Sunday, March 27, 2016

Random Observations


*  If you decompose market behavior into a linear, trend component and one or more cyclical components, you can identify optimal time horizons for holding positions.  Those optimal horizons are much longer than traders and portfolio managers typically hold.  Most people hold positions for shorter-than-optimal periods because of such factors as the need to trade, the fear of losing money, overly tight risk management, etc. 

*  Trends and cycles become more regular and noticeable when measured in event time, as opposed to chronological time.  A simple example of event time would be bars denominated in volume rather than minutes, hours, days, or weeks.  When assessing long-term cycles, the key is finding a way to measure event time given the variability of volume across instruments, across decades.  Such event cycles are least stable over short time horizons; much more stable over longer horizons, per the first point above.

*  Who is in the market determines how the market will move.  If we look at each transaction in each stock traded and assess whether the trade occurs on an uptick or downtick--and if we cumulate those data daily for many years--what we see is that a) buying and selling (activity on upticks and downticks) are different activities with their separate impacts on future market behavior; b) the total activity of participants (total upticking and downticking) is a good proxy for institutional participation in markets; and c) the market is significantly more likely to rise following periods of high institutional participation.  To simplify a bit, cycles occur because of the waxing and waning participation of agents that move the market.

*  Strong buying (activity on upticks) tends to fuel future buying (momentum).  Heavy selling (activity on downticks) tends to fuel future buying (value).  Weak buying and selling (little activity) tends to lead to subnormal returns per the point above.  Shifts in buying and selling over even a short time horizon can impact returns days and even weeks forward.

*  The worst traders are permabears and permabulls.  By definition, they do not adapt to market cycles.  They trade their predilections, not the objective activity of the marketplace.  

*  An argument can be made that the lows put in during January and February were part of a longer-term cyclical bottom and that the recent upswing has significantly more (event) time to run.  Frankly, I can think of many reasons for the market to weaken from here, but that's not what my analyses are telling me.  

Further Reading:  Trading Coaches As Whores
.