Wednesday, May 27, 2009
Is the Supplemental Liquidity Provider (SLP) Program Affecting Trade in the Stock Market?
This post is going to require a bit of background reading. First, recall the post on the increased frequency of herding days in the stock market. Also for review is the post that highlights the NYSE TICK as a measure of institutional sentiment. The Zero Hedge blog noted the shift in the NYSE TICK late yesterday and attributed the earlier buying strength to algorithmic trading, specifically the program trading associated with the NYSE's Supplemental Liquidity Provider program (SLP), much of which can be attributed to Goldman Sachs. More on SLP can be found here.
I think Zero Hedge is on to something, but I suspect there's more to the story than the directionality of TICK values. If the distribution of TICK values is elevated over time, it means that programs are being executed with an upward directional bias. SLP is supposedly a non-directional market making program; if it were actually run in that manner, it should not persistently elevate or depress TICK, since the market making would be a two-sided trade.
However, if market making programs were unusually influential in the marketplace, we should expect to see a higher standard deviation of TICK values. That is, more baskets executed should yield a wider variability in one-minute ranges of TICK readings. Those tracking TICK should see a large number of elevated TICK readings alongside a large number of depressed readings, as the liquidity provider(s) operate on both sides of the market executing their basket trades.
Recently, this increased variability of TICK readings (along with elevated volume in the ES contract) is exactly what we've been seeing during late day trade. The above chart provides an example from yesterday's trade, with a moving 20-minute standard deviation of one-minute high-low-close NYSE TICK readings. Notice how we get more programs executed as the session moves toward the New York close. Unlike the readings from earlier in the day, which were skewed positively, the late day readings featured many extreme one-minute readings on the buy and sell side--precisely what you'd expect if market makers were active in program trading.
A quick disclaimer: I am a psychologist who works with hedge fund portfolio managers, bank traders on proprietary desks, and proprietary trading firms; I make no claim to expertise when it comes to the microstructure of the market. Nor do I possess the macro perspectives of my club fighting colleague from Zero Hedge. What I can tell you, as one who trades and works with active intraday traders (including several that account for a meaningful percentage of total volume in the CME S&P 500 (ES) e-mini contract), is that the late day trade has changed recently and those changes have directly affected active market participants. Those changes include increased herding behavior (a historically unusual number of days skewed toward buying or selling) and increased market volume and volatility during the last hour of trading.
My sense, from the data above and my observations of the market, is that program trading has been affecting this market, but the impact appears to be particularly concentrated late in the trading day. If SLP were providing supplemental liquidity throughout the day, one would expect a general elevation of the TICK standard deviation compared with, say, a year ago. My analysis of historical data suggests that that has not occurred. Rather, it appears that liquidity providers are exploiting anomalies that occur toward the end of the day. I suspect these anomalies are related to three factors:
* Increased participation of proprietary (directional) traders who typically cover their positions by the market close;
* Increased intraday management of positions and portfolios by portfolio managers hesitant to hold overnight risk, given the diminished risk appetites of investors;
* Portfolio rebalancing among increasingly popular leveraged index ETFs, which may amplify existing directional moves.
Given the concentration of this program activity in the last hour, a skeptic might be led to conclude that this trading is more designed to aid the liquidity of participating trading firms than the liquidity of the marketplace. There is nothing wrong with this--unless the activity is funded in part or whole by a stock exchange working in conjunction with financial institutions backstopped by the government, creating a less than level playing field for independent traders and investors.