Institutional Buying and Selling During Weak Market Days

Following up on the previous post, we're now looking at occasions in which SPY is down .50% or more on the day since July, 2003 (N = 154). Recall that we're looking at a basket of large cap stocks favored by institutions and the tendency to transact business at the offer price (Institutional Buying) vs. the bid (Institutional Selling).

When we have a down day in SPY but Institutional Buying is relatively strong (N = 77), the next day in SPY averages a gain of .22% (50 up, 27 down). When we have a down day in SPY, but Institutional Buying is relatively weak, the next day in SPY averages a gain of only .10% (45 up, 32 down). When Institutional Buying is positive on a weak SPY day (N = 55), the next day in SPY averages a gain of .27% (37 up, 18 down).

Thus, as we saw with the strong days in SPY, heavier Institutional Buying is associated with more bullish short-term outcomes.

How about Institutional Selling? On a next day basis, it doesn't seem to make a difference. Three days out, however, we begin to see a pattern. When SPY is weak and there is light Institutional Selling (N = 77), the next three days in SPY average a gain of .33% (50 up, 27 down). When SPY is weak and there is strong Institutional Selling (N = 77), the next three days in SPY average a gain of .18% (45 up, 32 down).

Once again, we see that Selling is less robustly related to short-term outcomes than buying. Once again, also, combining Buying and Selling into a Composite loses information in the short run. When SPY is weak, the Composite of Buying and Selling does not have a meaningful impact 1-3 days out.

The moral of the story is, as Ayn Rand admonished, to check your premises. My premise that a Composite measure would outperform the individual components seems to be in at least partly in error. Oddly, when we consider the Composite independent of price change, patterns do emerge. I'll summarize those in the next post, with comments in tonight's Trading Psychology Weblog.