Friday, September 26, 2008

Why Pundits Can’t Predict Short-Term Stock Movements

There is no shortage of pundits who offer predictions of what will happen to particular stocks or the stock market in general. You’ll find them on television, radio, and innumerable blogs. I’m open to the possibility that some investors can guess the long-term success of a particular business, but I simply don’t believe any short-term predictions that I hear.

Here is why: if these pundits actually get it right a significant fraction of the time, then they could make much more money investing based on their predictions instead of wasting their time as pundits. To prove this, I decided to run some Monte Carlo simulations.

Suppose that our pundit Peter predicts whether the S&P 500 will beat inflation each month, and he gets it right 80% of the time. So, 20% of the time when he picks stocks to win, he is wrong, and 20% of the time when he picks inflation to win, he is wrong. This may seem like only a mildly impressive record, but I’ll show how Peter can make himself fabulously wealthy.

Let’s say Peter decides to give up his life of wearing a bow tie on inane financial television shows and starts trading based on his insight. He starts by taking a $100,000 second mortgage on his home. (Fortunately, Peter bought well before the housing bubble began and has been conservative with his money until now.)

Each month, Peter will make his prediction and do one of two things:

1. If he thinks stocks will not beat inflation, he just keeps his money in cash for the month. In this case, we’ll assume that the interest he collects just keeps up with inflation.

2. If he likes stocks for the month, he leverages his money by a factor of 3 and buys an S&P 500 index fund. By “leverage” here, I mean that if he has $100,000, he borrows an additional two times this amount ($200,000), and invests the whole $300,000 in the index fund.

Before I can run the simulations to see what happens to Peter’s money, we need to factor in some assumptions about real-world stock performance and costs:

- The expected compound return of the S&P 500 will be 6% above inflation each year, with a 20% standard deviation.

- The cost of borrowing money is 5% above inflation.

- The cost of trading in and out of the index fund is 0.5% each time.

- Peter must pay 40% tax on his gains each year.

- Inflation is 4%. (This is only needed for the tax calculation. You have to pay tax on all gains, even the gains needed to keep pace with inflation.)

All of these costs are a huge burden for Peter to overcome with his market timing strategy. The simulations will tell us whether Peter’s 80% prediction rate can win out.

I piled all this data into my program and ran millions of possible futures for Peter’s money. After 20 years, there is a 90% chance that Peter will have between $1.6 million and a whopping $69 million! The median outcome was $10.3 million. These figures are in present day dollars taking into account inflation.

So, why would Peter bother being a pundit and give away his valuable insight? The answer is that he wouldn’t. I don’t pay attention to pundits who make short-term stock market predictions because I don’t believe they can get it right significantly more often than someone who tosses a coin.


  1. I agree, but have never run the numbers.

    Looking at the problem from the customer's end, it's easy to decide not to subscribe.

    But how should Peter operate? Can he take the chance of taking out a mortgage and use 3:1 margin? He could easily go broke quickly - before the probabilities kick in.

    Wouldn't Peter be better off selling his opinions in your scenario?

    All I'm suggesting is that I believe you need a more conservative investment process for Peter to make the numbers real. How about zero margin - just using his own (borrowed) capital?

  2. Mark: My analysis took into account your concerns. The truth is that Peter will likely go broke because he can't make his predictions with 80% accuracy. If he really could make these predictions correctly 80% of the time, his optimum margin ratio is much higher than 3:1.

    One thing to keep in mind about a margin ratio is that it must be maintained. If your stocks drop, you must sell some, and if they rise, you borrow more and buy more to maintain the ratio. This significantly reduces the odds of going fully broke.

    Don't get me wrong -- I'm not a fan of margin. All of this underscores how powerful an 80% accuracy on monthly predictions would be.

    Without any margin, Peter's there is a 90% chance that after 20 years Peter's $100,000 would grow to between $280,000 and $940,000 in today's dollars. The median is $510,000. This makes it clear that Peter is best off investing borrowed money to exploit his unusual gift.

  3. Hey Michael I have a question. Im new to investing and I think that the Markets are pretty much like legal gambling. However I don't think you can go wrong with buying Canadian bank stocks or I-Units ( I think they are Barclays I-units ) what are your views on I units in particular. I was told they are good because they invest in a broad spectrum of the markets like some mutual funds and they come with very low MER's. Thanks John

  4. John in Toronto: Sorry for the delay -- I was gone for the weekend. Keep in mind that my financial predictions aren't likely to be any better than other pundits' predictions. Canadian banks have done well for a long time, and I've bet some of my own money on them. I'm a believer in index funds as long as they have low fees and don't have too narrow a focus. The most well-known index ETFs are broad stock market indexes, but there are index ETFs that invest only in one narrow industry. As with any other type of investment, it's a good idea to spend some time to understand the I-units you plan to own.