Printer Friendly

Does Inflation Lower Productivity? Time Series Evidence on the Impact of Inflation on Labor Productivity in 12 OECD Nations.


According to proponents of zero-inflation policies, even low rates of inflation create distortions in capital allocation and in price signals, which result in lower rates of productivity growth. This paper tests the hypothesis that inflation has a causal impact (in the Granger sense) on labor productivity growth in manufacturing for 12 countries of the Organization for Economic Cooperation and Development (OECD). In bivariate tests of inflation and productivity and in multivariate tests using controls for cyclical effects, there is no evidence of a consistent relationship between inflation and productivity growth with regard to either sign or magnitude. Therefore, the present analysis does not support the view that further reductions in inflation from already low single-digit levels would have a positive impact on labor productivity growth for major industrial countries. (JEL E31)


The disinflation of the 1980s and 1990s is widely regarded as a macroeconomic policy success story. Many industrial nations reduced inflation from double to low single digits, and some developing countries have seen inflation reductions of three or even four orders of magnitude. Now there are calls to reduce inflation even further, perhaps to zero, m expectations of improved macroeconomic outcomes, especially with regard to growth and productivity. This paper examines whether there is empirical support for the position that reducing already low rates of inflation to zero will lead to increased labor productivity growth and, thereby, increased overall economic growth.

While economists generally agree that high rates of inflation are associated with lower rates of economic growth, there appears to be less agreement on a causal mechanism. Some arguments in literature include the following:

1) Even fully anticipated inflation acts as a tax on money balances and causes agents to incur shoe-leather and menu costs.

2) Unanticipated inflation biases the tax system against capital assets by understating true depreciation costs. Since the bias increases with time, both the level and the composition of capital are likely to be affected, with short-lived assets more heavily favored.

3) Unanticipated inflation tends to generate price uncertainty, distorting relative price signals and increasing decision-making errors.

4) Uncertainty about prices can induce firms to increase inventories of buffer stocks and reduce expenditures on long-term basic research [Feldstein, 1982; Fischer, 1986; Briault, 1995; Thornton, 1996].

Advocates of low-inflation policies point to these distortions as rationale for price stability, and their arguments have been forceful enough to persuade at least one congressman to introduce a resolution in the U.S. House of Representatives, supported by four Federal Reserve Bank presidents, instructing the Federal Reserve to lower the inflation rate to zero over a five-year period. [1]

It is not clear, however, that reductions from already low rates of inflation will add to productivity and economic growth. The empirical evidence of a consistent relationship between inflation and productivity is inconclusive. Cross-section studies tend to support the position that high inflation (usually something greater than 2 percent per month) has adverse consequences on economic growth and stability [Bruno, 1995], and the coincidence of high inflation and the decline in productivity growth in the 1970s has been adduced as evidence of the negative effects of inflation. In the time series realm, most of the relevant research demonstrates a negative correlation between inflation and productivity, but the results are sensitive to issues of causality and to the presence of other economic variables. The range of inflation experienced by the sample countries may not have been large enough to produce conclusive evidence of reductions in productivity growth or observed statistical correlations between inflation and productivity growth may have been spurious, a byproduct of cyclical comovements with other macroeconomic variables.

This paper extends the existing time series-based empirical research on the relationship between inflation and productivity in several ways. Unlike previous studies that test for correlation and causality in a bivariate framework, the present analysis includes cyclical variables to test for spurious comovements between inflation and productivity growth, using a broader set of industrialized nations (12 countries) than any prior study of which the authors are aware. The range of average inflation experienced by these countries varies from less than 3 percent to more than 8 percent over the 1961-94 period, and from 5 percent to more than 15 percent over the inflationary 1973-82 subperiod. So the impact of inflation will be tested under a variety of inflationary environments. Because the time series in this paper begins in the late 1950s to early 1960s and extends to the low-inflation 1990s, the analysis captures both sides of the 1970s to early 1980s inflation spike experienced in most industrial nations.

In addition to the broader country coverage and the extended time series, the results from the present study, which employs causality tests using the variable lag-length technique suggested by Hsiao [1981, 1982] to examine the inflation-productivity nexus, may be more general than those of prior research. Allowing lag lengths in causality tests to vary across included variables conserves degrees of freedom and potentially increases efficiency.

This paper's primary conclusion is that there is minimal evidence of significant negative effects of inflation on productivity growth for the 12 nations in the study. In fact, the inflation/productivity nexus varies so considerably across the present sample that no clear statement is possible about the sign, much less the magnitude, of the relationship. Therefore, there is little support for the argument that reducing inflation from already low levels to zero will have a measurable effect on productivity growth.

Relevant empirical literature is reviewed in the second section. The third section presents the data and methodology, the fourth section presents the analysis of the results, and the fifth section concludes.

Literature Review [2]

Prior to the early 1980s, research on inflation and productivity assumed the relationship to be unidirectional, running from (exogenous) productivity growth to inflation. Spurred by the dramatic decline in productivity growth in most industrialized countries during the 1970s and the runup in inflation that appeared to precede it, investigators in the early i 980s began to more closely examine the issue of causality. They used recently developed theoretical rationale for the negative effects of inflation on productivity. Using Granger-type tests, Clark [1982] and Ram [1984] found that inflation Granger-caused (negative) productivity growth in the U.S., and Jarret and Selody [1982] exhibited the same effect for Canada. In all three studies, the productivity measure was based on a highly aggregated (for example, private business sector) index of output per hour of labor. Clark and Jarret and Selody used standard Granger causality tests. Ram used both the Granger tests and Hsiao's technique.

Additional studies since the early 1980s produced results broadly consistent with the earlier work. Simos and Triantis [19881 analyzed U.S. data through 1986, while Saunders and Biswas [1990] analyzed the United Kingdom's manufacturing productivity through 1985, with both studies finding a significant negative effect of inflation on productivity growth. Smyth [1995a, 1995b] analyzed both German and U.S. multifactor productivity data and found a significant negative effect from contemporaneous inflation.

The conclusions to be drawn from these studies are limited, however, by several factors. First, these studies for the most part include only the inflation run-up and not its subsequent decline. Given the dramatic slow-down in productivity growth starting in the early 1970s, it would be surprising not to find a negative association between inflation and productivity growth over this period. Second, in most of the papers, causality testing fails to control for potentially relevant business cycle effects, including the possible endogeneity of contemporaneous inflation and the impact of variations in aggregate demand growth on measured productivity. Jarret and Selody [1982] is the only study that tries to control for relevant omitted variables in the inflation/productivity relationship. Clark [1982] acknowledges that omitted variables may be a valid problem, and more recently, Sbordone and Kuttner [1994] present evidence for the U.S. which suggests that any observed correlation between inflation and labor produc tivity is due to cyclical comovements rather than to any causal link. A third problem is that none of the papers test for the stationarity of the data. Since stationary variables are necessary for valid causality testing, this is an important omission.

More recent papers question the findings of a negative causal impact of inflation on productivity growth. Cameron et al. [1996] test for the existence of a cointegrating relationship between the relevant aggregate price and productivity series for Canada, Germany, the United Kingdom, and the U.S. They find that inflation is I(1),[3] nonstationary, for all four nations, while productivity growth is I(0), stationary. Under these conditions, standard causality tests of inflation and productivity are not valid since the series are not integrated of the same order. Cameron et al. [1996] fail to reject the null hypothesis of no cointegration for inflation and the level of productivity, both I(1) variables, and interpret their results as not supporting a link between inflation and productivity. Freeman and Yerger [1997, 1998] include cyclical variables, including real gross domestic product and short-term interest rates, to control for the possible endogeneity of contemporaneous inflation in a reexamination of Smyt h's [1995a, 1995b] U.S. and German data. With the business cycle proxies in the model, Freeman and Yerger find no statistically significant effect of inflation on productivity growth.

However, these negative findings do not resolve the issue entirely. Cameron et al. [1996] depend on unit root tests in order to establish an inflation/productivity linkage (Engle-Granger cointegration tests are, in essence, a unit root test), but the power of the test to reject a false null is known to be low, especially with highly autocorrelated data [West, 1988]. The work by Freeman and Yerger is of limited use for making broad generalizations about the impact of inflation on productivity since the U.S. and Germany have had two of the lowest average inflation rates over the past 50 years of any of the industrialized nations. The absence of any significant findings may be due simply to the test countries' relatively low average inflation rates, which may not give estimating procedures sufficient independent variation to distinguish the effect of inflation on productivity from the effects of the other independent variables [Kennedy, 1992].

This study extends previous research on the relationship between inflation and productivity in several important directions:

1) The set of countries is expanded to all of the major OECD nations and includes a wider range of inflation environments;

2) the time period is lengthened to cover the more recent disinflationary period;

3) a comparable measure, labor productivity in manufacturing, compiled by a single agency, the Bureau of Labor Statistics (BLS), is used for the entire sample; and

4) a more general test of causality is incorporated, one that economizes on the limited degrees of freedom available when taking lags of relatively short annual time series.

Consequently, the results of this study should be more generalizable than prior work.

Data and Methodology


The data set consists of annual data from 1955 to 1994 for 12 OECD nations, listed in Table 1, for which an index of output per labor hour in the manufacturing sector was available. The indexes are constructed as average productivity measures by the U.S. Department of Labor's [1998] BLS as part of their International Comparison of Manufacturing Productivity and Unit Labor Costs Program. Manufacturing labor productivity is computed as gross product originating in manufacturing (value added), in constant dollars divided by total hours worked. Manufacturing output is defined by the International Standard Industrial Classification.

While the BLS attempts to construct uniform measures across all nations, there are minor variations in the productivity measures . [4] Because of these differences, and because of the lack of an industry-specific exchange rate to use for conversion purposes, the BLS reports the productivity measure in index and percent change. The BLS has determined that the indexes of productivity are sufficiently comparable for measuring comparative trends across countries [Sparks and Greiner, 1997]. Moreover, unit root tests indicate the use of growth rates rather than levels for estimation purposes, improving the comparability of the data.

The use of productivity growth rates also eliminates differences in the levels of labor productivity across countries, which may be affected by institutional factors such as minimum wage laws, restrictive hiring/firing practices, or strong union representation. The simplifying assumption initially is that these factors are relatively stable over the sample period, although this assumption is relaxed somewhat by including real wages and unemployment rates, which may be symptoms of institutional differences, in the multivariate analysis. [5]

The use of this productivity measure has several advantages. First, it provides a consistent measure of manufacturing productivity across the nations, thereby facilitating international comparisons of the results. Second, manufacturing's capital intensity relative to other sectors of the economy may make it more vulnerable to some of the potentially deleterious effects of inflation upon productivity noted earlier. Finally, the use of labor productivity, as opposed to a total factor productivity measure, does not require the assumption of any specific production function in order to compute the productivity growth variable.

The aggregate price variable used for each nation is the log of the consumer price index (CPI), in either first or second difference form (DCPI or D2CPI, respectively). In order to capture any effects on measured productivity from business cycle effects, the first difference of the log of an index of real gross domestic product (GDP) is used in some model specifications. In order to allow for second differences of the CPI variables in the unit root tests and an a priori maximum lag length of four in the causality tests, the sample period is confined to 1961 to 1994. [6]

Figure 1 displays labor productivity growth in manufacturing and CPI inflation for the 12 countries used in this paper. The data are smoothed in each case via a four-year moving average to reduce high frequency variation.

Comparison of trend and cyclical components in the series is made somewhat difficult because of differences in the vertical scale, necessitated by the relatively high rates of inflation in some of the countries. However, two well-known phenomena stand out for all countries. The first is the secular decline in productivity growth and the rapid run-up in inflation following the 1973 oil shock, and the second is the drastic decline in inflation beginning in the early 1980s and continuing into the 1990s.

For the individual countries, however, the relationship between inflation and productivity growth appears to be stronger in some than in others. The U.S., the United Kingdom, Sweden, and Canada display perhaps the closest to the conventional view of a negative correlation of inflation and productivity growth, with peaks in inflation associated with troughs in productivity growth at both high and low frequencies. While all countries exhibit at least some periods of higher inflation and lower productivity growth, the simple correlation coefficient for two of the countries (Germany and Italy) was near zero, and for two others (Belgium and the Netherlands) was strongly positive. At the very least, the mixed nature of the visual evidence indicates that the underlying relationship between inflation and fluctuations in productivity growth is not a simple one in either the short or medium term and that any relationship may be affected by omitted factors.

To test the stability and the robustness of the inflation/productivity relationship, this paper employs Granger-type causality tests for the bivariate case and for the multivariate case incorporating other cyclical variables, proxying omitted factors. The inclusion of other cyclical variables will help distinguish between a true structural relationship between inflation and productivity and a statistical artifact of comovements of the variables over the business cycle.


The method of causality testing used here is that of Hsiao [1981, 1982]. The primary advantage of Hsiao's approach is that it allows the variables to enter the estimating equation with differing lag lengths. This helps retain degrees of freedom, especially in the trivariate models, with the potential of improving the power of the testing methodology relative to the standard Granger-causality method of utilizing equal lag lengths on all variables. With Hsiao's methodology in a bivariate framework, variable X is said to have a causal impact on variable Y if the final prediction error (FPE) from the FPE-minimizing bivariate model with Y and X is less than the FPE for the FPE-minimizing version of the univariate model for Y. Similarly, in a trivariate model, variable Z is said to have a causal impact on variable Y, after controlling for the impact of X, if the FPE-minimizing version of the trivariate model with Y, X, and Z has a lower FPE than the FPE for the FPE-minimizing bivariate model with Y and X. [7] In t his study, the maximum number of lags is set at four.

After completing the Hsiao causality testing using both DCPI and D2CPI as the PRICE variables, the results are compared against those obtained from standard Granger causality testing. In addition, Engle-Granger cointegration tests on DCPI and log productivity are conducted to compare the present results against those of Cameron et al. [1996]. All of the findings are summarized in the fourth section.


Prior to testing for causality, the order of integration for each of the variables was investigated and the results are summarized in Table 2. The Phillips-Perron (PP) test statistic was utilized in the unit root tests for two reasons: the PP test imposes fewer restrictions on the residuals in the unit root tests than do the augmented Dickey-Fuller (ADF) tests, and the findings of the present study using the PP test statistic were more robust than were those using the ADF statistic. The decision to reject the unit root hypothesis using the ADF statistic was much more sensitive to one-period changes in the lag length in the unit root estimating equation away from the optimal lag structure [8] than in the decision using the PP test statistic. Given the relative imprecision of any of the rules for selecting optimal lag lengths, the lower fragility of the conclusions based upon the PP test argue for its use over that of the ADF.

As indicated in Table 2, the conclusions for both the log productivity and log real GDP variables are straightforward: the first differences of both variables are stationary for every nation. The findings for log CPI, however, are not as clear. If the PP test results are strictly interpreted, then log CPI is I(1) for five nations (France, Germany, Japan, Norway, and Sweden) and I(2) for seven nations (Belgium, Canada, Denmark, Italy, the Netherlands, the United Kingdom, and the U.S.). There are, of course, valid reasons to be somewhat skeptical of the findings that log CPI is I(2) for many nations. As noted previously, the unit root tests are known to have low power when sample sizes are small or when the variable is stationary with a near unit root process [West, 1988]. Moreover, large persistent changes in inflation, characteristic of nonstationary processes, are not observed for this sample of nations and time period. Given these ambiguities regarding the stationarity of inflation, the question arises as to how to proceed. The decision for the present analysis is to complete causality testing on all nations using both DCPI and D2CPI and then to examine the results for sensitivity to inflation's order of integration. Fortunately, the primary conclusions are not sensitive to the assumption made on the stationarity of inflation.

Hsiao Causality Tests

The results of the Hsiao tests for causality from inflation (DCPI) to productivity growth (PRD) are summarized in Table 3, which shows the results of bivariate Hsiao causality tests between inflation and productivity growth and multivariate causality tests including GDP growth.

For the bivariate analyses, inflation is found to have a significant causal impact on productivity in eight nations: Belgium, Canada, Denmark, Italy, Germany, the Netherlands, Sweden, and the U.S. The advantage of allowing unequal lag lengths in the variables is demonstrated by the fact that of the eight nations showing a significant causal effect from inflation, the variables have an unequal lag length in five (Canada, Denmark, Italy, the Netherlands, and the U.S.). Somewhat surprisingly, however, and in contrast to much of the earlier research on the bivariate relationship between inflation and productivity, the impact from inflation is negative only for Canada (-.05), Denmark (-.09), and the U.S. (-.24). Thus, even before controlling for business cycle effects, there is evidence of a significant negative effect of inflation on labor productivity in only three of the twelve countries in the present sample.

It may be possible, however, that bivariate tests are biased by the omission of other macroeconomic variables that account for common cyclical factors explaining comovements in inflation and productivity. For example, higher inflation may induce an increase in interest rates that, in turn, reduces output. If employers hoard labor, the end result may be a temporary reduction in labor productivity. Controlling for these types of cyclical responses may provide a better test of any structural relationship between inflation and productivity.

The expanded model controlling for potential business cycle effects, reported in Table 3, reduces the findings of significant causality from inflation to productivity growth to only six nations and causes the sign of the inflation variable to change from negative to positive for the U.S. As in the bivariate analyses, the impact from inflation is negative for only three nations: Canada (- .06), Denmark (-.10), and the United Kingdom (-.13). The advantage of allowing unequal lag length is again demonstrated, as none of the nations showing a significant impact from inflation have all three variables entering the equation with equal lag lengths. [9]

Given the uncertainty of the stationarity of inflation for several of the nations in the study, Hsiao tests were also conducted for the causal impact of the first difference of inflation (DCPI) on productivity growth (PRD). These results provide even less evidence of a significant negative impact of prices on productivity than the tests using inflation levels. [10] In the bivariate model, Canada, Sweden, and the U.S. are the only nations showing a significant negative causal impact. When the model is extended to include the potential effects from real GDP growth, D2CPI has a significant effect in five nations, but the impact is negative only for Canada and the U.S. Also, it is once more the case that in nearly every instance where inflation is found to have a significant effect, the variables are entering the equation with unequal lag lengths."

Potential Confounding Effects of Wage Policy

It is possible that the tests in this paper may not be sufficient to differentiate among policy changes occurring simultaneously with changes in inflation. For example, suppose a nation has labor market restrictions (for example, a minimum real wage, [W.sub.min], greater than the market clearing wage) and assume standard labor demand conditions are such that [W.sub.min] MPL, where MPL is marginal product of labor, and all variables are measured as growth rates. An increase in inflation can cause marginal productivity to decline through two channels: a shift to the left in the MPL curve, due to the distortions discussed earlier, and a movement along and down the MPL curve, due to falling real wages. An interventionist government, however, may increase nominal wages to restore the original growth rate of real wages and, thereby, the original growth rate of MPL. Thus in this scenario, government policy may mitigate the negative effects of inflation on labor productivity growth. [12]

The restoration of marginal productivity growth notwithstanding, the nation should still experience a decline in average productivity growth. To see this, imagine the usual S-shaped total productivity curve of labor, with marginal product measured as the slope and average product as a ray from the origin to a point on the curve. In the scenario described above, inflation will cause the total product curve to rotate down (both curves have a common point at the origin). The point on the new curve with slope equal to the equilibrium point on the old curve will thus have lower measured average product of labor. Because average productivity is the measure used in this paper, a systematic negative effect of inflation on productivity should be revealed by the data even in the case where government attempts to intervene.


This paper uses Hsiao causality tests to examine the empirical evidence of a negative effect of inflation on manufacturing labor productivity growth using time series analysis for 12 OECD countries, a broader set of countries than in any prior published study of which the authors are aware. The Hsiao methodology economizes on degrees of freedom, a decided advantage given the introduction of variables that capture business cycle effects into the inflation/productivity growth relationship.

The results of this paper demonstrate the pitfalls of previous studies' focus on a single country or a narrow subset of countries. For those nations for which prior work exists (Canada, Germany, the United Kingdom and the U.S.), the findings here are roughly consistent with previous work. Any evidence of a negative effect of inflation on productivity growth is sensitive to the inclusion of cyclical variables. For the remainder of the countries, the case for a negative inflationary impact on productivity is even weaker. The correlation of inflation and productivity is just as likely to be positive as negative.

Consistent with Sbordonne and Kuttner [1994] and Freeman and Yerger [1997], including cyclical variables in the estimation eliminated a significant effect of inflation on U.S. and German productivity growth, respectively. The finding of a negative impact of inflation on Canadian and United Kingdom productivity growth is consistent with Jarret and Selody [1982] and Saunders and Biswas [1990], respectively.

These conclusions depend slightly, but not critically, on whether inflation is regarded as stationary. A strict interpretation of the stationarity tests finds evidence of a negative impact of aggregate price movements on productivity growth for only two of the 12 nations in the sample.

Even if the unit root test results are disregarded because of their low power, the evidence of a negative impact from aggregate price movements is slight. Inflation has a significant negative causal effect on productivity growth only for Canada, Denmark, and the United Kingdom after controlling for the impact of changes in real GDP growth rates (shown in Table 3). Moreover, the magnitude of the effect is quite modest for both Canada and Denmark. [13]

In summary, the analysis in this study does not support the view of a widespread and materially significant adverse effect of inflation on manufacturing labor productivity growth across major industrialized nations over the past 30-plus years. This is not to say that inflation has had no negative effects nor to imply that the inflation rates experienced by most nations in the latter 1970s and early 1980s were benign. Rather, whatever the productivity-reducing effects of inflation may be, they are relatively modest and difficult to isolate in the data.

(*.) Sam Houston State University and Lycoming College--U.S.A.


(1.) Refer to H.J. Resolution 409, introduced by Representative Stephen L. Neal [U.S. Congress, 1990].

(2.) This review focuses on prior time series-based research. There is separate literature on cross-section or panel data, with similar findings of inconsistent effects of inflation on productivity or economic growth. Englander and Gurney [1994] are representative of this literature.

(3.) The notation I(*) refers to the number of times, *, the data must be differenced to achieve stationarity, also referred to as the order of integration [Box and Jenkins, 1976].

(4.) Some differences across countries are as follows. Output for Japan prior to 1970 and for the Netherlands between 1969 and 1977 are indexes of industrial production, which include mining. Also, composition of price indexes used to deflate output can lead to differences in productivity levels. Most countries link fixed-weight price indexes covering various periods (about five years, on average), except for the U.S. and Japan which use fixed-weight price indexes for longer periods. Labor hour measures are developed from statistics of manufacturing employment and average hours and refer to all employed persons (including self-employed) in the U.S., Canada, Japan, France, Germany, Norway, and Sweden and to employees in all other countries. Compensation costs include all payments in cash or kind made directly to employees, plus insurance and benefits.

(5.) The authors thank an anonymous referee for pointing out that differences in institutional factors might affect the regression results.

(6.) Due to missing data, the sample period for Belgium is 1965 to 1993.

(7.) For a complete explanation of Hsiao's technique, see Hsiao [1981, 1982].

(8.) The optimal lag length was selected on the basis of the AIC+2 criterion, as described in Pantula et al. [1994].

(9.) As a test on the robustness of the Hsiao causality findings, standard Granger causality tests were conducted using all four of the models from the Hsiao analysis. The details of the analysis can be summarized in two points. First, with few exceptions, for each model, the Hsiao causality findings would nest the findings of causality using the Granger technique. Second, for those nations showing a significant effect from DCPI or D2CPI, the sum of the coefficients on the price variable usually would be quite comparable across the two methods.

(10.) These results are not reported here but are available from the authors upon request.

(11.) As a comparison against the results of Cameron et al. [1996], Engle-Granger cointegration tests were conducted on inflation and the log of the productivity index for each nation (both variables taken to be I(1)). Regardless of which variable was the dependent variable in the cointegrating regression, the null hypothesis of no cointegration was not rejected for every nation except Germany and the U.S. For Germany and the U.S., the null could be rejected when inflation was the dependent variable but not when log productivity was the dependent variable. This inconsistency should not exist if the two variables are cointegrated. Hence, the conclusions from the Engle-Granger cointegration tests are the same as those of Cameron et al. on Canada, Germany, the United Kingdom, and the U.S. As noted previously however, the failure to find cointegration well may be due to the weakness of the test rather than the absence of a true cointegrating relationship for these variables.

(12.) The authors are grateful to the referee for pointing out the possibility of government intervention as a possible source of mispecification. As a check, the trivariate models were estimated with unemployment and real wages as cyclical controls. Because the results were little changed from those including GDP growth, they are not reported here.

(13.) A sense of the relative magnitude of the estimated impact of inflation for these three nations can be obtained by using the coefficient estimates on inflation from the trivariate model in Table 3 to compute the predicted change in average productivity growth rate for a nation over the 1973-94 period if its inflation rate had maintained its average over the 1961-72 period. The average inflation and productivity growth rates are listed in Table 1. Canada's actual average productivity growth rate for the 1973-94 period was 2.07 percent. The predicted change in its average productivity growth rate due to the higher average inflation over this period is -.204 percentage points (= -[.06.sup.*] (6.20 - 2.86)). So of the observed decline in average productivity growth of -2.23 percentage points from the 1961-72 to 1973-94 periods, only -.204 percentage points is associated with higher average inflation rates. This represents only 9.1 percent of the observed productivity growth decline. Repeating this analysis f or Denmark finds that of the observed average productivity decline of 3.23 percentage points, -0.084 percentage points is associated with higher average inflation, which accounts for just 2.6 percent of the observed productivity growth reduction. The estimated impact of inflation is large only for the United Kingdom, where -.515 percentage points of the observed .65 percentage points of the decline in average productivity is associated with the higher average inflation rates.


Box, George; Jenkins, Gwilym. Time Series Analysis, revised ed., Oakland, GA: Holden Day, 1976.

Briault, Clive. "The Costs of Inflation, "Bank of England Quarterly Bulletin, February 1995, pp. 33-45.

Bruno, Michael. "Does Inflation Really Lower Growth?," Finance and Development, September 1995, pp. 35-8.

Cameron, Norman; Hum, Derek; Simpson, Wayne. "Stylized Facts and Stylized Illusions: Inflation and Productivity Revisited," Canadian Journal of Economics, 29, 1, February 1996, pp. 152-62.

Clark, Peter. "Inflation and the Productivity Decline," American Economic Review, 72, 2, May 1982, pp. 149-54.

Englander, A. Stephen; Gurney, Andrew. "Medium-Term Determinants of OECD Productivity," OECD Economic Studies, 22, Spring 1994, pp 49-109.

Feldstein, Martin. "Inflation, Tax Rules, and Investment: Some Econometric Evidence," Econometrica, 50, 1982, pp. 825-62.

Fischer, Stanley. Indexing, Inflation, and Economic Policy, Cambridge, MA: MIT Press, 1986.

Freeman, Donald; Yerger, David. "Inflation and Total Factor Productivity in Germany: A Response to Smythe," Weltwirtschlaftliches Archiv, 133, 1, 1997, pp. 158-63.

_____. Inflation and Multifactor Productivity Growth: A Response to Smythe, "Applied Economics Letters, 5, 1998, pp. 271-4.

Hsiao, Cheng. "Autoregressive Modeling and Money-Income Causality Detection," Journal of Monetary Economics, 7, 1981, pp. 85-106.

_____. "Autoregressive Modeling and Causal Ordering of Economic Variables, "Journal of Economic Dynamics and Control, 4, 1982, pp. 243-59.

Jarret, Peter; Selody, Jack. "The Productivity-Inflation Nexus in Canada," The Review of Economics and Statistics, 64, 3, August 1982, pp. 361-7.

Kennedy, Peter. A Guide to Econometrics, 3rd ed., Cambridge, MA: MIT Press, 1992.

Pantula, Sastry; Gonzalez-Farias, Graciela; Fuller, Wayne. "A Comparison of Unit-Root Test Criteria," Journal of Business and Economic Statistics, October 1994, pp. 449-59.

Ram, Rati. "Causal Ordering Across Inflation and Productivity Growth in the Postwar United States," The Review of Economics and Statistics, 66, 1984, pp. 472-7.

Saunders, Peter; Biswas, Basudeb. "An Empirical Note on the Relationship Between Inflation and Productivity in the United Kingdom, "British Review of Economic Issues, 12, 28, October 1990, pp. 67-77.

Sbordonne, Argia; Kuttner, Kenneth. "Does Inflation Reduce Productivity?, "Federal Reserve Bank of Chicago Economic Perspectives, November/December 1994, pp. 2-14.

Simos, Evangelos; Triantis, John. "A Note on Productivity, Inflation, and Causality," Rivista Internazionale di Scienze Economiche e Commerciali, 35, 9, 1988, pp. 839-46.

Smyth, David. "Inflation and Total Factor Productivity in Germany," Weltwirtschlaftliches Archiv, 131, 2, 1995a, pp. 403-5.

_____. "The Supply Side Effects of Inflation in the United States: Evidence from Multifactor Productivity," Applied Economics Letters, 2, 1995b, pp. 482-3.

Sparks, Christopher; Greiner, Mary. "U.S. and Foreign Productivity and Unit Labor Costs," Monthly Labor Review, February 1997, pp. 26-35.

Thornton, Daniel. "The Costs and Benefits of Price Stability: An Assessment of Howitt's Rule," Federal Reserve Bank of St. Louis Review, March/April 1996, pp. 23-38.

U.S. Congress, House Committee on Banking, Finance, and Urban Affairs, Subcommittee on Domestic Monetary Policy. Hearing on H.J. Resolution 409, 101st Congress, 2nd Session, Part 2, Serial 101-71, February 6, 1990.

U.S. Department of Labor, Bureau of Labor Statistics. "Foreign Labor Statistics: Description of Methods,", 1998.

West, Kenneth. "On the Interpretation of Near Random Walk Behavior in GNP," American Economic Review, 78, 1, 1988, pp. 202-9.
COPYRIGHT 2000 Atlantic Economic Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Atlantic Economic Journal
Geographic Code:00WOR
Date:Sep 1, 2000
Previous Article:European Stagnation and American Growth.
Next Article:Does Rising Consumer Debt Signal Future Recessions?: Testing the Causal Relationship Between Consumer Debt and the Economy.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |