©2005 All content on site protected by copyright
Diffusion of Innovations
The diffusion of ideas, markets, industries and companies that involve innovations is one of the most fruitful areas for the spec-investor to study. The approach should start with a good book on methodology and in this context I have found the 86 page book Models for Innovation Diffusion by Vijay Mahajan and Robert Peterson very helpful. Another outstanding resource is the website of Roger Clarke, devoted to all manner of research on how technology, ideas, artifacts, and techniques "migrate from creation to use."
The basic model is drawn from the spread of epidemics. A germ or idea is unleashed. Its diffusion is related to the number of susceptibles (S) in the population. But this number keeps getting smaller as the number exposed accumulates through time.
N(t+1) - N(t) = A + B ( S - N(t) ), where N(t) = cumulative number infected at time t.
The above equation is a good difference equation model of the process that leads to many insights for the non-random diffusions. A basic part of all such theoretical and empirical models in innovation diffusion (ID) is the familiar S-shaped curve with growth starting out slowly, speeding up at the ~20% of susceptible point and then leveling off at ~50%, with usually a symmetric descent to a 0% growth rate as the number of unexposed susceptibles approaches zero.
Preliminary study of the subject reveals a gold mine of applications with applications to predicting the whole stock market, industry rotation and impact studies of patent and research spending jumping right off the Googled pages on the subject. More generally, every day in the market there is some idea that seems to have the mojo. The idea gets adopted with a slow or fast start and then it either levels, explodes or recedes. Where does the pitch in the pinch come in?
Such thoughts revolve and enliven the creative process while sleeping and trading but are still very preliminary and I call on my readers for insights and lines of approach in this area.
The number of shares held per shareholder, or shareholding density, might be an interesting variable to evaluate the idea diffusion of owning a stock. High density would be visible in either low floating stock companies or those that have a universal or near bearish outlook consensus. Similarly a low density of shares per share-holder would be a case visible in large floating stock companies, or where a universal or near bullish outlook consensus exists.
Low and High density are descriptive terms. Maybe a beginning point is a study that evaluates a correlation between shareholder density, (over a reporting period), and returns over the same period.
If meaningful correlations are found on certain stocks then this indicator can be exploited further, even if leading to dangers of curve-fitting by identifying shareholder density ranges that have turned out to be bullish and bearish consensus territories. Instead of using these ranges as a trigger to trade or invest, it might be useful to have these at the back of the mind as the coming change in bias.
Alternatively, and possibly for the better, one could study the same data-sets to identify between what levels of share-holder density the most rapid move in prices happened, such that those levels would roughly correspond to 20%-50% of the idea stage. That would take care of blocked shareholding such as owned by managements or investment managers who never knew how to sell.
Two other similar variables of study on this paradigm could be defined as:
The beauty of the approach of diffusion of innovations appears to be the S curve that encapsulates roughly the 20% to 50% of the susceptible population.
The equation in a single expression harbors the aspirations of the trendists as well the contrarians. Turn either S or Sum of N(t) to zero and you have contrarians lapping up the situation while the trendists' endeavor remains to ride the 20% to 50% range of susceptibility.