Wednesday, February 11, 2009

Don’t Invest Based on Emotional Accounting

We’re often told that we feel losses twice as strongly as gains. This tends to make us risk averse, which is a good thing when crossing the road, but may not work out so well for long-term investing. I tried a little experiment with Walmart stock to see how emotional accounting would affect the perception of this stock.

Walmart has been one of the greatest success stories in business. A $10,000 investment at the start of 1975 would be worth over $25 million today even ignoring dividends! This is an average compound return of over 25% per year for just over 34 years.

Imagine that our hero, Dave, invested his $10,000 savings in Walmart stock at the start of 1975. He then checked the previous day’s percentage change in stock price each morning. We’re assuming that all he looks at is the percentage change each day and not his portfolio value.

I downloaded a spreadsheet of daily Walmart closing prices (ignoring dividends, but incorporating stock splits) and calculated the daily percentage returns. To gauge how Dave feels about Walmart stock, I doubled the percentage on negative days.

So, a day with a 1% return would give Dave 1 unit of happiness, but a day with a 1% loss would give Dave 2 units of unhappiness. To capture this, the 1% loss is treated like a 2% emotional loss.

Because we know that Walmart had such phenomenal returns since 1975, we might think that even doubling the negative days to produce the “emotional return” would still leave the stock with a healthy return. Let’s look at the results.

It turns out that the average daily unemotional return was 0.11%, but the average daily emotional return was a 0.55% loss! Compounding the daily emotional returns for a year gave an average yearly emotional loss of 77%. Even the best one-year period had an emotional loss of 30%. If Dave never looked at anything but daily returns, he must be near suicidal even though he’s a multimillionaire now.

It’s clear that how we feel about short-term stock market moves can have little correlation with reality. It’s very important to work out the numbers rationally and not succumb to emotions when investing.


  1. Very interesting experiment. I never would have thought of it. I guess the lesson is to look at your quotes infrequently. I would assume a monthly check of Dave's WMT with the same 2:1 bias would be more positive, yearly even better. The most positive would to be checking only on the last day of the entire period.

    I check my stocks daily, but compare them to the major indexes. Thus, even a negative day could feel decent, if the benchmarks did worse. Still, I'm sure your thesis is valid in my case too.

    Now, if I didn't have access to the internet, or a regular newspaper, I would be a happier individual. Interesting thought.

  2. The underlieing (sp?) assumption is you have bought a stock worth keeping as well.

    Just think if Dave bought Nortel? I think he'd be up on the roof


  3. Gene: I repeated the experiment with monthly figures and was surprised to see how bad they look. The average compound emotional return is an 8% loss per year! This shows how important it is to look at your absolute financial position rather than just look at the change since you last checked.

    Big Cajun Man: That was the idea behind picking a stock that performed so well. An average stock (or worse, Nortel) would look much worse.

  4. Brilliant post! I'm stumbling it and encourage others to do the same. :)

  5. Preet: Thanks. I wasn't sure how this one would go over, but apparently a few people liked it.