2 Q1 2011
firms will respond to increaseddemand by raising prices. Thus,positive output gaps can signal futureinflationary pressures. Since monetarypolicy operates with significant lags, itis important that policymakers have anaccurate inflation forecast.Second, some economists arguethat employment gaps are a usefulguide for policy aimed at achievingmaximum sustainable employmentand price stability, which are themandated objectives of monetarypolicy in the United States. Over themedium term, employment is drivenby fundamentals such as productivityand labor supply growth, and thesemedium-term measures are used toinfer simultaneously the natural rate of unemployment and the unemploymentgap. Most economists do not thinkmonetary policy is part of thesemedium-term fundamentals. Instead,attempts to drive employment oroutput above their fundamental levelswould result in unwanted inflation andno employment gains.Economists have developedsophisticated measures of resourceslack that are grounded in economictheory, yet remain workable inpractical terms. We will brieflydiscuss two important examples:the nonaccelerating inflation rate of unemployment (NAIRU) model andthe output gap measure publishedquarterly by the CongressionalBudget Office (CBO). However,even the latest models recognizethat there is a large amount of uncertainty about output andemployment gaps. Moreover, thereremain competing definitions of the output and employment gaps.Alternative measures sometimes offercontrasting implications for monetarypolicy. Therefore, it is important tounderstand the limitations of currentoutput gap estimates for both inflationforecasting and output stabilization.
MILTON FRIEDMAN ANDTHE NATURAL RATE OFUNEMPLOYMENT
Economists have long believed ina relationship between money, prices,and employment some say since the18th century!
However, it was notuntil 1958 that A.W. Phillips providedthe first statistical analysis comparingwage inflation and unemployment, us-ing data for the United Kingdom since1861.
Phillips found that when unem-ployment was high, inflation was low.This negative relationship now bearshis name: the Phillips curve. A fewyears later Paul Samuelson and RobertSolow imported the Phillips curve tothe United States.
Samuelson andSolow used price inflation instead of wage inflation, a choice now preferredby most researchers. Figure 1 displaysa typical Phillips curve plot for the pe-riod 1948-1965. Each dot correspondsto the inflation and unemploymentrate during a quarter. The solid linedisplays the statistical relationship.Nowadays economists recognizethat the Phillips curve is more thana statistical relationship between twovariables. The modern view of thePhillips curve is rooted in the ideasthat Milton Friedman developed atthe University of Chicago during thelate 1960s.
Friedman believed thatattempts to increase employmentby increasing inflation were mis-guided. Only unanticipated inflation,Friedman argued, has the ability tostimulate employment. For example,if households and firms expect aninflation rate of 2 percent, a 3 percentinflation rate would effectively increaseoutput and employment by boostingreal demand. However, as workers andfirms came to expect a 3 percent infla-tion rate, they would embody such ex-pectations in wage demands and pricesetting. As a result, an inflation of 3percent would increase nominal outputcompared with 2 percent inflation, butsince all prices and wages adjusted by3 percent as well, there would be nochange in real output and employment.Friedman’s view was validatedwhen inflation rose persistently in the1970s despite a marked slowdown inemployment growth. Indeed, analystscoined the term stagflation to describethe combination of stagnant growthand inflation.The old view of the Phillips curvecould not explain the stagflationphenomenon. Figure 2 plots unemploy-ment and inflation rates for the period1970-1979. We include the Phillipscurve as given in Figure 1 for theperiod 1948-1965. The actual observa-tions are all northeast of the Phillipscurve: Both inflation and unemploy-ment rates were higher than the theorypredicted. This was not immediatelyrecognized, and many policymakersmistakenly believed that inflationwould reverse course and moderate.
Robert Lucas, in his Nobel lecture, traced theobservation back to the writings of David Hume.
See the article by A.W. Phillips.
See the study by Samuelson and Solow.
For a detailed discussion of the so-called NewKeynesian Phillips curve, see the article byKeith Sill on page 17.
Nowadays economists recognize that thePhillips curve is more than a statisticalrelationship between two variables.
The book by Thomas Sargent discusses theexperience of the 1970s in rich detail.