The use of market price data in the formulation of monetary policy.
RECENT YEARS have witnessed the demise of several indicators or targets for monetary policy. By the late 1960s, real variables such as the rate of unemployment were deemed inappropriate policy targets, whereas the experience of the 1970s suggested interest rate targets were misleading policy guides. More recently, narrow monetary aggregates have proven much less reliable than they had earlier.
We operate in a fiat money floating exchange-rate regime and live in a world of limited information and hence in an environment where intermediate indicators or targets are appropriate and perhaps * Robert E. Keleher is an economist, Board of Governors, Federal Reserve System, Washington, DC, 20551. The views expressed are those of the author and not necessarily those of the Board of Governors or members of its staff. This paper is adapted from one presented at the Annual Meeting of the National Association of Business Economists, September 24-27, 1989, San Francisco, CA. necessary. Therefore, what can monetary policy-makers do? If interest rate levels, real variables, and monetary aggregates do not work, what types of intermediate indicators are appropriate for monetary policymaking?
This paper suggests that the answer to these questions is to use market price indicators consistent with the prescriptions of monetary theory and prescribed in previous fiat money/flexible exchange rate regimes by classical monetary writers. These prices, from centralized auction markets, are characterized by their flexibility rather than the stickiness exhibited by indexes of sampled prices, such as the CPI or various GNP deflators. SOME IMPORTANT CLUES
The case for the use of market prices as intermediate indicators for monetary policy is made more convincing by identifying two very important clues. The first major clue relates to the fact that, while the monetary aggregates may no longer accurately measure the thrust of monetary policy (due, perhaps, to the effects of deregulation), still many valid lessons or principles of monetary thought suggest what might be appropriate indicators for monetary policy.
For example, in spite of the apparent breakdown of the reliability of the money-income relationship, there is no reason to reject the notion that inflation is ultimately a monetary phenomenon. While we may not be able to measure relevant transactions balances accurately, it is still likely that, when the nominal prices of all goods are persistently increasing, too much money is chasing too few goods; the value of money is decreasing. Similarly, in spite of problems with measuring relevant monetary aggregates, there is no reason to believe that monetary stimulation can permanently influence real variables or relative prices; the long-run neutrality of money is likely still quite valid.
Accordingly, the following notions remain well-grounded and valid in spite of the recent poor performance of monetary aggregates:
1. The targeting of interest rates, relative prices, or
real variables should be avoided;
2. Policymakers should use nominal variables as intermediate
guides to policy;
3. Price stability should be a key goal of monetary
Similarly, there is no convincing evidence to reject the view that, because we live in an uncertain world, an intermediate indicator approach to policymaking is appropriate.
In short, we should remember that some key principles of both monetary theory and policy remain valid. We also should recognize that these valid lessons suggest what are inappropriate policy indicator targets, i.e., real variables, relative prices, and interest rate levels. Accordingly, these principles limit the set of viable policy indicators.
But these lessons also indicate what types of variables might be appropriate policy guides. In particular, the use of nominal prices is consistent with virtually all of these principles. The first important clue, then, is that valid principles of monetary thought suggest that nominal prices may be appropriate monetary policy indicators.
A second major clue as to what might be appropriate policy guides comes from the history of monetary thought. More specifically, it emerges from an examination of previous monetary regimes like the one we live in today: a fiat money, flexible exchange-rate regime.
Historically, in only two major episodes prior to our current experience did most prominent countries of the world go off some form of commodity standard: the period of the Bullionist debates (17971821), and portions of the interwar period (especially immediately after World War 1). In both of these earlier periods, the view that market prices should be used as guides to policy emerged as viable policy prescriptions from leading monetary theorists of the era.
More specifically, during the Bullionist debate, Henry Thornton, other authors of the Bullionist Report, as well as David Ricardo, all explicitly advocated the use of market prices as guides to Bank of England monetary policy.(1) This prescription was reiterated, elaborated, and refined by several monetary theorists but particularly by Knut Wicksell in the late nineteenth century. See footnotes at end of text. A list of references mill be provided by the author on request.
During the interwar years, John Maynard Keynes (of both the Tract and the Treatise), Knut Wicksell, Henry Simons, and others supported the use of market prices as guides to monetary policy.(2) In short, the second major clue is that during previous monetary regimes that approximate current circumstances, leading monetary theorists explicitly prescribed the use of market prices as guides to monetary policy. Thus, both valid principles of monetary thought and the writings of leading monetary theorists living in circumstances that approximate our current experience suggest that market prices are appropriate guides to monetary policy. THE MARKET PRICE APPROACH TO MONETARY POLICY
The use of market prices as intermediate indicators for monetary policy is briefly described in the following sections. In examining this approach, it is useful briefly to describe its component parts: its ultimate policy goals, intermediate indicators, and operating instruments.
1. The ultimate policy goal of the market price approach
is price stability.
2. The intermediate indicators are nominal market
prices. Indices of commodity prices and the price
of foreign exchange both serve as such intermediate
indicators. Because commodity prices and the price
of foreign exchange are both impacted by changes
in monetary policy, they may provide useful information
to policymakers about the effects of their
action; they may serve as proxies for the value of
money. Such information is particularly useful if
market prices respond more quickly to policy
change than do other indicators and if market price
indicators are assessed in conjunction with one another
over a period of time. Interest rate spreads,
or, more specifically, the spread between the federal
funds rate and the long-term Treasury rate,
serve as a measure not of the effects of policy but
rather of the thrust of monetary policy. Accordingly,
this spread should be treated somewhat differently
than the other two indicators.(3)
3. The federal funds rate serves as the operating instrument,
not target, of monetary poliCy.(4) In short,
the approach approximates a Wicksellian approach
to policy whereby the federal funds rate is employed
as an operating instrument that is adjusted
on the basis of information from the commodity,
foreign exchange, and bond markets to achieve
price stability. Should these price indicators, assessed
in conjunction with one another, suggest the
Federal Reserve is too easy or too tight, the federal
funds rate is adjusted accordingly. ADVANTAGES OF PRICE DATA
As described, the approach uses market price data.(5) The advantages of employing price data are best described by recognizing the many problems of quantity data. Data measuring quantities such as the National Income and Product Account (NIPA) data or monetary statistics are necessarily based on samples. Accordingly, such quantity data are subject to revisions and rebenchmarks that can often be substantial. An inherent lag occurs in the reporting of such data, and the data are often constructed on the basis of accounting concepts rather than economic principles. To be useful, the data necessarily must be seasonally adjusted. And should redefinitions of the data occur, due, for example, to deregulation or evolving institutional or technological change, the altered measurements and changed behavior of particular variables can be substantial. In sum, significant measurement, timing, and sometimes definitional problems are associated with the use of sample-based quantity data. The market price approach, on the other hand, uses price data from centralized auction markets. Using observable data, the approach does not depend on unobservable variables, such as real or equilibrium" interest rates that depend on accurate measurements of future price expectations or the productivity of capital. Such price data are easy to understand, timely, and readily available, literally by the minute. They are accurate, being less subject to sampling error, and are not affected by revisions, rebenchmarks, seasonal adjustments, or "shift-adjustments" that sometimes plague quantity data. Several formal studies (e.g., Morgenstern, 1963; Zarnowitz, 1980) investigating the usefulness of various forms of economic statistics conclude that market price data are superior to other forms of data.
The use of market price data is premised on the notion that market prices are summaries of or aggregators of information embodying the knowledge and expectations of large numbers of buyers and sellers who have incentives to make informed decisions in an uncertain world. Accordingly, these prices are inherently forward looking, offering a distinct advantage over any form of quantity data and particularly pertinent for monetary policymakers who necessarily must also be forward looking.
The market price approach recognizes that we live in a world of limited information; the approach attempts to use best the limited information that exists. In a sense, this approach allows policymakers to take advantage of the information aggregating nature of prices, particularly those flexible prices that best serve as proxies for the value of money. In short, the market price view employs data that make better use of dispersed information than alternative forms of data. EVIDENCE
A growing body of research supports the use of market price indicators and demonstrates that each of the indicators mentioned embodies relevant information for monetary policymaking. The research suggests that these particular market prices are relevant indicators for monetary policy.
Recently, for example, empirical evidence has accumulated supporting the yield spread (the difference between federal funds rate and the yield of thirty-year Treasury bonds) as a useful measure of the thrust of monetary policy. Laurent (1988) has presented theoretical perspective and empirical support for the yield spread as a measure of this thrust. In particular, he provided empirical support for this particular measure relative to other well-known measures. Moreover, in a more recent paper, Laurent (1989) assessed the performance of this yield spread since the publication of his earlier article. These results pertaining to the federal funds - long bond spread have been fully corroborated in a recent paper by Bemanke and Blinder (1989). After an extensive empirical examination, they conclude that Laurent's yield spread is the preferred measure by which to gauge changes in monetary policy.
Other recent research has also corroborated Laurent's results. While not employing exactly the same interest rate spread, for example, Stock and Watson (1989) found that the yield spread is a potent leading indicator of economic activity that outperforms many alternative economic and financial variables. The predictive power of interest rate spreads has also been documented by Bernanke (1983), Estrella and Hardouvelis (1989), Friedman and Kuttner (1989), as well as by Harvey (1989). These studies generally find that yield spreads outperform other proxies often used to measure changes in monetary policy.
Similarly, empirical evidence supporting the use of commodity prices as indicators for monetary policy has also been accumulating. While the empirical evidence is not one-sided, recently many researchers have found empirical support for the use of commodity prices as a useful indicator for monetary policy. Several studies, for example, find that commodity prices significantly improve the fit of regressions explaining movements in consumer prices and are useful in predicting consumer prices.(6) And the evidence does suggest "causation" runs from movements in commodity prices to movements in consumer prices.(7) Commodity price movements, then, do contain information about the future path of inflation. Moreover, the evidence also indicates that changes in commodity prices tend to lead changes in consumer prices.(8) While evidence pertaining to cointegration is mixed, some evidence does indicate that there is a reliable long-term relationship between the level of commodity prices and the level of consumer prices.(9)
Similarly, both theory and empirical evidence suggest that exchange rates may serve as useful indicators for monetary policy. Ample empirical evidence indicates that changes in monetary policy influence the exchange rate and that exchange rate movements normally translate into price movements.(10) Accordingly, the exchange rate may serve as a useful indicator for monetary policy, particularly for the monetary policy of a reserve currency country.
In sum, research suggests that, because each of these indicators makes theoretical sense, has practical advantages, and has growing empirical support, they likely are relevant indicators for monetary policy. TWO ADDITIONAL CONSIDERATIONS
While empirical evidence supporting this approach accumulates, two considerations must be emphasized: (1) To be useful, these price indicators should be analyzed in conjunction with one another; and (2) there appear to be important biases against finding strong empirical support for this strategy.
Use Market Prices in Conjunction
with One Another
The market price indicators discussed above should neither be used as policy targets nor assessed independently. These indicators can be volatile and are sometimes influenced by factors other than monetary policy. Accordingly, if these prices are targeted or assessed in isolation, they can sometimes be misleading. Instead, as the classical bullionist writers suggested, these price indicators should be used in conjunction with one another because each is influenced by both monetary and nonmonetary factors. The job of the policymaker is to decipher when a disturbance is of monetary origin and when it is not. Assessing movements in these indicators in combination can aid in making such a determination. Each of these indicators can provide useful information and each allows a somewhat different perspective to be used in assessments of the overall picture. When used cautiously and together with one another so as to piece together a consistent picture of overall policy, these indicators can contribute to monetary policymaking by minimizing the probability of policy error. They can help policymakers to discriminate better between monetary and nonmonetary disturbances and thereby help policymakers determine when to react to relevant monetary disturbances and when to ignore those disturbances of nonmonetary origin. Synergistic effects can occur; evidence supporting each individual price indicator may not be so powerful, but assessed together these indicators may prove potent contributors to policymaking. Because these indicators are nominal variables, if used in a strategy to foster price stability, they will work to prevent any major policy mistakes.
The use of the three indicators in conjunction with one another can be briefly illustrated. If the dollar is depreciating while the yield spread is steepening and commodity prices are rising, monetary policy is likely expansionary and perhaps overly so. On the other hand, if the dollar is depreciating while commodity prices and the yield spread are stable, the dollar may reflect external factors, such as restrictive foreign monetarypolicy. (12) Furthermore, if the dollar is declining and the yield spread steepening but commodity prices are stable, foreign funds may be flowing out of the U. S. bond market for reasons other than inflationary expectations.
Some additional implications of the strategy should be mentioned. The assessment of these indicators in conjunction with one another makes the approach very difficult to model; no hard and fast policy rules seem obvious or emerge from the strategy. This quandary relates both to the difficulty of assessing three variables simultaneously but also to the fact that some price movements will be ignored while others will elicit a policy response. Additionally, while empirical support for individual market price indicators may not be strong, the overall strategy may prove to be potent. Thus, the individual contribution of each price indicator is difficult to assess and the overall strategy is difficult to evaluate empirically.
Biases Against Finding Empirical
Support for this Strategy
While existing empirical evidence does lend support to the market price strategy, there are several reasons why empirical tests may not find strong support for the approach? Empirical tests may be biased against supporting the approach even though it may work quite well in practice. (Indeed, at least one historical episode suggests that it in fact did work quiet successfully.(13)
First, of course, the well-known Lucas critique suggests that the approach cannot genuinely be assessed or rigorously tested with data from alien regimes.
Second, because market price indicators are affected by both monetary and nonmonetary influences, the strategy prescribes that policymakers respond to movements in indicators in some cases and ignore such movements at other times. Accordingly, empirical relationships between individual price indicators and Federal Reserve policy instruments may be weak, as may relationships between individual price indicators and actual inflation. Unfortunately, no straightforward empirical test capturing the simultaneous contribution of all relevant movements in these price indicators is available.
Third, because market prices embody expectations of the future, they are inherently forwardlooking. Accordingly, movements in commodity prices, foreign exchange rates, and bond yields may at times reflect expectations of future inflation. If the monetary authority responds to these movements before such expectations affect or become embedded in actual prices, then such expectations may recede without any observable effects on actual inflation. If these events should occur, there may be little empirical relationship between such market price indicators and measured inflation even though market prices may be excellent policy indicators in a price-stabilizing monetary policy strategy. Again, no appropriate empirical test to capture these effects seems obvious.
In sum, although each of the individual market price indicators has received empirical support, there are reasons to believe that empirical tests may not yield strong support for this strategy. Using market prices in conjunction with one another may yield more potent results than suggested by the results of formal empirical tests on individual market price indicators. All of these considerations suggest that despite its potential effectiveness, the market price approach is not readily modeled and does not easily lend itself to conventional empirical evaluation. SUMMARY AND CONCLUSIONS
The market price approach offers an alternative strategy for monetary policymakers. It may resolve the current dilemma facing monetary policy. Not a novel approach, its essentials were outlined by classical bullionist writers in the early nineteenth century. It was reiterated, clarified, and refined by subsequent monetary theorists, particularly by Knut Wicksell. In previous extended episodes of flexible exchange rate regimes, variants of the approach emerged as a viable policy option and were endorsed by leading monetary theorists of the day.
There are important philosophical underpinnings as well as many advantages of using market price data for monetary policymaking, particularly when such price data are contrasted with common forms of quantity data and the many problems plaguing such quantity data.
The approach works best when the indicators are assessed in conjunction with one another. While the strategy is not easily amenable to modeling, empirical evidence supporting the approach appears to be growing in spite of several important biases against finding such support. In sum, the strategy does appear to offer promise as a wellgrounded, viable alternative for monetary policy. FOOTNOTES
(1) For documentation of this argument, see Keleher 1989 (b).
(3) For a more detailed discussion of this measure, see Keleher 1989 (c).
(4) See ibid for a discussion of the role of the Fed funds rate.
(5) For a more detailed discussion of the advantages of using price data, see Keleher 1989 (a).
(6) See, for example, Branson and Boughton (1988), Cody and Mills (1989), Furlong (1989), Gamer (1988) (1989), Horrigan (1986), Webb (1980), Whitt (1989).
(7) See, for example, Horrigon, Cody and Mills, Webb, and Whitt.
(8) See, for example, Branson and Boughton, Klein, Roth, and Von Zur Muehlen.
(9) See, for example, Marquis and Cunningham, and Whitt.
(10) See, for example, Glick and Hutchison, Keleher (1989 d), and McKinnon.
(11) In assessing movements in commodity prices and the dollar, it should be mentioned that commodity prices denominated in foreign currencies should also be monitored in order to assess whether commodity price movements reflect world inflation or exchange-rate-related domestic price movement.
(12)There does appear to be one successful "case study" of the approach: the Swedish experience in the early 1930s. The Swedish central bank, The Rikesbank, adopted such a Wicksellian approach and found the approach resulted in more stable prices and economic performance superior to most other countries in the 1930s. See, for example, Jonung (1979), (1981) and Fisher (1934).
|Printer friendly Cite/link Email Feedback|
|Author:||Keleher, Robert E.|
|Date:||Jul 1, 1990|
|Previous Article:||Debt and business cycles.|
|Next Article:||The business economist at work: Inland Steel Industries.|