• No results found

The Impact of Fiscal Policy on Inflation : A panel data analysis on government spending and the price level

N/A
N/A
Protected

Academic year: 2021

Share "The Impact of Fiscal Policy on Inflation : A panel data analysis on government spending and the price level"

Copied!
45
0
0

Loading.... (view fulltext now)

Full text

(1)

The Impact of Fiscal

Policy on Inflation

BACHELOR

THESIS WITHIN: Economics NUMBER OF CREDITS: 15 PROGRAMME OF STUDY:

International Economics &

A panel data analysis on government spending and the

price level

(2)

Acknowledgements

Firstly, we would like to thank our tutor Rafael Barros De Rezende for giving us helpful comments on our work and guiding us throughout the whole process. We are greatly thankful for your support and knowledge within the area of research.

Further, we thank all professors within the various courses of macroeconomics and

econometrics we have taken at JIBS, who have provided us with the knowledge needed to conduct this thesis.

Finally, we would like to give our thanks to Emma Lappi who has coordinated this course well, provided valuable information, and been available for questions all through the course.

Thank you.

Axel Nåbo & Oscar Wahlgren

Jönköping International Business School June 2021

(3)

Bachelor Thesis in Economics

Title: The Impact of Fiscal Policy on Inflation Authors: Axel Nåbo & Oscar Wahlgren

Tutor: Rafael Barros De Rezende Date: 2020-06-12

Key terms: Inflation, Fiscal policy, Panel data, Money supply, Monetary Policy, Primary cash balance

Abstract

The issue of consumer price levels and the time value of money has for the longest time been at the forefront of monetary and fiscal policy decisions. The following study investigates to what effect, if any, the central government’s primary cash balance changes the level of inflation. After a year of a being struck by a pandemic the world is slowly recovering, and thus a sound fiscal policy is useful to repair the worldwide economy. Using historical data and basing our study mainly on the Keynesian Economic Theory and the Fiscal Theory of the Price Level, we used a panel least squares econometric method to conclude the effects previously mentioned. Our findings were that a primary cash deficit along with a broad

money supply increase significantly increase inflation. When implementing dummy variables, we also found the effect of a primary deficit is most significant in lower-middle to

low-income countries, and countries in Sub-Saharan Africa, Latin America, and the Caribbean. These findings may aid governments on their path forward to economic stability and responsible fiscal policies, a prominent issue as the coronavirus pandemic is coming to an end.

(4)

Table of Contents

Table of Contents ... iii

List of Tables ... v

List of Graphs ... v

List of Models ... v

1 Introduction ... 1

2 Theory and Literature Review ... 5

2.1 Ricardian Equivalence ... 5

2.2 Keynesian Economic Theory ... 6

2.2.1 Keynesian Multiplier ... 6

2.3 Fiscal Theory of the Price Level ... 7

2.4 The Quantity Theory of Money ... 8

2.5 Previous Research ... 8

3 Data ... 12

3.1 Variables included ... 12

3.1.1 Inflation, Consumer Prices (annual %) ... 12

3.1.2 Primary Cash Balance (PCB) as Percentage of GDP ... 12

3.1.3 Broad Money Growth (M2) ... 12

3.1.4 World Region ... 13

3.1.5 Income Group ... 13

3.1.6 Distinction between Regions and Income Groups ... 14

4 Methodology ... 17

4.1 Panel Analysis ... 17

4.2 Why Panel Data? ... 17

4.3 Panel Least Squares Model ... 17

4.4 The Least Squares Dummy Variable (LSDV) Model ... 18

4.4.1 Dummy Variable Trap ... 18

4.5 The Interaction Between Fiscal and Monetary Policy ... 18

4.6 Gram-Schmidt Orthogonalization ... 19 4.7 Interaction Term ... 20 4.8 Hypothesis ... 20

5 Results ... 21

5.1 Descriptive Statistics ... 21 5.2 Regression Analysis ... 21 5.3 Basic Model ... 22 5.4 Generalized Models ... 24 5.5 Robustness Check... 25 5.6 Discussion ... 26 5.7 Limitations ... 29

6 Conclusion ... 30

Reference list ... 32

(5)

Appendix A ... 37 Appendix B ... 39

(6)

List of Tables

Table 1. Descriptive Statistics ... 21 Table 2. The impact of primary cash balance on inflation in 135 countries during

the period 1993-2019 ... 22 Table 3. The impact of primary cash balance and broad money growth on inflation

in 121 countries during the period 1993-2019 ... 23 Table 4. The impact of primary cash balance on inflation in different regions

during the period 1993-2019 ... 24 Table 5. The impact of primary cash balance on inflation in different income

groups during the period 1993-2019... 25

List of Graphs

Graph 1. Inflation, consumer prices (annual %) – High income countries .... 14 Graph 2. Inflation, consumer prices (annual %) – Low-income countries .... 15 Graph 3. Inflation, consumer prices (annual %) – Europe & Central Asia .... 16 Graph 4. Inflation, consumer prices (annual %) – Sub-Saharan Africa ... 16

List of Models

Model 1. PCB as explanatory variable ... 18 Model 2. PCB and M2 growth as explanatory variables ... 18 Model 3. PCB and M2 growth as explanatory variables with region dummy

variables ... 18 Model 4. PCB and M2 growth as explanatory variables with income group dummy

variables ... 18

(7)

1 Introduction

The question of inflation and how to control it is an ever-relevant issue in society. Economists around the world have for a long time been trying to affect the inflation through various economic tools, both successfully and unsuccessfully. The perception that national macroeconomic developments have a connection with international

conditions is not a new idea. Ciccarelli and Mojon (2005) mean that a few determinants of inflation are global. On the other hand, it is not possible to overlook the fact that inflation also depends on national conditions and policies. The most well-known theories regarding inflation consists of the Monetarist view and Keynesian economics. In the matter of Keynesian economics, government intervention, i.e., fiscal policy, is normally seen as the solution to control inflation which goes in line with this report. An important part of fiscal policy is government spending. Historically, the structure of governments and government spending has amended globally. In the last century, government spending especially in early industrialized nations has expanded strikingly (Mauro et al., 2015). Furthermore, Nguyen (2019) means that there are three essential approaches that explain the principles and dynamics of inflation which comprises the approaches of fiscal, monetary, and public finance. As mentioned earlier, one segment in the Fiscal policy approach is government spending which consists of different parts such as welfare benefits, pension spending and infrastructure investment. General government spending indicates the magnitude of governments across nations. The great variation in this indicator emphasizes the variety of nations' approaches to deliver public goods and services and supply social protection. This does not necessarily highlight differences in resources spent by governments (OECD, 2021). However, this report does not focus on where government funds are prioritized. On the other hand, we find it alluring to examine whether fiscal policy is a relevant tool for controlling inflation, as well as its impact on it.

Regarding government spending there is a great heterogeneity between world regions. Central governments in countries with high income, especially European countries, tend to authorize a much greater share of national production than governments in countries with low income. In countries similar to France, central government spending stands for

(8)

analogous figure is close to 6% (Ortiz-Ospina & Roser, 2016). Because of this it is of great interest to look at data from several different regions and countries to examine patterns and similarities.

Looking at the monetary approach, Friedman (1963) means that inflation is and has always been a monetary event, and Fischer et al. (2002) argued that inflation is highly correlated with growth rates of money supply in both the short run and the long run. From a perspective of fiscal policy, the relation between fiscal deficit and inflation has been documented in both theoretical and empirical research. In a theoretical discussion by Sargent and Wallace (1981) the authors found that fiscal deficit can affect inflation indirectly via money creation. Empirical evidence is provided by Catão and Terrones (2005) that supports the hypotheses of positive correlation between fiscal deficit and inflation. Nevertheless, the Fiscal theory of the price level developed by Leeper (1991), Woodford (1994), and Sims (1994), states that fiscal authority alone can predominantly affect inflation, irrespective of monetary policy.

Moreover, Fiscal policy is used to fulfill macroeconomic targets such as employment growth and rise in output. These goals are achieved by either an increase in government spending or by a tax raise, or an adopted mixture of the two tools. The government spending is largely financed by loans from lender institutions (Hussain & Zafar, 2018). This could impede the economic growth. Furthermore, borrowing from the Central Bank improves the money supply, causing the inflation and raising the uncertainty in the economy (Landau, 1985). Thus, fiscal policy could be used for several reasons and targets as mentioned earlier, but the focus on this report will not lie on sustainable government prioritizations. The aim for this report is to examine the impact of fiscal policy implications, more specifically primary cash balances, on inflation. We must stress, that while the term primary cash balance (PCB) will be used frequently

throughout this report as a metric of government spending, these two concepts are not one and the same and shall not be mistaken as such. Rather, government spending is a vital component of PCB, where it is set in relation to government revenues, such as tax income and borrowed money. Why use PCB then? Because government spending as share of GDP tells us nothing of the relationship between the governments spending and revenues, thus PCB painting a more accurate depiction on a government’s budget

(9)

balance. Previous studies have proved a relationship between inflation and fiscal deficits (Catão & Terrones, 2005; Lin & Chu, 2013), which stresses that primary cash balance is the most proper indicator for our purpose.

One of the most fundamental goals of most central banks is to maintain price stability, which is often denoted as low inflation. The inflation target lies in the interest of the government as well as it affects the country, both firms and consumers. One question that needs to be answered is then, why is it important to have a low inflation? There are several answers that fits in to the question such as that high inflation pits the value of the wages. An increase in prices leads to less consumption as we can buy less on a given nominal wage. Thus, everyone becomes poorer, and a wage increase must follow in a similar rate as the price increase to keep the real wage unchanged. There are several macroeconomic variables that affects an economy both at the country level, and

globally. Since inflation is prominent in macroeconomic discussions, it is of great interest to know what, if and how different variables affect it. Normally inflation is seen as a monetary matter. However, our interest lies in the area of fiscal policy, more specifically primary cash balances and government spending as determinants of inflation. In contrast to the monetarist view on inflation where the relation between money supply and growth of national income determines inflation, we believe it is important to investigate other variables that may affect inflation. Governments and tax rates are different across the globe; thus, it is of importance to examine data of different world regions where both national income and wage rates differ.

The purpose of this thesis is to examine whether primary cash balance as percentage of GDP has had any effect on the inflation rates of countries across the globe during the last 27 years, 2020 excluded. We are also keen to examine if there is any observable difference in these effects, if any can be found, between different world regions as well as between countries of different income groups. There is arguably a social interest at this time for this question to be asked. Considering the coronavirus pandemic that has plagued the world for the past year, many states have considered different approaches in fiscal policy, with for example the US having voted in favor for stimulus checks for its

(10)

To get a statistically reliable result for our research question, data is retrieved from countries all over the world, including different regions and income groups. Both recent and historical data with a time span of 27 years (1993-2019) is used. The independent variable in the regression is inflation, while primary cash balance is the explanatory variable. Since it is known that fiscal and monetary policy are connected, a control variable for broad money growth is included in the regression. Variables of world region and income group will be implemented as different dummy variables, this to get a clearer view of differences with respect to these criteria.

Panel data analysis is the method used to examining patterns and obtain the results. We consider this as a relevant method in this research since it is used for investigating the behavior of individual subjects over a continuous period. The subject in this research is countries. Furthermore, panel data presents more degrees of freedom and further variability when combining time-series with cross-sectional data which is adequate when studying the dynamics of change, which is the case in this research. Another advantage of panel data over time-series or cross-sectional data is more accurate conclusion of model parameters (Hsiao, 2007). To conduct the full regression, the ordinary least squares model of panel data is implied. The dummy variables are

implemented through the least squares dummy variable model. Through these steps, we arrived at the conclusion that primary cash balance has a significant impact on a

country’s inflation, in the sense that a surplus leads to a lower inflation rate. This impact is particularly evident in lower-middle to low-income countries and in Sub-Saharan Africa, Latin America, and the Caribbean.

(11)

2 Theory and Literature Review

2.1 Ricardian Equivalence

The theory of Ricardian equivalence was developed by the British economist David Ricardo in the nineteenth century. The theory briefly states that government expenditure needs to be financed by taxes in one way or another. It is not of importance if the taxes are appointed now or in the future. In the theory people are seen as rational consumers, and thus have rational expectations about the future. This implies that pursuits to invigorate the economy by increases in debt-financed government expenditure will not be sufficient since consumers and investors will figure out that the debt will sooner or later have to be paid in form of future taxes. Therefore, a change in government budget may affect private saving (Gottfries, 2013). For example, if the government runs a budget deficit, people may save more now because they expect to pay higher taxes in the future to finance the budget deficit. The same logic works for the opposite, if the government runs a budget surplus people are not as keen saving money now because of expectations of lower taxes in the future.

Ricardian equivalence holds true if any increase in government spending that inflates the budget deficit leads to analogous decrease in consumption expenditure. This is because households save more with respect to their expectations of their future tax liabilities. The result is zero effect on aggregate demand and fiscal policy is completely ineffective. Furthermore, the amount of government debt does not matter in a reality with complete Ricardian equivalence with non-distortionary taxes (Gottfries, 2013). Barro (1989) argues that there are five major objections to Ricardian equivalence:

• People do not live forever. Hence, they do not care about taxes that come into force after their death.

• Private capital markets are imperfect.

• Taxes and incomes in the future are uncertain.

• Taxes are not a “lump sum” since they depend on factors such as income and wealth.

(12)

Moreover, it is necessary to not only know whether the Ricardian view remains intact, but to see what substitute conclusions arise (Barro, 1989).

2.2 Keynesian Economic Theory

The Keynesian Economic Theory was established in 1936 by English economist John Maynard Keynes with his publication of the text “The General Theory of Employment, Interest, and Money” (Keynes, 1936). In a Great Depression backdrop, Keynes argued that to boost demand for goods and services, government spending would need to increase. This was contrary to the classical laissez-fair policy of economics with limited government interference. The Keynesian theory contains three components:

• The decisions of the private and public sectors influence aggregate demand, such as if a decrease in aggregate spending causes a recession, the government may intervene with a fiscal or monetary stimulus.

• Prices usually respond slowly to changes in demand and supply, causing surplus or shortage in labor supply.

• The greatest impact on output and employment of the economy is caused by changes in aggregate demand. Consumer and government spending,

investments, and exports all increase the level of output.

2.2.1 Keynesian Multiplier

The Keynesian Multiplier suggests that an increase in private consumption expenditure, investment expenditure, and net government spending increases the gross domestic product at a proportionally larger rate than the increase of spending. As such, the output of the economy is a multiple of the increase or decrease in spending. The Keynesian Multiplier is the reciprocal of the propensities to withdraw from the circular flow of income (Rutherford. 2012). There are two factors determining the value of the multiplier:

• Marginal Propensity to Save (MPS): The MPS, for an increase in an individual’s income, measures the proportion of income the individual saves rather than spends on goods and services, such that MPS = ΔS/ΔY.

• Marginal Propensity to Consume (MPC): The MPC measures the change in total consumption with a change in income, such that MPC = ΔC/ΔY.

(13)

The MPC is then used to calculate the value of the Keynesian Multiplier with the formula: 1/(1-MPC), where the value of the multiplier is weighted to one unit of new income (Corporate Finance Institute, n.d.b).

2.3 Fiscal Theory of the Price Level

An important goal in public policy is to maintain price stability. According to

Christiano and Fitzgerald (2000), two key questions must be addressed to reach the goal of price stability:

• How can price stability be achieved? • How much price stability is desired?

Furthermore, Christiano and Fitzgerald (2000) means that the normal monetarist view to answer the first question suggests that central banks have an adamant commitment to price stability. An alternative view is that an independent central bank is not sufficient to assure price stability. With respect to this view, price stability does not only require an appropriate monetary policy, but an appropriate fiscal policy as well. Because of the focus on fiscal policy in this point of view regarding the determinants of the price level, the theory has become to be known as the Fiscal Theory of the price level (FTPL). FTPL states that the price level is resolved by the necessity to establish fiscal solvency, hence it is set in the way that the market value of outstanding debt is equal to the

expected present value primary surpluses in the future (Fan et al., 2016). Furthermore, it describes fiscal and monetary policy decrees such as that the price level is definite by government debt and fiscal policy alone. In this, monetary policy plays possibly, but not necessarily an indirect role. The FTPL collides with the monetarist view which points out that money supply is the primal determinant of the price level and inflation (Bassetto, 2008). Empirical evidence for this theory has been found by various

researchers. Tanner and Ramos (2003) argues that FTPL provides a cogent explanation for Brazil´s high inflation in the late twentieth century. The American economist John H. Cochrane (1998; 2001; 2005) has argued several times that there is a simple

identification problem that affects the FTPL. He means that in the FTPL, fiscal policy is exogenous and compels inflation to produce fiscal solvency.

(14)

2.4 The Quantity Theory of Money

The Quantity Theory of Money is about the relationship between the price level and the stock of money. The model asserts that there is a stable and anticipated relationship among the amount of money that circulates in an economy and the price level (Moles & Terry, 1997). The theory is occasionally described as postulating strict proportionality between the money supply and the price level (aggregate) and brings the corresponding neutrality of money (Gordon & Leeper, 2002). The equation is usually written as 𝑴𝑽 = 𝑷𝑻, where M equals the quantity of money, V equals the velocity of circulation, P equals the price level, and T equals the number of transactions in the period. The theory was first proposed by the English philosopher David Hume, and later on, Milton

Friedman made the equation the central fulcrum of monetarism, with an additional assumption that the velocity of money (V) is predictable. Therefore, for a given number of transactions, there is a direct relationship among M and P. This entails that any increase in the money supply thus will result in an increase in the price level, in other words, inflation. At last, the Quantity Theory of Money gives an understanding of price changes in relation to money supply in a given economy (Law, 2014).

2.5 Previous Research

Similar research has been done before within this area of research. Mainly with the limitation to see the effects of fiscal policy on inflation in specific countries or regions in the long run. This section contains of a selection of articles and previous research that we have used to find inspiration and knowledge for this work.

Nguyen (2019) investigates both long-run, and short-run effects of government spending on inflation in three emerging economies in Asia, more specifically China, India, and Indonesia. This by exercising the approaches of cointegration and the Vector Error Correction Model (VECM) in time series data. The data runs from 1970 to 2010. The variables used in the VECM include consumer price index (CPI) and government spending (as a percentage of nominal GDP). Firstly, they ran unit root tests to assess the property for the time series data including variables as CPI growth rate, government spending (as a percentage of GDP), GDP per capita and nominal exchange rates. The authors highlight that certain variables are excluded from the study such as money supply, interest rates, and government debt, variables that relate to inflation. The results from the study support a cointegrating link betwixt government spending and inflation

(15)

in the long-run perspective. This regardless of the countries’ institutional differences. However, in the short run, the government spending as a percentage of GDP turns to have negative effect on inflation in China, whilst there is a positive effect in India and Indonesia. The conclusion of their work is that governments in emerging economies should be cautious with decisions regarding government spending.

Moreover, Magazzino (2011) provides empirical evidence of the link between public spending and inflation for some Mediterranean countries during the period of 1970-2009 using time-series data. The findings include the long-run relationship between the share of public spending and inflation for Cyprus, Greece, France, and Portugal using cointegration analysis. The study also includes evidence in the short run of directional flow from inflation to public spending for Cyprus, France, and Spain. Furthermore, some evidence suggest that public expenditure causes prices dynamics. Hence, the data supports the original Clark’s proposition which means that excessive public spending is a consequence of pressure on prices in the economy.

Another article within the subject of fiscal policy and inflation has been written by Fan et. al (2016). The article examines if the Fiscal Theory of the Price Level justifies inflation in the UK in the 1970s. In that period fiscal policy mayst have been set without any thought of future solvency implications, and monetary policy may have been

completely accommodative, which indicates a case of FTPL. The FTPL is set up as a structural model setting it against an orthodox model. In the Orthodox theory monetary policy is set by the Taylor Rule to reach the inflation target. Moreover, fiscal policy is appointed to attain fiscal solvency at the inflation rate. The two theories in their research are then stated as rival structural models and tested against the behavior contained in the data by the Indirect Inference method. According to the data, neither model is rejected but the FTPL-model accounts better for the behavior of the data than the Orthodox model and substantially outperforms it. The superior account of the period assumes that expectations of the two regimes were a probability weighted combination. One

conclusion is that fiscal policy has a considerable role in determining inflation in the weighted model.

(16)

growth. Notable in this study is that the authors use the variables of money supply and economic growth in addition to the variables used in our research (inflation and primary cash balance). However, the study examines data from Pakistan from 1972 to 2015. Their research contains empirical evidence in both the long run and the short run. The autoregressive distributed lag (ARDL) was used to see the long run relationship amongst the variables where the findings include two cointegrating vectors in the long run; one between economic growth and the other variables, and the other between money supply and the other variables. The Granger causality test was used to find the direction of causality. The results show an association among economic growth,

government expenditure and inflation in the long run. The Granger causality test reveals that there is a causality between government expenditure and inflation. The authors argue in conclusion that both monetary and fiscal policy impacts economic growth (Hussain & Zafar, 2018). From our point of view, the most interesting finding is the one about causality between government expenditure and inflation which goes in line with our research.

A recent study by Oyerinde (2019) shows the relationship between government spending and inflation in the one country case, more particularly in Nigeria. The findings of the study underline the tendency that inflation has a significant effect on government spending, and vice versa. Clarified, both variables can cause the other. The article shows a positive relationship where a rising inflation causes an increase in government spending. Implicating this result means that an increase in the inflation rate has a habit of increasing prices of goods (and services) on which government funds are spent upon. Furthermore, the study shows the causality flow from both government expenditure and inflation. This implicates that government expenditure exercises a great influence on inflation pressure in the country of Nigeria and the other way around. Oyerinde (2019) recommends a policy where the government should target output increase on its expenditure so that it can dampen the effect of the resulting trend in inflation.

In the case of developing countries Kandil (2005) investigated the effects of government spending shocks in multiple developing countries. Regarding the impact on inflation of government spending shocks, the author found that government spending shocks increases the trend price inflation compellingly. The variability of government

(17)

expenditure shocks increases trend price inflation compellingly across nations as well. Continued, the inflationary impact of government spending shocks increases the variability of price inflation bidirectionally, in other words both positive and negative.

(18)

3 Data

Two data sets were retrieved from The World Bank: the rate of annual inflation of consumer prices (2021c), and the annual growth of broad money supply (2021a). Another data set is the primary cash balance as a percentage of GDP, retrieved from the International Monetary Fund (2021). A sample of 135 cross-sections are used, with a 27-year time-period ranging from 1993 to 2019. Since data for the variables are missing at certain years for some countries, only observations with satisfactory data are

included.

3.1 Variables included

This section shows the variables included in the regression.

3.1.1 Inflation, Consumer Prices (annual %)

The inflation rate is the dependent variable, which is calculated annually by the

consumer price of a given year divided by the consumer price of the previous year. The variable is noted as INFLATIONit in the regression being run.

3.1.2 Primary Cash Balance (PCB) as Percentage of GDP

The explanatory variable in the regression refers to the budgetary central government balance of liquid assets, excluding things like interest payments. It is measured by the difference between the government’s revenue and its expenses. As such, each

observation is noted with a positive sign in the case of a surplus, where revenue outweigh expenses, or with a negative sign in the case of a deficit, vice versa. In the model, it is noted as PCBit.

3.1.3 Broad Money Growth (M2)

Furthermore, we added a monetary policy control variable; the annual growth of the broad money supply, more popularly noted as M2. It is distinguished from narrow money (M1) by the fact that it includes not only currency in circulation, but also savings deposits and money market deposit accounts (Corporate Finance Institute. n.d.a). M2 is less liquid than M1 but can otherwise easily be converted into cash. Furthermore, the growth of M2 is often tracked by central banks to forecast inflation (The Economist. 2010). Why include a variable pertaining to monetary policy? Because monetary policy is primarily concerned with money supply adjustments as to achieve a

(19)

combination of inflation and output stabilization (Mathai. 2020). Thus, it is

a worthwile indicator of the development of inflation along with fiscal policy. The variable will be noted as M2Growthit.

3.1.4 World Region

One of two types of dummy variables used is for world region, with the purpose of differentiating results between countries pertaining to different geographical regions. There are seven regions as categorized by The World Bank:

• East Asia & Pacific (EAP) • Europe & Central Asia (ECA) • Latin America & Caribbean (LAC) • Middle East & North Africa (MENA) • North America (NA)

• South Asia (SA)

• Sub-Saharan Africa (SSA)

These dummy variables take values of 1 or 0. For example, the dummy variable for East Asia & Pacific, DummyEAPi (note the absence of the t index, as geographical regions

do not alter over time), which for any individual country, takes the value 1 if the country belongs to that region, and 0 otherwise. There are seven categories of world regions, where we use six dummy variables to avoid the dummy variable trap of perfect collinearity.

3.1.5 Income Group

The second type of dummy variable, being used to measure the difference of the outcomes between countries of different levels of income. The groups included in our dataset are categorized, also by The World Bank, as follows in descending order:

• High income

• Upper middle income • Lower middle income • Low income

(20)

The dummy variables behave in the same manner as the dummy variables for world region, even if it shall be noted that income levels might fluctuate over time. Although, for this research, each country during all years pertains to its status as of 2021.

3.1.6 Distinction between Regions and Income Groups

There is evidence to suggest that the cause of inflation fluctuates between countries of different regions and income groups. High-income countries have an established history in not letting inflation rise too high, with monetary policy adjusting with changes in short run aggregate supply and demand. Middle- and low-income countries are worse off regarding inflation rates, typically stemming from budget deficits financed by domestic currency printing by governments, adhering to the fallacy of “just print more money” (Greenlaw et al., 2017).

Empirically, there are notable differences. In Graph 1 and Graph 2 below are the inflation rates by consumer prices for high income countries and low-income countries from 1993 to 2019 (The World Bank, 2021c). Inflation in low-income countries has been significantly higher in general. Despite this, we can observe similarities in trends, as both income groups have seen spikes in the last thirty years, albeit during different times, and both groups have seen inconsistent inflation rates during recent years. Of course, the financial crisis of 2008 impacted both income groups significantly.

(21)

Graph 2. Inflation, consumer prices (annual %) – Low-income countries

Next are the inflation rates for Europe and Central Asia and Sub-Saharan Africa. There are thirty-four out of forty-four European countries in the high-income group, and several low-income countries in Sub-Saharan Africa, making the two regions valuable polar comparisons. See Graph 3 and Graph 4 below. The data for Sub-Saharan Africa is very similar to that of the low-income group, signifying the poverty of that region. The data for Europe and Central Asia is also quite like that of the high-income group. One outlier is the spike in 1993 in Europe and Central Asia, which is not at all as noticeable for high income countries in general. This is likely due to the emergence of new

sovereign states in Eastern Europe at the time, mostly not significantly affecting the richer countries of Western Europe where inflation is concerned.

(22)

Graph 3. Inflation, consumer prices (annual %) – Europe & Central Asia

(23)

4 Methodology

4.1 Panel Analysis

The method used for conducting this research is panel data analysis, which is used for examining the behavior of individual subjects over a continuous period. As previously mentioned, there is heterogeneity in government spending between world regions and countries in different income groups, where panel data takes heterogeneity over time into account since it allows for subject-specific variables to be used. The subject in question is countries. With panel data we are given more degrees of freedom and more variability when combining time series of cross-section observations, suited for

studying the dynamics of change (Gujarati & Porter, 2009).

4.2 Why Panel Data?

In this report we have constructed a Panel Data regression analysis, but what are the advantages of panel data over time series or cross-section data? Baltagi (2013) mentions some advantages of panel data:

• Panel data refers to firms, individuals, governments, and countries over time, therefore there is bound to be heterogeneity.

• The combination of time series of cross-section observations panel data provides more: informative data, variability, degrees of freedom and efficiency. It also provides less collinearity among variables.

• Panel data suits better to study the dynamics of change due to studying repeated cross section of observations.

• Panel data better detects and measures effects that naturally cannot be observed in pure cross-section or timer series data.

Although there are plenty of advantages of using panel data, these are not panacea. The potential of panel data analysis to isolate the results of particular actions heavily

depends on the conformity of the assumptions of the statistical tools and the process of collecting data (Hsiao, 2003).

(24)

Model 1. PCB as explanatory variable

Inflation

it

= α + βPCB

it

+ ε

it

We then add the monetary variable, M2 growth:

Model 2. PCB and M2 growth as explanatory variables

Inflation

it

= α + β

1

PCB

it

+ β

2

M2growth

it

+ ε

it

4.4 The Least Squares Dummy Variable (LSDV) Model

Dummy variables can be used in dynamic panel data models to explain unobserved effects of individual units of a cross section. The individual effect in each individual is assumed to be fixed over time, with the fixed effects model being useful for explaining cross section heterogeneity in panel data.

We use two models to account for the dummy variables, one for regions and one for income groups:

Model 3. PCB and M2 growth as explanatory variables with region dummy variables

Inflation

it

= α + β

1

M2growth

it

+ β

2

DummyEAP*PCB

i

+

β

3

DummyECA*PCB

i

+ β

4

DummyLAC*PCB

i

+ β

5

DummyMENA*PCB

i

+

β

6

DummySA*PCB

i

+ β

7

DummySSA*PCB

i

+ ε

it

Model 4. PCB and M2 growth as explanatory variables with income group dummy variables

Inflation

it

= α + β

1

M2growth

it

+ β

2

DummyHigh*PCB

i

+ β

3

DummyLower-middle*PCB

i

+ β

4

DummyLow*PCB

i

+ ε

it

4.4.1 Dummy Variable Trap

A concern of integrating dummy variables into regressions is that, if not used correctly, they might create perfect collinearity, where we have an exact linear relationship

between variables. This is what is called a dummy variable trap. To circumvent this, the number of dummy variables must be one fewer than the categories of that variable. So, for n numbers of categories, (n-1) dummy variables shall be introduced.

4.5 The Interaction Between Fiscal and Monetary Policy

The interaction between fiscal and monetary policy is fundamental since they can influence each other noticeably through different decisions. These decisions made by one institution may have adverse effects on the other institution, which may result in

(25)

welfare losses for the whole society (Saulo et al., 2013). In the matter of shocks, fiscal and monetary policies work differently depending on the nature of a shock. Thus, for supply and demand shocks, these policies are predicted to be complementary, whilst for shocks of monetary and fiscal nature the two policies work as substitutes (Ankargren & Shahnazarian, 2019). In a recent study by Afonso et al. (2019) the authors found that inflation has a compelling impact on monetary policy. The authors also found that when facing increases in government debt, governments usually raise their primary balances, and that there is a substitution relationship between monetary and fiscal policies. Clearly there is an intercorrelation between fiscal and monetary policies and especially when looking at inflation one or another policy simply cannot be ignored (Afonso et al., 2019). Thus, even when investigating the impact of fiscal policy on inflation the

monetary aspect must somehow be included.

4.6 Gram-Schmidt Orthogonalization

As established, there exists an intercorrelation between fiscal and monetary policy. When measuring the effects of PCB and M2 growth on inflation concurrently, one runs the risk of one independent variable influencing the other in a way that may lead to questionable results. One way to amend this issue is to make the independent variables orthogonal to one another. In other words, eliminating any correlation in between that may occur. As the primary interest for this report lies in the effects of PCB on inflation, we treat PCB as the main variable. We do this by regressing PCB as an explanatory variable of M2 growth. The residual series of this regression is then used as a substitute for M2 growth when measuring its effects on inflation, and still has the same coefficient as M2 growth. What has changed, however, is that the coefficient for PCB is no longer affected by M2 growth, thus its coefficient remains as it would be when only PCB is applied as an explanatory variable and issues with multicollinearity between variables are lessened. This method is similar to the Gram-Schmidt process, a sequence of operations allowing a set of linearly independent vectors to be transformed into a set of orthonormal (orthogonal and normalized) vectors spanning the same space as the original set (Taboga, 2017).

(26)

4.7 Interaction Term

As we want to investigate the effect of PCB on inflation in separate regions and income groups, we must apply an interaction term between PBC an all implemented dummy variables. The dummy variables themselves only tell us the rate of inflation weighted against a base group. Now, this is not uninteresting, but it does not further our knowledge on the subject of fiscal policy. Rather, we allow for different coefficients for PCB depending on the dummy variable categories in question, where only observations pertaining to a category of interest are included. Simply, the quantitative variable and a qualitative variable influence the dependent variable multiplicatively (Gujarati & Porter, 2009).

4.8 Hypothesis

The hypothesis testing for this paper regards if the budgetary central government primary cash balance as share of GDP has a significant effect on a country’s inflation rate by consumer price. Thus, the null hypothesis and alternative hypothesis are as follows:

H0: β1 = 0 H1: β1 ≠ 0

The null hypothesis states that the slope coefficient for the PCB variable is not significant, and government spending is not proven to have a noticeable impact on inflation. Meanwhile, the alternative hypothesis states that there exists a significant relationship. In words, the null and alternative hypotheses can be stated as:

H0: Government spending does not have a significant impact on inflation. H1: Government spending has a significant impact on inflation.

(27)

5 Results

This section presents our results gained from the regressions followed by a discussion of the results.

5.1 Descriptive Statistics

Table 1. Descriptive Statistics

Variable Inflation M2 growth PCB

Mean 25.26692 23.18409 -156.6220

Median 5.074743 12.74409 0.490018

Maximum 23773.13 6968.923 6717.160

Minimum -30.24316 -45.47297 -276073.6

Std. Dev. 584.0896 185.4373 6757.412

Note: All units are measured in percentage points.

The average inflation is 25% and the average growth of broad money supply is 23%, while the average PCB is a low -157%. Both the highest rate of inflation and broad money growth can be found in The Democratic Republic of the Congo (DRC) in 1994. The maximum value of PCB (i.e., the largest primary surplus) is found in Ecuador in 1993. The minimum value for inflation is found in The Bahamas in 2015, the

corresponding value for broad money growth is found in Costa Rica in 1997, while the minimum value of PCB (the largest primary deficit) is found in The DRC in 1993. Clearly, The DRC during ’93 and ‘94 is a large statistical outlier, the effects of which we will discuss later.

5.2 Regression Analysis

With data from The World Bank and the International Monetary Fund for inflation, primary cash balance as share of GDP, and annual broad money growth, we analyzed 135 countries and a total of 1758 individual observations during a period between 1993 and 2019. The reason for this time-period is due to the massive number of republics in Eastern Europe and Asia gaining sovereignty after the collapse of the Soviet Union at the start of the 1990’s. Another reason for choosing this time-period is the lack of data

(28)

institutions such as the IMF and the World Bank data from certain years and countries are missing for various reasons. As mentioned earlier, inflation is the dependent variable denoted as annual percentage calculated by the consumer price of any given year divided by the consumer price of the previous year. The main explanatory variable refers to the budgetary central government balance, or primary cash balance (PCB) of the government. Primary cash balance is measured by the difference between the

government´s revenue and expenses. Finally, broad money growth (M2) is measured by the annual percentual increase of broad money supply. Table 1 will demonstrate the regression output, using only PCB as an indicator.

5.3 Basic Model

Table 2. The impact of primary cash balance on inflation in 135 countries during the period 1993-2019

Dependent Variable: Inflation Method: Panel Least Squares Sample: 1993 2019

Periods included: 27

Cross-sections included: 135

Total panel (unbalanced) observations: 1758

Variable Coefficient Std. Error t-Statistic Prob.

C 27.19657 13.65051 1.992347 0.0465

PCB -0.007096 0.002072 -3.424699 0.0006

Root MSE 571.8744 R-squared 0.006635

Mean dependent var 28.25167 Adjusted R-squared 0.006069

S.D. dependent var 573.9443 S.E. of regression 572.2000

Akaike info criterion 15.53799 Sum squared resid 5.75E+08

Schwarz criterion 15.54422 Log likelihood -13655.89

Hannan-Quinn criterion 15.54029 F-statistic 11.72856

Durbin Watson stat 2.089794 Prob(F-statistic) 0.000630

Starting with only PCB as an independent variable, its effect on inflation stands at -0.007. What does this mean? It means that for a one unit increase of a primary surplus we see a 0.007 decrease in inflation. Likewise, it suggests that a primary deficit increases inflation by the same unit. This is in line with previous research. While the coefficients are largely significant, the model has a very low fit with an R2 at only 0.66%. On its own, the primary cash balance explains a very small portion of the variation in inflation. Thus, to get a more accurate and representative estimation, we then must add the control variable M2 growth into the model.

(29)

Table 3. The impact of primary cash balance and broad money growth on inflation in 121 countries during the period 1993-2019

Dependent Variable: Inflation Method: Panel Least Squares Sample: 1993 2019

Periods included: 27

Cross-sections included: 121

Total panel (unbalanced) observations: 1671

Variable Coefficient Std. Error t-Statistic Prob.

C 24.15389 1.744795 13.84340 0.0000

PCB -0.007106 0.000258 -27.52185 0.0000

eM2 3.358270 0.010143 331.1083 0.0000

Root MSE 71.24031 R-squared 0.985115

Mean dependent var 25.26692 Adjusted R-squared 0.985097

S.D. dependent var 584.0896 S.E. of regression 71.30434

Akaike info criterion 11.37359 Sum squared resid 8480628

Schwarz criterion 11.38332 Log likelihood -9499.630

Hannan-Quinn criterion 11.37719 F-statistic 55195.07

Durbin-Watson stat 0.963452 Prob(F-statistic) 0.000000

We implement the orthogonalization process described in 4.6, and we get the variable eM2 to use as the monetary variable. From the output in table 3, we can observe an almost perfect fit to the model (R2 is close to 1), signifying how monetary policy is also a valuable determinant of inflation. Unsurprisingly, there is a significant relationship between broad money growth and inflation also.

Afterward, we examine the effects of PCB on inflation of countries pertaining to different regions and income groups. The results can be observed in the following two tables (Table 4 & Table 5).

(30)

5.4 Generalized Models

Table 4. The impact of primary cash balance on inflation in different regions during the period 1993-2019

Dependent Variable: Inflation Method: Panel Least Squares Sample: 1993 2019

Periods included: 27

Cross-sections included: 121

Total panel (unbalanced) observations: 1671

Variable Coefficient Std. Error t-Statistic Prob.

C 24.56289 1.764442 13.92105 0.0000 eM2 3.359155 0.010086 333.0384 0.0000 DUMMY_EAP*PCB 0.026350 0.827137 0.031857 0.9746 DUMMY_ECA*PCB -2.642516 1.759164 -1.502143 0.1332 DUMMY_LAC*PCB -0.043914 0.007888 -5.566961 0.0000 DUMMY_MENA*PCB -0.144379 0.698071 -0.206826 0.8362 DUMMY_SA*PCB -0.893854 2.259947 -0.395520 0.6925 DUMMY_SSA*PCB -0.007066 0.000257 -27-50620 0.0000

Root MSE 70.72733 R-squared 0.985328

Mean dependent var 25.26692 Adjusted R-squared 0.985267

S.D. dependent var 584.0896 S.E. of regression 70.89724

Akaike info criterion 11.36512 Sum squared resid 8358935

Schwarz criterion 11.39107 Log likelihood -9487.555

Hannan-Quinn criterion 11.37473 F-statistic 15955.10

Durbin-Watson stat 0.971668 Prob(F-statistic) 0.000000

We first examined the different effects between the world regions. The final region, North America, is excluded from the regression as to avoid perfect collinearity. We can observe that there are two regions where PCB has a significant impact on inflation, Latin America & Caribbean and SubSaharan Africa, with coefficients at 0.04 and -0.007, respectively. This suggests that these two regions are more likely to be hurt by a primary deficit.

(31)

Table 5. The impact of primary cash balance on inflation in different income groups during the period 1993-2019

Dependent Variable: Inflation Method: Panel Least Squares Sample: 1993 2019

Periods included: 27

Cross-sections included: 121

Total panel (unbalanced) observations: 1671

Variable Coefficient Std. Error t-Statistic Prob.

C 24.71359 1.774899 13.92394 0.0000

eM2 3.359016 0.010100 332.5616 0.0000

DUMMY_HIGH*PCB 0.425650 0.604036 0.704677 0.4811

DUMMY_LOW_MID*PCB -3.040909 0.606109 -5.017095 0.0000

DUMMY_LOW*PCB -0.007065 0.000257 -27.46363 0.0000

Root MSE 70.88994 R-squared 0.985261

Mean dependent var 25.26692 Adjusted

R-squared 0.985226

S.D. dependent var 584.0896 S.E. of regression 70.99624

Akaike info criterion 11.36612 Sum squared

resid 8397415

Schwarz criterion 11.38234 Log likelihood -9491.392

Hannan-Quinn criterion 11.37213 F-statistic 27841.71

Durbin-Watson stat 0.984536 Prob(F-statistic) 0.000000

The countries in the upper middle-income group are omitted on the same basis as North America in Table 4. With dummies for the income groups, we see similar results as in Table 4. The inflation rates of two income groups are significantly impacted by PCB, in this case, the lower-middle income group and the low-income group. Again, we have a case of a primary deficit raising inflation, with coefficients of -3.04 and -0.007,

respectively.

5.5 Robustness Check

For the results to be reliable, robustness checking needs to be conducted. A Johansen Fisher panel cointegration test is run to check for cointegration between variables (See Table B6 in Appendix B). The test shows that more than two variables are cointegrated in the long run, significant at any conventional level of alpha. The regression is henceforth not spurious and can be proceeded with. We also tested cointegration between only primary cash balance and broad money growth (See Table B7 in Appendix B), with both being cointegrated in the long run, echoing the theory that fiscal policy and monetary policy sometimes interact.

(32)

To check for autocorrelation, we look at the Durbin-Watson statistic. Values between 0 and 2 indicate positive autocorrelation while values between 2 and 4 indicate negative autocorrelation. In the model with PCB as the sole explanatory variable the DW statistic is nearly 2, suggesting that there is close to no sign of autocorrelation. However, when adding eM2, the statistic drops to almost one, indicating that positive autocorrelation may be present in that series. As we used orthogonalization to amend for multicollinearity, the coefficients for PCB and its interpretations are not affected by this, but it is nonetheless noteworthy and should be approached with caution.

5.6 Discussion

There are a few takeaways from the results we have arrived at. Primary balance having a negative effect on inflation is adjacent with the notion that a fiscal deficit as such raises inflation. As theorized, that particularly applies to lower-middle to low-income countries due to the historically volatile nature of the fiscal policies seen in the developing world.

The primary results align with previous research about the relationship between government expenditure and inflation. While scholars like Nguyen (2019), Magazzino (2011) and Oyerinde (2019) have examined certain regions or one certain country, our study at a global scale echoes their assessments. There is a significant relationship between government spending and inflation, and in our findings, we are shown that an excess spending causing a primary deficit is linked to a rise in inflation. There are also links to be found to the FTPL, which connects price stability to proper fiscal policies where the primary balance is included. Further reference can be made to the Ricardian equivalence theory, that states that in the event of a government budget deficit, people save more money with the assumption of higher taxes in the future. With our results, this assumption has some merit. With a budget deficit follows higher consumer prices, hence it is reasonable to expect consumers to decrease their consumption and start saving money. If this is a conscious choice in the precedence of a tax increase, however, is difficult to validate, and might be explained more by the instinct that more expensive goods and services are less attractive. Importantly, let us not forget the significance that monetary policy plays in inflation as well. As assessed by Hume and Friedman in the

(33)

quantity theory of money, and reiterated in our empirical research, an increase in the central bank’s money supply significantly increases the price level.

There is a great heterogeneity between the world regions and its countries regarding the effects of government spending. As mentioned earlier, governments with a high share of income tend to authorize a larger share of national production than governments with low income. Thus, central government spending in richer countries usually stands for a greater share of national output in comparison to poorer countries. Now looking at the results where we implemented a dummy variable in our regression for world regions and integrating them with the PCB variable, we observed that two regions significantly stand out. Those being Latin America & Caribbean and Sub-Saharan Africa. These results are hardly surprising due to countries in these regions historically suffering from inflation following excessive spending.

The regression with the income group dummies showed another result with two statistically significant effects, this time for lower-middle income and low-income countries. This result was also expected since low-income countries often struggle with inflationary problems referring to LIC´s financing budget deficits by printing domestic currency by the government, leading to the deception of “just print more money” to solve the problem. In fact, it only worsens the inflation. Most recent example is

Venezuela who suffers from this hyperinflation. An earlier example is Zimbabwe in the mid-2000s. It is surely not a guarantor for LIC´s to have a large inflation but according to our results and earlier research it is more common in these countries. Why is that a fact? Governments and Central Banks in LIC´s struggle to maintain or to keep the inflation relatively low. This partly due to apace globalization affecting countries in several areas. There are some more specific reasons to why LIC´s reacts more devasting than other country groups. Firstly, LIC´s core inflation is more strongly affected by the global core inflation than others. Secondly, LIC´s core inflation is more strongly

affected by the global food inflation than others. Thirdly, LIC´s core inflation is affected more clearly (but more variably) to the global energy inflation than others (Ha, et al. 2019).

(34)

made. One must address “the elephant in the room”, of which there are a few. First, PCB only explains 0.66% of the variation in inflation, which is very little. Secondly, you may note that the coefficients for PCB in SSA and LIC’s are nearly identical both to each other and the coefficient of PCB at a worldwide aggregate. Recall that in 5.1, the DRC stands for many of the most extreme observations (highest inflation, M2 growth, and deficit), and happens to belong to both these groups. It is fair to assume that the two observations for this country for the years ’93 and ’94 influence the PCB

coefficient greatly, and one of the motives behind the usage of dummy variables is to account for these effects. Now, the DRC is not alone in indicating the relationship found, as the coefficients for LAC & LMIC’s imply likewise. Still, LAC is also a region with historical cases of hyperinflation caused by money printing. One cannot ignore, then, the possibility that extreme cases of hyperinflation shift the PCB coefficient not only on a worldwide aggregate, but also within the groups where the relationship is significant.

When considering determinants of inflation there are plenty of indicators to keep an eye on. Several indicators are excluded from the regressions in this report. Nevertheless, these may still have an impact on inflation. Variables that are left out are for example public debt, unemployment, differences in welfare systems, and investment in

infrastructure. Čaklovica and Efendic (2020) found evidence for some structural variables, including unemployment rate, affecting inflation in 28 European countries. Briefly, when unemployment decreases, the inflation rate rises, and vice versa. This inverse relationship between inflation and unemployment is former and first explained by Phillips (1958) and has become to be known as the ‘Phillips Curve’. Fiscal

policymakers can face challenges due to the positive correlation between the variables of inflation and unemployment. This since policy actions that aims to boost economic output and decrease unemployment often aggravates inflation. Alternatively, policy actions that aims to dampen inflation often increase unemployment and stifles the economy.

Public debt can lead to an increase in inflation as well. When governments borrows money domestically, a higher inflation leads to a devaluation of the local currency which implies that the amount owed increases. Furthermore, increasing debts results in

(35)

larger money supply which can be a forerunner to a rising inflation. Public debt can be inflationary in countries that already has a high debt and who increases it furthermore (Romero & Marin, 2017).

Even though M2 stands for a higher R2 along with a larger coefficient, this is not the main focus of this thesis. Still, what we know from our results is that M2 presents a greater impact on inflation than PCB, even if PCB represents a part as well. As

mentioned, fiscal policy actions may have greater impact in certain countries in specific times. Hence, referring to our results obtained, monetary policy still should be more widely considered when dealing with inflation while fiscal policy can serve as a complement.

5.7 Limitations

In finding data for inflation, government expenditure, and monetary policy, we found that the most extensive data can be found from the past 30 years, although for some countries, the data is limited, and certain countries lack data for inflation completely. Reasonably, a lot of new countries have been added since 1993, with, as previously referenced, the fall of the Soviet Union bringing sovereignty to many republics.

Furthermore, the Least Squares Dummy Variable model may be inconsistent if the time-period is too short. As we wanted a large number of cross-sections, a balanced panel data set was difficult to satisfy, as certain observations lacked satisfactory data. Hence, when accounting for M2 growth within the regression, 14 cross-sections are omitted due to data for M2 growth not being satisfied within these cross sections. Many of those are European countries, and the coefficient for PCB in that region (along with Central Asia) may have been impacted. Ideally, an even larger sample size would have been used. Data for inflation was found in almost all the world, while data for M2 growth and especially PCB was scarcer. There are also many other historical cases of inflation spikes that could also have been included were the data more available. Our results do as such not give a general interpretation of all countries at all times. Furthermore, while the worldwide aggregate shows a significant relationship between PCB and inflation, one must be careful with drawing too many conclusions, as the coefficient may be shifted by extreme statistical outliers.

(36)

6 Conclusion

First and foremost, the main conclusion for this report is that an increased primary deficit raises inflation, and the same logic works the other way around, meaning that an increased primary surplus dampens inflation. This supports our theory that government spending has a significant impact on inflation.

Due to the vast number of financial crises witnessed in history, with several just during the past century, it is evident to all the fickle nature of the global economy. While banking and trading famously play a large part in inflation, there is reason to have an eye on the public sector and realize the significance of a solid budget balance. As of writing, the covid-19 pandemic is still ongoing and has arguably set the economy in a more uncertain situation than has been witnessed for many generations. Making predictions of what the future holds is, even in the best of times, tricky.

The classic monetarist view on inflation suggests that changes in money supply is the most compelling determinant of the rate of economic growth. It is steered by Central Banks which controls the monetary policy and making decisions on adjustments and maintenance of the inflation rate. However, the Keynesian model states that total spending in an economy affects output, employment, and more specific in this report, inflation. When a Central Bank lacks in power in steering the inflation, the government should consider policy actions focusing on affecting the inflation rate. We believe that the Keynesian view, and more specific, government spending should be considered more frequently than just put all mandate to the Central Bank. In times of inefficient Central Banks, fiscal policy implications might be the solution.

As PCB shows a very small fit when explaining the variation in inflation, dwarfed by the corresponding fit for M2 growth, one can view fiscal policy as a complement to monetary policy when controlling inflation. Government spending as a source of inflation mostly does not hold much weight on its own. But when it does, it is arguably in the cases of hyperinflation noted throughout this report; a rare occurrence in

developed countries but has historical precedence in countries of lower income in certain regions. While primary deficits may not be a concern in much of the world, these examples serve as a cautionary tale of the worst-case scenario.

(37)

Moving forward, pandemic or not, it is of interest that government budgets are kept at a balanced level, and preferably at a surplus, in order not to influence inflation in the wrong direction. Especially LIC’s and LMIC’s face a risk of inflation from an excess public expenditure. The time value of money is more than a monetary issue, the course of which may in certain cases depend on the budget balance of the central government. That does not necessarily imply that taxes need to be raised or for more money to be borrowed for the sake of higher revenue, but rather a more fiscally conservative policy where spending like welfare and stimulus checks is more closely monitored.

Suggestion for further research within this area is to look more closely at high income countries since these generally has not seen significant effects on inflation by primary deficits according to our results. Furthermore, these high-income countries represent a large share of the world’s cumulative GDP and has a great impact on the world

economy. High-income countries with effective policies for controlling the inflation could then work as role-models to less developed countries in these questions regarding inflation and government spending. On the opposite end, it would be worthwhile to more closely study the cases where hyperinflation has occurred, like in Venezuela and Zimbabwe, to better determine when, why, and how fiscal deficits are linked to hyperinflation. Additionally, one can study price changes at an industry-specific level, as to how different industries are affected by fiscal policy decisions. Finally, factors determining inflation like debt and unemployment can be studied alongside government spending as these factors are sometimes cointegrated.

(38)

Reference list

Afonso, A., Alves, J., & Balhote, R. (2019). Interactions between monetary and fiscal policies. Journal of Applied Economics, 22(1), 132-151.

https://doi.org/10.1080/15140326.2019.1583309

Ankargren, S., & Shahnazarian, H. (2019). The Interaction Between Fiscal and Monetary Policies: Evidence from Sweden (Working Paper Series No. 365). Sveriges Riksbank.

Baltagi, B.H. (2013). Econometric analysis of panel data (5. ed.). John Wiley & Sons, Inc.

Barro, R.J. (1989). The Ricardian Approach to Budget Deficits. The Journal of Economic Perspectives, 3(2), 37-54. https://doi.org/10.1257/jep.3.2.37

Bassetto, M. (2008). Fiscal Theory of the Price Level. In S. N. Durlauf., & L.E. Blume (Eds.) The New Palgrave Dictionary of Economics (2nd ed., pp. 2215-2218). Palgrave Macmillan. https://doi.org/10.1057/978-1-349-95121-5_2575-1

Čaklovica, L., & Efendic, A. S. (2020). Determinants of Inflation in Europe – A Dynamic Panel Analysis. Financial Internet Quarterly, 16(3), 51–79.

https://doi.org/10.2478/fiqf-2020-0018

Catão, L., & Terrones, M. (2005). Fiscal deficits and inflation. Journal of Monetary Economics, 52(3), 529–554. https://doi.org/10.1016/j.jmoneco.2004.06.003

Christiano, L.J., & Fitzgerald, T.J. (2000). Understanding the Fiscal Theory of the Price Level. National Bureau of Economic Research.

Ciccarelli, M., & Mojon, B. (2005). Global inflation (European Central Bank Working Paper No. 537). European Central Bank.

https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp537.pdf

Cochrane, J.H. (1998). A frictionless view of U.S. inflation. NBER Macroeconomics Annual, 13, 323-384. https://doi.org/10.1086/ma.13.4623752

Cochrane, J.H. (2001). Long-Term Debt and Optimal Policy in the Fiscal Theory of the Price Level. Econometrica, 69(1), 69-116. https://doi.org/10.1111/1468-0262.00179

(39)

Cochrane, J.H. (2005). Money as stock. Journal of Monetary Economics, 52(3), 501-528. https://doi.org/10.1016/j.jmoneco.2004.07.004

Corporate Finance Institute. (n.d.a). Definitions of Money.

https://corporatefinanceinstitute.com/resources/knowledge/finance/definitions-of-money/

Corporate Finance Institute. (n.d.b). Keynesian Multiplier.

https://corporatefinanceinstitute.com/resources/knowledge/economics/keynesian-multiplier/

Fan, J., Mindford, P., & Ou, Z. (2016). The Role of Fiscal Policy in Britain's Great Inflation. Economic Modelling, 58, 203-218.

https://doi.org/10.1016/j.econmod.2016.05.027

Fischer, S., Sahay, R., & Végh, C. (2002). Modern Hyper- and High Inflations. Journal of Economic Literature, 40(3), 837–880. https://doi.org/10.1257/jel.40.3.837

Friedman, M. (1963). Inflation: Causes and Consequences. Asia Pub. House.

Greenlaw, S. A., Taylor, T., & Shapiro, D. (2017). Principles of Economics (2nd ed.). OpenStax College, Rice University.

Gordon, D., & Leeper, E. (2002). The price level, the quantity theory of money, and the fiscal theory of the price level (NBER Working Paper Series No. 9084). National Bureau of Economic Research.

https://www.nber.org/system/files/working_papers/w9084/w9084.pdf Gottfries, N. (2013). Macroeconomics. Palgrave Macmillan.

Gujarati, D.N. & Porter, D.C. (2009). Basic econometrics (5th ed.). McGrawHill. Ha, J., Ivanova, A., Montiel, P., & Pedroni, P. (2019). Inflation in low-income countries (Policy Research Working Paper No. 8934). World Bank.

https://documents1.worldbank.org/curated/en/410071562700985189/pdf/Inflation-in-Low-Income-Countries.pdf

(40)

Hsiao, C. (2007). Panel Data Analysis - Advantages and Challenges. Test, 16(1), 1-22. https://doi.org/10.1007/s11749-007-0046-x

Hussain, M. I., & Zafar, T. (2018). The Interrelationship between Money Supply, Inflation, Public Expenditure and Economic Growth. European Online Journal of Natural and Social Sciences, 7(1), 1-24. https://european-science.com/eojnss International Monetary Fund. (2021). Government Finance Statistics [Data set].

https://data.imf.org/?sk=A0867067-D23C-4EBC-AD23-D3B015045405&sId=1393552803658

Kandil, M. (2005). On the Effects of Government Spending Shocks in Developing Countries. Oxford Development Studies, 33(2), 269-304.

https://doi.org/10.1080/13600810500137970

Keynes, J.M. (1936). The General Theory of Employment, Interest and Money. MacMillan.

Landau, D. (1985). Government Expenditure and Economic Growth in the Developed Countries: 1952–76. Public Choice, 47(3), 459-477.

https://doi.org/10.1007/BF00182148

Law, J. (2014). A dictionary of finance and banking (5th ed.) Oxford University Press. Leeper, E. (1991). Equilibria under “active” and “passive” monetary and fiscal policies. Journal of Monetary Economics, 27(1), 129–47.

https://doi.org/10.1016/0304-3932(91)90007-B

Lin, H.-Y., & Chu, H.-P. (2013). Are fiscal deficits inflationary? Journal of International Money and Finance, 32, 214–233.

https://doi.org/10.1016/j.jimonfin.2012.04.006

Magazzino, C. (2011). The nexus between public expenditure and inflation in the Mediterranean countries. Theoretical and Practical Research in Economic Fields, 2(1), 94–107.

Mathai, K. (2020). Monetary Policy: Stabilizing Prices and Output. https://www.imf.org/external/pubs/ft/fandd/basics/monpol.htm

References

Related documents

The variable family second blockholder (2NDFAM2) is a dummy taking the value 1 if the second largest shareholder holds at least 10% of the firms total voting rights, and is a

Monetary policy, inflation risk premia, term structure, zero bound, quantitative easing, inflation compensation, inflation expectations, break-even inflation, inflation swaps,

Using the GVAR model and Generalized impulse response functions we find that a positive standard error shock in Money Supply, Discount Rate and Monetary Policy uncertainty

Key words: Deficit bias, European semester, Fiscal consolidation, Fiscal policy councils, Forecasts, Independent fiscal agencies4. Title: Preaching to

In the case of Sweden, there is a decline in positive attitudes towards non-EU immigrants that correlate with the implementation of the more restrictive policy

They include strengthening of the EU fiscal rules, broader macroeconomic surveillance, better connect between the policy processes at the European level and the national level, and

(2010), The Rationale for Fiscal Policy Councils: Theory and Evidence, Paper to Conference on Independent Fiscal Institutions, March 18-19, Fiscal Council Republic of

issues including tax, employment and regulatory policies as well as resource depletion and financial crises. 9) The Council is not instructed to provide own forecasts but may base its