Thursday, May 14, 2009

Yield Curve to Predict Equity Markets? Yes and No

Interesting article in Bloomberg about the yield curve accurately forecasting economic conditions:

  • First, it’s a leading economic indicator, officially added to the index designed to predict the economy’s ebbs and flows in 1996. It was a leader well before that, even though it was unofficial.
  • Second, what you see is what you get. The spread is never revised, always available and in no way proprietary.
  • Third, and most curious, the majority of economists don’t get it. They see rising bond yields in isolation -- without paying attention to what that price-setter, the Fed, is doing at the front end of the curve.
  • It’s the juxtaposition of short and long rates, not their level, that conveys information about monetary policy.
Crossing Wall Street makes the case that the yield curve can also help predict the stock market:

Two years ago, I looked at the impact of the yield curve on the stock market and I was stunned to find:

Probably the most fascinating stat is that all of the stock market’s net capital gains have come when the 10-year yield is 65 or more basis points above the 90-day yield (that happens about 70% of the time). The yield curve hasn’t been that positive in 15 months.

Anything less than 65 basis points, including a negative yield curve, works out to a net equity return of a Blutarsky. Zero Point Zero.

Today the spread is out to nearly 300 basis points.

Yes, over the past 25 years (all the data I was able to pull) a flat yield curve did equate to a poor performing equity market over the subsequent two year period. However, a high spread between the 10 year and 90 day yield did not necessarily mean strong returns were on the horizon (see 1992 and 2002 noting that the red line indicate the two year FORWARD return).

What did? Sustained periods of a steep yield curve.

So the question becomes, what type of economic conditions usually persist in order to have SUSTAINED periods of a steep yield curve? My quick answer... expectations of higher growth for an extended period that is reinforced by a rebound in the economy (allowing long rates to STAY higher than short term rates).

My concern this time around isn't around short-term rates (they will likely stay low), but long term rates may now be influenced not only by expectations of economic growth, but by issuance and inflation expectations.


  1. Very interesting piece.

    To address the issue in your last para., could you discount the long-term spreads by the % change in M and/or inflation expectations to arrive at an "adjusted" LT rate that would help provide some predictive insight?

    Just a thought.

  2. so a "real" yield spread (i.e. nominal less inflation) for the 10 year. interesting. will do when i have the time.

    the interesting thing... although i think inflation will be a problem, it is not currently priced in, though i do think:

    1) long rates will come down in the short-run
    2) deflation will be a larger than expected problem in the short-run
    3) inflation will be higher than expected in the long run

  3. I took the daily 90/10 credit spread from early 2000 thru today and ran it through a self created oversold/overbought indicator.

    I then compared it to the S&P500 idx in the same indicator.

    There's an apparent leading indicator quality of 3-5 weeks on the part of the credit spread.

  4. VA Voter- sounds interesting. Any data you can share?

  5. Jake--

    To follow up on my earlier comment on "adjusted" (aka: real) interest rate comparisons, it strikes me that one would also have to discount the short-term rate as well to be reasonable. This might be done with current CPI or GDP deflator.

    Hope you get a chance to put the numbers together--especially if they provide some predictive insight.