7
$\begingroup$

The inverted yield curve has often been cited as a good predictor of impending economic recession. What are the reasons to explain why is it a good predictor? Is the inverted yield curve a cause or symptom of recession?

$\endgroup$
  • 1
    $\begingroup$ Please back up "The inverted yield curve has often been cited as..." with some sources. $\endgroup$ – Giskard Dec 31 '17 at 9:02
  • $\begingroup$ This asks two different questions "why?" and "is?" $\endgroup$ – Fizz Apr 7 at 12:41
5
$\begingroup$

There is a fair amount of ambiguity to this question. The first question is: what is the yield curve? A fixed income investor may refer to the yields across all maturities as the yield curve, while economists pick the difference in yields between two arbitrary maturities as the yield curve. Which two? Typical choices are the spread between the 2-year and 10-year bond, or the spread between the 3-month Treasury bill rate and the 10-year. The quality of a recession indicator depends which pair you choose.

The second question is: what country are we talking about? The yield curve is often described as a good recession indcator for the United States (which can be seen by looking at the time series of the 2-year/10-year Treasury spread, and compare them to recession dates. It is less clear how well it works for other countries.

There is a simple theoretical reason why it works. Under any modern financial model of the risk-free interest rate curve, we can see that term interest rates are roughly equal to the average of the expected overnight rate to its maturity, plus a term premium. (Technically, the average is a geometric average, since we are compounding an investment at the overnight rate.) As long as the term premium is not significantly negative, an inverted yield curve implies that the expected path of short-term rates is heading lower. The usual reason for a central bank to cut rates significantly is a recession (either domestically, or in trading partners).

In other words, if an inverted yield curve is a good recession indicator, it is just telling us that bond investors may be good at forecasting recession odds. Do we have reason to expect that they are perfect forecasters? Probably not. There’s any number of things that could distort bond market pricing, or else bond investors could herd into a poor forecast. Attempting to do statistical analysis of the yield curve without taking into account the changing institutional structure of the bond market is problematic.

I am not aware of any reason that suggests the inverted yield curve causes recessions under current institutional arrangements. (I am admittedly biased.) This is because developed country banking systes have much tighter control of interest rate risk now. This was not the case earlier (e.g., 1970s), and it may be that for some countries now, interest rate risk is not being hedged.

If you are interested in that question, I suggest asking it in a separate question, and you should specify what countries and eras you are interested in. The title of the question refers to prediction; a lot of readers would not relate that to causing recessions.

$\endgroup$
  • $\begingroup$ Hi, I'm interested in your views and have added another answer to continue the discussion. $\endgroup$ – S Meaden Feb 5 '18 at 12:14
  • $\begingroup$ I added a few comments to my answer. If you want to discuss causality, you really need to specify what countries and eras you are interested in. The answer probably depends upon how much interest rate risk the banking system takes. $\endgroup$ – Brian Romanchuk Feb 5 '18 at 15:05
3
$\begingroup$

[I can't possibly compete with the credentials of Brian Romanchuk so I write this answer simply to continue the discussion.]

Isn't there an argument based on maturity transformation?

Maturity transformation is that process by which banks create long terms loans from short term deposits.

So, equity share price analysts will take about how a steeper yield curve is good for the profits of banks because banks pay interest at a 90-day rate whilst making loans at a longer rate, say 10yr on average (think of a mix of 5yr personal loans and up to 30 yr mortgages). When the yield curve flattens, i.e. when long rates minus short rates falls then making loans becomes less profitable.

If the yield curve inverts and shorter rates are more expensive than longer rates then (ceterus paribus) what is the incentive for banks to make an unprofitable loan?

Further, at any given time loans are being paid off and loans are being created, if there is a cessation of new loans whilst old loans are being paid off then one can see how there will be a drop in M3 (the higher money supply indicators). This is capital shallowing aka. disinvestment and that would (could?) cause a recession.

So there are issues with this theory. Loans do not take place at the risk free rate, each borrower pays a rate higher based on counterparty risk. Other factors include corporate strategy where one bank intends to drive competitors out of business with deliberately predatory (and unprofitable) pricing. That is why I added ceterus paribus in the italicised paragraph above.

This is my pet theory; I have not read it in a textbook. Yield curve flattening and potential yield curve inversion arises recurrently on business television but they hardly ever explain the transmission mechanism to a recession (if one exists). Recently, I heard one economist propose my theory on Bloomberg but he was disputed by another economist, so this is not orthodox economics view.

I'd love to know the real answer.

UPDATE: In this interview with Bloomberg Ray Dalio says a "flatter yield curve causes a constraint in the economy and a constraint in lending" ~4:10 to ~4:50.

UPDATE2: New Fed Chair Jerome Powell in his first press conference says the effects of inverted yield curve on inter-mediation are hard to find in the research data but nonetheless those are issues that they'll be watching carefully.

$\endgroup$
  • $\begingroup$ If the yield curve slopes down, all that will happen is that the bank will make long-term loans at an appropriately defined average of short-term rates over the maturity of the loan. If the yield curve remains inverted for a long time, then the bank makes losses on some loans on its portfolio while making profits on others, so there's little to no net effect on bank profits. It's unexpected movements of interest rates which cause trouble for lending, since a long-term loan can be viewed as a bet on the future path of short-term interest rates. $\endgroup$ – Starfall Feb 5 '18 at 13:56
  • $\begingroup$ There's no reason to think that a fall in M3 (or any money stock variable, however defined) would cause a recession - correlations of macroeconomic variables with monetary aggregates broke down post-Volcker. There's also an ambiguity here about whether we're talking about the real (TIPS) or nominal yield curve - the two are related, but if you're going to talk about nominal yields you have to say something about monetary policy in the discussion. $\endgroup$ – Starfall Feb 5 '18 at 13:59
  • $\begingroup$ @Starfall: do you think you might add your own answer, I think that's preferable to extended comments. My contribution is meant to be discussive not didactic. $\endgroup$ – S Meaden Feb 5 '18 at 14:29
  • $\begingroup$ @S Meadan It might be best to structure a new question, or try chat. In the post-1994 era, interest rate risk is tightly controlled. This was not the case historically (e.g., start of S&L Crisis in US in late 1970s). As I noted, whether the yield curve causes recessions is a separate issue, which I want to avoid. As is obvious, the answer in 1982 may be different than 2008. $\endgroup$ – Brian Romanchuk Feb 5 '18 at 15:00
  • 1
    $\begingroup$ Returning to this answer above, if it is unprofitable for any bank to make the loan, the market rate of interest will rise. As you note, the spread between the long-term government bond will rise, and so the private sector curve can be positive if even if the govvie curve is inverted. However, the answer gets very complicated very quickly. The format of this website is that it is best to create a new specific question (with a question title that matches the question in the body of the text, which is a problem that comes up). $\endgroup$ – Brian Romanchuk Feb 5 '18 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.