Thursday, May 31, 2012

Back to the [Economic] Future

Telephone Switchboard Operators - a vintage circa 1914 photo (cropped)
The economy of the future isn't going to look like what you think it will. In fact, it very well may look more like the economy of the past than the economy we live in today.

And that won't be because of retrogression or stagnation. It will be because what we identify as "the future" -- the omnipresence of tech and information jobs -- simply is not going to happen. American business will not transform into the economy of the Jetsons.

Instead, expect to see the return of manufacturing. That's where "technology" will be the relevant future; it will make the United States, for decades now priced out of manufacturing due to high labor costs, competitive in high-productivity, high-end manufacturing. (See here and here for more of my thoughts on this question.)

For proof, just look for the jobs. I've constructed a broad measure of employment in the supposedly "new economy" -- the sum of payrolls in telecommunications; data processing, hosting, and related services; consumer electronic durable goods manufacturing; computer systems design and related services; and computer and peripheral equipment manufacturing. (Please let me know if I missed anything.)

Those are all of the tech jobs I could think of in the BLS numbers. Guess what: except for computer systems designers, they are all shrinking, and most of them enormously so. Here's the combined measure, graphed from 1990 to today.Karl Smith recently put up a graph which showed a more concise version of this point, comparing information technology shipments to primary metals shipments. Respectfully, though, I don't exactly agree. Smith uses the "value of manufacturer's shipments," but those are nominal, and obviously there has been historic deflation in technology products. Comparatively, you wouldn't point to high nominal spending on health and education services to say that those industries are booming -- you'd look at the slow real growth and say that costs are going up faster than price indices.

The nominal value of aggregate "new economy" goods and services is falling, and that's connected to the contraction in labor input costs, and thus the employment contractions -- but it does not follow that the real output numbers look anything like the graph Smith shows, which seems to be his implication.The blue line is the measure he used. The problem is that the "new economy" can grow and grow in terms of real output, as it has here, without any apparent gains in employment. Structural change, insofar as we look at the labor side, which is "what is seen" to most people, won't happen.

In fact, it leads us to an interesting, perhaps even paradoxical, conclusion about the American economy: when industries become more productive here, they shrink their employment even as output grows, making it seem like they are dying out or contracting, and employment shifts into the less productive industries. In short, productivity gains means that workers end up in lower-productivity-gains industries.


  1. MoneyBox had an interesting response to this post this morning

    Productivity and Employment—It Can Go Either Way

    The basic idea being that depending on the elasticity of demand, productivity increases can lead either to increases or decreases in employment in a given sector.

    I put together a spread sheet to lay out the math. I can email it to you if you like ( Keep up the good work, this is a great blog.

    productivity increase price change price sensitivity exponent output increase employment change
    10% =(1/(1+A2))-1 1 =(1/(1+B2)^C2)-1 =(1+D2)/(1+A2)-1
    10% -9% 1 10% 0%
    10% -9% 0.5 5% -5%
    10% -9% 2 21% 10%
    10% -9% 0.333333 3% -6%
    10% -9% 3 33% 21%
    10% -9% 0.25 2% -7%
    10% -9% 4 46% 33%
    10% -9% 0.2 2% -7%
    10% -9% 5 61% 46%

  2. productivity increase,price change,price sensitivity exponent,output increase,employment change

  3. The problem here is that I think this is a bit of a straw man. Most people don't restrict the term "new economy" to tech-based. Most consider it to mean "service based"

    So look at something like health care, which is about roughly 18% of GDP. That's an industry that is not only going to continue to grow as the population ages and the types and varieties of treatments expand, it's also one that is ripe for continued tech-based productivity gains. That's the new economy.

    Productivity is not really anything other than output/labor hours. It's just an equation.

    But ultimately, the mix that constitutes the economy of the future isn't going to be determined by productivity. Productivity is one way to increase the final value of GDP. The other is to simply produce more things. My understanding is that affluent, mature economies in welfare states orient toward service employment. That's just comparative advantage. Manufacturing will become automated and that will make it highly productive, but it won't be what most people intersect with in daily life. I'm sure manufacturing plants in the Jetsons were highly productive as well.

    Last thing: you include two high end durable manufacturing series in your "new economy" aggregate. I'm not really sure what that is supposed to be saying.

  4. For tech employment, I would encourage you to look at occupations rather than industries. "Tech" jobs have been spreading across many industries, so looking only at trends in these particular industries is incomplete and misleading.