In reviewing the Report of the of the staffs of the CFTC and SEC to the Joint Advisory Committee on Emerging Regulatory Issues,  and the prepared testimony of SEC Chairman before the Subcommittee on Securities, Insurance and Investment of the United States Senate Committee on Banking, Housing, and Urban Affairs,  I was enthralled by the thoroughness of the investigation into the market mechanisms and structure that caused the precipitous drop and sudden loss of liquidity in United States Markets. 

The complexity of this problem is highlighted in the Next Steps section of the report near the end, and it concludes that a detailed reconstruction of the markets involving hundreds of millions or records comprising an estimated five to ten terabytes of information must be analyzed so that cross-market patterns can be detected and then the behavior of stocks and traders can be analyzed in detail in order to solve the problem of “what caused the drop?”. 

While I unquestioningly agree with the complexity involved in answering the question “what caused the drop?” I also wonder if focusing solely on that question is a bit myopic.

I find that when I am faced with a very complex problem, that if I work through a number of restatements of the problem, I will often identify other worthwhile problems to solve that will allow me to see the data differently. Often, the restatements are less complicated than the original problem. 

“Who profited from the drop?” seems to me like a very good question to ask and could bring different answers and solutions into focus.

First, since the possibility of a very large sell order precipitating the “flash crash” cannot be discounted, it is possible that following the money could lead to the origination of said sell order(s).

Second, and far more important in my opinion, is the issue raised by the SEC chairman in her prepared testimony regarding whether market professionals fully met their best execution obligations during the “flash crash”. I believe that the answer to this question and more can be more accurately discovered by following the money.  

If the markets during normal volatility can be compared to a game of musical chairs with market participants following a somewhat orderly procession until an economic factor occurs (the music stops) and there is a rush for the chairs (exits), then by comparison the afternoon of May 6^th was a game of musical chairs where the music stopped and the lights went out at the same time.  While it is important to understand why the lights went out, it is also important to thoroughly investigate the conduct of professional market participants while it was dark.  

Below are tables 1&2 combined, taken from the Joint Report on page 20.  The tables show the # trades, volume, and dollar volume for gains and losses between 2:40 pm and 3:00 pm in percentage terms. I added the columns percent of trades and percent of dollars.  

70% of all trades during this period were losses and 30% of trades were gains. Interestingly though, the dollar losses were 67% versus 33% gains. Furthermore, what stands out as a red flag is that the majority of the extra dollar gains were trade profits >90% away from the 2:40pm price.  This is a very good place to focus the lens. 

Here I show only the trades between 2:40 and 3:00pm on May 6^th with gains or losses greater than 10%.  Again, I added the percent of trades that were gains and percent of trades that were losses.

The trade data of gains and losses greater than 10% between 2:40pm and 3:00pm on May 6^th is decidedly skewed toward losses with 88% of the trades being losers verses 70% for all trades during this period.

It would be very interesting to see this subset of data (and the dollar volume data below) broken out by accounts categorized into retail and professional for comparison. I think that could provide additional insight into how the professionals conducted themselves while the lights were out.

Here is a breakdown of the red flag data from above.  Notice the marked difference in dollar gains starting around 70%-80% through > than 90%. I don’t think it is too much of a stretch to surmise that this was the buy-to-close zone for sophisticated short sellers. In a free falling market this data leaps off the page as a pretty tight grouping.

Of particular note is the detail of the > than 90% red flag data where only 9% of the trades reaped greater than 99% of dollars gained.  By examining these buy-to-close trades, and matching them with their sell-to-open counterparts, the CFTC / SEC might discover interesting information regarding market participant involvement on May 6, 2010.

Examining the events of the “flash crash” through the lens of market mechanism and structure will very likely provide additional clues about specific catalysts that contributed to the exaggerated price swings and liquidity vacuum as measured by broken trades, bid/offer spreads, self-help declarations, and outsized ETF factors.  However, in order to identify whether or not market participants conducted themselves professionally, ethically, and even legally is better viewed through a money lens. 

By focusing the money lens on a very small sampling of the market data that I was able to glean from the report, I believe I uncovered helpful information about market participants on May 6^th, 2010.