HAS THE CREATION OF THE
EURO LED TO AN INCREASE IN
THE USE OF POLITICAL
BUDGET CYCLES?
Ben Roberts (606064) – BSc Economic...
P a g e | i
Abstract
This paper investigates the existence of political budget cycles (PBCs) in the European Union
(EU) us...
P a g e | ii
Contents
List of Figures .......................................................................................
P a g e | iii
List of Figures
Figure 1 – Greece Budget Balance………………………………………………………………………………………7
Figure 2 – Finland Budget...
P a g e | iv
List of Tables
Table 1 – Equation 1 Output…………………………………………………………………………………………….……………..12
Table 2 – Equation 2 ...
P a g e | v
List of Abbreviations
CPI – Consumer Price Index
EC – European Commission
ECB – European Central Bank
EMU – Eu...
P a g e |1
1. Introduction
The term Political Budget Cycle (PBC) refers to cycles in a domestic government’s budget
balanc...
P a g e |2
2. Literature Review
This paper begins with a comprehensive review of the current literature on the topic of
PB...
P a g e |3
state governments do indeed abstain from raising deficits in the pre-election period as a
result of fiscal tran...
P a g e |4
3. Theoretical Considerations
The existence of political cycles was first introduced by Rogoff and Sibert in 19...
P a g e |5
endogenous. By using data on British elections from 1992 – 1997 they find evidence
supporting the contention th...
P a g e |6
4. Data
To empirically test for the existence of PBCs within the Eurozone, annual time series data is
collected...
P a g e |7
5. Budget Balance and Election Timing – A Brief Overview
The graphs displayed below show the budget balance of ...
P a g e |8
6. Expected findings
Based on the results of previous studies and the theory highlighted above, the expectation...
P a g e |9
7. Methodology – Empirical Strategy
The main hypothesis to be tested by this investigation is whether or not th...
P a g e |10
This variable has been lagged due to voters assessing economic performance retrospectively
as mentioned above....
P a g e |11
8. Empirical Results
8.1. Evidence of political budget cycles over the sample period
The empirical analysis is...
P a g e |12
level of significance and hence it is possible to reject the null hypothesis in favour of the
alternate and as...
P a g e |13
8.2. Evidence of political budget cycles in the Eurozone
Following the discovery of a structural break in the ...
P a g e |14
is insignificant at the 10% level of significance it provides no explanatory power when
testing for the hypoth...
P a g e |15
prove that the SGP has been unsuccessful in its task to control fiscal spending and limit the
levels of debt o...
P a g e |16
Table 2 – Equation 2 Output
Dependant Variable: Budget
Method
Panel Ordinary Least
Squares
Sample Period: 1971...
P a g e |17
Table 3 – Chow Test Output
Table 4 – Wald Test Output
Wald Test
Test Equation: Equation
Test Statistic Value d...
P a g e |18
8.3. Robustness Checks
The results presented above provide clear, statistically significant, empirical evidenc...
P a g e |19
Equation 4
𝐵𝑢𝑑𝑔𝑒𝑡𝑖,𝑡 = ∁0 + 𝐷1 𝑃𝑟𝑒𝑖,𝑡 + 𝐷2 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 + 𝐷3 𝑃𝑜𝑠𝑡𝑖,𝑡 + 𝐷4 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐷5 𝑃𝑟𝑒𝑖,𝑡 ∗
𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐷6 𝑌𝑒𝑎...
P a g e |20
reinforces both the importance of including the lagged budget variable and the fact the
current year’s budget ...
P a g e |21
Table 6 represents the results of the OLS estimation based on equation 3, which
incorporates the effects of th...
P a g e |22
Table 6 – Equation 4 Output
Dependant Variable: Budget
Method Panel Ordinary Least Squares
Sample Period: 1971...
P a g e |23
9. Conclusions
This paper makes a significant and innovative contribution to the existing literature on PBCs.
...
P a g e |24
10. Bibliography
Alt, J. E., & Lassen, D. D. (2006). Transparency, political polarization, and political budge...
P a g e |25
Hallerberg, M., de Souza, L. V., & Clark, W. R. (2002). Political business cycles in EU accession
countries. E...
P a g e |26
11. Appendix 1 – Chow Test
Chow Test = ((Sc - (S1+ S2)) / (k))⁄((S1 + S2 ) /(N1+N2- 2k))
Where Sc is the sum o...
of 32

Political Budget Cycles

Published on: Mar 4, 2016
Source: www.slideshare.net


Transcripts - Political Budget Cycles

  • 1. HAS THE CREATION OF THE EURO LED TO AN INCREASE IN THE USE OF POLITICAL BUDGET CYCLES? Ben Roberts (606064) – BSc Economics Neeraj Somaiya (651412) – BSc Economics Harry Hayden (654442) – BSc Economics, Finance and Banking Jessie Gore (604372) – BSc Economics, Finance and Banking Taylor Woodhams (623136) – BSc Business Economics Supervisor – Petros Sekeris 23RD MARCH, 2015 UNIVERSITY OF PORTSMOUTH
  • 2. P a g e | i Abstract This paper investigates the existence of political budget cycles (PBCs) in the European Union (EU) using panel data from 15 large European economies from 1970 – 2013, generating 534 panel observations. The research further goes on to explore the Euro specific effects on the engineering of PBCs within the sample countries. The empirical analysis finds no evidence to support the use of PBCs throughout the sample period. However as the analysis is extended further to examine the effects of the Euro, there is statistically significant evidence supporting the hypothesis that since joining the Euro, the governments’ of the sample countries have manipulated fiscal policy in the periods leading up to and during an election in attempt to gain re-election. These findings lead to questions being drawn on the effectiveness of the Stability and Growth Pact (SGP). The paper is organized as followed. The subsequent section discusses the related empirical literature on PBCs while the methodological issues are discussed in Section three. A description of the data is provided in Section four and section five discusses the empirical results. Section six concludes.
  • 3. P a g e | ii Contents List of Figures ......................................................................................................................................... iii List of Tables .......................................................................................................................................... iv List of Abbreviations ............................................................................................................................... v 1. Introduction ....................................................................................................................................1 2. Literature Review............................................................................................................................2 3. Theoretical Considerations.............................................................................................................4 4. Data.................................................................................................................................................6 5. Budget Balance and Election Timing – A Brief Overview................................................................7 6. Expected findings............................................................................................................................8 7. Methodology – Empirical Strategy..................................................................................................9 8. Empirical Results...........................................................................................................................11 8.1. Evidence of political budget cycles over the sample period.................................................11 8.2. Evidence of political budget cycles in the Eurozone.............................................................13 8.3. Robustness Checks................................................................................................................18 9. Conclusions ...................................................................................................................................23 10. Bibliography ..............................................................................................................................24 11. Appendix 1 – Chow Test............................................................................................................26
  • 4. P a g e | iii List of Figures Figure 1 – Greece Budget Balance………………………………………………………………………………………7 Figure 2 – Finland Budget Balance………………………………………………………………………………………7 Figure 3 - Graphical representation of expected results…………………………………………….……….8
  • 5. P a g e | iv List of Tables Table 1 – Equation 1 Output…………………………………………………………………………………………….……………..12 Table 2 – Equation 2 Output…………………………………………………………………………………………………………….16 Table 3 – Chow Test Output……………………………………….………………………………………..………………………….17 Table 4 – Wald Test Output………………………………………………………………………………….…………….……………17 Table 5 – Equation 3 Output………………………………………………………………………………….…………….…….……20 Table 6 – Equation 4 Output……………………………………………………………………………………….…….….….……..22
  • 6. P a g e | v List of Abbreviations CPI – Consumer Price Index EC – European Commission ECB – European Central Bank EMU – European Monetary Union EU – European Union GDP – Gross Domestic Product IMF – International Monetary Fund OECD – Organisation for Economic Cooperation and Development OLS – Ordinary Least Squares PBC – Political Budget Cycle SGP – Stability and Growth Pact
  • 7. P a g e |1 1. Introduction The term Political Budget Cycle (PBC) refers to cycles in a domestic government’s budget balance that result from changes in fiscal policy induced by the political calendar. Motivated by the incumbents desire for re-election, expansionary fiscal policy may be implemented as an election approached with the aim of maximising a constituent’s welfare with the objective to win re-election. This opportunistic behaviour would lead to variations in budget balances weighted towards election years (Drazen, 2008). This study will research the presence of PBCs within the highly developed and democratic countries of the Eurozone. The presence of PBCs within the Eurozone is ambiguous. Firstly, by adopting the Euro, the member nation relinquishes its ability to control its domestic monetary policy to the European Central Bank (ECB). This leaves fiscal policy as the only remaining instrument available to influence voters’ perceptions before an election. Conversely, Euro members must adhere to the Stability and Growth Pact (SGP) which sets out clear restrictions for the use of fiscal policy, limiting an administration’s ability to manipulate spending for political gains. Due to the discrepancies in conclusions achieved by various studies, the topic of the existence of PBCs, especially within the EU, is a highly debated and a relatively unexplored area of interest. Little research has been conducted on the change in composition of PBCs as a result of the single currency union, with no existing datasets comparing the strength of existences of PBCs before the adoption of the Euro to their strengths today. This research will make a new contribution to existing literature by exploring the evidence of partisan fiscal spending cycles within Euro members. By comparing evidence for the existence of PBCs before and after the adoption of the Euro, this research will also investigate the impact the implementation of the single currency had on the ability of incumbent governments to use fiscal policy for political gain. This comparison allows this research to further make an original contribution by drawing conclusion on the capability of SGP, the ECB and the European Commission (EC) to ensure that fiscal discipline is enforced within the European Monetary Union (EMU).
  • 8. P a g e |2 2. Literature Review This paper begins with a comprehensive review of the current literature on the topic of PBCs. Much of the empirical evidence on the existence of PBCs is contradictory. Hallerberg et. al (2002) study 10 Eastern European accession countries between 1990-1999, focusing on the role of the exchange rates and central bank independence in restricting or encouraging PBCs. Their results find that a country with a flexible exchange rate relies upon monetary expansion whereas countries that fix their exchange rates to the euro are likely to experience fiscal PBC. A drawback of these findings is that over time, the 10 European accession countries will all presumably join the euro zone, fixing the exchange rate and surrendering all monetary policy to the EBC (Hallerberg et. al, 2002). The empirical evidence through the EU is also conflicting; Andrikopoulos et. al (2004) found no evidence of PBCs within 14 EU members through the period 1970 - 1998. However a more recent study by Bruck and Stephan (2006) concluded that fiscal policies have had an expansionary bias in correspondence to political elections. In contrast to the euro zone, there is also literature on the existence of PBCs in the developing world (Schuknecht, 1998; Block, 2002). Schuknecht (1998) studies the fiscal policy instruments used by governments to influence election outcomes in 24 developing countries for the period 1973-1992. The empirical evidence strongly suggests the prevalence of public investment cycles used by governments in order to influence their popularity around elections (Schuknecht, 1998). These findings are consistent with those of Block (2002) in his study of 44 Sub-Saharan African countries between 1980 and 1995; politicians interested primarily in re-election manipulate policy instruments to win support of voters. Furthermore Klomp & De Haan (2013) comprise a total of 70 developed and developing countries in their study, spanning 1970 to 2007. Alike with many of the studies on PBCs, Klomp & De Haan use the budget balance as the fiscal indicator for the dependent variable. They also find that PBCs only occur in the short-run and are conditional on a number of factors such as: the level of development and democracy; government transparency; the country’s political system and its membership of a monetary union. Shi and Svensson (2006) argue that PBCs are ‘significantly larger’ in developing countries and believe depending on whether a country is developing or developed plays a ‘large extent’ in explaining the differences in PBCs. Gonzalez (2002) and Brender and Drazen (2004) studied democracies and their link with PBCs and found democratic countries show stronger evidence for the existence of PBCs. Brender and Drazen (2004) find in their research that PBCs are ‘stronger in weaker democracies’ where governments are less transparent. Schneider’s (2010) study uses data of public spending and deficits in the West German states from 1970 to 2003. While a number of studies find that PBCs only occur if fiscal transparency is low, Schneider argues that this is not the case. Schneider uses the annual growth of the budget deficit as one of his main dependent variables. The results show that
  • 9. P a g e |3 state governments do indeed abstain from raising deficits in the pre-election period as a result of fiscal transparency (Schneider, 2010). Brender and Drazen (2004) argue that voters in new democracies are less aware of political strategies used in order to try and encourage people to vote in favour for that party. This supports Gonzalez’ work (2002) in that most new democracies are linked with developing countries, as ‘established democracies’ (Brender and Drazen, 2004) are usually set up in developed countries as has been found between the years of 1960 and 2001. Shi and Svensson (2002) examine the relationship between electoral dates and the use of fiscal policy by different governments. Using a panel of 91 countries, which include developed and developing nations between the years of 1975 and 1995, the authors collect for a number of economic factors, including fiscal balance and expenditure. From their results they find the existence of PBCs in both developed and developing countries, but the effects of a PBC to be much greater in developing nations compared to developed. The authors also find that the governments tend to increase their spending before an election, whilst revenue decreases. Bender and Drazen (2004) embark upon a similar study, where the authors empirically test for the existence of PBCs in 106 countries, again for both developing and developed. The countries are split into 68 democracies with the remaining non-democracies, with results showing only democracies accountable for the existence of PBCs. With democracies removed from the model, the authors find no evidence for PBCs. The 68 democracies consist of both developed and developing economies, where the strength of the PBC is dependent on the electoral system and government. In contrast to the works of the previous papers, Alt and Lassen (2006) show through their sample of 19 Organisation for Economic Cooperation and Development (OECD) countries that many ‘advanced industrialised economies’ feature PBCs. The author’s state PBCs are most noticeable in politically polarized countries1. As there is no common ground between political ideologies, the government will increase fiscal manipulation in order to win the public vote. Furthermore, Dreher and Vaubel (2001) put the existence of PBCs mainly down to the International Monetary Fund (IMF). They used panel data for 106 countries between 1971 and 1997 and found the facilitation of PBCs were down to lending from the IMF. It appears as long as a country is ‘credit worthy’ (Dreher and Vaubel, 2001) the IMF would contribute to funding. This would create a ‘boom’ during the time of election, therefore influencing voters to favour the incumbent party. The argument highlighted here against the IMF lending is the moral hazard that is created. However, Dreher and Vaubel (2001) do also find ‘strong evidence’ that the contribution of the IMF lays in the ‘more democratic borrowing countries’ therefore agreeing PBCs are most likely found in democratic countries. 1 Politically Polarised Countries – Countries that face extreme opposing views, with little or no common ground.
  • 10. P a g e |4 3. Theoretical Considerations The existence of political cycles was first introduced by Rogoff and Sibert in 1988. They proposed a model that assumed each political party has a competence type that was unknown to the voters. Rational voters would vote for the political party that maximised their expected utility and can only assess a party competence through fiscal outcomes. Therefore an incumbent would attempt to signal their competency by engaging in expansionary fiscal policy before an election. The model is essentially one of adverse selection that emphasised the idea of competence, coupled with asymmetric information. Rogoff (1990) builds upon this theoretical framework to develop a model that focused on the composition of government spending relative to election timings, specifically a more competent policymaker engineers a PBC to shift government outlays to more favourable transfers and visible programs (Rogoff, 1990). However the implications of this model conflicts with empirical evidence, suggesting key variables were missing from the model. Persson and Tabellini (2000) and Shi and Svensson (2006) proposed a new generation of PBC models that stated desire for re-election creates incentives for politicians to appear competent preceding an election. This incentive to manipulate fiscal policy varies from election to election and depends on how sensitive the probability of winning re-election is. Thus it is argued that perceived favourable economic conditions can result in the loss of desire to manipulate a constituent’s welfare through fiscal policy as government popularity is already strong. Therefore when investigating how the timings of elections affects the positions on fiscal policy, it is important to consider which macroeconomic indicators have a significant impact on both the support of the incumbent and the budget balance directly. A governments’ budget balance can be directly impacted by the economic conditions faced under the incumbent’s tenure. The growth in the domestic demand or output, as measured by GDP, will have a substantial effect on the government’s budget balance through the direct influences economic growth has on tax revenues via changes in demand. The rate of growth experienced can also have an impact on the labour market conditions, where changes in employment levels will considerably impact income tax revenues and net social transfer payments. The economic conditions experienced by a constituency throughout the tenure of a government can also have a significant impact on election outcomes. While voters do not evaluate a government’s success exclusively through the performance of macroeconomic indicators, they are arguably the most important and therefore weigh in more heavily relative to other factors (Healy & Lenz, 2014). Lewis-Beck and Stegmaier (2000) evaluate evidence from economic and election research focused on national elections. They conclude much of the variance in government support can be attributed to the economic indicators; unemployment, inflation and GDP growth, which are specifically shown to have the strongest relationship with public approval levels (Lewis-Beck & Stegmaier, 2000). Conversely Evans and Anderson (2006) argue that the voting public’s general understanding of economic performance is weak and therefore economic perception should be treated as
  • 11. P a g e |5 endogenous. By using data on British elections from 1992 – 1997 they find evidence supporting the contention that political partisanship systematically influences economic perception, therefore lagged economic indicators better predict government popularity (Evans & Anderson, 2006). Consequently, perceived utility experienced by residents is based primarily on the lagged performance of the specific variables indicated by Lewis-Beck and Stegmaier (2000). As a result of these findings, the one year lagged performance of growth, inflation and unemployment will be considered in the analysis.
  • 12. P a g e |6 4. Data To empirically test for the existence of PBCs within the Eurozone, annual time series data is collected from 15 major European economies, including the original 12 Eurozone countries over the period 1970 – 2013. These nations were selected as they represent the largest economies within the European borders where the most accurate and reliable data is available. The collected data was then pooled to create a balanced panel dataset, consisting of 43 sample periods and 15 cross-sectional countries to yield a total of 534 panel observations. Data is collected on key economic indicators that correspond with existing literature as mentioned above. The key economic indicators selected are; budget balance as a percentage of GDP, annual GDP growth rate, annual inflation rate as measured by the consumer price index (CPI) and the annual unemployment rate as measured by the claimant count. Data for each of these indicators were collected from reputable secondary sources, with data for the budget balance as a percentage of GDP collected from the Statistical Annex to European Economy, a publication created by the EC. To cover all years under observation, each periodical was used over 2002 – 2013. The budget balance has been computed as a ratio of GDP to negate the effects of different size economies. This is also in alignment with the requirements of the SGP. The annual GDP growth rate in current prices for each country, was collected from the OECD database. Similarly, the annual CPI rate used to calculate inflation was also collected from OECD, with 2010 as the base year. The data for the unemployment rate, which is calculated as percentage of the total labour force was collected from the World Bank website. The 43 year time frame covers on average 12.4 electoral elections per country, with data collected from the World Bank’s Database of Political Institutions, and complemented by the electronic version of the Europa Year book. There is also scope to test for the effectiveness of the governing bodies of the EU to enforce fiscal discipline and prevent volatility in the domestic government’s budget, in line with the criteria of the SGP. This test requires one additional factor; the date in which the Euro was adopted as a domestic currency, if at all in each of the selected countries. This data is collected from the European Commission publication, ‘Timeline: The Evolution of EU Economic Governance in Historical Context’.
  • 13. P a g e |7 5. Budget Balance and Election Timing – A Brief Overview The graphs displayed below show the budget balance of Greece and Finland respectively over the period 1970-2013. As can be seen the minimum points of budget balance tend to coincide with the year in which an election occurs. This leads further into our empirical study whereby the relationship between budget balance and election timings is examined. Figure 1 – Greece Budget Balance Figure 2 – Finland Budget Balance
  • 14. P a g e |8 6. Expected findings Based on the results of previous studies and the theory highlighted above, the expectation is that the value of the coefficient 𝛿1 will be positive and significant. It is expected that the value of 𝛿2 will be positive and significant, and greater than 𝛿1, indicating a peak in the cycle of fiscal spending in the year of an election. The value of 𝛿3 is expected to be significantly less than 𝛿2 and either be low or negative, representing a fall in fiscal spending in the year after an election and signalling the existence of PBCs within the EMU. A further expectation is that the values of the coefficients 𝛿4and 𝛿5 will still be positive and significant, but will be of less magnitude than 𝛿1 + 𝛿2 respectively. With the value of 𝛿6 either small or negative but of less magnitude than 𝛿3. This would signal that the EC and SGP have been effective in limiting fiscal volatility. However numerous studies have found no evidence of PBCs in developed economies so it may be the case that the value of the coefficients 𝛿1, 𝛿2 and 𝛿3 are insignificant, signifying no evidence for the existence of PBCs within the EMU. Figure 3 graphically represents these hypothetical three outcomes. Figure 3 – Graphical representation of expected results
  • 15. P a g e |9 7. Methodology – Empirical Strategy The main hypothesis to be tested by this investigation is whether or not the domestic governments of major European economies adopt an expansionary fiscal policy in the years before an election, signalled by a deterioration of the budget balance, and contractionary fiscal policy in the years after. A further hypothesis to be tested is whether or not the creation of the EMU, specifically the implementation of the Euro, has resulted in an increase in the use of PBCs. In order to evaluate the relationship between the timing of domestic parliamentary elections and the behaviour of fiscal policy, an empirical specification has been developed building on work previously conducted in this field, specifically that of Shi and Svensson (2002 & 2006) and Andrikopoulos et. Al (2004). However as this study is specifically investigating the major European economies it has been possible to increase the timeframe under examination from previous studies and also to include variables that are more specific to highly developed, democratic European economies. In order to test the hypothesises, this study estimates the following version of the empirical model which takes the form: Equation 1 𝐵𝑢𝑑𝑔𝑒𝑡𝑖,𝑡 = ∁0 + 𝛿1 𝑃𝑟𝑒𝑖,𝑡 + 𝛿2 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 + 𝛿3 𝑃𝑜𝑠𝑡𝑖,𝑡 + 𝛽1 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖,𝑡−1 + 𝛽2 𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1 + 𝛽3 𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑖,𝑡−1 + 𝜀𝑖,𝑡 Hypothesis 1 – Evidence of PBCs 𝐻0: 𝛿1 = 0 , 𝛿2 = 0, 𝛿3 = 0 𝐻1: 𝛿1 ≠ 0 , 𝛿2 ≠ 0, 𝛿3 ≠ 0 Where the dependent variable to be tested is Budget i, t,which represents the government’s budget balance as a percentage of GPD in country i and year t, with a budget deficit represented by a negative value and a budget surplus represented by a positive value. The independent variables will be tested to assess the impact that they have on the budget balance. ∁0 represents a constant term and will measure the value of the y-axis intercept when all other independent variable values are 0. Pre i,t is a binary dummy variable that is included to capture the effect of the year prior to a parliamentary election in country i and takes the value of 1 in the year before the election and 0 otherwise. For simplicity, only the year in which an election occurs will be recorded, regardless of which month the election took place in and this process is extended for the year before and the year after an election. Year of i, t is a binary dummy variable that is included to capture the effect of being the year of an election on budget balance in country i and takes a value of 1 in the year of an election and 0 otherwise with again only the year of occurrence being recorded. Post i, t is a dummy variable to capture the effect of the subsequent year of an election in country i, and takes a value of 1 post-election and 0 otherwise. Three control variables have also been included for their theoretical impact to changes in fiscal policy and for their significant impact on a constituent’s utility as mentioned above. Inflation i, t-1 represents the annual inflation rate as measured by CPI in country i and year t.
  • 16. P a g e |10 This variable has been lagged due to voters assessing economic performance retrospectively as mentioned above. Growth i, t - 1 is the annual percentage growth rate of GDP in country i and year t -1. This variable has been lagged as to remove the strong correlation with the independent that arises from budget being a ratio of GDP in year t, however multicollinearity may still be present and will be tested for using the variance inflation factor. Unemployment i, t-1 is the rate of unemployment as a percentage of total labour force in country i and year t. This variable has again been lagged due to the retrospective nature of the formation of economic perceptions. As these three variables are controls, the values of the coefficients 𝛽1, 𝛽2 and 𝛽3 are not of significance in testing the chosen hypothesises. i represents the error term and incorporates all other factors that affect budget deficit that this model excludes. The method used for estimating the values of each coefficient will be Ordinary Least Squares (OLS), with each hypothesis tested at the 1%, 5% and 10% significance levels. This approximation method and significance levels are consistent with previous models developed by Lane (2003), Andrikopoulos et. al (2004), and Brender and Drazen (2004) and deemed appropriate in accordance with the existing literature in this field.
  • 17. P a g e |11 8. Empirical Results 8.1. Evidence of political budget cycles over the sample period The empirical analysis is directed at the detection of electoral cycle regularities as specified by the aforementioned hypotheses. Table 1 reports the results of the OLS estimation based on Equation 1 specified above. As can be seen, PRE has a coefficient of -0.342733, which implies that the budget balance decreases by approximately 0.34% of GDP in the year before an election. YEAR OF has a coefficient of -0.473421 indicating that the budget balance decreases by approximately 0.47% of GDP in the year of an election. POST has a coefficient of -0.414599 suggesting that the budget balance decreases by approximately 0.41% in the year after an election. However as the probabilities of these coefficients exceed the critical value implied by the maximum 10% significance level it is not possible to reject the null hypothesis presented in hypothesis 1. Therefore we must accept the null and assume that there is no evidence of PBCs over the entire sample period. This is consistent with Andrikopoulos et al. (2004) and Efthyvoulou (2011) who failed to provide any evidence of the electoral cycles within the Eurozone. This is also backed by Brender and Drazen (2005) and Shi and Svensson (2002 & 2006) who also could not provide evidence of PBCs in developed economies. The lack of empirical evidence for PBCs within our data set is likely to be due to several contributing factors. One being the majority of governments in the sample pursued fiscal policies with the purpose of reducing relatively large budget deficits, accumulated from the economic downturns of the 1970s oil crisis, the recessions of the 1980s and 1990s and the recent global “credit crunch” in 2008, eroding the impact of election based cycles. A common element of the sampled countries is that there is a high level of transparency amongst constituents which, as highlighted by Brender and Drazen (2005) means that voters are more aware of fiscal manipulation and as a result they should punish the incumbent party for such policies, therefore resulting in fewer PBCs. Shi and Svensson (2006) also find that a higher level of transparency in established democracies which results in fewer PBCs. One reason for the lack of evidence for the existence of PBCs in our sample is likely to be due to the fact that the countries in our model consist of established democracies and in such democracies the incumbent government faces a more observant voter. In response to this the incumbent government seeks to actively avoid creating a PBC, as a high level of transparency can result in constituents punishing government parties rather than rewarding them (Schneider, 2010). Based on the theoretical considerations mentioned above and the results of previous studies such as those by Bruck and Stephan (2006), it is proposed that there is a structural breakpoint in the data, that is considered to occur around 2001, in alignment with the introduction of the Euro. Therefore a Chow Test (See Appendix 1 – Chow Test) was conducted so as to validate the existence of a structural break in the year 2000. As can be seen from Table 2 the computed Chow F-statistic value is statistically significant at the 1%
  • 18. P a g e |12 level of significance and hence it is possible to reject the null hypothesis in favour of the alternate and assume that a break does occur. Table 1 – Equation 1 Output Dependant Variable: Budget Method Panel Ordinary Least Squares Sample Period: 1971 - 2013 Sample Periods included: 43 Cross-Sectional Countries included: 15 Total panel (unbalanced) observations 534 Variable Coefficient Standard Error Probability C -0.129109 0.459035 0.7786 PRE -0.342733 0.365878 0.3493 YEAR OF -0.473421 0.377530 0.2104 POST -0.414599 0.361060 0.2514 INFLATION (-1) -0.270085 0.031372 0.0000*** GROWTH (-1) 0.589497 0.056207 0.0000*** UNEMPLOYMENT (-1) -0.349782 0.035777 0.0000*** R-Squared Statistic 0.390195 Adjusted R-Squared Statistic 0.383252 Significance Levels: * 10% level of significance ** 5% level of significance *** 1% level of significance
  • 19. P a g e |13 8.2. Evidence of political budget cycles in the Eurozone Following the discovery of a structural break in the data using our Chow Test it is anticipated that, in line with aforementioned theory, the introduction of the Euro has had a significant effect on the use of PBCs. Consequently, the model has been developed to incorporate the Euro specific effects on the relationship between budget balances and election timings. The inclusion of a Euro variable yields a model with greater explanatory power in the form of equation 2, shown below. Equation 2 𝐵𝑢𝑑𝑔𝑒𝑡𝑖,𝑡 = ∁0 + 𝛿1 𝑃𝑟𝑒𝑖,𝑡 + 𝛿2 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 + 𝛿3 𝑃𝑜𝑠𝑡𝑖,𝑡 + 𝛿4 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛿5 𝑃𝑟𝑒𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛿6 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛿7 𝑃𝑜𝑠𝑡𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛽1 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖,𝑡−1 + 𝛽2 𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1 + 𝛽3 𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑖,𝑡−1 + 𝜀𝑖,𝑡 Hypothesis 2 – Evidence of PBCs 𝐻0: 𝛿1 = 0 , 𝛿2 = 0, 𝛿3 = 0 𝐻1: 𝛿1 ≠ 0 , 𝛿2 ≠ 0, 𝛿3 ≠ 0 Hypothesis 3 – Evidence of PBCs after the adoption of the Euro 𝐻0: 𝛿5 = 0 , 𝛿6 = 0, 𝛿7 = 0 𝐻1: 𝛿5 ≠ 0 , 𝛿6 ≠ 0, 𝛿7 ≠ 0 To test how the acceptance of the Euro has affected the budget balances of each individual countries it is necessary to observe in which year each country adopted the Euro as its domestic currency. Only the year in which the adoption occurred will be recorded regardless of the month. To approximate the effect of adopting the Euro on budget balance in country i, the binary dummy variable Euro i, t is introduced, which takes the value of 1 in the years where the euro has been adopted and 0 otherwise. The Euro variable will be multiplied with Pre, Year Of and Post independently, allowing 𝛿5, 𝛿6 and 𝛿7 to represent the magnitude of cycle’s in fiscal policy in years before, during and after an election, after the adoption of the Euro. These variables will allow some conclusions to be drawn on the ability of the EC and the ECB to limit fiscal volatility within its member states. It will also be possible to draw some conclusions on the ability of European governing bodies to enforce that the members of the SGP adhere to the strict guidelines imposed for fiscal discipline. Equation 2 is estimated using the OLS method of estimation, with the results presented in Table 3. The coefficient associated with the inclusion of the variable EURO is equal to 1.1863372, carrying the implication that since the adoption of the Euro as a domestic currency, budget balances have increased by approximately 1.19% of total GDP. It can also be seen that the coefficient corresponding with POST*EURO is equal to -0.930014. When this interaction coefficient is combined with the existing effect of POST, the implication is that this variable offers an explanatory magnitude of -1.053983 [-0.123969 + (-0.930014)]. This indicates that an incumbent government inclines towards decreasing their budget balance by approximately 1.05% of total GDP in the year after an election, which could indicate an increase in public spending. However, as the probability value of this coefficient
  • 20. P a g e |14 is insignificant at the 10% level of significance it provides no explanatory power when testing for the hypothesis and therefore can provide no evidence for the existence of PBCs. Conversely, the coefficient of the explanatory variable PRE*EURO is equal to -1.520255, when coupled with the existing effects of the variable PRE, this equates to the explanatory power of being the year before an election given the use of the euro being equal to - 1.381662 [0.148593 + (-1.530255)]. This demonstrates that the incumbent government reduces their budget balance by approximately 1.38% of total GDP, implying the implementation of expansionary fiscal policy in the year before an election. The coefficient of the variable YEAR OF*EURO is equal to -1.339580 which, when combined with the existing effect of YEAR OF, leads to the effect of being in the year of an election, given Euro adoption, equal to -1.375833 [-0.036253 + (-1.339580)]. This represents a reduction in the budget balance by approximately 1.38% of total GDP, which may suggest a significant increase in government expenditure or a reduction in the rate of taxation in the year of an election. Both of the coefficients, PRE*EURO and YEAR OF*EURO provide statistical significance to the model at the 10% level of significance threshold, enabling the rejection of the null hypothesis. This provides statistically significant empirical evidence to support the hypothesis that since the creation and implementation of the euro, European governments have manipulated fiscal policy in an attempt to increase their party’s support preceding an election. This provides sufficient evidence in support for the existence of PBCs within the Eurozone. This is consistent with the research conducted by Lane (2003) and Bruck and Stephan (2006), where the authors found an increase in government expenditure leading up to an election. One contributing factor is that with the adoption of the Euro, the sample European countries have forfeited authority over their own domestic monetary policy to the ECB, as is a prerequisite to membership of the euro. Compensating for this, European governments have moved to manipulate fiscal policy in order to give false perceptions to the voters in the run up to an election, increasing the occurrence of PBCs (Hallerberg, de Souza & Clark, 2002). Another reason for the evidence of PBCs is that the sample countries included are established and civilised democracies, where voters support the use of expansionary fiscal policy and lower tax rates and it tends to be the deciding factor when it comes to the vote, thus government’s incline towards using expansionary fiscal policy one year before an election and in the year of an election (Efthyvoulou, 2012). It is important to note that in order to join the Eurozone and adopt the Euro as the domestic currency, an economy must first adhere to the criteria laid out in the SGP, which provides specific guidelines for fiscal spending and the accumulation of national debt. Therefore to join the single currency and throughout membership of the Euro it should be imperative that fiscal policy is stable and consistent throughout the tenure of each government. However, the results achieved by this investigation provide contradictory evidence. Before the Euro was introduced there is no indication to suggest the existence or the use of PBCs, but following the introduction of the Euro, evidence from this research provides statistically significant empirical confirmation for the implementation of fiscal manipulation during election periods, providing evidence for the existence of PBCs in the EU. These findings
  • 21. P a g e |15 prove that the SGP has been unsuccessful in its task to control fiscal spending and limit the levels of debt one nation can accumulate over time. Instead there is evidence to suggest that since the European governments have been restricted by the motions of the SGP, they have acted more uncontrollably in terms of maintaining sustainable levels of debt and fiscal spending. This could be the consequence of the SGP not being implemented or policed rigorously enough and countries not being punished severely when they do not meet or regularly break the set requirements of the SGP. To test the validity of the augmented model, the Durbin-Watson test is applied to detect for the presence of autocorrelation within the residuals. As presented in table 4 the Durbin- Watson statistic is equal to 0.561540. This indicates a high level of positive autocorrelation is present in the model and would normally cause issue for alarm. However this autocorrelation is expected due to the current year’s budget balance being highly dependent on its previous years in an autoregressive process and is likely to be present only through the control variables. Therefore this autocorrelation has little impact on the conclusions drawn from the variables under investigation and the hypothesis tests are deemed acceptable. The Wald test is also applied to test for the true value of the parameter based on the sample estimate, with the difference between the two assumed to conform to normal distribution. As shown in table 4 the test of the joint multiple hypothesis that the coefficients under examination are equal to 0. Based on both the F-Stat and the Chi-Squared it is not possible to reject the null hypothesis as the probability exceeds the critical value. However through the robustness checks that proceed, the validity of the model is tested further, limiting the conclusions drawn from the Wald test. The R-Squared value of 0.395929 indicates that the model only explains approximately 40% of the data, and when coupled with the Durbin-Watson test, it becomes necessary to conduct two robustness checks which are implemented to cater for some of the remaining factors.
  • 22. P a g e |16 Table 2 – Equation 2 Output Dependant Variable: Budget Method Panel Ordinary Least Squares Sample Period: 1971 - 2013 Sample Periods included: 43 Cross-Sectional Countries included: 15 Total panel (unbalanced) observations 534 Variable Coefficient Standard Error Probability C -0.532138 0.516620 0.3035 PRE 0.148593 0.442304 0.7370 YEAR OF -0.036253 0.460174 0.9372 POST -0.123969 0.437023 0.7768 EURO 1.1863372 0.593851 0.0463** PRE*EURO -1.530255 0.784016 0.0515* YEAR OF*EURO -1.339580 0.801665 0.0953* POST*EURO -0.930014 0.773870 0.2300 INFLATION (-1) -0.265899 0.033970 0.0000*** GROWTH (-1) 0.593704 0.056865 0.0000*** UNEMPLOYMENT (-1) -0.352533 0.035795 0.0000*** R-Squared Statistic 0.395929 - - Adjusted R-Squared Statistic 0.384379 - - Durbin-Watson Statistic 0.561540 - - Significance Levels: * 10% level of significance ** 5% level of significance *** 1% level of significance
  • 23. P a g e |17 Table 3 – Chow Test Output Table 4 – Wald Test Output Wald Test Test Equation: Equation Test Statistic Value df Prob. F-Statistic 1.24118 (4, 523) 0.2924 Chi-Squared Statistic 4.964718 4 0.2909 Null Hypothesis: C(5)=0, C(6)=0, C(7)=0, C(8)=0 Null Hypothesis Summary: Normalised Restriction(=0) Value Std. Err C(5) 1.186372 0.593851 C(6) -1.53026 0.784016 C(7) -1.33958 0.801665 C(8) -0.93001 0.77387 Restrictions are linear in coefficients Chow Test - A Test For Structural Break * ** *** Significance Levels: 10% level of significance 5% level of significance 1% level of significance Equation Sample: 1970-2013 Alternate Hypothesis: Breakpoint at 2000 Chow F-Statistic Cumulative prob. F(11,512): P(F < 0.268883703) 0.268883703 0.009*** Null Hypothesis: No breaks at specified breakpoints Chow Break Point Test: Break at 2000
  • 24. P a g e |18 8.3. Robustness Checks The results presented above provide clear, statistically significant, empirical evidence for the existence of PBCs within the European nations since the implementation of the single common currency. However it could be argued that there are other important factors that will significantly affect changes in budget balance that are not considered in the model specification. Therefore to test the robustness of the results, two other model specifications are presented below and will add reliability to results achieved. Firstly it is important to note that fiscal spending on commitments made by a government do not necessarily occur in the year in which the commitment is made. Spending can be spread over a number of years, particularly in the instant of a dramatic policy change, and the true change in spending patterns can be hidden over a number of years. A study by Milesi-Ferretti (2004) concluded that fiscal rules and regulations encourage the use of “Creative Accounting” within governments to hide the true spending figures of fiscal policy in order to achieve the fiscal criteria. Additionally, the end of the working tax year varies depending on the country which can also distort fiscal spending. Local and national governments will seek to exhaust their budgets at the end of the tax year and public and corporate tax revenue will also increase in this period. Accordingly when performing the first robustness check, it is considered that particularly around the time of an election, some fiscal spending in the current year may be related to commitments made in previous years or some spending hidden through creative accounting. Therefore the one year lagged budget variable is added to the model to capture the effect that the previous year’s budget has on the current year. For the second robustness check, it is considered that the most notable factor not incorporated in the model is time, particularly as the timeframe examined will include the 2008 economic crisis. This crisis will have a significant effect on budget deficits, not only through falling tax revenues but also through the impact of austerity measures imposed by the ECB. Conversely, uncertainty with the Euro has led to falling popularity of governments and a greater need for opportunistic expansionary policy. Very little research has been conducted in this field that examines a timeframe to include this economic downturn, further emphasising the need to take this crisis into consideration. The inclusion of a crisis dummy variable, Crisis i, which takes a value of 1 in the years after the crisis and 0 otherwise, will allow the impact of this crisis to be estimated and taken into consideration. The models now take the form: Equation 3 𝐵𝑢𝑑𝑔𝑒𝑡𝑖,𝑡 = ∁0 + 𝛼1 𝑃𝑟𝑒𝑖,𝑡 + 𝛼2 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 + 𝛼3 𝑃𝑜𝑠𝑡𝑖,𝑡 + 𝛼4 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛼5 𝑃𝑟𝑒𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛼6 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝛼7 𝑃𝑜𝑠𝑡𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝜃1 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖,𝑡−1 + 𝜃2 𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1 + 𝜃3 𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑖,𝑡−1 + 𝜃4 𝐵𝑢𝑑𝑔𝑒𝑡𝑖, 𝑡−1 + 𝜖𝑖,𝑡
  • 25. P a g e |19 Equation 4 𝐵𝑢𝑑𝑔𝑒𝑡𝑖,𝑡 = ∁0 + 𝐷1 𝑃𝑟𝑒𝑖,𝑡 + 𝐷2 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 + 𝐷3 𝑃𝑜𝑠𝑡𝑖,𝑡 + 𝐷4 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐷5 𝑃𝑟𝑒𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐷6 𝑌𝑒𝑎𝑟 𝑜𝑓𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐷7 𝑃𝑜𝑠𝑡𝑖,𝑡 ∗ 𝐸𝑢𝑟𝑜𝑖,𝑡 + 𝐵1 𝐼𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛𝑖,𝑡−1 + 𝐵2 𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1 + 𝐵3 𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑖,𝑡−1 + 𝐵4 𝐶𝑟𝑖𝑠𝑖𝑠𝑖,𝑡 + 𝜖𝑖,𝑡 Hypothesis 4 – Evidence of PBCs after the adoption of the Euro with budget lagged 𝐻0: 𝛼5 = 0 , 𝛼6 = 0, 𝛼7 = 0 𝐻1: 𝛼5 ≠ 0 , 𝛼6 ≠ 0, 𝛼7 ≠ 0 Hypothesis 5 – Evidence of PBCs after the adoption of the Euro with crisis considered 𝐻0: 𝐷5 = 0 , 𝐷6 = 0, 𝐷7 = 0 𝐻1: 𝐷5 ≠ 0 , 𝐷6 ≠ 0, 𝐷7 ≠ 0 Table 5 represents the results of the OLS estimation based on equation 3, with the lagged budget variable included. This has the effect of removing the element of the impact of last year’s budget position on this year’s. As can be seen, although the coefficients of YEAR OF*EURO and POST*EURO demonstrate the budget patterns of a PBC, the probability of achieving these coefficients exceeds the maximum 10% significance levels and therefore does not provide any evidence supporting PBCs. Interestingly the coefficient of PRE*EURO is significant at the 10% level. The magnitude of this coefficient demonstrates that in a country that adopted the Euro, the budget balance is 0.83% of GDP lower in the year before an election than it otherwise would be. This is particularly important as the effects of the previous balances have been removed and therefore this provides further evidence that this dramatic increase in spending is a direct result of being the year before an election occurs. This large fall in budget balance is likely to be due to the incumbent government altering policy objectives and increasing public spending in order to maximise voter’s utility in the build up to an election to appear more favourable. This would also explain why YEAR OF is insignificant as the increase in spending that results directly from an election has already occurred in the year before. At the 10% significance level it is possible to reject the null and assume budget balance is directly related to elections. This provides further evidence in support of the original findings. This model is again tested for autocorrelation using the Durbin-Watson statistic which is presented in table 5. As can been seen the Durbin-Watson statistic is now computed to be 1.833110, this provides sufficient statistical evidence that there is no autocorrelation present in the error terms. This confirms that the autocorrelation present in equation 2 was directly related to the fact that the current year’s budget balance is highly dependent on its previous years, most likely in an autoregressive process. As can also be seen in table 5, the R-squared and adjusted R-squared value is now equal to 0.769391 and 0.764531 respectively. This indicates that approximately 77% of the data is now explained by the model. This is a dramatic improvement to equation 2 and further
  • 26. P a g e |20 reinforces both the importance of including the lagged budget variable and the fact the current year’s budget directly depends on its previous values. The strong performance in both of these test statistics combined with achieving a statistically significant result that leads to a rejecting of the null further supports the results achieved in equation 2. Table 5 – Equation 3 Output Dependant Variable: Budget Method: Panel Ordinary Least Squares Sample Period: 1971 - 2013 Sample Periods included: 43 Cross-Sectional Countries included: 15 Total panel (unbalanced) observations 534 Variable Coefficient Standard Error Probability C -0.614044 -0.31952 0.0552* PRE 0.008183 -0.273589 0.9761 YEAR OF -0.273532 0.284715 0.3371 POST 0.097206 0.270387 0.7194 EURO 0.047644 0.369354 0.8974 PRE*EURO -0.829729 0.485479 0.088* YEAR OF*EURO -0.081144 0.497681 0.8705 POST*EURO -0.254877 0.479168 0.5950 INFLATION (-1) -0.09546 0.021812 0.0000*** GROWTH (-1) 0.222861 0.03741 0.0000*** UNEMPLOYMENT (-1) -0.003069 0.02519 0.9031 BUDGET (-1) 0.769923 0.026481 0.0000*** R-Squared Statistic 0.769391 - - Adjusted R-Squared Statistic 0.764531 - - Durbin – Watson Statistic 1.833110 - - Significance Levels: * 10% level of significance ** 5% level of significance *** 1% level of significance
  • 27. P a g e |21 Table 6 represents the results of the OLS estimation based on equation 3, which incorporates the effects of the 2008 financial crisis. Further to the actual statistical results it can be shown with the PRE*EURO coefficient of -1.454603 the indication is that in the year preceding an election the incumbent party increases budget deficit by 1.45% of GDP, this is compounded by the fact that YEAR OF*EURO and POST*EURO have coefficients of - 1.202161 and -0.827712 respectively, the indication being that during the year of an election budget deficit increases by 1.20% and also increases by 0.8% in the year following an election. However the probability associated with achieving the coefficients of YEAR OF*EURO and POST*EURO exceed the critical value implied by the maximum 10% significance level and therefore can provide no support in testing for the existence of PBCs. Most tellingly the PRE*EURO probability value is significant at the same level. This leads to the rejection of hypothesis 4 and gives credibility to the notion of a PBC, in the form that there is an increase in fiscal spending in the year preceding an election supposedly in the attempt of manipulating the voter. This adds support to the argument and only leads to further confirm the results achieved, despite excluding the effects of the 2008 financial crisis. After completing both robustness checks, it is still possible to reject the null hypothesis at a 10% significance level and sufficient evidence has been provided in support of the existence of PBSs within Eurozone countries.
  • 28. P a g e |22 Table 6 – Equation 4 Output Dependant Variable: Budget Method Panel Ordinary Least Squares Sample Period: 1971 - 2013 Sample Periods included: 43 Cross-Sectional Countries included: 15 Total panel (unbalanced) observations 534 Variable Coefficient Standard Error Probability C -0.112379 0.529864 0.8321 PRE 0.066749 0.43947 0.8793 YEAR OF -0.150199 0.457873 0.7430 POST -0.182479 0.433851 0.6742 EURO 1.502892 0.597714 0.0122** PRE*EURO -1.454603 0.777975 0.0621* YEAR OF*EURO -1.202161 0.796325 0.1317 POST*EURO -0.827712 0.768325 0.2818 INFLATION (-1) -0.281444 0.034061 0.0000*** GROWTH (-1) 0.519928 0.061184 0.0000*** UNEMPLOYMENT (-1) -0.352079 0.035502 0.0000*** CRISIS -1.439605 0.462851 0.0020*** R-Squared Statistic 0.406920 Adjusted R-Squared Statistic 0.394422 Significance Levels: * 10% level of significance ** 5% level of significance *** 1% level of significance
  • 29. P a g e |23 9. Conclusions This paper makes a significant and innovative contribution to the existing literature on PBCs. The nature of this contribution is manifested in the fact that little existing research has attempted to identify the existence of PBCs in Europe, specifically since the introduction of the Euro and accompanying regulations such as the SGP. Initially; we use data from 15 European countries and find that prior to the introduction of the Euro there was no significant evidence that incumbent governments manipulate fiscal policy to increase re- election opportunities. We then implement a Euro specific variable with the suggestion being that due to contributing factors such as the introduction of a single currency and relinquished monetary policy controls, the effect has been the creation of more significant PBCs. This is sustained by our empirical results, which indicate that PBCs in the Eurozone have been more prominent despite attempts of the SGP to restrain fiscal manipulation and create more legitimate voting patterns in established democracies. We find that the creation of the EMU and the implementation of a single currency has caused significant ramifications for incumbent government parties, either consciously or subconsciously, as it appears from the results that they now tend to create PBCs more often as an apparent consequence of the reduced policy tools at their disposal. This study lays the foundation for further research into the subject area, most notably the effectiveness of the SGP with significant evidence questioning its merits.
  • 30. P a g e |24 10. Bibliography Alt, J. E., & Lassen, D. D. (2006). Transparency, political polarization, and political budget cycles in OECD countries. American Journal of Political Science, 50(3), 530-550. Andrikopoulos, A., Loizides, I., & Prodromidis, K. (2004). Fiscal policy and political business cycles in the EU. European Journal of Political Economy, 20(1), 125-152 Block, S. A. (2002). Political business cycles, democratization, and economic reform: the case of Africa. Journal of Development Economics, 67(1), 205-228. Brender, A., & Drazen, A. (2008). How do budget deficits and economic growth affect re- election prospects? Evidence from a large panel of countries. The American Economic Review, 2203-2220. Brender, A., & Drazen, A. (2005). Political budget cycles in new versus established democracies. Journal of monetary Economics, 52(7), 1271-1295. Brück, T., & Stephan, A. (2006). Do Eurozone countries cheat with their budget deficit forecasts?. Kyklos, 59(1), 3-15. Dreher, A. (2001). Does the World Bank cause moral hazard and political business cycles? Evidence from panel data. Universität Mannheim, mimeo, http://www.axel-dreher.de. Efthyvoulou, G. (2011). Political cycles under external economic constraints: Evidence from Cyprus. Journal of Economics and Business, 63(6), 638-662. Efthyvoulou, G. (2012). Political budget cycles in the European Union and the impact of political pressures. Public Choice, 153(3-4), 295-327 European Central Bank (2002), Convergence Report 2002, European Central Bank. ISBN: 92- 9181-282-X European Commission (1998). Council Decision of 3 May 1998 in accordance with Article 109j(4) of the Treaty. Official Journal of the European Union, 139(30). European Commission (2015). Stability and Growth Pact. Retrieved from: http://ec.europa.eu/economy_finance/economic_governance/sgp/index_en.htm Eurostat (2015). General Government Deficit (-) and Surplus (+) – Annual Data. Retrieved from: http://ec.europa.eu/eurostat/tgm/table.do?tab=table&plugin=1&language=en&pcode=teina200 Evans, G., & Andersen, R. (2006). The political conditioning of economic perceptions. Journal of Politics, 68(1), 194-207. Gonzalez, M. D. L. A. (2002). Do changes in democracy affect the political budget cycle? Evidence from Mexico. Review of Development Economics, 6(2), 204-224.
  • 31. P a g e |25 Hallerberg, M., de Souza, L. V., & Clark, W. R. (2002). Political business cycles in EU accession countries. European Union Politics, 3(2), 231-250. Healy, A., & Lenz, G. S. (2014). Substituting the End for the Whole: Why Voters Respond Primarily to the Election‐Year Economy. American Journal of Political Science, 58(1), 31-47. Inter-parliamentary union (2014), Chronicle of Parliamentary Elections, Vol. 48, Inter- Parliamentary Union. ISSN: 1994 – 098X Klomp, J., & De Haan, J. (2013). Do political budget cycles really exist?. Applied Economics, 45(3), 329-341. Lane, P. R. (2003). The cyclical behaviour of fiscal policy: evidence from the OECD. Journal of Public Economics, 87(12), 2661-2675. Lewis-Beck, M. S., & Stegmaier, M. (2000). Economic determinants of electoral outcomes. Annual Review of Political Science, 3(1), 183-219. Milesi-Ferretti, G. M. (2004). Good, bad or ugly? On the effects of fiscal rules with creative accounting. Journal of Public Economics, 88(1), 377-394. OECD (2014), National Accounts of OECD Countries, Vol. 2014/2, OECD Publishing. DOI: 10.1787/na_ma_dt-v2014-2-en OECD Database (2015). Gross-Domestic Product. Retrieved from: http://stats.oecd.org/index.aspx?DatasetCode=SNA_TABLE1 Persson, T., & Tabellini, G. (2000). Political economics: explaining economic policy. Cambridge: MIT Press. Rogoff, K. (1990). Equilibrium political budget cycles. American Economic Review, 80, 21–36. Schneider, C. J. (2010). Fighting with one hand tied behind the back: political budget cycles in the West German states. Public Choice, 142(1-2), 125-150. Schuknecht, L. (1998). A simple trade policy perspective on capital controls (No. ERAD-98- 11). WTO Staff Working Paper. Shi, M., & Svensson, J. (2002). Political budget cycles in developed and developing countries. Institute for International Economic Studies, Stockholm University. Shi, M., & Svensson, J. (2006). Political budget cycles: Do they differ across countries and why? Journal of public economics, 90(8), 1367-1389. The World Bank Database (2015). Unemployment, total (% of total labour force). Retrieved from http://data.worldbank.org/indicator/SL.UEM.TOTL.ZS
  • 32. P a g e |26 11. Appendix 1 – Chow Test Chow Test = ((Sc - (S1+ S2)) / (k))⁄((S1 + S2 ) /(N1+N2- 2k)) Where Sc is the sum of squared residuals from the total data sample, S1 is the sum of squared residuals from 1st group of selected samples, S2 is the sum of squared residuals from 2nd group of selected samples, N1 and N2 are the number of observations in each group respectively and k is the total number of parameters.

Related Documents