^{1}

^{2}

^{3}

The purpose of this paper is to test the relationship between budget deficits and inflation for nine EU countries during the period of 1990-2013 using the quarterly data. Recently, the public deficits and inflation have had an increasing importance for developing/emerging and developed countries to build the stability macroeconomic performance in the long run. This study is the first attempt to determine the relationship between the inflation and budget deficits for nine EU countries using different bootstrap causality tests. We employ the bootstrap causality and Granger causality test in the frequency domain analysis which allows us to distinguish short and long-run causality. We do not find a relationship between these variables when we employ bootstrap causality analysis. While the frequency domain causality shows that there is no relationship causality from budget deficits to inflation for all countries, causality from inflation to budget deficits indicates a permanent (long-run) relationship for Belgium, and France.

There is a growing body of the literature which examines the relationships between the budget deficits and inflation using different methods. Not only researchers but also policymakers take interest in this topic and it is also important for monetary policy and macroeconomics area. There is a wide range of theory about the inflation and budget deficits in the economic literature. In the light of theoretical discussions, the one of the well-organized works in this area was made by Sargent and Wallace [

According to Tanzi [

Catão and Terrones [

In the literature, as a result of government expenditure increase, expansionary fiscal policy appears and this result increases aggregate demand and goes up the inflation. This theory agrees with both classical and Keynesian economic school. Ricardian equivalence hypothesis, which is supported by classical economic thought, argues that budget deficits neither affect aggregate demand nor cause inflation due to cut in taxes. But Keynesian view supports the non-Ricardian regime in which tax cut goes up the aggregate demand and inflation. On the other hand, Ricardian regime shows that monetary authority has an “active behavior”; treasury has a “passive behaviour”. The fiscal theory of price level (FTPL) developed by Leeper [^{1}, Sims [

The contribution of our empirical study is quart. Firstly, unlike previous time series studies, we examine the relationship between the budget deficits and inflation for 9 countries as a time series. So, we can fulfill a need in this area. We can compare different results of these countries for causality. Secondly, we test both frequency domain causality and the Toda and Yamamoto [

There is growing literature about the fiscal deficits and inflation. In this area, there have been plenty of studies in relationship between fiscal deficits and inflation in this literature, still no consensus is reached. In the light of theoretical discussions, first works in this area were made by Sargent and Wallace [

There is also empirical relationship between budget deficits and inflation for both developed and developing countries. Empirical studies show statistically significant connection between fiscal deficits and inflation in economic literature. It has been growing literature in this area and most papers in these topics focus on the comprehensive examination of the relationship between fiscal deficits and inflation. The main purpose of this analysis can understand the determiners of the inflation. There are some studies for the US (Hamburger and Zwick [

There are too empirical studies in the relationship between budget deficits and inflation. Hamburger and Zwick [

King and Plosser [

Cottarelli, Griffiths and Moghadam [

The purpose of this paper is to test whether there is a relationship between budget deficits and inflation by using two methodological ways for nine countries for period from 1990 to 2013 using a quarterly data. These countries have been selected according to the data availability. The data is drawn from the Eurostat Database and IMF International Financial Statistics (IFS). Countries and time spans used in this analysis are shown in

Countries | Time span | Countries | Time Span |
---|---|---|---|

Belgium | 1991Q1-2013Q1 | Spain | 1995Q1-2013Q1 |

Cyprus | 1995Q1-2013Q1 | Sweden | 1993Q1-2013Q1 |

France | 1991Q1-2012Q4 | UK | 1990Q1-2013Q1 |

Germany | 1995Q1-2012Q4 | ||

Netherlands | 1991Q1-2013Q1 |

In the Granger-causality literature for time series data there are several tests developed such as Engle-Granger approach, Sim’s causality approach, VAR Granger-causality approach, and last but not least Toda and Yamamoto (TY, [

1) Determine the optimal order of the VAR process (p).

2) Determine the maximal order of integration occurring in the true generation process for variables.

3) Estimate the augmented VAR (p + dmax) model, where dmax is the maximal order of integration of variables determined in step 2.

4) Test Granger-causality by utilizing modified Wald (MWALD) test in an estimated augmented VAR model in step 3.

Say following is a augmented VAR (p + dmax) model to be estimated:

where the circumflex above a variable denotes its Ordinary Least Square (OLS) estimates and the d is the maximal order of integration of the variables.Following Hacker and Hatemi-J [

Using these notations, the estimated VAR (p + d) model will be as follows:

To test for the non-Granger causality the MWALD test developed by Toda and Yamamoto [

where

where

Granger [

Let

where,

Let

where

The measure of causality suggested by Geweke [

If _{ }at frequency

where

To test the hypothesis that BD does not cause Inf at frequency

within a bivariate framework. Breitung and Candelon [

The null hypothesis tested by Geweke,

where

The ordinary

This paper is to examine literature by extending the relationship between the budget deficits and inflation. The contribution of this study is to investigate in two methodological ways for nine countries for period from 1990 to 2013. At first, we will employ both the Toda-Yamamoto method based on asymptotic critical values and the bootstrap corrected causality test suggested by Hacker and Hatemi-J [

Before analyzing causality between the budget deficits and inflation, we have to determine maximum integration order of variables. So, the Toda-Yamamoto procedure needs the additional lag(s) in terms of determining the maximum integration degree of our variables. In order to ensure reliability for our variables, we use several unit root tests developed by PP and the KPPS test.

We will choose optimal lag order using Monte Carlo simulation in the VAR models. Herewith, we illustrate different information criteria which are Akaike Information Criteria (AIC), Hannan-Quinn Criteria (HQC), Schwarz Bayesian Information Criteria (SBC), Hatemi-J Criteria (HJC) in

It is important to determine the selection of lag length for correct results in our analysis. In our analysis, we

Countries | Budget Deficit | Inflation | ||
---|---|---|---|---|

PP | KPSS | PP | KPSS | |

Belgium | −3.70^{**} | 0.31^{***} | −3.89^{**} | 0.10 |

Cyprus | −4.51^{**} | 0.09 | −4.27^{***} | 0.03 |

France | −2.59 | 0.19^{**} | −3.47^{**} | 0.13^{*} |

Germany | −9.72^{***} | 0.09 | −3.38^{*} | 0.04 |

Netherlands | −1.84 | 0.24^{***} | −2.66 | 0.07 |

Spain | −2.06 | 0.19^{**} | −2.37 | 0.09 |

Sweden | −1.71 | 0.24^{***} | −2.52 | 0.07 |

UK | −2.19 | 0.19^{**} | −2.06 | 0.26^{***} |

Notes: The critical values for PP test are −4.07, −3.46 and 3.16 and the critical values for KPSS test are 0.22, 0.15 and 0.12 at 1%, 5% and 10% significance level, respectively. ^{***}, ^{**} and ^{*} show the rejection of the null hypothesis at 1%, 5% and 10% significance level, respectively.

Countries | AIC | SBC | HQC | HJC |
---|---|---|---|---|

Belgium | [ | [ | [ | [ |

Cyprus | [ | [ | [ | [ |

France | [ | [ | [ | [ |

Germany | [ | [ | [ | [ |

Netherlands | [ | [ | [ | [ |

Spain | [ | [ | [ | [ |

Sweden | [ | [ | [ | [ |

UK | [ | [ | [ | [ |

Notes: Akaike Information Criteria (AIC), Hannan-Quinn Criteria (HQC), Schwarz Bayesian Information Criteria (SBC), Hatemi-J Criteria (HJC). The numbers in brackets are the optimal lag lengths and minimum test statistics are in the parenthesis.

chose four different information criteria^{2}, namely Akaike Information Criteria (AIC), Hannan-Quinn Criteria (HQC), Schwarz Bayesian Information Criteria (SBC), Hatemi-J Criteria (HJC). It is shown two different lag length selected by HJC and HQC to test between our variables in

Different from linear causality analysis, this procedure allows examining causality between these variables in different frequencies. We find no evidence of causal relationship running from budget deficits to inflation for all countries in

Countries | H_{0}: BD does not Granger cause INF | H_{0}: INF does not Granger cause BD | ||||||||
---|---|---|---|---|---|---|---|---|---|---|

MWALD | TY Prob. | Bootstrap critical values | MWALD | TY Prob. | Bootstrap critical values | |||||

1% | 5% | 10% | 1% | 5% | 10% | |||||

Belgium | 0.183 | 0.913 | 9.674 | 6.083 | 4.673 | 0.243 | 0.886 | 10.343 | 6.373 | 4.801 |

Cyprus | 0.794 | 0.672 | 10.083 | 6.378 | 4.854 | 2.553 | 0.279 | 9.801 | 6.288 | 4.790 |

France | 0.566 | 0.753 | 10.414 | 6.361 | 4.776 | 0.182 | 0.913 | 9.304 | 6.121 | 4.592 |

Germany | 0.517 | 0.472 | 7.237 | 3.925 | 2.799 | 0.039 | 0.844 | 7.352 | 4.159 | 2.788 |

Netherlands | 0.907 | 0.341 | 7.891 | 3.927 | 2.649 | 0.029 | 0.865 | 7.903 | 3.865 | 2.637 |

Spain | 0.086 | 0.770 | 8.638 | 4.078 | 2.735 | 0.362 | 0.547 | 7.887 | 4.076 | 2.720 |

Sweden | 0.387 | 0.534 | 7.868 | 4.012 | 2.749 | 0.762 | 0.383 | 7.990 | 4.087 | 2.790 |

UK | 1.803 | 0.179 | 7.455 | 4.175 | 2.781 | 0.218 | 0.640 | 7.665 | 4.207 | 2.833 |

Notes: In this table, INF signifies inflation rate and BD denotes budget deficits. It is used lags based on Hatemi-J Criterion (HJC). The notations ^{***}, ^{**}, and ^{*} imply rejection of the null hypothesis at the 1%, 5% and 10% levels of significance, respectively. They based on the bootstrap critical values. TY probability is determined probability value by the Toda-Yamamoto procedure for the MWALD stat.

Countries | H_{0}: BD does not Granger cause INF | H_{0}: INF does not Granger cause BD | ||||||||
---|---|---|---|---|---|---|---|---|---|---|

MWALD | TY Prob. | Bootstrap critical values | MWALD | TY Prob. | Bootstrap critical values | |||||

1% | 5% | 10% | 1% | 5% | 10% | |||||

Belgium | 1.273 | 0.736 | 12.442 | 8.068 | 6.439 | 5.398 | 0.145 | 12.177 | 8.259 | 6.535 |

Cyprus | 0.794 | 0.672 | 10.083 | 6.378 | 4.854 | 2.553 | 0.279 | 9.801 | 6.288 | 4.790 |

France | 0.566 | 0.753 | 10.414 | 6.361 | 4.776 | 0.182 | 0.913 | 9.304 | 6.121 | 4.592 |

Germany | 0.517 | 0.472 | 7.237 | 3.925 | 2.799 | 0.039 | 0.844 | 7.352 | 4.159 | 2.788 |

Netherlands | 0.907 | 0.341 | 7.891 | 3.927 | 2.649 | 0.029 | 0.865 | 7.903 | 3.865 | 2.637 |

Spain | 0.086 | 0.770 | 8.638 | 4.078 | 2.735 | 0.362 | 0.547 | 7.887 | 4.076 | 2.720 |

Sweden | 0.387 | 0.534 | 7.868 | 4.012 | 2.749 | 0.762 | 0.383 | 7.990 | 4.087 | 2.790 |

UK | 1.328 | 0.515 | 10.595 | 6.409 | 4.951 | 0.477 | 0.788 | 10.383 | 6.346 | 4.775 |

Notes: In this table, INF signifies inflation rate and BD denotes budget deficits. It is used lags based on Hannan-Quinn Criteria (HQC). The notations ^{***}, ^{**}, and ^{*} imply rejection of the null hypothesis at the 1%, 5% and 10% levels of significance, respectively. They based on the bootstrap critical values. TY probability is determined probability value by the Toda-Yamamoto procedure for the MWALD stat.

BD to INF causality | |||||||
---|---|---|---|---|---|---|---|

Countries | Long term | Medium term | Short term | ||||

0.01 | 0.05 | 1.00 | 1.50 | 2.00 | 2.50 | ||

Belgium | 0.869 | 0.869 | 0.741 | 0.432 | 0.327 | 0.301 | |

Cyprus | 0.876 | 0.878 | 1.229 | 0.826 | 0.411 | 0.251 | |

France | 0.322 | 0.323 | 0.777 | 0.642 | 0.538 | 0.491 | |

Germany | 0.493 | 0.492 | 0.130 | 0.171 | 0.504 | 0.748 | |

Netherlands | 0.074 | 0.077 | 2.021 | 2.034 | 1.990 | 1.965 | |

Spain | 1.135 | 1.144 | 5.849 | 5.511 | 5.025 | 4.788 | |

Sweden | 1.099 | 1.097 | 0.497 | 0.589 | 0.638 | 0.659 | |

UK | 2.356 | 2.361 | 2.512 | 1.957 | 1.749 | 1.672 |

Notes: The lag lengths for the VAR models are determined by SIC. F-distribution with (2, T-2p) degrees of freedom equal 5.99. BD signifies budget deficits and INF denotes inflation rate. ^{*}Indicates the relationship of causality between these variables.

INF to BUD causality | |||||||
---|---|---|---|---|---|---|---|

Countries | Long term | Medium term | Short term | ||||

0.01 | 0.05 | 1.00 | 1.50 | 2.00 | 2.50 | ||

Belgium | 7.062^{*} | 7.068^{*} | 2.153 | 0.853 | 0.977 | 1.098 | |

Cyprus | 1.001 | 1.005 | 3.291 | 3.142 | 2.323 | 1.965 | |

France | 7.072^{*} | 7.074^{*} | 3.004 | 2.180 | 2.258 | 2.341 | |

Germany | 1.505 | 1.508 | 3.416 | 3.473 | 3.193 | 3.048 | |

Netherlands | 1.412 | 1.411 | 4.135 | 4.796 | 4.791 | 4.752 | |

Spain | 1.135 | 1.144 | 5.849 | 5.511 | 5.025 | 4.788 | |

Sweden | 3.524 | 3.512 | 0.142 | 1.127 | 1.733 | 1.996 | |

UK | 4.356 | 4.216 | 0.059 | 0.088 | 0.209 | 0.272 |

Notes: The lag lengths for the VAR models are determined by SIC. F-distribution with (2, T-2p) degrees of freedom equal 5.99. BUD signifies current account deficits and INF denotes inflation rate. ^{*}Indicates the relationship of causality between these variables.

This paper is the first attempt to determine the relationship between the inflation and budget deficits for nine EU countries using different bootstrap causality and frequency domain causality tests. Empirical test results using bootstrap causality indicate that there is no causality between budget deficits and inflation in general. Further, frequency domain Granger causality analysis provides the evidence that the null hypothesis of no Granger causality running from budget deficit to inflation is accepted at 1%, 5% and 10% significance levels across all the frequencies. On the other hand, we find significant evidence of a long-term Granger causal relationship where results show unidirectional causality from inflation to budget deficits in Belgium, and France. We do not find for the relationship running from inflation to budget deficits in medium and short term. Finally, this paper can be extended by investigating the effect of public deficit problems on inflation and other macro-economic variables and may offer new insights for the problems.