Master Thesis
Inflation mean and variability interaction:
The role of institutions, fossil energy, trade openness and level
of development
Author: Class Code: Student number: Email University: Faculty: Specialization: Date: Supervisor: Co-assessor Erwan DujeancourtMaster Thesis (EBM868A20) S3398897
e.dujeancourt@student.rug.nl University of Groningen
Faculty of Economics and Business International Economics and Business June 2018
Prof. Dr. Jakob De Haan Dr. Andreas C. Steiner
Abstract
This Master Thesis examines the multilevel effect of subsidiary conditions on the inflation and its variability relationship from 1965 to 2015. The objective of this paper is to determine whether the institutions, fossil energy, trade openness and level of development conditions affect the inflation level variability nexus. The research reveals the existence of national regulation and trade openness effects on the inflation relationship nexus.
Table of content
1. Introduction ... 4
2. Literature review and hypotheses development ... 5
2.1 Correlation between higher inflation average and higher variability ... 6
2.2 Relationship between GDP per capita and the nexus inflation level and variability ... 7
2.3 The association between the crude oil price and the inflation level variability nexus ... 8
2.4 The relation between the efficiency of institutions on inflation and its volatility ... 9
2.5 The effect of trade openness on the inflation mean and variability interrelationship ... 10
3. Methodology ... 11
3.1 Specification of the data sources and the final sample ... 11
3.2 Specification of the measures and variables in the theoretical model ... 13
3.3 Testing specifications ... 18
3.4 Econometrics models ... 20
4. Empirical results and analyses ... 22
4.1 Model A estimations ... 22 4.2 Model B estimations ... 23 4.3 Model C estimations ... 24 4.4 Model D estimations ... 27 4.5 Discussion of results ... 27 5. Final notes ... 30 5.1 Conclusion ... 30 5.2 Further discussions ... 32 6. References ... 33 7. Appendix ... 36
7.1 Formal List of Country ... 36
7.2 Variable descriptions ... 38
7.3 Components of the Economic Freddom Index ... 43
7.4 Summary Statistics ... 44
7.5 Collinearity Diagnostics ... 44
7.6 Multicollinearity Matrix ... 46
1. Introduction
On June 14, 2017 (The Fed, 2017), the former president of the Fed, Janet Yellen, urged to rethink the four percent inflation target not only in the US but in all monetary institutions all over the world. According to Yellen, the economy has recently been more constrained by the Zero Lower Bound (ZLB) than at the time when the two percent goal was adopted. In her opinion, the most crucial point is to analyze the benefits and potential costs of a higher inflation objective. She finally suggests in the speech that they ‘very much look forward to seeing research by economists that will help inform our future decisions on this’ (The Fed, 2017). Most central banks, such as the ECB, the BoJ or the Fed, consider the optimal inflation level to be around two percent. However, several economists already questioned if this is optimal inflation level (Krugman, 2013; Ball, 2014; Blanchard, 2016). These economists fully support the four percent inflation target, claiming that with a higher baseline for inflation, the economy would avoid the ZLB problem, the liquidity trap, massive unemployment, and economic downturn. Because recent recessions predominantly occur in a low-inflation environment, a higher inflation target would significantly diminish real interest rates, overcome the ZLB problem and its severe consequences. However, the idea remains unpopular among monetary economists (Mishkin, 2010; Akhtar Hossain and Arwatchanakarn, 2016). As the former Fed president already stated, potential costs could occur and overcome the benefits. The opponents of a higher inflation target affirm that the higher inflation objective would undermine the credibility of monetary institutions, change the expectations and increase the variability of inflation. Bernanke (2010) highlighted that a higher inflation goal could “likely entail much greater costs than benefits.” Previous empirical studies emphasized the robust relationship between the variability of inflation and the average inflation rate. Nevertheless, these studies suffer from an endogeneity bias, as countries struggling with high inflation variability are more likely to have also high inflation. Thus, previous literature undertook studies by comparing inflation level and variability from developed to developing countries (Logue and Willet, 1976; Taylor, 1981; Alpanda and Honig, 2010). But even if they discovered significant differences between nations, they did not identify which factors could explain these differences between institutions, fossil energy prices, the degree of development and national trade openness. Thus, additional research is needed to surely ascertain the results.
Consequently, this Master Thesis will investigate the impact of institutions, fossil energy prices, the degree of development and the openness to trade regarding the relationship between inflation level and inflation variability. This Master Thesis has two principal objectives. Firstly, to review and update the empirical results on the relationship between inflation and inflation variability. Secondly, to investigate whether the efficiency of national institutions, the level of development, fossil energy prices or openness of the country have a potential influence on the link between inflation and inflation variability. Thus, the following research questions will be addressed:
Does the level of development, the fossil energy prices, national institutions and openness of countries affect the relationship of inflation and inflation variability? Therefore, this Master Thesis aims to answer the current question regarding the effect of a higher inflation target that several economists such as Janet Yellen try to determine.
The thesis is structured as follows: Chapter 2 discusses the main concepts and theories, reviews the literature on the elements that affect the inflation level variability nexus, and will state the hypotheses. Thereafter, chapter 3 will describe the data used in the empirical research, explain the methodology and present the empirical model. Chapter 4 will provide and discuss the results. Finally, chapter 5 will conclude the principal research question.
2. Literature review and hypotheses development
The effect of institutions, level of development, the variability of crude oil price and the openness of countries on inflation and its variability are essential topics for the monetary policy theory. Indeed, the two principal economic goals usually are full employment and reasonable price stability (Gordon, 1971). In addition, Okun (1971) considers that inflation is a strong indicator to apprehend. He supports that, between both objectives, the society would prefer higher inflation and variability to unemployment which is considered as more significant evil for a community. Therefore, in order to analyze the topic, a concise summary of the previous research on the topic will be provided.
2.1 Correlation between higher inflation average and higher variability
Since Taylor (1981) and Okun (1971), most scholars have emphasized the relatively stout relationship between a higher average rate of inflation and higher variability (Logue and Willet, 1976; Foster, 1978; Fischer, Sahay and Végh, 2002; Hyeon and Chin, 2013). Indeed, as Taylor (1981) underlines, several international comparisons and historical studies observed correlations between price level and highly variable inflation. He notices with Friedman (1977) the tremendously high variability of inflation which accompanies the high average of the inflation rate. According to Friedman, when inflation is high, the population is uncertain whether policymakers would prefer to increase inflation rather than constraint economic sustainability. This uncertainty leads to more future uncertainty and thus causes variability of inflation.
Furthermore, even if voluminous literature notified the strong association between inflation level and inflation variability, Logue discovered that the variability was less responsive to the average level of inflation (Logue and Willet, 1976). In their paper, the authors examine the relationship between the rate and predictability of inflation by studying the past experiences of a large number of countries. Indubitably, past experiences are not always safe guarantees for future performance. Logue and Willet argue that this situation is particularly accurate in the 1970’s because much of the observed variability of past inflation rates may have been induced by deliberate government policies aimed at stabilizing the price level rather than alleviating expectations.
2.2 Relationship between GDP per capita and the nexus inflation level and variability
As previously mentioned, research of monetary economics demonstrates that a higher inflation rate is correlated with less steady inflation (Okun, 1971). However, an extensive literature review by Bruno and Easterly (1998), and Chiu and Molico (2011) reports that the high variability of inflation aggravates the welfare of nations.
Interestingly, instead of revealing the simple correlation of GDP per capita and inflation variability, this Master Thesis will examine the impact of the same indicator on the relationship of the variability of inflation and the inflation mean. Indeed, while some economists perceive the correlation between higher inflation and volatility. Other economists may observe the association between subsidiary economic condition and volatility. In other words, these economists estimate that prior studies could suffer from an endogeneity bias, as countries struggling with high inflation variability are more likely to have high inflation as well. The relationship between average inflation and higher variability might be a result of the literature’s concentration on advanced economies for low variability and on developing countries for high variability (Grier and Perry, 2000; Daal, Naka, and Sanchez, 2005; Fountas and Karanasos, 2007; Jiranyakul and Opiela, 2010). Therefore, this Master Thesis will maximize the sample in order to include developing nations. Subsequently, this work has the objective to determine whether the previous research on the question does have biased data and what the real effect of higher inflation on volatility may be, after controlling for GDP per capita.
Other scholars (Logue and Willet, 1976; Taylor, 1981) discover that highly industrialized nations are more capable of conducting internal monetary and fiscal policy, thus limiting the inflation level and variability relationship. Even if they discover significant differences, they do not stipulate which factors could explain this distinction. They suppose that highly developed countries have better institutions. Moreover, these scholars categorize countries into different levels of development. This research design could be problematic if nations are heteroskedastic. Moreover, these economists neglect the possibility that the level of development and GDP per capita are dynamic. Thus, additional research is needed, and this thesis aims to fill the gap of previous literature.
developed financial markets. Lastly, developing countries have slightly more massive government budget deficits on average (Foster, 1978).
Overall, taking into consideration the abovementioned, the following hypothesis is developed:
Hypothesis A: GDP per capita has an adverse effect on the inflation level variability nexus.
2.3 The association between the crude oil price and the inflation level variability nexus
Economists assume that inflation is determined by numerous parameters. In their work, Chinn and Leblanc (2004) demonstrate that the oil price significantly changes and impacts inflation. Taylor (1981) states that a dominant supply shock was the prominent cause of the variability of inflation in the 1970’s. Voluminous literature (IMF, 2006; Kilian, 2009; Álvarez et al., 2011; Busse, Bernhard and Ihle, 2011) found a significant association between the oil price and inflation. However, this analysis will assess the oil price effect on the correlation between inflation and its variability. Oil price shock is the prominent example of exogenous supply shock that affects the inflation average and its variability. In order to mainly apprehend the impact of oil price that several researchers underlined the higher average of inflation and its variability, this thesis will include the world oil price.
Moreover, Chinn and Leblanc (2004) confirm that the significant and immediate effect of oil price on national inflation. Therefore, the model will assume this specification. In the empirical model, including the crude oil price is substantial to determine the multilevel effect of the different independent variables.
Taking into consideration previous work, the following model assesses that crude oil price does not affect inflation mean and inflation variability relationship but remains an exogenous parameter. Consequently, the second hypothesis is formulated as followed:
2.4 The relation between the efficiency of institutions on inflation and its volatility
As mentioned beforehand, even if the adverse effects of inflation variability on economic prosperity are generally recognized, the causes are not sufficiently examined (Aisen and Veiga, 2008; Chiu and Molico, 2011b). Indeed, several researchers have found a correlation between higher inflation and volatility. Fischer et al. found in their paper that higher inflation conduct to extreme volatility (Fischer, Sahay and Végh, 2002).
Nevertheless, this correlation would be simple to interpret if the number of institutions conducting monetary and fiscal policy was distributed independently and randomly. In other words, the result would be significant and apparent, if both efficient and the less efficient national institutions, were found in equal proportion among the high inflation variability sample, as among the low variability one. Indeed, one could assess that the data is biased. Consequently, these studies could suffer from an endogeneity bias, as countries struggling with high inflation variability are far more likely to have high inflation as well. Hence, the effect of the efficiency of institutions on the mean and variability of inflation nexus represents a gap in the literature.
Throughout the years, various groups of international scholars show that the decline in inflation and its variability is attributed to the improvements of the central bank practice (Okun, 1971; Taylor, 1981; Bomberger and Makinen, 1993; Fischer, Sahay and Végh, 2002). In his paper, Rogoff (2003) distinguishes four practices that improved the quality of central banks: first, the enhanced central bank independence; second, the conservative orientation of the institutions; third, the enhanced communication strategy such as “the moving forward,” and finally, the new monetary channels and tools that the monetary institutions possess. However, determining the efficiency of a central bank or a government is a difficult challenge. Several scholars proposed an approach to calculate and measure the efficiency of the central bank. For example, Dreher, Sturm and de Haan (2008) Klomp and de Haan (2010) only find a significant negative effect of central banks in specific cases and with a minority of countries. Central bank efficiency indicators remain limited. Therefore, this study could apprehend other institutional aspects that could influence the inflation level variability relationship.
Moreover, erstwhile scholars underline the interest of some governments to realize surprise inflation in order to acquire additional revenue. The more public spending augments and the higher public debt is, the more the government will be tempted to provoke surprise inflations (De Grauwe, 1996). Thus, the influence of the governmental institution performance could be an excellent indicator to analyze the inflation level variability relationship.
Furthermore, Hayo and Voigt (2008) found that a higher degree of legal system independence, as well as trust towards legal systems, tends to decrease the inflation via two channels. First, by fostering potential growth, second, by increasing the independence of the central bank.
In addition, regulations have a prominent effect on inflation average and volatility. Previous literature highlights the impact of labor, capital and business regulation on inflation. The scholars argue that nations with more labor or capital market constraints are less flexible and more sensitive to external shocks. Similarly, the empirical research supports an adverse influence of business regulation on national inflation and inflation volatility (Copelovitch and Singer, 2008; Echeverri-Carroll, E. and Ayala, 2009; Jaumotte and Morsy, 2013).
Previous macroeconomic performance of a country, fiscality, exchange rate development, regulation, the judicial system, the government structure or the structure of its financial system, and the power of the monetary institution have a substantial positive impact on inflation. In the remainder of this study, the efficiency of institutions will be defined and its impacts examined. Consequently, based on the above indication, the third hypothesis is built:
Hypothesis C: Efficient Institutions have a negative effect on the inflation level variability nexus.
2.5 The effect of trade openness on the inflation mean and variability interrelationship
Lastly, openness to trade is frequently seen as a beneficial factor to avoid a higher level of inflation mean and variability. However, the results of the existing literature are ambiguous. Cooke (2010) argues that surprise monetary expansions cause real exchange rate depreciation and this real depreciation is more harmful if the country is more open to trade. Subsequently, the government will not have any interest to create surprise inflation and provoke high price volatility, when trade openness is high.
negative effect of openness on inflation level and inflation volatility for countries that were initially more trade closed economies with high inflation level (Ghosh, 2014). Conversely, it could be argued that a more open economy would be more sensitive and vulnerable to external shocks and will face higher variability of inflation (De Grauwe, 1996).
These findings regarding the interaction between inflation mean and inflation variability lead to the formulation of the following hypothesis
Hypothesis D: The degree of trade openness negatively affects the relationship between level and variability of inflation
3. Methodology
This chapter depicts the data and the methodology that was used in order to formally test the hypotheses established in the precedent chapter. This section explains the econometric model specification performed and describes the estimation techniques applied.
3.1 Specification of the data sources and the final sample
1965 in order to enrich the analysis. However, several states have been excluded because the data was neither constant, nor sufficient to be incorporated in the analysis.
In addition, given the influence of energy price and specifically of crude oil price, the annual crude oil price data provided by the Bank of Saint Louis (2018) has been added in order to better understand the result and analyses of the several models.
Furthermore, an extensive dataset regarding the level of institutions has been applied to determine if institutions have an impact on level and variability of inflation. This dataset has been provided by Fraser Institute (2018) for the time period of 1970-2015. This dataset measures the degree to which the policies and the institutions of countries are supportive of economic freedom.
The five pillars of economic freedom are: personal choices, voluntary exchange, freedom to enter markets and compete, and security of the individual and privately owned property. The authors of the index used 42 data points to construct the summary index and to measure the degree of economic freedom in five broad areas: Size of Government (SizeofGovernment), Legal System and Property Rights (LegalSystemPropertyRights), Sound Money (SoundMoney), Freedom to Trade Internationally (Freedomtotradeinternationally) and Regulation (Regulation). A description of the 42 data measures and the five variables are provided in Appendix 7.2 and 7.3. The Economic freedom economic index captures the five pillars into a single index.
However, given that Sound Money is already apprehending standard deviation and inflation mean in its index, this study will exclude the Sound Money to avoid any collinearity. Subsequently, the creation of a new index (Institutions) is needed to proxy the level of institutions. This index captures the essential pillars of the economic freedom summary index, excluding the Sound Money and the Freedom to Trade Internationally which will be analyzed independently. Merging both datasets, several observations have been excluded due to the lack of data in one of the dataset. Besides, given inflation mean, inflation standard deviation and Fraser institute variables cover the five more recent years, numerous observations were excluded in order to prevent data overlaps and will be explained in the following section. An understanding overview of all the data used in the model is proposed in Appendix 7.2 and 7.3. The descriptive statistics regarding the data for the inflation mean and variability is displayed in Appendix 7.4.
3.2 Specification of the measures and variables in the theoretical model
This chapter will depict the data and the source in greater detail. Furthermore, it will describe which variables are used in the model and what are the sources of the data. A complete overview of all the variables is provided in the Table 1 below and in Appendix 7.3 and 7.4.
Table 1: Summary Statistics
(1) (2) (3) (4) (5)
VARIABLES N mean sd min max
Inflation 1,183 41.41 473.1 -4.008 11,750 CrudeOilprice 1,305 34.11 21.53 1.210 77.38 SizeofGovernment 1,254 5.598 1.987 0 9.748 LegalSystemPropertyRights 1,088 5.202 1.926 0.956 9.278 Freedomtotradeinternationally 1,215 5.859 2.483 0 9.970 Regulation 1,142 6.083 1.646 0 9.430 Country 1,305 86.43 48.12 2 172 GDPpercapita 1,241 7.724 13.13 0.0586 105.0 Institutions 1,036 5.834 1.084 2.373 8.840 Inflation_sd 1,208 47.52 547.8 0.00178 13,696 Inflation_mean (lagged) 1,188 40.58 366.5 -2.987 8,504 Inflation_mean 1,218 39.03 353.1 -3.016 8,603 varCrudeOilPrice_sd 1,298 10.89 8.486 3.250 26.86 Number of Countries 151 151 151 151 151
3.2.1 Inflation mean and Inflation variability
Thus, the following equation is obtained: Inflation_mean i,t = 1 𝑛∑Inflationit−n 5 𝑖=1 where : ▪ i – Country ▪ t – Year
▪ n – Total number of years
The second component will measure the standard deviation of the inflation rate over the five more recent years and is displayed below:
Inflation_sdi,k = √∑ (Inflation𝑖,𝑘−Inflation_mean𝑖 )
2 5 𝑘=−5 5 where : ▪ i – Country ▪ k – Year
▪ n – Total number of years
Therefore, this study collects the annual inflation mean and standard deviation. In order to avoid overlapping data, the study will only apply the inflation mean and standard deviation over a period of five years. The reasoning line is as follows. Given Inflation_mean year 5 captures the inflation rate from year 0 to year 4 and the Inflation_mean year 6 captures the inflation rate from year 1 to year 5. The Inflation_mean year 5 and 6 will apprehend the identical observations on year 1, 2, 3, and 4. This process is continuing along the years and yields to inconsistent estimators. Similarly, the Inflation_sd will encounter the identical problem.
3.2.2 GDP per capita
The GDP per capita is used in order to measure the level of development of a country. The GDP per capita is obtained from the World Bank Database (2018) in current thousands of US$. The World Bank defines GDP per capita as the gross domestic product divided by midyear population. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. The use of GDP per capita to determine the level of development is straightforward because it is used for the World bank to categorized countries the by level of development.
Several scholars found the necessity to categorize countries to study the relationship with inflation and inflation volatility (Cukierman, Webb and Neyapti, 1992). However, categorizing countries by group of level of development will limit the analysis and the result. First, due to internal heteroscedasticity within the group. Second, because the levels of development are not immutable. Indeed, some countries have distinguished growth rate per capita and can outreach each other. For instance, South Korea had a lower GDP per capita than Mexico before 1970. This is not the case anymore. Categorizing South Korea in the same group of country for 40 years will be irrelevant to better understand the inflation nexus. Therefore, a static categorization of country to apprehend a dynamic variable will limit the analysis. Third, taking into account the individual GDP per capita instead of a group of country will augment the consistency of the estimators.
Nevertheless, given that the Inflation mean has only results in every five years, the inflation mean and, thus, the GDP per capita will be lagged of five years. Consequently, the number of observations has been reduced.
3.2.3 Crude Oil Price
To determine the effect of external energy and specifically the oil price on inflation level and variability, the oil crude price will be included. Even though this is an imperfect estimator, oil crude price significantly captures the effect of other fossil energies and has a substantial correlation with inflation (IMF, 2006; Kilian, 2009; Álvarez et al., 2011; Busse, Bernhard and Ihle, 2011). This variable acquired from the Saint Louis Bank database (2018) is straightforward; it measures the annual crude oil price level at the current USD. One variable will be created from the world crude oil price: varCrudeOilPrice_sd
varCrueOilPrice_mean measure the weighted average crude oil price over the five more recent years. One can obtain the equation below:
varCrudeOilPrice_mean i,t = 1
𝑛∑ Crude oilpriceit−n
5
𝑖=1
where :
▪ i – Country ▪ t – Year
Whereas varCrudeOilPrice_sd quantifies the standard deviation of the crude oil price over the five more recent years. One can acquire the following equation:
Inflation_sdi,k = √ ∑5𝑘=−5(CrudeOilprice𝑖,𝑘−varCrudeOilPrice_mean𝑖,𝑘 )2 5 where : ▪ i – Country ▪ k – Year
▪ n – Total number of years
One can notice the absence of lags on this variable. Indeed, to investigate the principal political, institutional and economic determinants of inflation volatility across countries and time, Aisen and Veiga (2008) estimated panel data models for standard deviations of inflation for consecutive 3-year periods. They found that the lagged dependent variable is not statistically significant, which means that inflation volatility is not persistent along three-year periods. Moreover, Aisen and Veiga (2008) confirmed that the prominent effect of oil price on national inflation is significantly short-term. Moreover, Choi et al. (2018) demonstrated that the effect on oil price was immediate and vanished after two years. Then, a static panel data model can be estimated without problems of inconsistency. As a result, the model will assume that the impact of oil crude price on higher inflation average is strictly immediate.
Apprehending the five more recent years of crude oil price, the research will take the identical time frame structure as inflation to avoid overlapping data. Moreover, given the oil price is a world estimator, the national inflation effect on the world oil crude price is negligible. The endogeneity effect will be limited.
In order to determine whether the institutions have an influence on inflation level and variability, an extensive dataset with information regarding the level of institutions is used. The three following sub-sections will describe them in greater detail. A complete description of the Fraser institute variables and their component is displayed appendix 7.3
3.2.4 Size of Government
3.2.5 Legal system and property rights
Fraser Institute (2018) gives the definition of LegalSystemPropertyRights as a variable reflecting the legal framework that allows individuals to accumulate private property freely, secured by clear laws that the government enforces effectively. It computes indicators which assess judicial independence, impartial courts, protection of property rights, military interference in rule of law and politics, integrity of the legal system, but also legal enforcement of contracts, regulatory costs of the sale of real property, reliability of police and finally business costs of crime. Protection of persons and their rightfully acquired property is a central element of both economic freedom and civil society. This variable is scaled from 0, inexistent legal system and property rights to 10 performing legal system and property rights.
3.2.6 Regulation
Fraser Institute (2018) defines Regulation as a variable that assesses the regulatory and infra-structure environments constrain and the efficiency regarding operation of business, labor and capital market. According to the institute, governments not only use a number of tools to limit the right to exchange internationally, they may also develop onerous regulations that limit the right to exchange, gain credit, hire or work for whom you wish, or freely operate your business. This variable is scaled from 0, inefficient regulation framework and property rights, to 10 performing regulation framework.
3.2.7 Freedom to trade internationally
In order to determine the impact of the openness of a country on the inflation level and variability, the Freedomtotradeinternationally variable is an essential indicator. According to Fraser Institute, the Freedomtotradeinternationally could be defined as the effects of capital control and people, the black market exchange rates, tariff and nontariff barriers that affect imports and exports of goods and services. In its broader sense, buying, selling, making contracts, and freedom to exchange is essential to economic freedom. The latter diminishes if a country augment trade barriers with other nations. This variable is scaled from 0, close to the world to 10 fundamentally open to the world.
3.2.8 Institutions
The equation is constructed as: InstitutionsI,t =
α1SizeofGovernment𝑖,𝑡 + α2 LegalSystemPropertyRights𝑖,𝑡 + α3Regulation𝑖,𝑡 3
where :
▪ α – Weighted coefficient of the parameter ▪ i – Country
▪ t - Year
This variable is scaled from 0 constraining policies and institutions to economic freedom to 10 supportive policies and institutions to economic freedom.
3.2.9 Inter-analysis
All Fraser institute variables (2018) should be analyzed cautiously. Indeed, these variables are relatively highly correlated. Thus, the estimations of Institutions and other variables of Fraser Institute should be analyzed independently in order to test the hypotheses posed in this thesis. The tests will give extended insights. The results and the analyses will be covered by chapter 4.
For some independent variables, the number of observations is lower for other. This difference is due to the following reason. The World Bank Database (2018) is more furnished than the database from Fraser Institute (2018). In addition, the country-level data is only available from latest years for two prominent reasons: first, some countries did not exist beforehand, second for some countries data was available only from recent years.
3.3 Testing specifications
Using panel data, it is essential to establish whether to use random or fixed effect with a Hausman test. The test performed on the model indicates heteroskedasticity. Thus the estimated coefficients are different, and some random effect estimates will be inconsistent. Moreover, some characteristics of countries persist over time, such as climate or religion. These specifications are unobserved effects. Consequently, correlation with variables is expecting and could have an influence on the analysis. Therefore, all the models will be controlled for country specifics. Furthermore, year fixed effects capture the influence of aggregate inflation trends. Models A, C, and D will be tested with year fixed effects. The table could be found in Appendix 7.7. Lastly, given the world crude oil price does not differ between countries, year fixed effects could not be included in model B.
performed and are presented in appendix 7.5. A multicollinearity presentation is displayed in Appendix 7.6. As expected, the Economic Freedom Index indicators are relatively highly correlated with each other without being higher than 0.70. Nevertheless, it is decided to test these estimates individually. Moreover, possessing a certain degree of correlation with the residuals in the regression model, the estimated models will use the generalized least squares technique in order to avoid statistical inefficient results.
The high correlation between not lagged inflation mean and standard deviation stands out. This study needs to highlight the high VIF of every model of interaction effect with their estimators. While collinearity is a problem to analyze, high correlation is expected with interaction effects. Thus, such as control variables, interaction variables are not problematic if they are used in the model. The other estimators would not be impacted, and the result would not be undermined. In this model, the VIF of the variables are all relatively low and could undoubtely be analyzed. 42 countries and several periods of other states have been depleted due to lack of data. This omission could have a substantial endogeneity effect. Indeed, the absence of observations is often from developing countries with poor performing institutions or countries that had multilevel crisis and hyperinflation and were not able to estimate their economics variables. As a result, the data possessed an extensive data of low inflation developed countries and a lack of data of developing countries with high or hyperinflation. Thus, it is expected that the results will have reduced coefficient, knowing that omitted observations are from low GDP per capita and weak institutions in countries with often hyperinflation. In this research, the number of countries and year had been maximized in order to limit the effect of data omission.
However, Model B will differ. First, it is preferable to analyze the oil crude price regarding its effect are immediate and vanish after two years (IMF, 2006). Second, the reversed causality argument is limited. Indeed, the national variability of inflation has low effect on the worldwide oil price variability. Therefore, it is anticipated that this source of reversed causality does not have a significant effect and diminishes the possibility of endogeneity. Subsequently, this study will use non-lagged inflation mean and a non-lagged variability of oil crude price to better determine if the variability of crude oil price has an impact on the inflation level variability nexus.
Nonetheless, section 3.4 will present the econometric models performed in order to test the four hypotheses previously posed.
3.4 Econometrics models
Established on precedent researches (Foster, 1978; Hyeon and Chin, 2013; Eisenstat and Strachan, 2016), the first model (Model A) is constructed in order to test whether the level of development (HA) affects the inflation mean and variability association.
Interactions effect between GDP per capita with Inflation Mean will be created in order to precisely capture the effect of GDP per capita. This is important because several previous literatures underlined the interaction effect between GDP per capita and inflation even if the presence of lag will limit the reverse endogeneity effect (Koulakiotis, Lyroudi and Papasyriopoulos, 2012; Barro, 2013). In order to test hypothesis A, a test for significance of β4 will be completed.
In the following models, β1 capture the intercept. β2 measures the effect of inflation on inflation variability. β3 determine the effect of the tested variable on the inflation standard deviation. Lastly, β4 determine the effect of the tested variable on the inflation level variability nexus.
(A) Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 GDPpercapitai,t-5 +
β4 GDPpercapitat-5*Inflation_meani,t-5 + εi,t where :
▪ βn – Parameters for intercept/slope ▪ i – Country
▪ t - Year
▪ εi,t - the error term
reversed causality in this model is limited. Thus, non-lagged inflation mean and crude oil price standard deviation will be used to improve the precision of the results. Model B is built to quantify the effect of oil crude price variability on the inflation mean and variability relationship. Thus, β4 significance will be assessed in order to test hypothesis B
(B) Inflation_sdi, t = β1 + β2 Inflation_meani,t + β3 varCrudeOilPrice_sdi,t +
β4 varCrudeOilPrice_sdi,t *Inflation_meani,t + εi,t where :
▪ βn – Parameters for intercept/slope ▪ i – Country
▪ t – Year
▪ εi,t - the error term
In these additional specifications, the Fraser Institute data will be used to test hypothesis C. According to Fraser Institute and confirmed by calculating the VIF and the correlation matrix, appendix 7.5 and 7.6. The four indicators of Economic Freedom Summary index are relatively highly correlated. Thus their analysis should be careful and be estimated independently. This study will employ four approaches to test, firstly, the effect of Institutions and secondly the independent effect of the three institutional variables as follows: Size of Government, Legal System Property Rights, and Regulation. The identical lag will be implemented to determine the impact of the dependent variables on the inflation mean variability association in order to limit reversed causality effect. To test hypothesis C, a test of β4 significance will be conducted.
(C) Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 Institutionsi,t-5
+ β4 Inflation_meani,t-5 * Institutionsi,t-5 + εi,t
Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 SizeofGovernmenti,t-5
+ β4 Inflation_meani,t-5 * SizeofGovernmenti,t-5 + εi,t
Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 LegalSystemPropertyRightsi,t-5
+ β4 Inflation_meani,t-5 * LegalSystemPropertyRightsi,t-5 + εi,t
Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 Regulationi,t-5
+ β4 Inflation_meani,t-5 * Regulationi,t-5 + εi,t
where :
▪ βn – Parameters for intercept/slope ▪ i – Country
▪ t – Year
In a similar vein, this research uses model (D) to test the effect of the trade openness of countries (HD). A five-year lag will be applied for the reason mentioned before. In order to test hypothesis D, a test for significance of β4 will be completed.
(D) Inflation_sdi, t = β1 + β2 Inflation_meani,t-5 + β3 Freedomtotradeinternationallyi,t-5
+ β4 Inflation_meani,t-5 * Freedomtotradeinternationallyi,t-5 + εi,t
where :
▪ βn – Parameters for intercept/slope ▪ i – Country
▪ t – Year
▪ εi,t - the error term
The four models are relevant and innovative to answer the research question. Indeed, the level of development, an energy price indicator the efficiency of institutions, and the degree of trade openness will be captured in a dynamic and linear structure with a large sample of countries and time period in order to measure more precisely the impact of these subsidiary economic and financial conditions on the inflation mean and inflation variability relationship. All measurements of variables are described in the next chapter and defined in Appendix 7.2 and 7.3.
4. Empirical results and analyses
This chapter will present the results of the estimating models described in the prior chapters. Furthermore, this section will study and construct the potential implications. Firstly, it will examine the results of models A, B, C, and D respectively.
4.1 Model A estimations
The first section will discuss the empirical results of GDP on Inflation variability. These lagged fixed effects estimations are performed in order to test the first hypothesis A, which states that GDP negatively affects the inflation nexus.
development are insignificant. First, no statistically significant evidence for the occurrence of Inflation mean seems to exist. Second, no evidence was established in favor of a GDP per capita influence hypothesis (HA), which predicts a negative association between GDP per capita effect on the inflation mean variability nexus. This result could be due to the fact that the inflation mean and GDP per capita have a lag. Overall, the results do not provide enough evidence to reject the null hypothesis of model A, which supposes a negative relationship of GDP on the inflation nexus.
Table 2: Estimations result of equation A with Inflation_sd as the dependent variable
(1) (2) (3) (4)
VARIABLES Model A1 Model A2 Model A3 Model A4
Inflation_mean 0.122 0.134 0.0511 (0.178) (0.189) (0.197) GDPpercapita -0.163 -0.291 -0.420 (1.537) (1.505) (1.441) I_mean*GDPpercapita 0.138 (0.138) Constant 0.156 -4.667 -7.936 -8.117 (24.02) (30.30) (27.36) (27.53) Observations 1,180 1,171 1,146 1,146 Adj-R-squared 0.020 0.014 0.020 0.024 Number of Countries 151 151 151 151
Note: Dependent variable is the inflation standard deviation. It measures the standard deviation of the inflation rate over the five more recent years where inflation represents the growth rate of the consumer price index.
Robust standard errors in parentheses. Inflation mean is lagged.
Level of significance: *** p<0.01, ** p<0.05, * p<0.1 Country and time fixed effects.
Corrected robust standard errors for heteroscedasticity.
However, no definitive conclusion could be made yet, since the results do not cover the influence of institutions, the fossil price energy and openness to the world effect. Further estimations results are necessary and will be examined in the ensuing subsections.
4.2 Model B estimations
significant positive association of inflation mean on inflation standard deviation at 1% significance level. As previously discussed, these results confirmed the key interaction between the two variables. These coefficients of models B2 and B4 present the insignificant effect of crude oil price standard deviation and its interaction effect on the dependent variable. While controlled by inflation mean, the crude oil price standard deviation is statically significant at 1% significance level. Finally, column 4 indicates no decisive significance. Overall, the estimation validates hypothesis B that established the existence of a crude oil price variability affecting the variability of inflation, without affecting the nexus.
Table 3: Estimations result of equation B with Inflation_sd as the dependent variable
(1) (2) (3) (4)
VARIABLES Model B1 Model B2 Model B3 Model B4
Inflation_mean 1.533*** 1.533*** 1.440*** (0.0552) (0.0545) (0.109) varCrudeOilPrice_sd -0.543 0.841*** 0.577*** (2.272) (0.281) (0.163) I_mean*varCrudeOilPrice_sd 0.00704 (0.00478) Constant -12.60*** 53.57** -21.99*** -17.80*** (2.165) (25.33) (3.120) (3.345) Observations 1,208 1,208 1,208 1,208 Adj- R-squared 0.973 0.000 0.973 0.975 Number of Countries 151 151 151 151
Note: Dependent variable is the inflation standard deviation. It measures the standard deviation inflation over the five more recent years where inflation represents the growth rate of the consumer price index. Robust standard errors in parentheses.
Inflation mean is not lagged.
Level of significance: *** p<0.01, ** p<0.05, * p<0.1 Country fixed effect.
Corrected robust standard errors for heteroscedasticity.
4.3 Model C estimations
The third section will discuss the estimations results on Inflation mean and variability. Particularly, this part will analyze the institutions impacts of the dependent variable. This model is performed to test the hypothesis C posed in a preceding chapter. Results of these fixed effects estimations are revealed in table 4 beneath.
In table 4, column 1 indicates the effect of inflation mean on the variability of inflation. Then, the table is separated into four group of coefficients for four indicators: Institutions models C2-C4, Size of Government C5-C7, Legal System and Property Rights C8-C10, and Regulations C11-C13. In each part, the first column describes the linear effect of the coefficient on the variability of inflation, the second column includes the lagged inflation mean, and finally, the third column adds the interaction effect.
These results appear to be indicative of the existence of institutional effect on inflation variability. In the following part, the coefficient for all three institutions indicators will be discussed. First of all, one can notice different magnitudes of the linear models. Institutions are negatively significant at 10 % significance level, and Regulation is negative and significant at 5% significance level.
Column 2 and column 11 signal the positive association between more performing institutions, regulations and lower variability of inflation. Models C5, C6, and C7 indicate no negative effect of government-controlled on the inflation nexus. These coefficient estimations could not validate prior literature stating the interest of inflation surprise for highly controlled governments (De Grauwe, 1996).
Additionally, institutional coefficients C10 and C13 are both negative with 10% significance level. These two models underline the adverse effect of the Legal System and Property Rights, the Regulation on inflation variability, controlled by inflation mean and the interaction effects. Finally, in column 13 one can notice that the inflation mean is statistically significant at 1% significance level if the model includes the interaction effect with Regulations column 13. Moreover, the same column estimations reveal a fundamental finding. Model C13 denotes a significantly adverse interaction effect on inflation nexus at 1% significance level. Β4 endorses researchers supporting that the effect of inflation mean on inflation variability is higher if countries are substantially free from regulatory and environmental constraint. However, these results furnish an imperfect convincing evidence of hypotheses C.
4.4 Model D estimations
Finally, table 4 displays the estimation results of equation D. These lagged fixed effect estimations are performed in order to test the final hypothesis D. This model measures the effect of openness of international trade on the inflation mean variability nexus.
Model D2 denotes a significantly negative effect of freedom to trade internationally on inflation variability at 10% significance. Moreover, column 4, inflation mean coefficient indicates a significant positive effect of inflation mean after including the interaction effect with Freedom to trade internationally. Finally, the key finding remains the interaction effect coefficient which is significantly negative at 1% significance level. In highly trade open nations, inflation means have a lower effect on the inflation variability. Overall, these results are in line with hypothesis D.
Table 5 Estimations result of equation D with Inflation_sd as the dependent variable
(1) (2) (3) (4)
VARIABLES Model D1 Model D2 Model D3 Model D4
Inflation_mean 0.122 0.101 1.576** (0.178) (0.187) (0.620) Freedomtotradeinternationally -31.64* -27.89 -16.83 (18.88) (18.55) (17.99) I_mean* Freedomtotradeinternationally -0.282*** (0.102) Constant 0.156 147.2* 137.8 82.12 (24.02) (82.11) (84.02) (85.58) Observations 1,180 1,149 1,126 1,126 Adj-R-squared 0.020 0.021 0.023 0.046 Number of Countries 151 151 151 151
Note: Dependent variable is the inflation standard deviation. It measures the standard deviation inflation over the five more recent years where inflation represents the growth rate of the consumer price index.
Robust standard errors in parentheses. Inflation mean is not lagged.
Level of significance: *** p<0.01, ** p<0.05, * p<0.1 Country and time fixed effect.
Corrected robust standard errors for heteroscedasticity.
4.5 Discussion of results
This section will summarize and examine the principal results displayed beforehand.
of inflation mean on inflation variability seems imperfect in the long term and has only been observed by controlling trade openness and regulation indicators.
Second, the findings regarding the level of development will be reviewed. While academics found an essential key relationship between GDP per capita and the nexus (Logue and Willet, 1976; Foster, 1978; Taylor, 1981), Model A reveals that GDP per capita does not have an influence on the long-term relationship between inflation mean and inflation variability. Model A even shows no evidence that could establish the effect of GDP per capita on inflation variability. Overall, the findings do not support the results performed by former economists that discovered a diminishing effect of the inflation mean variability nexus for highly developed countries (Logue and Willet, 1976; Taylor, 1981). In this model, this result could be due to the presence of a lagged which eliminates the endogeneity issue. It is possible that the GDP per capita has only an immediate effect on the nexus.
Third, the results of the country fixed effect estimation revealed a significant impact of oil crude price on the variability of inflation. Thus, the latter confirms the results of the prior literature (IMF, 2006; Busse, Bernhard and Ihle, 2011; Choi et al., 2018). Moreover, considering the insignificance of the interaction effect, the result validates the hypothesis B. In other words, crude oil price volatility alters the variability of inflation, but does not impact the inflation mean variability relationship. Overall, these results regarding the impact of fossil price energy are in line with the IMF (2006) who established an immediate, exogenous effect of variability of crude oil price on the variability of inflation.
Fourth, the results regarding the institutions effects will be discussed. It can be established that institutions have a significant negative effect on variability of inflation. However, there is not significantly negative evidence that the government control augments the variability of inflation. Subsequently, the findings do not support the line of De Grauwe (1996). His reasoning was as follows. Highly indebted and highly spending countries are more tempted to generate surprise inflation and tend to have higher inflation variability. Furthermore, the results from the specification that test the effect of Legal System and Property Rights confirm the findings of Hayo and Voigt (2008) sustaining that a higher degree of legal system independence as well as trust towards legal system, tend to decrease the inflation through two prominent effects. First by affecting the output and second by increasing the independence of the central bank. Subsequently, model C partially authenticates hypothesis C.
constraints in labor, business or capital market augment the sensitivity and the effect of external shocks. Moreover, Posen (1995), demonstrated the link between inflation mean variability nexus and political independence. This researcher concludes that both are the product of fundamental economic and social interest. The presence of lobbying groups against inflation – such as large financial sectors, affect countries. These nations tend to possess more politically independent central banks, and thus less inflation mean and variability. Conversely, countries with pressure group less opposed to inflation will have less political independence monetary institutions and higher inflation mean and variability. Therefore, Posen underlined diminishing capital market constraints foster financial and banking sectors and could result in a more independent monetary institution. Finally, this process will reduce the inflation mean and variability relationship effect. Overall, the thesis results corroborate previous literature which states that an independent and efficient legal system and performing regulation are negatively correlated with higher inflation volatility (Copelovitch and Singer, 2008; Hayo and Voigt, 2008; Echeverri-Carroll, E. and Ayala, 2009; Jaumotte and Morsy, 2013). Lastly, fewer constraint regulations and business environment reduce the inflation mean variability nexus effect. Finally, this Master Thesis presents evidence of a negative relationship of trade openness in line with Ghosh (2014). It concludes that the more the degree of openness of a country is high, the more the variability of inflation decrease. An explanation of this observation can be found in the diminishing interest to generate surprise inflation. The reasoning is constructed as follows. The more a country is globally integrated, the more an inflation phenomenon will lead to a real depreciation, the more the process will be harmful to the economy. As a result, the benefits of unanticipated inflation and volatility of inflation is declining in a higher level of openness. (Cooke, 2010). Besides, an open country could diversify its economy and reduces economics shocks that could amplify the volatility of inflation (De Grauwe, 1996). Additionally, Hyeon and Chin (2013), claimed that domestic firms might encounter price uncertainty, losing sales and profits to international competitors. Thus, the government and monetary institutions would be less tempted to generate undisciplined policy. Overall, the estimation coefficients support the findings that in highly opened to trade countries, the effect of inflation mean on inflation variability is lower.
All in all, the regulations and the degree of trade openness are both critical variable decreasing the effect of inflation mean on inflation variability.
5. Final notes
5.1 Conclusion
This Master Thesis sets out to determine whether the interaction between inflation level and inflation variability differ with different economic, institutional, and structural conditions. It actually challenges the impending nonlinearity in this fundamental relationship. The study interrogations are established in political and monetary economics theory. Several conclusions can be drawn from the study.
This research used data from the World Bank and Fraser Institute in order to test the hypotheses. The empirical test used robust standard errors, country and time fixed effects.
In order to respond to these research interrogations, several variables such as Institutions or Inflation mean have been estimated. Four hypotheses have been posed. First, this Master Thesis examined whether GDP per capita has a negative effect on the relationship between inflation level and variability of inflation. Second, the variability of the world crude oil price is affecting inflation variability but is not associated with the relationship between level and variability of inflation. Third, performing institutions have a negative effect on the association between level and variability of inflation. Lastly, the study investigated whether the degree of openness negatively affects the association between level and variability of inflation.
Based on panel data analysis for the period between 1965 and 2015, this research sustains the negative association between GDP per capita and inflation level and in less significant measure on variability. Moreover, the study highlights the partial positive effect of crude oil price on inflation level and variability. Regarding the institutions, the study provides evidence concerning the importance of oil crude price on the variability of inflation. Additionally, this Master Thesis concludes that both, regulations and freedom to trade internationally have a significant adverse effect on the inflation nexus. Thus, policies could first take the form of an open trade agreement. It could also reinforce central bank independence, institutions, and improve the regulations, by reducing constraint on labor, capital and business market.
Final notes need to be evoked regarding the policy implication of this research. It was established that oil crude price level has no significance on the nexus. Thus, it could be advised to eliminate the impact of fossil energy in the target inflation level. Including the exogenous energy price will distort the monetary policy without improving the state of the economy and with low results given the effect vanishes within two years (IMF, 2006). This is a key notion because most central banks could, therefore, overreact to energy inflation shock, causing an economic recession. Secondly, the Master Thesis revealed the adverse effect of legal system property rights and regulations on the level of variability and the nexus. After examining the findings, one can suggest to improve the legal system, simplify the labor, capital and business regulation in order to reduce the adverse effect of the inflation relationship. Moreover, the study advocates a greater openness of trade such as the NAFTA or Schengen area in order to limit the negative effect of inflation mean variability association.
Furthermore, the Master Thesis demonstrates that the relationship between inflation mean and inflation volatility is not immutable.
Moreover, the global financial crisis highlights the concept of 4% significance level of inflation target and the necessity to reconsider higher inflation target, its costs and advantages. This study partially answers the questions posed by the political and monetary economic theory.
5.2 Further discussions
First of all, the study focuses on the variability of inflation as the standard deviation of inflation, but future research might investigate whether standard deviation is the "Ideal" measure of variability. Indeed, to calculate the variability, (Foster, 1978), the standard error could be an interesting indicator of variability of the inflation if the sample has an independent and random variable.
Second, the study focuses on institutions, level of development, openness to trade and crude oil price. However future research might investigate and find other parameters that could have an impact on inflation level and variability.
6. References
Aisen, A. and Veiga, F. J. (2008) ‘Political instability and inflation volatility’, Public Choice. doi: 10.1007/s11127-007-9254-x.
Akhtar Hossain, A. and Arwatchanakarn, P. (2016) ‘Inflation and inflation volatility in
Thailand’, Applied Economics. doi:
10.1080/00036846.2015.1130215doi.org/10.1080/00036846.2015.1130215.
Alpanda, S. and Honig, A. (2010) ‘Political monetary cycles and a de facto ranking of central
bank independence’, Journal of International Money and Finance. doi:
10.1016/j.jimonfin.2010.02.001.
Álvarez, L. J. et al. (2011) ‘The impact of oil price changes on Spanish and euro area consumer price inflation’, Economic Modelling. doi: 10.1016/j.econmod.2010.08.006.
Ball, L. (2014) ‘The Case for a Long-Run Inflation Target of Four Percent’, IMF Working Papers. doi: 10.5089/9781498395601.001.
Barro, R. J. (2013) ‘Inflation and economic growth’, Annals of Economics and Finance. doi: 10.1086/450067.
Bernanke, B. S. (2010) The Economic Outlook and Monetary Policy.
Blanchard, O. (2016) ‘The Phillips curve: Back to the ’60s?’, in American Economic Review. doi: 10.1257/aer.p20161003.
Bomberger, W. and Makinen, G. (1993) ‘Inflation and Relative Price Variability : Parks’ Study Reexamined’, Journal of Money, Credit, and Banking, 25(4), pp. 854–861.
Bowdler, C. and Malik, A. (2017) ‘Openness and inflation volatility: Panel data evidence’, North American Journal of Economics and Finance. doi: 10.1016/j.najef.2017.03.008.
Bruno, M. and Easterly, W. (1998) ‘Inflation crises and long-run growth’, Journal of Monetary Economics. doi: 10.1016/S0304-3932(97)00063-9.
Busse, S., Bernhard, B. and Ihle, R. (2011) ‘Emerging linkages between price volatilities in energy and agricultural markets’, in An analysis of price and volatility transmission in butter, palm oil and crude oil markets.
Chinn, M. and Leblanc, M. (2004) ‘Do High Oil Prices Presage Inflation ?’, Business Economics, 39(2), p. 38. doi: 4aabcd045a8ec14c0d08ecd330968821.
Chiu, J. and Molico, M. (2011) ‘Uncertainty, Inflation, and Welfare’, Journal of Money, Credit and Banking, 43(7).
Choi, S. et al. (2018) ‘Oil prices and inflation dynamics: Evidence from advanced and developing economies’, Journal of International Money and Finance. doi: 10.1016/j.jimonfin.2017.12.004.
Cooke, D. (2010) ‘Openness and Inflation’, Journal of Money, Credit and Banking. doi: 10.1111/j.1538-4616.2009.00287.x.
Cukierman, A., Webb, S. B. and Neyapti, B. (1992) ‘Measuring the independence of Central Banks and Its Effects on Policy Outcomes’, The World Bank Economic Review. doi: 10.1093/wber/6.3.353.
Daal, E., Naka, A. and Sanchez, B. (2005) ‘Re-examining inflation and inflation uncertainty in developed and emerging countries’, Economics Letters. doi: 10.1016/j.econlet.2005.05.024. Dreher, A., Sturm, J. E. and de Haan, J. (2008) ‘Does high inflation cause central bankers to lose their job? Evidence based on a new data set’, European Journal of Political Economy. doi: 10.1016/j.ejpoleco.2008.04.001.
Echeverri-Carroll, E. and Ayala, S. (2009) ‘Regulation and American business’, Policy Review. doi: 216434604.
Eisenstat, E. and Strachan, R. W. (2016) ‘Modelling Inflation Volatility’, Journal of Applied Econometrics. doi: 10.1002/jae.2469.
Fischer, S., Sahay, R. and Végh, C. A. (2002) ‘Modern Hyper- and High Inflations’, Journal of Economic Literature, 40(3), pp. 837–880. doi: 10.1257/002205102760273805.
Foster, E. (1978) ‘The Variability of Inflation’, The Review of Economics and Statistics, 60(3), pp. 346–350.
Fountas, S. and Karanasos, M. (2007) ‘Inflation, output growth, and nominal and real uncertainty: Empirical evidence for the G7’, Journal of International Money and Finance. doi: 10.1016/j.jimonfin.2006.10.006.
Fraser Institute (2018) Economic Freedom Summary Index.
Friedman, M. (1977) ‘Nobel Lecture: Inflation and Unemployment’, Journal of Political Economy. doi: 10.1086/260579.
Ghosh, A. (2014) ‘How do openness and exchange-rate regimes affect inflation?’, International Review of Economics and Finance. doi: 10.1016/j.iref.2014.08.008.
Gordon, R. J. (1971) ‘Steady Anticipated Inflation: Mirage or Oasis?’, Brookings Papers on Economic Activity.
De Grauwe, P. (1996) ‘Monetary union and convergence economics’, European Economic Review. doi: 10.1016/0014-2921(95)00117-4.
Grier, K. B. and Perry, M. J. (2000) ‘The effects of real and nominal uncertainty on inflation and output growth: Some GARCH-M edivence’, Journal of Applied Econometrics. doi: 10.1002/(SICI)1099-1255(200001/02)15:1<45::AID-JAE542>3.0.CO;2-K.
Hayo, B. and Voigt, S. (2008) ‘Inflation, Central Bank Independence, and the Legal System’, Journal of Institutional and Theoretical Economics JITE. doi: 10.1628/093245608786534578. Hyeon, D. and Chin, S. (2013) ‘Inflation and Inflation Volatility Revisited’, International Finance, (1977), pp. 327–345. doi: 10.1111/j.1468-2362.2012.12001.x.
IMF (2006) ‘Oil Prices and Global Imbalances’, in World Economic Outlook, Globalization and Inflation.
Jha, R. and Dang, T. (2012) ‘Inflation variability and the relationship between inflation and growth’, Macroeconomics and Finance in Emerging Market Economies. Available at: https://crawford.anu.edu.au/acde/asarc/pdf/papers/2011/WP2011_08.pdf (Accessed: 13 May 2018).
Jiranyakul, K. and Opiela, T. P. (2010) ‘Inflation and inflation uncertainty in the ASEAN-5 economies’, Journal of Asian Economics. doi: 10.1016/j.asieco.2009.09.007.
Kilian, L. (2009) ‘Oil Price Shocks , Monetary Policy and Stagflation’, Inflation in an Era of Relative Price Shocks.
Klomp, J. and de Haan, J. (2010) ‘Inflation and central bank independence: A meta-regression analysis’, Journal of Economic Surveys, 24(4), pp. 593–621. doi: 10.1111/j.1467-6419.2009.00597.x.
Koulakiotis, A., Lyroudi, K. and Papasyriopoulos, N. (2012) ‘Inflation, GDP and Causality for European Countries’, International Advances in Economic Research. doi: 10.1007/s11294-011-9340-1.
Krugman, P. (2013) ‘The Four Percent Solution’, The New York Times.
Logue, D. and Willet, T. (1976) ‘A Note on the Relation between the Rate and Variability of Inflation’, Economica, New Series, 43(170), pp. 151–158.
Mishkin, F. S. (2010) ‘Monetary Policy Strategy : Lessons from the Crisis’. doi: 10.1037/0022-0663.92.1.191.
Okun, A. M. (1971) ‘The Mirage Inflation of Steady’, Brookings Papers on Economic Activity, 1971(May 1960), pp. 485–498.
Posen, A. S. (1995) ‘Declarations Are Not Enough: Financial Sector Sources of Central Bank Independence’, NBER Macroeconomics, 10(January), pp. 253–274. doi: 10.1086/654279. Rogoff, K. (2003) ‘Globalization and global disinflation’, Economic Review-Federal Reserve Bank of Kansas City.
Taylor, J. B. (1981) ‘On the relation between the variability of inflation and the average inflation rate’, Carnegie-Rochester Confer. Series on Public Policy, pp. 57–85. doi: 10.1016/0167-2231(81)90019-1.
The Fed (2017) Yellen’s Press Conference. Available at:
7. Appendix
7.1 Formal List of Country
The Appendix reports all the countries present in the model
Albania Algeria Angola Argentina Armenia Australia Austria Azerbaijan Bahamas, The Bahrain Bangladesh Belgium Belize Benin Bhutan Bolivia Bosnia and Herzegovina Botswana Brazil Brunei Darussalam Bulgaria Burkina Faso Burundi Czech Republic Denmark Dominican Republic Ecuador
Egypt, Arab Rep. El Salvador Estonia Ethiopia Fiji Finland France Gabon Gambia, The Georgia Germany Ghana Greece Guatemala Guinea-Bissau Guyana Haiti Honduras Kuwait Kyrgyz Republic Lao PDR Latvia Lesotho Lithuania Luxembourg Macedonia, FYR Madagascar Malawi Malaysia Mali Malta Mauritania Mauritius Mexico Moldova Mongolia Morocco Mozambique Myanmar Namibia Nepal Qatar Romania Russian Federation Rwanda Saudi Arabia Senegal Serbia Seychelles Sierra Leone Singapore Slovak Republic Slovenia South Africa Spain Sri Lanka Suriname Swaziland Sweden Switzerland
Syrian Arab Republic Tajikistan
Cabo Verde Cambodia Cameroon Canada Central African Republic Chad Chile China Colombia
Congo, Dem. Rep. Congo, Rep. Costa Rica Cote d'Ivoire Croatia Cyprus
Hong Kong SAR, China
Hungary Iceland India Indonesia
Iran, Islamic Rep. Ireland Israel Italy Jamaica Japan Jordan Kazakhstan Kenya Korea, Rep. Netherlands, The New Zealand Nicaragua Niger Nigeria Norway Oman Pakistan Panama
Papua New Guinea Paraguay Peru Philippines Poland Portugal Timor-Leste Togo
7.2 Variable descriptions
The Appendix describes all the variables used in the study
Variable Description Source
Inflation
Inflation, consumer prices (annual %)
Inflation as measured by the consumer price index reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed at specified intervals, such as yearly. The Laspeyres formula is generally used.
World Bank
Inflation_mean
Weighted average Inflation:
The component measures the weighted average inflation over the five more recent years
Inflation_meani =–1 𝑛∑ α𝑖Inflationit−n 5 𝑖=1 where : ▪ i – Country ▪ t – Year
▪ n – Total number of years
Created
Inflation_sd
Inflation Standard deviation:
The component measures the standard deviation of the inflation rate over the five more recent years.
Inflation_sdi,k = √ ∑5 (Inflation𝑖,𝑘−Inflation_mean𝑖,𝑘 )2 𝑘=−5 5 where : ▪ i – Country ▪ k – Year
▪ n – Total number of years
