In all likelihood no one reading Rajan’s essay was a impacted by his casual dismissives regarding residential construction as I was. This has been a point of mine for over a year now but let me make a few comments because they are still relevant to how people think about the world.

Rajan wrote:

The status quo ante is not a good place to return to because bloated finance, residential construction, and government sectors need to shrink, and workers need to move to more productive work.

There are three ways you could take this – all of them I think misguided.

  1. The residential construction sector is currently bloated and needs to shrink
  2. At the onset of the financial crisis the residential construction sector was bloated and needed to shrink.
  3. At its peak the residential construction sector was bloated and needed to shrink.

Here is a chart of housing starts since 1960. Some folks like to use Residential Fixed Investment as a percentage of GDP. However, there are at least two issues with this. One RFI is not the same as construction. Real Estate brokers are a part of RFI for example and they experienced a significant boom during the 2000s related to the home buying increase. A distinction I will make more of later. Two, RFI as a percentage of GDP will rise when the some other part of GDP falls but this does not necessarily imply that all of GDP should fall.

So, lets look at starts which are a proxy for construction itself


I think it’s a fairly straight forward case that the current level of housing starts is well below what is likely to be the long run average for the United States and indeed is currently falling behind household formation, which itself is falling behind population growth. This is reflected in rapidly declining rental vacancies and rapidly increasing rents.

What may be more surprising to folks is how depressed residential construction was going into the financial crisis. As Lehman Brothers declared bankruptcy, housing starts were already nearing record lows and were well behind population growth.

Indeed, the majority of the fall in starts happened even before the US entered recession. By the time the recession began housing construction was close to what some analysts at the time thought would be the cycle low. I disagreed, of course, but I do remember this was a major point of contention. There were not unintelligent folks  suggesting in 2008 that home construction had bottomed.

Had the US remained at the construction level it had at the beginning of the recession it would not have been able to keep up with population growth especially given what had been a healthy influx of Mexican immigrants. Recently, the influx has died off and housing demand is actually weaker than we would have expected.

However, if the recession and immigration crackdowns had not come we would have been well below trend even at the dawn of the recession.

Now, what about the peak. Surely it wasn’t sustainable?

A few things though

First, it wasn’t even the all time peak. Starts in the 1970s were more robust as the baby boom generation was leaving the nest. And in the 2000s with the Eco-Boomers and Mexican immigration we were facing an adult population surge greater than that seen with the baby boom.

Here is the year-over-year absolute change in men and women over the age of 20. You can see the baby boom generation come of age in the 70s and you can also see the recent spikes, in part due to updates in the count of immigrants.

FRED Graph

Now, lets layer this over starts

FRED Graph

Given the relative changes in the adult population the surge in starts is not way out of line. Indeed, its not clear that it is out of line at all.

It may be that given the rapid drop off in immigration that we saw in the last half of the the 2000s that the huge ramp up in housing starts was ill-conceived. However, in 2004 when starts where surging towards their peak it was not clear that this was going to happen.

Thus, arguing that what was wrong with America is that we went on a crazy homebuilding binge is not clearly supported by the data. It is clear that we went on a crazy home buying binge and the number of rentals in the US collapsed. However, that is not the same as homebuilding.

Rental construction is residential construction and I am particularly sensitive to folks acting like the only residential construction in the world is detached single family homes. There is an important market in rentals and its not at all clear that once this market is taken into account that America was ever building an unreasonable number of homes.

However, even given all of that the US reduced home construction by over 50% even while the unemployment rate fell from 2005 to 2007. There was no problem at all shifting resources out of residential construction and into other things.

And THIS is precisely my point. People routinely ignore this key fact. The US had absolutely no problem whatsoever rapidly shifting the structure of production out of residential construction.

There was no problem because as this happened the dollar was falling and US Net Exports were surging. Indeed, part of the reason Net Exports had collapsed was because residential construction was surging. There was a smooth rebalance between the sectors and it occurred because PRICES were changing to make it happen.

This is why I simply can’t get over conversations about the recession and the mass increase in joblessness that ignore the key role of prices in shifting patterns of supply and demand.