Forecast Folly

One of my favorite events each year is the “CFA Annual Forecast Dinner” that takes place every January. At this dinner, there is an array of professionals in the money management business who get the chance to “talk shop” and reflect on what happened in the previous year and how they navigated that environment. In addition, they usually have a few high-profile guests that are invited to give their thoughts on the past environment, but much more importantly, on what is to come.

All this is well and good, but one of the fun parts of each year is the awards ceremony. Each year, several analysts are asked to give a year-end target for the S&P 500 (also the year end yield on treasury and price of gold, but the S&P 500 target is most important.) At the following year’s dinner, the analyst with the closest estimate wins a prize. It’s nothing major, just a small trophy and bragging rights in front of your peers.

I say this part is fun, because it’s really the most lighthearted part of the whole night. All these guys are very smart, experienced, and highly regarded analysts. They’ve spent their lives analyzing businesses and the economy more broadly. But even they know that predicting what happens in the next year is a joke. The funny part is everyone is in on it. Everyone in the crowd laughs as the host shows the range in estimates and how far off they are from what actually happened. Whenever the person wins the reward, they have a bit of a smirk, almost in embarrassment. Their table erupts in laughter and pats the winner on the back as if to say “you lucky son of a gun.” Often, the winner is someone who either had the largest or smallest guess, as the actual result fell outside of all the analysts’ range.

Sadly, there’s a pretty good chance that dinner will be cancelled this year with everything going on. However, we still have the top market strategists’ estimates from last year, so we can find out who would have won this year’s award. Below is a chart of all the top strategist’s estimates in blue, followed by what actually occurred in orange.

As is so often the case, the actual gain fell outside of the range of all estimates, so that the highest guess wins (Goldman Sachs in this case.)

In addition to the year-end price, they also estimate what earnings per share (EPS) will be. Below are the estimates followed by the most recent EPS estimate.

As you can see, the estimates for EPS don’t vary too widely. They average around $131. Again, if there was a prize for forecasting EPS, the lowest forecast would win as the actual EPS was outside of the range.

Here’s where it gets interesting. Did you notice that the market’s price ended above all the estimates despite EPS ending way below every estimate? What do you think the analyst’s estimates would have been had they known earnings were going to fall so much for the year?

To some degree, they tell you. Implied in their price and EPS estimates is how they think the market will be valued, using an implied P/E (price/earnings) ratio.

Implied PE ratio = Price Target / EPS Target

For example, Goldman Sachs estimated that the year-end price target would be $3700, and that EPS would be $136. This implies a P/E ratio of about 27 (3700/136). Had Goldman Sachs known that the years actual EPS was going to be $120, and they applied that same P/E ratio of 27, they would have wound up with a price target of $3,240. That implies a loss for the year (vs. the 15% price gain that actually happened.)

This is why these forecasts are just entertainment. Even if you would’ve guessed the EPS correctly, you still have to guess what the market is willing to pay for them. At the end of 2019, the S&P 500 price was $3,250 and EPS was $139.5, implying a P/E of about 23. Another way to say this is the market was willing to pay $23 for each dollar of earnings from the S&P 500.

As of this writing, the S&P 500 price is about $3,740 with an estimated EPS of about $120. This implies a P/E ratio of about 31. That means that for whatever reason, the market is currently willing to pay about 31% more (31 vs. 23) for a dollar of earnings today than they were last year.

The reason why the market was willing to value the S&P at 31 in 2020 vs 23 in 2019 is another story altogether. You could argue all the stimulus gave people a more aggressive appetite for stocks, or that all the new retail speculators on apps like Robinhood aren’t as sensitive to valuation. Both probably have some truth to them. Either way, the reasons for that happening were pretty hard to foresee in 2019, and none of the forecasters did. I’d suspect the same for next year’s forecasts.