Using Autocorrelation Characteristics To Pick Stocks
Autocorrelation characteristics of Petroleo Brasileiro [PBR.A]
Autocorrelation characteristics of Statoil [STO]
Autocorrelation characteristics of Google [GOOG]
* click on images for full-size and details.
It is a fact that a stock's past price affects its current price. i.e. A stock's price is correlated with itself for any given period of lag. A less technical explanation of this phenomenon would be that investors look to a stock's past price as a guide for future prices. The 'collective memory' of a stock's investors thus manifests itself as its autocorrelation characteristic.
The degree of autocorrelation [AC], and the total effect over time varys greatly with each stock. In the context of ValuEngine's Valuation model, a high autocorrelation level indicates high persistence in over or undervaluation, and a long period to correct the mispricing. If we know the AC characteristics of a stock, it can be of help in our stock-picking. For example, if we compare the three Oil companies above:
On average XOM takes a long time to correct any mispricing. So even if you bought XOM while it is Undervalued, it may stay Undervalued for a very long time. On the other hand, with a stock like Brazil's PBR.A, mispricing gets corrected in a much shorter time. Therefore it is a good stock for shorter term trading, buying low and selling high. Stocks with characteristics of XOM are usually big Blue Chips or Index components. Their huge daily trading volume, and the fact that they are being held by Mutual Funds, and other institutionals whose copycat Buy/Sell signals are triggered and executed by programs causes them to have high persistence in overvaluation/undervaluation. You will find that other big caps like Wal-Mart, General Electric, Pfizer also have XOM's AC characteristics.
The AC characteristics of STO, the Norwegian State Oil Company and GOOG are also shown. In the case of STO, it has strong tendencies of high negative autocorrelation- i.e. it's current price is negatively correlated with its past price!
Comments
Post a Comment