Empirical Studies in Social Sciences 6th International Student Conference,
Izmir University of Economics, Izmir Turkey
A PANEL COINTEGRATION ANALYSIS OF BUDGET DEFICIT AND
INFLATION FOR EU COUNTRIES AND TURKEY
Marmara University/ Department of Economics
Marmara University Goztepe Campus, 34722, Kadiköy, Istanbul
0090 538 3950143
Marmara University/ Department of Economics
Marmara University Goztepe Campus, 34722, Kadiköy, Istanbul
0090 554 8881718
This paper aims to analyze the empirical relationship among budget deficit and inflation for Turkey and European countries. According to theory, fiscal imbalances result in inflation problem as 1990s experience of Turkey has shown. The major outcome from the empirical studies indicated strong evidence that a budget deficit financed through monetarisation and a rising money supply could lead to inflation. The inflationary effect of budget deficits depends on the means by which the deficit is financed and the impact of that on aggregate demand. In this paper, budget deficits and inflation relationships are studied by utilizing the Larsson et. al. test
approach for Turkey and other sixteen European countries, including Czech Republic, Hungary, Poland, Austria, Belgium, Greece, Denmark, France, Germany, Italy, Holland, Norway, Slovakia, Spain, Sweden and England over the period 1990-2008, annually. Apart from the traditional studies, this paper investigates the relationship by using panel data cointegration analysis. Firstly, LLC test, IPS test and Hadri test are employed to test the presence of unit roots among series. After that, cointegration analysis is done. Our results will point out the fact that budget deficit and inflation has long run positive relationship in some of the countries, while it has negative relation in other countries. At the end of this analysis, we will try to understand if there is difference tendency towards to budget deficit-inflation relation among developing and developed countries.
Key Words: Budget Deficit, Inflation, Panel Data, Cointegration Analysis
JEL Classification: C23, E31, H62
The impact of fiscal policies on inflation has been widely discussed since 1990s. After the studies done for closed economies in 1990s, in 2000s the fiscal theory of price level was being debated within monetary unity of European countries framework. In the discussion of this monetary system’s sustainability, the public finance phenomenon has gained importance. In the Maastricht criteria (Maastricht treaty, 1992), while defining the basic conditions to enter euro area for the new entries, the finance of fiscal deficits is prioritized by European countries. Therefore, one can argue that the FTPL has been dominating the literature recently.
It is argued that the monetization of deficits is the fundamental reason for the high inflation problems in developing countries. Turkey, as a developing country, experienced extremely high inflations during 1990s. The underlying reasons for such a tremendous amount of inflation can be miscellaneous. However, to large extent, economists concluded that the main cause of inflation is high budget problems. The restriction of government to central bank resources which is implemented with the liberalization program of 1980s, made governing authorities focus on domestic debt financing. That is, the government avoided compensating the deficits through corresponding money supply. Meanwhile, the domestic banks became the major source of domestic debt financing that resulted in an increase in the assets of banking system. Thus, the interrelationship of budget deficit and inflation is a vital point for Turkish economy, despite the fact that the direction of relationship is uncertain.
The basic purpose of this paper is to analyze this relationship between price level and fiscal imbalances. The paper unfolds firstly by giving a theoretical framework for that relationship. In the following part, a brief literature review is provided. In the third part, empirical evidence from EU countries and Turkey is stated by using panel cointegration analysis. In the last part, the results of empirical study and its compliance with the theory is discussed.
2. Theoretical Framework
2.1. Budget Deficit and Inflation Relationship in General
The term of budget deficit can be defined as the difference between budget revenue and budget expenditure. Budget revenue includes three important components which are tax revenue , tax-exempt revenues and private revenues. The most important component of the budget revenue is tax revenue. However, budget expenditure involves four important elements that are current expenditure, investment expenditure, real expenditure and transfer payments. Current expenditure is a kind of expenditure which is related to nondurable goods. It is usually used for short term expenses. Investment expenditure is stated as expenses related to investment and efficient use of resources. Transfer payment is an unrequited payment that has an indirect effect on GDP. Real expenditure consists of production factors and production expenditure. If budget deficit shows the disharmony and imbalance between revenue and expenditure, both the revenue and expenditure side of budget should be analyzed in detail.
It is notable about budget deficits that there is a significant difference between developed and developing countries about budget balances. Most developed countries don’t have budget deficit problem because of their strong fiscal structure. In developed countries, low level of foreign indebtment prevents the debt payment to be burden on the budget. Moreover, most developed countries have a trade surplus due to having more export than import. On the other hand, developing countries usually have high inflation, lower per capita income compared to developed countries, high current account deficit and high public expenses. All these progresses cause increases in budget deficit and deterioration of macroeconomic stability. Developing countries have four different ways to finance their high budget deficit which are printing money, running down foreign exchange reserves, borrowing from abroad and domestic markets.
As stated above, developing countries have more budget deficit problems compared to developed countries. The reasons of budget deficit can be stated as unstable public revenue, low degree of economic development, low acceleration of public revenue, deficient government auditing and the regulatory role of government in the economy. Countries which have low degree of economic development, have high level of budget deficit owing to three important reasons which are high spending pressure, deficient tax revenue and low private savings. High employment cost is very crucial problem of public economy in developing countries and these governments don’t have any
chance to reduce it. Also, deficient public revenue leads to increase in budget deficit. In developing countries, private saving level is so low and deficits are financed by borrowing which cause to borrowing-interest spiral by increasing budget deficit more.
In general, inflation has raising effect on budget deficit by raising nominal interest rate. According to Fischer Effect; nominal interest rate consist of real interest rate and expected inflation rate. If the inflation expectation increases, it causes to rising nominal interest rate which leads to the public debt to go up. Interest payment covers the big part of public payment in developing countries. If interest rate increases because of inflation, it leads to raise interest payment as well as budget deficit by causing the Debt/ GDP ratio to increase. Thus, high interest rate and interest payment lead to instability between budget and public deficit acceleration and tax revenue acceleration. Budget and public deficit always increase faster than public revenue so budget deficit increase as well.
In spite of the positive relationship between inflation and budget deficit as stated above, in some cases inflation and budget deficit move in reverse direction. Inflation tax is important for this. If inflation tax is higher than normal level, as inflation increase people avoid holding money because the cost of holding money is high. Thus, real monetary base tends to decrease as inflation tax correspondingly. Holding money would be a costly activity. Inflation tax would be a type of tax revenue which makes the budget deficit decline. Another type of negative relation between inflation and budget deficit occurs because of public borrowing stocks. If borrowing is not indexed to the inflation, as the inflation rise the real value of public borrowing stocks would decline. As the public borrowing stock fall, budget deficit is expected to decrease.
2.2. Budget Deficit and Inflation Relationship in Different Economic Thoughts
There is a considerable debate on budget deficits and its inflationary effects in economic theory literature. In the preceding period of Keynes, the classical economists gave importance to a balanced budget, yet they didn’t analyze its impact on price levels. Apart from classical economics, Keynes saw the fiscal imbalances and budget deficits as internal components of aggregate national demand.(Corsetti and Roubini,1997:27) The underlying reason is that when budget expenditures increase, aggregate demand curve responds it by shifting right, leading to an increase in both prices and production.(Assuming aggregate supply is not perfectly elastic/inelastic) The increasing nominal income will come up with rising transactional demand for money, which is compensated by speculative demand for money, i.e. increasing real interest rates.(Anusic,1991) In the Keynesian economic thought, the budget deficits can be tolerable in the crisis times. Moreover, Keynes saw the budget deficits as an indicator of the impact of fiscal policy on aggregate demand. Thereby, due to the fact that the budget deficit can affect economic performance, it has been perceived as an endogenous factor. (Blanchard and Fischer, 1989). As a result, in Keynesian theory, because the main aim of the governments is to sustain high overall economic performance in the long run, the budget deficits can be acceptable to some degree. (Altıntaş, et.al.,2008)
In the neoclassic theory, the debate of Sargent and Wallace enlightens the discussion on the relationship among fiscal imbalances and inflation. Sargent and Wallace discuss two types of the coordination between monetary and fiscal authorities which is effective in controlling the inflation. In the first type of coordination in which the monetary authority is dominant, monetary authorities announce the monetary base growth and fiscal policy sets its budget by considering the revenue created by monetary policy. In the second type of coordination, in which the fiscal authorities are dominant, fiscal policy sets its budget and announces the amount of money needed for monetary authorities through seignorage and bond sales. (Sargent and Wallace, 1981) The second type of coordination provides insight to inflation problem which is led by fiscal imbalances, since the fiscal authorities sometimes demand more revenue than tolerable amounts and this creates inflation. This argument has been debated widely in the literature. The ‘fiscal view of inflation has been especially
prominent in developing country literature which has long recognized that less efficient tax collection, political instability and more limited access to external borrowing tend to lower the relative cost of seignorage and increase dependence on inflation tax.’ (Catao and Torrentes, 2003) Thus in the neoclassic theory the effect of fiscal theory is significant especially in developing countries.
In the neo classical approach, increasing budget deficit, which is compensated by borrowing instead of taxes, results in incrementing private sector wealth, consumption and aggregate demand, in turn. Nevertheless, this rising wealth is accompanied with a misperception by private sector about which the budget deficit will be paid by taxes in the future. Buiter (1983) argues that ‘if deficits are financed by printing money, this will fuel inflation. If they are financed by borrowing this will put upward pressure on interest rates, leading to "crowding out" of interest sensitive spending.’ As this kind of financing arises the real interest rates, the neoclassical theory suggests that increasing budget deficit can lead crowding out of investment and capital.
The new classical economists oppose the misperception part of the theory and asserted that such an assumption is inconsistent with rational expectation theory. That is, the demand for goods is based on expected present value of the future taxes. Fiscal policy can influence the price level through aggregate demand changes; it should change the expected value of the future taxes, which occurs by altering the spending. ‘In this sense, budget deficits and taxation have equivalent effects on the economy-hence the term, "Ricardian equivalence theorem."’(Barro, 1989) i.e. there is no change in national saving, since an increase in private saving as faced by an equivalent decline in public saving. Because national savings, in turn, investment and aggregate demand do not change, one can argue that the budget deficit doesn’t affect price levels.
As we can see from miscellaneous economists from different economic schools, the financing of fiscal deficits has a key role in inflationary effects of them. To this end, the type of deficit is can be either bond-financing or monetization. In case of monetization, as monetarist approach puts forward, the price levels are directly affected. In addition, if the deficits are financed by borrowing, i.e. selling bonds, then the interest rates must be lower than the monetary base growth to prevent the unexpected
inflationary effects. Thus, one can assert that budget deficits are an important policy tool to be taken into account in inflation targeting policies.
3. Literature Review
The relationship between budget deficit and inflation is a very common debate in economic literature. Lots of economists have analyzed the relationship among these two variables for years by using different countries, different econometric technique and different time period. While some economists found negative relation, most of the economists found positive relationship between budget deficit and inflation. Some of these studies are ordered below with their critical results.
Fischer (1989) analyzed the budget deficit-inflation relationship in different countries . He found that the countries with high inflation have strong relationship among inflation and budget deficit. Fischer noted that high inflation has reducing effect on tax revenue which is known as Tanzi-Olivera Effect. Also, high rate of inflation increases budget deficit by declining seignorage revenue.
Hondroyiannis and Papapetrou (1997) studied on the direct and indirect effect of budget deficit on inflation in Greece and found the result that budget deficit has an indirect raising effect on inflation. However, they also stated that an increase in inflation results in an increase in budget deficit.
Cardoso(1998) worked on the relationship between budget deficit and inflation in Brazil by following Patinkin’s(1993) study. He found reverse relationship between budget deficit and inflation in Brazil.
Kıvılcım(1998) analyzed the long run relationship among budget deficit and inflation in Turkish Economy between the years 1950-1987. At the end of his study, he concluded that a change in budget deficit cause to change in inflation on the same direction. He also highlighted that this budget deficit-inflation spiral is one of the most important problems of Turkish Economy.
Tanzi (2000) researched the tax revenue and budget deficit relation in Latin American countries. He emphasized that even though the tax revenue rises, the budget deficit and public deficit also increase. He stated that this imbalance results from the deficient and inefficient social programs.
Egeli(1999) studied relations among inflation tax, budget deficit and public spending. His result was reverse relation among inflation tax and budget deficit. He also stated that increasing public spending leads to increase in budget deficit. Egeli concluded that this disequilibrium results from governments’ wrong policies such as using borrowing in order to finance the deficit.
Şen (2003) analyzed the relations among tax revenue and inflation. Şen stated that high inflation cause to decrement in tax revenue in crisis time. Low level of tax revenue cause to tax loss which leads to high budget deficit. He also interrogated the time of tax collection. He concluded that short term tax collection is better than long term tax collection. In the long run the real value of tax revenue tends to decline because of high inflation.
Catao and Terrones (2000) worked on the relationship between inflation and budget deficit by using data from different countries. The result that they reached is being weak relationship in developed countries and being strong positive relationship in developing countries.
Yabal, Baldemir and Bakımlı (2004) made a study about imbalance between public spending and public revenue in Turkey. They highlighted that the government finances budget deficit by using short term advance money. It also results in the money supply to increase which results in inflation to go up. They concluded that high budget deficit leads high inflation in Turkish Economy.
Solomon and Wet (2004) made a research on Tanzania. They found a strong positive
relationship between inflation and budget deficit. They stated that budget deficit has a significant effect on inflation. They also concluded that developing countries should give more importance to inflation because inflation tends to be affected from many economic shocks such as high budget deficit. According to them inflation should be controlled by efficient fiscal policies.
4. Econometric Methodology
After considering relevant theory and reviewing literature, it is time to give empirical evidence to test the consistency of the theories with the real world. To this end, firstly the series should be taken logarithms and differenced to avoid spurious regression problem that means that ‘regressing a non-stationary variable on a deterministic trend does not yield a stationary variable.’(Harris, 1995:20).
4.1. Panel Unit Root Test
As the involvement of macroeconomic applications in the panel data analyses has been growing recently, the Dickey-Fuller and Augmented Dickey-Fuller tests are required to be extended for testing stationarity in panel data analysis. When dealing with panel data, because the procedure is more complex, the ADF and DF tests can result in inconsistent estimators. Thus, the stationarity of the series should be tested by using three different types of tests, namely LLC (Levin,et.al.,2002), IPS (Im,et.al., 2003) and Hadri(2000).
In the analysis, firstly the LLC test is employed to test the stationarity. Levin et. al. model
allows heterogeneity of individual deterministic effects and heterogeneous serial correlation structure of the error terms assuming homogeneous first order autoregressive parameters.(Barbieri,2004) In addition the model provides two-way fixed effects, one of which comes from the term α and the other i
one emanates from δMoreover, these two parameters allow for heterogeneity, as the coefficient of t. .
lagged Y is limited to be homogenous through all individual units of the panel. i
LLC model tests the hypothesis of non-stationarity, i.e. the presence of unit roots. That is,
H ?;：;?：;?，，，，，：;?：;?：;！ 0:，(;；;~;;
H?;：;?：;?，，，，，：;?：;?;，;！; 1: ，(;；;~;;
There are two major shortcomings of the LLC test. Firstly, it relies on the assumption of the independence across units of panel where a cross sectional correlation may be present.(Barbieri,2004) Secondly and more importantly, Autoregressive parameters are considered to be identical across the panel in this model.
Im, Persan and Shin (2003) broadened the LLC test to overcome the second limitation of it by presenting a more flexible and computationally simple test structure that permits the ?;?; to differ
among individuals, i.e. by allowing heterogeneity. The IPS test made the estimation for each of the i section possible. As a result their model is such that;
Im et. al. tests the null of non-stationarity. That is;
H ?= 0 for all i 0:i
H;;?< 0 for i= 1, 2,…..,N1: i 1
;;;;;;;;?= 0 for i= N+ 1,…,N i 1
This alternative test clarifies that a fraction of the panel can have unit roots. This is the contrasting point of IPS to LLC. The IPS model is constructed under the restrictive assumption that T should be the same across individuals. That is to say, there should be a t-bar statistic which is the
mean of ADF t-statistics for testing ?= 0 for all i such that i