Friday, October 1, 2010

Evaluating investments: theory vs. practice

Let’s say you or your company is deciding whether to make some sort of new investment. How do you go about deciding whether it makes sense? The typical advice from economics / finance professors is to perform some sort of net present value (NPV) analysis. Project all of the proposed cash flows from the project, properly discount future cash flows to reflect the fact that a dollar received in the future is worth less than a dollar received today, and calculate the net. If the NPV is positive, you’ve got yourself a good investment. Live long and prosper.

Despite the masses of MBAs bludgeoned with the importance of NPV analysis over the years, other techniques remain in practice. In a very interesting paper that appeared in the Journal of Financial Economics, two Duke professors examined how real CFOs make decisions and found that over half of surveyed CFOs used the payback period as a technique to evaluate projects. Interesting because, as any first year MBA who has completed core Finance can tell you, the payback period technique can lead to incorrect decisions.

Theory sure is nice, but practice is often very different.

[I should add a few caveats related to NPV analysis. First, it’s not entirely clear what discount rate you should use – the standard story from finance is to use your weighted average cost of capital, but not everyone agrees on how this should be calculated. Second, real options theory would argue that simple NPV analysis misses some critical values flowing from the option nature of projects.]


  1. This is a very interesting phenomenon that we see in investment decisions of all types. Cash Cab, a TV show I watch on occasion, is a game show where unsuspecting cab customers are asked trivia questions for money. If they make it to their destination without 3 strikes, they get one final video bonus question. The payoff is a classic double-or-nothing. If they get the question right, they double their winnings. If the get the question wrong, they walk away with nothing. If the contestant has gotten to the final question, then he or she has answered many more questions correctly than incorrectly. Since there is evidence that the contestant is fairly decent at trivia, he or she should assume that the odds of getting the next question correct are fairly high. If we assume getting a correct answer with a probability of 0.75, the contestant should attempt the bonus question. From my experience with the show, nearly walks without trying the bonus question. The expected value of attempting the bonus question should be higher than just walking.

    Assume the same scenario but this time the probability is 0.50 of getting a correct answer. You should be indifferent to either decision. But ask yourself, which would you take? This is where risk aversion comes into play. The result of your decision is known almost immediately. Upon answering the question you are told you have or have not answered the question correctly.

    Lottery lump sum pay-outs are based on a discount factor. A rational actor would take the pay-out if the discount factor used to calculate the payout is lower than his or her cost of capital and take the scheduled payments if the discount factor is higher than his or her cost of capital. Taking the data from you can see that the going discount rate for lump sum pay-out is roughly 3% (there are tax consequences also). The historic real total return on the S&P 500 (adjusted for inflation) from the year 1950 through 2009 was 7.0%. So, why would anyone take the structured payments instead of taking the lump sum and investing it in the stock market?

    The outcomes of the structured payments, walking away without attempting the video bonus, and the CFO accepting the investment with the shortest pay-back period all have more narrow probability distributions. In these decisions, you are protected from what Nassim Taleb calls Black Swans. If the improbable happened in any of these situations, the outcome would be devastating. The cab show contestant would lose a wad of cash, the lottery winner could be broke again, and the CFO could be fired and may have a hard time finding gainful employment. Discount rates account for known risks. However, we are still vulnerable to unknown risks. Therefore, the risk averse chose a bird in the hand instead of the two in the bush.

  2. I suspect this has more to do with the incentives managers have than what is "best" for the company's long-run value. The typical first question a CFO has is: "Is it dilutive in this year?" If the answer is yes, good luck. That means someone else is going to have to make up the "miss" in order to manage earnings...

  3. This is an awesome post, and a great site. I thought posting a ton of little things would engage my readers