This analysis was inspired by recent posts from World Beta and Bespoke pointing out that the spread between corporate bonds and treasuries are approaching historic levels. In this post I want to look at how the stock market reacts when credit spreads are rising versus falling (and in a follow up post I’ll look at how the market reacts when spreads are relatively high or low).
First, a look at monthly average credit spreads between BAA rated corporate bonds and 10-year Treasuries over the last 50+ years. We are clearly in an upward trend and nearing all-time highs.
All interest rate data from St. Louis FRED Database
Next, the following graph shows the S&P 500 (blue), and strategies that are long the S&P 500 following months when the average credit spread either widened (green) or narrowed (red). Additional stats follow.
All data is in monthly intervals. Chart is logarithmically scaled.
As the data makes clear rising or falling credit spreads have by no means been a magic bullet – the market has at times consistently risen or fallen regardless of the direction of spreads, but very generally speaking, rising credit spreads have been more bullish for the market over the last 50+ years (especially in terms of the size of average winning vs losing months).
In a follow up post I’ll be looking at how the stock market reacts when credit spreads are relative high or low.
Happy Trading,
ms
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Trading Strategies, Treasuries & Interest Rates | 5 Comments
Some very nice work, Michael. Not particularly surprising to this VIX aficionado (similar principles at work, methinks), but I’m sure the numbers will catch some off guard.
Cheers,
-Bill
Thanks Bill. You know I’m not a “fundamentals” guy so I didn’t put my 2 cents in on this, but I saw this as a contrarian play following months when stock market doom-n-gloom drove the spread up. Is that how you read it? michael
I like the thought process in your analysis, but I think you’re missing some important points that may distort the conclusion.
Let’s disaggregate a corporate bond into a series of Treasury strips due on the coupon and maturity dates, and the sale of a series of daily puts on the issuer’s equity. We collect spread income from these puts, but do not participate in the upside of the equity (if any). So a widening credit spread should mean greater implied volatility for the puts, which suggests a lower price on the underlying equity. Therefore your conclusion seems to be contrary to the foundations of corporate finance. Ideas why:
1. Not all S&P 500 companies have debt outstanding. Many S&P 500 companies’ debt is rated higher than BBB. Consider huge index weights MSFT (tiny sliver of debt from AQuantive buy, otherwise none) and XOM (AAA rated), which have no fundamental link to the Baa spread, and in fact may move countercyclically to it. So before we conclude that widening spreads tend to accompany above-average equity returns, we should separate the debt-free and debtor components of the index and test them separately.
2. Many equity investors use Treasurys as safe haven during equity bear markets. When a manager holding Treasurys wants to re-enter the equity markets (a better buyer, driving up equity prices), s/he is a better seller of Treasurys, driving down their prices.
3. The term structure of interest rates may need to be investigated. Many corporations borrow opportunistically based on the shape of the yield curve. When rates are low, they may emphasize floating-rate bank loan debt or commercial paper, rather than issue 10- or 30-year fixed-rate bonds. Controlling for curve shape may be helpful.
RE to Bond Investor: I really like your sentence when you say “your conclusion seems to be contrary to the foundations of modern finance”. I think I might make that my new blog slogan, but I’m going to change it to “EVERYTHING I do is contrary to the foundations of modern finance”.
I agree with everything you wrote above, but theories are not really what this blog is about. I’m a slave to the data – my job is to show that these relationships exists, regardless of why or whether we think they should exist. I leave it to much smarter people such as yourself to provide the “why” =)
Thanks for the comment.
ms