Monetary policy is boring. That’s good.30th August 2018
You’ve probably noticed that I no longer do as many posts on monetary policy. That’s partly because I put my better posts over at Econlog and partly because there’s currently not much to talk about. NGDP growth has been in the 3% to 5% range for 9 years, and I see no sign of anything changing in the near future (except perhaps a bit slower NGDP growth after this year, which is currently featuring above 4% growth.) While boring is bad for me, it’s good for the economy.
Today, interest has turned to the question of when the next recession will occur. The short answer is the same as always—no one can predict recessions. But I’d like to talk about the issue anyway, since everyone seems interested in the question.
I’ve recently been catching up on David Beckworth’s podcasts, and listened to a very interesting discussion he had with Michael Darda. Michael is a market monetarist who works in the investment area, and is known for having an excellent grasp of macroeconomics. He knows the data quite well and he has a rare ability to interpret macro data correctly. Lots of people are good at one, but he’s extremely good at both—no doubt partly due to his upbringing in Madison, Wisconsin.
At one point they began discussing the yield spread, which has been one of the better recession forecasting tools. I’d like to put in my two cents worth.
As you can see from the following graph, the yield curve often inverts before a recession. Here it’s important not to over interpret the correlation, as US expansions never last more than 10 years, and yield curves typically don’t invert until well into an expansion, and the lag between inversion and recession varies somewhat over time. Furthermore, yield curves did not invert before recessions in the 1933-58 period. Still it’s one of our most reliable forecasting tools.
Here are a few observations:
1. Over at Econlog, I argued that while the yield spread is pretty good at predicting recessions, it’s much less good at indicating when money is too tight.
2. It’s possible that the yield spread is reacting to changes in the unemployment rate. Consider the following hypothesis. The yield spread gets relatively flat whenever the unemployment rate falls to a level close to the natural rate of unemployment. Let’s also assume that the unemployment rate falling close to the natural rate is a good predictor of recessions:
In that case we should be worried, as the unemployment rate has recently fallen to a level that is probably close to the natural rate.
Interestingly, there is one notable case when unemployment falling close to the natural rate did not lead to a recession. In 1966, unemployment fell to 3.8% and there was no recession until 1970. But that false signal is equally true of the yield spread, which also inverted in 1966. Coincidence?
[On the other hand, the natural rate of unemployment is time varying (higher during 1975-95) and hard to measure, which makes the yield spread a better forecasting tool.]
When the unemployment rate is trending lower, it’s rational to expect the expansion to continue. It may not always continue (consider 1981) but it’s a rational forecast. And when unemployment is trending lower, the yield spread tends to be positive. That’s because investors expect the future economy to be stronger than the current economy, and interest rates are highly correlated with the strength of the economy.
Conversely, when unemployment has fallen close to the natural rate, it’s no longer rational to expect lower unemployment in the future. Indeed it’s quite likely that we’ll soon enter a recession. That’s why the yield curve gets flat, and sometimes inverts. In plain English, we never seem to achieve soft landings. But Australia frequently does, and thus it’s not impossible. The UK achieved a soft landing in 2001, and hence their 1990s expansion lasted until 2008. If they can do it, so can we.
Memo to Jerome Powell: Your mission, should you decide to accept it, is to avoid the hard landings that so often occur, but also avoid the 1966 scenario, where the Fed pulled up too soon, never landed at all, and instead soared off into the Great Inflation. To do this, you need to avoid an excessively expansionary policy, which sometime triggers the inflation that later causes overly tight policy, and also avoid overly tight policy, which can directly cause a recession. Let’s see . . . how about 4% expected NGDP growth?
Given that we’ve never had an expansion last for more that 10 years, you might wonder why I expect this one to go beyond a decade. Well, until 2006-12 we’d never had a housing crash. Until 2016, we’d never elected a lunatic as President. Until 2018, we’d never adopted a wildly expansionary fiscal policy during a period of peace and prosperity. None of those examples have any bearing on how long this expansion will last, rather they show that it’s really common for things to happen that have never happened before in the US. And notice that other countries have had housing crashes before 2006, and lunatic presidents, and reckless fiscal policies during peace and prosperity. That’s why I mentioned Australia and the UK (and there are many other such examples.) They provide an important clue that there is no inherent age limit on expansions.
Inductive reasoning can be useful, but must be handled with care.
PS. The Hypermind NGDP prediction market is currently forecasting 4.8% NGDP growth from 2018:Q1 to 2019:Q1. However, since the 2nd quarter is already in and growth was quite strong, this implies (annualized) 3.9% NGDP growth from 2018:Q2 to 2019:Q1. Monetary policy is right on course.
PPS. Because money is now boring, and my Trump posts are always stupid, please feel free to recommend other topics. On the other hand, money is the only topic on which I have anything interesting to say.
Read more about eu binary options trading and CFD brokers
Powered by Copy Trading binary options