In the NYT Upshot section Neil Irwin had an interesting piece assessing which sectors are most responsible for the weakness of the economy. His culprits (in order) were residential invesment (housing), state and local government, durable goods consumption, business equipment investment, and federal spending. Irwin's methodology was to take the Congressional Budget Office's estimate of potential GDP (roughly 5 percent higher than the current level) and then assume that each component has the same share of this potential as its average of GDP over the two decades from 1993 to 2013. The difference between this hypothetical level of demand from a component and the actual level of demand from that component in the second quarter of 2014 is the basis for determining the shortfall.

I decided to do a similar exercise with a couple of minor differences. The table below shows the difference between each component's average share of GDP in the period from 1990-2013 (this was an accident -- misread Irwin's start point) and the average for the first two quarters of 2014. The two quarters are taken together because for many components a strong second quarter offset a weak first quarter. I have also lumped components together (e.g. the categories of consumption are all together). The categories in bold are the major components that together add to GDP.

  Percentage Point Change
  Average 1990-2013
  Minus 2014
Consumption expenditures -2.3
Durable goods 0.7
Nondurable goods -0.1
Services -2.9
Nonresidential investment 0.0
Structures 0.0
Equipment 0.4
Intellectual property products -0.4
Residential 1.1
Change in inventories -0.1
Net exports 0.3
Exports -2.7
Imports -3.0
Government 1.1
Federal 0.5
State and local 0.5

 Source: Bureau of Economic Analysis, National Income and Product Accounts, Table 1.1.5.


There are a few points that can be made from this table. First, the items that have fallen substantially as a share of GDP are government spending, which had roughly equal dropoffs at the federal and state and local levels, and residential construction. Net exports are also down as the import share had grown more than the export share. Non-residential investment is at its average level for the 1990-2013 period. The big gainer in shares is consumption, which had a 2.3 percentage points larger share of GDP in 2014 than its average in the prior period.

The second point is that it is worth keeping the categories together because components can be substitutes. To be specific, the large growth in consumption of services noted by Irwin more than offset the drop in durable goods consumption that he notes. The story here is that if households are going to spend more on services, primarily health care, then they will have less money to spend on other consumption goods. Taken together we can see that the consumption share of GDP is higher in 2014 than it was on average in the earlier period.

We see the same story with equipment investment investment. While Irwin lists this component as one of the culprits, the drop in equipment investment was fully offset by a rise in investment in intellectual property products. Taken together, non-residential investment is just keeping even with its average over the prior twenty three years.

This brings up the third point. We can think of some components as being largely a function of GDP, whereas others are to a substantial extent independent of short-term changes in GDP. Consumption falls into the former category. If GDP were to increase by $100 billion, it is a safe bet that consumption would increase by around $70 billion. Imports also largely follow GDP. On the other hand, investment to some extent follows GDP, but also a fair degree of independent determinants. This is also true for housing, and even more so for exports and federal spending.

This is important, because we envision getting back to potential GDP it will depend on increases in one or more of the components that are not primarily determined by GDP. Our candidate list is then non-residential investment, housing, exports, and federal government spending.

Taking these in turn, it would be great to increase non-residential investment, but we don't have any great tricks to accomplish the task. Lower interest rates from the Fed can help, but the impact is limited and it is not easy to make interest rates go lower than they have been.

We can and will see some upturn in residential construction, especially as the vacancy rate falls back closer to normal levels. However it may not rise back to its two decade average. The fact that households are spending so much more on health care now than they did over the longer period means that they will have less money to spend on housing. As a result, we may never see residential construction rise to the same share of GDP as its average in the period from 1990-2013.

Skipping to government, we can see more spending, but that will mean larger budget deficits. That's fine by me, but doesn't sit well with the people calling the shots in Washington. (We could raise taxes, but that is hardly an easy political sell and it would also reduce consumption.)

This leaves net exports, my pet peeve. We can boost exports by lowering the value of the dollar against other currencies. A lower valued dollar will also lead people to substitute consumption of domestically produced goods for imports. That could do the trick for getting us back to full employment, but we would first need to get a lower valued dollar on the nation's political agenda. Don't hold your breath.