Institutions matter. Economists of the classical period knew that well. In recent years, economists have increasingly included institutional variables in their empirical work. The economic freedom indexes from the Fraser Institute and the Heritage Foundation have been among the most widely used institutional indicators.
The purpose of an economic freedom index, according to Heritage, is to “document the positive relationship between economic freedom and a variety of positive social and economic goals.” Many studies support that claim, finding that countries with high economic freedom scores are, in fact, more prosperous and dynamic than those that are less free. However, not all aspects of economic freedom turn out to be equally important. In earlier posts, I have been critical of the components of the economic freedom indexes that focus on the size of government and regulation. Here, I take on those that focus on price stability—the Fraser “Sound Money” component and the Heritage “Monetary Freedom” component.
I find that these price stability components add little to our understanding of economic freedom. Furthermore, because they incorporate an exaggerated fear of even moderate inflation, an attempt to achieve maximum price stability, as defined by these indicators, would be less likely to bring prosperity than to undermine it.
Is price stability an institution or a policy outcome?
The first problem with the Fraser and Heritage price stability indicators is that they do not really measure economic freedom in the sense that it underlies other components of the indexes. Robert Lawson, a senior member of the team that publishes Fraser’s Economic Freedom of the World reports, once wrote that an economic freedom index is, or should be, only that—“not an index of economic growth policies, efficient government provision of public goods, macroeconomic stabilization policies, or ideal income distribution policies.” If so, then the sound money indicators, as indexes of the success of monetary policy, are just what an economic freedom index should not be.
Measures of inflation and money growth account for three-quarters of the data that go into the Fraser Sound Money indicator and four-fifths of the Heritage Monetary Freedom indicator. Rather than measuring the freedom of individuals to work, produce, consume and invest as they see fit, they measure the performance of central banks. To be sure, if we are going to have central banks, we want them to do a good job, but conceptually, good central banking and economic freedom are different things.
Instead, it seems to me that the natural meaning of economic freedom, as applied to monetary matters, ought to be the freedom to use and exchange whatever kind of money one wants for whatever purpose. To their credit, neither Fraser nor Heritage entirely neglects that notion of monetary freedom. Fraser’s Sound Money includes a subcomponent, weighted at 25 percent, that measures the extent of restrictions on ownership of foreign currency bank accounts. The Heritage Monetary Freedom indicator gives a weight of up to 20 percent to a measure of the extent of price controls. If I were to try to build a better measure of monetary freedom, I would begin by including both of those, but I would not stop there.
Why not, for example, include a measure of central bank independence, much as both the Fraser and Heritage indexes include measures of judicial independence in evaluating the quality of legal institutions? There is a large literature linking central bank independence to the desired policy outcome of price stability. (This IMF working paper discusses measurement issues and provides links to previous literature.) In a similar vein, it might make sense to rate central banks on their use of policy rules. Many economists have argued that using policy rules to constrain or even fully replace central bank discretion can improve price stability. (A recent conference sponsored by the Boston Fed provides a detailed airing of the case for policy rules and contained discretion.)
Another indicator for possible inclusion would be the presence or absence of black markets for foreign currency, and, where black markets exist, the spread between the official and black-market exchange rates. (Fraser does include a measure of black-market exchange rates, although it appears as a subcomponent of the Freedom of Trade category rather than of Sound Money.) The multiple official exchange rates for different types of transactions that some countries maintain also represent limitations on freedom.
Finally, we might try to measure the degree to which a country’s monetary system is open to competing forms of money—a subject that Larry White, among others, has written about at length. An up-to-date treatment would want to cover regulatory limits on the use of cryptocurrencies.
A greater emphasis on these and other institutional variables, and less on the rate of inflation itself, would make the monetary components of the Fraser and Heritage indicators more consistent with the conceptual framework that underlies their other components.
An exaggerated fear of inflation
Setting aside the conceptual question of whether sound money is a form of freedom, an institution, or simply good macroeconomic policy, we come to the second major problem with the Fraser and Heritage price stability measures. By enshrining zero percent inflation as the ideal, both of them reflect an exaggerated fear of even moderate inflation that is not supported by the preponderance of evidence.
To understand why, we need to dig into some technical details, starting with Fraser’s Sound Money. This indicator has four subcomponents: the growth rate of the M1 money stock (currency plus checkable deposits) relative to the growth of real GDP; the standard deviation of inflation over the previous five years; the rate of inflation in the most recent year; and a measure of the right to own foreign currency bank accounts. The raw data for each subcomponent are converted to a scale of 0 to 10, with higher numbers representing greater price stability, and then averaged to get the full Sound Money indicator.
Of the four subcomponents, M1 growth appears to be a holdover from Milton Friedman’s personal involvement in the early stages of Fraser’s Economic Freedom of the World project. Few, if any, economists today consider holding M1 growth equal to the long-run rate of real GDP growth to be a reasonable target for monetary policy. One of the main reasons is that the difference between the rate of money growth and the rate of real economic growth equals the rate of inflation only if M1 velocity (the ratio of nominal GDP to the money stock) is constant. In practice, however, velocity can be highly variable. For example, from 2005 to 2008, when velocity was rising, the Fraser money growth index averaged an excellent 9.7 even though inflation averaged 3.2 percent. From 2009 to 2014, when velocity was falling, inflation slowed to 1.6 percent but the Fraser money growth index, perversely, fell to 8.45, its lowest level in the 45 years for which it has been calculated.
Beyond the problems with M1 growth, there are problems with the way Fraser’s Sound Money indicator measures inflation itself. The inflation subcomponent is based on the rate of CPI inflation in the most recent year, but manipulates the raw data extensively. First, the distribution of inflation rates is truncated at 50 percent per year. All countries that experience hyperinflation, no matter how rapid, are assigned the same score. Next, any deflation, that is, any negative rate of inflation, is converted to its absolute value. Finally, the data are converted to a scale of 0 through 10. This procedure can be summarized by the formula
FINFi = (50 – |пi|)/5
where FINFi is the Fraser inflation score for country i and пi is the raw inflation rate. By this method, then, an inflation rate of 0 percent becomes a perfect 10, a rate of either +2 percent or -2 percent gives a score of 9.6, and any inflation rate of 50 percent or higher earns a zero.
This procedure exaggerates the harm done by moderate rates of inflation in three ways. First, it assumes a priori that the optimal rate of inflation is zero. Second, by arbitrarily truncating the distribution at an inflation rate of 50 percent, it makes the scores of countries with moderate inflation look worse than they otherwise would. (For example, if the maximum permitted value of inflation were 100 percent rather than 50 percent, the score for a country with 2 percent inflation would be 9.8 rather than 9.6.) Third, by using absolute values of inflation rates, it implicitly assumes that a rate of +2 percent inflation is just as harmful as deflation of -2 percent, something that few economists believe to be the case. (See here for a full discussion of deflation and its effects.)
The Heritage Monetary Freedom indicator uses a somewhat different method, but one that similarly exaggerates the harm done by moderate inflation and understates the risks of deflation. Like Fraser, Heritage uses absolute values to convert deflation to an equivalent rate of inflation. Rather than truncating the distribution of observed inflation rates, Heritage reduces the statistical impact of extreme hyperinflation values by basing its score on the square root of observed inflation rates, multiplied by a constant. Rather than the most recent year’s inflation, it uses a three-year weighted average of inflation rates, and rather than a scale of 0 to 10, it uses a scale of 0 to 100, with higher values indicating less inflation. The full procedure is given by the formula
HINFi = (100 – 6.33*√п*i) – CONTROLS
where HINFi is the Heritage inflation score for country i and п*i is a weighted average of the absolute values of the last three years’ inflation rates, and CONTROLS is a penalty of up to 20 points depending on the pervasiveness of price controls.
The effect of this formula is to reduce the scores of countries with moderate inflation rates by even more, in comparison to those with zero inflation, than does the Fraser method. For example, raising the inflation rate from zero to 2 percent lowers the Heritage inflation score from 100 to 91 (assuming no price controls), compared to a smaller relative reduction from 10 to 9.6 using the Fraser formula.
Rather than making a priori assumptions that idealize zero inflation, penalize moderate inflation, and understate the risks of deflation, it would make more sense to look at the actual effects of inflation as revealed by empirical studies. There is an abundant literature to draw on.
Probably the most extensively studied question is the relationship between price stability and economic growth. (See here and here for recent surveys of the literature.) The question is not simply one of whether inflation is good or bad for growth. Three frequent findings suggest that a more nuanced view is appropriate.
- Many studies find a nonlinear relationship, in which inflation up to a certain rate is associated with higher growth and with lower growth after that.
- The optimal level of inflation—that is, the rate associated with the highest growth rates—appears to be lower in more developed economies.
- Country-specific effects are strong. Both the optimal inflation rate and the degree of harm from excess inflation vary according to circumstances.
Some of the studies that reached these conclusions were done years ago, raising the question of whether they continue to hold for the period leading up to the global financial crisis and the recovery from it. For a quick-and-dirty check, I examined the most recent ten years of IMF data on inflation and growth. The findings were generally consistent with those of past studies:
- Nonlinear formulations of the inflation-growth relation produced significantly better fits than linear formulations.
- The relationship of inflation to growth varied according to income level. The rate of inflation associated with the best growth performance was lower for countries in the highest income quartile than for middle- and lower-income countries.
- Although there was a statistically significant relationship between growth and inflation overall, the fit was far from tight. As in previous studies, country-specific factors accounted for a large part of the cross-country variations in growth.
In addition to the literature on the relationship between inflation and growth, there are many studies of the relationship between inflation and unemployment, both in the short run and the long run. One strand of that literature suggests that there is a “backward bending Phillips curve.” The idea is that there is some moderate, positive rate of inflation that produces the lowest minimum unemployment rate that can be sustained without accelerating inflation. Proponents of this view emphasize behavioral factors, especially downward rigidity of nominal wages. The most widely cited paper proposing a backward-bending Phillips curve was published in 1996 by George A. Akerlof, George L. Perry, and William T. Dickens. Thomas Palley has proposed a simpler version of their model that reaches the same conclusion.
In short, whereas the Fraser and Heritage price stability indicators assign the highest possible scores to countries with zero inflation, economists who have studied the effects of inflation in the real world have found, more often than not, that a moderately positive rate produces better results.
Taking all of the above into account, I reach the following conclusions regarding the price stability components of the Fraser and Heritage economic freedom indexes:
- Conceptually, the idea of measuring economic freedom in the first place is motivated by the hypothesis that good institutions produce good outcomes. In my view, explorations of that hypothesis are best conducted using economic freedom indicators that focus on institutional quality. Including indicators of policy outcomes, such as inflation rates, only confuses matters.
- The price stability components of Fraser and Heritage economic freedom indexes, which assign the highest possible scores to countries with zero inflation, reflect an exaggerated fear of the effects of moderate inflation. By using the absolute value of inflation rates in their indicators, they also understate the detrimental effects of deflation.
These conclusions have important implications both for researchers and for policymakers.
Researchers should treat the price stability components of the Fraser and Heritage indexes with caution. The results of any statistical tests that use the economic freedom indexes as a whole should be checked against tests that disaggregate those indexes into their principal components. The results of statistical tests that use the Sound Money and Monetary Freedom indicators as independent variables should be checked against results that use raw inflation data, instead, or using tests that strip out price stability data altogether. Researchers interested in exploring the relationship between macroeconomic performance and the quality of monetary institutions should consider augmenting the Fraser and Heritage data with additional institutional indicators, such as measures of central bank independence, the use of monetary policy rules, freedom to use competing forms of money, and exchange rate regimes.
At the same time, policymakers should be cautioned against using the Sound Money or Monetary Freedom indexes as performance benchmarks. There is little evidence to support these indexes’ implicit assumptions that zero is the optimal rate of inflation or that a given rate of deflation is no more damaging than the same rate of inflation. Far from promoting freedom and prosperity, any attempt to maximize the soundness of money as defined by Fraser or monetary freedom as defined by Heritage would be more likely to place economies in a low-growth, high-unemployment straitjacket.