Tuesday, April 13, 2010

What economies want

It can hardly have escaped people's attention that much of the interest around education policy and its relationship to the economy is based on the idea that we need more graduates in Science, Technology, Engineering and Mathematics (aka STEM). It appears to be taken for granted that these folk are the engine of growth. Perhaps because I am a social scientist I am wary of this STEM-fetish. Rather than the thrash the argument about: what is the scientific evidence for this proposition?


Leigh Caldwell said...

Presumably the evidence you want is not scientific but economic (unless economics is a science, which might render the whole question moot!)

1. Is there a correlation between numbers of STEM graduates and economic growth?

2. Do firms that employ more STEM graduates invest more?

3. More simply, do STEM graduates command a higher wage in the market? If so, does the market already provide the correct number of STEM graduates - or are there market failures, or cognitive biases, which stop the market's price signal from properly influencing the supply [/quantity supplied] of new graduates?

All of these questions could plausibly be answered 'yes' but I'm afraid I don't know of research that answers them.

Kevin Denny said...

Leigh,yes economics is a science as it happens. But thats beside the point: we want good evidence that there is something special about STEM graduates.
1. Thats not evidence.Correlation doesn't prove anything. An economy that is growing might well produce more STEM graduates as a result. More Humanities ones too.
2.Again, thats not a great line of argument. Because it doesn't mean that they are investing more because of their STEM graduates.
3. I think this is a more interesting test. But you are saying there's nothing special about STEM, its just where the money is. Fair enough. So if we find that graduates of say, Law, earn more, then we need more law graduates?
What an economist would look for is some sort of direct evidence that there was market failure and that if we parachuted in STEM graduates output would rise by more than if we parachuted in some Arts/Humanities types.
Maybe that is the case: I don't know, but I am not clear that anyone else knows either. All I'm asking for is serious evidence for a policy that is important.
I find it ironic that policies about science can be so unscientific.
I am skeptical, even cynical. I think the STEM lobby have been quite successful at convincing policy makers that there is something special about their graduates & their discipines while other disciplines have not. And saying that Engineering is more important to the economy than Sociology seems pretty plausible. But is it actually true and by how much?

Martin Ryan said...

My first comment is that we are not alone: there is also an emphasis on STEM in the States. My second comment is that we are not the States: we are much smaller and a more open economy; (one of the most open in the world, it hs been said).

More pointedly, do we just need more graduates in STEM (and hope that they find jobs in the private-sector?)... or do we also need to employ these graduates as post-doctoral researchers (possibly for about six+ years, and then hope they get a lecturing position or move to the private-sector?).

To move on, if we can take "scientific and technological advances" as what STEM graduates are meant to achieve (whether they go onto post-doc or work in the private/public sector)... then I have a few more comments to make (which others may have heard from me before, though I think not on this blog).

I’ve looked into the Dale Jorgensen et al. (JEP) paper, which to my knowledge, is the source material for the Holdren quote to Congressional Committee, that “between 50 and 85 percent of the growth of the U.S. economy over the past half-century –and two-thirds of our productivity gains in recent decades –are directly attributable to scientific and technological

Jorgensen et al. paper: "A Retrospective Look at the U.S. Productivity Growth Resurgence"


It turns out that the Jorgenson et al. paper covers the period: 1959-2006. The central part is really Table 1 on page 13, which presents the growth of output. Following the paper’s terminology (from the top of page 8), the focus is on “total factor productivity” - defined as output per unit of both capital and labor inputs - and which “primarily reflects innovations in both products and processes” - or more simply, innovation. Total factor productivity (TFP), contribution of labour quality and contribution of capital deepening all feed into average labour productivity - which (along with hours worked) feeds into private output. So TFP as a percentage of private output (from 1959-2006) is 0.75/3.58, which is almost 21%.

Even if we consider TFP as part of average labour productivity, which (from 1959-2006) is 0.75/2.14, this is still only 35%. The other possibility I thought of is that Holdren’s figure is coming from a calculation which is average labour productivity divided by private output. This (from 1959-2006) is 2.14/3.58, which is 60% - somewhere between 50 and 85. Experimentation with this ratio in Table 1’s additional columns (over the sub-group timeframes) led me to believe that this is where Holdren’s figure is coming from.

Going back to the definitions on page 8, we know that capital deepening is defined as “the increase in capital services per hour worked and captures the fact that workers become more productive if they have more and better capital with which to work”. Labor quality is defined as “labor input per hour worked and reflects changes in the composition of the workforce.” It is possible that Holdren is emphasising the importance of higher education for the latter (labor quality) but I am unsure about what Holdren is thinking with regard to the former (capital deepening). My understanding is that technology does play a role in capital deepening, so this may be what Holdren has in mind.

However, none of this goes beyond the "correlation" that you mention Kevin. of course, the problem in relating economic growth to "scientific and technological advances" is that there is a fairly difficult endogeneity problem. In that regard, many of use are familiar with the seminal papers by Paul Romer on “Endogenous Technological Change”, and “Too Much of a Good Thing? The Economics of Investment in R&D” by Jones and Williams.