Possible History/Economics Lesson from Japan in the 1990s

Professor Diego Comin of Harvard Business School has a working paper called An Exploration of the Japanese Slowdown during the 1990s. Here is the blurb from HBS:

Why was the 1990s a lost decade for Japan? HBS professor Diego Comin argues that it was the combination of some shocks that lasted for about three years and the response of companies that drastically reduced their expenses in adopting new technologies and developing new ones. Though the severe shocks that hit the Japanese economy did not persist, the investments that Japanese companies and entrepreneurs did not undertake to improve technology and production methods during the 1990s propagated those shocks and made their effects very long-lasting.

I am sure economists will have more to say than I about the paper and its validity (and to be fair, the paper notes several variables and issues future research should take into account). Here, however, I am interested in the idea that one can have a better sense about the relationship between investment in technology and moving an economy out of a shock event. Models can be broken of course. Nonetheless, it may be that work such as this paper provide insight about factors that open the door to possibly better ways to proceed even if they do not have an absolute prescription for current economic ills. In other words, this paper may not fully support this claim: especially now, broad investment in R&D is something our country must do. Yet, as I digest the paper, I hope it explains the possible truth of that claim.

You may also like...

3 Responses

  1. A.J. Sutter says:

    From a quick skim of the paper, the model is framed within a narrative that emphasizes the “decline” of Japan. This is a typical narrative from Western scholars and media, and even from the Japanese press like my publishers at the Nikkei (who can be just as disaster-loving as the NY Times and others in the States, if it’ll spark sales).

    However, as I mentioned in a comment to one of Lawrence Cunningham’s post recently, there is a very different take on this from Richard Koo, who is a US-educated former economist for the NY Fed and now chief economist for Nomura Securities. (He’s also advised several administrations of the LDP, the ruling right-wing party here.) When I heard him talk last week, plugging his new book, he pointed out that Japan’s GDP grew throughout its “lost decade”. The reason, he said, was government spending: until households and businesses started borrowing again, government borrowed the money that they weren’t. Also, the government injected capital to the banks (though at nothing like the scale in the US). In his view, tax cuts and buying bad assets are counterproductive, since they don’t inject any money into the economy.

    Koo’s take-away point is that the usual media narrative of the Japan’s experience is all backwards — actually, Japan’s experience was successful. Koo thinks Japan’s experience should be encouraging to other countries, instead of being used as a role model of gloom.(BTW, although the economy here now is hurting somewhat, it’s nothing like what’s going on in the States, even adjusting for the difference in population size.) I haven’t yet read his book, but I expect you’ll find this theme in it.

  2. Deven says:


    Thanks much for the tips. I am not sure that the paper is gloom/doom. Instead I think it is trying to see what might have moved Japan faster. I suppose that view suggests the status quo was gloomy. Still, I think the paper was saying that spending of a certain type, i.e., R&D, (arguably by the government or the private sector or both) can have large benefits.

    In any event, thanks again for the book tip. I will check it out.


  3. A.J. Sutter says:

    Hi Deven — the doom and gloom comes in the framing of the issue: the piece begins by referring to the Japanese experience as “stagnation,” a phrase repeated several times, the economy is said to have “continued to deteriorate” (@and the conclusion refers to the 1990s as a “lost decade”. Moreover, the very question the paper asks presupposes a norm of adoption of technology, and judges Japan deficient: “The slow speed of computer adoption in Japan is reflected by the fact that Korea, a significantly poorer economy, surpassed Japan in computers per capita. An almost identical picture emerges from the discussion of the internet. The question that these findings beg to ask is why did R&D and the speed of adoption decline in Japan” (@4).

    Turning to the substance of the analysis: The author concludes that “[T]he model does a surprisingly good job in explaining the Japanese slowdown during the 90s specially given that I have calibrated it using the US parameter values and that there are a number of factors completely missing from the analysis which must play a role.” (Id., emphasis added.) I suggest that at best all the model has done is to model the graph of certain quantities, but not to explain anything. Among other reasons, the output could be a fortuitous output of the underlying assumptions (including both equations and initial conditions); this is all the more likely if “factors that must play a role” have been omitted. Philosophically, there’s also an ambiguity in the notion of “explaining” a “slowdown”: is that an explanation of its causes, of what it was like to experience it, of its effects, or what?

    The thesis of the model is that “temporary declines in the investments in improving the existing technology through R&D and adoption” lead to permanent declines in the amount of technology, which fact in turn provides “the propagation power of the model” (@9). The model also links total factor productivity growth with the growth rate in new technologies (id. @ Eqs. 8 & 9). So it’s interesting that in fact the model underestimates the amount of actual R&D spending (Fig. 14 and @13-14; though to be fair, the amount of the miss, ~20%, is exaggerated in the visual by the misleading truncation of the ordinate axis).

    So even though this model is claimed to have “explained” the slowdown in TFP, which is framed as a result of a decline of R&D spending, actually R&D didn’t decline as much as the model predicts. (Note that the discrepancy between model and actual in the TFP graph, Fig. 13, is much less (~3-5%) than that in the R&D graph). I’m not in the habit of quoting Robert Solow favorably, but I’m reminded of his famous remark, “The idea of endogenous growth so captures the imagination that growth theorists often just insert favorable assumptions in an unearned way; and then when they put in their thumb and pull out the very plum they have inserted, there is a tendency to think that something has been proved.” Only in this case, I think the plum is still in the pie.