I wanted to take a minute to talk about the GDP report. It was much worse than I expected. No big deal for us, as we expected it to be a one-off positive, similar to how oil spiked up 6 weeks ago and has now been tailing off. Our answer to the expected report was to make no changes, and just accept that things may go against us for a bit, and maybe try to take advantage where we could on the other side.
First, what happened? To be totally honest, I put a lot of faith in our principal macro data research provider Hedgeye’s treatment of GDP analysis. We both discovered pretty similar things in the GDP data, but they do far more detailed, thorough work than we can do on our own. Nonetheless, even for them there are GDP swing factors that are pretty hard to predict. A lot of these factors saw some dramatic changes. The biggest change was the deflator (the inflation number used to get from nominal to real GDP), which got much higher this quarter. If the GDP deflator from Q2 was used in Q1, it would have been a negative GDP print. Maybe that’s why they didn’t use it? The GDP deflator is subtracted from nominal GDP to come up with real GDP, which is what gets reported. This greatly reduced the print.
The next big factor was inventories. According to the GDP report, inventories were strongly liquidated in the quarter, which depresses production. If you just liquidate inventories, you don’t need to produce so much, so GDP goes down.
Third, and smallest, is that government spending went down. That’s pretty unusual in an election year. Nonetheless, it happened.
This explains the vast majority of the miss. Fortunately, we didn’t make any changes due to our expectation of the report, as our attitude was that the Q2 GDP would probably mark a top in economic numbers. As it happens, that top didn’t happen. No big deal, though I think it’s worth pointing out that we were arguably right for the wrong reasons. We got the Q2 GDP report wrong (we expected a one-time GDP improvement), but it didn’t hurt our results due to our broader process. The broader process doesn’t overreact to one-off data points, a popular hobby for many, as we prefer to focus on longer-term trends. I think that helps us stay in line with what we’re good at.