How to Predict Stock Prices: A New Tool for Most Investors

The overarching goal for all stock market investors is to predict stock prices. Retail (a.k.a. individual) investors try to do it, institutional investors try to do it, analysts try, academics try, anyone who is at all interested in the stock market inherently is trying to figure out what the stock market will be worth in the future. Today’s post will give you a new way to look at stock prices in the future (well new to most of you).

For analysts and most institutional investors, this is usually done through some combination of discounting cash flows, and talking to management to try and get a read on future cash flows. For individual investors it usually means relying on the work and advice of analysts.

Unfortunately, analysts have a mixed record of success in general. In particular, analysts frequently tend to be overly optimistic. That’s not my opinion – it’s a fact that has been borne out by dozens of empirical studies. I will write more about that in a future post. But today, I wanted to talk a little bit about how financial economists, like me, look at stock prices.

Now before anyone gets up in arms over the idea that academics are better stock pickers than other investors, let me ease your mind – They aren’t! In fact, financial economists rarely even attempt to “pick” specific stocks. Instead, what we do usually involves trying to determine the characteristics of companies that will be good investments in the future, and the characteristics of companies that will be bad investments in the future.

It’s a lot like your car insurance company. Car insurance companies don’t try and decide if you personally are going to get in an accident. Instead, they have identified a bunch of characteristics associated with drivers more likely to get in an accident – stuff like DUIs, speeding tickets, experience in driving, etc. Then they use these characteristics to say on average whether a certain driver is more or less likely to get into an accident.

In the same way financial economists identify characteristics of firms and then determine how these characteristics have affected stock price in the past and how they are likely to affect stock price in the future. Financial economists have literally looked at dozens of firm characteristics, perhaps even hundreds. Some of the main ones for any given company include: their growth rate, cash flow per share, level of R&D investment, amount of intangible assets, book value of the firm compared with market value, earnings per share, market capitalization, correlation with the broader market, macroeconomic conditions in the country where the company is located, industry, past price movement, analysts ratings, and how long it has been since the firm hit a 52 week high.

So why do you, the average investor care? Well bear with me a minute, and I’ll tell you.

The vast majority of financial economics research involving the stock market goes something like the following. Step 1: get data on a bunch of characteristics for thousands of companies over a long period of time – usually at least 10 years, sometimes 30-50. Step 2: Put the data for each company each quarter (or year, month, day, whatever) into a mathematical equation called a regression that looks something like the following:

Stock Return This Year = B0 + B1*P/E Ratio + B2*Growth Rate + B3*Market Capitalization (In Billions of Dollars) + B4*Return Last Year +… etc.

Step 3: Use all of the regressions for all of the companies and a computer program to figure out what B0, B1, B2, etc are. Basically, what B1 then tells you is how much an increase in the PE ratio will change stock returns on average. So if B1 is -0.25%, then a stock that has a PE of 14 will have a stock return that is a quarter of a percent less than a stock with a PE of 15. Step 4: Write down all the numbers your equations gave you, and how accurate your model is in a long and confusing paper, and then get the paper published in an academic journal that pretty much no one except other academics reads.

So what was missing from this scenario? … How about using the data to actually predict which stocks will go up in the future? Most academics never do this because academic journals simply aren’t interested in publishing papers that are about a specific stock. The idea is that no one cares if $GE, or $BAC, or $AAPL, or $CSCO is going to go up or down this year. Now of course that’s not true, but academic journals want their articles to be meaningful for a large number of other academics, not just the few who care about these specific stocks.

So how can this help you? Well, financial economists publish the estimates from their models for B0, B1, B2, etc. If you have these numbers, and you have the data for one specific company – say that companies, P/E ratio, market cap, growth rate, and return last year, you can plug these numbers in, and get a pretty accurate predictor of what your stock’s price will be in the future. How accurate? – About 85% or so if you do it right. Your equation with numbers plugged in should look something like this:

Stock Return this Year = 8.00 + -0.25*14 + 1.25*3% + -0.50*4 + 0.6*10%

Stock Return this Year = 8 + -3.5 + 3.75 + -2 + 6

Stock Return this Year = 12.25 => This stock should return 12.25% this year.

Now I am making up the numbers for B0, B1, B2, B3, and B4 here, and of course I am making up the P/E ratio, etc for a hypothetical company, but you can definitely get numbers for B0, B1, B2, B3, and B4, that have been vetted by dozens of financial economists. How you ask? Well, either by subscribing to the very expensive and dull academic finance journals, or by following my profile and reading the future articles that I’m going to post which will give you these numbers.

As I said when I started, this method definitely isn’t perfect, but lots of economists have spent their careers trying to find a better way to predict stock prices (which would result in them winning the Nobel Prize), and no one has been able to come up with a better method yet. This should tell you that this tool is pretty darn good.