Olivier Blanchard’s recent speech at the Brookings Institution event “What’s (not) up with inflation?” encouraged me to write this post.
Blanchard is still my second favourite economist (after Matt Nolan of course 🙂 ). But despite that I felt that some of the important elements of the discussion was missing, and I didn’t fully agree with some of Blanchard’s arguments on GDP inflation being used in the same way as the CPI inflation, which needs a separate post perhaps.
Today I want to discuss wage inflation, price inflation, and central bank targets.
In Blanchard’s presentation, he showed that the relationship between wage inflation and unemployment is strong, while the relationship between price inflation and unemployment has been much weaker.
Given this he concluded that wage inflation may be a better indicator for monetary policy than price inflation has been. As the “Phillips curve is now flat”.
Although he did not fully articulate this, the inference would be that wage inflation provides a clearer indication of spare capacity in the economy than the growth in prices – and as a result, this is what indicates what the output gap is (through the unemployment rate and price growth in this market, wages).
This flows naturally from earlier work by Blanchard and Gali (2010) and Gali (2010) discussing how optimal monetary policy includes the unemployment rate, and the unemployment rate can be used as a powerful summary statistic to measure the output gap.
But what are we saying here when we use terms like “the Phillip’s curve” and “target wage inflation” – I think it would be useful to think about some of the measures, and the history, of what is being said. Overall, the idea of using wages to understand monetary policy settings is much less novel than Blanchard appears to be saying in his presentation.
What did Phillips actually mean in his curve?
As noted above, Blanchard stated the Phillips curve of being flat – after comparing an inflation measure to unemployment.
To see if that is the right way to think about it, let’s go back to the original Philips curve framework. William Philips (1958) found an inverse relationship between rates of unemployment and corresponding rates of rises in money wages that resulted within an economy. Yes, this was increases in wages not consumer prices.
Following by Phillips’s paper, Samuelson and Solow related “wages increases” to “cost pushes”. As a result, they posited that the same relationship would exist between the prices charged in an economy and unemployment. The relationship they coined there was defined by a Phillips curve – and the possibility for the relationship to be unstable or variable was, even at that time, accepted.
The inflation of the 1970s showed the importance of expectations, and the instability of the Phillips curve between consumer prices and unemployment (in much the same way the opposite issue of limited inflation now is generating similar questions).
But it never undermined the idea that certain prices could contain information about the underutilisation of resources and the potential for a general shortage or general overutilisation in the economy.
Mankiw and Reis later discussed how different inflation measures could be used to signal the stance of monetary policy, depending on their size (large), sensitive to the cycle (high), sluggishness of price adjustment (slug like), and the degree of sectoral shocks (low).
They concluded that “ a central bank that wants to achieve maximum stability of economic activity should use a price index that gives substantial weight to the level of nominal wages”.
If everyone has been saying target wages, why don’t we?
Let’s go back to the original question posed in the header: “Should a central bank target wage inflation instead of price stability?”.
Despite prominent economists’ recommendations to target wage inflation for monetary policy, central banks still use price stability as a target measure. Why?
- Because it is very challenging to communicate wage inflation to the public: Stating that they will use a measure of “wage inflation” (perhaps even differing from wages directly) may make the public believe that the central bank sets their wages! This is a world where increasing political demands would be made of a central bank, in areas where they have no real long-run power.
- Because part of a central banks role is to anchor expectations of inflation: By announcing that final goods prices are expected to grow, at average, at X% households and firms can make choices knowing that this should be the average experience. Switching to a wage measure alone would sacrifice this communication device.
Communication and the information used for setting monetary policy are two separate things. To my mind, the central bank should be using wage measures (such as NZs Labour Cost Index) as a core indicator of the stance of monetary policy – but perhaps it is best to keep communicating it in terms of price stability for the reasons mentioned above.