1 natural rate of unemployment Milton Friedman defined the natural rate of unemployment as the level of unemployment that resulted from real economic forces, the long-run level of which could not be altered by monetary policy. Macroeconomic policymakers continue to view the natural rate as a key benchmark due to the belief that monetary policy can counter short-run deviations of the unemployment rate from the natural rate. It is important, however, that policymakers focus as much attention on understanding the real determinants of the natural rate, and the policies that can affect it, as they do trying to identify and counteract deviations from it. In his 1968 presidential address to the American Economics Association, Milton Friedman famously defined the natural rate of unemployment as . . . the level that would be ground out by the Walrasian system of general equilibrium equations, provided there is imbedded in them the actual structural characteristics of the labor and commodity markets, including market imperfections, stochastic variability in demands and supplies, the cost of gathering information about job vacancies and labor availabilities, the costs of mobility, and so on. (1968, p. 8) This definition is incomplete, however, because it conspicuously lacks any mention of inflation. A more complete definition emerges from the remainder of Friedman‟s presidential address, in which he extensively examined the relationship between the unemployment rate and inflation. He argued that, whereas the natural rate of unemployment is determined by the real factors described in the passage quoted above, deviations from the natural rate are monetary phenomena: „I use the term “natural” for the same reason Wicksell did – to try to separate real forces from monetary forces‟ (Friedman, 1968, p. 9). 2 The unemployment–inflation tradeoff Friedman‟s „natural rate hypothesis‟ maintained that „. . . there is a „natural rate of unemployment‟ which is consistent with the real forces and with accurate perceptions; unemployment can be kept below that level only by an accelerating inflation; or above it only by accelerating deflation‟ (Friedman, 1976, p. 458). This view of the relationship between the unemployment rate and inflation grew out of the experiences of the previous decades. In 1958, Phillips had observed a negative empirical relationship between the unemployment rate and the growth rate of wages (Phillips, 1958). Understanding that high wage growth would ultimately translate into inflation, policymakers believed that there was a stable tradeoff between unemployment and inflation that they could exploit. In other words, monetary and fiscal policy could be used to drive down unemployment at the cost of a certain degree of inflation. Experience showed, however, that the relationship was not stable. As individuals started to anticipate the inflation that resulted from attempts to exploit the tradeoff, stimulative policy ceased to lower unemployment. Consequently, the Phillips curve appeared to have shifted outward, with higher inflation accompanying higher unemployment. Friedman provided an explanation for this apparent shift. Over the long run, there is an unemployment rate determined by real factors that cannot be affected by monetary policy: the natural rate. In the short run, unanticipated inflation can temporarily push the unemployment rate below its natural rate. If workers do not perceive the higher inflation, then they will respond to higher nominal wages by increasing labour supply; similarly, employers who do not immediately perceive the higher inflation will respond to a higher price for their product by demanding more labour. This temporarily lowers unemployment, but the unemployment rate returns to its natural level when workers and employers begin to perceive the inflation. As emphasized in the literature on rational expectations (for example, Lucas, 1973) that followed Friedman, inflation has no impact on real variables like the unemployment rate once individuals have already built the level of inflation into their expectations. In other words, as expectations about inflation change, the Phillips curve shifts. Although the absence of any long-run tradeoff between inflation and unemployment has gained wide acceptance, the possibility of a short-run tradeoff has 3 kept the natural rate of unemployment at the centre of policymaking. In particular, policy rules such as the Taylor rule (see Taylor, 1999) maintain that central banks can stabilize the inflation rate by assessing where the economy stands relative to economic benchmarks such as the natural rate of unemployment, „potential output‟, or the „natural rate of interest‟. When unemployment is high relative to the natural rate, and when output is below potential output, the policy rules call for stimulative monetary policy. However, several important questions arise when one contemplates the usefulness of the natural rate of unemployment as a policy benchmark. First, although the natural rate clearly cannot be observed directly, can it be estimated with enough accuracy to be useful for policy? Or do movements in the natural rate itself make it too difficult to distinguish the natural rate and deviations from the natural rate in a sufficiently timely manner to be useful for policymakers? Second, rather than focusing so much on deviations from the natural rate, should policymakers also focus on policies that would alter the natural rate, either at low frequencies or perhaps even at business cycle frequencies? What would those policies be? Identifying the natural rate Although the natural rate is often simplistically described as the long-run average unemployment rate, economists widely recognize that this rate varies over time. Friedman (1968, p. 9) was clear on this point: To avoid misunderstanding, let me emphasize that by using the term „natural‟ rate of unemployment, I do not mean to suggest that it is immutable and unchangeable. On the contrary, many of the market characteristics that determine its level are man-made and policy-made. . . Improvements in employment exchanges, in availability of information about job vacancies and labor supply, and so on, would tend to lower the natural rate of unemployment. Friedman (1968, p. 10) further argued that the mutability of the natural rate of unemployment significantly reduces its policy usefulness: 4 What if the monetary authority chose the „natural‟ rate – either of interest or unemployment – as its target? One problem is that it cannot know what the „natural‟ rate is. Unfortunately, we have as yet devised no method to estimate accurately and readily the natural rate of either interest or unemployment. And the „natural‟ rate will itself change from time to time. Since Friedman‟s work, however, economists have achieved additional understanding of some of the factors that contribute to low-frequency fluctuations in the natural rate of unemployment. It is now generally understood that demographic changes can have a significant impact on the natural rate of unemployment (see Shimer, 1998). For instance, young workers experience substantially more job turnover than more experienced workers, with the spells between jobs often spent in unemployment. Accordingly, when younger workers make up a larger fraction of the workforce (as they did in the 1970s when the baby boom generation entered the workforce in significant numbers), unemployment will be higher on average. Nevertheless, it is not clear whether this greater understanding of the factors that affect the natural rate can be translated into an estimate of the natural rate that is accurate enough to be useful for policy. Often changes in the natural rate can only be detected with a significant lag, after which time a policy response may actually increase volatility by causing the economy to overshoot its target. Further complicating the question of the natural rate‟s usefulness as a policy benchmark is the question of whether even higher-frequency (that is, business cycle) fluctuations in the unemployment rate could in fact represent movements in the natural rate. For example, modern search theory views unemployment fluctuations at business cycle frequencies as movements in the natural rate, in the sense that they result from real rather than monetary forces. Evidence from data on job flows shows that jobs are constantly being reallocated across firms, industries, geographical regions, and so on (see Davis, Haltiwanger, and Schuh, 1996). Moreover, periods of above-average unemployment rates tend to coincide with an increased level of this reallocative activity. In this sense, unemployment rate fluctuations at business cycle frequencies can be viewed 5 as the outcome of real phenomena of the type described in Friedman‟s famous quote – that is, as cyclical movements in the natural rate. This emphasis on the real determinants of movements in the unemployment rate is part of the broader view that a significant portion of economic fluctuations reflects real factors as opposed to monetary phenomena. The vast real business cycle literature has explored this proposition since the seminal paper by Kydland and Prescott (1982). Hall (2005b) argues that real fluctuations, and the difficulty of distinguishing them from monetary phenomena, render useless the various benchmark concepts such as the natural rate of unemployment, potential output, and the equilibrium real interest rate. Optimality of the natural rate and policies to alter it If real sources of unemployment fluctuations are in fact as important as monetary sources, then the proper response by monetary policymakers to the fluctuations is much less clear. However, even if unemployment fluctuations are primarily driven by real factors, it would be incorrect to conclude that either the level or fluctuations of the natural rate are optimal. Accordingly, there may be a role for policy to improve welfare by affecting the natural rate (either at low frequencies or perhaps even at high frequencies). This suggests that research on the optimality of the natural rate, and on policies that can affect it, is as important as research aimed at detecting and proposing policies to counteract deviations from it. The idea that the natural rate can be either too high or too low has been a primary focus of modern search and matching models of the labour market. In those models, the process whereby workers and firms meet may be subject to various externalities. When a worker chooses to search for a job, it has a positive externality on the probability that employers will find a suitable worker and a negative externality on the probability that other workers will find a job. Employers‟ search decisions cause similar externalities. Hosios (1990) analyses the conditions under which, in a broad class of search and matching models, the various externalities result in an unemployment rate that is either too high or too low. He finds that in general there is no economic force that draws the unemployment rate towards its optimal level. One suspects that the wage might play that role. When employers decide whether to open job vacancies (the number of which 6 ultimately determines the unemployment rate), they anticipate the wages that they will have to pay and the profits that they will earn when they form an employment relationship. However, the level of those wages and the resulting profits are determined after the fact by bargaining between workers and firms who have been matched, and who are not contemplating the impact that their bargain has on firms posting new vacancies. If the wages that result from bargaining are too low (high), firms anticipate this and create many (few) vacancies, and the unemployment rate is inefficiently low (high). As a complement to this more theoretical examination of the optimal level of the natural rate, there is a more applied literature that tries to understand cross-country differences (particularly between continental Europe and the United States) in the average unemployment rate and how those differences relate to various policies. For example, Hopenhayn and Rogerson (1992) examine the impact of firing costs on unemployment and on productivity. They find that, in addition to increasing average unemployment, firing costs reduce productivity by impeding the reallocation of workers towards more productive employers. Ljungqvist and Sargent (1998) argue that the interaction between generous unemployment insurance in many western European countries and an increased turbulence in labour markets can explain the secular rise in European unemployment rates relative to the US rate over the last several decades. In addition to this work on the determinants of average unemployment rates in the long run, recent work has also focused on trying to better understand the sources of non- monetary movements in the unemployment rate over the business cycle, and whether they are efficient. What real factors contribute to spikes in unemployment, and why is the subsequent recovery so slow? Pries (2004) argues that the slow recovery occurs because workers who lose their job in the initial spike may pass through several short-lived jobs, and several intervening unemployment spells, before ultimately settling into more stable employment. In this environment, policies that try to accelerate a recovery may be counterproductive if they encourage worker–firm pairs to hang on to low-quality matches. Shimer (2005), on the other hand, argues that the slow recovery of the unemployment rate during economic downturns results from a significant reduction in posted vacancies and, consequently, a decline in workers‟ job-finding rates. More 7 research is needed to understand the causes of the decline in posted vacancies. The canonical Mortensen–Pissarides (1994) matching model, in which wages are flexibly renegotiated as part of a Nash bargaining solution, struggles to produce a sizeable decline in vacancies during recessions. In the model, wages fall considerably during economic downturns, and the lower wages mean that firms still find it quite profitable to post vacancies. This model‟s failure to deliver the observed cyclicality in vacancies leads Hall (2005a) to suggest that in fact wages are much less flexible than assumed in Mortensen– Pissarides (1994). If so, then should the fluctuations be seen as monetary in nature, and is stimulative monetary policy the correct policy response? Or are tax incentives for investment, which may spur the creation of new jobs, a better policy response? As with countercyclical monetary policy, tax incentives may take effect with a lag and exacerbate fluctuations. Milton Friedman‟s assertion in 1968 that there is a natural rate of unemployment that is determined by real economic forces and is impervious to monetary policy has become relatively uncontroversial. Nevertheless, important unresolved questions about the natural rate remain. What is the optimal natural rate? To what extent do unemployment rate fluctuations reflect movements in the natural rate as opposed to deviations from it? What policies, if any, are appropriate for counteracting movements in the natural rate or deviations from it? Michael J. Pries See also Friedman, Milton; Phillips curve; real business cycles; search models of unemployment; Taylor rule Bibliography Davis, S., Haltiwanger, J. and Schuh, S. 1996. Job Creation and Destruction. Cambridge, MA: MIT Press. Friedman, M. 1968. The role of monetary policy. American Economic Review 58, 1–17. Friedman, M. 1976. Nobel lecture: Inflation and unemployment. Journal of Political Economy 85, 451–72. Hall, R. 2005a. Employment fluctuations with equilibrium wage stickiness. American Economic Review 95, 50–65. 8 Hall, R. 2005b. Separating the business cycle from other economic fluctuations. In The Greenspan Era: Lessons for the Future Proceedings of the Federal Reserve Bank of Kansas City Symposium, August. Hopenhayn, H. and Rogerson, R. 1992. Job turnover and policy evaluation: a general equilibrium analysis. Journal of Political Economy 101, 915–38. Hosios, A. 1990. On the efficiency of matching and related models of search and unemployment. Review of Economic Studies 57, 279–98. Kydland, F. and Prescott, E. 1982. Time to build and aggregate fluctuations. Econometrica 50, 1345–71. Ljungqvist, L. and Sargent, T. 1998. The European unemployment dilemma. Journal of Political Economy 106, 514–50. Lucas, R. 1973. Some international evidence on output-inflation tradeoffs. American Economic Review 63, 326–34. Mortensen, D. and Pissarides, C. 1994. Job creation and job destruction in the theory of unemployment. Review of Economic Studies 61, 397–415. Phillips, A. 1958. The relationship between unemployment and the rate of change of money wage rates in the United Kingdom, 1861–1957. Economica 58, 283–99. Pries, M. 2004. Persistence of employment fluctuations: a model of recurring job loss. Review of Economic Studies 71, 193–215. Shimer, R. 1998. Why is the U.S. unemployment rate so much lower? In NBER Macroeconomics Annual, vol. 13, ed. B. Bernanke and J. Rotemberg. Cambridge, MA: MIT Press. Shimer, R. 2005. The cyclical behavior of equilibrium unemployment and vacancies. American Economic Review 95, 25–49. Taylor, J. 1999. Monetary Policy Rules. NBER Conference Report series. Chicago and London: University of Chicago Press.