Auto Inventory Cycle Adjustment: More Pain Ahead

Mike Smitka
Washington and Lee University

During the extended housing boom of the early 2000s, consumers used home equity lines to go on a consumption binge. Light vehicle sales remained strong, while the model mix richened. Then the housing bubble reached its peak (and gas prices rose). The good times were over, and at the February 2009 trough sales fell to that of the trough of the 1981 recession. In relative terms the situation was far worse, because in the interim the US population grew by almost 80 million. Relative to employment, sales were a full 15% below the previous post-1976 (before which consistent data are unavailable).

Cars are durable goods, and even though the average vehicle now lasts 12 years, they eventually need to be replaced. Meanwhile the US population continues to increase and incomes have recovered (though for most Americans, not risen). From an excess of vehicles per household going into the recession, the number fell below what households wanted, and has now recovered. In February 2007 SAAR was 16.7 million units; that level wasn't hit again until March 2014. Fueled by low interest rates, longer loan maturities, and high used vehicle prices, sales crept back up. Indeed, Paul Traub of the Federal Reserve Bank of Chicago argues that they have been higher than sustainable.

...it's payback time...

It's payback time. Over the past several months the Light Vehicle SAAR (seasonally adjusted annual rate of sales) fell 7.5% from its December 2016 peak of 18.051 million units. This represents a drop of over 1 million units. Now  inventories rise and fall as a normal response to short-term swings in sales. With a steady fall, though, they mushroom, and mushroom they have. (Thanks to Paul and his May 2017 presentation to my W&L auto seminar for the graph on the left.) Now that the drop appears to be more than a transitory blip, it's time to bring inventories under control and pare production.

The supply chain is like a snake....

The supply chain is like a snake. If it's to eat bigger prey, it has to bulk up, and that takes time as suppliers rehire and otherwise add capacity. (This is on top of the normal investment to replace tooling for old models with that for new. Given the hit balance sheets took from the Great Recession, they also had to repair balance sheets to add capacity.) Going into the recession, production peaked in June 2007. It took over 6 years, until November 2013, to reach the previous peak.

....that's consumed too big a meal.

Now it’s like a snake that’s just consumed too big a meal: it's sluggish, because it takes time to get it out of its system. The Production Index hit 131 in October 2016, and bounced around that level through April 2017. It’s since fallen 7.6%. But that at best brings production in line with sales. It doesn’t pare inventories. That will require either an uptick in sales (fueled by assorted incentives) or a further cut in output. GM for example has chosen the latter, with extended summer vacations – not that they aren’t discounting, after all they can’t totally ignore price cuts by competitors. Output will surely fall more; it’s better to overshoot a bit, given uncertainly about whether sales will fall further. And today labor is no longer a fixed cost. As per Econ 101, the flip side of higher marginal costs makes it relatively more profitable to cut production than prices.

That leaves two questions. One: what is the stable level of output? I’ll save that for another post; too many graphs already! The second is employment. Here the bottom line is clear: productivity continues to rise, a trend visible as far back as data are available. Industry employment is certain to fall, and on a permanent basis. As a dismal scientist, I close with that graph, and with that of auto industry manufacturing employment.

I leave out many pieces. One is that while manufacturing productivity increased, that's much less true of automotive retail: employment there is at a historic peak. Then there are the finance, household formation, population data, household formation, depreciation and other pieces of the peak sales story that Paul Traub sets forth. I have not tried to put together the last couple months of inventory data, or compiled the various media reports of reduced production at GM factories, eg Lake Orion where the slow-selling compact Sonic (and the Chevy Bolt EV) are produced. Finally, while I can't provide any details, I do participate when my schedule permits in the monthly sales analysis roundtable that BWG hosts for their clients. My reading of these various sources is that at a more detailed level, some manufacturers are slower in responding than others, while the impact of the sales slowdown is uneven across manufacturers/brands, with for example the shift towards light trucks amplifying the impact on producers that are sedan-heavy, such as some of the Asian brands.