American Economic Policy, as told by Martin Feldstein at Harvard University: Lecture 7, Monetary Policy: Business Cycles and Inflation
Today we are continuing to talk about monetary policy. We have been discussing the issue of how the Federal Reserve Bank should respond to the extent of slack in the economy: when there is high unemployment, the Fed should be increasing aggregate demand, but at full employment more aggregate demand would lead to inflation. It is a difficult problem to find this balance because of lags in the process and host of measurement issues.
On Friday the Department of Labor of released its monthly report. What did we learn? Here we are at the Fed trying to make sense of the data. [Professor Feldstein is referring to the February 5, 2016 release by the Bureau of Labor Statistics.]
Well, the good news is that the unemployment rate came down from 5 to 4.9%. For college grads it remained at a low 2.5%. And the participation rate increased, as did the employment-population ratio. These data lead you to say things have gotten stronger, and we have scope to increase interest rates in March.
But another survey on establishments also came out, and was confusing. Though employment increased this was lower than in previous months, indeed there was a big drop in the increase of employment, and this business-based survey is sending the opposite picture than that of the household-based survey.
So we dig down a bit deeper. Where did these declines come from? They all came from the service sector, with half from temporary employment and another half from transportation and warehousing. I have no clue why that happened.
Does this tell us the economy is weaker? It is hard to know.
But also from establishment data we get information on wages. If economy is at full employment we expect wages to rise, and indeed in January the annual rate was 6%, a significant rise.
All of this makes for a mixed set of factors, and is difficult for the Fed.
What tends to be the focus of monetary policy in managing these trade-offs is the short run Phillips curve. This curve offers guidance as to whether we are at the full employment point or not, and is named after the British economist Phillips who taught us that the change in the inflation rate is directly related to the unemployment rate. A higher unemployment rate implies lower inflation rates.
[Actually Phillips was from New Zealand, a very colourful character who made his way to the London School of Economics. You can get a quick portrait of him, and of his ideas, in the introductory chapter of Tim Harford’s lovely book The Undercover Economist Strikes Back: How to Run—or Ruin–an Economy. Harford calls Phillips the Indiana Jones of economics. A 2011 Journal of Economic Perspectives article by Sleeman discusses the Phillips Curve.]
The unemployment rate at which inflation is not changing is the so-called NAIRU, the “non-accelerating inflation rate of unemployment.” When the unemployment rate is at the NAIRU, then inflation is not going up, nor going down. This is estimated to be about 5 %, though it could be as low as 4.5% or as high as 5.5%.
With an unemployment rate higher than the NAIRU, then there is slack in the economy and the Fed should increase aggregate demand by lowering interest rates. But it is too simplistic to focus just on the unemployment rates. A more sophisticated version of the Phillips curve would add lags in the unemployment, oil prices, and what is happening to the exchange rates since factors such as these may also have independent impacts on the inflation rate.
Even the unemployment rate itself is not a good measure because of demographic changes resulting in changes in the age composition of the labour force, and changes in retirement behaviour. Changes in unemployment insurance rules may also independently influence the NAIRU (for example, the extension of unemployment insurance benefits during the recession is likely to have led to more unemployment, which would increase the NAIRU).
It also turns out that the unemployment rate of people who have been without a job for less than 6 months is really what matters for wage-setting and inflation. A Phillips curve estimated on the short-term unemployed and long-term unemployed separately finds that it is only the short-term unemployed is related to inflation. Alan Krueger, the Princeton University economist and former chairman of the Council of Economic Advisors, found that the short-term rate was statistically significant but not long-term unemployed. Those out of work for more than six months are detached from the wage setting process. [Here is a link to the paper, co-authored with Judd Cramer and David Cho.]
What Krueger found is that NAIRU based on the short-term unemployment rate was between 4 and 4.5 %, and right now that is down to 3%. So if Alan’s work is right then accelerating rates of price inflation is what we should be seeing.
The Fed keeps saying it has an inflation target of 2%, and currently the core PCE [personal consumption expenditures] inflation rate is about 1.2% . If energy prices stop falling, then the Fed believes inflation will return to a rate consistent with its target.
The Fed has chosen to define a target inflation rate of 2%, and other Central Banks also do this: the European Central Bank, The Bank of England, The Bank of Canada. This raises three questions we need to address: (1) why have a low inflation rate? (2) why not an inflation rate of zero as a target? and (3) Why a target as opposed to a general statement, or as in Paul Volcker’s terms an inflation rate low enough that people did not think about it?
1. Why moderate or low inflation
In early 1980s inflation rate was more than 10%, in the United Kingdom way above 10%. Why is double-digit inflation a problem?
- There is substantial evidence that high inflation rates reduce economic growth and incomes because businesses spend their time on wasteful speculation, for example trying to profit from managing inventories rather than growing the business.
- Moderately high inflation seems historically to lead to higher inflation: once you let the cat out of bag it gets going faster as it steps up expectations of future price increases, and leads to a leap frogging of wage and price setting.
- The public dislikes declines in real wages caused by even moderate rates of inflation. A typical earner may interpret a pay increase as the result of higher productivity [and therefore deserved], and view that it is stolen away by higher unexpected inflation.
- Higher inflation makes for more uncertain inflation, and this uncertainty makes planning more difficult. At an inflation rate of 2% nominal prices double in 36 years, but at six years they double in 12 years. So saving and planning for the future is more difficult. Savings can be eroded for those who can’t manage things in a sophisticated way.
- There is an interaction between inflation and taxes. Historically, tax brackets were fixed nominally, and people were moved into higher brackets by inflation in their incomes. This was fixed in 1986 when Congress automatically adjusted tax brackets.
- But there is also an associated problem of the measurement of real incomes associated with interest income and capital gains. These incomes are distorted by inflation. An advantage of lower inflation is we no longer have this problem. The tax rate applies to nominal interest rate, not to the real interest rate. This is the root of the problem with modest or not so modest inflation. Interest rates adjust to maintaining the same real rate but taxes are applied to interest earnings based on the higher nominal interest rates and this reduces the after tax return to the saver. The effective tax rate has increased. The same thing happens for a mortgage borrower, but int the opposite way leading to a lower after tax rate. This distorts the demand for housing and borrowing.
All six of these reasons suggest why a higher inflation rate is a bad thing.
But there is also a cost of bringing the inflation rate down: higher unemployment rate, and a loss of real output.
How much unemployment is needed to lower the inflation rate, and is it worth it?
The short run Phillips curve tells us that a 1 to 1.5% increase in unemployment for a year will bring down the inflation rate by 1%. But the cost is a temporary one-year increase in the unemployment rate, while the benefit is a permanent reduction in the inflation rate.
Say we are starting at an inflation rate of 4%, and want to reduce it to 2%. Then we need a 2 to 3% increase in the unemployment rate. Is this worth it? Is it worth imposing this higher unemployment rate? What is the cost of this?
Well Okun’s Law tells us that a 1 percent increase in the unemployment rate decreases GDP by 2.25%. So we need to pay a price of roughly 3% x 2.25% or about a 7% reduction in GDP. Remember GDP is a flow so this is a one time reduction.
There are different estimates of what a fall in inflation will do to GDP in the long run. A 2% decline in inflation leads to a permanent gain of real income of 1%. Is this worth it? Economists across the world argued that it was. The price was a one time loss of GDP of 7%, but a permanent rise of 1%. The cost pays for itself in 7 years. That is a pretty good rate of return. The decision to do this was a good thing.
2. Why not target stable prices, an inflation rate of 0%?
Why stop at 2%, why not push further to 0%? Wouldn’t this extra investment pay off as well? In fact, right now central banks are saying we need to pursue an easier monetary policy to get inflation up to 2%. There is a fear that pursuing 0%, literal price stability, might lead to deflation and this is viewed as having a couple of serious threats.
Deflation is seen as a danger by central banks for two reasons: it raises real interest rates, and it raises the value of household debt. Lets look at both of these
If inflation is say 2% and interest rates are at 4%, then the real interest rate is 2% (that is, 4% – 2%). What happens if inflation drops to 0%, then nominal interest rates will fall to 2%, and the real rate is still 2%. Well that’s okay. But what if there is at this point a shock of some sort and inflation falls to -3%. Well the most monetary authorities could do is drive the interest rate to zero, but that leaves us with a higher real rate of 3%.
Economist’s talk about a zero lower bound to interest rates, the view that they can’t lower rates below 0% because people would be paying the banks to hold their money. They’d be better off holding cash. Well some central banks have now actually gone to less than zero—Japan and Germany for example—but even so you can’t go much below zero without leading some entrepreneur set up shop and offer savers something better to keep their money secure.
In this scenario, real rates are now higher and this would lead to a decline in real economic activity, this decline in economic activity would lower incomes and further reduce GDP, and the process would feed on itself. There would be a downward spiral, and this is the Central Banker’s fear of deflation.
Another worry is that deflation raises the real value of debt, say the debt associated with a mortgage. My wages parallel what happens to prices, but my debit is nominal and does not change. So if there is deflation and my wages and income are falling, then my debts are harder to pay, and this leads to further downward spiral in incomes and GDP because my real debt load is higher as I cut back spending.
Is this fear of deflation a real problem or a technical problem? We don’t have much experience with this, but Japan offers an example. That country experienced mild deflation for a decade, but it did not get worse, and we didn’t see this process of a downward spiral of economic activity. This makes me think that this fear of dropping into deflation among central bankers is an exaggerated fear. Maybe central banks talk about it because it is a cover for what they really care about: employment. Perhaps they use the fear of deflation as a cover, or an excuse, to increase aggregate demand in order to boast employment.
3. Why have a public target?
Paul Volcker, the Federal Reserve chairman in the 1980s, talked about “lower” inflation in his attempt to move from double-digit to single digit inflation. He wasn’t concerned with a specific target, but just a rate of inflation low enough so that people didn’t care about it.
Economists have come to believe that a target will help businesses, who will take the target into account when making their pricing and wage decisions. But this only works if the Fed can deliver on the target, and so far central banks have not done a good job of delivering.
It just ain’t happening. The Japanese have been promising 2% for a long time, but it is not happening. Same thing in the US.
I have another worry about targeting. I think we do a terrible job of measuring inflation because we can’t account for changes in the quality and kinds of products consumers can buy. The true rate of inflation is probably lower than the actual, and we have no idea what it is.
The true inflation rate is unknown and maybe we should not be having such a fuss about a target, and simply use a broader band. Maybe like Volcker, we should just look at the drift of the inflation and adjust accordingly rather than being hung up on a particular number.
Other banks have a target, but in practice they all care about employment. Everyone has two goals but only one instrument to hit those targets. (Unconventional policy suggests that there are other instruments, but essentially we can think about just having open market operations to influence interest rates.)
If the Fed is the only player in macro-economic policy (that is, fiscal policy is not playing a role), then as the Dutch economist Jan Tinbergen taught us, we have to have one instrument for each objective we are pursuing, and we can only have one target because we have only one instrument. Next lecture we will pick up on this, and start talking about unconventional monetary policy.