Sunday, January 30, 2011

Cowen on Innovation

I think Tyler Cowen is one of those people who Shiller thinks is out of touch with the "real world." Cohen thinks that advances we have seen since 1962 are small potatoes compared to what happened between 1905 and 1962.

Let's go back to 1959, which is when I started to notice technology in a serious way. My father bought a new car that year - a 1959 Oldsmobile. My father wanted a radio in the car, and seatbelts, and those had to be installed by the dealer. The car came with an automatic transmission, but for us that was a novelty. This car required attention. It quickly developed oil leaks, its tires had to be replaced frequently, and it did not always want to start in the cold weather. Compared to this, the car that is within reach of the average person in the US today is a marvel. It is enormously safer, more efficient, easier to maintain, and with many devices that did not exist in 1959.

How did we communicate in 1959? We had telephones, but they were clunky devices that did not record messages, and required dialing. Long distance calling was something of a luxury, and calling overseas was virtually unheard-of for the average person. We received our news from the television (two broadcast channels) and via newspapers delivered to the door. The post office was very important. Everything was billed my mail, and transactions technologies were primitive. Transactions were all made with cash and checks, and withdrawing cash from the bank or depositing a check required waiting in a line, sometimes for a long time.

For entertainment, we had movies. We had one movie theater in town. If you did not like what was showing, you could stay at home and watch Ed Sullivan. The event of the year was an annual showing of "The Wizard of Oz" on TV. We had records. Of course you had to purchase the physical vinyl and bring it home to play it. Stereo was relatively new. My father bought one, and all the neighbors came over to hear it. Radio was primarily AM.

Was father was an engineer, and the tools he had were quite primitive. He did most of his calculations with a slide rule (three significant digits, four if you had good eyes), but he also had a machine that could do basic arithmetic. The transistor had been invented, but most of my father's electronic equipment used vacuum tubes.

In the town where I lived, we had a couple of department stores in the downtown: a Woolworths and a Five-and-Dime store. Some things we could buy locally, typically in specialty stores, but there was much that we could not find in town. For those things, there was the Eaton's catalog, i.e. mail order.

It's quite possible that Tyler Cowen lives like his grandmother. I know that I don't. While the technological change we have seen since 1959 is sometimes less obvious than the changes wrought by the automobile and electrification, it is no less phenomenal. Particularly notable are advances in information technology, communication, retailing, and biotechnology. I don't have my fingertips on the latest measurements of secular technological change and change in the quality of goods and services, but maybe some readers could provide us with references. However, standard national income accounting tells us that average growth in real per capita GDP in the US was about the same, pre-1962 and post-1962. What is Cowen on about?


  1. FYI, it's "Cowen".

  2. I agree 100% with you and am baffled by the attention that Tyler Cowen's 'great stagnation' hypothesis has received. Part of, I think, is a mutual admiration you-scratch-my-back-I-scratch-your-back society being run by various 'top bloggers.' Among many in that set there's a tendency to declare 'brilliant,' 'deep,' 'insightful,' etc. many ideas that are quite orthogonal to such notions. Blogs and the technology to easily self-publish are wonderful developments for the exchange of ideas, but the discussion/fascination around the 'great stagnation' claim also demonstrates that the constraints on silliness imposed by rigorous peer review in quality scholarly journals are also useful.

  3. Oops. Cohen error corrected.


    I hadn't heard about Cowen's book, or I heard about it and forgot. Here's a review:

    I'm not sure what data he looks at in there, but at first blush, I don't see that we have some kind of secular stagnation at work.

  4. Steve,

    I also haven't read Cowen's book, but my understanding is that, among other things, he focuses on the serially misunderstood 'stagnant' measure of U.S. real median household income. This piece from the Minneapolis Fed, I think, does a good job on explaining the nature of the misunderstanding I have in mind:

    Apparently, there also are claims about why large health care expenditures in the U.S., in some revisionist thinking about the NIPA system, somehow creates an upward bias in measured GDP. More serious thinking on this particular issue, I think, is provided by Hall and Jones in:

  5. Steve,

    I haven't read Cowen's book, just reactions to it. I get the impression that he's talking about marginal technological improvements, not absolute improvements. For example, see Krugman:

    Or if that's too much to ask, read Bob Gordon's JEP article (2000-ish) on whether the internet is as good as major innovations of the late 1800's (that's a paraphrase of the title). He gave that paper in Melbourne while I was there, and I assign it when I teach intermediate macro. I took a vote once on whether my students would rather have (actually, give up) a fast internet connection or a flush toilet. As I recall, the internet only got one vote.


    PS: all that said, I agree with you that the improvements since 1959 have in fact been pretty awesome (I was in the "slide rule club" in 8th grade).

  6. Is your argument that we have a wider variety of similar things (more movie theaters, more TV channels, more stores) and our cars don't leak oil as soon? This improvement is on the same scale as technological progress between 1909-1959?

    We still drive gas powered cars to the local stores, mail stuff in envelopes, dial the phone, go to the bank, radios, own CDs, etc.

    The life of 1959 sounds very close to my life in 2009. Tell me if the life the average person had in 1959 was close to the life people had in 1909. What compares with, say, the change to society due to the invention of the automobile? What has saved a family as much time, as say, a washing machine?

    I also just replaced my 1965 kitchen with a brand new 2010 one...i'm not seeing much of a difference besides the microwave...still have a stove, cabinets, fridge...they all work almost exactly the same...

    When looking at a graph of life expectancy in the US, i see that from 1900 to 1950ish seems to be on one slope, and from 1950ish to now seems to be on another...where are all the great improvements from medicine and technology? I'm seeing about a 15% increase in the second half of the last century...after a 45% increase in the first half.


  7. Pete,

    There's an element in some of this of what we all feel sometimes. I actually prefer my car with a manual transmission, and I was perfectly happy with window cranks and opening the door with the key. However, if you actually think carefully about what the technological improvements are doing in terms of the production and distribution of goods and services, and measure these things, it's hard to make the case that there is no progress.


    A time machine would be handy. We could send you to 1959. You would find it interesting.

  8. I don't need a time machine, there are plenty of records of history.

    I think you're missing the main point - you say "it's hard to make the case that there is no progress." No one is making that claim. The claim being made is that the rate of progress has been declining. IE, we're still driving cars that eventually leak oil, but not horses. The biggest improvement in cars? Fuel injection in stead of carburetion...yet, the efficiency of automobiles has remained virtually unchanged from the production of the model T: 3 mpg in total over the last century. We have had marginal improvements in most areas, incremental gains in productivity have been about it.

    Life today is a lot closer to life in 1959 than life in 1959 was to life in 1909.


  9. So how do you reconcile that with the measurement? Except for the Great Depression and WWII, US GDP per capita has been growing on trend at a more or less steady 2% for more than 100 years. Our standard tool, the neoclassical growth model, says long run growth is driven by TFP. Apparently Cowen wants to make same claims about the behavior of median income, but in the comments above, some other people cite evidence that he is looking at the wrong numbers. Do you have some other evidence?

  10. These sorts of claims seem to have no basis. Standard measurement suggests living standards are rising steadily over time. Further, a variety of research argues that standard measurement underestimates -- to a considerable degree -- the growth in living standards because much of standard measurement fails to take into account the rapid quality growth for many of the products that people purchase.

  11. "much of standard measurement fails to take into account the rapid quality growth for many of the products that people purchase."

    Yes, it's easy to see that the measurement problems are formidable. A year 1988 386 IBM personal computer running MS-DOS with a black and white screen and without internet access, is a fundamentally different object from what is sitting in front of me in my office. Yet, national income accountants treat these as the same good, and heroically try to make adjustments for quality. Even though these national income accountants are well aware of the problems, it's hard to see how you are not going to underestimate the quality improvement.

  12. There is perhaps one simple explanation of Tyler Cowen's view of the recent developments - he is an economist, a member of the dismal science-:), a little bit of pessimism is expected and in many cases very much encouraged especially now after the crisis. Joking aside, there are many arguments that refute his view of the Great Stagnation, I find those best summarised in Matt Ridley's book The Rational Optimist,

  13. Based on my readings of neoclassical price index theory, I thought that measuring changes in the standard of living was impossible.

    I mean, let’s say that real income increases at a constant rate over time. There’s no way of telling whether welfare or utility has increased at the same rate without access to individual utility functions for that period. Right?

    A price index like the CPI is not a cost of living index, though the BLS (for example) regards the CoL index as a kind of Platonic ideal and a framework for the construction of the CPI. Rather than measuring changes in the cost of living, the CPI evaluates a hypothetical consumer fixed in space and time and facing two alternative price systems:

    “Given an indifference map, we compare two hypothetical situations, A and B. We ask how much income the consumer in B would require to make him just indifferent between facing B’s prices and facing A’s prices with a stated income. Note that the question of whether the consumer has the same utility in A as B never arises”.

    [Fisher, Franklin M. and Karl Shell. (1968). "Taste and Quality Change in the Pure Theory of the True Cost-of-Living Index". Value, Capital, and Growth: Papers in Honour of Sir John Hicks. J. N. Wolfe. Edinburgh, University of Edinburgh Press: 97-139.]

  14. vimothy,

    Yes, but the answer is not to throw up our hands and say that any assessment of economic conditions is a waste of time, of course. We have to be aware of what we are measuring, and think about ways to make the measurement better. Theory can tell you how to do the measurement, and measurement can lead you to better theories.

  15. Stephen,

    I agree with all of that. Sorry, I guess I’m making a very obvious point, but haven’t we just reconciled the two sides of the argument here? You’re talking about two different but related things—the rate of technological progress does not seem to have slowed down much, but that does not mean that welfare has increased at the same rate.

  16. A key question has to do with the income distribution, and whether Cowen is correctly measuring the implications of technological change for the median income earner.

  17. This comment has been removed by the author.

  18. Really excellent post. I was inspired to write my own Cowen takedown...