Each day in the Trading Psychology Weblog, I post a reading on an indicator called the Institutional Composite. This is a measure of buying vs. selling activity within a basket of large cap institutional favorites that are frequently included in index arbitrage strategies. The method for computing the Composite consists of analyzing the relative activity at the bid vs. offer prices for each of the component stocks and then summing across the basket. As the indicator name suggests, the Institutional Composite is a single reading of the balance between buying and selling activity.
For this post, however, I will pull apart these components and treat Institutional Buying separately from Institutional Selling. My reason for doing so pertains to the recent postings here and on the Weblog in which I have become interested in buying/selling that is due to arbitrage activity vs. buying/selling that is directional in intent. Institutional Buying and Selling correlate at a near zero level going back to July, 2003 (N = 732 trading days). This suggests to me that the Buying and Selling are reflecting a mix of arbitrage and directional trade.
First let's look at strong days in SPY. We have had 104 days since July, 2003 in which SPY has been up by .75% or more. When the strong SPY days have been accompanied by high Institutional Buying (N = 52), the next three days in SPY average a gain of .29% (36 up, 16 down). When the strong SPY days have been accompanied by low Institutional Buying (N = 52), the next three days in SPY average a loss of -.01% (29 up, 23 down). It thus appears that high levels of Institutional Buying on strong market days greatly favors the odds of price continuation.
Now let's take those same strong days in SPY and look at Institutional Selling. When the strong SPY day has been accompanied by a high level of Institutional Selling (N = 52), the next three days in SPY average a gain of .18% (32 up, 20 down). When a strong SPY comes with a low level of Institutional Selling (N = 52), the next three days in SPY average a gain of .11% (33 up, 19 down). Clearly, Selling is less important to future market prospects after a strong day than Buying.
When Buying and Selling are combined in the Institutional Composite measure, information appears to be lost with respect to those strong market occasions. High and low Composite readings during strong market days don't have a significant impact on market performance three days out. Another way of looking at this is that a low degree of Selling is very different from a high degree of Buying in its market implications. In my next post, I will see if this dynamic holds up for weak market occasions.