- by New Deal democrat
Twenty years ago Prof. Edward Leamer gave an important speech at the Fed’s Jackson Hole, WY, retreat called “Housing IS the Business Cycle.” In that speech he discussed the fact that, historically, private residential construction as a share of GDP on average peaked 7 quarters before the onset of recessions, followed by motor vehicle and other durable goods sales, followed by consumer durable sales, and then the coincident indicators of recessions.
Let’s take a look at what that progression looks like at the moment.
First, here is Leamer’s noted measure of housing as of Q2’s GDP, both in nominal (blue) and real (red) terms:
In real terms, housing as a share of GDP peaked in Q1 2021. Nominally it peaked one year later, in Q1 2022. After stabilizing for awhile, both measures had secondary peaks in Q1 2024. The present readings are the lowest since before the pandemic.
My preferred way of looking at housing is the historical progression of peaks during expansions: first, single family permits, then housing units under construction, and then employment in residential construction and new housing units for sale. Here’s what that looks like currently:
Single family permits peaked in 2021. Units under construction did not peak until October 2022. Residential construction employment probably peaked in March, although it is only down -0.2% since then. And new homes for sale have not turned down yet.
Next in Leamer’s line of progression are durable goods, starting with vehicles. Here is the most recent data for heavy truck sales (red) and light motor vehicles (blue):
In past cycles, heavy truck sales have declined earlier and far more unambiguously than car sales. The same appears to be the case at present, as heavy truck sales peaked two years ago and have declined more than -10%, a typical pre-recession decline. Car and light truck sales actually increased in late 2024, and even moreso with the front-running of tariffs earlier this year.
Looking at the broader durable goods picture, manufacturers’ new orders for durable goods increased in 2024 and have trended generally flat so far this year. Real purchases of durable goods appear to have trended similarly, although it is more ambiguous given the big jump last December and slump in January, which may just reflected unresolved seasonality:
Finally, manufacturers’ new orders for non-durable consumer goods has trended generally sideways for almost two years, while real spending on non-durable goods has continued to trend higher, albeit at a very attenuated pace since March of this year:
Putting it all together, housing is clearly down, even counting from its secondary peak 18 months ago, suggesting a likely time for recession by roughly the end of this year. Heavy truck sales are also down enough to signal a recession is near, although the signal from light vehicles is unclear to say the least. Broader durable goods orders and sales may have been in the process of peaking in the first half of this year; but consumer goods orders and sales are still trending higher.
Among the signs I am looking for to determine if my current “recession watch” should turn into a “warning” is more definite evidence that durable goods orders and purchases have turned down, and that consumer goods orders and purchases are at least trending more sideways.
