Sunday, July 14, 2019

What do we mean by well-anchored inflation expectations?


The recent Globe and Mail commentary  by Bill Robson and Jeremy Kronick of the C.D, Howe Institute was in response to remarks made in the Canadian Senate by Sen Diane Bellemare in which she recommended that full employment be added to the Bank of Canada’s policy targets.  Sen. Bellemare recentlyresponded in the Globe to Robson and Kronick’s argument.  Much of her Globe piece consists of things like observations that other countries central banks have dual mandates, but as part of her argument,  Sen. Bellemere refers to:

 “………the view expressed on Jan. 4 at the annual meeting of the American Economic Association, by two former chairs of the U.S. Federal Reserve, Janet Yellen and Ben Bernanke, and by the current chair, Jerome Powell. They said that the trade-off between inflation and employment is no longer seen as a problem. Therefore, there is no need to achieve price stability at the expense of employment.”
The idea that the inflation/unemployment trade-off is not a problem comes out of the fact that unemployment, in both Canada and the US, has been falling to long term lows without inflation being pushed up above the central bank targets of 2%.   Actually, the business about the trade-off between inflation and employment longer being seen as a problem is not quite what was said (you can watch the AEA session here.)  What Janet Yellen actually said was that the Phillips Curve was currently very flat, and Ben Bernanke emphasized that inflation expectations are currently well anchored. 


If we’re being honest about it, saying that the Phillips Curve is currently very flat is nothing more than a formal way of saying that unemployment has been falling without inflation rising significantly.  Unless we have some idea why the curve has flattened, we’re just hiding our ignorance behind technical terminology.

The reason for the flattening of the Phillips Curve may well lie in Bernanke’s observation about inflation expectations being firmly anchored, presumably at about 2%.  The concept of an expectations-augmented Phillips Curve is a well established one.  It means that people (individuals, firms, union representatives and policy makers) form an expectation about how much prices are going to rise in the next year (say) and build that into their plans.  Multi-year wage contracts, for example, will tend to include an allowance for the rate of general inflation – the rate at which some kind of average price level is going to rise.  Firms which expect to see the demand for their product rising relative to the overall rate of increase in aggregate demand will expect to be able to raise their prices by more than the expected rate of increase in the overall cost of living.  The expected rate of inflation is, in a sense, the rate at which the general price level is expected to rise even when there’s no growth in aggregate demand, so employment and unemployment are unchanging.  Prices of individual commodities and wages of specific types of labour will then rise faster or more slowly than the expected general rate of inflation depending on the stage of demand in those markets.  You can settle into an equilibrium in which the actual (average) rate of inflation is equal to the expected rate of inflation.

The question, then, is where the expected rate of inflation comes from.  At one point we assumed that people tended to expect that inflation next year would be the same as it had been last year, which meant that if you could get inflation to drift down, expected inflation would move with it.  That seemed too backward-looking, though, so some analysts replaced it with rational expectations, which basically meant that the rate of inflation which people expected would hold next year (and remember that they’re forming their expectations this year) would be equal to next year’s true rate of inflation plus a random error term, an error term which for different people could be positive or negative but which averaged out to zero.  Again, if that expected rate of inflation declined, it would be built into pricing plans and actual inflation would decline with it (obviously I’m oversimplifying – the math could get kind of hairy looking).

But there’s still the question of where the rational expectation comes from.  It’s not enough to say that people on average have perfect foresight about inflation. 

One possibility, which seems quite plausible these days, is that people (meaning, remember, the whole range of economic agents) believe that the Central Bank is serious about hitting its target and that they believe it has the tools to do it.  So if the Central Bank says that its target is 2% inflation, that’s what gets built in as the expected rate of inflation.  If that’s so, price and wage setters are going to be wary about raising their own prices too far above 2% for fear of getting out of line with the rest of the economy and watching their customers (who are regarded as pretty price sensitive these days) head off somewhere else.  That means that when Ben Bernanke says that inflation expectations are well anchored, he may well be saying that people in general believe that the Fed is not going to let prices rise by more than 2%.

If you’re not a fan of the Phillips Curve, we can build the anchoring notion into an AD/AS model – specifically into the AS curve.

Think about a Keynesian AS curve.  It has a long, virtually horizontal segment associated with levels of output below potential output, then as it nears potential output it starts to get steeper, eventually becoming vertical.  Broadly speaking, two factors determine the rate at which the slope of the curve gets steeper as GDP rises.  One is the rate at which the value of the marginal product of labour declines and the other is the rate at which wages rise.  Most discussions of the current recovery have focused on how slow wages have been to rise.  Instead of wage increases cutting off the quantity of labour demanded, firms have proceeded to hire from the part of the population which had been classed as out of the labour force.  Part of this might have been a statistical artifact, since we usually judge labour market tightness by the unemployment rate, and there are a number of studies around suggesting that at the very least the concept of “not in the labour force” might well have changed.  But the result has been that firms have been hiring and expanding output without the kind of wage increases which had been expected, suggesting a more elastic than expected aggregate labour supply curve.  That means that such price increases we have seen have been due to diminishing productivity of labour without the add-on of rapid wage increases, and that might be happening because firms and labour both believe that the Bank is serious about the 2% inflation target.

None of this, of course, means that every small businessman is sitting there thinking: “if I raise my process by more than 2%, Stephen Poloz will tell my bank to jack up the interest rate on my loan”.  It does mean, though, that when that small businessman is doing his planning for the next year, and thinking about what kind of input price increases he’s likely to be facing, he is likely to be influenced by what he’s heard about the central bank’s inflation target, and also by how serious the Bank seems to be about that target.  If the Central Bank is clearly determined to keep inflation below 2%, that is likely to be the rate he will anchor his own plans on.
 
None of this is proof of anything, of course, but it does offer the beginnings of a testable hypothesis, to the effect that the Bank’s focus on the inflation target has, in effect, flattened the positively-sloped segment of the AS curve and enabled employment to rise by more than it otherwise would in response to rising AD.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.