The opinions and perspectives presented in this article belong solely to the author(s). Kalshi does not provide investment or trading advice or make any other claim to the veracity of the contents described herein and provides this article solely for the convenience of its members.
It’s 8:15 on November 10th, 2020, when a letter arrives on your doorstep. It reads:
In 15 minutes, the Bureau of Labor Statistics is going to announce that inflation last month was 0.48%, above 7% year over year. Use this information wisely.
My god, you think to yourself, I must take advantage of this information immediately. 7% inflation in just one year sounds high to you, so, expecting imminent economic collapse (and too rushed to consider the legality of your inside-information), you aggressively short the S&P 500. 15 minutes later, your source is confirmed — the CPI estimate was precisely on point.
You breathe a sigh of relief, but your elation is short-lived. For some reason, the S&P has climbed rapidly since the announcement. Why? What could explain this behavior?
Unfortunately, you’ve fallen victim to one of the classic blunders — while inflation was indeed quite high, the market was expecting an even worse outcome. That is, when you bought the S&P 500, inflation of more than 0.5% was already priced in. The difference between the true value of a data release and the market’s expectation is known as “surprise,” and since the market was positively surprised in this case (by lower than expected inflation), it rallied. In fact, if we plot month-over-month CPI against S&P movements after the data release, we see that there’s no clear relationship:
Taken by itself, the value of a CPI release is a very poor predictor of stock market movements.
This brings us to a tricky point about trading — to profit, it’s often not enough to be a perfect predictor; you also have to have a good sense of what others in the market are thinking. The problem isn’t just limited to inflation either, you could have made the same mistake with the Fed rate, unemployment, or GDP. How then, could you have secured a profit using your alpha?
One option would be to look at expert reports and use them as a proxy for the market’s broader consensus. Unfortunately, these are pretty noisy — experts often disagree with one another, and it may be unclear who is most trusted. Other traders, like yourself, also have estimates of their own which aren’t shared publicly.
A better alternative would be to look at a tradable derivative that directly prices CPI. By tracking the price of the derivative, you can see exactly what others think the CPI release will be, and with enough participants, a public market would provide a great estimate of broader consensus. In fact, economic indicators are traded through instruments called “event contracts.” There are a few exchanges which offer event contracts on CPI. Today, we’ll look at Kalshi’s markets, but the idea is generally applicable to any CPI market.
If, rather than plotting S&P movements against CPI releases, we instead plot S&P movements against surprise (the difference between the released data and the forecasted value), we see a much stronger correlation:
When we adjust CPI data releases for the market’s expectation, and calculate surprise, CPI becomes a strong predictor of thestock market.
As you can see, when inflation exceeds Kalshi’s estimate, the S&P falls, and when inflation falls short of the estimate, the S&P rises (in fact, the “reaction function” here is a bit more complicated, but this linear model is good as a crude approximation).
To go back to the earlier example: if after receiving your letter you’d looked up the public forecast, you would have seen an estimate of 0.65% inflation. As that’s significantly larger than your prediction, you’d have expected the market to be positively surprised, and you’d have known to take a long position on the S&P.
If you’re interested in using Kalshi’s predictions to inform your own trades, you can access them publicly on Kalshi.com. An open-source CPI calculator which walks through this exact example using data fetched from Kalshi and Yahoo Finance is also available here.
If you’re interested in learning more about surprise and the impact of economic data releases on the USD, you can check out this article by Daniel Dubrovsky.