The failure of the few and the effect on
the many
The effect of cabinet termination on stock market returns for 15
parliamentary democracies in Western Europe
Name: Junior Ike
Student number: S1978624
Study program: Msc IFM Supervisor: Dr. J. O. Mierau
1
Abstract
Using a time series cross section sample of 15 Western European countries for the period of 1980-2011, this study aims to determine whether cabinet terminations significantly affect stock market returns. I investigate several reasons for cabinet termination and other additional determinants. In particular, I find that only lack of parliamentary support significantly explains returns. Overall, the results suggest that cabinet termination does not significantly explain stock market returns.
2
1. Introduction
On Tuesday, the 9th
of December, Greece saw the most significant stock market crash since November 1987. Nearly every stock lost value, large companies such as Public Power (the biggest utility company in Greece) and the National Bank of Greece lost between a fifth and a quarter of their value. The figure below shows the change in stock price of the Athens exchange around December 9. The sudden shock in the stock market was caused by the news that Athens wishes to elect a new president before the end of 2015. Since Dimas – the preferred presidential candidate of the seated government – is unlikely to be chosen, new elections will have to be held. With the radical left-wing party Syriza heading polls, investors are afraid of a new political crisis and how a government headed by Syriza might affect their future cash flows (Witteman, 2014). We can see that a change in power, even before it has occurred, already has a real effect on stock prices.
Figure 1. Greece stock market reaction around December 9
Source: Index Historic Athens Exchange Group. www.helex.gr
3
Previous research that comes closest this paper is by Arin et al. (2013) and Bernhard and Leblang (2006). Arin et al. use a Bayesian Model Averaging approach to investigate whether cabinet changes affect the volatility and if these generate excess returns using a sample of 17 parliamentary democracies for the period 1945-1995. They find that the effect on excess returns is rather weak, but there is evidence that certain political variables, such as reasons for cabinet termination or types of governments, do affect the volatility of returns. Amongst the 17 democracies used, most are set in Europe. Similarly, Bernhard and Leblang also look at excess returns on a monthly basis, but test the effect of cabinet termination as one of many political variables. Their analysis focuses on 10 parliamentary democracies between 1970 and 2003. They find no significant relationship between cabinet termination and excess returns with a median comparison or a panel data approach. Interestingly, changing the benchmark for economic effects does not change the significance of the results. Looking at both articles, the main in contrast with this paper is that I focus on returns rather than excess returns.
Political stability is an important measure for managers and investors. Julio and Yook (2012) argue that a change of power is relevant for taxation, regulations, monetary and trade policies. Roe and Siegel (2009) argue that political instability hinders the development of financial markets. In addition, having connections to politics is beneficial (Bertrand et al., 2006). However, the question remains how markets are affected by political instability on the longer term. The papers mentioned in the previous paragraph find little evidence for the effect of political uncertainty on stock returns. However, Erb et al. (1996) find a weak relationship between the political risk and future stock returns. On the other hand, managers do take political uncertainty seriously. Even in relatively more stable political systems, Julio and Yook find that political uncertainty reduces firm investments. Managers are also aware of the information of their own company’s fundamentals’ included in the stock price, and base investment decisions on the stock price, the so-called investment-to-price sensitivity (Chen et al., 2007). Durnev (2010) finds election years decrease the level of investment-to-price sensitivity significantly. He argues that stock prices become noisier and include less firm-specific information in election years, giving managers less informative signals. Moreover, this drop in investment-to-price sensitivity is associated with lower post-election performance.
4
variable to account for economic shocks. I find that in general, cabinet termination does not significantly affect returns of the stock market.
Besides the introduction, the paper has six other sections. Section 2 looks at the relevant literature. Section 3 explains what data I use and where it comes from. Section 4 explains the methodology used for the results, reported in section 5. Section 6 discusses the results and how these reject or accept hypotheses. Finally, section 7 concludes the paper.
2. Theory and Literature
Previous studies and theory on the link between government failure and a country’s economic performance are inconclusive at best. Theory and research gives reasons that contradict, and so for there is no clear resolution on the matter. In this section I look at the relevant literature and aim to provide a balanced view of possible determinants of the relationship.
Before examining the theory behind the relationship, it is interesting to look at the theory behind election timing and the implication these might have for the results. There are two approaches in the critical events schools of thoughts, corresponding to the belief that the timing is either endogenous or exogenous. The exogenous view is taken by Browne et al. (1984; 1986). The authors consider critical events threatening cabinet survival. These critical events are exogenous shocks. The Lupia and Strøm (1995) model represents a more endogenous view. According to their model, cabinet dissolution does not only depend on external factors, but also on internal factors. They develop a game-theoretical model that takes the critical event approach by Browne et al. and uses this as a starting for the bargaining process between political actors. Termination depends on the attractiveness of the outside option, the critical event,
compared to keeping staying in the current coalition. Empirical evidence for this is given by
5
2.1 Main literature
At a general level, a relationship between political uncertainty and asset valuation can be expected due to the general uncertainty caused by political uncertainty. A cabinet serves a country by developing policies that ensure that a country economically prospers. It is acknowledged that policies put in place by elected officials do affect macroeconomic conditions (Grier, 2008). If a cabinet fails, a new cabinet has to be elected. This brings forth uncertainty about future policies due to uncertainty about election results. Investors will revise their beliefs about future policies based on the information they have. Uncertainty, in turn, affects asset valuation. Ozoguz (2009) argues that beliefs about future cash flows induce increased price sensitivity to news when uncertainty is greater, creates higher asset price volatility. In line with investors demanding higher returns in uncertain times, Ozoguz finds a negative relationship between uncertainty and asset prices when investors believe that times are bad. As such, political uncertainty causes macroeconomic uncertainty due to uncertainty about future policies, which causes investors to revise the information available to them and in turn affect asset valuation.
A negative relationship could exist if investors perceive uncertainty as negative. If investors feel that they have assets whose values are vulnerable to a change in policies, they will likely remove these assets from their portfolio in favour of assets whose value is less correlated to future changes. One of the main findings that do show the negative relationship between cabinet termination and the economy, besides the article mentioned in the introduction, is the link between currency crashes and cabinet termination (Bernhard and Leblang, 2008). The authors study the possibly endogenous relationship between currency crashes and cabinet termination. They hypothesise that asset market performance affects cabinet termination but cabinet termination also affects how market participants evaluate the government’s commitment to economic policies. They find support for both relationships.
6
that cabinet termination is an endogenous event, a party inside the cabinet that aims to break the coalition and force new elections will try to ensure positive stock market returns first in order to ensure re-election. Thus, a cabinet termination could be associated with at least a run-up period of positive stock returns. Moreover, based on the assumption that investors are rational and recognise such schemes, this could mean that there is no observable effect at all. Rational investors will correct for excess returns when they signal that a cabinet is close to termination, thus the market will move towards a regular equilibrium.
There is documentation of a positive relationship between political uncertainty and asset valuation. However, these results are either found before or after the actual event. Pantzalis et al. (2000) find a positive effect before elections. Vuchelen (2003) and Beaulieu et al. (2006) find a positive relationship afterwards. These results are most likely driven by the resolution of political uncertainty. Resolving uncertainty likely drives investors back to assets they perceived as being at risk due to the uncertainty. Moreover, to my knowledge there is no academic research that tests political uncertainty and asset valuation on a yearly basis.
Theory does not give a clear answer on the question whether a positive or a negative relationship can be expected. Therefore, I aim to test the following hypothesis:
(1) Is there a significant relationship between cabinet failure and stock market returns?
The relationship is likely not as clear cut as the hypothesis above. Additional determinants might reveal the actual relationship. The following sub-section deals with these additional determinants.
2.2 Additional determinants
7
effect of cabinet termination on currency crashes. If we accept reasons other than a general election at face value, we can assume that investors will not see a cabinet termination coming. This leads to the following additional determinants:
(a) A year in which a cabinet change occurs due to general elections, a positive change in the stock market is expected.
(b) A year in which cabinet failure occurs, for a reason other than general elections, a negative change in the stock market is expected.
Arguably multiple terminations are more unexpected than the other reasons variable,
since cabinet termination could be signalled before the actual event itself. This gives investors time to review their options in case the event actually takes place. With multiple in a single year, these signs are likely less visible or the termination is a complete surprise. Consequently, uncertainty will be perceived as negative by risk-averse investors. Using the same assumptions as the previous paragraph, the third additional determinant is:
(c) A year in which cabinet failure occurs, for a reason other than general elections, a negative change in the stock market is expected.
8
uncommon in Belgium). In contrast, Rogoff (1990) arguments that the political business cycle can cause lagged negative effects. If policy makers artificially inflate the economy to increase the voters welfare, short-term opportunistic policies cause negative results in the long run since these are not optimum. Again, contrasting views make it difficult to predict a coefficient sign beforehand. Research does demonstrate the possibility of lead and lags. Using daily data, the following additional determinants are:
(d) In a year where a cabinet change occurs, an effect might be witnessed a year before the actual event
(e) In a year where a cabinet change occurs, an effect might be witnessed a year after the actual event.
9
partisan effect is measurable in European economies. Nevertheless, the following two additional determinants are:
(f) A termination shifting power to left-wing policy makers positively affects stock market returns.
(g) A termination shifting power to right-wing policy makers positively affects stock market returns.
3. Data
This paper studies 180 cases of cabinet failure for 15 countries. The analysis focuses on cabinet failures from 1980 - 2011. Data is collected from two major sources. Data on cabinet failure has been collected the Comparative Political Data Set I (CPDS I), created by Armingeon et al. (2011). This dataset covers the period 1960 – 2011, and includes all cabinet failures in this period on an annual basis for 23 democratic countries. A second major source is DataStream. All data on country indices is obtained from this source. In order to create comparable indices, the return indices are taken from the Morgan Stanley Capital International (MSCI). Although data for every country is not available for the entire sample period, this ensures that the return indices are calculated in the same way.
There are 15 countries in the sample. Selection is done in order to create a homogenous group of countries. The countries included in the sample are Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Norway, Spain, Portugal, Spain, Sweden and the United Kingdom. All countries are Western European, with the most notable exemption being Switzerland. The reason Switzerland is left out is the bi-yearly obligatory legislative elections. Besides Finland and France, both hybrid systems, all countries in the sample have a parliamentary executive legitimacy. Julio and Yook (2012) consider a country parliamentary if the prime minister is the chief of state and head of government. Additionally, a country where a hereditary monarch is the chief of state and a prime minister is the head of government. The list of countries and sample periods is provided in the Appendix.
10
to four years, so general elections are required at least every four years. The second reason is voluntary resignation by the prime minister. A quick overview of the number of cases for each specific reason tested is presented in Table 1 below. The full list is presented in the Appendix. Some reasons are not tested due to a low number of observations, rendering the possibility of results by chance too high. In other cases, a reason is not identified at all. If a reason for termination cannot be identified, and under the assumption that legislative elections are easily identified, it must be different from legislative elections. The cases for which a reason cannot be
identified are captured by other reasons dummy. The main idea of this dummy is that it
includes all cases that are different from legislative elections.
Another possible determinant of cabinet termination on the stock market is partisanship of government. Partisanship is also reported in the CPDS I. The authors calculated the percentage of right-wing, centre and left-wing parties for each cabinet composition. Based on these percentages, they calculate five categories of the dominant ideology. These range from a hegemony of right-wing parties (a cabinet including only right-wing parties and centre parties) to a hegemony of left-wing parties (a cabinet including only left-wing parties and centre parties. In addition to defining partisanship, defining the ideological categories as numbers allows a computation of a government gap. By calculating the difference between the cabinet ideology of the outgoing and incoming party, the change in ideology can be quantified. The authors already provide a dummy for significant changes in ideology, meaning that the change in ideology must be different from zero. Additionally, the gap between ideologies allows the creation of two other variables, if the change was towards right-wing parties or towards left-wing parties.
11
Table 1. Descriptive statistics
Mean Median Std. Dev. Minimum Maximum Observations
Market returns 15.529 16.009 31.939 -70.252 196.782 430 Nasdaq returns 12.667 15.319 17.741 -37.136 38.187 430 Cabinet failure 0.384 0 0.487 0 1 165 Legislative elections 0.188 0 0.391 0 1 81 Other reasons 0.195 0 0.397 0 1 84 Voluntary resignation 0.037 0 0.189 0 1 16 Dissension 0.051 0 0.221 0 1 22 Lack of support 0.035 0 0.184 0 1 15 Multiple failures 0.037 0 0.189 0 1 16 Left-wing failures 0.065 0 0.247 0 1 28 Left-wing elections 0.042 0 0.201 0 1 18
Left-wing other reasons 0.033 0 0.178 0 1 14
Right-wing failures 0.058 0 0.234 0 1 25
Right-wing elections 0.037 0 0.189 0 1 16
13
4. Methodology
In order to analyse whether cabinet failure affects stock market return, the paper employs a very basic analysis. The paper similar to this paper by Akin et al. (2013) utilises Bayesian Model Averaging. The authors have a valid point stating that although the sample is large, the number of cabinet failures is relatively small in comparison. However, by using yearly data and only looking at the cabinet terminations since the 1980s, assuming all cabinets are seated for four years at least 25% of the sample should include a cabinet termination. Thus I do not believe that this is an issue in the case of yearly data. The method used in this paper is based on the method Zussman and Zussman (2006) use to analyse whether Israeli assassinations of terrorist targets have an effect on the Tel Aviv 25 index.
Analysing yearly date for the first model uses has the following specification:
where and and
The model used it estimate yearly data is a fixed effect time series cross section model.
The dependent variable is MARKET, which represents the percentage change in the return
index for country i at year t by deducting its previous value from its current value. The first
independent variable is GOVCHAN, a dummy variable that takes on the value of 1 when there
is a change of cabinets for country i at year t. The second independent variable
14
Since cabinets have multiple reasons for termination, it is interesting to look at the effect of these different types of terminations. As mentioned in the data section, the CDSP I defines seven types of termination. The second model looks at the reasons of termination in order to determine if one particular reason is statistically significant. In particular, it is interesting to study whether reasons other than general elections make a difference. An ending of a four year term should be anticipated so the market should be more concerned with the results of the elections rather than the reason of termination. As noted in the section 2, there is sufficient evidence to assume that legislative terminations are different from premature terminations. Therefore, the models (2), (3), (4) and (5) investigate this relationship. Model (2) looks at the effect if only
legislative elections are taken into account. Model (3) includes the variable other reasons and
looks at every reason for cabinet termination other than legislative elections. This variable also captures any termination for which a reason could not documented by Armingeon et al. (2013). Model (4) takes specific reasons, proposed by Woldendorp et al. (2008), from the database and tests these to see whether one specific reason is significant. Not all reasons are included, lack of parliamentary support, resignation due to health reasons and an intervention by the Head of State are left out since there are too few observations. The final model, model (5), looks at multiple terminations in a single year. Only cases where there are two or three failures in a single year are included.
The variables are similar to the first model. The difference between these models and the general model is the inclusion of interaction dummies to replace the original cabinet termination variable. For example, if the reason for cabinet termination is a legislative election, every cabinet termination in the sample that was terminated for this particular reason gets the value of 1; all other cases get a 0. By using interaction dummies rather than regular dummies, the test does not lose power due to losing a number of observations.
Another interesting issue is timing of the shock. Investors might anticipate a government change and the negative or positive shock might therefore be before the year of the change. Moreover, stock markets take time to process news, so we can also assume that any effect from a cabinet termination might only be visible after the actual event. The third model is exactly the same as the first model, the only two major differences are possible leads or lags and
the data analysed. Leads or lags are done for a single year. The first three variables, all
terminations, legislative elections and other reasons are analysed.
To analyse partisanship, all terminations, legislative elections and other reasons are
15
right and central parties (1) (see data section). As such, a left- to right-wing change is any termination where the difference between the old coalition and the new coalition is negative. For example: the 1983 pre-election cabinet of Austria only had left wing parties (5), and after the elections right-wing parties (4). This equals a change of -1 for ideology. Right- to left-wing ideology calculations are the opposite. Again, I code the dummies as interaction dummies to prevent loss of power.
5. Results
The tables below show the estimation results from the models in the methodology. Table 3 has estimation results for the all models except the lagged effects. The models are estimated with a panel data cross section fixed effects approach. Although the Haussmann test supports the use of random effects (rejecting H0 with p = 0.85), the cross section fixed effects keeps the estimation for countries constant. This also means that the R² reported is the within estimator rather than the between estimator. All standard errors presented in the table are robust standard errors, since estimation results show that they differ from regular errors.
The main empirical question of this paper is whether cabinet terminations significantly affect stock market returns. Column (1) shows that cabinet terminations in general are not significant at all. The p-value obtained from the regression is 0.587, which is not even close to the maximum threshold of 10%. The only thing column (1) shows is that the NASDAQ is a very significant control variable to capture external economic influences.
Moving on to column (2), (3), (4) and (5), we see similar results as column (1). The results show that additional determinant (a) does not help to show significant results in the
regression. Continuing with column (3), the other reasons variable is insignificant. Initially, we
should draw the conclusion that determinant (b), reasons other than legislative elections help to explain the effect, is rejected as well. However, Column (4) shows these reasons as specified in a regression analysis. Although column (3) is insignificant, lack of parliamentary support is weakly significant but with a strong positive effect. While there is no evidence to assume that reasons other than legislative elections significantly explains the relationship between cabinet termination and stock market returns, the support variable does add a side note. Finally,
column (5) also fails to show significant results for the multiple changes dummy variable. The
results underline that multiple changes in a single year are not an additional determinant, rejecting any assumption made by determinant (c).
16
18
Table 4. Lags and leads
(Dependent variable: percentage change in MSCI index for selected countries)
Lag (N=430) Lead (N=429) (1) (2) (3) (4) (5) (6) All terminations -0.719 (2.653) 3.620 (2.291) Elections -0.746 (2.787) 5.856 (2.603) Other reasons -0.349 (2.624) -0.347 (2.779) Percentage change in NASDAQ 1.011* (0.036) 1.011* (0.036) 1.011* (0.036) 1.007* (0.036) 0.999* (0.036) 1.010* (0.035) Constant 3.012 (1.046) 2.879 (0.648) 2.811 (0.645) 1.395 (0.999) 1.781 (0.762) 2.815 (0.638) R² 0.320 0.320 0.319 0.323 0.325 0.320 Number of cases 164 81 83 165 81 84
Notes: *, **, *** represent significance levels of 1%, 5% and 10% of a two sided p-value. All standard errors are robust standard errors and adjusted for 15 country clusters.
6. Discussion
The question remains what drives the results shown in the previous section. How can these results be interpreted using literature mentioned earlier in the paper? This section tries to deal with this issue. Using literature presented earlier, I will try to explain what drives these results and why they are no surprise.
19
measure on a yearly basis. Stock market returns on a daily or monthly basis could be affected, but the relationship is not strong enough to influence yearly returns.
Additional determinants (a) (b) and (c) do not help to uncover a relationship. For determinant (a), taking into account the lagged effects, the coefficient signs support the political business cycle and the anticipation of the stock market. Returns are positive in a termination year, and negative a year before. However, results are insignificant so it is impossible to draw a
definite conclusion based on the current results. This is partly underlined by the multiple
changes variable. If a strong reaction can be expected, it should be from sudden shocks. Since this variable is insignificant as well, the effect is likely to be weak on a yearly basis. As for other reasons, only the lack of parliamentary support variable is significant. In part, this makes additional determinant (b) true. The positive sign is the opposite of expectations earlier in this paper. A positive effect can come from investors trust in the current policies. If future policies negatively affect investors (Ozoguz, 2009), current policies certainly will as well. Withdrawing support or a successful vote of no confidence means that politicians themselves perceive current or future policies as harmful, so harmful that they are willing to call early elections. Another plausible explanation comes from the political business cycle (Lewis and Black, 1988). Investors consider their financial position when voting. As such, the party opting for premature elections benefits when the current economic conditions are good since investors will likely link these conditions to their efforts. Unfortunately, the CPDS I does not offer information on the conditions behind a withdrawal of support, so there is currently no data available to test this assumption.
Looking at the leads and lags, additional determinants (d) and (e) prove to be insignificant. The result of the lagged effect does not show support for Vuchelen (2003), although the coefficient is the correct sign for except for early elections. The leading effect opposes the findings of Pantzalis et al. (2000) that the pre-termination period sees positive changes. A likely explanation is that again, the effect is not strong enough to be measured on a yearly basis.
20
on a consensus, so a change in ideology does not necessarily mean that the new coalition will change policies.
Overall, I do not find support for the hypotheses or the additional discussed in Section 2. Only a single variable is significant under the highest threshold. The lack of strength of the effect could be an explanation. The effect might not be measurable on a yearly basis. However, a yearly analysis for investments does show results (Julio and Yook, 2012; Durnev, 2010). Another explanation is the state of the economies. All countries in the sample are open economies, some of them relatively small as well. It is possible that these economies are so correlated that specific country issues hardly affect these open economies. This explanation seems more plausible since the correlation matrix (see table 2) shows that the correlation between European market returns and the NASDAQ returns is over 0.5. The general conclusion is that cabinet termination for a country does not cause changes in yearly stock market returns.
Although this paper offers no significant results, there are still managerial implications. Durnev (2010) reports that investment-to-price sensitivity decreases in cabinet termination years as opposed to non-termination years, and that this decrease affects company profitability. On a yearly basis, managers can question whether this behaviour is most beneficial for the company. Stock returns seem unaffected by cabinet changes, and should therefore not be as noisy as Durnev suggests. Cabinet changes should not be ignored entirely; there are enough reasons why managers should be concerned with cabinet changes (Julio and Yook, 2012). However, concerns for stock returns on a yearly basis should – based on the results – not be one of them. Investors can derive implications as well, although these are mentioned with precaution. We must assume that investors do not correct for cabinet terminations, isolating any effect that could be present. Under this assumption, investors do not need to correct these indices as these are not affected by cabinet termination.
7. Conclusion
This paper examines the relationship between cabinet termination and stock market returns. This relationship hardly received any attention, most of the research focussing on terminations in general and excess returns. In contrast, I focus on general returns for 15 Western European countries and look at qualitative characteristics of cabinet terminations.
21
effect is too weak to measure on a yearly basis. Another explanation, following Vuchelen (2003), is that parliamentary democracies are less likely to change policies since these are the result of a consensus between different parties.
General implications for managers are in the area of risk management and investment decisions. Managers are less likely to take stock prices of their company into account during termination years, which seems a non-optimum strategy. Stock returns are not affected by cabinet termination, so they might contain less noise than expected.
22
Appendix
Table 5. Countries, sample periods and means
Country Period Mean Mean termination Years Mean non-termination years
Austria 1980 - 2011 12.375 25.012 12.375 Belgium 1980 - 2011 15.287 8.426 17.116 Denmark 1980 - 2011 18.129 26.126 18.129 Finland 1989 - 2011 20.390 22.974 20.390 France 1980 - 2011 14.935 11.277 18.824 Germany 1980 - 2011 12.656 7.225 12.976 Greece 2003 - 2011 4.452 -9.951 4.452 Ireland 1989 - 2011 7.945 23.006 7.945 Italy 1980 - 2011 16.709 24.840 15.557 Netherlands 1980 - 2011 15.092 17.495 15.589 Norway 1980 - 2011 20.334 20.163 21.459 Portugal 1989 - 2011 9.017 -6.303 9.017 Spain 1980 - 2011 18.340 17.123 18.962 Sweden 1980 - 2011 22.973 22.968 23.084 UK 1980 - 2011 13.740 19.364 13.740
Table 6. Reasons for government termination
Reason Explanation Instances
Elections General elections due to ending of the
governing term. 81
Voluntary resignation Prime minister resigns voluntarily. 16
Health reasons Prime minister resigns due to health
reasons. 1
Dissension A coalition breaks up without external
pressure. 22
Lack of support A party withdraws from the coalition or a
successful vote of no confidence occurred. 6
Intervention Head of state intervenes, breaking up the
coalition. 2
Broadening A termination of the government to allow
23
Bibliography
Alesina, A., 1987. Macroeconomic policy in a two-party system as a repeated game. The Quarterly Journal of Economics 102, 651-678.
Arin K., Molchanov A., Reich O., August 2013. Politics, stock markets, and model uncertainty. Empirical Economics 45, 23-38.
Arnold, I. J., Vrugt, E. B., 2010. Treasury bond volatility and uncertainty about monetary policy. Financial Review 45, 707-728.
Armingeon, K., Knöpfel, L., Weisstanner, D., and Engler, S., 2013. Comparative Political Data Set I 1960-2011. Bern: Institute of Political Science, Berne.
Beaulieu, M. C., Cosset, J. C., & Essaddam, N., 2006. Political uncertainty and stock market returns: evidence from the 1995 Quebec referendum. Canadian Journal of Economics 39, 621-642.
Berlemann, M., Markwardt, G., 2006. Variable rational partisan cycles and electoral uncertainty. European Journal of Political Economy 22, 874-886.
Bernhard, W., & Leblang, D., 2006. When Markets Party: Stocks, Bonds and Cabinet Formation. In: Bernhard, W., & Leblang, D., 2006. Domestic Processes and Financial Markets: Pricing Politics. Cambridge University Press, New York, pp. 49-85.
Bernhard, W., Leblang, D., 2008. Cabinet collapses and currency crashes. Political Research Quarterly 61, 517-531.
Bertrand, M., Kramarz, F., Schoar, A., Thesmar, D., 2006. Politicians, firms and the political business cycle: evidence from France. Unpublished working paper. University of Chicago, Chicago.
Browne, E. C., Fendreis, J. P., Gleiber D. W., 1984. An 'events' approach to the problem of cabinet stability. Comparative Political Studies 17, 167-97.
Browne, E. C., Fendreis, J. P., Gleiber D. W., 1986. The process of cabinet dissolution: An exponential model of duration and stability in Western democracies. American Journal of Political Science 30, 628-50.
Chen, Q., Goldstein, I., Jiang, W., 2007. Price informativeness and investment sensitivity to stock price. Review of Financial Studies 20, 619-650.
Döpke, J., Pierdzioch, C. 2006. Politics and the stock market: Evidence from Germany. European Journal of Political Economy 22, 925-943.
Durnev, A., December 2010. The real effects of political uncertainty: Elections and investment sensitivity to stock prices. In Paris December 2010 Finance Meeting EUROFIDAI-AFFI. Erb, C. B., Harvery, C. R., Viskanta T. E., 1996. Political risk, economic risk, and financial risk, Financial Analysts Journal, 29–46.
24
Grier, K., 2008. US presidential elections and real GDP growth, 1961–2004. Public Choice 135, 337-352.
Julio, B., & Yook, Y., 2012. Political uncertainty and corporate investment cycles. The Journal of Finance 67, 45-83.
Lewis-Beck, M., 1988. Economics and Elections. University of Michigan Press, Ann Arbor, Michigan.
Lupia, A., Strøm, K., 1995. Coalition termination and the strategic timing of parliamentary elections. American Political Science Review 89, 648-665.
Miller, T. W., Sabbarese, D. M., 2012. Stock market’s anticipation of information and reaction to the release of information. Intertnational Research Journal of Applied Finance 10, 1384-1559.
Ozoguz, A., 2009. Good times or bad times? Investors' uncertainty and stock returns. Review of Financial Studies 22, 4377-4422.
Pantzalis, C., Stangeland, D., Turtle, H., 2000. Political elections and the resolution of uncertainty: the internationalevidence. Journal of Banking and Finance 24, 1575–1604.
Pástor, Ľ., Veronesi, P., 2013. Political uncertainty and risk premia. Journal of Financial
Economics 110, 520-545.
Roe, M., Siegel J., 2009. Finance and politics: A review essay based on Kenneth Dam’s analysis of legal traditions in The Law-Growth Nexus. Journal of Economic Literature 47, 781-800.
Rogoff, K., 1990. Equilibrium political budget cycles. American Economic Review 80, 21-36. Sagrera, P. R., 2010. To anticipate or not to anticipate? A comparative analysis of opportunistic early elections and in-cumbents economic performance. Documentos de Trabajo FUNCAS 537, 1-60
Santa‐Clara, P., Valkanov, R. 2003. The presidential puzzle: Political cycles and the stock market. The Journal of Finance 58, 1841-1872.
Stevens, D., 2007. Mobilization, demobilization and the economy in American elections. British Journal of Political Science 37, 165-186.
Veronesi, P., 1999. Stock market overreaction to bad news in good Times: A rational expectations equilibrium model. Review of Financial Studies 12, 975–1007.
Vuchelen, J., 2003. Electoral systems and the effects of political events on the stock market: the Belgian case. Econ Politics 15, 85–102.
Warwick, P. V., 1994. Government Survival in Parliamentary Democracies. Cambridge University Press, Cambridge.
Witteman, J., 9 December 2014. Ongekende beurskrach Griekenland. De Volkskrant, retrieved from http://www.volkskrant.nl
25
