Division of Research Graduate School of Business Administration The University of Michigan July 1983 GOVERNMENT BUDGETS IN MAJOR DEVELOPING COUNTRIES Concepts and Comparision for Political Risk Analysis Working Paper No. 333 Thomas L. Brewer The University of Michigan FOR DISCUSSION PRUPOSES ONLY None of this material is to be quoted or reproduced without the expressed permission of the Division of Research.

GOVERNMENT BUDGETS TN MAJOR DEVELOPING COUNTRIES Concepts and Comparisons for Political Risk Analysis This article contributes to the political risk literature in two ways. (1) It extends the domain of political risk analysis into the previously unexplored area of government budget policies. The analysis is specifically focused on dimensions of government budget policies that have important implications for international business. Data are presented for nine of the largest developing countries for the 1972-1980 period. Three industrial countries are also included for reference purposes in a cross-national comparative analysis, (2) The other principal contribution of the article to the. political risk literature is that it introduces the use of an operational definition of risk that can be used in systematic empirical research. Risk is thus indicated by variability over time in budget policies, as measured by standard deviations and coefficients of variation. The use of such a precise, operational, and theoretically-derived concept of risk can provide a basis for political risk analysis that is empirically and theoretically rigorous. II ~a~I..........IIF~~.I IJ L

1 INTRODUCTION Government budgeting is not only a central process in the political economies of countries; it is also a major concern of business firms. Thus, the budgets of foreign countries as well as home countries have important implications for corporations with international operations. -For instance, the debt repayment problems of several developing countries provide current reminders of the role of government budgets in the "country risks" and!'sovereign risks" faced in international banking. This article provides an empirical analysis of key aspects of the government budgets of nine large. developing countries. However, the article extends beyond the mere description of patterns in government budgets;. those patterns are analyzed in the context of a broader concern with political risks faced by international firms. The article therefore expands political risk assessment into a previously neglected analytic domain, and it presents new concepts and evidence that can be used for comparative assessments of such risks. Most international political risk research has been focused on violence, governmental instability, and expropriation.l However, in recent years there has been increasing recognition that more research is needed on specific governmental policies that affect international business since violence, regime instability, and expropriation are not the only forms of political risk faced

2 by firms.2 Indeed,, there are many other forms of politically related risks that are often more important. Thjs article continues the expansion of empirical political risk research by analyzing instabilities in government budget policies in particular. Political Risk, Country Risk:, and Sovereign Risk As used in the academic literature, the terms "political risk," "country risk," and "sovereign risk" overlap with one another.3 "Political risk" can be defined broadly to refer to uncertainty about the impact of the political environment on firms. "Sovereign risk" is normally used to refer more specifically to debt repayment problems posed by governments for banks. "Country risk" includes sovereign risk, but it also includes other risks that may arise from economic and/or political conditions in a foreign country. Sovereign risk and country risk are therefore in substantial part politically related risks. The article is thus intended to be a contribution to the closely related topics of political risk, sovereign risk, and country risk, However, the terms "political risk" and "politically related risks" will be used below since the emphasis of the article is on the political aspects of government budgeting. Scope and Outline The balance of the article will first discuss the relationship of government budgets to political conditions in countries, and the relationship of budget policy instability to other forms of ins - -----— ~- ---— ~Jl_- IL____

3 stability in political systems. Then, the potential effects of budget policy instability on firms will be discussed, and the specific hypotheses that have guided the research will be presented. Next, the research. methods and findings will be discussed; and finally, the concluding section will consider the implications for management, methodology, and.theory, POLITICS AND GOVERNMENT BUDGETS A government's budget policies are embedded in a complex set of relationships with other government policies, and with-political and economic conditions in the country. Those relationships are summarized schematically in Figure Fig. 1 1. It is unnecessary and beyond the scope of this article, however, to discuss all those relationships in detail; instead, we need to focus our attention here only on the relationship between political conditions and government budget policies (relationship a in Figure 1). The specific implications of government budget policies for international firms (relationship b in Figure 1) are discussed in subA sequent section below. Government budgets are political in two respects: budgets are affected by political conditions, and budgets 4 in turn influence political conditions. Political conditions affect government budgets through political processes in which there are conflicts among group interests. Since government budgets allocate scarce resources in ways that impinge on group interests, there is a struggle

FIGURE 1 Government Budgets and Risks in International Business: An Analytic Framework Govt. budget policies Economic conditions Political conditions Risks for international firms Other govt. policies

4 among competing groups to exercise power in the budget-making process. Indeed, the exercise of power in conflicts over the allocation of scarce resources among competing groups is one of the principal determinants of budget policy. Government budgets are therefore. outcomes of political processes. Government budgets, in turn, influence other political processes. The allocation of resources represented by a government budget has widespread implications for the interests, the power, and the conflicts among groups and individuals in the political system. Because government budget decisions do not affect all individuals and groups equally, budget decisions inevitably have political consequences. The unequal distribution of the benefits and costs that are inherent in government budget decisions means that the interests and the power of different groups and individuals are affected differently. Conflict in the political system is the result. In short, a government's budget is both a consequence and a cause of political conditions in a country. Political Instability and Budgets It is plausible to suppose, therefore, that political instability in a country would be reflected in its government's budget, Political instability, after all, is partly an overt manifestation of processes in which competing groups in a society are trying to exercise I I,-s I__ _

. 5 power as they pursue their conflicting goals. The. changes in the political system marked by political instability, should ' consequently be accompanied by instability in government budgets. Given the relative instability of developing countries' governmental regimes, the argument might continue, we should expect relatively high degrees *of instability in their budgets. On the other hand, several considerations suggest that perhaps there would be relative stability in at least some developing countries' budgets. In the first place, many developing countries have been marked by considerable stability in their governmental regimes. In some developing countries, individual heads of government or. political factions have remained in power for a decade or more. Indeed, it is precisely this kind of stability that often prompts the riots, guerilla wars, and coups d'etat that commonly occur in developing countries. Thus, in some developing countries, we might expect the stability of governmental regimes to be accompanied by stability in government budget policies, even though other forms of instability may be present. Moreover, the strong tendencies toward inertia jin bureaucratic procedures plus the need for political conpromising in legislative and executive. policymaking suggest that stability and incremental change in budgets might be the norm, rather than instability and irregular change.

6 In any case, there is mounting evidence that the many different forms of political instability in developing countries need to be distinguished from one another and analyzed individually. Political instability is a multidimensional phenomenon that needs carefully focused research on specific individual dimensions of political systems. DIMENSIONS OF GOVERNMENT BUDGETS There are at least four dimensions of government budgets in developing countries that have significant implications for the activities of international firms.6 Macro-Economic Policy First, government budgets are important elements in macro-economic policy. In particular, the size of the deficit (or surplus) in the government budget is a principal component of fiscal policy. The effects of fiscal policy on GNP, prices, and unemployment are of direct interest to all firms operating in a given country, Fiscal policy is also related to other government policies — such as monetary policy —that affect firms either directly or indirectly. Furthermore, fiscal policy is ultimately related to exchange control risks, exchange rate risks, and loan default risks. Role of Government in the Economy Second, government budgets are an important determinant of the role of government in the economy. At the aggregate level, the size of government expenditures

7 relative to the rest of the economy is an indicator of the 7 extent of the government's role in the economy7 Changes in that role may affect internal support for a regime as well as business-government relations. Sectoral Priorities Third, government budgets establish sectoral priorities in government expenditures and thereby influence the structure of countries' economies. The proportions of government expenditures devoted to agriculture, defense, transportation, health, education, and other sectors thus affect the sales prospects of firms in various industrial sectors. Government spending priorities also influence the prospects for economic development and social conflict. Project Funding Fourth, government budgets establish project funding levels, and those projects may offer sales opportunities to individual firms. These four budget policy dimensions and their impliTable 1 cations are summarized in Table 1. HYPOTHESES The analysis of instability in these dimensions of government budget policies has been guided by three hypotheses. (1) The correlations among the instabilities in the different budget policy dimensions are positive but only weak or moderate. This hypothesis is consistent with the

TABLE 1 Summary of Government Budget Implications for Risks Faced by International Firms Dimensions of Budgets Operationalized Indicators Risk Implications for Firms Fiscal policy Govt. deficit/GNP Internal political conflict Internal support for govt: mass & elite Exchange control policies Loan defaults Internal support for govt. by business elites Business-government relations Govt. role in economy Govt. exp./GNP Sectoral priorities Functional exp./total exp. Internal political conflict Sectoral sales growth Project support Project funding levels Project sales opportunities

8 previous political risk studies that have suggested the importance of distinguishing among various forms of political instability because they are frequently independent of one another. This hypothesis extends that theme to the specific dimensions of government budgeting included in the present study. (2) The correlations between budget policy instability and government regime instability are. also positive but only weak or moderate. The rationale for this hypothesis is the same as for the first hypothesis: Developing countries which are relatively unstable in some respects may be relatively stable in other respects. This hypothesis, furthermore, is consistent with the findings of a previous study of the association between policy and regime instability.8 (3) Developing countries' budget policies are not uniformly less stable than those of industrial countries, There are numerous and important exceptions to the generalization that developing countries tend to be relatively unstable, compared with industrial countries. These hypotheses have been tested with data concerning the budget policies of major developing countries, RESEARCH DESIGN AND METHODS Countries The budgets of nine developing countries have been analyzed —Argentina, Brazil, India, Indonesia, Korea, Mexico, South Africa, Turkey, and Venezuela. These are nine

9 of the twelve largest developing countries in terms of GNP.9 (Iran, Nigeria, and Saudi Arabia were excluded from the analysis because of inadequate data.) These particular developing countries were selected because of their economic size, and hence their inherent importance to firms with extensive international operations. Since they are not a representative sample of all developing countries, the findings concerning them clearly cannot necessarily be generalized to all developing countries. Three industrial countries have also been included so that comparisons can be made between developing countries and industrial countries. One of the industrial countries is the United States, which was included because it is a commonly used reference country in international economic and political comparisons. Germany is included because of its relative economic and political stability among industrial countries. Finally, Italy has been included because of the relative instability in its political economy in comparison with other industrial countries. In short, we have collected data on these nine developing countries because of their economic importance, and these three industrial countries because they provide convenient reference points for comparative perspective. Data Collection Government budget data made available by the International Monetary Fund have been used for the analysis.0 These are the most reliable and readily available data that

10 can be used for cross-national comparative analysis of countries' budgets. Data have been collected for the nine years from 1972 through 1980.11 We thus have data for a fairly lengthy time frame, which includes periods of global economi.c growth and recession. The time period encompassed by the data is also long enough to cover a lengthy period in each country's political history. Indicators of Ins-tability in Budget Policy Dimensions The operational indicator of fiscal policy is the government deficit as a percent of GDP. This measure of the size of the government deficit explicitly takes into account the changing size of the national economy, and it implicitly takes into account inflation. The same is also true of the indicator of the size of government that has been used -- the ratio of government expenditures to GNP. The choice of operational indicators for those two policy dimensions is therefore straightforward and in keeping with common practice. The choices of operational indicators for the other two budget policy dimensions are more problematic, There are numerous sectors and programs that might be of interest to any given international firm. The IMF publishes data on many standardized functional categories in governments' budgets so that sectoral priorities can be computed for each of them.2 However, it would not be feasible or appropriate to report findings for all such priorities since the analysis of

11 sectoral priorities is only one of several concerns of the present study. Thus, only one functional category has been 13 used —agriculture.3 Sectoral priorities are therefore indicated by agricultural expenditures as a percent of total government expenditures. All things considered, agriculture is one of the more homogeneous and pertinent functional categories for present purposes. Only further research, though, will reveal whether other functional categories would yield similar findings about sectoral priority instabilities in these countries' budgets. Similar methodological issues are involved in the operational indicator used for project funding levelsexpenditures on roads. In this instance, the choice was substantially limited by the available data since expenditures on roads are the most specific programmatic budget category in these IMF data sources. Expenditure levels for roads were converted into constant prices by using the wholesale price index for each country. Thus, changes over time in the program funding level reflect real changes, not nominal changes. Expenditures have been left in local currency terms, but the measure of instability used permits comparability across countries. Two common statistical measures of variability have been used to describe the degree of instability over time in the budget policy dimensions, The standard deviation has been used for the first three dimensions, For program funding levels, the coefficient of variation

12 has been used because it normalizes the measure of instability across countries according to differences in their mean levels of expenditures. The use of such measures for budget policy instability thus provides a familiar and operational measure of risk as deviations0 14 about an expected (mean) value.4 Governmental Regiime Instability The indicators of governmental regime instability are simpler, but still serviceable. They are based on the frequency of governmental change. Thus each country's regime instability score is the number of governmental changes that transpired during the period covered by the study. Three specific types of governmental instability have been measured —personnel, factional, and systemic. These three types of governmental instability can be defined as follows: Personnel instability refers to change in the head of the government. Changes in the prime minister or president or other head of government therefore constitute personnel changes. Personnel changes can be the result of elections or other political processes.

13 * Factional instability involves changes in the composition of the group(s) holding governmental power. In countries with competitive party systems, shifts in the party controlling the government are factional changes. If a coalition of parties rules, a change in the party composition of the coalition is a factional change. In countries with military regimes, coups are factional changes. In spite of their many differences in comparison with party shifts in competitive civilian systems, coups nevertheless share the common characteristic of a change in the. ruling group, Systemic instability involves a change in the basic form of the government, including a change in the constitutional basis of the government's authority. This form of regime instability therefore includes changes from a civilian to a military regime and vice versa. On the other hand, it excludes "cosmetic" constitutional changes not accompanied by changes in the head of the government. Data on these kinds of governmental changes have been collected from standard reference sources —Political Handbook of World, Statesman's Yearbook, and International Yearbook and Statesmen's Who's Who. Further information about the rationale for developing these indicators of governmental instability and the data collection procedures is available in a previous publication.15

14 Table 2 Table 3 FINDINGS Corre'lations Among In stabil'ities 'in Budgeta Poclicy DPimens ions The first hypothesis is that the correlations among the instability scores of the four budget policy dimensions are positive but only weak or moderate. The pertinent correlation coefficients for this hypothesis are presented in Table. 2. Some of the evidence in that table is consissome is nQt tent with the hypothesis;. In particular, only one of the positive correlations is strong, and statistically significant at the.05 level. Most of the correlations are positive, but they are not strong or statistically significant. On the other hand, two of the correlations are negative —one at a statistically significant level. Correlations Between Budget Policy Instability and Governmental Regime Instability Table 3 contains evidence for testing the second hypothesis that the correlations between policy instability and governmental regime instability would also be positive but weak or moderate. In fact, none of the correlations are strong enough to be statistically significant at the.05 level. Moreover, most of the relationships are negative. (The negative relationshipsbetween the sectoral priority instability scores and the personnel and factional instability scores are significant at.10. ) /

TABLE 2 Correlations Among Instability Scores for Budget Policy Dimensionsa Deficit Govt. 'Agri. Road Size Size Priority Funding Deficit Size 1.00 Govt. Size -0.28 1.00 Agri. Priority -0.63b +0.74b 1.00 Road Funding +0.54 +0.26 +0.05 1.00 aBased on developing countries only. b at Statistically signficant at.05 level. &

TABLE 3 Correlations between Governmental Instability and Budget Instability Form of Government Instability Form of Budget Policy Instabili ty Personnel Factional Systemic Fiscal Policy +0.36 +0.41 +0.40 Size of *"Sectoral Project Govt. Priorities Funding -0.12 -0.53 -0.38 -0.07 -0.55 -0.18 -0.32 -0.29 -0.37 aDeveloping countries only.

15 Comparisons Across Coun:tries 'for Each Policy Dimension The comparative degrees of instability in the four budget policy dimensions for each of the developing and industrial countries are presented in Table 4, Those Table 4 comparisons enable us to test the third hypothesis that the developing countries are not uniformly less stable than industrial countries. The first column of Table 4 presents evidence concerning the degree of instability in government deficits, Most of the developing countries are. actually more stable than the industrial countries in this respect. Only Argentina and Venezuela exhibit substantially greater degrees of instability than the United States and Germany — but about the same degree as Italy. The degree of instability in total government expenditures as a percent of GNP is indicated in the second column of Table 4. Only Mexico among the developing countries exhibits the same substantial degree of instability as Italy on this dimension. However, not only Mexico but also Indonesia, Turkey, Venezuela, and South Africa are all less stable than either Germany or the United States. On the other hand, four of the developing countries —Korea, India, Brazil, and Argentina —are all more stable than Germany on this dimension. The third column of Table 4 contains evidence concerning the variability over time in budget priorities, as indicated by the percentages of total government expenditures

TABLE 4 Variability Over Time in Four Dimensions of Government Budgets 1 2 3 4 Size of Size of Sectoral Project Deficita Governmentb PrioritiesC Fundingd Std. Dev. Rank Std. Dev. Rank Std. Dev. Rank -C. of V. Rank Developing Countries Argentina 2.29 (2) 1.50 (11) 0.63 (6) 0.15 (6) Brazil 0.63 (12) 1.58 (10). 1.43 (4) 0.18 (4) India 1.01 (8) 1.66 (9) 1.41 (5) na (na) Indonesia 0.82 (10) 3.44 (3) 1.95 (3) na (na) Korea 0.98 (9) 1.72 (8) 1.97 (2) 0.19 (3) Mexico 0.71 (11) 4.72 (2) 4.69 (1) 0.29 (2) South Africa 1.18 (7) 2.11 (6) na (na) na (na) Turkey 1.67 (4) 2.88 (4) 0.49 (7) 0.17 (5) Venezuela 3.18 (1) 2.14 (5) 0.42 (8) 0.46 (1) Industrial Countries Germany (FRG) 1.52 (5) 1.98 (7) 0.09 (10) 0.05 (8) Italy 2.25 (3) 4.74 (1) na (na) na (na) United States 1.26 (6) 1.09 (12) 0.39 (9) 0.07 (7) aStandard deviation of budget deficit as percent of GDP, 1972-1980 (except India, 1974-1980). bStandard deviation of government expenditures as percent of GNP, 1972-1980. CStandard deviation of government expenditures on agriculture, forestry, fishing, and hunting as a percent of total government expenditures, 1974-1980. dCoefficient of variation (standard deviation/mean) of constant local currency expenditures on roads.

Table 4 - Continued Data computed from International Monetary Fund, Government Finance Statistics Yearbook, Volume VI, 1982 lines S.8, CJI, B.8.2., and B.8.5; and from International Monetary Fund, International Financial Statistics,Yearbook, 1981; Yearbook, 1982, and Volume XXXV, Number 10; lines 63 and 99a.

16 that were devoted to agricultural programs. All of the developing countries exhibit more instability than the United States. (Germany and Italy are not very useful for comparative purposes on this dimension since so much of the European countries' expenditures on agriculture occur through the European Community's Common Agricultural Program.) However, the instability scores of Venezuela, Turkey, and Argentina are not substantially higher than the American score.. Mexico's score is much greater than any other country's score. Finally, in the fourth column of Table. 4, there is evidence concerning the variability in the funding level for roads. The pattern there is clear: greater instability among the developing countries than among the industrial countries. Country Profiles These findings can be further illustrated by country profiles. Three developing countries have been selected for diagrammatic presentations of their profiles in terms of the four budget dimensions. Brazil is presented as a relatively stable case, Korea as a moderately stable/unstable case, and Venezuela as an unstable case. It is apparent that each country exhibits a mixture of stability and instability across the four budget dimensions. None of the countries is uniformly stable across all policy dimensions. Whereas Brazil's profile in Figure 2 exhibits substantial stability in the fiscal policy Fig. 2

FIGURE 2 Country Profile for Brazil 72 73 74 I I I 75 76 77 78 79 )80 I 1 I 1 i' I tI I. I Government Deficit as Percent of GDP.4 _ deficit 0 - surplus 1 I 1 i ~: 1 t: t I I I I I:.. r i I -~, 4 24 _ Government Expenditures as Percent of GNP 20,~ * I\ **^" 16 *O ^ 91 Expenditures on Agriculture as Percent of Govt. Expend.,Nl *0 5 N. S-I \ 1 - 14 Expenditures on Roads in Constant Cruzeiros (bil.) * e 4 10 - 6

FIGURE 3 Country Profile for South Korea - - --- - - -L 72 73 74 75 76 77 78 79 1 I I 1 I I I I 80.; I I. I I. i I I. i i tI I I I I I. I Government Deficit as Percentof GDP 4 defici t 0 -surplu 4 _ 19 15 s I I t, 1 7 1 i I. I I I I I Government Expenditures as Percent of GNP./ ------ I - 'i II I * I \: 0.. Ti:r II i - I — K'It-.I-- I*iit I'1 -t4i 11 9 Expenditures on Agriculture as Percent of Govt. Expend. I I 5 / S 1 66 50 - I I I, i I:..., '., ~i~~~~~... * / 3,, 4~~~~~~; ' '. i:, ', 1~~~~~~. * *,, * _ ~ ~ r t ~~~**:, '. I * * ~ *. I, *,* r I I:.. 1 1 r /I r i I I I I i I Expenditures on Roads in Constant Won (bil.) * IaI tI I I II i i I i I II i * II. 34!t.

FIGURE 4 Country Profile for Venezuela.72 I 73 74 I, I 75 76 77 78 7~,1 I [ i I I, Government Deficit as Percent of GDP 4 - defici 9 80,:; j * ~ '; i t I:; '~ ' i i L / ': ' i ' ~ * * \ * ' * i;!j ' i:.' ] *. *. I;"I. ilf.,! r I i j. I I I_ r * | r~~~~ i i 1~~~~~~~~~~~~~;II surplu.: 4 ' 29 _ Government Expenditures as Percent of GNP.25 \% &.I I'^ - -'\: @ _* 21 9 Expenditures on Agriculture as Percent of Govt. Expend. 5 0 - - - 0 - * -. a — 1 2.5 Nl ~c~- ~9. - Expenditures on Roads in Constant Bolivares (bil.) 1.5. 0 0 0.5

17 dimension, it exhibits much less stability in the other three dimensions. South Korea, in Figure 3, also exhibits Fig. 3 relative stability in fiscal policy (at least after 1973), but considerable instability in its sectoral priorities and program funding levels. Finally, in Figure 4, it is Fig. 4 evident that Venezuela i.s rather unstable in three dimensions, but quite stable in the priority accorded to agricultural programs, CONCLUSIONS These findings have a variety of managerial, methodological, and theoretical implications for the assessment of politically related risks in international business. Management One conclusion is that the riskiness of countries should be assessed systematically in a cross-national comparative perspective. Since all countries pose political risks to firms! international operations, the issue is not whether a given country is or is not risky, Instead, the issue is whether it is more or less risky than other countries. Some countries that seem quite risky when considered in isolation may turn out to be much less risky when compared with other countries. We found, for example, that many developing countries' budget policies have actually been relatively stable in

18 some respects compared with American budget policies, A second conclusion, therefore, is that instability in developing countries is not necessarily more problematic than it is in industrial countries. In fact, similar or lesser degrees of instability are evident for these particular developing countries, when compared with industrial countries in terms of their budget deficits and their total government expenditures. The degree of instability in fiscal policy and in the role of government in the economy was therefore not uniformly more problematic in the developing than in the industrial countries. Thirdly, though, the degrees of instability for other dimensions of budget policy were quite different for the developing and industrial countries. The specific sectoral priorities and program funding levels included in the study were generally less stable in the developing countries than in the industrial countries. A fourth implication, then, is that the numerous specific forms of instability and risk need to be distinguished from one another and analyzed separately. Any one form may be more or less problematic for a particular activity of an individual firm in a given country. Political risks are not only firm-specific, they are also form-specific. That is, a given country may be relatively risky in terms of one form of politically related risk but not so risky in terms of another form. Although South Africa, for instance, has exhibited relative regime

19 stability, some aspects of its budget have been relatively unstable. Further research will of course be needed to determine whether the patterns discovered in this study also pertain to other time periods, countries, and aspects of budget policies. But the methodological implications of this study should be relevant in any case, Methodology An important methodological implication of this analysis is that politically related risks can be precisely defined and measured in ways that are analogous to traditional notions of risk. We have seen that the variability of government budget policies about mean levels can be measured by such traditional risk indicators as standard deviations and coefficients of variation. At least some forms of politically related risks can therefore be subjected to rigorous and systematic empirical analysis based on operationalized concepts. Theory The combination of (a) the possibility of measuring risks precisely in keeping with theoretical notionst plus (b) the desirability of analyzing those risks comparatively across countries has important theoretical implications.

20 It may be possible, for example, to apply the concepts and implications of portfolio theory to the analysis of 18 political risks. On intuitive and a priori grounds, anyway, portfolio theory appears to offer promising directions for the future of political risk research.

FOOTNOTES 1. Reviews of the political risk literature are available in Green (1972), Kobrin (1979), and Simon (1982). 2. See, for example, Blake (1982), Brewer (1981), Robock, Simmonds, and Zwick (1982), and Simon (1982). 3. Political risk has been defined several different ways in the academic literature. The most extensive discussions of definitional issues are in Kobrin (1979) and in Robock, Simmonds, and Zwick (1982). The definition adopted in the present article is somewhat more encompassing than either of those two. However, the present definition is not inconsistent with either their usage or more general usage in academic or managerial circles. The same core elements are evident here as elsewhere: uncertainty about the future, especially because of instability in the past; and the impact of politics on firms' activities, especially the possibility of losses. For distinctions among "political risk," "country risk," and "sovereign risk," see Nagy (1982). 4. Among the most important items in the extensive literature on the politics of budgeting are Davis, Dempster and Wildavsky (1966) and Wildavsky (1974). Items that deal specifically with budgeting in developing countries include Caiden and Wildavsky (1974), and Wildavsky (1975). Also see the annotated bibliography by Joseph (1982). 5. See, for example, Brewer (1983). 6. Coplin and O'Leary (1982) have noted the importance of government budget policies. In their "User Note," Number 11, for Frost & Sullivan's World Political Risk Forecasts, they observe: "to the extent that marketing includes a] concern about the stability and growth of existing markets and the potential for new markets, political events as well government policies can have a direct impact [6n firms' marketing activities]. Several Lfirmsi use political risk] analysis to anticipate government procurement patterns and those trends that are influenced by government decisions —such as economic growth, unemployment, and inflation.... While looking at the potential threats or political events in their markets, [marketing analysts] assess the degree to which political events may increase marketing opportunities. Particularly in some Third World countries, where government priorities frequently shift and sales are highly sensitive to those priorities, a new government or a new emphasis in policy can open up significant markets."

fn 2 7. In addition, funding levels for particular regulatory and subsidy programs that affect business are also clearly important determinants of the role of government in the economy. 8. Brewer (1983) found positive but not very strong relationships among indicators of government instability and policies concerning controls on remittances of funds by multinational corporations. 9. The GNPs of all the developing countries included in the study exceeded $40 billion in 1978 and in 1979. After the largest dozen developing countries, there was a gap of approximately $20 billion before the next cluster of countries. Thus, $40 billion GNP was a convenient cut-off point for inclusion in the study. In addition, about ten countries was a manageable number given the resources available for the study. It also seemed to provide a large enough and diverse enough group of countries. The GNP data were taken from United States (1981) and World Bank (1980). 10. The particular sources used were the Government Finance Statistics Yearbook and International Financial Statistics. 11. Data were not available for the years before 1972 or after 1980 for many countries. It is important to note that data for those later years would reflect several important and pertinent changes in the early 1980s: global recession, oil price declines, and the Reagan administration changes in the U.S. budget. Whether incorporation of this later period would substantially alter the results can only be determined by a subsequent study. In the meantime, it can be noted that the degree of instability in the U.S. budget as a reference base would surely be much greater if the 1981-present period were included. 12. The IMF functional (sectoral) categories include: defense, social security and welfare, education, health, housing, economic services, and others. 13. The agricultural category includes forestry and fishing as well as agriculture. 14. One limitation of these measures of instability is that they do'not separate long-term and short-term instability. In this particular study, however, this is not a serious problem. In only a few instances are there notable underlying trends for the time period in the study. In any case,the existence of trends raises a set of complex questions about conceptualizations of instability and its measurement which are beyond the scope ofithis 'article.However, they are being addressed in a separate study.

fn 3 15. Data on governmental regime instability were collected from Banks (1981), International Yearbook (1982), and Paxton (1983). See Brewer (1983) for further details on the concepts and data concerning governmental instability. 16. Mexico exhibits rather clear trends over the period. 17. Kobrin (1982) has emphasized the firm-specific nature of risks. 18. A paper in progress explores the implications of portfolio theory for political risk analysis.

REFERENCES Agodo, Oriye. "The Determinants of U.S. Private Manufacturing in Africa." Journal of International Business Studies, (Winter 1978), 95-107. Banks, Arthur S., ed. Political Handbook of the World. New York: McGraw-Hill, 1981. Bennett, Peter D., and Robert T. Green. "Political Instability as a Determinant of DirectfForeign Investing in-Marketing." Journal of Marketing Research, (May 1972), 1983-187. Blake, David H. "The Political Environments of Multinational Corporations," in Ingo Walter and Tracy Murray, eds., Handbook of International Business. New York: Wiley, 1982. Section 7. Brewer, Thomas L. "The Instability of Controls on MNEs' Funds Transfers and the Instability of Governments." Journal of International Business Studies, (Fall 1983),. "Political Risk Assessment for Foreign Direct Investment Decisions." Columbia Journal of World Business, 16, 1 (Spring 1981), 5-12. Caiden, Naomi, and Aaron Wildavsky. Planning and Budgeting in Poor Countries. New York: Wiley, 1974. Coplin, William D., and Michael K. O'Leary. "User Notes," No. 11, World Political Risk Forecasts. New York: Frost & Sullivan, 1982. Davis, Otto A., M. Dempster, and Aaron Wildavsky. "A Theory of the Budgetary Process." American Political Science Review, 60 (September 1966), 529-547. Friedman, Irving, et al. Assessing Country Risk. London: Euromoney, 1982. Green, Robert T. Political Instability as a Determinant of U.S. Foreign Investment. Austin: Bureau of Business Research, University of Texas at Austin, 1972. International Monetary Fund. Government Finance Statistics Yearbook. Washington, D.C.: IMF, 1982.. International Financial Statistics. Washington, D.C.: IMF, various volumes.. A Selected Bibliography on Budgeting and Expenditure Controls. Washington, D.C.: IMF, 1979.

International Yearbook and Statesmen's Who's Who, 1982. London: IPC Business Press, 1982. Joseph, Richard M., Jr. Budgeting in Third World Countries: An Annotated Bibliography. Monticello, Illinois: Vance Bibliographies, 1982. Kobrin, Stephen J. "The Environmental Determinants of Foreign Direct Manufacturing Investment: An Ex-Post-Empirical Analysis." Journal of International Business Studies, (Fall/Winter 1976), 29-43. ____- Managing Political Risk Assessment. Berkely: University of California Press, 1982. ~_____*_ "Political Risk: A Review and Reconsideration." Journal of International Business Studies, 10, 1 (Spring/Summer 1979), 67-80. Kurian, George Thomas, ed. Encyclopedia of the Third World, revised edition. New York: Facts on File, 1982. Mascarenhas, Briance. "Coping With Uncertainty in International Business." Journal of International Business Studies, 13, 2 (Fall 1982), 87-98. Moran, Theodore H., ed. International Political Risk Assessment: The State of the Art. School of Foreign Service, Georgetown University, 1981. Nagy, P. J. Country Risk. London: Euromoney, 1979. Overholt, William H. Political Risk. London: Euromoney, 1982. Paxton, John, ed. The Statesman's Yearbook, 1982-83. New York: St. Martin's, 1983. Robock, Stefen H., Kenneth Simmonds, and Jack Zwick. International Business and the Multinational Enterprise. Homewood, Ill.: Richard D. Irwin, 1982. Root, Franklin. Foreign Market Entry Strategies. New York: AMACOM, 1982. Simon, Jeffrey D. "Political Risk Assessment: Past Trends and Future Prospects." Columbia Journal of World Business, 17, 3 (Fall 1982), 62-71. United States. Statistical Abstract. Washington, D.C.: Government Printing Office, 1981. Walter, Ingo. "Country Risk Assessment," in Ingo Walter and Tracy Murray, eds., Handbook of International Business. New York: Wiley, 1982. Section 21.

Wildavsky, Aaron. The Politics of the Budgetary Process, 2nd edition. Boston: Little, Brown, 1974. Wildavsky, Aaron. Budgeting: A Comparative Theory of Budgetary Processes. Boston: Little, Brown, 1975. World Bank. World Atlas. Washington, D.C.: World Bank, 1980. C