Saturday, January 14, 2006

Less Talked About Properties Of Simple Moving Averages

Institutions who are buying a stock don't put their entire position on in one day. They buy over time, therefore their cost basis isn't a quoted price, it is the average cost of all their buys. While any given institution won't buy the same ammount every day, I think an assumption we can make is that in aggregate all the institutional buying is at prices close to the average. From this assumption it follows that the institutions who have been buying a stock for the last 50 days will in aggregate have a cost basis close to the 50 day simple moving average.

An important implication of this conclusion is that we've identified a reason why moving averages will often act as support for a rising stock. If an institution has been buying a rising stock for 50 days and has a cost basis near in price to the 50 day simple moving average, buying at any price under the 50 day simple moving average will lower the firm's cost basis. Buyer firms will compete with each other for shares below their current cost basis, often rapidly driving prices higher.

Another factor to consider is that firms who have been purchasing a rising stock for less than 50 days will most likely have a higher cost basis and are likely settle for higher prices that still reduce their own cost basis, thus providing additional buying competition. The reason newer buyers probably have a higher cost basis is that the shorter the moving average interval, the higher it will typically be for a rising stock.

There are obviously many other forces at work on the price. Chief among these in my opinion is whether or not the price will get cheaper before it gets more expensive from a trader's perspective on various timeframes, but that is a subject for another post.


At 8:37 PM, Blogger Brett Steenbarger, Ph.D. said...

You make an interesting observation. Let's assume, however, that time frames are relatively evenly divided among participants: shorter-term traders and longer-term ones. Each thus has a different average price/moving average that serves as "value" and a motivation to buy below that price.

If this is true, we should see a tendency for price declines to reverse across multiple timeframes.

We should also see the greatest reversals when: a) there is an expansion of volume at or below moving average points (as this would suggest an influx of value-oriented buyers), and b) when the volume transacted at or below those levels is predominantly transacted at the offer rather than the bid (indicating aggressiveness of buying by the value-oriented traders).

Both of these hypotheses are consistent with your interesting observations and both test out in my research, FWIW.

Thanks for the thought-provoking post.

Brett Steenbarger

At 1:29 AM, Blogger jontait said...


Thanks for sharing your observations. Comments like this add a lot of value to the blog.



Post a Comment

<< Home