Political Cycles and the Macroeconomy


Alberto Alesina

Nouriel Roubini

with Gerald Cohen

MIT Press

Publication date: November 1997


Chapter 1. Introduction.

Chapter 2. Opportunistic Models of the Political Business Cycle.

Chapter 3. Partisan Models of the Business Cycle.

Chapter 4. Political Cycles in the United States:

Evidence from Macroeconomic Policies and Outcomes.

Chapter 5. Electoral Probabilities and Their Impact on United States: Macroeconomic Cycles and Asset Prices

Chapter 6. Political Cycles in Industrial Democracies.

Chapter 7. Political Cycles and Macroeconomic Policies (Monetary and Fiscal): Evidence from Industrial Democracies.

Chapter 8. Political Cycles, Institutions and Monetary Policies.

Chapter 9. Political Parties, Institutions and Budget Deficits

Chapter 10. Conclusions: Summary of Results and Normative

Implications for Monetary and Fiscal Policy.

Chapter 1


1.1 Introduction
This book investigates the relationship between political cycles and economic cycles, namely how the timing of elections, the ideological orientation of governments and the nature of competition amongst political parties influence unemployment, economic growth, inflation, and various monetary and fiscal policy instruments.The topic of "Political Cycles" is one of the most widely studied subjects in political economics. Many believe that one can summarize this area of research with the idea that politicians artificially create unusually favorable economic conditions before an election, and that voters reward incumbent governments for doing so, even though the economy will take a turn for the worse immediately after the election. We argue, instead, that there is much more than this story to be told: in fact, this book illustrates a complex pattern of relationships between macroeconomic policies, timing of elections, ideological orientation of different governments, type of party structure, electoral laws, and degrees of Central Banks' political independence.
The literature on political and economic cycles developed in two clearly distinct phases. The first one, which flourished in the mid-Seventies, uses traditional macroeconomic models, in which, taking advantage of an exploitable Phillips curve, governments can systematically and predictably influence macroeconomic outcomes. A strand of this literature (Nordhaus (1975), Lindbeck (1976)) emphasizes the "opportunistic" motivation of policy-makers; politicians are viewed as individuals with no policy preferences of their own, so that they choose the policies which maximize their chances of electoral victory. The second strand of the literature (Hibbs (1977)) emphasized, instead, the partisan motivation of policy-makers. Specifically, according to this view, left-wing parties are relatively more concerned with unemployment than with inflation, while right-wing parties have opposite concerns. Moreover, parties stick to their electoral platform when in office.
The second phase of the literature took off in the mid-Eighties as a branch of the game-theoretic approach to macroeconomic policy, an approach pioneered by Kydland and Prescott (1977) and Barro and Gordon (1983a). These models of political cycles incorporate Rational Expectations and emphasize the limits that a rational public places upon the extent to which policy-makers can influence the economic cycle. Cukierman and Meltzer (1986), Rogoff and Sibert (1988), Rogoff (1990) and Persson and Tabellini (1990) develop "opportunistic" models with rational expectations. Alesina (1987) suggests a "partisan" model with rational expectations.
Table 1 frames the above mentioned contributions in a simple two-by-two matrix. Two critical points differentiate these models. First, partisan models focus upon differences in policies and outcomes as a result of the ideological orientation of different governments. On the contrary, opportunistic models claim that every government behaves in the same way: opportunistically, in order to win reelection. Second, Rational Expectations models emphasize the limits in the policy-makers' ability of influencing, permanently and predictably, the state of the economy. The "rationality constraint" affects both the policy (via voting behavior) and the economy (via the expectations augmented Phillips curve).
Our first goal (which we achieve in the next two chapters) is to illustrate these four types of models. (Box 1 succinctly summarizes the empirical bearings of these models.) We then test them on data drawn from the United States and almost all the other OECD countries. We analyze both policy outcomes (unemployment, output growth, and inflation), and policy instruments (monetary and fiscal variables). In addition, we use evidence from financial markets as well as economic and political indicators such as pre-electoral polls. Also, we discuss how opportunistic and partisan motivations interact with various institutional features, such as government structure (coalition versus single-party governments), electoral laws, and degree of independence of Central Banks.
This book reaches several conclusions: First, the most recent theories of political cycles, based upon the paradigms of Rational Choice and Rational Expectations, are empirically more successful than their predecessors.
Second, the partisan model outperforms the opportunistic one as an explanation of macroeconomic fluctuations in output growth, inflation, and unemployment. Opportunistic effects are confined to short-run, occasional manipulations of policy instruments around elections; for instance, in election years, one might expect "loose" monetary and fiscal policies, but not a significant surge in output growth or sharp reductions in unemployment.
Third, the lack of a systematic surge of growth in election years does not imply that economic conditions are not an important determinant of election results. We argue, in fact, that these two observations are not inconsistent.
Fourth, the characteristics of the political cycle depend upon the nature of the party system. A clear partisan cycle emerges in "two-party" or "two-bloc" systems. Countries with these party systems typically have majoritarian, or at least, not strictly proportional electoral laws. In contrast, partisan cycles are less clearly identifiable in systems with coalition governments, the typical product of proportional electoral systems. On the other hand, proportional electoral systems which produce coalition governments tend to be associated with excessive deficits and deadlocks in policy-making.
Fifth, we discuss how the benefits of independent monetary institutions are enhanced by the explicit considerations of opportunistic and partisan behavior of elected politicians.
These results point toward a consistent and rather clear picture. Our first result establishes the usefulness of the research efforts of economists and political scientists who have taken seriously the notion of rationality in political economy. However, even though we emphasize the role of rational behavior and expectations, we believe that wage and price rigidities are crucial ingredients for a realistic model of the economy. The combination of rational expectations and wage-price rigidities leads to models which exhibit short-run non-neutralities, and medium to long-run neutrality of aggregate demand policies. In fact, our empirical results on political cycles are consistent with a view of the economy based upon sticky wages and prices with rational expectations. Thus, although our empirical results are primarily concerned with political cycles, they have broader implications concerning the relative merits of different macroeconomic models.
Our second set of results highlight the nature of party policies. We argue that political parties are not all alike in industrial democracies. Different parties follow distinct macroeconomic policies when in office, and, therefore, partisan differences have been important determinants of macroeconomic policies and outcomes in the last thirty years. While the idea that parties are not all alike may strike many of the non professional political scientists as obvious, an established tradition in political science emphasizes the tendency of parties in two-bloc systems to converge to the middle.
This observation raises several issues relevant to current events. For example, one may wonder whether partisan cycles will become less marked in the future. Both left- and right-wing parties may realize, looking back, that partisan cycles destabilize domestic and world economies. For instance, in the late Seventies the political barometer of OECD tilted toward the left: in the United States, Germany, and the United Kingdom, left-leaning governments pursued expansionary policy, leading to a temporary boom and a build-up of inflation (aggravated by the second oil shock). In the early Eighties, the same three major economies moved to the right of the political spectrum, leading to decisive anti-inflationary policies and sharp recessions. Of the largest economies, only France followed the opposite political cycle, turning to the left in May of 1981. The Socialist government in France in 1982-83 followed typically partisan policies which were not very successful, partly as a result of the fact that the country was moving in the opposite direction of the world business cycle. In fact, the case of France is a good example of an important theme of this book, namely how external constraints limit the influence of partisan governments in open, integrated economies.
A related question is whether governments in OECD countries are learning to be less partisan and follow more "middle of the road" policies in macroeconomic management. For instance, the policies of the French and Spanish Socialist governments in the second half of the Eighties point in this direction. Also the process of European integration will affect partisan cycles. Will a monetary union and economic integration prevent national partisan cycles? How will the domestic politics of macroeconomic policy interact with European level politics?
Our third result highlights an apparent puzzle. The rate of economic growth is an important determinant of election results, but incumbent governments do not seem to systematically raise the rate of growth in election years. We address this puzzle directly by showing how recent theoretical developments and, particularly, the work by Alesina and Rosenthal (1995) are consistent with this observation.
Our fourth result, which contrasts two-party systems with coalition governments, links our book to a recent literature on the impact of institutional design on macroeconomic policy. Both academics and legislators around the world are debating the relative merits of proportional versus majoritarian systems. We believe, and our results confirm, that one faces a trade-off between these two systems. Proportional systems, which lead to coalition governments, are likely to insure moderation in policy-making and reduce the magnitude of partisan cycles. However, coalition governments are slow in reacting to shocks because each member of the coalition has a veto power over the choice of policies. In contrast, majoritarian systems exhibit the opposite features; they do not delay policies, but emphasize partisan cycles. Theoretical results by Alesina and Drazen (1991) and Spolaore (1993) and empirical results by Roubini and Sachs (1989a,b), Grilli, Masciandaro and Tabellini (1991), Alesina and Perotti (1995a) and the findings of this book support this view. Proportional systems achieve moderation, but create potentially inefficient delays in policymaking, and fiscal imbalances. Majoritarian, single-party governments do not delay policy implementation, but create partisan cycles.
A second institutional variable which we emphasize is the degree of independence of Central Banks. A vast literature has discussed the advantage and disadvantage of this institution, and pointed out empirical regularities and puzzles. Developing the recent work by Alesina and Gatti (1995) and Waller (1995), we extend the discussion of Central Bank independence to an explicitly political model, where partisan and opportunistic motivations of politicians are accounted for. This model resolves several empirical puzzles, and shows that the benefits of Central Bank independence may be even larger, when political incentives are explicitly taken into account.
The attentive reader may have noted that in summarizing our results, we did not make a distinction between those which apply to the United States and those which hold for other countries. In fact, one of our most striking results is that the United States are not exceptional. The same correlations between macroeconomic policies and outcomes on one side, and elections and government changes on the other, are remarkably similar in the United States and in other democracies, particularly those with a two-party system. A widely held view in political science suggests, instead, that the polity in the United States is unique, and hardly comparable with that of other parliamentary democracies. On the contrary, we argue that, at least as far as macroeconomics and politics are concerned, the similarities between the political economy of the United States and that of other OECD democracies are much more important than the differences. In fact, the key differences are not those between the United States and other parliamentary democracies as a whole, but between two-party systems and systems of large coalition governments.
1.2 Organization of the Book
This book is organized in ten chapters, including this introduction. Chapters 2 and 3 illustrate different theories of political cycles. Chapters 4 to 7 discuss empirical evidence from both the United States and other OECD economies on political cycles. Chapters 8 and 9 cover questions of institutional design; in particular, the relative performance, in terms of macroeconomic management, of different party structures generated by different electoral laws, the role of Central Bank independence and the issue of fiscal budget rules. The last chapter concludes and summarizes our results. We now briefly review the content of each chapter so that the reader has a sense of how the book unfolds.
Chapter 2 covers opportunistic models. We begin with what we define "traditional" approaches, and, specifically, with the Nordhaus' (1975) model, the most widely cited in this area. We then move to "rational opportunistic models", in which voters are rational, rather than naive as in Nordhaus' model, but are imperfectly informed about some characteristics of the policy-makers. Because of this information asymmetry, politicians have incentives to be strategic and opportunistic. These models also rationalize the widely observed phenomenon of "retrospective voting" (Fiorina (1981)). That is, they suggest why a rational electorate should evaluate the incumbent politician by looking at the state of the economy immediately before elections. This chapter ends with a discussion of the relationship between opportunistic behavior of politicians and rational opportunistic voting, a point which is quite important for interpreting the following empirical evidence.
Chapter 3 discusses partisan models. We begin again with what we refer to as the "traditional" approach by Hibbs (1977) and discuss party motivations and preferences. Then we proceed to a "rational" version of the partisan model, originally proposed by Alesina (1987 and 1988b). This is a model where, rather than a Phillips curve with adaptive expectations, as in Hibbs, we have rational expectations with wage stickiness. This chapter also includes an extension of the partisan model to incorporate retrospective voting behavior. We conclude with a summary of the empirical implications of all the theoretical models, which we then test in the following three chapters.
The empirical chapters begin with a comprehensive study of the United States, on both macroeconomic policies and outcomes as a function of the electoral cycle: Chapters 4 and 5 present several tests of political cycles on United States data from 1947 to 1993. We reach several conclusions which are then broadly confirmed by our study of all the other OECD countries. We devote two chapters to the United States not, as we argued above, because the results for this country are special. The reason is that we can study this country in more detail than any of the other countries, and we can compare our results with those of the previous literature on the subject, which has largely focused on the United States. For instance, Chapter 5 includes several tests using financial market data and pre-electoral polls, which are not available for other countries.
The next two chapters (6 and 7) consider a sample of almost all the OECD countries. Chapter 6 tests the implications of all the different models of political cycles on macroeconomic outcomes, namely, unemployment, growth, and inflation. We study whether the cyclical movements of these three variables are systematically affected by the timing of elections and by the partisan nature of different governments. Our sample includes quarterly observations on inflation, unemployment, and output growth in the last three decades for the following 18 OECD democracies: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Japan, Ireland, Italy, the Netherlands, New Zealand Norway, Sweden, Switzerland, the United Kingdom, and the United States.
While the previous chapter considered evidence of political cycles by studying economic outcomes, Chapter 7 follows a similar empirical methodology to test for the presence of political cycles in policy instruments. We conduct a detailed analysis of opportunistic and partisan cycles in monetary and fiscal policy.
Chapters 8 and 9 explore the interaction of various institutional features with party competition and economic policy making. In Chapter 8, we study the role of Central Bank independence. Then we develop, in more detail, an idea which emerged in previous chapters, namely, that macroeconomic policies are systematically different in countries with a two-party system relative to countries with coalition governments. In particular, we document the trade-off between coalition governments and single-party governments discussed above. In Chapter 9, we discuss how the evidence on political biases in fiscal policy affects the debate on the desirability of fiscal budget rules (such as a balanced budget amendment or a Maastricht type of fiscal rule).
The last chapter concludes by summarizing our main results and by discussing whether we should expect to see in the future more political cycles. In particular, we speculate on how certain institutional developments, European integration, increasing capital mobility and financial integration, learning from past experience, and ideological changes in attitude toward economic policy are likely to reduce or increase the magnitude of partisan and/or opportunistic cycles in the next decade.
1.3 How does this Book Relate to Previous Research?
This book is the result of several years of research, partly joint efforts and partly individual efforts of the three authors. Some of this research appeared in academic journals in the last decade, but this book is very far from a collection of previously published articles. First of all, the theoretical part of the book provides a comprehensive treatment of the theory of political cycles which is unavailable, as a whole, in any previously published work. Second, all of the empirical results previously published have been updated to the most recent available data. This procedure not only checks on the robustness of our previous results, but also allows us to study whether in recent years some systematic changes in governments' behavior have occurred. Third, we present several new tests never previously published. Fourth, and foremost, this book provides a comprehensive treatment of this area of research; a goal which escapes any individual paper.
In recent years, several authors have published books on related topics to the present volume. In terms of coverage, the closest work to our own is probably Lewis-Beck (1988). Our book differs from his in two significant ways. First, Lewis-Beck (1988) focuses more on the effects of economic conditions on voting behavior, rather than on the impact of elections on the economy, which is, instead, the main argument of our book. Second, at the time during which Lewis-Beck's volume was written, the new wave of rational expectation models had not appeared. Thus, his book deals only with the "traditional" models. In comparison, a critical point of the present volume is to show the usefulness of these more recent rational models, as a guide to empirical research.
Alesina and Rosenthal (1995) present an original model of the political economy of the United States. Even though one of the authors of the present book is very sympathetic to the message of Alesina and Rosenthal , the purpose, coverage, and goal of the two books are very different. First, Alesina and Rosenthal (1995) focus exclusively on the United States. Second, their goal is not to test a broad range of models, which is our purpose here. They develop and test a specific politico-economic model with particular reference to American institutions, and to the issue of divided government. Third, in this book we consider a much broader range of economic variables, while Alesina and Rosenthal (1995) focus almost exclusively on output growth.
Persson and Tabellini (1990) also discuss political-economic models in the context of their review of recent developments in macro-political-economy. However, they do not have any discussion of empirical evidence, while one of the main purposes of this book is to empirically evaluate different theories.
Keech (1995) critically evaluates recent developments in the field of macro-political-economy from the point of view of a political scientist. His book is quite different from ours for many reasons. First, he mainly focuses on the United States. Second, he addresses more normative issues than we do. Third, his book has much less formal modeling than we have in ours.
Tufte (1978) and Hibbs (1987a) discuss empirical evidence drawn almost exclusively on the United States concerning the opportunistic (Tufte) or the partisan (Hibbs) model. Both books, particularly the one by Hibbs, are richer in institutional details than ours. However, we draw evidence on a much larger sample of countries, and we use more economic variables. In addition, neither Tufte nor Hibbs discusses rational models, which, instead, are at the center of our attention.
1.4 Technical Level of this Book
We have tried to write a book which is at the same time rigorous and accessible to a much wider audience than the typical readership of academic journals. Our goal is to reach not only professional economists and political scientists, but also the educated general reader with an interest in this field.
In order to achieve this goal, in the theoretical part of the book, we will first develop the various models intuitively, with very little use of mathematical tools. We believe that the basic assumptions and results of these models can be easily illustrated in this way. Then, we develop a more formal treatment of the models. These sections (and/or appendices) can be skipped by the less technically inclined reader. The empirical part of the book makes use of econometric techniques, which, for the most part, are fairly standard. A deep and detailed understanding of technical aspects is never essential to grasp the basic message that we want to convey.

Table 1
Models with an exploitable  

Phillips Curve

Nordhaus (1975) 

Lindbeck (1976)

Hibbs (1977)
Rational Models
Cukierman-Meltzer (1986) 
Rogoff and Sibert (1988)
Rogoff (1990)
Persson and Tabellini (1990)
Alesina (1987)

Box 1
Empirical Implications of Different Models of Political Cycles  
i) Opportunistic traditional models
-expansion in the year or two before the elections; GNP growth above normal, unemployment below normal in the election year;
-inflation begins to increase immediately before or immediately after the election;
-recession (or downturn) after the election, with gradual reduction of inflation;
- no differences in policies and outcomes between different governments;
-incumbents reappointed when growth is high and unemployment low in election years.
ii) Rational opportunistic models
-short run manipulations of policy instruments immediately before elections: increase in deficits, inflation, money growth in the two-three quarters before each election;
-tightening of monetary and fiscal policies after elections;
-no systematic, multi-years effects on growth and unemployment except for, possibly, some minor effects immediately before the election;
-incumbents reappointed when growth is high and unemployment low in election years.
iii) Traditional partisan model
-unemployment permanently lower, growth and inflation permanently higher during the tenure in office of left-wing governments than with right-wing governments.
iv) Rational partisan model
-short-run partisan effects after elections: unemployment temporarily lower than normal and growth temporarily higher than normal for about two years after an electoral victory of the left; the opposite outcome after an electoral victory of the right;
-inflation permanently higher when the left is in office relative to when the right is in office.