In our debate earlier this week Rob Atkinson talked about how US companies weren’t investing like they used to. He’s not the only one concerned about a dropoff. Tyler Cowen had a post in July of this year titled “Why is US corporate investment not more robust?” It points to an article in the FT that contains much head-scratching over the fact that “Profits in the US are at an all-time high but, perversely, investment is stagnant.”
I agree about the profits (Atkinson doesn’t, but his errors of fact are a subject for another post), but I’m not convinced there’s any deep puzzle here. I just don’t see the alleged investment shortfall.
US companies make capital investments in three categories of stuff: plant (i.e. buildings); equipment like machine tools, trucks, PCs, and servers; and software. My digital research assistant FRED tracks these in different combinations and for different time periods. So let’s look at what they tell us about corporate investment over the years.
In all cases I’ll express investment as a percentage of GDP.
Here’s total (nonresidential) corporate investment — plant + equipment + software — over just about the entire post-war period:
It certainly seems to show that things are in the doldrums, at least compared to the investment levels of previous decades. When we isolate the three different components of investment, though, a very different picture emerges.
Here’s total investment in ‘information processing equipment and software:.’ Corporate spending on digital stuff peaked in the late 1990s thanks to the twin fiascos of the dot-com bubble and the Y2K ‘crisis’ and has backed off some since then. But it’s still only a couple tenths of a percentage point lower than it was between the last two recessions of the 2000s:
This slight dropoff could spring from the fact that companies have honestly lost some of their enthusiasm for information technology. But it could also be due to the fact that the cloud and competition have made enterprise software a lot cheaper than it used to be. Cloud computing has also greatly reduced the amount many companies are spending on servers, and powerful PCs and other end-user devices are now dirt cheap.
These developments have been bad news for IT producers like Dell and HP, but they’re great news for technology consuming companies because they lower the total investment required for a given level of tech. For some companies they’ve probably lowered the total tech bill.
Another fundamental development is that non-computer equipment these days very often has a strong digital component. Machine tools and trucks now are full of processors, controlled largely by code, and networked. So how’s the total corporate appetite for equipment + software? FRED gives us data going back to 1995:
(this one’s a tiny bit sloppy because investment is given in 2005 dollars, while GDP is in 2009 dollars. This isn’t ideal, but it doesn’t make any substantive difference to the conclusions)
So this graph shows us that the American companies’ appetite for physical and digital gear is at an 18-year high. I wish I could say that it’s at an all-time high, but the data don’t go back far enough for me to make that claim. I’ve got my human RA looking for this data going back further in time so we see if that early 80s spike in the top graph above is still present. I’m betting it won’t be; stay tuned.
The overbuilding prior to the Great Recession certainly cooled off enthusiasm for all forms of real estate investment today, but when I look at this last graph I don’t see any diminished interest in gear. Companies are investing in it at historically high levels, not historically low ones.
So, what investment shortfall are we so puzzled or worried about? I just can’t see it.
One final note: Cowen correctly points out that there’s a big difference when looking at gross vs. net (of depreciation) investment. The data above show gross levels. There are the right ones to look at because they show corporate willingness to get out the checkbook and make capital investments instead of hoarding cash or returning it to shareholders. What’s more, companies don’t trash all their hardware and (especially) software once they’re fully depreciated, so the fact that these assets are off the books doesn’t mean that they’re off the premises.