Capital stock growth rates are middling

For the past few days, I have had my head buried in the BEA’s capital stock data.  And I am here to tell you that they confirm the message from the financial-balances and cyclical-spending-share perspectives.

That is, the cyclical position of the economy is meh.  There is not a lot of pent-up demand, but nor are there the sorts of glaring imbalances that have typically preceded recession, particularly in the last three episodes. If you want to save time, just go look at the charts below. I have already given you the point, as I see it.

Looking at the capital stock data is not generally considered a very good use of time, I think for two reasons. First, the figures are usually roughly what one would expect them to be, and therefore not very informative.

Second, and perhaps more to the point, the data are produced with a long reporting lag, which often means that the economy has moved on by the time we get them. For example, the full-year 2015 numbers were made available only this month. The economic activity that generated those results was centered on July 1 of last year.

There is also the thorny question of whether that the capital stock is even a real thing, worth considering.  Maybe I will return to that later, because IMHO it is interesting. But for the purpose of this post, I am going to stay at the level of my pay grade and suspend disbelief.

My effort these past couple days was focused on trying to come up with proxies for the capital stock growth rates to 2016.  This involved, first, estimating capital spending (flow) growth rates for the remaining two quarters of the year, for which I penciled in 2% (ar) for all items, simply to be “neutral.”  Second, I came up with a guess for depreciation by assuming the depreciation rate relative to the prior year capital stock was constant.   And I finally, I just strung together the resulting real net investment figures into an implied capital stock, in this case available to 2016. This seemed to work pretty well for the in-sample period I had.  So I figure my estimates for 2016 are probably in the ballpark.

So what do we got, suspending disbelief and assuming the capital stock is even a real thing?  The main thing we got is that the fun, from a forecasting perspective, ended a while ago.

The fixed capital stock data follow the same stock-adjustment logic as the inventory data. When the flow of investment is depressed (or overextended) relative to what is required to keep the ratio of the accumulated stock to real output on its trend line, then we can say that the flow is too low and that there is pent-up demand (or too high and that there are imbalances.)

I remember this being a satisfying set-up during the early stages of the economic recovery. People were dissing the investment spending upswing on the grounds that it amounted “only” to replacement capex. True, but that is what made the upswing so durable and so easy to forecast continuing.  Zero growth of the accumulated capital stock is not an equilibrium condition.  Just as a steeper-than-trend decline of the inventory/sales ratio is not. (Incidentally, this implies there is now some minor pent up demand in inventories.)

But now it is kind of blah, on the fixed investment side.  Take a look at this chart.


The top left panel compares the annual rate of growth of the “cyclical total” capital stock with the CBO’s estimate of potential GDP growth. The “cyclical total” comprises the non-residential stock of private structures (2/3) plus the business equipment stock (1/3).

I exclude the stock of “intellectual property” because the concept seems particularly dubious – and more to the point – because spending on intellectual property development is not that cyclical. Back in the day, much of this new category was considered an expense within the National Accounts and not even form of capex, although I will touch on it briefly below.

The relationship between the capital stock growth rate and the economy’s potential growth rate involves causal relationships running both ways. Capital accumulation affects potential growth and potential growth should affect the rate of capital accumulation.

I think the chicken-egg problem can be solved if we take the rate of tech advance to be determined elsewhere (although not reliably). So if aggregate capital stock growth is running above the economy’s potential growth, then we might say – very roughly – that the capital spending cycle is overextended.

The top left panel is not Nobel Prize stuff, obviously. The only inference I would really draw is that the trend slowdown of capital stock growth is not an anomaly and that you can probably forget about mean reversion in this context.

For now, though it seems as though the moderate rate of growth of the accumulated capital stock implies that the flow of capital spending is neither a source of pent-up demand nor a major real imbalance pointing towards recession. I would like to say more, but this story right now is just not that interesting. You read it here first!

In the panel top left, I do not show an estimate of the relevant capital stock growth rate to 2016, but it seems likely to be roughly unchanged — at 1.8%.  The odds would seem to favor a marginal slowdown if anything.  The reason is that the structures stock growth rate seems likely to have quickened slightly, while the equipment stock growth rate probably slowed a bit more dramatically.

These data are largely self-explaining, assuming you believe they measure something real, so there is no need to belabor this too much. But I will conclude with three quick points, one on each of the components of the non-residential capital stock.

The accumulated stock of business equipment was growing a bit quickly, relative to the pace seemingly implied by technology advance, in 2015.  But it looks to have slowed in 2016. The growth rate still looks slightly elevated, which can be taken as evidence of optimism (good) or as evidence that things are marginally overextended there (bad).  In my view, it is not sufficiently out of line to be interesting, unlike in the early period of the recovery, where the growth rate was clearly too slow and where a forecast of a quicker capex seemed to be a no-brainer, assuming the financial system could stabilize.

Somewhat related, Paul Krugman has often made the point that those claiming the expansion has been held back by a lack of business confidence have been full of it.  If business confidence has been a drag, then why has the flow of business spending on equipment scored a roughly normal recovery, asks Krugman? I agree.  I would just say the better measure of this is the capital stock growth rate, not the pace of recovery in the flow of capital spending.

The rate of growth of the accumulated stock of total business structures has remained subdued, although mostly in line with what one might expect given the economy’s slower potential growth rate.

This may seem like an odd thing, given that there was clearly an overinvestment cycle in the energy patch, where gas and oil wells are considered “structures.”  At the peak of the energy boom, investment in mines and wells was about 30% of total non-residential structures investment.  The flow has since fallen by almost 70% to near its historical low relative to GDP.  And I see a note from the street this AM showing that the Baker Hughes rig count stopped collapsing recently.

But the flow of investment in commercial and health care structures has continued to rise quite strongly throughout.  As a result, the rate of growth for this overall category — involving apples and oranges as it does — looks set to have firmed just slightly in 2016.  The commercial/health subcategory probably deserves separate attention, but not here.

Last and least, the rate of growth of the accumulated stock of intellectual property remains low by historical standards, although it seems to have picked up in 2016.  This category is dominated by software development which has continued to do well, although it also includes pure ideas, like those expressed in movies and books.

I don’t pay much attention to this category because it is not particularly cyclical, at least recently, and because I can’t keep a straight face thinking about the rate at which ideas depreciate.  Some ideas are really bad and it would probably speed growth if they were to “depreciate” completely. A lot of quite important things in macro are not necessarily measurable. So deep.