The Irrelevance of Political Party Differences for Public Finances – Evidence from Public Deficit and Debt in Portugal (1974–2012)

André Corrêa D’almeida; Paulo Reis Mourao

doi:10.1017/S1062798717000291

The Irrelevance of Political Party Differences for Public Finances – Evidence from Public Deficit and Debt in Portugal (1974–2012)

Published online by Cambridge University Press: 11 September 2017

André Corrêa D’almeida and

Paulo Reis Mourao

Show author details

André Corrêa D’almeida: Affiliation:
Columbia University, School of International and Public Affairs, 420 West 118th Street, Room 1435, New York, NY 10027, USA. Email: andre.dalmeida@columbia.edu
Paulo Reis Mourao: Affiliation:
Department of Economics, Economics & Management School, University of Minho, 4700 Braga, Portugal, & NIPE. Email: paulom@eeg.uminho.pt

Article contents

Abstract
Introduction
Deficits, Debts and Partisan Politics: An Overview of the Literature
Methodology
Empirical Models and Bivariate Analysis: The Irrelevance of Inter-party Difference
Conclusions and Further Research
Footnotes
References

Rights & Permissions

Abstract

This paper attempts to empirically test whether inter-party political differences impact public finances in Portugal differently. Focused on public debt and on government budget deficit, and using data since 1974 for several variables, this paper applies econometric modelling to show that inter-party differences have had, until now, no significant impacts on the public finances’ performance in Portugal. In this context, this paper aims at dispelling some myths regarding the ‘value’ of a policy process based on political intrigue, enmity and a discourse of confrontation around differentiated political parties’ merits in modern democracies.

Information

Type: Articles
Information: European Review , Volume 25 , Issue 4 , October 2017 , pp. 560 - 587

DOI: https://doi.org/10.1017/S1062798717000291 [Opens in a new window]
Copyright: © Academia Europaea 2017

1 Introduction

Since 1976, the policy process in Portugal has been dominated by a competition between socialists and social democrats, which are the groups that have been alternating in power for nearly equal proportions of time. ¹ The Socialist Party (PS) ruled the country for a total of 18 years after the first elections in April of 1976. Within the same period, the Social Democratic Party (PSD) ruled the country for a total of 21 years. ²

The current analysis will show that, for almost four decades, the political differences between the two main political parties had no impact on the public finances and economic performance of Portugal. That is, actual long-term policy outcomes are indifferent to the differences in the political ideologies and discourse of the socialists and the social democrats. Within this period, both parties have been equally associated with (i) cycles of growth, stagnation and recession (see Figure 1) and (ii) continuous budget deficits (see Figure 2).

Figure 1

Growth, stagnation and recession and ruling parties (%GDP). Sources: GDP data (Ref. 3), Ruling Parties (Ref. 4). Note: Years in the darker shade correspond to periods ruled by a Socialist government. Years in the lighter shade correspond to periods ruled by a Social Democrat government. The period 1970–1974 (until 25 April) corresponds to the last years of Salazar’s dictatorship. The period 1974–1976 corresponds to the so-called ‘Provisory Governments’ which ruled the country prior to the 1976 new Constitution. The second half of 1978 was ruled by an independent government.

Figure 2

Public deficit and ruling parties (%GDP). Source: Public Deficit (Ref. 3), Ruling Parties (Ref. 4). Note: The authors chose to use the Primary Deficit indicator, which excludes interests paid to cover the cost of the debt, to focus the analysis on actual expenditure decisions.

This research adopts two of the most important public finance variables, budget deficit and public debt, as proxy variables for the country’s performance to test the hypothesis that political parties’ differences have no impact in fiscal outcomes. Testing whether political differences impact the country’s performance is useful to shed light on the merits of a policy process that is based on competitive intrigue, enmity and confrontation between the two parties. Revenues are not considered to be proxies for performance (i.e. a dependent variable) in this analysis because tax policies are a much more resilient variable over time.

While the problems of inter-parties’ competition in the policy process have been highlighted, the merits of an alternative approach still need to be explained. Almond and VerbaReference Almond and Verba ⁵ argue that civic political culture is what a country needs to advance its democratic society’s political structure. More specifically, they consider political culture to be the element that connects individual attitudes with the overall political system structure. Of the ten key characteristics of a civic political culture identified by these authors, at least half of these refer directly to political parties’ behaviour: (i) expectation of fair treatment from government authorities; (ii) emotional involvement in elections; (iii) tolerance toward opposition parties; (iv) valuing of active participation in local government activities, parties and in civic associations; (vi) civic cooperation and trust.

Since Almond and Verba,Reference Almond and Verba ⁵ several other authors have emphasized the key role of cooperation to promote development and advance democratic organization, e.g. ColemanReference Coleman ⁶ with the concept of social capital; OstromReference Ostrom ⁷ with the idea of collective action; NordhausReference Nordhaus ⁸ with the political business cycles models; PutnamReference Putnam ⁹ with the concept of networks of civic engagement; WrightReference Wright ¹⁰ with the non-zero sum logic of human destiny; and HardinReference Hardin ¹¹ with the role of trust and reciprocity in solving social dilemmas.

The next section explores the literature capable of explaining the relationship between deficits, debts and partisan politics. Special attention is paid to authors who focus on variables that are most relevant to the Portuguese case. Section 3 details the techniques, variables and datasets used to test our research hypothesis. This section also offers some preliminary results regarding the role of inter-party differences in a country’s performance, while analysing the main results obtained from this research regarding a country’s performance, Section 4 pays special attention to the distinction between external macroeconomic effects and domestic politics. Section 5 concludes with the implications of the paper’s findings for the democratic process and suggestions for future research as well.

2 Deficits, Debts and Partisan Politics: An Overview of the Literature

There already exists a substantial literature on the differences in policies applied by different parties. CusackReference Cusack ¹² (1997) studied a set of 14 OECD countries for the period 1961–1991, but did not include Portugal because it was not a democracy during part of the period. MourãoReference Mourão ¹³ found that the long-term determinants of the Portuguese public expenditure (1947–2002) are the number of unemployed, the number of public employees (i.e. the size of the public sector), the rate of openness of the economy, and the real current transfers per capita. However, Mourão left unanswered the following question: are there any partisan-biased differences in the expenditure policy in Portugal?Reference Mourão ¹³ Our research aims at answering this question.

Our review of the literature (detailed below) identified four main topics on the issue of the influence of partisanship on public finances: due to differentiated redistributive policies, due to voters’ attraction, due to different reactions to exogenous forces such as globalization, and due to political business cycles.

2.1 Partisanship and Redistributive Policies

There are several theoretical models that try to explain partisan-biased differences in public finance, especially related to redistributive policies.

Usually, the number of unemployed suggests that an increase in expenditure is more associated with socialist-led governments. This may be the result of the commitment of leftist parties to the Keynesian paradigm of expanding government programmes and the welfare state,Reference Alesina, Cohen and Roubini ¹⁴ ^, Reference Hahm, Kamlet and Mowery ¹⁵ which is pushed by a relatively more economically vulnerable electorate.Reference Cusack ¹² CowartReference Cowart ¹⁶ and CarlsenReference Carlsen ¹⁷ argue that leftist governments are far more responsive to problems of unemployment than are right-wing governments that foreswear interfering with the self-adjusting ‘invisible hand’ of the market.Reference Cusack ¹⁸

CarlsenReference Carlsen ¹⁷ also argues that right-wing parties are more prone to pro-cyclical policies by which fiscal policy is tightened as demand slackens and unemployment rises. If real current transfers per capita are associated with economic redistribution processes, then, for the same reasons described above, this variable seems to be more associated with left-leaning governments.

According to Meltzer and Richard,Reference Meltzer and Richard ¹⁹ the redistributive role of the government is determined by voters’ demand for income redistribution. Voters with income below (above) the mean will favour higher (lower) taxes and more (less) redistribution. Thus, under universal suffrage and majority rule, a major determinant for explaining the fact that the real government debt per capita increases is income inequality.

Right-wing governments try to avoid (significant) budget deficits because they believe that structural deficits raise interest rates, impede investment, reduce national savings, slow down long-run economic growth, and thus undermine the social democratic supply-side economic policy.Reference Boix ²⁰

2.2 Public Sector and Voters

The size of the public sector follows the Leviathan hypothesisReference Brennan and Buchanan ²¹ according to which politicians are politically unconstrained agents who aim to maximize their power and the rents associated with holding office. In this context, power and rents are directly associated with the size of the public sector, and both socialists and democrats, who have been alternating in power in quasi-equal proportions since 1976, could equally be responsible for the increase in public expenditure (i.e. the Leviathan hypothesis emphasizes the parties’ behaviour homogeneity). Other authors point to the principle-agent dilemma as the justification for an inbuilt bureaucratic preference against any reduction of acquired rights by public employees and as supporting ever larger sizes of public agencies in a large public sector such as that of Portugal.Reference Frey and Pommerehne ²² ^, Reference Castles ²³

2.3 Other Mediating Factors: Competition and Globalization

Several authorsReference Besley, Persson and Sturm ²⁴ ^, Reference Padovano and Ricciuti ²⁵ suggest that party competition improves economic outcomes but what they do not discuss is how outcomes could improve further with collaborative approaches to the policy process. In addition, the baselines of their analyses are situations of monopoly of power by one party, which obviously cannot be compared with the multiparty framework studied in this paper.

Given the competitive setting that has been driving the political confrontation between socialists and social democrats over the past 39 years, the Partisan modelReference Boyne ²⁶ ^, Reference Solé-Ollé ²⁷ can also be relevant for this analysis. According to this theoretical framework, increased competition between political forces reduces deficits for left-leaning governments and increases them for right-leaning governments (i.e. the Partisan hypothesis emphasizes the heterogeneous parties’ behaviour). This result contradicts previously analysed theoretical principles that point to higher public expenditure in socialist-led governments.

Similar to the Partisan model, which introduces political competition as a mediating factor between ideology and policy outcomes, the rate of openness introduces macroeconomic interdependence as a mediating factor. CarlsenReference Carlsen ¹⁷ argues that parties on the left will follow a counter-cyclical expansionary fiscal policy when demand slackens and a contracting one when demand surges.

Garret and LangeReference Garrett and Lange ²⁸ argue that the ever-greater openness of national economies and the related interdependencies decrease the relevance of the governing parties’ ideology; i.e. there has been a convergence of right and left with no notable distinctions between their fiscal policies. Because of interdependence,

[G]overnments no longer possess the autonomy to pursue independent macroeconomic strategies effectively, even if they were to seek to do. In anything but the short run, the fiscal and monetary policies of governments of the left and the right should converge. (Ref. Reference Garrett and Lange28, p. 543)

Other authors agree that the economic demands generated by the global financial markets have weakened any linkage between partisanship and fiscal policy.Reference Artis ²⁹ ^– Reference Helleiner ³³ This applies even more to a small, open economy such as the Portuguese.

Contradicting the rate of openness-driven convergence argument, GarrettReference Garrett ³⁴ ^, Reference Garrett ³⁵ shows that strong left party governments with the presence of strong labour unions run larger deficits under the conditions of great capital mobility and trade because those governments try to mitigate the market dislocations that are experienced by their populations in the integrated international economy.

2.4 Non-partisanship and Political Business Cycles

Albeit from a different perspective, Nordhaus,Reference Nordhaus ³⁶ Frey and SchneiderReference Frey and Schneider ³⁷ and Rogoff and SibertReference Rogoff and Sibert ³⁸ argue that public expenditure drivers are more associated with the structural components of democratic models than with the differences in party ideology. For these authors, the temporal proximity to electoral moments is what explains public expenditure’s cyclical peaks. In fact, as illustrated by Figure 3, there seems to exist a temporal ‘coincidence’ in Portugal between election years and expenditure peaks.

Figure 3

Public expenditures, ruling parties and electoral cycles. Source: Government Expenditures (Ref. 3), Ruling Parties (Ref. 4), Electoral cycles (Ref. 39).

NordhausReference Nordhaus ⁸ ^, Reference Nordhaus ³⁶ explains these cyclical peaks with his political business cycles (PBC) theory. According to this author, these peaks result from the opportunistic behaviour of political incumbents who strategically change public expenditure in pre-electoral periods to signal competence (i.e. the ability to produce a given level of government services with less revenue) to voters and thus increase their chances of re-election. This opportunistic-cycles approach departs from the models of economic performance that are explained by political ideology and ideological cycles discussed earlier. For Nordhaus, partisan effects are of secondary importance to opportunistic behaviour.

Advocates of PBC argue that, in facing choices between present and future welfare, the behaviour of democratic political systems is biased against future generations.Reference Nordhaus ⁸ ^, Reference Coelho, Veiga and Veiga ⁴⁰ Elections are viewed as a source of political instability, and they affect growth through the shortening of governments’ horizons and the disruption of long term economic policies that are conducive to a better economic performance.Reference Aisen and Veiga ⁴¹

Veiga and VeigaReference Veiga and Veiga ⁴² find strong evidence for opportunistic cycles at the municipal expenditure level in Portugal and very little evidence of ideological cycles between left-wing oriented incumbents and right-wing ones. Coelho et al.Reference Coelho, Veiga and Veiga ⁴⁰ also find strong evidence of PBC in municipalities’ employment in Portugal.

2.5 In Summary

The literature that has been explored does not clearly answer the question raised at the beginning of this section: do partisan politics matter for public finances in Portugal? All types of answers – yes, no and maybe – have been found for this question. Studies about Portugal, where evidence for PBC was found and/or no evidence for ideological/partisanship cycles was found, do not adopt the whole country as their level of analysis, do not cover the entire period of democracy in Portugal, and do not clarify whether there is any partisanship effect in the amplitude of the cycle.

Next, we will test whether differences between different parties in charge of the government make any difference for the Portuguese public deficit and public debt.

3 Methodology

To test our research hypothesis, we use descriptive statistics and bivariate analyses that will provide initial insights into the role of inter-party differences in fiscal outcomes and the country’s performance. In this context, descriptive statistics and empirical evidence will also dispel some myths circulating in the public opinion regarding the role of parties in fiscal performance.

3.1 Variables: Datasets and Sources

Our first methodological step was to collect data for some of the most important variables that are associated with a country’s performance. We collected yearly data since 1974 ⁴³ from official and available sources for the following socio-economic variables:

∙ Public expenditure (% GDP)
∙ Public revenue (% GDP)
∙ Number of public employees
∙ Public investment (% GDP)
∙ Real GDP
∙ Budget deficit (i.e. public expenditure minus public revenue as a % of GDP) ⁴⁴
∙ Public debt (% GDP)
∙ Unemployment rate
∙ Number of unemployed people

Data for the following electoral and political variables were also collected for the same period:

∙ Electoral (legislative and municipal elections) year
∙ Number and share of seats belonging to left-wing parties in the Portuguese National Parliament
∙ Year of left-wing majority in the Portuguese National Parliament
∙ Tenure duration

Several sources of information were used to build the time series for these variables. For public expenditure, public revenue, public investment, real GDP, openness rate, unemployment rate, and the number of unemployed people, we used PinheiroReference Pinheiro ⁴⁵ for data until 1995 and the International Financial Statistics ³ for data from 1995 to 2012. NevesReference Neves ⁴⁶ was the source for the number of public employees for data before 1994. For the period that followed, the information was collected from the Portuguese Institute for Employment and Professional Training (IEFP). ⁴⁷ For the Portuguese public budget deficit and public debt, the data source was the IMF (2012). ³ The data for electoral years that detail the nature – municipal or legislative – of the popular consultation, was provided by the official National Electoral Commission (CNE). The information about the ruling parties at the national government level for each year, the breakdown of the number of Parliamentary seats between the left- and right-wing parties, and the tenure duration of each government was collected from PORDATA. ⁴⁸

3.2 Descriptive Statistics

To discuss our first descriptive statistics, we built two tables (see Tables 1 and Table 2). Each table shows the mean, standard deviation, maximum values, and minimum values for each of the previously introduced socio-economic variables and for the entire study period (1974–2012, N=39). In addition, these statistics have also been analysed within both those sub-periods in which left-wing parties held the majority of the seats in the National Parliament (N=18) and those in which right-wing parties had the majority of the seats (N=21). ⁴⁹

Table 1

Descriptive statistics, log scale (Portuguese data, 1974–2012).

Table 2

Descriptive statistics, absolute scale (Portuguese data, 1974–2012).

Table 1 shows the descriptive statistics for the socio-economic variables expressed in logarithmic form, and Table 2 shows the same statistics for the raw variables.

A preliminary inspection of Tables 1 and 2 allows us to state that the years in which left-wing parties ruled the country exhibit higher values for the following variables (if compared with the years ruled by right-wing parties):

∙ Public expenditure (% GDP)
∙ Public revenue (% GDP)
∙ Number of public employees
∙ Public investment (% GDP)
∙ Real GDP
∙ Openness rate
∙ Budget deficit (% GDP)
∙ Public debt (% GDP)

The exceptions to this statement were found for the unemployment rate and the number of unemployed people. For these variables, the years ruled by left-wing governments exhibited lower values than the years ruled by right-wing governments.

Table 3 shows the same descriptive statistics that were analysed above but for the yearly growth rates of the studied variables.

Table 3

Descriptive statistics, yearly growth rates (Portuguese data, 1975–2012).

Note: The yearly growth rate for two consecutive values of the variable X (X _t, X _t+1) is given by the formula (X _t+1 – X _t)/(X _t)

Table 3 suggests that the years in which there was a right-wing majority in the Portuguese Parliament are characterized by higher growth rates for the following variables (if compared with the years that are associated with left-wing governments):

∙ Public expenditure (% GDP)
∙ Public revenue (% GDP)
∙ Public employees
∙ Real GDP
∙ Openness rate
∙ Unemployed

In contrast, the years that are associated with left-wing governments tend to have higher growth rates for the following variables (if compared with the years that are associated with right-wing governments):

∙ Public investment (% GDP)
∙ Budget deficit (% GDP)
∙ Public debt (% GDP)
∙ Unemployment rate

However, these preliminary results require further inspection. Many of the comparisons arguably are commonplaces in the Portuguese media. However, a more accurate assessment of these commonplaces requires data variability and that the different number of observations collected for each political wing are factored in using more complex approaches. A commonly used technique to address these limitations is the non-parametric Mann-Whitney U-test, which is similar to the Two-sample t test with unequal variances. This type of test allows the equality of two means to be tested (in our case: the mean observed for the years that were ruled by left-wing parties and the mean observed for the years that were ruled by right-wing parties). The rejection of the null hypothesis (of equal means) can be interpreted as the existence of significant differences between the financial performances that are assumed by the different Portuguese political wings.

Tables 4, 5 and 6 exhibit the asymptotic significance (2-tailed) p-values for the variables of this study. ⁵⁰ Because p-values lower than 0.100 were not found, it cannot be concluded that there are significant differences in our variables vis-à-vis the party wing in power.

Table 4

Two-sample t test with unequal variances (Yearly growth rates for Portuguese data, 1975–2012).

Note: The yearly growth rate for two consecutive values of the variable X (X _t, X _t+1) is given by the formula (X _t+1 – X _t)/(X _t)

Table 5

Two-sample t test with unequal variances (log values for Portuguese data, 1974–2012).

Table 6

Two-sample t test with unequal variances (absolute values for Portuguese data, 1974–2012).

In conflict with popular beliefs and other misconceptions of the role of political parties in modern democracies, our results find no evidence for a differentiated impact of the Portuguese parties on the country’s macroeconomic and fiscal performances. This is the case for all variable transformations performed: absolute values, log and yearly growth rates.

4 Empirical Models and Bivariate Analysis: The Irrelevance of Inter-party Difference

The next step is to build two models to test and discuss the hypothetical influence of political wings on the two macroeconomic variables that are most commonly used to describe country performance: the budget deficit and the public debt (each of which are measured both as a percentage of GDP and as yearly changes). The results obtained from the Mann-Whitney U-tests suggest that there are no differences among the ruling parties when considering each variable in isolation. However, we also need to evaluate this relationship by considering multivariate regressions. Therefore, the following two equations will be estimated (equation (1) for public budget deficit and equation (2) for public debt):

(1)

$$budget_{t} =const{\plus}A\,{\asterisk}\,X_{t} {\plus}B\,{\asterisk}\,Elect_{t} {\plus}C\,{\asterisk}\,Polit_{t} {\plus}e_{t} $$

(2)

$$debt_{t} =const{\plus}A\,{\asterisk}\,X_{t} {\plus}B\,{\asterisk}\,Elect_{t} {\plus}C\,{\asterisk}\,Polit_{t} {\plus}e_{t} $$

On their right-hand side, both of these equations use a column-vector of socio-economic determinants that is suggested by the literature (X _t, with the estimated coefficients represented by the line-vector A), a column-vector of electoral determinants (Elect _t, with the estimated coefficients represented by the line-vector B), and a column-vector of political-wing variables (Polit _t, with the estimated coefficients represented by the line-vector C). The errors are assumed to follow white-noise processes. These variables follow authors who have studied the budget deficit and the public debt in PortugalReference Mourão ¹³ and authors who have analysed the influence of partisan politics on national public finances.Reference Cusack ¹² ^, Reference Cowart ¹⁶ ^, Reference Carlsen ¹⁷ Table A1 exhibits the results from the ADF tests on the stationarity of these series.

Tables 7 and 8 show the coefficients that were estimated by the Static Ordinary Least Squares for the various specifications of our models by considering the two transformations of the Portuguese public deficit – logs and yearly growth rates – and their stochastic explicative variables. Table 9 shows the coefficients that were estimated for the log variables but now using different estimators – the Dynamic Ordinary Least Squares (DOLS), ⁵¹ a Vector of Error Correction Models (VECM), and a model derived from autoregressive distributed lags (ARDL).Reference Pesaran, Shin and Smith ⁵² These estimators lead to a better understanding of the nature of the non-stationarity of the variables, and their estimates provide a clearer idea of the direction and significance of the influence of our explicative variables on public debt and the public deficit.

Table 7

SOLS results for Portuguese budget deficit (1974–2012).

Constant estimated but omitted in the table. Standard errors between parentheses. Significance levels: 1%, ***. 5%, **. 10%, *. ADF (1): Augmented Dickey-Fuller test for unit root on the residual series.

Table 8

SOLS results for Portuguese budget deficit (1974–2012).

Table 9

DOLS, VECM and ARDL results for Portuguese budget deficit (1974–2012).

Table 10

SOLS results for Portuguese Public Debt (1974–2012).

Table 11

SOLS results for Portuguese Public Debt (1974–2012).

Table 12

DOLS, VECM and ARDL results for Portuguese Public Debt (1974–2012).

Table 13

Causality tests on budget deficit.

Table 14

Causality tests on public debt.

To measure the influence of political wings, three alternative variables have been introduced in the model:Reference Castles ²³ ^, Reference Cameron ⁵³ ^, Reference Elkins ⁵⁴

∙ A binary variable that signals each year in which there was a majority of left-wing parties in the national parliament.
∙ The share of seats for left-wing parties in the Portuguese parliament.
∙ Length of tenure.

We recall that the descriptive statistics of these variables are shown in Table 1 (log scale) and in Table 3 (yearly growth rates).

Finally, we want to analyse the statistical significance of the political wings on the Portuguese budget deficit and on the Portuguese public debt. To do this, we are going to use the traditional binary variable that identifies those years that were ruled by a left-wing coalition of parties as ‘1’ and those years that were ruled by a right-wing coalition of parties as ‘0’. Additionally, we are going to use the share of seats that belong to left-wing parties in the national parliament. Therefore, the null hypothesis of the statistical non-significance of the estimated coefficients for these political variables shall be interpreted, if accepted, as the confirmation of the irrelevance of the parties’ differences for public finance management in Portugal. We control these political variables using the tenure length, which is measured as the number of years that the incumbent party holds the office. This is a common procedure for two main reasons. First, the procedure recognizes that the tenure length can be used to control for the awareness effect that is more visible in governments that have been in power for more years,Reference Ferris and Moia ⁵⁵ i.e. more mature governments tend to have more knowledge of national public finances. Second, the procedure identifies that more mature governments have a greater propensity to run consolidation programmes on national public finances than do governments with less experience.Reference Merlo ⁵⁶ ^, Reference Désiré ⁵⁷

Tables 10–12 show the results for Portuguese public debt as the dependent variable. Table 10 shows the results for the SOLS estimation on the logs of our stochastic variables and our set of binary independent variables. Table 11 shows the results for the SOLS estimation using the yearly growth rates of the stochastic variables and our set of binary independent variables. Table 12 shows the DOLS, the VECM and the ARDL results as done for the budget deficit in Table 9 but now using public debt as the dependent variable.

Briefly analysing Tables 10–12, we observe again that electoral moments or political wings do not influence Portuguese public debt. However, we find again that Portuguese public debt (as a share of the national income) tends to react to the number of unemployed people (which confirms the dependence of the structure of the Portuguese Welfare State) and to the real GDP per capita (i.e. to the economic cycle of the country).

Again, these results are supported by Meltzer and Richard,Reference Meltzer and Richard ¹⁹ Garret and Lange,Reference Garrett and Lange ²⁸ PetersReference Peters ³⁰ and ScharpfReference Scharpf ³¹ who have identified that the national public finances of highly globalized and small economies are driven mostly by socio-economic dimensions and not by inter-parties differences.

To test the causality of the determinants that were found to be statistically significant, we ran Granger Causality tests and Instantaneous Causality tests using JMulti. Table 13 reports the results for the budget deficit and for public debt.

As reported in Table 13, we confirm that there is a statistically significant causal effect of the Portuguese trade openness, the number of unemployed people, and the real GDP per capita on the public budget deficit. This again shows that the past changes in the Portuguese level of globalization, its labour market and its economic growth induce future changes in the Portuguese public budget deficit. Although it is significant in the cointegration equation, the number of Portuguese public employees is not shown to anticipate the budget deficit. To interpret these results, we view this as a case of short-term causation that tends to lose statistical significance in the long-term. Table 14 also confirms that the past changes in the Portuguese number of unemployed people and GDP lead to subsequent changes in the Portuguese level of public indebtedness.

We recognize that these results have pioneering implications for the study of the relationship between structural economic and social realities, electoral moments, political wings and public finances in Portugal. They clarify which factors truly influence the Portuguese public budget deficit and the Portuguese level of public debt.

5 Conclusions and Further Research

There is no empirical evidence to support the hypothesis that differences between political parties matter for the public finances or for the overall financial and economic performance of Portugal. It is simply not true that one party is better for the country than the other, at least not in the last 39 years, which is the entire period of democracy in the country. ParkerReference Parker ⁵⁸ finds similar results for the US when comparing the performances of the Republican Party and the Democratic Party during the period 1949–2005.

The policy processes and political rhetoric based on competitive intrigue, enmity and confrontation between parties keep promising differentiated economic and social outcomes depending on who is in power: PS or PSD (either alone or in coalition). While commonplace in popular belief and partisanship, this research found no statistical significance to support this differentiated effect. Using various techniques and procedures such as graphic analysis, descriptive statistics, and Mann-Whitney tests, it cannot be claimed that the years that were ruled by one of the Portuguese political wings produced significantly different outcomes than the years that were ruled by the opposite political wing. Future research could focus on whether these undifferentiated results between parties also apply to the composition of expenditures.

This conclusion applies not only for the main public finance variables studied – budget deficit and public debt – but also for the GDP growth rates, unemployment rates, size of the public sector and public investment. These indicators have performed indifferently to the political party differences.

Three main strong implications for Portuguese democracy and the policy process can be derived from these results. First, the party differences do not matter but the parties matter. In a democracy, political parties try to build unique identities and support bases to be elected for office. While these differences should be emphasized during the official electoral campaign periods, outside these periods, political leaders should behave according to a certain ethical code that would allow them to embrace legislation from a national priority standpoint and not from a competing democracy paradigm. For example, it was always possible during the three times that the IMF intervened in the country since 1974, and during the Euro process, to have a broad inter-party political pact regarding institutional development and international commitment. Future research could try to understand why these domestic pacts are rare in Portugal without international intervention. Why do specific configurations of situations and conditions increase or decrease political cooperation? What institutional changes would enhance the likelihood of cooperation?

Second, Portugal, and modern democracies in general, need new institutional designs that aim at creating a better incentives structure to foster collaboration between political parties. The central theme here is the gains from association that are achieved when politicians are able to develop trust and reciprocity, and coordinate action. If politicians are the legitimate representatives of the people, the question is how to sustain agreements that counteract individual temptations to select short-term self-interested actions when all parties would be better off if each party selected actions that lead to higher collective and individual returns. Short-term competing approaches to the policy process, as illustrated by the intensity of political quarrels that every day are brought to us in several formats (e.g. the press, social media, political discourse), consume too much of the energy, attention and political vision that a long-term strategic debate for the country would require from all group and individual actors.

Third, collaboration and coordinated action require the development of trust and reciprocity between political parties. As empirically shown, the current schemes of adversarial democracy have failed in Portugal and, apparently, in other southern European countries such as Spain, Italy and Greece. Trust building is an evolutionary process of multiple iterations through which parties generate cognitive systems and decision rules that lead to outcomes that are greater than those of alternative designs over time.

Party differences, ideology and identity matter for the diversity of political ideas, the engagement of civic participation in the lives of communities and the advancement of societies. What the results of this research show is that, because neither of the two major parties has been better for the country than the other, differences should be reconciled for policy enhancement and should not undermine the collective good. Ideas should still compete in the political arena, but collaboration that is driven by an ethics of trust and reciprocity and/or the need for a broader social and political support base should be the norm in policy design and implementation. Moving forward, the focus should be on leveraging the overlaps between the parties’ agendas and negotiation skills.

Two last suggestions are given for further research that can be derived from this study. The first is the role that the regulation of the activities of political parties could play in fostering collaborative work between political leaders. Indeed, the failure to cooperate for mutual benefit does not always signal ignorance or irrationality. Game theorists explain in various ways how contextual conditions such as history, rules and norms within which games are played, influence players’ behaviour and outcomes. NorthReference North ⁵⁹ ^, Reference North ⁶⁰ and OstromReference Ostrom ⁷ ^, Reference Ostrom ⁶¹ argue that if rules change, so will outcomes. The second suggestion is the use of multivariate econometric models, which will clarify the significance of domestic inter-party differences for the performance of public finances within the broader context of economic globalization. For this step, several economic, institutional, political and social variables should also be considered.

Acknowledgements

The work of Mourao for this article was carried out within the funding with COMPETE reference no. POCI-01-0145-FEDER-006683, with the FCT/MEC’s (Fundação para a Ciência e a Tecnologia, I.P.) financial support through national funding and by the ERDF through the Operational Programme on Competitiveness and Internationalization – COMPETE 2020 under the PT2020 Partnership Agreement.

André Corrêa d’Almeida PhD, MSc, is the Assistant Director and Adjunct Associate Professor for the School of International and Public Affairs and The Earth Institute’s Master of Public Administration in Development Practice, Columbia University. He has more than 20 years of entrepreneurial, management, academic, and leadership experience with applied multidisciplinary programmes, teaching, research and consultancy carried out in the US, Europe, Central Asia, Africa, the Middle East and China. He is Senior Advisor to the United Nations Development Program (UNDP) for programme evaluation and Founder of ARCx-Applied Research for Change Center. He has published his research in several areas of Social Economics, Development Economics and Institutional Development, and is the scientific coordinator and editor of Smarter New York City: How City Agencies Innovate (Columbia University Press, 2018). He is also a regular contributor to the Huffington Post and The Guardian. Follow him on Twitter @AndreCdAlmeida.

Paulo Reis Mourao is an Assistant Professor at the University of Minho, Braga - Portugal. He holds a PhD in Economics. His main research focuses on Applied Economics, Public Finances, and Sports Economics. He is the author/co-author of articles published in Journal of Economic Literature, Environment and Planning A, Environment and Planning C, Public Finance Review, Kyklos, Applied Economics.

Table A1

Augmented Dickey-Fuller Statistics, 1974–2012.

Footnotes

Note: Significance level: 10%, *; 5%, **; 1%, ***. Between parentheses, optimum number of lags using Schwarz Criteria.

d identifies the level of differentiation of y _t.

References

1.For a detailed description of each party’s ideological inclinations, consult www.ps.pt and www.psd.pt Google Scholar

2.The authors classify the eight-month long government of Carlos Mota Pinto (22 November 1978 – 2 August 1979) as Social Democrat and the five-month long government of Maria de Lourdes Pintasilgo (2 August 1979 – 3 January 1980) as Socialist.Google Scholar

3.International Monetary Fund. IMF (Several years). IMF Data Mapper. Available at http://www.imf.org/external/datamapper/index.php (accessed 28 May 2013).Google Scholar

4. Portugal Gov. (2013). Government of Portugal. Available at http://www.portugal.gov.pt/pt/o-governo/arquivo-historico/chefes-de-estado/chefes-de-estado.aspx (accessed 21 August 2013).Google Scholar

5. Almond, G. and Verba, S. (1963) The Civic Culture: Political Attitudes and Democracy in Five Nations (Princeton, NJ: Princeton University Press).Google Scholar

6. Coleman, J. (1988) Social capital in the creation of human capital. American Journal of Sociology, Supplement 94, pp. S95–S120.CrossRef Google Scholar

7. Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions for Collective Action (New York: Cambridge University Press)Google Scholar

8. Nordhaus, W. (1990) Alternative approaches to the political business cycle. Cowles Foundation Paper No 748.Google Scholar

9. Putnam, R. (1992) Making Democracy Work: Civic Tradition in Modern Italy (Princeton, NJ: Princeton University Press).Google Scholar

10. Wright, R. (2001) Nonzero: the Logic of Human Destiny (New York: Vintage Books).Google Scholar

11. Hardin, R. (2005) Gaming trust. In: E. Ostrom and J. Walker (Eds.), Trust and Reciprocity: Interdisciplinary Lessons from Experimental Research (pp. 80–101) (New York: Russel Sage Foundation).Google Scholar

12. Cusack, T. (1997) Partisan politics and public finance: Changes in public spending in the industrialized democracies, 1955–1989. Public Choice, 91, pp. 375–395.Google Scholar

13. Mourão, P. (2007) Long-Term Determinants of Portuguese Public Expenditures. International Research Journal of Finance and Economics, 7, pp. 153–167.Google Scholar

14. Alesina, A., Cohen, G. and Roubini, N. (1993) Electoral business cycles in industrial democracies. European Journal of Political Economy, 9, pp. 1–23.Google Scholar

15. Hahm, S., Kamlet, M. and Mowery, D. (1995) Influences on deficit spending in industrialized democracies. Journal of Public Policy, 15(2), pp. 183–197.Google Scholar

16. Cowart, A. (1978) The economic policies of European governments, Part II: Fiscal policy. British Journal of Political Science, 9, 425–440.Google Scholar

17. Carlsen, F. (1997) Counter fiscal policies and partisan politics: Evidence from industrialized countries. Applied Economics, 29, pp. 145–151.Google Scholar

18. Cusack, T. (2001) Partisanship in the setting and coordination of fiscal and monetary policies. European Journal of Political Research, 40, pp. 93–115.Google Scholar

19. Meltzer, A. and Richard, S. (1981) A rational theory of the size of government. Journal of Political Economy, 89(51), pp. 914–927.Google Scholar

20. Boix, C. (1998) Political Parties, Growth and Equality: Conservative and Social Democratic Economic Strategies in the World Economy (Cambridge, UK: Cambridge University Press).Google Scholar

21. Brennan, G. and Buchanan, J. (1980) The Power to Tax: Analytical Foundations of a Tax Constitution (Cambridge, UK: Cambridge University Press).Google Scholar

22. Frey, B. and Pommerehne, W. (1984) The hidden economy: State and prospects for measurement. Review of Income and Wealth, 30(1), pp. 1–23.Google Scholar

23. Castles, F. (Ed.) (1982) The Impact of Parties: Politics and Policies in Democratic Capitalist State (Beverly Hills: Sage).Google Scholar

24. Besley, T., Persson, T. and Sturm, D. (2005) Political competition and economic performance: Theory and evidence from the United States. NBER Working Paper No. 11484.Google Scholar

25. Padovano, F. and Ricciuti, R. (2008) Italian Institutional Reforms: A Public Choice Perspective (New York: Springer).Google Scholar

26. Boyne, G. (1994) Party competition and local spending decisions. British Journal of Political Science, 35, pp. 210–222.Google Scholar

27. Solé-Ollé, A. (2006) The effects of party competition on budget outcomes: Empirical evidence from local governments in Spain. Public Choice, 126(1/2), pp. 145–176.Google Scholar

28. Garrett, G. and Lange, P. (1991) Political responses to interdependence: What’s ‘left’ for the left? International Organization, 45, pp. 539–564.Google Scholar

29. Artis, M.J. (1987) Deficit spending. In: J. Eatwell, M. Milgate and P. Newman, (Eds), The New Palgrave: A Dictionary of Economics (London: Macmillan).Google Scholar

30. Peters, B. (1991) The Politics of Taxation: A Comparative Perspective (Oxford: Blackwell).Google Scholar

31. Scharpf, F. (1991) Crisis and Choice in European Social Democracy (New York: Cornell University Press).Google Scholar

32. Kurzer, P. (1993) Business and Banking (Ithaca: Cornell University Press).Google Scholar

33. Helleiner, E. (1994) States and the Reemergence of Global Finance: From Bretton Woods to the 1990s (Ithaca: Cornell University Press).Google Scholar

34. Garrett, G. (1996) Capital mobility, trade, and the domestic politics of economic policy. In: R.O. Keohane and H.V. Milner (Eds.), Internationalization and Domestic Politics (pp. 79–107) (Cambridge, UK: Cambridge University Press).Google Scholar

35. Garrett, G. (1998) Partisan Politics in the Global Economy (Cambridge, UK: Cambridge University Press).Google Scholar

36. Nordhaus, W. (1975) The political business cycle. The Review of Economic Studies, 42(2), pp. 169–190.Google Scholar

37. Frey, B. and Schneider, F. (1981) A politico-economic model of the U.K.: New estimates and predictions; The Economic Journal, 91(363), pp. 737–740.Google Scholar

38. Rogoff, K. and Sibert, A. (1988) Elections and macroeconomic policy cycles. Review of Economic Studies, 55, pp. 1–16.Google Scholar

39. Comissão Nacional de Eleições, CNE (2013) Resultados Eleitorais. Available at http://eleicoes.cne.pt/sel_eleicoes.cfm?m=raster (accessed 13 May 2013).Google Scholar

40. Coelho, C., Veiga, F. and Veiga, L. (2006) Political business cycles in local employment: Evidence from Portugal. Economic Letters, 93, pp. 82–87.Google Scholar

41. Aisen, A. and Veiga, F. (2011) How does political instability affect economic growth? IMF Working Paper, International Monetary Fund.CrossRef Google Scholar

42. Veiga, L. and Veiga, F. (2007) Political business cycles at the municipal level. Public Choice, 131(1/2), pp. 45–64.Google Scholar

43.This year (1974) is the year of the Carnation Revolution. Our data also include the years when Portugal had to fulfil the Maastricht criteria related to the Stability and Growth Pact (since 1997) and the periods of financial and economic crisis when Portuguese fiscal autonomy has been highly influenced by the International Monetary Fund (1978–1979 and 1983–1985), the European Central Bank and, the European Commission (since 2011). Therefore, we have to recognize that, since 1974, a considerable number of years can be characterized by these external influences, which put special pressures on the parties’ strategies.Google Scholar

44.We also used the cyclical adjusted deficit (as a percentage of GDP) by using, as a source, Eurostat (2013) Cyclical adjust balance (Portugal, several years). Available at http://eurostat.eu/ (accessed 4 October 2013). The results that we obtained are similar to those here discussed. Full details are available upon request.Google Scholar

45. Pinheiro, M. (Coord.). (1997) Séries Longas para a Economia Portuguesa, Vol. I (Lisboa: Séries Estatísticas do Banco de Portugal).Google Scholar

46. Neves, J. (1994) The Portuguese Economy: A Picture in Figures (Lisboa: Universidade Católica Editora).Google Scholar

47. Instituto do Emprego e Formação Profissional, IEFP (Several years) (Lisboa: Relatórios anuais do Mercado do Emprego).Google Scholar

48. PORDATA (2013) Base de dados do Portugal Contemporâneo. Available at http://www.pordata.pt/en/Home (accessed 20 August 2013).Google Scholar

49.We considered the following parties and movements/coalitions to be Portuguese left-wing parties: PS (Partido Socialista, the main Portuguese Socialist Party), PCP (Partido Comunista Português, the main Portuguese Communist Party), MDP (Movimento Democrático Português), UDP (União Democrática Portuguesa), APU (Aliança do Povo Unido), FRS (Frente Republicana e Socialista), CDU/PCP-PEV (a coalition of PCP with the environmentalist party Os Verdes), and BE (Bloco de Esquerda, a coalition of heterogeneous leftist movements). Source: http://eleicoes.cne.pt/raster/index.cfm?dia=16&mes=12&ano=1979&eleicao=am Google Scholar

50.For a more complete description, please contact the authors.Google Scholar

51.Using the case of the public budget (also replicated for public debt), the respective DOLS is given by budget _t =const+A*X _t+B*Elect _t+C*Polit _t +D*L _t+e _t where L _t represents a set of leads and lags of the first differences of the explicative variables. J. Stock and M. Watson (1993) A simple estimator of cointegrating vectors in higher order integrated systems. Econometrica, 61(4), pp. 783–820.Google Scholar

52. Pesaran, M., Shin, Y. and Smith, R. (2001) Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), pp. 289–326.Google Scholar

53. Cameron, D. (1978) The expansion of the public economy: A comparative analysis. American Political Science Review, 72, pp. 1243–1261.Google Scholar

54. Elkins, D. (1974) The measurement of party competition. American Political Science Review, 68, pp. 682–700.Google Scholar

55. Ferris, F. and Moia, M. (2009) What determines the length of a typical Canadian parliamentary government? Canadian Journal of Political Science, 42(4), pp. 881–910.Google Scholar

56. Merlo, A. (1998) Economic dynamics and government stability in postwar Italy. The Review of Economics and Statistics, 80(4), pp. 629–637.Google Scholar

57. Désiré, O. (2007) Oil rents and the tenure of the leaders in Africa. Economics Bulletin, 3(42), pp. 1–12.Google Scholar

58. Parker, E. (2006) Does the party in power matter for economic performance? UNR Economics Working Paper Series. Working Paper No. 06-008.Google Scholar

59. North, D. (1990) Institutions, Institutional Change and Economic Performance (New York: Cambridge University Press).Google Scholar

60. North, D. (2005) Understanding the Process of Institutional Change (Princeton, NJ: Princeton University Press).Google Scholar

61. Ostrom, E. (2005) Understanding Institutional Diversity (Princeton, NJ: Princeton University Press).Google Scholar

Figure 1 Growth, stagnation and recession and ruling parties (%GDP). Sources: GDP data (Ref. 3), Ruling Parties (Ref. 4). Note: Years in the darker shade correspond to periods ruled by a Socialist government. Years in the lighter shade correspond to periods ruled by a Social Democrat government. The period 1970–1974 (until 25 April) corresponds to the last years of Salazar’s dictatorship. The period 1974–1976 corresponds to the so-called ‘Provisory Governments’ which ruled the country prior to the 1976 new Constitution. The second half of 1978 was ruled by an independent government.

Figure 2 Public deficit and ruling parties (%GDP). Source: Public Deficit (Ref. 3), Ruling Parties (Ref. 4). Note: The authors chose to use the Primary Deficit indicator, which excludes interests paid to cover the cost of the debt, to focus the analysis on actual expenditure decisions.

Figure 3 Public expenditures, ruling parties and electoral cycles. Source: Government Expenditures (Ref. 3), Ruling Parties (Ref. 4), Electoral cycles (Ref. 39).