Monday, October 15, 2018

Real retail sales lay an egg in September

 - by New Deal democrat

Real retail sales is among my favorite economic indicators. Measured per capita, it is a long leading indicator for the economy as a whole, while its simple measure is a short leading indicator for employment.

Let's take a look at each.

Last week I noted that retail sales had grown nominally by +1.9% last September, likely a rebound from last year's troika of major hurricanes, including the one that dumped up to 50" of rain on parts of metro Houston. While this year Florence was bad, it did not shut down major metro areas, so I anticipated the YoY number would suffer in comparison.

And that is what happened, as last month nominal retail sales grew only +0.1%. Ex-autos, they actually declined -0.1%. Adjusting for inflation there was no growth at all, as shown in the graph below:

Real retail sales have grown by +2.4% over the last 12 months.

This is a big deceleration from their previous YoY reading of +3.8% (blue in the graph below). If this marks the start of a substantial YoY downshift (and I suspect it does), then the stellar jobs reports we've seen for most of this year (red) are likely to end in a few months:

Real retail sales per capita have also decelerated to +1.7% YoY, and have declined in absolute terms for the last two months:

Since the monthly data is somewhat noisy, we can usefully tune out that noise to get a decent real time signal by measuring in quarterly increments:

The quarterly change has not been problematic unless it is negative for two consecutive periods. Since with the exception in Q1 this year it has been positive, there is no negative signal yet.
On  the other hand, if the weakness of the last two months persists further, and if housing continues to turn south, that would be a potent sign that the consumption side of the economy is under pressure, and be yet another harbinger of a slowdown that I have suggested is likely by about midyear 2019.