Political business cycles 40 years after Nordhaus

This paper necessitated reading 530 articles or books. The complete list of references is available upon request. For convenience, I have disregarded unpublished papers (with a few exceptions) and papers not written in English. Despite all our efforts, I was not able to locate approximately 20 published articles or books related to the topic. In some contributions, the study of political business cycles is reduced to a simple electoral dummy introduced as a control variable among others and sometimes not related to Nordhaus (1975). I have probably missed some studies like these. Finally, I note that PBCs, explicitly related or not to Nordhaus (1975), have been investigated in somewhat exotic areas (compared to economics): Wasserman (1983) in relation to suicides, Nincic (1990) and Gaubatz (1991) in international relations, Shughart and Tollison (1985), Lagona and Padovano (2008) and Brechler and Geršl (2014) in legislative activity, Block and Vaaler (2004) in ratings published by agencies, Thies and Porche (2007) in agricultural producer protection, Ladewig (2008) in the housing market, Horgos and Zimmermann (2010) in the activity of interest groups, Kächelein et al. (2011) in electricity supply, and Wehner (2013) in legislative budgetary decisions. Unusual applications in the economic area include Paiva (1996) on prices in regulated industries, Castro and Veiga (2004) on the timing of stabilization programs, Dreher and Vaubel (2004) on credits from the IMF and IBRD, Özatay (2007) on public sector prices, Faye and Niehaus (2012) on foreign aid, Sukhtankar (2012) on prices paid to farmers for sugar cane, Lami et al. (2014) on household consumption spending, and Klien (2014) on water tariffs.

Although it is the canonical expression, “political business cycles” can lead to possible confusion as it in fact refers to a particular type of cycle, namely opportunistic. An opportunistic cycle is the result of political manipulations before an election. In contrast, partisan cycles appear as a result of manipulation after the election. A generic term to designate both opportunistic and partisan cycles could be “politico-economic cycles” or “electoral cycles”. The first author to use the expression political business cycle even if not directly linked to elections was Kalecki (1943, p. 330). Michal Kelecki used this expression to qualify the inflation–unemployment cycle, which reflects fluctuations in the political power struggle between conflicting social classes. Moreover, he was rather silent on the government’s behavior. In his work, the government appears as a passive relay in class struggles, favoring the different classes alternately through stop and go policies.

All the same, one notes that the reelection motivation is present, for example, in Tocqueville (1835).

However, in these papers (and in subsequent papers by the same authors and their coauthors), one should note that the concept of cycle is missing, the political pressure being permanent and not only during elections. I will disregard here the literature concerning this “satisficing” version of the PBC model, but this idea will later be taken up and combined in the traditional electoral cycle (see Sect. 3.4).

One might note that other articles or books dealing with the idea of PBCs, sometimes written independently of Nordhaus (1975), exist: Schumpeter (1939, in particular p. 391), Breton (1974, in particular p. 49), Ben-Porath (1975), Lindbeck (1976), Frey and Ramser (1976), Umstead (1977), Wagner (1977) and Lächler (1978).

See de Haan (2013), de Haan and Klomp (2013), Drazen (2000, chapter 7; Drazen 2001), Franzese (2000, 2002), Franzese and Long-Jusko (2006), Olters (2001) and Shi and Svensson (2003) for the most recent references.

Naïve rather than adaptive expectations, as in the Nordhaus model, are used here without loss of generality.

For example, the specification of Nordhaus’s (1975) vote function has been criticized (e.g., Frey and Ramser 1976; Paldam 1981). I do not develop this point here because the debate about vote function was older and the critique of Nordhaus (1975) did not give rise to a new strand of literature in this area. However, this issue is related to the effectiveness of PBCs. Do the PBCs pay off in terms of votes? Examining the impact of pre-election budget deficit on reelection probability, Brender and Drazen’s (2008) conclusion is negative. Aidt et al. (2011) challenge this view and find that greater expenditures in the election year lead to greater vote differences between the incumbent and the main opponent (see also Veiga and Veiga 2007b, in this line). Moreover, some refinements of the Nordhaus model are not shown here because they have not yet given rise to further developments.

One notes, however, that in partisan theory, the cycle is involuntary: the electoral calendar is a source of uncertainty and disrupts the agents who, not knowing which policy will be implemented after the elections, anticipate an inflation rate between the parties’ preferred rates. After the elections, expectations necessarily turn out to be false, which generates a cycle.

Another way, as chosen by Nordhaus (1989), is to defend adaptive expectations by showing that people do not have rational expectations. Early critics of Nordhaus’s view of expectations include, among others, McCallum (1978), Keech (1980) and Paldam (1981). In these studies, voters’ sophistication and memory are also debated.

Cukierman and Meltzer (1986) are also often quoted, but if they emphasize the role of information asymmetries, they do not deal explicitly with PBCs. Persson and Tabellini (1990) translate Rogoff (1990) into an inflation-unemployment setting.

One should note, however, that an old and persisting literature exists that highlights the presence of PBCs in stock markets. For early references, see Umstead (1977) and Allvine and O’Neill (1980); for more recent references, see Sturm (2013) and Kräussl et al. (2014).

The legitimacy of tests in terms of instruments rather than outcomes is reinforced by the fact that instruments are the action variables in rational PBC models. Besides all of their properties, their use for political purposes is, moreover, grounded theoretically. One also notes that budgetary variables are privileged over monetary variables because of the conjunction of two elements: (1) for one-country studies, there is a need for a degree of freedom which implies resorting to local data; (2) monetary policy is centralized at the national (or supra-national) level and is therefore not an instrument available to local governments.

Conversely, when the incumbent expects to lose the election, it may prefer to use the debt strategically instead of generating a PBC. In doing so, it can constrain the next government by reducing the latter’s room for maneuver. In Borooah and van der Ploeg (1983), the “kamikaze” government, knowing it will lose power, pursues a ruinous economic policy, penalizing its successor to the point that the latter will not be reelected, which puts the suicidal government back at the helm. While Alesina and Tabellini (1990) predict an increase in debt whatever the party, Persson and Swensson (1989) highlight partisan differences according to the preferences of each party. The lowest spending party reduces taxes and in doing so, increases debt, thus forcing the next party to spend less. However, the highest spending party raises taxes, which creates a surplus and makes the next party’s policy easier to implement. This theory was tested with success by Pettersson-Lidbom (2001). In Tabellini and Alesina (1990), as the preferences of the next majority concerning the composition of expenditures are uncertain, the current median voter prefers higher debt to be sure that the next majority will have sufficient scope to implement its desired policy. Finally, in Aghion and Bolton (1990), the conservative party accumulates deficits when in power and voters reelect it as they deem the opposition (liberal) party unable to fight against deficits.

Popularity can be considered a proxy for economic conditions, which can also have a direct impact on PBC. In Efthyvoulou (2012), the magnitude of a PBC is weaker when non-economic voting (measured by replies to a survey about households’ financial situation) is high. In Hanusch (2012), economic disturbances blur competence signals, which dampen PBCs. In Canes-Wrone and Park (2012), the traditional PBC is invisible because it is cancelled by another cycle in private investment which drops before the election owing to electoral uncertainty. In Alpanda and Honig (2010), the government is constrained in generating a PBC by its borrowing capacity. Finally, endowment in natural resources can play a role (Vergne 2009).

This can be related to the growing literature on leviathan behavior (Brennan and Buchanan 1980), in which politicians are viewed as power-maximizing agents whose only purpose is to maximize the size of the public sector. In this literature, the electoral constraint modelled by the margin (the difference in vote shares or the distance from a 50 % vote share) affects budget outcomes. Ultimately, the empirical test looks very much like a classic PBC test in which the margin is substituted for the traditional electoral dummy. However, the interpretation is different. In the leviathan model, the margin is expected to be related positively to expenditure and taxes. The larger the margin, the greater the government’s capacity to increase expenditures and taxes (Dubois et al. 2007; Solé-Ollé 2003, 2006). In the PBC model, the larger the margin, the lower the incentive to intervene to secure reelection and the lower the expenditures and the higher the tax.

Several other institutional variables that may exert an impact on the magnitude of PBCs have been studied: the quality of institutions by Shi and Svensson (2006), government transparency by Akhmedov and Zhuravskaya (2004), corruption by Shi and Svensson (2006), globalization by Efthyvoulou (2011), social capital measured by the level of generalized trust by Kouvavas (2013), direct versus indirect elections by Sjahrir et al. (2013) and the degree of decentralization (Vergne 2009; Shelton 2014). Among these miscellaneous issues, the special role of the media emerges. As one has seen, information is crucial in the formation of PBCs as informational asymmetries could be at their heart. The better informed the voters, the better their evaluation of government competence, which diminishes the incentives for the government to generate a PBC. Shi and Svensson (2006), Akhmedov and Zhuravskaya (2004), Alt and Rose (2007), Vergne (2009) and Shelton (2014) have found that voters’ access to media or media freedom does indeed have an impact on the magnitude of a PBC.

Some of these studies do not deal explicitly with term limits but investigate an issue that is quite close, i.e., what happens to the PBC when the incumbent is seeking reelection (on the understanding that it is allowed to do so).

See also the panel data studies of Alesina et al. (1997) and Palmer and Whitten (2000). It should be noted that besides these empirical studies, analytical models of the choice of election date have been proposed (see, among others, Smith 1996; Kayser 2005).

In fact, through all of the settings that reduce inflationary bias: implementation of a rule (Barro and Gordon 1983), contract between the government and the central bank (Walsh 1995; see al-Nowaihi and Levine 1998, for the impact of such a contract on PBCs) or the appointment of a conservative central banker (Rogoff 1985).

One might ask why the government could choose to give up the monetary instrument which could be so important in generating PBCs and therefore helpful in reelection. First, monetary policy would be less useful because it is less adapted to targeted interventions toward groups of voters. Fiscal instruments are more easily manipulated and have more immediate and clear impacts. Political gains from inflation are relatively blurred. Second, the adoption of an independent central bank could be a sign of good will towards the financial market to obtain more credibility and a better reputation. Credibility and reputation are crucial to obtain the confidence of the financial market and this confidence should help finance the public debt, which is also a tool for generating PBCs. Third, adopting an independent central bank could reduce volatility; volatility generates uncertainty and penalizes investment and therefore economic growth, which is the key indicator in the voters’ eyes.

Usually measured by an index of the political constraints facing the executive when implementing policy.

Usually measured by an index coding several aspects of the budget process (e.g., budget cycle frequency, the existence of multi-year expenditure forecasts, the requirement of published performance measures). In a somewhat related approach, Mourao (2008) shows that fiscal illusion, in hiding the real fiscal situation, aggravates the PBC. Finally, financial transparency should not be mistaken for transparency understood as the absence of corruption, which can also influence the PBC (Shi and Svensson 2006, Klomp and de Haan 2013c).

Similarly, some scholars have studied PBCs in targeted taxes (e.g., Karran 1985; Ashworth and Heyndels 2002; Chaudhuri and Dasgupta 2005; Ehrhart 2013).