Hack the Fed - July 2019
I haven't blogged in a while, been enjoying the summer! But since the Fed meets this week, and this meeting could be momentous, I thought I'd put out a quick post with my latest model results.
If you remember my first post on the Fed, you'll recall we examined how the markets react after a Fed cut. And, while on the surface a cut seems positive, in the longer-term it generally leads to lower equity market returns.
The second thing I posted about was how to predict the Fed. We developed a model using some well-established factors with reasonable predictive power in-sample. The weakness of the model was/is that the data set is small and that makes it difficult to establish a good in and out of sample test. So, as before, the model should be taken with a grain of salt.
The below plot shows which factors were important in the model and hence how the Fed seems to weight them in their historical thinking. Unemployment, Equity returns, and Non-Farm Payrolls seem to be at the top of their list.
If we put current data into the model ...
we get a prediction of a +.1936 percentage point change.
The markets, judged by CME Fed Watch tool are 100% predicting a rate cut of at least 25 basis points, not a raise. So, either our model is 180 degrees out, or there is some exogenous factor(s) influencing the current environment that is not captured by this model.
Markets are irrational and markets are just a collection of decisions made by people (or machines programmed by people). And, while we can hack the Fed's decision making just like Google can predict the next ad we'd like to see, sometimes Google is wrong. Because people are irrational.
One could argue that IF the Fed followed the Taylor Rule (and the latest version published by Bullard in 2018) a 50 bps cut is warranted. However, the Fed rarely seems to follow any version of the Taylor rule.
So, we'll see. It seems to me a cut is perhaps a touch irrational at this point given the current state of the market and the Fed won't do anything, like the last meeting.