De Grauwe and Polan - 2005 - Is Inflation Always and Everywhere a Monetary Phen.pdf

Is Inflation Always and Everywhere a Monetary Phenomenon? Author(s): Paul De Grauwe and Magdalena Polan Source: The Scandinavian Journal of Economics , Jun., 2005, Vol. 107, No. 2 (Jun., 2005), pp. 239-259 Published by: Wiley on behalf of The Scandinavian Journal of Economics Stable URL: https://www...

Is Inflation Always and Everywhere a Monetary Phenomenon? Author(s): Paul De Grauwe and Magdalena Polan Source: The Scandinavian Journal of Economics , Jun., 2005, Vol. 107, No. 2 (Jun., 2005), pp. 239-259 Published by: Wiley on behalf of The Scandinavian Journal of Economics Stable URL: https://www.jstor.org/stable/3441104 JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms Wiley and The Scandinavian Journal of Economics are collaborating with JSTOR to digitize, preserve and extend access to The Scandinavian Journal of Economics This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms Scand. J. of Economics 107(2), 239-259, 2005 DOI: 10.1111/j.1467-9442.2005.00406.x Is Inflation Always and Everywhere a Monetary Phenomenon?* Paul De Grauwe University of Leuven, B-3000 Leuven, Belgium [email protected] Magdalena Polan University of Leuven, B-3000 Leuven, Belgium [email protected] Abstract Using a sample of about 160 countries over the last 30 years, we test for the quantity theory relationship between money and inflation. When analysing the full sample of countries, we find a strong positive relation between long-run inflation and the money growth rate. The relation is not proportional, however. The strong link between inflation and money growth is almost wholly due to the presence of high- (or hyper-) inflation countries in the sample. The relationship between inflation and money growth for low-inflation countries (on average less than 100/ per annum over the last 30 years) is weak. Keywords: Inflation; money; quantity theory of money JEL classification: E40; E50 I. Introduction Is inflation always and everywhere a monetary phenomenon?1 economists today will argue that when analysed over a sufficiently long of time, inflation is indeed everywhere a monetary phenomenon. "monetarist" view has not always been widespread, however. Prior t upsurge of inflation in the 1970s, many economists were not inclined to at the money stock when analysing the sources of the (low) inflation rat that time. In this paper, we return to this issue using a sample of count spanning the whole world over a period of 30 years. The key questio analyse concerns the link between inflation and the growth rate of mon how it depends on whether countries experience low or high rates of inf The view that inflation is always and everywhere a monetary phenome has a long tradition based on the quantity theory of money (QTM). * We are grateful to Steinar Holden and to two anonymous referees for comme suggestions. 1 Friedman (1963) wrote these now famous words, not as a question but in the affirma also Friedman and Schwartz (1963). ? The editors of the Scandinavian Journal of Economics 2005. Published by Blackwell Publishing, 9600 Garsington Oxford, OX4 2DQ, UK and 350 Main Street, Maiden, MA 02148, USA. This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms 240 P. De Grauwe and M. Polan simplest form, the QTM says that changes in money supply growth followed by equal changes in the inflation rate and, through the force of t Fisher effect, in the nominal interest rate. The QTM is a measure of the ext to which inflation movements can be explained by purely monetary forces The starting point of the QTM is the well-known identity: MV= YP, (1) where Mis money supply, Vis the price level. If we move to m+v=y+p,2 (2) where lowercase letters denote growth rate of the price level-can be expresse p=m-y+v. (3) These identities are transforme following two propositions. Fir relation between inflation and of inflation on money growth, Second, over a sufficiently long are orthogonal to the growth r Thus, there prediction are says two that a aspects o permane increase in the rate of inflat proposition-also referred to a permanent increase in the grow unaffected in the long run. If output, it only holds in the sho In this paper, we analyse these proceed is to transform the iden propositions. Since we do not ha velocity in the error term.3 As Pi = Ao + Pmi + i2Yi + /i, (4) where pi, mi and yi are the rate of inflation, the money grow growth of country i, respectively, measured over a sufficien 2 These are, of course, instantaneous rates, not average rates. For low gro not pose a problem; for high growth rates, however, the inflation rate by just adding growth rates of money, velocity and output. 3 We could, of course, use the definitional equation to derive velocity. B very sensible as we would then estimate an identity. ? The editors of the Scandinavian Journal of Economics 2005. This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms Is inflation always and everywhere a monetary phenomenon? 241 time (30 years). The QTM theory then predicts that /3 = 1, /2 < 0, and mi and yi are uncorrelated. We then test these propositions. Note that a potential bias may arise if the independent variables mi and yi are correlated with the error term (velocity). We provide indirect evidence that such a bias might exist. The QTM does not specify which definition of money supply should be used in empirical tests of the theory. There is no theoretical reason why Ml or M2 should be used as the appropriate variable. Accordingly, many authors use both or other monetary aggregates to compare the results obtained for various definitions of money. Since the empirical literature is not consistent in its opinion as to which monetary aggregate is more correlated with the price level, we use both Ml and M2 in our study. II. Review of the Empirical Literature There is a vast empirical literature concerning the long-run relation between money growth and inflation. We begin by briefly describing some of its aspects. This literature can be divided into three groups. The first uses crosssection data on a large number of countries over a long time span. Usually, a long-run average of money supply (or its growth rate) and price level (or the inflation rate) is calculated and used to compute the correlation between the two. All countries are treated equally, and there is no distinction according to monetary or economic regimes. Authors in the second group use long series of higher frequency data (annual or quarterly) referring to only one country to describe a long-run relationship between money and the price level. Sometimes, the results are compared with other single-country findings. The third group takes the shape of a historical investigation, sometimes reaching more than 200 years into the past. These studies often focus on one country only, but they suffer, as do studies of the second type, from the incomparability of the economic systems of a country across centuries. Table 1 gives an overview of the representative articles of the first type of empirical studies, based on cross-sections of countries. The table also describes the data sets and the results. Authors of the articles listed in Table 1 try to either analyse data on the largest possible number of countries or focus on a smaller group of countries with similar economic systems. In the latter case, the results are only applicable to this particular group of countries, while the first method is supposed to yield universal results. In most cases, the relation between money supply and price level is strong and positive. A common finding of these studies is that countries with low money growth (and low inflation) tend to create a horizontal cluster in a plot where inflation (vertical axis) is set against money growth (horizontal axis). ? The editors of the Scandinavian Journal of Economics 2005. This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms v Table 1. Main multi-country studies of long-run relationship between mon &-St~ ~ Monetary High Author, year aggregate Prices Data set Time span d | Vogel (1974) Currency + CPI 16 Latin Amer demand deposits inflation ,0Bg~~~~~~~~~~~~~~~~~~~~ ~~~~~~~ ^i$t~~~~~~~~~~~~~~~~ ~~~~~~~ Dwyer and M2 GDP 62 countries 1979-84, five-ye I Hafer (1988) deflator averages o Barro (1990) Hand-to-hand CPI 83 countries 1950currency "asso Pakko (1994) Currency + CPI 13 former Soviet republics 1992 an bank deposits four-quarter averag Poole (1994) Broad money n.a. All countries in 1970-80 and 1 World Bank tables annual McCandless and MO, M1, M2 CPI 110 countries reported 196 Weber (1995) in IMF IFS c Dwyer and n.a. GDP 79 countries reported 1987-97, t Hafer (1999) deflator in IMF IFS five-year avera Note: "High-Low differentiation" indicates whether the autho This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms Is inflation always and everywhere a monetary phenomenon? 243 However, none of the papers surveyed here has attempted to analyse this phenomenon or study how the level of inflation affects the relation between money growth and inflation. An interesting conclusion can be drawn from the article by Dwyer and Hafer (1999). These authors compare the relation between average money growth and average inflation rate in two periods, 1987-1992 and 1993-1997. In the second period, the average inflation rate (across all countries in the sample) is lower. The reduction in the average inflation rate leads to the creation of two horizontal clusters of observations close to the origin. Thus, the weakening relation between money growth and inflation, as we progress towards zero money growth, may be associated with the average money growth of a country. The second type of empirical study uses single country time-series analysis. Within this class of studies, an initial approach has been to analyse the long-term quantity theory relationship after transforming time series into the frequency domain. Representative papers are Lucas (1980) and Fitzgerald (1999). These studies tend to confirm the proportionality prediction of the quantity theory, although their methodology has been criticised by McCallum (1984) and Rolnick and Weber (1995). McCallum (1984) warns us that associating high-frequency time series with long-run economic propositions is not always warranted. More recently, researchers have adopted another-more satisfactoryapproach in analysing the time-series properties of inflation, output and money. This consists of explicitly testing coefficient restrictions implied by the quantity theory in vector autoregression models. Important papers using this approach are Geweke (1986), Stock and Watson (1988), Boschen and Mills (1995), and King and Watson (1997). These authors confirm the long-run neutrality of money on output for the US economy. Similar results for G7 countries were obtained by Weber (1994). In this context, the empirical studies using the P-star model should also be mentioned. This model, suggested by Hallman, Porter and Small (1991), was further explored by Vega and Trecroci (2002) and Gerlach and Svensson (2004); see also Jansen (2004) for a recent exposition. The P-star model may be regarded as a moder monetarist approach to modelling inflation. It starts by defining the price gap as the difference between the price level and the long-run equilibrium price level, which is implied by the long-run quantity relation. The model then specifies a direct effect from the lagged price gap and the current price level. Studies designed to test the QTM using data on one or a few countries (the second group) often overlap with the third type of studies-very long-term historical analyses of the relation between money and prices, or investigations of this relation over a particular period in the past. One such long historical analysis was carried out by Smith (1988), who explores the relation between money and prices in the British colonies. ? The editors of the Scandinavian Journal of Economics 2005. This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms 244 P. De Grauwe and M. Polan Studies analysing a large set of countries typically do not take differen between countries into account. However, Rolnick and Weber (1995) that such disregard can change the results of estimations. They prove tha strength of the long-run relationship between money and prices differs countries operating under different monetary standards. When com with fiat standards, commodity standards result in lower correlation money growth and inflation, a higher correlation with output growth an lower correlation of various monetary aggregates with each other. Inflat money growth and output growth are generally lower under commo standards than under fiat standards. III. Cross-section Evidence: The Long Run We now turn to tests of the quantity theory using cross-section data on 30-year averages of money growth, inflation and output growth. (Later on, in Section IV, we use panel data to test the quantity theory.) We expect 30 years to be a sufficiently long period to be considered as "long run". Therefore, we assume that our sample of data is sufficiently long to detect the type of relationship predicted by the quantity theory. We begin by presenting the data and then proceed to the regression analysis. The Data To explore the relationship between money growth and inflation, we cho the largest available sample of countries, covering the years 1969-1999. W used the International Financial Statistics of the IMF as the source of our data and tested the theory using two monetary aggregates, Ml and M2. Inflation is measured as a percentage increase in the consumer price index. Not all observations are shown in our graphs; five observations with an average inflation rate above 200% per annum were omitted. Including them would have compressed the remainder of the chart too much. Figure 1 shows the full sample of observations on average annual inflation and money growth rates. As in the studies reviewed above, the observations are clustered around the 45? line. The correlation between average inflation and average M1 growth is 0.877, and 0.89 for the correlation with M2. Thus, the results are very similar to those obtained by Vogel (1974), Dwyer and Hafer (1988, 1999), Barro (1990), Poole (1994) and McCandless and Weber (1995). Note that our sample of countries is larger than the samples used in these studies. Most of the observations are grouped in the lower-left part of the chart, close to the origin. To detect whether the relation between money supply growth and the inflation rate may vary between subsamples, we divided the set of all ? The editors of the Scandinavian Journal of Economics 2005. This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms Is inflation always and everywhere a monetary phenomenon? 245 /UU UVU g 160 5 160- 2c4 a 120. 0 0 o I 80. a) 80 a O. v. a) 40 d o (a c 120 D. 0s,* o IO 40 * ,= 0o n ^' 0 40 80 120 160 200 0 40 80 120 160 200 Average M1 growth (%) Average M2 growth (%) Fig. 1. Inflation and the average growth of Ml and M2, 196 observations into groups in the following way. We s consisting of countries with inflation and money gro we progressively expanded the sample by adding the obs classes, i.e., 10% to 20%, 20% to 30% and so on. A se diagrams is shown in Figures 2-4. It is immediately evide scatter diagrams that the positive relation between inflat seems to become more pronounced as observations of hi are added to the sample. For low-inflation countries (les diagram forms a shapeless, almost horizontal cloud. Thu inflation and money growth obtained for the lowest infl to be quite different from the results for the full samp cross-section analysis which, to our knowledge, has not existing literature, is the focus of our analysis. Cross-section Empirical Analysis Here, we test both the proportionality and the neut propositions of the QTM. We begin by examining t then try to obtain additional insights into the QTM rela different subsamples. Estimation over the Whole Sample We start by estimating the regression equation (4) r average inflation rate to the long-term average mone the long-term growth rate of output (where the long te 4 Some of the time series used in the calculations of averages dif estimated all equations using a sample consisting of time series w The results are very similar to those obtained for the full sample a They can be obtained from the authors on request. ? The editors of the Scandinavian Journal of Ec This content downloaded from 92.202.12.4 on Thu, 02 May 2024 05:30:59 +00:00 All use subject to https://about.jstor.org/terms 246 P. De Grauwe and M. Polan I|. 10 9 Y 8 - a o -9 0 7 6 o 0 ? O 5 co a) a o o 6 0 000 4 4 aD X , ? 3 o o a 0 e 5 4 CU 0 e at : o o 3 2 2 1- 0 1 0 2 4 6 8 1( Average Fig. 2. 3 M1 o and 'E rt o 10 0) e 20 a 10 o 20 o o o? a eg?? ? p - 6 o ' 0 CD 0 102'0 0 5 0 money.? 15 a gr 0 ? at 15 M2 a 14 c 30 25 Ca 10 6 16 Ca o ~~30 --~o o. 0 4 18 o 35 1: 2 20 a 45 Ca 0 Average growth Inflation 50 40 n 0o? C? o0eo e a% 4 $? o '? 2 30 5( 3 40 0 20 15 Average M1 growth 0 5 10 Average M2 growth Fig. 3. Inflation and money supply growth from 0 ;u o 50 a 16 aE - o 8 ca o- A47tX ? e @ 9 0 *o 10 i 0 5 I w 00 o 8 200 20 - oa? 0, o I 4 o 'O , ? to0 "a ' 2 o 0 2 5 0 00