Stock market participants can be divided into three groups. One set are foreign institutional investors (FIIs) — this list includes hedge funds as well as normal long-term investors. FIIs generate the largest percentage of trading volumes. The second set are domestic institutional investors (DIIs). They, by and large, stick to traditional methods of long-term investment. The third set is all classes of non-institutional players, ranging from the large operators to small day-traders.
Many traders track institutional attitudes. Some traders mirror it too — they buy when FIIs are buying, or when domestic institutions are buying. Some blindly buy any stock the institutions are buying. Some slice and dice institutional numbers to read nuances and develop more elaborate trading rules.
There are several assumptions underlying this trading style. One big assumption is that the institutions know what they are doing. The logic is, institutions have research teams and access to managements and hence, superior knowledge and better decision-making ability.
The assumption that institutions can translate superior knowledge and better number-crunching ability into better returns is not always borne out by performance statistics. Most actively-managed funds under-perform market indices. Some funds do consistently outperform, however. But this is over the long-term — traders are more focussed on the short-term.
Another assumption is more short-term. If an institution with deep pockets is buying a specific share in large quantities or buying equity in general, that imparts at least temporary upwards momentum as well as creating a floor for prices. This is definitely useful for a trader, if it is true.
There are many ways to look at institutional buy-sell data. Many short-term traders simply track net FII sales numbers. If the FIIs are net-positive, they assume the market will go up and vice-versa if FIIs are net-negative. Others look at DII numbers with a similar eye.
A third set of traders only go long if both FIIs and DIIs are separately net positive and only go short if FIIs and DIIs are separately net-negative. A fourth set sum the DIIs’ and FIIs’ numbers and see what the net institutional attitude is. They go long if the institutional perspective is long.
There are other more sophisticated questions that FIIs-DIIs data analysis may help to answer. For example, how often do we see a trend of sustained buying over say, 5-10-15 sessions? Do institutions persistently buy when the Nifty is below a certain valuation? Are there situations when the institutions are consistently wrong?
A lot of “data-massaging” may be required to fit individual needs. A short-term trader for example, will be interested in knowing how often the market goes up in the session immediately following one with net institutional buying. A trader with a longer-term perspective may use a 7-day or 30-day sum of net buy/sell as an indicator if that suits his time-frame. One may use cut-offs and set threshold buy-sell levels of +/- Rs 100 crore.
Obviously, one cannot even list the multiple possibilities of mining institutional buy-sell data in a simple column due to space restrictions. However, here are a few points to ponder for the interested trader.
Data is available for the 491 sessions since April 16, 2007. The market was up on 281 sessions out of those 491. There seems a very strong chance the market will go up on a session where both FIIs and DIIs attitude is separately positive. Both sets of institutions have been separately net buyers on 117 sessions. On 91 of these 117 sessions, the market has gone up.
Of course, to act upon the above involves guesswork because a trader doesn’t have real-time institutional data. However, an experienced trader can guess right about institutional attitude on the basis of order-flow. One-session lagged correlations are not strong. The market has gone up only 19 times in the session immediately following a session with net institutional buys.
The connection is strong (though not as powerful as 91/117), when we consider combined net buy-sell positions. There were 267 sessions when the sum of institutional positions was net positive though one of the two institutional sets was net-negative in 150 of these 267 sessions. The market went up on 186 out of these 267 sessions. That seems significant.
Obviously these observations are just touching the surface in terms of institutional data examination. But it does suggest that digging for deeper relationships between institutional positions and short-term market movements may be worthwhile.