In today's post, I'll update the composite indicator heat map for May. See this post for an introduction to the composite indicators and the heat map. I'll also introduce a new composite indicator based on the COMP system I introduced here.
Below is the composite indicator heat map as of Friday, April 28, 2017. A few changes for this month. My partner in crime, Tony, takes care of all these updates and is responsible for the new COMP model as well.
- We deleted the Capital Spectator's CRPI and MMRI, since he doesn't publish them on a regular basis; they're only available with reasonable lag time if you subscribe to his newsletter.
- We created a Probit model for COMP, the composite of the six individual indicators that we keep track of here, and added that to the Composite heat map table. (See below for more info on this model.)
Generally, all green in April and all of 2017. No imminent recession is signaled. This confirms what the more timely individual indicators and the COMP indicator are telling us. When the composite score exceeds 62 the scoreboard will flash yellow, and when it exceeds 106 it will go red. The composite score hasn't exceeded 40 since the beginning of 2016.
In the second line of the table, you will see the new composite indicator based on the COMP system, ARA-PWN Composite. The COMP system is just a simple combination of the top 6 economic indicators that triggers when any one of the 6 triggers. Therefore, it is quite sensitive and triggers often. Turns out that this sensitivity is OK when you require a confirming indicator - in our case, SPY - to take action. The SPY-COMP system is ultimately concerned with the market and not the economy. But we wanted to see if we could build a better model just out of COMP that better tracks the economy. We built a Probit model with the COMP indicators. A Probit model is a multiple regression model where the dependent variable is binary - in our case, recession or no recession. Here is the result.
Pretty darn good. When the Probit model for COMP has gone above 20%, we've headed into a recession. We will keep this model updated with a trigger at > 20%. As of now, the model is nowhere near triggering a recession call.
That's it for this month's composite indicators. No signs of imminent recession are on the horizon (9-12 months). As a reminder, that doesn't mean the markets won't go down. Monthly drawdowns of 10-15% outside of recessions are quite common (with higher daily drawdowns) and should be expected. And no indicator is 100% accurate.