The Non-Vertical Long-Run Phillips Curve
Recent developments in macroeconomics contradict the widely held belief that permanently higher inflation cannot affect unemployment, which - according to Blanchard and Fischer (1989) - derives from a priori reasoning rather being consistent with empirical evidence.
Orphanides and Solow (1990) suggest that the long-run effects of money growth can depend on a number of assumptions, such as those of non- separable utility functions and cash-in-advance or money in the production function. Holden (2003) shows that they can also arise at low inflation rates in the case where wages are set through negotiations. According to Hughes Hallett (2000), regional or sectoral curves can have different slopes. If there is a mismatch between supply and demand in different sectors (or, by extension, in different countries or regions in a fixed-exchange-rate regime or monetary union), their aggregation always generates a long-run trade-off between inflation and output, which makes it possible to change the natural rate of unemployment by changing the sectoral or regional compositions of demand.In New Keynesian models, a long-run relationship between inflation and real activity is obtained based on price staggering (see, among others, Goodfriend and King 1997 and Woodford 2003), where the implied effects of inflation on unemployment are unambiguously adverse. This implies that the optimal long-run inflation rate should be zero, an issue that will be discussed further in Section 3.5. Various authors have demonstrated the non-linearity of the Phillips curve (e.g. Ascari 2000, 2003).
Graham and Snower (2008) show that the interaction of staggered nominal wage contracts with hyperbolic discounting inflation leads to a significant long-run effect of inflation on real variables via money growth. With their baseline calibration, which takes the length of the contract period to be 1 year, a permanent increase in inflation of 1 per cent is associated with an increase in output and employment of approximately 0.2 per cent for inflation rates of up to around 10 per cent.
This is roughly of the same order of magnitude as the empirical estimates of the long-run Phillips curve in the classical studies that have found a significant trade-off from Phillips (1958) and Samuelson and Solow (1960) to Akerlof, Dickens and Perry (1996, 2000). These authors derive a tradeoff between unemployment and inflation via a static model with downward wage rigidities. Fair (1999), instead, finds that there would have been a rather low price level and inflation costs with rather significant gains in output and decreases in unemployment if the Bundesbank had been more expansionary in the 1980s.In a dynamic stochastic general equilibrium (DSGE) model with RE agents optimally setting their wages, Benigno and Ricci (2011) investigate the macroeconomic implications of downward nominal wage rigidities in a low-inflation environment. They derive a closed-form solution for the long-run Phillips curve. This shows an inflation-output trade-off that is virtually vertical at high inflation rates but flattens at low inflation. The need for stabilisation policies derives from two circumstances: (1) the flat section of the curve implies progressively larger costs in terms of output of attempts at reducing inflation, and (2) macroeconomic volatility shifts the curve outward, generating output and employment costs. The results question the conventional view that argues against the presence of a long-run trade-off and in favour of price stability. They also reconcile models with real life, where no central bank adopts a policy of zero inflation and in the long run the inflation rate is positive. This can be explained as due to the zero nominal interest bound and, as in Benigno and Ricci's model, the presence of downward nominal rigidities.
Di Bartolomeo, Tirelli and Acocella (2014) share the view that modern monetary models may underestimate the benefits of inflation on wage mark-ups but highlight a disciplining channel based on the idea that inflation is a tax on money balances. To model this effect, they introduce money in the utility function, as in Christiano, Eichenbaum and Evans (2005).
Di Bartolomeo, Tirelli and Acocella (2014) then identify a channel that supports the inflation tax effect on money balances, which disciplines wage mark-ups. In line with a recent analysis of institutional features of wage bargaining in twenty-two European Union countries, the United States and Japan (Du Caju et al. 2008: 25), Di Bartolomeo, Tirelli and Acocella emphasise that wage contract renegotiations take place while expiring contracts are still in place, enabling wage setters to internalise their consequences for household choices. In their model, this is captured by the assumption that within each period wages are predetermined to macroeconomic variables (see Corsetti and Pesenti 2001 for a similar assumption). This allows setting a framework where wage setters internalise the effect of their wage choice on their own real money holdings. In the paper, the authors show that such an effect is negative and becomes stronger with the expected inflation rate, inducing wage setters to limit their wage claims. This is therefore a new justification for the existence of a non-vertical Phillips curve. Model simulations show that a moderate inflation rate can generate substantial output gains relative to both the Friedman rule and the commitment to price stability, popularised in standard New Keynesian models.From an empirical point of view, a growing literature has shown that New Keynesian models significantly improve their ability to replicate business cycle facts if monetary policy rules are assumed to target time-varying, non-zero long-run inflation rates (see Cogley and Sbordone 2008 and the references therein). Ireland (2007) estimates a New Keynesian model to draw inferences about the behaviour of the Federal Reserve's unobserved inflation target. His results indicate that the target soared from 1.25 per cent in 1959 to over 8 per cent in the middle to late 1970s before falling back to below 2.5 per cent in 2004. He provides evidence that is consistent with the view that shifts in the secular trend in inflation (i.e.
the expected long-term inflation rate) could be attributed to a systematic tendency for Federal Reserve policy either to limit the contractionary consequences of adverse shocks (Blinder 1982b; Hetzel 1998; Mayer 1999) or to exploit favourable economic conditions to eventually bring inflation down (Bomfim and Rudebusch 2000; Orphanides and Wilcox 2002). More generally, monetary policy should consider the possibility of unobserved shifts in sustainable growth rates (Lansing 2002). Blanchard (2016a: 31) re-examines the behaviour of inflation and unemployment andreaches four conclusions: (1) Low unemployment still pushes inflation up; high unemployment pushes it down. Put another way, the US Phillips curve is alive... (2) Inflation expectations, however, have become steadily more anchored, leading to a relation between the unemployment rate and the level of inflation rather than the change in inflation. In this sense, the relation resembles more the Phillips curve of the 1960s than the accelerationist Phillips curve of the later period. (3) The slope of the Phillips curve, i.e. the effect of the unemployment rate on inflation, given expected inflation, has substantially declined. But the decline dates back to the 1980s rather than to the crisis. There is no evidence of a further decline during the crisis. (4) The standard error of the residual in the relation is large, especially in comparison to the low level of inflation.
Each of the last three conclusions presents challenges for the conduct of monetary policy. Wisdom gained from the experience of the 1960s and later will be needed.
3.5