The Effects of Austerity, Inflation, and Growth Shocks on Greek Public Debt Papapanagiotou Maria
Student ID 10841385
papapanagiotou.maria1@gmail.com Supervisor: Mavromatis Konstantinos
15-07-2017 University of Amsterdam
Statement of Originality
This document is written by Student Maria Papapanagiotou who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This thesis examines how Greek public debt is affected by shocks of inflation, austerity, and GDP growth. I use a variation of VAR model with debt feedback to compute the coefficients associated with the endogenous variables, which is followed by impulse-response analysis. Main results show that after an austerity shock, the debt ratio first declines and then returns close to its initial path and stabilizes slightly lower than the pre-shock path. Despite the weak economy, during the majority of the period studied, an austerity shock seems not to be self-defeating when it is applied to the Greek economy. An inflation shock only slightly decreases the debt ratio before it returns to its pre-shock path. A positive GDP growth shock reduces the debt ratio before it slowly returns to the pre-shock value after eight quarters.
Contents
1. Introduction ... 5
2. Literature Review ... 7
3. Evolution of Greek Public Debt ... 11
Figure 1: Greece Government Debt-to-GDP ... 14
Figure 2: 1st and 2nd bailout programs money allocation ... 14
4. Empirical Model and Data ... 15
4.1 Empirical Model ... 15
4.2 Data and Descriptive statistics ... 17
Table 1: Descriptive Statistics ... 19
Table 2: Dickey-Fuller Test for Unit Root ... 19
Table 3: Lag Selection ... 20
Table 4: Stability Test ... 20
5. Analysis ... 21
5.1 Results from the VAR ... 21
Table 5: Main Results... 27
5.2 Debt Impulse Responses ... 28
Figure 3: Impulse-Response Functions of debt ratio ... 30
6. Conclusion ... 31
References ... 32
1. Introduction
After the global financial crisis rose in 2008, as a result of the 2007 US subprime mortgage loan crisis, many advanced economies have faced the deterioration of their public finances (Kouretas & Vlamis, 2010). According to the European Commission report of 2016, the public debt ratio for the EU-28 increased from 57.5 percent of GDP before the crisis to 85 percent of GDP in 2015. Regarding public debt, Greece is still experiencing one of the harshest and longest crisis after Second World War II, which led the debt-to-GDP ratio up to 181.5 percent of GDP in the second quarter of 2014. Besides, the country is suffering from the loss of the international competitiveness, the exclusion of the markets, and the increase of the unemployment (Nikiforos, Papadimitriou and Zezza, 2016).
In 2010, Greece asked financial assistance from the international institutions, which came by Troika (European Commission, European Central Bank, & International Monetary Fund). The main objectives of the bailout packages were to restore confidence, competitiveness and credibility for private investors, access to the
markets, and debt stabilization (European Commission, 2010). However, evidence has shown that the austerity measures that put in place, involving budget cuts, tax
increases, and privatizations, have resulted in a deeper recession (Truth Committee, 2015).
Greek public debt for 2016, after 6 years of austerity measures, was 179 percent of GDP, 32.8 percent more than before the bailout packages (Hellenic Statistical
Authority, 2017). Today the country has not yet returned to the international markets. Greece is still under austerity measures, going through the third stability support program. In the 7th of July 2017 and after the second positive review of the progress of the program implementation, Greece unlocked the third tranche of $8.5 billion of
the third bailout program which is planned to disturbed in two sub tranches (European Stability Mechanism, 2017).
As the austerity measures in Greece have only managed to cause an overall surplus of 0.7 percent in 2016 (Hellenic Statistical Authority, 2017), a crucial question about which are the best elements to reduce the debt ratio has arisen. The aim of this paper is to contribute to the existing literature by providing an analysis of the effects that austerity, inflation, and growth shocks have in Greek public debt. For that purpose, I will apply an empirical framework close to one that implemented by Cherif and Hasanov (2012) to the Greek reality. In order to analyze the US debt dynamics, they used a Vector Auto Regressive (VAR) framework which included the debt-to-GDP ratio as an exogenous variable and the primary deficit-to-GDP ratio, the real GDP growth rate, the inflation, and the nominal average interest rate as endogenous variables. These macro aggregates are included in the debt equation, which keeps track of the debt dynamics. Since my data cover the period from 1985 to 2015, I add to this model a dummy variable to indicate the period before and after the adoption of the euro in 2001, given that the adoption of the euro might have an important impact on the macro variables of the model. After the estimation of the VAR model, I perform impulse-response functions (from now on IRFs) for all the endogenous variables and the debt-to-GDP ratio.
At this point, I would like to highlight the importance of the debt feedback in
computing the impulse responses as it was emphasized by Favero and Giavazzi (2007, 2009). If we omit the level of the debt ratio then the responses of the endogenous variables will be included in the error terms. As a result, the coefficients that will be used to estimate the impulse responses will be biased (Favero & Giavazzi, 2007).
Finding the right measures to combat high debt and recessionary periods is crucial to advance in a complete understanding of the dynamics of the economic system and in that case, Greece stands a good opportunity to go forward in that direction.
The thesis is structured as follows. Section II discusses the existing literature on debt reduction and debt sustainability analysis. Section III presents the evolution of Greek public debt. Section IV contains the empirical methodology, the data, and the
descriptive statistics. Section V presents the results of the analysis. Finally, section VI concludes.
2. Literature Review
Many studies use VAR models to examine the effects of shocks on macroeconomic variables or for forecasting. The VAR methodology was first introduced by Sims with his seminal work Macroeconomics and Reality (1980). However, the importance of including the debt feedback while using the VAR methodology is highly
emphasized by Favero and Giavazzi (2007, 2009). According to the evidence, most of the studies that do not keep track of the debt dynamics result in a volatile debt path. There are a few papers that examined the measures that one country can take in order to reduce the public debt-to-GDP ratio. Hall and Sargent (2010) and Apergis and Cooray (2013) used U.S. data and Greek data, respectively, to calculate which
elements have the largest impact on debt ratio. The former research concluded that the enormous reduction of the U.S. debt ratio that happened between 1946 and 1974 occurred, up to 40 percent, due to growth in GDP. Also, the contribution of primary surplus accounts up to 40 percent, and only 20 percent of the reduction was because of inflation. In the same line, the latter research about Greece suggested that the best way to take the debt ratio down is an aggressive growth policy. Again, tax increases
and expenditure cuts or higher inflation can be used to decrease the debt ratio.
However, these measures might lead to loss of output or higher interest rate payments, respectively. Moreover, Abbas, Akitoby, Andritzky, Berger, Komatsuzaki, and Tyson (2013) supported that the key elements to reducing the debt ratio are the growth and the primary fiscal balance with the inflation and interest rates to be less effective. But since the economic environment is now weak and debt reduction is less likely to occur by normal growth measures, the countries should put extra effort into promoting GDP growth, such as fiscal consolidation measures to protect the programs that have large impact on growth.
Only a small number of research investigated how particularly inflation affects debt. Akitoby, Komatsuzaki, and Binder (2014) examined the effect of different levels of inflation on the public debt-to-GDP ratio in the G-7 countries. Their results suggested that there is a negative relationship between these two variables for the advanced economies and thus, a higher inflation would lead the debt-to-GDP ratio to decrease. On the other hand, Reinhart and Rogoff (2010) by studying the link between inflation and public debt for 44 countries, showed that advanced economies do not appear to have any systematic link between these two variables, but the emerging countries show a positive relationship between inflation and debt. In addition, the efficiency of inflation to reduce the debt ratio is linked to the maturity of the debt. Hall and Sargent (2010) showed that from 1946 to 2008 the effect of inflation on the US debt ratio decreased due to shorter maturity of the debt. Moreover, Aizemna and Marion (2009) stated that another factor that affects the impact of inflation on debt ratio is the amount of the debt that foreign creditors hold. The larger the volume of the debt that foreign creditors hold, the larger the temptation for higher inflation.
The results of the effects of fiscal consolidation on debt ratio during a period of crisis do not differ. Monastiriotis (2011), and Cavero and Cortés (2013) examined the impact of austerity measures in Greece. Their conclusions suggested that the results of these measures are the increase of unemployment, poverty, and inequality. Gechert and Rannenberg (2015) found that the main reasons for the collapse of Greek GDP are the tax increases and the cuts in the government spending. Also, they supported that if the austerity measures would not have been implemented during the crisis but after the country had recovered, then even if there was slow or no growth, at least the country would not have been led in a prolonged period of depression. Attinasi and Metelli (2017) using a panel VAR approach, studied the impact of a fiscal
consolidation on the debt-to-GDP ratio in the euro area countries. They concluded that if the austerity measures are implemented via budget cuts then the consolidation is successful since the debt-to-GDP decreases. Fiscal consolidation, though, through tax increments, might be proved self-defeating in recessionary periods. Moreover, the effect of austerity measures goes beyond the mere economic area. Austerity measures are directly connected with negative social impacts like the increase on the suicide rate (Antonakakis & Collins, 2014) and the crime rate (Cavero & Cortés, 2013). Liu (2013) studied the effects of the fiscal policy in seven European countries other than Greece. His findings showed again that the fiscal consolidation in these countries had vast negative effects. The debt-to-GDP increased, as also the unemployment, the suicide, and the depression rate, while the well-being of the citizens deteriorated. However, it is not enough only the reduction of the debt ratio. The decrease of this ratio on a sustainable level is important. The IMF, in order to help advanced countries to face the crisis and the public debt sustainability concerns, has developed a
Moghadam, 2011). This paper gives the structure of how a detailed DSA should be implemented. One has to examine the evolution of the public debt, and the
susceptibilities related not only to the debt level but also to the debt profile, like spreads and average maturity. Also, a DSA should contain an analysis of the fiscal risks and the assessment of the impact that risks might have.
A few studies on Greek public debt sustainability exist. Buiter and Rahbari (2010), Cline (2011), Garcia-Puente, Khoury, Grigoryan, and Yamada (2012), Apergis and Cooray (2014) and the country report that the IMF published in 2015 found the public debt dynamics to be on an unsustainable path. However, the debt can be sustainable if, indeed, the country follows a set of fiscal measures so that government revenues would rise and government spending would decrease at the same time.
The evolution of the Greek public debt is analyzed in many papers. Fouskas (2013) discussed the effect that the excessively prolonged military spending had on public debt and the fact that previous Greek governments chose to act in favor of some entrepreneurs by tax evasion. Kouretas and Vlamis (2010), Contessi (2012),Husson (2015), Truth Committee (2015) and Nikiforos et al (2016) conducted research over long periods to examine what factors contributed the most to debt accumulation. The results showed that excessive interest rates, primary deficits, illicit capital outflows and unjustified military spending are the main contributors to the debt accumulation. Since policy reforms should be done, Contarelli (2010), OECD (2010), Varoufakis (2015) and Rocholl and Stahmer (2016) have already suggested some realms where reforms should be implemented such as at the side of taxes, the budget preparation, the government spending, the pension system, the competition, the innovation and the Green growth.
Despite the overwhelming research in course in these areas, there is still a vast room improvement and many issues to be clarified. This is why I try to step forward to disentangle the complicated interactions and effects that macroeconomic variables have on the debt-to-GDP ratio.
3. Evolution of Greek Public Debt
Despite the fact that the Great Recession played a crucial role in the increase of debt-to-GDP ratio, Greece was already exposed, due to economic difficulties that have been accumulated for years. Nikiforos et al (2016), Truth Committee on public debt (2015), and Husson (2015) presented some of the factors that contributed to the accumulation of the public debt, the amounts of the bailout packages and the use of them. According to their findings, during the period 1980 to 1993, the debt to GDP was skyrocketed to 95 percent. The main factor for this increase was the difference between the implicit interest rate that the government was paying to service the debt and the nominal GDP growth rate, the so-called “snowball effect”. If this difference is positive then the debt ratio increases, even if the government spends exactly the same as it receives, and the debt gets more expensive (Husson, 2015). In less extent,
government deficits and illicit capital outflows also contributed to this increase (Truth Committee, 2015).
In 1992, Greece signed the Maastricht Treaty. In order to become a member of the Eurozone and adopt the euro at the end of 90’s, Greece had to stabilize the prices and to lower the government deficit under 3 percent. To achieve the price stabilization, the hard-drachma policy came into force by the Bank of Greece (Nikiforos et al, 2016). Thus, “the drachma was allowed to depreciate relative to the European Currency Unit by less than the full inflation differential” (Detragiache & Hamann, 1997). However,
the combination of the hard-drachma policy and the adoption of the euro in 2001 caused the real exchange rate to appreciate by 40 percent resulting on a current account deficit of 15 percent at the beginning of the crisis in 2008 (Nikiforos et al, 2016). As seen in Figure 1, in 2009 the debt-to-GDP ratio increased to almost 127 percent (Hellenic Statistical Authority, 2017).
As mentioned above, in 2010 Greece asked financial assistance from the international institutions which came by Troika (European Commission, European Central Bank, & International Monetary Fund). The objectives of the bailout packages were to restore confidence, competitiveness and credibility for private investors, access to the
markets, and debt stabilization (European Commission, 2010). The austerity measures that were put in place involved job cuts, tax increases and privatizations resulting in a deeper recession (Truth Committee, 2015). The first package started in May 2010 and lasted until March 2012. The initial amount of €110 billion that was scheduled to be distributed decreased to €107.3 billion as Slovakia finally did not take part and Ireland and Portugal was in need of financial assistance as well. From that amount, finally, the volume of €73 billion was distributed during the period of the first program. By the end of the first package, the debt-to-GDP ratio increased to 138.73 percent (Nikiforos et al, 2016).
The second program began in March 2012 and ended in June 2015. It included €34.3 billion un-disbursed from the first program and in total was planned to reach €172.6 billion. This time, €153.8 billion actually was distributed and from that amount €10.9 billion was paid back by the Hellenic Financial Stability Fund (HFSF) to European Financial Stability Facility before the program was finished. Only for 2012, we see a decrease of the debt ratio due to 50 percent haircut. The ratio continues to increase up to 179 percent (Nikiforos et al, 2016).
The third program which started in August 2015 and is scheduled until May 2018 reaches €86 billion. The first installment was including €26 billion long term government bonds of which €21.4 billion was actually used till the end of 2015 (Rocholl & Stahmer, 2016). During 2016, the second installment of €10.3 billion was disbursed (European Stability Mechanism, 2016). On the 7 of July 2017 and after the second positive review of the progress of the program implementation, the European Stability Mechanism Board announced the approval of the third tranche of $8.5 billion which is planned to disturbed in two sub-tranches (European Stability Mechanism, 2017).
According to Rocholl and Stahmer (2016), the sum of the first two adjustment programs was €215.9 billion and almost 95 percent of the money was used to pay back the creditors and to recapitalize the domestic banks. Figure 2 shows that €139.2 billion was used for debt repayments and interest payments, €37.3 billion was paid to HFSF for the recapitalization of the banks, €29.7 billion was paid for the Private Sector Involvement, and only €9.7 billion of the total amount was used for household residuals. Like that, the objective of the bailout packages turned to be a safe out for the European and Greek banks that were exposed to Greek bonds (Truth Committee, 2015).
As we can see, the Greek public debt increased dramatically immediately after 2008 and continues to increase during the austerity period. This is why it is important the implementation of a set of measures that will stop the upward route of the ratio.
Figure 1: Greece Government Debt-to-GDP
Figure 2: 1st and 2nd bailout programs money allocation
4. Empirical Model and Data 4.1 Empirical Model
Following Cherif and Hasanov (2012), in order to test how austerity, inflation and growth shocks affect Greek public debt, I include as endogenous variables in the vector Y in equation (1) the primary balance-to-GDP ratio (instead of the primary deficit, pb), the real GDP growth rate (g), the inflation rate calculated by the GDP deflator (π) and the average nominal interest rate based on interest payments on debt (i). Moreover, the debt-to-GDP ratio (d) will be included as an exogenous variable in equation (1). The exogenous dummy variable E takes value 0 for all quarters prior 2001 and takes value 1 for data after 2001. The reason that I include this dummy variable is to test if the adoption of the euro currency has a significant effect on debt.
= ∑ + + + (1)
The numbers of lags that I used are k = 4. The debt dynamics are described by equation (2):
= ( )
( )( ) − (2)
The concept is that the nonlinear government budget constraint enters the system as an identity in order to endogenize the feedback from the debt dynamics in the VAR (Stanova, 2015). Thus, following Stanova (2015) and Favero and Giavazzi (2009), I include the debt-to-GDP ratio as an endogenous variable.
Keeping track of the debt dynamics in equation (1) is important to compute the IRs. All the variables that will be used for the calculation of the debt in equation (2) are
included in equation (1) and thus, I have to estimate only equation (1). Once the VAR is estimated the IRFs are computed. I use orthogonalized IRs to ensure that each shock is independent of the other and I obtain isolated effects in each graph for each different shock. A bootstrap methodology is used for the calculation of the standard errors and the confidence intervals for the IRs.
Cherif and Hasanov (2012) used both narrative and structural identification methods to identify a causal shock in the initial period in order to estimate the IRs. They used the narrative approach of Romer and Romer (2011) and thus, they included
exogenous tax shocks as primary balance shocks in the VAR. As structural approach, to identify if there was a contemporaneous correlation between VAR residuals, they used Blanchard and Perotti (2002). The last identification method they used was GIR methodology where a shock to a variable is accompanied with shocks to the other variables. The order of the variables determines which variable is shocked first. For the purpose of this thesis, I only applythe narrative method of Romer and Romer (2010) and thus, I include exogenous tax shocks as primary balance shocks which are included in the VAR as exogenous variable. These tax shocks are defined as
exogenous by Romer and Romer (2010) given that the motivation to implement these changes was not to counterbalance changes on other factors likely to affect growth but to reduce deficit. For example, changes on taxes because of forecasted recession or because of changes on government spending are occurred to offset the impact on output, and thus they are not exogenous. On the other hand, tax changes that are taken to achieve higher normal growth or to reduce deficit that has been accumulated over years are exogenous.
4.2 Data and Descriptive statistics
The whole sample of the data that are used for the analysis are quarterly series and they are covering the period from the first quarter of 1985 to the last quarter of 2015. Only for the GDP deflator, the data are covering the period from the first quarter of 1984 to the last quarter of 2015 in order to be used for the calculation of the inflation. Data regarding the primary balance of the general government as percentage of GDP are taken from the Oxford Economics. Likewise, interest payments and debt-to-GDP ratio come also from the Oxford Economics. On the other hand, GDP deflator and real GDP growth rate data are taken from OECD. Inflation is calculated based on GDP deflator.
The data for the narrative approach have been retrieved from different sources, mainly from official papers published by the central bank of the country. Given the sources, I identified thirteen exogenous tax changes with main goals the decrease of the general government deficit and the increase of competitiveness. Most of these tax changes are identified in the Annual Reports that every year the Bank of Greece presents for the previous financial year, at the meeting of the shareholders every April. In these papers, the Bank of Greece set discussions, among others, about tax policy measures, the magnitude of tax changes and the motivation behind them. Due to the lack of Annual reports before 1997, data have also been retrieved from the paper written by Alogoskoufis (1992) who has been Minister of Economy and Finance in Greece. Also, tax changes have been identified in the paper written by Pelagidis (1997) who assessed the economic policies in Greece during the period 1990-1993.
The history of exogenous tax changes identified from the above sources is as follow: At 1985q1 the corporate tax increased by 4 percent from 45 percent to 49 percent. At 1993q2 the corporate tax decreased from 40 percent to 35 percent. At 2001q1 the same
tax decreased by 2.5 percent. At 2005q2 the standard Value-added tax (VAT) increased by 1 percent from 18 percent to 19 percent. At 2006q1 and 2007q1 the corporate tax decreased by 3 percent and 4 percent, respectively. At 2009q1 and 2010q2 the tax for cigarettes increased by 5 percent and the tax for alcohol increased by 20 percent, respectively. At 2010q2 and 2010q3 the VAT increased by 2 percent each quarter. At 2011q1 the reduced VAT increased by 2 percent. At 2013q3 the VAT for restaurants and catering decreased by 10 percent only for this quarter and finally at 2015q3 the corporate tax increased by 4 percent. As mentioned above, all these tax changes aimed mainly at reducing the general government deficit and increasing the competitiveness of the Greek economy.
The descriptive statistics in Table 1 show that during the thirty years of the data sample, Greece has an average debt of 104.8 percent of GDP. The statistics show that debt reached an all-time high of 181.5 percent of GDP in the second quarter of 2014, whereas the lowest level reaches 52 percent of the GDP in 1985. The primary balance has an average of -9 percent of GDP. The lower value is -30.6 percent of GDP in the second quarter of 2013 which is equivalent to say that 30.6 is the maximum value that primary deficit reached. The maximum value of primary balance is 1.8 percent of GDP in the third quarter of 2000 and thus the maximum value is a surplus. Real growth of the country is averaged 1.1 percent of GDP. The maximum and the minimum real growth rate of GDP is 11 percent in the third quarter of 1991 and -10.2 percent in the first quarter of 2011, respectively. The mean value of inflation is 7.3 percent while it is observed to have reached a maximum of 23.7 percent in the fourth quarter of 1990 and a minimum of -4.1 percent in the first quarter of 2014. Finally, Greek interest rate has an average of 8 percent. The highest value for the interest rate has been noted in the
fourth quarter of 1985 at 53 percent while the smallest interest rate is -47.5 percent in the third quarter of 2012.
Table 1: Descriptive Statistics
Variable Mean Std. Dev. Min Max
Primary Balance -0.09 0.041 -0.306 0.018
Growth 0.011 0.041 -0.102 0.110
Inflation 0.073 0.072 -0.041 0.237
Interest 0.080 0.192 -0.475 0.531
Debt 1.048 0.338 0.520 1.815
Before proceeding to the VAR analysis, two statistical tests are performed. First, I check the data for stationarity by using Dickey-Fuller test. Table 2 includes the results. I find that the t-statistic for growth is higher than the critical values in all confidence levels and thus the null hypothesis of unit root can be rejected, the variable is stationary. Likewise, I reject the null hypothesis for the primary balance-to-GDP ratio at 95 percent confidence level. However, inflation, interest rate, and debt ratio are not stationary given that we cannot reject the null hypothesis of unit root. Thus, inflation, interest rate, and debt ratio have been detrended.
Table 2: Dickey-Fuller Test for Unit Root
Variable Test Statistic P-Value
Growth -3.653 0.0048
Inflation -1.267 0.6443
Interest -2.242 0.1912
Primary Deficit -8.309 0.0000
Debt 0.016 0.9598
The next step is to test for the optimal number of lags that I should use to run the VAR model. Table 3 presents the results. The lag selection test shows that the optimal number of lags to be included is four. This is given by the number of asterisks assigned to the fourth lag for the different criteria.
Table 3: Lag Selection
Lag LL LR df p FPE AIC HQIC SBIC
0 1196.62 1.9e-15 -19.69 -19.55 -19.34
1 1268.17 143.1 25 0.000 8.9e-16 -20.47 -20.09 -19.5*
2 1292.25 48.146 25 0.004 9.0e-16 -20.45 -19.84 -18.94
3 1312.86 41.229 25 0.022 9.8e-16 -20.38 -19.53 -18.29
4 1389.07 152.43* 25 0.000 4.2e-16* -21.2* -20.2* -18.56
Finally, a post-estimation test is performed to assure the stability of the VAR model. Table 4 shows that all eigenvalues lie inside the unit circle, hence our model does satisfy the stability condition.
Table 4: Stability Test
Eigen Value Modulus
0.69 + 0.53 0.87 0.69 - 0.53 0.87 -0.58 + 0.63 0.86 -0.58 - 0.63 0.86 -0.83 0.83 0.8 + 0.14 0.82 0.8 - 0.14 0.82 0.02 + 0.74 0.74 0.02 - 0.74 0.74 -0.45 + 0.55 0.71 -0.45 - 0.55 0.71 0.52 + 0.48 0.7 0.52 - 0.48 0.7 0.18 + 0.6 0.62 0.18 - 0.6 0.62 0.58 0.58 -0.44 + 0.36 0.57 -0.44 - 0.36 0.57 -0.45 0.45 -0.09 0.09
5. Analysis 5.1 Results from the VAR
Table 5 shows the results of the estimated coefficients of the VAR. The first column represents the independent variable and its respective lag, from 1 to 4. If we assume a period t as starting point, then Lag 1 stands for period − 1, Lag 2 for − 2 and so on. From column (2) to (6) we have each of the dependent variables included in our endogenous system, starting with Debt until Primary Balance.
a) Effects on Debt
As column (2) in Table 5 shows, the effect of GDP growth and its lags is negative (except for lag 4). This goes in line with the result of Cherif and Hasanov (2012), given that an increase in GDP growth is positively associated with economic improvement which in normal times has negative effect on government debt. However, this effect is non-significant. In the case of interest rate, we observe a positive effect for the two first lags and a negative one for the third and fourth lags. The most immediate effect of a rise in interest rate is that government debt becomes more expensive, hence the positive sign is not surprising. Nevertheless, neither the positive effect nor the negative are statistically significant.
Inflation shows a steady negative effect on debt that goes from -0.106 percent to a maximum of -0.314 percent. A normal range of inflation is associated with good economic times, which justify the negative sign. Nevertheless, these effects are again not significant.
In the case of primary balance, the effect on debt seems to be negative for all the lags, but only for lags 1 and 4 it is statistically significant at 90 and 95 percent confidence respectively. Given the definition of primary balance, a one percent increase means
that either revenues increase or expenditures decrease. The negative signs then go in line with the mainstream recipe of austerity as a measure to lower debt. Increasing taxes or decreasing government expenditures have been one of the main
recommendations given to the Greek government. Thus, a one percent increase in the primary balance for lag 1 will reduce the debt ratio by 0.199 percent. Moreover, at the lag 4 the debt ratio reduces even more, by 0.224 percent.
As expected the lags of the debt ratio have a very significant effect on current debt ratio. This negative effect can be explained by the fact that if debt ratio in past periods has increased then it is more likely that current period’s debt will decrease. This follows the fact that government facing high debts will act in order to decrease it. The magnitude as well as the significance of the effect decrease for older lags.
Finally, the dummy variable E which control for the adoption of euro as currency in 2001, the exogenous tax change, and the constant term have no significant effect on debt. However, a one percent increase in the debt dynamics variable seems to increase the debt ratio by 0.003 percent at 95 percent confidence level.
b) Effects on GDP growth.
At the third column of Table 5, we see the effects on GDP growth. If lags of GDP growth increase by 1 percent, it means that the economy is doing well and all the others constant it is expecting that it will continue to grow. Only for the fourth lag, the increase of GDP growth rate has a significant negative effect on current GDP growth rate but as the economy is arriving at the starting point the effect becomes significant positive at 99 percent level and the magnitude increases. At lag 1 a one percent increase in GDP growth rate will increase current GDP growth rate by 0.577 percent.
The effect of a 1 percent increase in interest rate has a positive effect on GDP growth for lags 1 and 3 and a negative effect for lags 2 and 4 as it was expecting. Despite the fact that the positive effect is significant, the negative effect at lag 2 is even more significant at 99 percent level and the magnitude is bigger as GDP growth rate decreases by 0.152 percent since people spend less and save more due to higher interest rate and since there is less incentive to invest.
If inflation increases by 1 percent then the GDP growth rate goes up at lags 2 and 3 but decreases at lags 1 and 4. As I mentioned before, the increase of inflation is a positive sign of the progress of the economy and as that, we would expect the GDP growth rate would increase. However, none of the effects is statistically significant. The effects on GDP growth rate of a one percent increase in primary balance are not as we would expect. If primary balance increases then either government revenues increase or government spending decreases. Both of them would lead the economy to shrink as people would spend less. However, we see a positive effect on GDP growth rate which though is not statistically significant at any lag.
As debt ratio increases by one percent, the GDP growth rate will increase for all lags except lag 2. That results go against the theory that wants the increase of debt ratio to decrease growth rate. However, it might be the case the higher sustainable debt is adopted in order to boost the economy, which would explain the positive relation evidenced here. The effects of the debt ratio on GDP growth rate are not statistically significant for any lag.
Lastly, the adoption of euro variable significantly decreases the growth rate by 0.011 percent at 95 percent confidence level, while a one percent increase in constant term
will increase the GDP growth rate at 99 percent confidence level. The exogenous tax shock and the debt dynamics variable have almost no effect on growth.
c) Effects on Interest Rate
In column 4 of Table 5, we can see the effects on interest rate if the independent variable increases by one percent. As we can see interest rate increases for all GDP growth rate lags except from lag 2 where it decreases although this decrease is not statistically significant. The increase in the interest rate is in line with the intuition which wants higher GDP growth rate to increase the real money demand and thus the interest rate increases. However, only the increase at lag 1 and the decrease at lag 4 are statistically significant at 99 percent confidence level.
The effect of a one percent increase in inflation on the interest rate differs between the lags. Higher inflation means higher prices so the demand for real money increases and interest rate increases to counterbalance the effects. Thus, we see at lags 2 and 4 one percent increase of inflation to increase interest rate. On the other hand, at lags 1 and 3 the interest rate decreases. However, none of the lags are statistically significant. The effect of a one percent increase in primary surplus is significant and positive for the two first lags and negative for the third and fourth lag. Higher primary surplus means that the government does not need anymore to lure investors to buy the debt so interest rate decreases.
As we expected, higher debt ratio increases the interest rate for all lags. When the debt ratio increases the government needs to attract investors to buy the extra debt and thus interest rate increases. The effect is statistically significant for lags 1 and 4 at 95 percent confidence interval. A one percent increase in the debt ratio will cause the
interest rate to increase by 0.539 percent at first lag and this magnitude lowers at 0.42 percent at fourth lag.
The dummy variable E, the exogenous tax changes, and the constant term do not have any statistical significant effect on interest rate while the debt dynamics variable tends to decrease the interest rate at 99 percent confidence level, although the magnitude of this effect is very small.
d) Effects on Inflation Rate
As column 5 in Table 5 shows, a one percent increase in GDP growth rate lags has a positive impact on inflation except from lag 3. As economy is growing, the aggregate demand for goods and services increases resulting on a higher inflation. However, only for lag 1 the positive effect is significant where inflation rises by 0.086 percent at 90 percent confidence level.
On the other hand, we would expect a higher interest rate to decrease inflation as people save more and spend less so the aggregate demand decreases and prices decrease. However, we see exactly the opposite effect with the lag four to be statistically significant at 95 percent level.
The effects of the inflation lags have alternate signs. However, inflation falls at lag 2 and 4 at 90 percent and 99 percent significant level, respectively.
Inflation rises at all lags of primary balance except from lag 3. As we know, an increase in primary balance causes people to spend less, possibly due to higher taxes, and thus demand decreases and so the prices. However, the effects from the primary surplus increase are not statistically significant.
We would expect higher debt ratio to increase inflation. However, only at lag 4 inflation behaves like that. For the rest of the lags, the inflation decreases with the second lag to be even statistically significant at 95 percent confidence interval.
The effect of the dummy variable E is positive and significant at 90 percent level. The effects of the exogenous tax changes are also significant at 90 percent confidence level but its magnitude is very small. The debt dynamics variable and the constant effects are not statistically significant.
e) Effects on Primary deficit
In the last column of Table 5, we see that as it was expecting, higher growth by one percent leads to a lower primary balance as probably people pay less taxes. Only at the first lag the primary surplus increases without though the effect to be significant. Only the negative effect at fourth lag is statistically significant at 90 percent
confidence interval.
The lags of interest rate have a positive effect on primary balance except for lag 2 where this effect is negative and not significant. The effects of lag 3 and 4 are statistically significant at 95 and 90 percent confidence level respectively.
Inflation has a clear negative effect on primary balance evidenced by the negative sign of all its lags. However, only lag 1 is statistically significant at 90 percent confidence level.
One percent increase in the primary balance at the past will increase the current primary balance at all lags. Although only the last lag has a significant effect on current primary balance.
The effect of one percent increase in the debt ratio will cause the primary balance to decrease at all lags as the primary deficit presents the accumulation of debt and so it increases. The effect of the fourth lag is statistically significant at 90 percent level with one percent increase of the debt ratio to cause primary balance to decrease by 0.109 percent.
Neither of the exogenous variables or the constant term has a significant effect on the primary balance.
Table 5: Main Results
(1) (2) (3) (4) (5) (6)
Variable/Lag Debt Growth Interest
Rate Inflation Primary Balance Growth/1 -.260 .577*** .383** .086* .139 Growth/2 -.253 .284*** -.124 .058 -.003 Growth/3 -.130 .032 .139 -.195*** -.094 Growth/4 .235 -.175** .091 .069 -.227* Interest Rate/1 .065 .081* .255*** .019 .098 Interest Rate/2 .126 -.152*** .041 .015 -.085 Interest Rate/3 -.018 .077* -.005 .003 .141** Interest Rate/4 -.014 -.04 -.584*** .044** .249*** Inflation/1 -.162 -.011 -.148 .108 -.538** Inflation/2 -.314 .169 .006 -.062* -.14 Inflation/3 -.242 .119 -.154 .103 -.273 Inflation/4 -.106 -.071 .002 -.356*** -.065 Primary Bal./1 -.199* .088 .207** .002 .076 Primary Bal ./2 -.06 .011 .175* .003 .068 Primary Bal./3 -.015 .017 -.218** -.039 .006 Primary Bal./4 -.224** .09 -.173* .007 .458*** Debt/1 -.342*** .036 .539*** -.024 -.082 Debt/2 -.185* -.0084 .119 -.061** -.080 Debt/3 -.168* .01 .043 -.029 -.129 Debt/4 -.006 .024 .42*** .005 -.109* E -.007 -.011** .010 .005* -.004 Tax .000 .000 .000 -.000* -.001 Debt Dynamic .003** -.001 -.003*** -.000 -.000 Constant -.005 .029*** -.016 -.007 -.021
5.2 Debt Impulse Responses
At this section, I present the results from the impulse response functions given a positive initial shock in each of the homogeneous variables included in the VAR model. Figure 3 shows the IRFs of positive shocks in austerity (denoted by pbgdp), GDP growth (growth) and inflation (inf_det) using as dependent variable debt as a percentage of GDP (d_det).
The bottom right graph shows the response of debt ratio facing an austerity shock. Thus, either taxes increase or expenditures decrease so the primary deficit balance increases. The immediate response is a decreased debt of about 0.7 percent of GDP which supports the arguments used by the European institutions of austerity measures in order to decrease debt. However, during the two periods after the shock, the debt ratio increases getting close to the pre-shock state. From period three until fourth there is a sharp decrease of similar magnitude as the immediate response. Finally, from period four on the debt ratio stabilize close but lower than the pre-shock value. As seen in the bottom left graph, a one standard deviation increase in the inflation causes the debt ratio to fall slightly from quarter one to three, before it returns to its pre-shock path in quarter four. As the first reaction to the shock, the debt ratio increases slightly of about 0.5 percent of GDP. This initial positive impact declines sharply when reaching period one. Usually, inflation, when bounded in a normal range, is associated with a good economic situation. This might be the reason behind the slight decline of debt in periods one to three.
The top right graph shows the relationship between interest rate and debt as a percentage of GDP. The immediate effect is a very slightly increment in debt which keeps sluggishly growing until period two. This is direct given that a higher interest
rate on payments means a higher debt. From period two to three, there is a decrease that leads the debt ratio to a lower value than the pre-shock state. From period four onwards there are positive and negative effects but the trend fluctuates very close to the pre-shock value.
Finally, the top left graph shows that a positive shock on GDP growth causes the debt ratio to instantaneously fall by about 0.5 percent and keep under the pre-shock value from quarter one to four. The rise of debt ratio begins in quarter three but it is not until eighth quarter that debt ratio reaches its pre-shock path. This negative effect in debt ratio goes in line with the intuition given that a positive shock in GDP growth rate is closely associated with an improvement in the economy.
For a complete graphical view of all interactions among the dependent variables refer to Appendix 1.
Figure 3: Impulse-Response Functions of debt ratio -.02 -.01 0 .01 -.02 -.01 0 .01 0 2 4 6 8 0 2 4 6 8
res8, growth, d_det res8, i_det, d_det
res8, inf_det, d_det res8, pbgdp, d_det
95% CI
orthogonalized irf
step
6. Conclusion
In this paper, I investigate the dynamics of the Greek Public Debt in response to austerity, GDP growth, and inflation shocks. By using a VAR model and keeping track of the debt dynamics, I observe that an austerity shock is more likely to reduce the debt ratio than to increase it, as perceived by the Greek experience. Despite the fact that somewhere after the second quarter the debt ratio returns to its pre-shock path, after two years it will stabilize almost 0.5 percent lower than its initial path. Given that result, an austerity shock is not self-defeating when it is applied on the Greek data despite the weak economy and what previous evidence has shown. On the other hand, the effects of inflation and growth shocks are to reduce the debt ratio for the few first periods before it converges to its pre-shock path. In the case of inflation shock, after eight quarters the value of debt-to-GDP is even higher than the pre-shock level.
All in all, the results tend to support the austerity measures imposed by the European institutions, despite the fact that Greek people have perceived their effects differently. It is important to point out that future research could include other approaches besides the narrative one, to assure that shocks in every dependent variable are exogenous. .
References
Abbas, S., Akitoby, B., Andritzky, J., Berger, H., Komatsuzaki, T. and Tyson, J. (2013). Dealing with High Debt in an Era of Low Growth. Staff Discussion
Notes, 13(7), pp.27-28.
Aizenman, J., & Marion, N. (2011). Using inflation to erode the US public debt. Journal of Macroeconomics, 33(4), pp.524-541.
http://dx.doi.org/10.1016/j.jmacro.2011.09.001
Akitoby, B., Komatsuzaki, T., & Binder, A. (2014). Inflation and Public Debt Reversals in the G7 Countries. IMF Working Papers, 14(96), pp.1. http://dx.doi.org/10.5089/9781498369954.001
Alencontre.org. (2015). A l'encontre » Grèce: pourquoi une dette à 100% du PIB
avant la crise?. [Greece: Why a 100% debt-to-GDP before the crisis?] [online]
Available at: http://alencontre.org/europe/grece-pourquoi-une-dette-a-100-du-pib-avant-la-crise.html [Accessed 12 Jul. 2017].
Alogoskoufis, G. (1992). Fiscal Policies, Devaluations, and Exchange Rate Regimes. The Stabilization Programmes of Ireland and Greece. The Economic and Social
Review, 23(3), pp.225-246.
Antonakakis, N., & Collins, A. (2014). The impact of fiscal austerity on suicide: On the empirics of a Modern Greek tragedy. Social Science & Medicine, 112, pp.39-50. http://dx.doi.org/10.1016/j.socscimed.2014.04.019
Apergis, N., & Cooray, A. (2013). Forecasting Fiscal Variables: Only a Strong Growth Plan Can Sustain the Greek Austerity Programs - Evidence from
Simultaneous and Structural Models, CAMA Working Paper 25/2013, 3. http://dx.doi.org/10.2139/ssrn.2265694
Apergis, N., & Cooray, A. (2014). The dynamics of Greek public debt – evidence from simultaneous and structural VAR models. Applied Economics, 47(10), pp.967-980. http://dx.doi.org/10.1080/00036846.2014.985372
Attinasi, M., & Metelli, L. (2017). Is fiscal consolidation self-defeating? A panel-VAR analysis for the Euro area countries. Journal of International Money and
Finance, 74, pp.147-164. http://dx.doi.org/10.1016/j.jimonfin.2017.03.009
Bank of Greece. Various Annual Reports.
Blanchard, O., & Perotti, R. (2002). An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output. The Quarterly
Journal of Economics, 117(4), pp.1329-1368.
http://dx.doi.org/10.1162/003355302320935043
Buiter, W. H., & Rahbari, E. (2010). Greece and the fiscal crisis in the EMU. Centre
for Policy Research Paper (Policy Insight No. 51).
Cavero, G. & Cortes, M. (2013). The Truth Cost of Austerity and Inequality. Oxfam
GB case study, pp.1-9.
https://www.oxfam.org/sites/www.oxfam.org/files/cs-true-cost-austerity-inequality-greece-120913-en.pdf
Cline, W. (2011). Sustainability of Greek Public Debt. Washington: Peterson Institute
for International Economics, pp.1-17
Contarelli, C. (2010). From Stimulus to Consolidation: Revenue and Expenditure Policies in Advanced and Emerging Economies. IMF, April, 30, 2010.
Contessi, S. (2012). An Application of Conventional Sovereign Debt Sustainability
Analysis to the Current Debt Crises. Federal Reserve Bank of St. Louis Review,
94(3), pp.197-220.
Cottarreli, C. & Moghadam, R. (2011). Modernizing the Framework for Fiscal Policy and Public Debt Sustainability Analysis. IMF, August, 5, 2011.
Detragiache, E., & Hamann, A. (1997). Exchange Rate-Based Stabilization in Western Europe: Greece, Ireland, Italy and Portugal. IMF Working
Papers, 97(75), pp.1-29. http://dx.doi.org/10.5089/9781451849691.001
European Commission. (2010). The Economic Adjustment Programme for Greece. pp.1-93. Brussels: European Commission.
European Commission. (2016). Fiscal Sustainability Report 2015. Institutional Paper, pp.3.
Esm.europa.eu. (2016). ESM Board of Directors approves €7.5 billion disbursement
to Greece | European Stability Mechanism. [online] Available at:
https://www.esm.europa.eu/press-releases/esm-board-directors-approves-%E2%82%AC75-billion-disbursement-greece [Accessed 14 Jun. 2017].
Esm.europa.eu. (2017). ESM Board of Directors approves €8.5 billion loan tranche to
Greece | European Stability Mechanism. [online] Available at:
https://www.esm.europa.eu/press-releases/esm-board-directors-approves-%E2%82%AC85-billion-loan-tranche-greece [Accessed 10 Jul. 2017]. Favero, C., & Giavazzi, F. (2007). Debt and the Effects of Fiscal Policy. NBER
Favero, C., & Giavazzi, F. (2009). How large are the effects of tax changes? NBER Working Paper 15303, pp.8-10. http://dx.doi.org/10.3386/w15303
Fouskas, V. (2013). Whatever Happened to Greece? The Political Quarterly, 84(1), pp.132-138. http://dx.doi.org/10.1111/j.1467-923x.2013.12002.x
Garcia-Puente, P., Khoury, Z., Grigoryan, V., & Yamada T. (2012). Is Greece on the Right Track? Columbia University working paper 08/12-01E, pp.1-32.
Gechert, S., & Rannenberg, A. (2015). The Costs of Greece’s Fiscal
Consolidation. Vierteljahrshefte Zur Wirtschaftsforschung, 84(3), pp.47-59. http://dx.doi.org/10.3790/vjh.84.3.47
Hall, G., & Sargent, T. (2010). Interest Rate Risk and Other Determinants of Post-WWII U.S. Government Debt/GDP Dynamics. NBER, W15702, pp.13. http://dx.doi.org/10.2139/ssrn.1673451
Hasanov, F., & Cherif, R. (2012). Public Debt Dynamics: The Effects of Austerity, Inflation, and Growth Shocks. IMF Working Papers, 12(230), pp.1-13.
http://dx.doi.org/10.5089/9781475510553.001
Hellenic Statistical Authority. (2017). The Greek Economy (07 July 2017). Athens: Hellenic Statistical Authority. pp.6.
Husson, M (2015). Dette grecque: quatre ‘variables de fuite’ [Greek debt: four ‘leak variables’]
International Monetary Fund. (2015). Greece: Preliminary Draft Debt Sustainability Analysis. IMF Staff Country Reports, 15(165), pp.1-23.
Kouretas, G., & Vlamis, P. (2010). The Greek crisis: Causes and implications. Panoeconomicus, 57(4), pp.391-404.
http://dx.doi.org/10.2298/pan1004391k
Liu, L. (2014). The Austerity Trap: Economic and Social Consequences of Fiscal Consolidation in Europe. CreateSpace Independent Publishing Platform, pp.1-73 Monastiriotis, V. (2011). The Greek crisis in focus: Austerity, Recession, and paths to
Recovery. London School of Economics. pp.1-111.
Nikiforos, M., Papadimitriou, D., & Zezza, G. (2016). The Greek Public Debt
Problem (pp. 1-30). Levy Economics Institute of Bard College.
OECD. (2010). Greece at a Glance: Policies for a Sustainable Recovery. OECD Publishing. Paris. http://dx.doi.org/10.1787/9789264201781-en
Pelagidis, T. (1997). Economic Policies in Greece during 1990-1993: An Assessment. Journal of Modern Greek Studies, 15(1), pp.67-85. http://dx.doi.org/10.1353/mgs.1997.0007
Reinhart, C., & Rogoff, K. (2010). Growth in a Time of Debt. Cambridge: National
Bureau of Economic Research. pp.2. Retrieved from
http://www.nber.org/papers/w15639
Rocholl, J., & Stahmer A. (2016). Where did the Greek bailout money go? ESMT
White Paper No. WP–16–02.
Romer, C., & Romer, D. (2010). The Macroeconomic Effects of Tax Changes: Estimates Based on a New Measure of Fiscal Shocks. American Economic
Sims, C. (1980). Macroeconomics and Reality. Econometrica, 48(1), pp.1-48. http://dx.doi.org/10.2307/1912017
Stanova N. (2015). Effects of Fiscal Shocks in New EU Members Estimated From a SVARX Model with Debt Feedback. MPRA Paper No. 63148, pp.1-29.
Truth Committee. (2015). Truth Committee on Public debt. http://debt-truth.gr/english/
Varoufakis, Y. (2015). A Policy Framework for Greece’s Fiscal Consolidation, Recovery, and Growth. Athens: Ministry of Finance Hellenic Republic. pp. 1-57.
