- Consumer staples companies have fallen out of favor versus other industry stocks during the time, most likely technology stocks.
- FDFAX's management made poor decisions in the late 1990s that detracted from investment performance but they recovered by the early 2000s.
- The Consumer Staples Industry Investment Model is incomplete.
- Something else.
Let's look at each of the theories.
Consumer staples companies have fallen out of favor versus other industry stocks during the time, most likely technology stocks.
The dot com boom was known for the big share price increases of internet, software, and technology related companies and possibly a price drop in other types of companies, most likely those in consumer staples. I'm going to test the first part by comparing the Fidelity Select Technology Portfolio (FSPTX), Fidelity Select Software & Computer Portfolio (FSCSX), and Fidelity Select Consumer Staples Portfolio (FDFAX) with the S&P 500 Index. The chart slideshow below shows the high correlations of the technology and software portfolios but the very poor correlation of the consumer staples portfolio. This suggests that the S&P 500 Index, a proxy for stock market performance, is strongly influenced by dot com stocks. Stock market performance is an input into the Consumer Staples Industry Investment Model.
FDFAX's management made poor decisions in the late 1990s that detracted from investment performance but they recovered by the early 2000s.
This isn't true. The Morningstar Chart below shows the FDFAX and Morningstar's Consumer Defensive Index moving together. If poor decisions had a major impact we wouldn't see a good fit between FDFAX and the index.
The Consumer Staples Industry Investment Model is either incomplete or something else is happening.
There is a degree of raw truth in this statement. We can reason that if the model is complete we wouldn't see such an anomaly. In terms of consumer staples related data I used there could be a unique relationship between wholesaler and retail sales/inventory that I haven't discovered. For instance the slight improvement in accuracy of the prototype V2 model over the V1 model is due to an adjustment in the way the model interprets wholesale inventory data.
The dot com boom was a period of irrationality as any company technology related increased in price. This rapid ascent also influenced the S&P 500 as many technology companies were members. From July 1, 1996 to December 1, 2001, the S&P 500 Price to Earnings ratio went from 18.26 to 46.37. There exist a possibility that factors other than consumer staples related data hurt performance during this period such as inflation, interest rates, mutual fund flows, imports/exports of retail products. For future model enhancement I will consider these data points for testing.
Although the model fails during the dot com era, the chart below shows that the prototype model is an effective predictor of FDFAX after the dot com era. The model as usable despite its misses in the dot com era. No investment model is perfect and model results always needs to be reviewed before using. This is what I'll do on this website every week when I release model results. I plan on releasing model results either today if I have time or starting next week.