I’m going to do an analysis of an indicator called the Money Flow Index. You can find how to calculate the MFI on the Investopedia page I linked. A value over 80 indicates the stock is overbought, and expected to be weaker than normal going forward. A value under 20 indicates the stock is oversold, and expected to be stronger than normal going forward.
I’ve applied the MFI to four stocks below (AAPL, BA, CAT, and XOM), for the period 2007 – 2016. That’s a 10-year period that included a major sell-off, a major rally, and some time spent moving sideways.
Before looking at the specific results, let’s discuss what we would expect to see if the indicator has predictive value. We expect that whenever the indicator is less than 20, the average gain over the next N days will be greater than when the indicator is between 20 and 80. We also expect that whenever the indicator is greater than 80, the average gain over the next N days will be less than when the indicator is between 20 and 80.
In the table I’ve included values of N for 5 trading days, 30 trading days, and 110 trading days. Let’s see how the indicator worked out:
|Days < 20||132||326||439||332|
|Days > 80||297||516||456||431|
|Days 20 - 80||2209||1796||1743||1875|
|Avg 5 day Return, <20||0.93%||0.06%||0.02%||0.59%|
|Avg 30 day Return, <20||1.21%||1.84%||1.54%||1.98%|
|Avg 110 day Return, <20||10.67%||3.32%||4.86%||4.36%|
|Avg 5 day Return, >80||0.55%||0.46%||0.44%||-0.12%|
|Avg 30 day Return, >80||5.62%||2.91%||1.29%||1.16%|
|Avg 110 day Return, >80||14.12%||9.43%||7.24%||2.49%|
|Avg 5 day Return, 20-80||0.59%||0.32%||0.28%||0.07%|
|Avg 30 day Return, 20-80||3.64%||1.60%||1.89%||1.86%|
|Avg 110 day Return, 20-80||13.99%||7.82%||7.14%||1.18%|
Taking AAPL and the 5-day return, we see that for MFI < 20 the average return was 0.93%, and for MFI between 20 and 80 the average return was 0.59%. Potentially the MFI has predictive value when it’s less than 20. For XOM an MFI < 20 also had a greater return than when the MFI was between 20 and 80, but for BA and CAT it was actually less. So an MFI < 20 doesn’t look too promising to have any predictive value on 5-day returns.
30-day returns with an MFI < 20 doesn’t look any better. In fact, for 3 of the stocks the returns were almost identical to when the MFI was between 20 and 80, and for the fourth (AAPL) 30-day returns with MFI < 20 were actually less than when MFI was between 20 and 80.
110-day returns were even worse, with 3 of the stocks having worse 110-day returns when the MFI < 20 than when it was between 20 and 80. Indeed, it may be worth investigating if an MFI < 20 is a good indicator that the stock is going to perform poorly over the next 6 months.
Maybe an MFI > 80 will have better predictive value. Recall that a value over 80 is supposed to mean coming weakness in the stock price.
5-day returns when MFI > 80 are almost identical to 5-day returns when MFI is between 20 and 80 for all four stocks, so nothing there.
30-day returns when MFI > 80 were better for two stocks, and worse for two stocks, than when MFI was between 20 and 80. Again, there doesn’t appear to be any predictive value there.
Finally, 110-day returns were almost identical for two stocks, and slightly less for two stocks, comparing MFI > 80 to MFI between 20 and 80. Maybe there’s a very slight predictive value there. Testing on more stocks would be required.
In summary, the Money Flow Index indicator does not appear to have much predictive value.
This is an analysis of past performance, but past performance is not a guarantee of future performance.
Comments / Questions: joseph AT StockMarketMovement.com