countries began with the 30 "semi-industrialized" ones identified by Chenery (1980) and was complemented by five poorer countries. Our reasons for starting with the Chenery sample were basically threefold. To begin with, we estimated that our available resources would not allow us to collect data on more than about 35 countries, so the Chenery sample was about the right size. Second, we were interested in understanding variations in growth among developing countries rather than between developing and advanced industrial countries. We wanted a sample of countries that were still confronting the issue of industrial tranformation during the period under consideration. The Chenery sample provided a good range of developing countries. Third, we wanted a good range of variation on "Weberianness." While there was no systematic data available on Weberianness per se, we did know that variation in "bureaucratic performance" variables was much greater among developing countries. This latter consideration was also a motive for including some additional countries too poor to make it into the Chenery sample. Data on bureaucratic performance made it clear that excluding poorer countries would under sample the lower end of the distribution on bureaucratic performance and therefore might well under sample the lower ranges of Weberianness as well (see Rauch and Evans 1997:8-9 and figures 1b and 1c). The choices of the particular poor countries to be added were driven by a desire for increased representation of the Caribbean, South Asia and Sub-Saharan Africa and by our belief that there was a sufficient corps of experts on the bureaucracies of these countries to allow us to find at least three experts for each of them.[15] The resulting sample of 35 countries provides representation for all the major regions of the developing world as well as the southern European fringe of the OECD. It also offers a range of growth performance over the 1970's and 1980s from Korea and Singapore growing consistently at over 6% per capita per year to Zaire shrinking at a bit over 2%. At the same time, this sample of countries offers a good range in terms of "bureaucratic performance" as measured by commercial rating services (see Rauch and Evans 1997: figures 1b and 1c and discussion below).

Obtaining measures of different features of bureaucratic structure in each of these countries required the cooperation of a large set of experts, each with specific knowledge of the state bureaucracy of a particular country. The process of collecting responses was labor intensive, but the level of collaboration was both surprising and gratifying. We were ultimately able to get responses from a total of 126 experts, a minimum of three experts for 32 of our 35 countries and two for the remaining three (Morocco, Thailand and Uruguay). The experts were drawn from a combination of scholars known for their research and publications on the bureaucracies of each country, local officials with reputations for having a broad perspective on their country's administrative structures, and professionals working on these issues in multilateral organizations. Together the experts in our sample account for a substantial portion of the published literature on the state bureaucracies for the countries in question.[16]

The experts' evaluations were structured by means of a fixed-response questionnaire.[17] In answering the fixed response questions, the experts were asked first to identify the central state agencies which played the most important role in formulating economic policy[18] and answer questions regarding them. Responses concerning the state bureaucracy more generally then followed.[19] We reasoned that the structure of the core economic agencies was most likely to have an effect on economic growth, and since case studies of individual countries make it clear that there is usually substantial variation across agencies, getting a measure that focused on the most relevant agencies made sense.

Country experts were not asked to make evaluations of the performance or quality of the bureaucracy. Instead, the questionnaire focused on specific descriptive features of the bureaucracy that are subject to objective estimation. We then combined these descriptive features to construct a simple measure which would reflect a Weberian bureaucratic structure, built on meritocratic recruitment and rewarding predictable career ladders.

Since bureaucratic structures are notoriously resistant to change, we felt secure in assuming differences we discovered among of bureaucratic structures could be considered to characterize the beginning of the period (and, indeed, had probably been in place for some time prior to 1970) and were, therefore, temporally antecedent to growth during the period 1970-90 period. To check on this assumption, experts were asked not only to provide responses that characterized the entire period, but also to note any significant changes during the period. Despite some references to deterioration in the situation of bureaucrats over time (particularly in relation to relative salaries), it was clear from their comments that the bureaucratic structures they described antedated 1970-90 economic growth.[20]

Ten questions were used to create what we called the "WEBState Scale." An initial question indicated the importance of the agencies in question in generating economic policy. Two of the subsequent questions (Appendix B:2 and 9) measured the importance of exams in recruiting civil servants to the core economic agencies and more generally.[21] Three of the questions tapped issues relating to careers: whether civil servants, once recruited are likely to stay in the civil service (3 and 5) and whether staying in the civil service implies possibilities for moving up within a hierarchy (4). An additional four questions tapped the issue of career rewards, both in terms of salaries and prestige (6, 7, 8, 10). The resulting "Webstate" scale provides a succinct, substantively plausible measure of Weberianness. (See Appendix B for a full discussion of the scale and a listing of its items.)

To discover what relationship, if any, this WEBState Scale might bear to economic growth during the period 1970-1990, we then drew on measures from available standard data sets to create our dependent variable -- growth in GDP per capita over the period 1970-1990-- and our control variables -- initial income level and pre-existing human capital. (See Appendix C for a full discussion of the dependent and control variables.) The 1994 version of Summers and Heston's (1991) Penn World Tables was the source of our measures of real GDP per capita in 1965, 1970 and 1990. As our proxy for human capital we used an updated version of Barro and Lee's (1993) measure of average years of education for the population 25 years and older.

Analysis: The aim is to discover whether "Weberianness" appears to have an effect on growth that is independent of the effects of other variables classically associated with economic development. There is a strong and significant correlation between the WEBState Scale and total growth of real GDP per capital during the 1970-1990 period (r. = .674; p.<.001), but it could be argued that "Weberianness" is simply a proxy for overall level of development or existing stocks of human capital. Such an argument cannot be dismissed out of hand. We know that more highly developed bureaucracies are more likely to be found among developed countries (cf. World Bank 1997; Rauch and Evans 1997:figure1a). We also know that higher levels of human capital, which are in general associated with higher levels of development, are strongly associated with growth. In fact, however, there was virtually no association between the degree of approximation to Weberian characteristics and initial levels of per capita income in this sample of developing countries (r = .05). In this sample of developing countries at least, it is hard to argue that past growth or higher levels of income are important causes, in themselves, of more Weberian states. Put more optimistically, it seems that low levels of per capita income are not necessarily a barrier to achieving more competent and coherent state bureaucracies. Nonetheless, there is a modest (though not significant) correlation between the WEBState Scale in our sample and pre-existing level of human capital (r. " .25, p. = .15) and that human capital has in turn a significant positive effect on subsequent growth.

Our first key result then is that even after the effects of initial GDP per capita levels and pre-existing levels of human capital have been controlled for, the relation between the WEBState Scale and growth remains strong and significant. As the regression equations below indicate, the WEBState Scale continues to have a powerful and significant effect on growth. Weberianness is not simply a spurious proxy for effects of pre-existing levels of development or human capital.

_________________________

Regression of Growth on WEBState Scale

(controlling for initial GDP per Capita and pre-existing Human Capital)

Basic equation in unstandardized form:

GROW7090 = -44.54 - .02(GDPCAP65) + 15.77(HUMAN65) + 16.05(WEBSCALE) (1)

In standardized form:

GROW7090 = -.316(GDPCAP65) + .307(HUMAN65) + .615(WEBSCALE) (2)

(See Appendices B and C for definitions of variables)

_____________________

The scattergram of growth regressed on the WEBState Scale with the effects of initial level of GDP per capita and pre-existing levels of human capital controlled is shown in Figure 1. One of the things that stands out in this scattergram is the strong degree to which the regional distribution of Weberian characteristics parallels regional differences in growth performance.

[ INSERT FIGURE 1 ABOUT HERE ]

Regional differences in both growth and Weberianness are summarized in Figures 2 and 3. Figure 2 shows the range, interquartile range and median on growth for each of 4 regional groupings. When the four regional dummies are arranged in the predicted order they form almost a perfect regression line with sub-Sahara Africa at the bottom and the four Tigers at the top. The Latin American region exhibits a growth performance that is clearly inferior to any region except for sub-saharan Africa, largely because Latin America's growth experience in the period we are examining is dominated by the "lost decade" of the 1980's.

[ INSERT FIGURE 2 ABOUT HERE ]

Figure 3 provides a portrayal of regional differences with respect to the WEBState Scale and makes the parallel between regional variations in growth and regional variations in bureaucratic structure graphically clear. Just as sub-Saharan Africa defines the bottom of the scale in terms of growth, it is also the region in which state bureaucracies are least Weberian. Likewise, the 4 Tigers epitomize both high growth and Weberian bureaucratic traits.[22] The WEBState Scale appears to capture an important institutional element of the "high performing" East Asian economies while at the same time pointing to an institutional deficit that may be important in explaining low rates of growth in Africa.

.

[ INSERT FIGURE 3 ABOUT HERE ]

Table 1 adds a set of regressions using a range of regional dummies to the basic regression illustrated in the scattergram. The effects of the WEBState Scale still appear robust. Introducing the dummies for sub-saharan Africa and Latin America reduce the WEBState Scale coefficient only modestly (Regressions 2 and 3). Even more important, a "bureaucratic structure effect" (significant at the .05 level) remains even when an East Asian dummy ( the Four Tigers) is included in Regression (4).[23]

[ INSERT TABLE 1 ABOUT HERE ]

Given these initial results, the obvious next question is whether the effects of Weberianness continue to be significant when the WEBState Scale is included in regressions that introduce other variables used in the standard growth models which abound in the literature. Levine and Renelt (1992) survey 41 studies using cross-national regressions to explain growth. They conclude (1992:942) that even though "over 50 variables have been found to be significantly correlated with growth in at least one regression," the list of variables whose effects are truly robust is, in fact, very small. They identify (1992:947, table 1) three basic variables - initial level of GDP, investment, and human capital - as most robustly related to growth. Table 2, Regression (5) shows that the results using our sample, time period, and variable definitions are fully consistent with previous work, except that the effects of human capital fall just short of significance. Initial investment level is the most significant predictor of growth, initial level of GDP continues to have a negative relation to growth, and human capital is positive and significant at the p < .05 level (one-tailed test).

[ INSERT TABLE 2 ABOUT HERE ]

When the WEBState Scale is added to this basic equation (in Regression 7), it becomes the most powerful predictor of growth and the adjusted R² increases significantly. Table 2 also explores (in Regression 6) the effects of two additional variables which were found in Barro's (1991) analysis to have significant negative effects on growth -- government consumption and revolutions. When these variables are added to the basic Levine and Renelt set, their coefficients are not significant[24] and they result in only a trivial increase in the adjusted R². In Regression (8) the WEBState Scale is added to Regression (6). The results parallel those of Regression (7). The WEBState Scale once again becomes the most significant predictor, and the adjusted R² is again significantly increased.[25] Overall, these results suggest that if data could be collected for a broader range of countries, Weberianness would become a valuable addition to the existing literature on cross-national growth models.

Given the powerful and robust relationship between investment levels and growth in the standard cross-national literature and the fact that sociological work on cross-national models of growth has also emphasized the important role of investment (cf. Firebaugh, 1992:125), it makes sense to look at the effects of "Weberianness" on investment levels as well as on rates of growth.[26] Table 3 shows Weberianness to have a consistently significant positive effects on end of period investment levels, reinforcing the position generally taken in the case study literature on "developmental states," which emphasizes the role of the state in elevating levels of private investment as a principal mechanism through which states promote higher rates of growth (see Evans 1995). [27]

[ INSERT TABLE 3 ABOUT HERE ]

Discussion: The evidence presented here adds credence to the proposition that state bureaucracies characterized by meritocratic recruitment and predictable, rewarding career ladders are associated with higher growth rates. Since the data refer primarily to core economic agencies, the implication is not that the entire bureaucratic apparatus must be structured in this way in order to have positive effects on growth. Having "Weberian" structures in the strategic core of the bureaucracy may be sufficient.

"Weberianness" provides a parsimonious, analytically satisfying account of observed differences in regional growth performance. These findings support interpretations of high East Asian growth that have emphasized the contribution of competent, cohesive bureaucracies and offer a succinct, objective and replicable substitute for the unsatisfyingly amorphous and atheoretical idea of an "East Asian effect." They are also consistent with explanations of low African growth rates which underline problems of governance. More generally these findings suggest that an important contribution could be made to the existing literature on the cross-national analysis of growth if systematic evidence on state structures could be collected on a larger sample of countries. Weberianness is a potential sociological addition to the small set of robust predictors of growth that have been identified in recent cross-national studies by economists.

Despite the promising character of these results, it must still be underlined that the research presented here is only a beginning. A variety of avenues for further work are obvious. Extension of the data set to a larger sample of countries is the first step toward a better test of the robustness of the relationship found here.[28] The success of the initial effort reported here argues that the returns from gathering more and better evidence on cross-national variations in the structural characteristics of state bureaucracies would more than justify the effort required.

More ambitious, but increasingly essential, is extending the longitudinal coverage of data collection by looking at changes in the character of public bureaucracies in the 1990's.[29] This task is particularly relevant to understanding the roots of current economic problems in East Asia. Earlier case studies (e.g. Amsden 1989; Evans 1995) emphasized that it might be difficult to sustain the Weberian character of East Asian bureaucracies. Recent commentators have suggested that the declining integrity of public bureaucracies has played an important role in the deterioration of East Asian financial systems (e.g. Chang, Park and Yoo, 1998).

Finally, it should be illuminating to examine the relationship between Weberianness and a range of other political, social and economic variables that have been shown to be related to growth. Political regimes (Alvarez et al. 1996) and the policy outputs commonly used in cross-country regressions (e.g. black market premium, fiscal surplus, price distortions) are two examples.[30]

There is much to be done, but one incontrovertible conclusion transcends the exploratory character of the work reported here. The "Weberian state hypothesis" deserves more attention from sociologists and other social scientists, both empirically and analytically.

References

Alvarez, Mike, JosŽ Antonio Cheibub, Fernando Limongi and Adam Przeworski. 1996. "Classifying Political Regimes." Studies in Comparative International Development. 31(2)[summer]:3-36.

Amsden, Alice. 1989. Asia's Next Giant: South Korea and Late Industrialization. New York: Oxford University Press.

Barro, Robert J. 1991. "Economic Growth in a Cross-section of Countries." Quarterly Journal of Economics 106 (May):407-444.

Barro, Robert, and Jong-wha Lee. 1993. "International Comparisons of Educational Attainment." Journal of Monetary Economics 32 (December):363-394.

Barzelay, Michael. 1997. "Results Oriented Government: Theories and Research on the New Public Management," Presentation at the Aaron Wildavsky Forum for Public Policy, April 3-4, University of California, Berkeley.

Barzelay, Michael (with the collaboration of Babak Armajani). 1992. Breaking Through Bureaucracy: A New Vision for Managing in Government. Berkeley: University of California Press.

Bates, Robert H. 1989. Beyond the Miracle of the Market: The Institutional Foundations of Agrarian Development in Kenya. Cambrige, England: Cambridge University Press.

___________. 1981. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley, CA: University of California Press.

Buchanan, James M., Robert D. Tollison and Gordon Tullock, eds. 1980. Toward a Theory of the Rent-Seeking Society. College Station, Tx: Texas A&M University Press.

Campos, Jose Egardo, and Hilton L. Root. 1996. The Key to the Asian Miracle: Making Shared Growth Credible. Washington, DC: The Brookings Institution.

Callaghy, Thomas. 1989. "Lost Between State and Market: The Politics of Economic Adjustment in Gambia, Zambia, and Nigeria" Pp. 257-319 in Economic Crisis and Policy Choice: The Politics of Economic Adjustment in the Third World, edited by Joan M. Nelson. Princeton, N.J.: Princeton University Press.

Chang, Ha-Joon, Hong Jae Park, and Chul Gyue You, 1998. "Interpreting the Korean Crisis: Financial Liberalization, Industrial Policy, and Corporate Governance," unpublished ms. Cambridge, UK: Faculty of Economics and Politics, Cambridge University.

Cheng, Tun-jen, Stephan Haggard, and David Kang. 1995. "Institutions, Economic Policy and Growth in Korea and Taiwan. Paper presented at UNCTAD Seminar on "Development of East and South-east Asia and a New Development Strategy - the Role of Government" 30-31 October, Geneva, Switzerland.

Chenery, Hollis B. 1980. "The Semi-Industrialized Countries," Mimeo. Washington, DC: The World Bank.

Colclough, C. and J. Manor (eds.), 1991. States or Markets: Neo-liberalism and the Development Policy Debate. Oxford: Clarendon Press.

Collander, David, ed. 1984. Neoclassical Political Economy: an Analysis of Rent-Seeking and DUP Activities. Cambridge, MA: Ballinger.

Crowley, Angela, James Rauch, Susanne Seagrave and David Smith. 1998. "Quantitative Cross-National Studies of Economic Development: A Comparison of the Economics and Sociology Literatures." Studies in Comparative International Development 33 (Summer):30-57.

Delacroix, Jacque and Ragin, Charles. 1981. "Structural Blockage: A Crossnational Study of Economic Dependency, State Efficacy, and Underdevelopment." American Journal of Sociology 86:1311-1347.

Easterly, William and Ross Levine. 1997. "Africa's Growth Tragedy: Policies and Ethnic Divisions." Quarterly Journal of Economics 112(4) [November]:1203-1250.

Ernst, Dieter and David O'Connor. 1992. Competing in the Electronics Industry: The Experience of Newly Industrializing Countries. [An OECD Development Centre Study] Paris: OECD.

Evans, Peter. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton, NJ: Princeton University Press.

Feder, Gershon. 1983. "On Exports and Economic Growth," Journal of Development Economics 12 [February/April]: 59-73.

Firebaugh, Glenn. 1992. "Growth Effects of Foreign and Domestic Investment," American Journal of Sociology 98:105-30

Fligstein, Neil. 1996 "Markets as Politics: A Political-Cultural Approach to Market Institutions," American Sociological Review. 61[August]:656-673.

Gerth, Hans and C.W. Mills. 1958. From Max Weber: Essays in Sociology. New York: Oxford University Press.

Hood, Christopher and Michael Jackson. 1991. Administrative Argument. Aldershot, Herts.: Dartmouth.

Johnson, Chalmers. 1982. MITI and the Japanese Miracle: the Growth of Industrial Policy, 1925-1975. Stanford: Stanford University Press.

Keesing, Donald B. 1988. "The Four Successful Exceptions: Official Export Promotion and Support for Export Marketing in Korea, Hong Kong, Singapore, and Taiwan,China," UNDP‑World Bank Trade Expansion Program Occasional Paper 2.

Klitgaard, Robert. 1988. Controlling Corruption. Berkeley, CA: University of California Press.

Knack, Stephen and Philip Keefer. 1995. "Institutions and Economic Performance: Cross-country tests using Alternative Institutional Measures," Economics and Politics 7(3)[November]:207-227.

Koh, Gillian. 1995. "A Sociological Analysis of the Singapore Administrative Elite: The Bureaucracy in an Evolving Developmentalist State." Ph.D. Dissertation, University of Sheffield, England.

Krueger, Anne O. 1974. "The Political Economy of the Rent-Seeking Society." American Economic Review 64,3 (June):291-303.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert Vishny. 1998. "The Quality of Government," Discussion Paper No.1847 Harvard Institute for Economic Research, Harvard University, Cambridge, MA.

Levine, Ross, and David Renelt, "A Sensitivity Analysis of Cross-Country Growth Regressions," American Economic Review. 82(4):942-963.

Mauro, Paolo, 1995. "Corruption and Growth," Quarterly Journal of Economics 110 (August): 681-712.

Milgrom, Paul and John Roberts. 1992. Economics, Organization, and Management. Englewood Cliffs, NJ: Prentice Hall.

Olsen, Johan P. and B. Guy Peters (eds.). 1996. Lessons From Experience: Experiential Learning in Administrative Reforms in Eight Democracies. Oxford: Scandanavian University Press.

Page, John. 1996. "Does Governance Matter," Paper presented at International Conference on Governance Innovations. October, 20-23. Manila, Philippines.

Parsons, Talcott. 1964. Max Weber: The Theory of Economic and Social Organization. New York: The Free Press.

Polanyi, Karl. [1944] 1957. The Great Transformation. Boston: Beacon Press. PRS Group. 1995. "International Country Risk Guide" (brochure), Syracuse, NY: PRS.

Quah, Jonathan. 1993. "The Rediscovery of the Market and Public Administration: Some Lessons from the Singapore Experience." Australian Journal of Public Administration. 52(3):320-28.

Rauch, James and Peter Evans. 1999. "Bureaucratic Structures and Bureaucratic Performance in Less Developed Countries." University of California, San Diego Discussion Paper No. 99-06.

Rodrik, Dani. 1995. "Getting Interventions Right: How South Korea and Taiwan Grew Rich," Economic Policy (April): 55‑97.

Romer, Paul. 1994. "The Origins of Endogenous Growth" Journal of Economic Perspectives. 8(1)[Winter]:3-22.

_________. 1990. "Endogenous Technological Change." Journal of Political Economy 94(October):71-102.

_________. 1986. "Increasing Returns and Long Run Growth." Journal of Political Economy 94(October):1002‑1037.

Rubinson, Richard. 1977. "Dependence, Government Revenue and Economic Growth, 1995-1970," Studies in Comparative International Development. 12:3-28.

Stinchcombe, Arthur. 1974. Creating Efficient Industrial Administrations. New York, NY: Academic Press.

Summers, Robert and Alan Heston. 1991. "The Penn-World Tables (Mark 5): An expanded set of international comparisons, 1950-1988," Quarterly Journal of Economics, 106(9) (May), updated.

Talbot, John. 1997. "A Description of the 'Weberian State' Comparative Data Set," Project document, Bureaucratic Structure and Economic Performance Project, University of California, Berkeley.

Wade, Robert. 1990. Governing the Market: Economic Theory and the Role of Government in Taiwan's Industrialization. Princeton, NJ: Princeton University Press.

Weber, Max. [1904-1911] 1968. Economy and Society. Edited by Guenter Roth and Claus Wittich. New York: Bedminster Press.

World Bank (IBRD). 1997. World Development Report: The State in a Changing World. New York: Oxford University Press.

_________ . 1994. Adjustment in Africa: Reforms, Results and the Road Ahead. New York: Oxford University Press.

_________ . 1993. The East Asian Miracle: Economic Growth and Public Policy (A World Bank Policy Research Report) New York: Oxford University Press.

_________ . 1991. World Development Report. New York: Oxford University Press.

_________ . 1983. World Development Report. New York: Oxford University Press.

Zifcak, Stephen. 1994. New Managerialism: Administrative Reforms in Whitehall and Canberra. Buckingham: Open University Press.

Appendix A: Ranked scores of the 35 countries on the Weberianness scale

	Country	Abbreviation	WEBSCALE
	Argentina	ARG	3.80
	Brazil	BRA	7.60
	Chile	CHL	5.00
	Colombia	COL	8.50
	Costa Rica	CRI	9.00
	Cote d'Ivoire	CIV	8.00
	Dominican Republic	DOM	2.00
	Ecuador	ECU	4.00
	Egypt	EGY	7.80
	Greece	GRC	10.00
	Guatemala	GTM	3.00
	Haiti	HTI	4.00
	Hong Kong	HKG	11.00
	India	IND	10.00
	Israel	ISR	7.00
	Kenya	KEN	1.00
	Korea	KOR	13.00
	Malaysia	MYS	10.50
	Mexico	MEX	8.50
	Morocco	MAR	7.00
	Nigeria	NGA	3.00
	Pakistan	PAK	11.00
	Peru	PER	5.00
	Philippines	PHL	6.00
	Portugal	PRT	5.00
	Singapore	SGP	13.50
	Spain	ESP	10.00
	Sri Lanka	LKA	8.00
	Syria	SYR	3.80
	Taiwan	OAN	12.00
	Thailand	THA	8.00
	Tunisia	TUN	9.00
	Turkey	TUR	7.00
	Uruguay	URY	4.50
	Zaire	ZAR	4.00

Appendix B: Creation of the "Weberianness" Scale (WEBSCALE)

The Weberianness scale was created from ten items in the original questionnaire. The items with response alternatives abbreviated are shown below. The full questionnaire and the recodes used in compiling the scale are available at weber.ucsd.edu/~jrauch/webstate.

1. Which of the following descriptions best fits the role of these agencies in the formulation of economic policy.

1 = many new economic policies originate inside them.

2 = some new policies originate inside them.

3 = they rarely originate new policies.

2. Approximately what proportion of the higher officials in these agencies enter the civil service via a formal examination system?

1 = less than 30% 2 = 30 - 60% 3 = 60% -90% 4 = more than 90%

3. What is roughly the modal number of years spent by a typical higher level official in one of these agencies during his career?

1 = 1-5 years 2 = 5-10 years 3 = 10 -20 years 4 = entire career

4. What prospects for promotion can someone who enters one of these agencies through a higher civil service examination early in his/her career reasonably expect? Assuming that there are at least a half dozen steps or levels between and entry-level position and the head of the agency, how would you characterize the possibilities for moving up in the agency? [ NB. more than one may apply.]

1. in most cases, will move up one or two levels.

2. in most cases, will move up three or four levels.

3.     will move up several levels to the level just below top political appointees.

4.     in at least a few cases, will move up to the very top.

5. How common is it for higher officials in these agencies to spend substantial proportions of their careers in the private sector, interspersing private and public sector activity?

1 = normal 2 = frequent but not modal 3 = unusual 4 = almost never

6. How would you estimate the salaries (and perquisites, not including bribes or other extra-legal sources of income) of higher officials in these agencies relative to those of private sector managers with roughly comparable training and responsibilities?

1 = less than 50%2 = 50 - 80% 3 = 80% - 90% 4 = Comparable 5 = Higher

7. If bribes and other extra-legal perquisites are included what would the proportion be?

1 = less than 50% 2 = 50 - 80% 3 = 80% - 90% 4 = Comparable 5 = Higher

8. Over the period in question (roughly 1970-1990) what was the movement of legal income in these agencies relative to salaries in the private sector,

1 = declined dramatically 2 = declined slightly

3 = maintained the same position. 4 = improved their position.

9: This variable was created from the combined responses of all experts for each country, based on an assessment of the importance of civil service examinations for entry into the bureaucracy.

0 = No civil service exams, or exams are of trivial importance

1 = Ambiguous based on experts' responses

2 = Civil service exams are an important component of entry to the bureaucracy

10. Among graduates of the country's most elite university(ies), is a public sector career considered:

1 = the best possible option.

2 , 3 = depends on circumstances.

4 = a second best option.

WEBSCALE - The Weberianness Scale

Basic Measure of Bureaucratic Competence/Coherence

The individual responses to the above questions (except # 9) were aggregated to create a country-level data set, in which each country received a score equal to the average of the responses of all experts answering each question for that country. [Country ratings on #9 were based on the investigator's assessment of combined country expert responses to two questions regarding initiation and selectivity of civil service exams.] Country averages for each of the ten questions were recoded into two or three categories in such a way as to obtain as equal a distribution of countries over the categories as possible, and the ten questions were then combined to form a scale.

Appendix C: Definition of Variables

Sources:

Banks, Arthur S., "Cross-national time series data archive." Center for Social Analysis, State University of New York at Binghamton, (September 1979), updated.

Barro, Robert & Jong-hwa Lee, "International comparisons of educational attainment." NBER Working Paper, 1993.

Nehru, Vikram, Eric Swanson & Ashutosh Dubey, "A new database on human capital stock in developing and industrial countries: Sources, methodology, and results." Journal of Development Economics v46 (1995): 379-401.

Summers, Robert & Alan Heston, "The Penn-World Tables (Mark 5): An expanded set of international comparisons, 1950-1988," Quarterly Journal of Economics, v106 n9 (May 1991), updated.

Variables:

GROW7090: Total percentage growth of real per capita GDP from 1970 to 1990 ((GDPCAP90 - GDPCAP70) / GDPCAP70). Source: Penn-World Tables 5.5. (see definitions below).

INV8590: Average of the annual ratio of real domestic investment (private plus public) to real GDP over the period 1985 to 1990. Source: Penn-World Tables 5.5.

INV6570: Average of the annual ratio of real domestic investment (private plus public) to real GDP over the period 1965 to 1970. Source: Penn-World Tables 5.5.

WEBSCALE: Measure of state bureaucratic competence/coherence or "Weberianness" (see Appendix B).

GDPCAP65: Real per capita GDP, 1965. Source: Penn-World Tables 5.5, (RGDP/population).

GDPCAP70: Real per capita GDP, 1970. Source: Penn-World Tables 5.5, (RGDP/population).

GDPCAP90: Real per capita GDP, 1990. Source: Penn-World Tables 5.5, (RGDP/population).

HUMAN65: Average schooling years in the total population over age 25, 1965. Source: Barro & Lee, 1994. NOTE: Data for HUMAN65 was unavailable for Egypt, C™te d'Ivoire, Morocco and Nigeria. Observations for these four countries were estimated using data on total educational attainment (TOTLEDUC) from Nehru (1995). The Pearson correlation coefficient between HUMAN and TOTLEDUC was .7991 (p = .000). Regressing HUMAN on TOTLEDUC produced the following equation:

HUMAN = .176757 + ( .825982 * TOTLEDUC )

The coefficient on TOTLEDUC was significant at better than the .0001 level, and the (adjusted) R-squared was .63860.

GOVC7085: Average of the annual ratio of real government consumption expenditure net of spending on defense and on education to real GDP over the period 1970 to 1985. Source: Penn-World Tables 5.5, IMF Government Financial Statistics, SIPRI Yearbooks, UNESCO Statistical Yearbooks.

REVO7085: Average of the number of revolutions per year from 1970 to 1985. Source: Barro & Lee (itself sourced from Banks).

AFRICA: 0/1 dummy variable for the sub-Saharan African countries: C™te d'Ivoire, Kenya, Nigeria and Zaire.

EASIA: 0/1 dummy variable for the four "East Asian Tigers": Hong Kong, Korea, Singapore, and Taiwan.

LATINAM: 0/1 dummy variable for the Latin American countries: Argentina, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, Guatemala, Haiti, Mexico, Peru and Uruguay.

Table 1

Growth of Real GDP per capita and "Weberianness"

	Model Number
Independent Variable	(1)	(2)	(3)	(4)
WEBSCALE	.615** (4.649)	.537** (3.700)	.599** (4.268)	.247* (2.418)
GDPCAP65	-.317 (-1.749)	-.270 (-1.472)	-.324 (-1.753)	-.150 (-1.251)
HUMAN65	.307 (1.645)	.319 (1.725)	.290 (1.487)	.090 (.724)
LATINAM		-.180 (-1.242)
AFRICA			-.056 (-.367)
EASIA				.696** (6.638)
N	35	35	35	35
Adj. R2	.460	.469	.444	.774

Dependent variable is GROW7090: Growth in real GDP/capita from 1970 to 1990.

See Appendices B, C, and D for definitions of variables.

Figures presented are: standardized coefficients, T-scores (in parentheses).

*indicates significance at the .05 level.

**indicates significance at the .01 level.

Table 2

"Weberianness" and Some Basic Models of Economic of Growth

	Model Number
Independent Variable	(5)	(6)	(7)	(8)
WEBSCALE			.490** (3.647)	.482** (3.104)
GDPCAP65	-.644** (-3.171)	-.778** (-3.642)	-.458* (-2.554)	-.498* (-2.402)
HUMAN65	.401 (1.988)	.420* (2.068)	.258 (1.470)	.211 (1.115)
INV6570	.580** (3.498)	.434* (2.463)	.360* (2.362)	.298 (1.861)
GOVC7085		-.334 (-2.024)		-.179 (-1.169)
REVO7085		-.136 (-.950)		-.003 (-.023)
N	35	34a	35	34a
Adj. R2	.343	.373	.529	.521

^aCote d'Ivoire (CIV) lacks data on government consumption.

Dependent variable is GROW7090: Growth in real GDP/capita from 1970 to 1990.

See Appendices B, C, and D for definitions of variables.

Figures presented are: standardized coefficients, T-scores (in parentheses).

*indicates significance at the .05 level.

**indicates significance at the .01 level.

Table 3

Investment and "Weberianness"

(Dependent Variable = End of Period Investment Level)

		Model Number
Independent Variable		(9)	(10)	(11)	(12)	(13)
WEBSCALE		.489** (3.508)	.477** (3.044)	.456** (3.103)	.348* (2.146)	.370* (2.276)
GDPCAP65		-.179 (-.936)	-.172 (-.867)	-.194 (-1.003)	-.115 (-.603)	-.354 (-1.609)
HUMAN65		.443* (2.255)	.445* (2.225)	.407 (1.994)	.360 (1.814)	.436* (2.128)
GOVC7085						-.314 (-1.931)
REVO7085						-.136 (-.951)
LATINAM			-.026 (-.169)
AFRICA				-.119 (-.747)
EASIA					.267 (1.599)
N		35	35	35	35	34a
Adj. R2		.400	.381	.392	.429	.434

^aCote d'Ivoire (CIV) lacks data on government consumption.

Dependent variable is INV8590: average annual investment share of real GDP from 1985 through 1990.

See Appendices B, C, and D for definitions of variables.

Figures presented are: standardized coefficients, T-scores (in parentheses).

*indicates significance at the .05 level.

**indicates significance at the .01 level.