This Time is Different: Predicting the next Recession
Updated: Aug 21, 2020
11 years on, the calamitous events of the global financial crisis seem like a bygone time. As the Dow Jones continues to soar toward an all-time high and unemployment matches a 50-year low, the allure of riches has once again tipped the balance against caution. Yet films like 'The Big Short' warn of a resurfacing of opaque securitisation products, while books like 'Too Big to Fail' preserve a detailed record of the collapse and lessons therein. Which lessons should we heed?
This article analyses two metrics empirically correlated to recessions: the economic cycle and US Treasury yield curve.
The Vicissitudes of Economics
An economic cycle involves four stages: expansion, peak, contraction, and trough. Since 1950, the US has experienced 10 full business cycles, with an average duration of 5.5 years. The current state of expansion (since 2009) is the longest in recorded history.
Statistically, this means that a recession is long overdue. While regulators and investors can make future decisions with the benefit of hindsight, such improvements are hindered by factors such as procyclical fiscal policy, regulatory capture and most importantly, that the nature of each recession is different from its predecessor. In the aftermath of the Dot-com bubble, how many could have predicted the next crisis to be in the form of mortgage-backed securities?
I Yield! I Yield!
The yield curve has been one of the most consistent predictors of historical economic downturns, with an 'inversion' having preceded every economic recession since 1950 (more on that below). Essentially, the curve plots yield (rate of investment return) against maturity (length of time until the investor is paid a security's face value) of US Treasury Securities, as below.
In a normal period of economic growth, the yield curve is upward sloping, indicating that short-term bonds provide lower returns than long term bonds. This is consistent with the liquidity premium theory.
On rare occasions, however, the curve undergoes an 'inversion', resulting in some longer-term yields falling below shorter-term yields. This phenomenon occurs when the market predicts poor economic performance in the near future, causing investors to flock toward long term bonds, thereby driving up its price and lowering its yield.
Common comparisons exists between 2-month/3-month and 10-year securities, colloquially termed the '2-10' and '3-10' spreads. Notably, a 2-10 inversion has preceded every economic recession since 1950.
There are two factors to consider when using this information to decide on a future investment strategy. First, is there any chance that this time is different? If not, when is the best to adjust strategies? While recessions have followed inversions without fail, the lag time has in some instances been as long as 34 months. Moreover, markets continued to rally an average of 15% prior to a recession and after an inversion event.
The most recent 2-10 inversion occurred in August 2019, just shy of 7 months to date. A much deeper 3-10 inversion occurred around March 2019, just shy of 12 months. Data from the previous 7 recessions show an average lag time of 12 months with a standard deviation of 3.83 months. That's an awfully close call.
Or perhaps, in our fervent stupor, we will continue to proclaim that 'this time is different'.