Earlier today, Tyler Cowen had a post titled "Why are there so few computer science majors?", which was prompted by this Dan Wang post on the subject. Among other things, Tyler wonders if there are relatively few computer science majors simply because the tech sector is actually pretty small. Since I was already working with economic data today for another project, I thought it was worth taking a quick look to find out just how big the tech sector really is.
Using data on industry real gross value add (the industry equivalent of GDP) and employment from Moody's, I took a look at the size of the information technology sector—which includes computer hardware, software, services, and Internet, ignoring other high-technology industries such as biotech, medical devices, and aerospace (details on industry definitions are here)—between 1980 and 2015 for the United States. I also took at look the variation across metropolitan areas.
The information technology sector accounted for about 0.8 percent of output and 2.8 percent of employment in 1980, but growth in the two series changed dramatically after 1995 when industry output began a sharp upward rise while employment held relatively flat. By 2015, the tech sector had grown to 5.2 percent of total GDP and 2.7 percent of total employment.
With explosive growth in output and flat employment growth, then, productivity in the information technology sector surged during this period. Real gross value added per employee, a somewhat crude measure of labor productivity, grew 834 percent in the information technology sector over this 35-year period, for a compound annual growth rate of 6.6 percent. Compare that with the rest of the economy, where these same figures were 37 percent and 0.9 percent. In other words, the tech sector had a compounded annual productivity growth rate more than seven times the rest of the economy! The tech sector lifted-up total labor productivity growth by 6 percentage points over these period, or about 0.1 percentage point each year on a compounded basis.
So while the information technology sector may be relatively small in size, what matters more is that it is an engine for broad-based economic growth. Enrico Moretti of Berkeley established an approach for estimating "local jobs multipliers." In his paper, and later in his book The New Geography of Jobs (highly recommended by the way), Moretti demonstrates that for each job created in the tech sector, 4.9 additional jobs are created in the local services economy over the long run—this includes low-skilled work like taxi drivers, barbers, and cooks, as well as high-skilled roles like doctors, lawyers, and teachers. In my own work, I found similar results in the United States during an extended time period and with a broader definition of the tech sector than Moretti used, and also when applying his methodology across regions of 27 European countries (in both cases, the multiplier was 4.3 jobs). In all cases, the local multiplier for manufacturing was about one-third what it was for high-tech.
If many of these stimulative effects are captured at the local level, in which cities does the tech sector have high concentrations? In each of the 52 metropolitan or micropolitan areas listed below, the tech sector share of GVA was greater than the U.S. average of 5.2 percent in 2015.
In many of these places, the tech sector's contribution to economic output surpasses its contribution to local employment by a long shot. In Silicon Valley-San Jose, the employment share is slightly more than half of the output share, while in Corvalis, Oregon (home to Oregon State University) it's about one-eighth, and Seattle it's about one-third.
And that's the point. Sizing the information technology sector by employment alone is limiting. A better measure would be its contribution to output and productivity growth, and the resultant multiplier effects it has on other sectors—not to mention the productivity impacts that the goods and services produced by the tech sector have on the performance of industries across the economy, which are sizable, and the employment of workers in tech occupations in non-tech industries (such as quantitative analysts in finance or systems administrators in hospitals—I estimate these roles to be equal about 50 percent of employment in the information technology sector).
So, is Tyler right to say that the tech sector is relatively small? Yes, at 5.2 percent of economic output and 2.7 percent of employment, that is a fair assessment. But, the sector has a much bigger impact on the economy than those numbers would suggest. Tyler wasn't making that point in his post—in fact, his post was about something else—but I wanted to take a few moments here to put some firmer numbers behind that, and to remind readers, that in terms of contribution to economic vitality, not all jobs are the same.