Inflation in the Great Recession: The importance of a flatter Phillips Curve

Twice a year the IMF produces its World Economic Outlook (WEO). Along with data on the world economy, the accompanying report always addresses the important economic issues of the time. Last year, they produced an excellent chapter on private debt  and now they have another great chapter on the mystery of inflation in the Great Recession.[1]

As discussed in a recent IMF paper , when economies suffer a prolonged period of below potential growth, inflation will fall. The mystery of the Great Recession has been the inability of inflation to continue falling, as the large output gap would have implied actual deflation by now. So what is going on?

The WEO chapter examines the possible changing nature of the New Keynesian Phillips Curve – the curve that defines the relationship between inflation and unemployment. Using actual data from advanced economies, the IMF charted this relationship for different recessions in past decades. On the surface, at least, it seems clear that the relationship between inflation and unemployment has changed during the Great Recession (see the yellow line).

NK Phillips Curve

How they did it

After a preliminary examination of the data, the IMF looked at a more formal econometric model to test these issues. The specification included prior inflation and inflation expectations, import price inflations, to capture the impact from commodity prices for example, as well as cyclical unemployment. Inventively, to test the notion that this relationship has changed over time, the IMF allows these parameters to change over time. They do this via the technique of rolling regressions – regressions for different windows of the data. For example, the coefficient on inflation expectations will test whether inflation has become more anchored to future expectations. The coefficient on the cyclical unemployment parameter will test whether the relationship between unemployment and inflation has become more muted, in other words, that the Phillips curve has become flatter. The dataset consisted of 21 countries, with data beginning in 1960.

What they found

There are two results of this work that really stand out. Firstly, the responsiveness of inflation to unemployment has been gradually falling over time. This changing nature of inflation is shown in the chart below, which shows the parameter on cyclical unemployment. The chart reveals that this has declined over time – the Phillips Curve has become flatter.

Secondly, the relationship between current and past inflation has also weakened. This finding illustrates that over time, inflation has become much more ‘anchored’ to central bank targets and deviations from this target are having less of an impact.

 slope of the phillips curve

To further demonstrate these findings, the WEO presented the chart below. It shows the predicted results of the modelling work as compared to actual inflation. It does this in two ways, however. The red line uses parameters from the 1970s to describe inflation, which results in a prediction of deflation. This is one of Paul Krugman’s points, which he has admitted he was expecting – with such an output gap in a liquidity trap scenario, Krugman expected deflation by now. Conversely, the yellow line shows the predicted inflation uses parameters from the Great Recession, resulting in a much closer fit to the actual data.

actual versus predicted inflation

As the WEO, Krugman and others discuss, this changing nature of inflation probably reflects a couple of issues. Firstly, inflation has become much more anchored to central bank targets and this credibility has held during the recession. Secondly, there is evidence of downward nominal wage rigidity, meaning that inflation struggles to drop further at lower levels, as workers do not like cuts in nominal pay.

Citing work from Ball, Mankiw & Romer (1988), the WEO also discusses the importance of average inflation and ‘menu costs’. This is the notion that at higher levels of inflation, firms are used to changing their prices more often, so the cost of doing this is less when average inflation is low. This is yet another reason to explain the flat Phillips Curve at low levels of inflation.

Implications of a flatter Phillips curve: What does this mean for policy? Could inflation expectations become un-anchored?

A possible upside of a flatter Phillips Curve is the notion that aggressive monetary stimulus is unlikely to result in runaway inflation. This is not a given of course, but the WEO does discuss the concerns that inflation expectations could become ‘unanchored’. They looked at the US and Germany in the 1970s as they faced similar shocks but had different inflation experiences. One of the crucial differences of these two countries at this time was the institutional independence of the central bank. While the Bundesbank was set up with independence from war time powers, the US Federal reserve did not have independence due to ‘the lack of social consensus on the appropriate objectives of monetary policy’. Importantly, the WEO notes that the Bundesbank was not perfect – like the Fed, it miscalculated the output gap and the degree of cyclical unemployment. What set them apart however, and resulted in different outcomes for inflation, was the independence and ‘social consensus’ around inflation. On the surface, at least, this suggests that as long as central banks maintain their independence, inflation expectations are unlikely to become unanchored from aggressive monetary policy.

This seminal work also suggests that there is a risk of central banks under-responding to a downturn if they only focus on inflation. A flatter Phillips curve means that inflation will remain sticky even in the face of economic slack and cyclical unemployment. Consequently, this supports the framework of a dual mandate, like that of the Fed, compared to an inflation only mandate. The WEO points out this ‘flexible inflation targeting’ approach that gives weight to both inflation and output. This draws a central banks’ attention to output and unemployment gaps, not only inflation, and academics have suggested that the two are compatible in the same monetary framework.

This WEO chapter has been greatly received, and rightly so. Respected commentators like Krugman, Gavyn Davies, Martin Wolf and Ryan Avent have all spoken of the importance of this work. This is encouraging, as the more people that are aware of this WEO chapter, the better.

[1] IMF World Economic Outlook, April 2013. Chapter 3: The dog that didn’t bark: has inflation been muzzled or was it just sleeping?


1 Comment

Filed under Uncategorized

One response to “Inflation in the Great Recession: The importance of a flatter Phillips Curve

  1. Pingback: Wonk Wednesday: Phillip, Phillip, Phillipadelphia | Elastic Thought

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s