Krugman, Feldstein and POMOs: oh my! How QE impacts stock and bond markets

Zero Hedge makes an excellent point about the relationship of QE to interest rate levels. In it, he compares the analysis of Krugman and Feldstein over the relationship between Fed open market operations and the low interest rate environment that has plagued the fixed-income landscape for so long now. While Krugman analyzes the Fed’s balance sheet (dollars of debt purchases) against the nominal 10-year yield, demonstrating little relationship between the two, Zero Hedge points to analysis producing a 94% correlation between the Fed’s 10-year equivalent holdings and the real (inflation-adjusted) 10-year yield. Clearly, the economic reality of the thing, which seems so plain at first blush, is only obvious when analyzed properly.

Now this got me thinking: this is all well and good, but for the everyday investor deciding how to make allocations, the 10-year yield is indeed more important than the real 10-year yield, which doesn’t directly affect the price of assets he holds. And of course, neither says anything of the stock market, a never-more-important playground for investors when yields are so heavily depressed. So I set out to analyze the impact of the Fed’s open market operations on a.) the S&P 500 and b.) BND, the Treasury bond ETF.

The following points will compare the behavior of these two assets during two periods: the first is the year-and-a-half of active bond purchases by the Fed, and the second is the year-and-a-half in which bond purchases were paused or increasing only mildly. It is important to mention that part of this second period does include purchases that would, on a 10-year-equivalent basis, match the criteria for the first period, but don’t when not duration-adjusted.

Point 1:  average daily returns

During Fed buying, SPY returned an average of 10 bps with a standard deviation of 139 bps, and BND return an average of 1 bp with 29 bps sigma.

During the pauses, SPY return an average of 3 bps with a standard deviation of 139 bps, and BND returned an average of 3 bps with 23 bps sigma.

Consider the histogram of daily returns of SPY during Fed buying (left) versus during the pauses (right):


Clearly, when the Fed is buying actively, the stock market is skewed more to the positive side; when they’re not, the cheap tail is fat and the skew is closer to center.

Point 2: compounded total returns

For stocks, this is perhaps the easiest way to see the effect of Fed buying.

During the buying, SPY returns a total of 47%.

During the pauses, it returns a total of 10%.

Clearly the buying is a boon for stocks, and the reasons are somewhat obvious. Increasing bond prices twist investors’ arms to take risk, and presumably, they do. What’s less clear is the impact on bonds.

During the buying, BND returns capital gains of 5.3%.

During the pauses, it returns 12.4% — even more than stocks!

There are only a couple reasonable explanations for me — each successive round of quantitative easing makes the policy look more permanent, and therefore investors pick up the slack of the Fed in anticipation of the next round of QE.  This seems to jive with Krugman’s central point, which is that the bond market itself is not entirely driven by Fed purchases. However, he neglects to consider my explanation of investor reaction to the Fed’s continued support of that market.

Further in support of my explanation are the histograms of BND during the buying (left) and during the pauses (right):


BND has much more skew during Fed operations, while its behavior is more sedate — slow and steady, so to speak — during the pauses. These more palatable risk properties due to the Fed’s explicit absence could increase the investor’s subjective risk-adjusted utility analysis, making them more inclined to invest in bonds when they simultaneously believe the Fed will once more support the asset class, but isn’t actively manipulating it at present, thus making it a safe time for entry into a position.

Point 3: tail risk

For all the impact the Fed does have, it seems to do little for tail risk in stocks.

During the buying, the probability SPY would lose more than 100 bps in a day was 16.6%.

During the pauses, the same probability is 17.3%.

Bonds, however, enjoy slightly better tail risk properties during the pauses, for reasons that should be clear to all:

During the buying, the probability BND lost more than 20 bps was 19.7%.

During the pauses, the same probability was 14.4%.

Point 4: correlations

A final point seems worthy of mentioning.

During the buying, BND and SPY were correlated -18.6%.

During the pauses, they were correlated -56.7%.

This extreme change in correlation makes me wonder whether this is even more support of investors picking up the slack when the Fed leaves the bond market. The increased negative correlation points to funds flowing from stocks into bonds.


My conclusion is that QE does in fact, for good reason or otherwise, support the equity markets, particularly when the Fed is actively buying bonds. In doing so, however, they are making the bond market more risky: while bonds have only 6 bps of extra standard deviation during the buying vs. paused periods, their probability of outsize losses is a full 5% higher. To me, it seems this cycle of Fed intervention and investor reaction has caused the strange price patterns observed here, and by Krugman, and by Zero Hedge. However, does it mean that QE has little impact on rates markets? Certainly not! Just not exactly what you’d expect.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: