|
Article Excerpt
Abstract Using a two-regime model of the inflation-unemployment process for US data 1960:2 to 2000:2, this paper finds strong evidence to support the Eisner puzzle, which occurs when the short-run Phillips curve (SRPC) is flatter at low rates of unemployment than at higher rates. The puzzling aspect of this pattern is the expectation of excess demand to become apparent at very low rates of unemployment causing the SRPC to be steep rather than fairly flat. We show the puzzle can be resolved by estimating a three-regime model which reveals a steep SRPC at very low rates of unemployment. The estimates of the three regime model also reveal a horizontal SRPC at intermediate rates of unemployment at those intermediate rates.
Keywords Eisner puzzle. Range of equilibria. Inflation. Unemployment Natural rate
JEL E00
Introduction
Soon after its initial development by Friedman (1968) and Phelps (1967), the concept of the natural rate of unemployment became the dominant paradigm in macroeconomics. However, the empirical performance of natural rate models has not been satisfactory. (1) A glaring weakness has been the failure to identify the supply factors that may influence the natural rate, such as unemployment benefits and trade union power (2). Instead, abnormal theoretcal theoretical processes, such as the Kalman filter, have been invoked to generate movements in the natural rate. Examples are found in Gordon (1997, 1998) and Staiger et al. (1997, 2001) (3). These processes are theoretical because they do not identify the supply factors explicitly. (4) One implication is the resulting time-varying non-accelerating inflation rate of unemployment (TV NAIRU) may reflect, at least in part, the effects of aggregate demand as implied by non-natural rate approaches, such as hysteresis and range of equilibria. (5)
Prominent, mainstream economists have expressed doubts about the natural rate concept. For example, Robert Hall believed the natural rate hypothesis is "a theory whose acceptance, in his opinion, [has] been a victory of theory over fact" (De Long and Summers 1988, p.492). De Long (2000, p.84), in discussing the natural rate proposition that "business cycle fluctuations in production are best analyzed... as fluctuations around the sustainable long-run trend," confessed that "I ... cling to the belief--albeit with out much supporting empirical evidence--that policy is as much about gap closing as stabilization policy." The uneasy nature of the dominance of the natural rate is reflected in Benjamin Friedman's observation, while some accept the natural rate model, those that doubt it "accept it anyway as a working tool for policymaking purposes" (Solow and Taylor 1998, p. 56). (6)
Robert Eisner was a trenchant critic of the natural rate concept (Subcommittee on Domestic and International Monetary Policy 1997). Eisner argues the natural rate approach caused US policy makers to adopt an excessively restrictive aggregate demand policy for most of the post-second world war period. In Eisner's view, inflationary pressures at low rates of unemployment were, in fact, low. He argued this view was supported by his estimates of a two-piecewise linear short-run Phillips curve (SRPC) for US quarterly data for 1965:1 to 1996:3 (Eisner 1998). According to these estimates, the SRPC in the US is shallower at low rates of unemployment than at high rates. Empirical work by the Council of Economic Advisors, referred to by Stiglitz (1997), found a similar concave pattern. (7)
Eisner (1998) suggested that four factors are responsible for a shallower SRPC at low rates of unemployment. He argued that low unemployment may lead to a more intensive use of fixed factors of production, capital-labor substitution induced by the labor shortage, reduced wage pressure because of game theoretic considerations and price moderation by firms flush with profits but anxious about new competition from entrants. The first two reasons tend to increase labor productivity, thereby reducing cost pressure on prices. The last reason is also put forward by Stiglitz (1997, p.9).
However, Eisner's argument is not entirely satisfactory because it is hard to reject the idea that inflationary pressures should intensify at low rates of unemployment. (8) Surely there is a rate of unemployment at which excess demand must become strong, swamping the counter effects of higher productivity and moderation in wage and price setting for strategic reasons, causing strong upward pressure on inflation. Perhaps the explanation favored by Eisner is the US economy has not operated in excess demand region for any significant period of time. However, this is countered by some empirical evidence based on estimating a non-linear SRPC. In particular, Laxton et al. (1999) estimate a convex SRPC and find a minimum rate of unemployment at which demand pressures become intense and the SRPC becomes vertical (9). The original estimate of the Phillips curve, for the UK, was also convex with an intensification of demand pressures at low rates of unemployment (Phillips 1958) (10).
Barnes and Olivei (2003) consider, in detail, the data on inflation-unemployment for the US since the 1960s, focussing on the possibility that the relation has different slopes for different rates of unemployment. Their examination of the US inflation-unemployment data reveals a flat spot between the unemployment rates of 4% and 7%, as shown in their Fig.2. Their estimation of a threshold model with a three-piecewise linear SRPC confirmed this impression. Their estimated SRPC for the US has three sections: the peak and trough sections are negatively sloped and represent regions of low and high unemployment, respectively; and the intervening region is horizontal and covers intermediate regions of unemployment. Barnes and Olivei interpreted this intermediate region by suggesting an extension of the natural rate concept such that unemployment has to be outside of a threshold, centred on the natural rate, in order to have an influence on inflation. In their case of a constant natural rate, their estimates implied that the unemployment threshold was between unemployment rates of 4% and 7.5%. In an alternative specification using an exogenously-determined time varying NAIRU, Barnes and Olivei found the unemployment threshold was 2.8 percentage points of unemployment around the Congressional Budget Office (CBO) NAIRU.
Although they do not refer to the Eisner puzzle, the demonstration by Barnes and Olivei (2003) of three regions for the SRPC resolves the Eisner puzzle because Eisner's gently-negative sloped SRPC at low rates of unemployment can be interpreted as the combination of two sections of the SRPC, a negative-sloped section and a horizontal section. (11)
Driscoll and Holden (2004) also estimate a three-piecewise linear model of the inflation-unemployment relation for the US. They base their empirical model on the theoretical model of Bhaskar (1990). In Bhaskar's model, the evaluation by workers of wage outcomes, according to the concept of loss aversion, implies a tendency to resist wage cuts and, thus, a range of equilibrium rates of unemployment. With Bhaskar's model, the unemployment threshold of Barnes and Olivei (2003) becomes a range of equilibria with a loss aversion providing a behavioral foundation for a horizontal section in the SRPC, although Driscoll and Holden (2004) do not refer to Barnes and Olivei (2003). From the pattern of inflation persistence implied by their estimates of the three-piecewise linear model, Driscoll and Holden (2004) infer that there is a range of equilibria for the US using data for 1955:1 to 2000:4. This range occurs between the rates of unemployment of 4.7% and 6.5%.
In this paper we re-examine, using US quarterly data for the period 1960:2 to 2001:2, the existence of the Eisner puzzle in the two-piecewise linear SRPC...
|
