Building Things in America: Notes on Manufacturing

by: Modeled Behavior

By Karl Smith

At Ezra’s place I talked a little about what a big story deal manufacturing has been over the last decade. The persistent weakness and the way that it creates a negative undercurrent in the US job market.

I think this is also key to growing inequality as the forces that contribute to the hollowing out of US manufacturing jobs may also be the forces that are contributing to record profits on Wall Street.

In short, strong demand for US financial assets is matched against weak demand for US tradable goods.

One result is US financial institutions who can borrow artificially cheaply and use that leverage advantage to make higher than expected profits. In finance we sometimes talk about the Fed put. That’s the idea that the US financial system is ultimately backstopped by the Federal Reserves refusal to allow a collapse.

I am also interested, however, in the China Put. The idea that leverage is backstopped by the fact that there is always an enormous buyer for some form of dollar denominated debt.

This may have been even more important in the explosion in credit and leverage that we have seen. For example, when everything was falling apart in 2008 one of the more notable activities was that China was selling mortgage bonds like gangbusters in order to buy Treasury bills.

At the same time construction is in a truly deep whole that I just don’t have a reasonable explanation for. Look how far this has fallen.


Again, that’s real construction spending per capita.

There are a lot of ways we could look at this chart. We could say that there has been uneven growth since 1965, with a stall in the 1980s a return to trend in the 90s and then some sort of absolute disaster starting in 2005.

Another way, is to say that something very odd happened in the 1990s. Perhaps the “sustainable” level of construction was simply the 1980s level. Yet in the 90s we boomed away from that and are returning to trend.

Cutting construction into its pieces might help answer that. Lets look at non-residential construction.

Here is real per person non-residential construction spending

FRED Graph

There looks to be a solid growth trend as we would expect. Even in real per person terms the country spends more on non-residential investment as it gets wealthier.

So let's take the natural log. This means that steady growth will look like straight lines and volatility that results mechanically from higher spending will be damped down.

FRED Graph

That looks to me more like a slowdown that occurred in the 1999 and from which there has not yet been a recovery. On the other hand a similar slowdown occurred throughout the 1980s.

Still not sure what to make of this.

Let's look at real residential construction spending per person.

FRED Graph

One interesting thing here is that we do see a very different pattern, not only from commercial spending but from units. The 90s really were different. It makes one wonder if the great moderation wasn’t really the great residential construction. Or perhaps, if moderating interest rates and employment spurred on households to expand construction spending.

Briefly let's look at units.

FRED Graph

Above is new housing units completed per capita in the US. In some sense this does look like a Great Moderation in home building. The wild swings turn into a slow steady climb.

Yet, there is also the fact that even at the peak of this climb were were producing weren’t producing many more homes per person than in the 1970s bust. This tells us the increase in real construction spending meant we weren’t getting more housing units, we were getting more expensive housing units. Pimp my house, apparently.

Through the magic of FRED, its easy for me to calculate average spending per unit over time.

FRED Graph

So we can say that through the 1980s we saw a sustained decline in the number of units but an increase in the spending per unit. The in the 90s the decline in units ended and even mildly reversed, but the spending per unit did not.

What was odd about the 1990 to 2005 run was not the number of homes that were built but the number of expensive homes that were built.

I want to check quickly against per average per capita income.

FRED Graph

This graph I just plotted is construction spending per new home as a fraction average disposable personal income. What I am looking for is whether or not increasingly expensive home matches some pattern in disposable income.

Indeed, it does. Let's natural log it to be sure though.

FRED Graph

Less pronounced effect but it is still there.

What we are seeing is that up until 1982, average new home construction costs were shrinking as a percentage of average real disposable income. This means the average cost of building a new home was shrinking as a fraction of average disposable income. That trend flattened out in the 1980s. Then resumed in a more minor form in the 1990s.

Put together, however, it does look like an increasing fraction of America’s increasing wealth went into more expensive homes. That is unlike food for example, new homes did not keep shrinking as a portion of the family budget.

There is still a lot we don’t know. This could simply be because the housing stock of the 1960s was of such good quality that all the new homes were bought by rich people and poorer people simply moved into the still good quality homes the rich were vacating.

I am going to stop there for now, leaving you with no conclusion but hopefully a sense that there is something interesting going on in the story of construction and it started well before 2002 and subprime.