Various rumblings started at Zero Hedge and a few other fringe sites, and now essentially mainstream (not to mention emanating from such firms as, oops, Goldman Sachs) as pertains to a rather curious correlation between POMO days and market outperformance, appear to have finally gotten to such institutional stalwarts as Bank of America and its traditionally imperturbable Jeff Rosenberg (whose opinion we tend to respect). In a piece released tonight titled appropriately enough, "The POMO Conspiracy Theory", Rosenberg (not to be confused with former M-Lyncher David) sets off to debunk that POMO days have an impact on risk assets. Alas, he fails. The conclusion: "Our analysis points to the correlation, but not causality of POMO with rising stock prices." Sure enough, if one could confirm definitive "causality" of Fed intervention in the stocks markets, that would pretty much be the ballgame right there. And it appears that even his correlation results force Rosenberg to step back: "We likely are about to get a lot more days of POMO if the market’s expectations of $500bn further expansion of the Fed’s balance sheet is confirmed at the conclusion of Wednesday’s FOMC meeting. If the correlation of POMO purchases and stock prices were to continue to hold going forward as it has since August, than we should expect more frequent days where stocks go up as the Fed pumps in liquidity into the financial markets." Thank you for proving our point Jeffrey. Amusingly, at the end of his "debunking", Rosenberg, in typical banker fashion inverts the argument by 180 degrees, and says essentially that even if POMO is goosing markets, it basically creates a self-fulfilling prophecy that "can contribute to a better economic outcome" as it boosts inflation expectations. Jeffrey: a better outcome yes, but for you. And nobody else.
Anyway, here is the "debunking." We leave it up to our readers to make up their own mind.
The POMO Conspiracy Theory
Like all good conspiracy theories, the POMO Conspiracy Theory starts with an element of fact: Of the 24 days in which Permanent Open Market Operations (POMO)1 were conducted by the New York Federal Reserve since August 17th to affect the Fed’s policy of reinvesting paydowns from its mortgage portfolio into Treasuries, on 15 of these days, or 63% of the time, stocks rose. Does that mean stocks rise because the Fed is purchasing Treasuries? We think not, and find such claims in line with other theories of the conspiratorial type. Yet enough investors have asked us this question to prompt a bit more analysis. And while the facts do suggest some correlation between POMO purchases and stock market performance, other facts belie the argument of causality.
Consider the simple two-way table analysis of Figure 1. The basic observation supports the theory: on POMO days, stocks tend to go up more than they go down. Specifically on 63% of the time (15/24) stocks tend to rise when the Fed is purchasing Treasuries. Moreover, on days when stocks go down, that tends to be more associated with non-POMO days - 57% of the time (12/21).However, by the same set of observations we can also argue that of the days when stocks go up, there is nearly an equal chance that increase is associated with POMO purchases as not 48% (15/31).
Furthermore, if we focus on the 10 POMO days with a significant stock market increase (greater than 0.5% increase in the S&P 500), 6 occurred on days with economic data that met or exceeded expectations. Further casting doubt on the POMO effect on the market, on 8 out of the 15 POMO days with rising stock markets, bond prices were falling. Hence how the Fed’s purchases of Treasuries could support the stock market but not be supportive of the bond market appears at odds with the “theory” (though one further twist of the argument could be that after selling Treasuries to the Fed, investors continued selling bonds in order to – you guessed it – buy stocks.)
A corollary to the POMO theory should also apply: if POMO purchases of Treasuries positively impacts the stock market, then the size of those purchases should also be positively related to the size of the positive impact. Yet the opposite is true: the rank order correlation of POMO purchase size and SPX performance on the 15 days when the stocks went up was actually a negative 20%. That means stocks went up less on days with larger POMO purchase amounts and went up more on days with smaller POMO amounts.
While we may be skeptical of the merits of the POMO Conspiracy Theory, its mere existence is testament to the fact that “Conspiracism” is not limited to the realm of History. Our analysis points to the correlation, but not causality of POMO with rising stock prices. But we likely are about to get a lot more days of POMO if the market’s expectations of $500bn further expansion of the Fed’s balance sheet is confirmed at the conclusion of Wednesday’s FOMC meeting. If the correlation of POMO purchases and stock prices were to continue to hold going forward as it has since August, than we should expect more frequent days where stocks go up as the Fed pumps in liquidity into the financial markets. Based on the analysis above that may simply mean stocks are more likely to go up if economic performance going forward ends up generally better than expected.
Whether QE2 is causal or merely correlated to that economic performance remains to be seen. And while we have our doubts, the extent stock prices increase is what matters as the transmission of QE (both 1 and 2) into the real economy occurs through the wealth effect of rising stock prices. And that effect itself, if successful, can contribute to a better economic outcome. Such a scenario of asset inflation – as we discussed earlier (see research links in sidebar) – remains at the heart of our favorable outlook towards continued compression in risky asset spreads.