On The Predictive Power Of The "Submitted-To-Accepted" Ratio In POMO

Late in October, before QE2 was fully launched we penned a post, with the help of John Lohman, titled "The POMO Submitted-To-Accepted Ratio: A Tell On How To Frontrun The Frontrunning Primary Dealers" whose focus was the "Submitted-to-Accepted" ratio in any given POMO operation. While by now everyone is aware that POMO days (at least historically) have had a huge positive impact on stock returns (since they have been virtually daily since the beginning of the market meltup), and created their own self-fulfilling prophecy, it is the nuances in POMO that still catch people unaware. Namely, we claimed 3 months ago that the Submitted-To-Accepted ratio is a critical tell in how the market will perform through close, finding that "generic market effect on POMO days (i.e. stocks and yields up relative to non-POMO days) should be pronounced when the submitted-to-accepted ratio is relatively low (“meets expectations”) and muted when the ratio is high (“a negative surprise”, particularly if said Dealers had already positioned themselves in pre-POMO trading, based on a set of expectations regarding the outcome)." Following the surge in the S/A ratio in yestedrday's POMO, which effectively predicted the market rout, we decided to rerun the analysis. We found that recent incremental data merely reinforces the original conclusion: namely, watch out for days that have a substantially above average Submitted to Accepted ratio.

John Lohman elaborates:

The first analysis of the Submitted/Accepted ratio pointed out that above average ratios (suggesting the POMO did not ‘meet expectations’) have a negative effect on the equity market.  Since anything that might cause a red close is directly opposed to the Fed’s third mandate, they attempted to remedy the situation a short 6 days later with their November 3rd announcement that full details of upcoming POMOs would be released well in advance.  While this attempt to reduce uncertainty around the operations has managed to lower the average S/A ratio, it is interesting to note that the overall phenomenon remains: stock prices do better on days when the ratio is below average and worse on days when it is above.

The following table measures the S&P 500’s performance over the 44 POMOs since the first analysis, based on the S/A ratio of each day’s operation relative to the average. 

The last table is similar, except that it compares the ratio with the median in an attempt to account for any potential skew in the S/A series.

The chart below confirms all one needs to know empirically based on POMO S/A performance.