For people who purport to be skeptical of normative analysis, economists are awfully fond of morality tales. One of our favourites is The Legend of the Auto Quota.
Once upon a time, the US imposed a quota on imports of Japanese automobiles. Japanese automakers responded by upgrading; packing every car they sold full of features. Thus the quotas, designed to protect the American auto industry, had an unintended consequence: they pushed Japanese auto makers into manufacturing high quality mid-range and luxury vehicles. The rest is history.
The story is a staple of economics textsbooks (for example, this one and this one and this one), because it illustrates economists' favourite moral: any attempt to interfere with markets will come back to haunt you.
But is it really true?
The evidence that auto quotas caused upgrading comes from a classic QJE article by Robert Feenstra, ungated here. In it he shows that Japanese cars gained weight, width, height and horsepower around the period when the 1981 auto quotas were imposed, and added on features like automatic or five-speed transmissions, air conditioning, and power steering.
I wanted more specifics, so I asked auto guru Nick Rowe – exactly what features were added after the quotas, and in which models? He sent me this link, complete with photos and video clip:
There's only one problem: these links show that upgraded features like power steering and five-speed engines were introduced before the import quotas were imposed in 1981. This is consistent with Feenstra's results, which shows that the biggest increases in the quality of Japanese automobiles happened in the 1981 (7.6 percent quality improvement) and 1982 (7.0 percent improvement) model years.
But the 1981 model year went on the market in the summer of 1980; the 1982 model year went on the market in the summer of 1981. The quota was imposed in April, 1981. There is no way that the 1981 quality upgrades could be a response to the quota; even attributing the 1982 upgrades to the quota is somewhat problematic. Feenstra acknowledges this point in footnote (p. 143) saying "In attributing the 1980-1981 quality upgrading to the trade restraint, we are supposing that Japanese producers anticipated the quota."
That's possible. It's also possible that innovations in engine design were a response to the emission regulations Japan enacted in the 1970s, and that increased vehicle size was a response to the decline in real oil prices throughout the 1980s. The introduction of the Honda Accord five years before the quota was imposed and the Lexus and other lines several years afterwards suggests that quality upgrades were part of a long-term plan by Japanese auto-makers to penetrate every segment of the American car market.
Do I believe that, faced with import restrictions, Japanese auto manufacturers filled their freighters with more Honda Accords and fewer Honda Civics? Yes. Do I believe that taking into account quality adjustments is important when estimating the welfare loss due to import quotas? Yes.
But do I believe that the story told in Nick Wilkinson's Managerial Economics textbook, that "Japanese firms redesigned their cars to make them more luxurious and expensive." No. The technological innovation was happening anyways, as was the pursuit of the higher-end market.
A rework of Feenstra's original analysis using modern regression discontinuity analysis techniques might be one way of getting at the truth behind The Legend of the Auto Quota, as would a more institutional approach, that looked at the process of model design in Japanese auto manufacturers. I suspect that such an analysis would find that the real story of the auto quotas is a lot more mundane than managerial economics textbooks make it out to be.
Worthwhile Canadian Initiative 
Hey – its a good story – who cares if its true?
More seriously, I think this reflects something that I try to be conscious of: a good narrative explanation is tempting, but its truth value is often limited.
Isn’t there data? You could check how many Accords had, say, cruise control, compared with, say, Chevy Citations.
Phil – the data is what Feenstra used in his analysis. The question is: base model v. upgraded model, and is it possible to get data on the # of upgraded Citations sold v. the # of base model Citations sold? Feenstra used base model data only, so part of what he’s picking up is a marketing decision – the US manufacturers tended to have less loaded into their base models, and then used high pressure sales techniques to get consumers to upgrade, while the Japanese manufacturers deliberately chose to employ a more all-in, less-high-pressure sales strategy. I think. Hopefully Nick is reading and will correct me if I’m wrong.
Whitfit – that’s true!
It isn’t clear to me how you would use a regression discontinuity design for this problem. For one thing, there’s fuzziness about the actual dates of the effect, as you mention, so the estimates would be imprecise. But additionally, it seems to me that the hole point here is that pre and post quota cars have very different characteristics, which may or may not be caused by the quota. i don’t see how an RD design resolves that discontinuity. A difference-in-differences design might, if you can find a suitable control variable.
I seem to remember that both Nissan and Toyota began to market bigger vehicles in the USA earlier in the 1980s. Throughout the 70s it was mostly 4 bangers with 4 speed manual transmissions.
The Maxima and Cressida with more powerful sixes with 5 speeds and automatics came in around 1984-85. The bigger engines could run AC too.
Anyway, here’s one Canadian initiative that was worthwhile (and apparently still is):
http://blog.seattlepi.com/seattlepolitics/2013/08/15/a-tax-that-is-taxing-the-ports-of-seattle-tacoma/
Reading between the lines of this official history of Toyota, it does appear than the Voluntary Quotas had some effect on how they did business in not only the U.S., but Canada and Europe too.
Patrick, that official history gives one example of Toyota’s strategy here:
“Firstly, it reviewed the model mix of the limited number of vehicles available for sale in order to efficiently improve profitability. The best example of that initiative was the redesigned Celica, launched in autumn 1981. TMS promoted the Celica’s high performance and quality, succeeding in expanding the vehicle’s sales to the extent that the Celica stood side-by-side with the Cressida in terms of profitability. The Celica Supra, in particular, which won Motor Trend magazine’s “’82 Import Car of the Year” award, helped improve the image of Toyota’s technology in the public’s eye.”
I’d argue that’s still a “putting different cars on the freighter” story rather than an “increased rate of technological change” story.
Yep, they were already redesigning. But the quotas left them no choice: they had to ship the higher-value products.
Matthew: “It isn’t clear to me how you would use a regression discontinuity design for this problem.”
Feenstra did a hedonic regression, where the dependent variable was quality. Hence it is precisely getting at the point “that pre and post quota cars have very different characteristics.” If the upgrading story is correct, the rate at which quality is changing over time should increase around 1981.
Unfortunately Canada imposed import quotas too, otherwise the Canada/US comparison would be a perfect diff-in-diff strategy. If the Canadian controls were less binding than the US ones, it might still work. It sounds like Europe had restrictions around the same time too, so that wouldn’t be an ideal a diff-in-diff scenario either.
My understanding was that the quotas biggest effect was the building of Japanese auto plants in the US and Mexico- something that surely hurt the Big 3 far more than the change in quality of the cars that were loaded for export in Japan itself.
AIUI Toyota built its Cambridge plant in 1988, Honda started in Alliston in 1986. From that time on, they were no longer “imports”, in my view, but domestic manufacturers, on a par with GM, Ford and Chrysler. The auto bankruptcy episode should have made it clear that the Big 3 are foreign (American) corporations which operate in Canada, no different from Toyota, rather that “domestic” manufacturers.
‘I’d argue that’s still a “putting different cars on the freighter” story rather than an “increased rate of technological change” story.’
I don’t see why it has to be ‘either/or’. Toyota clearly changed strategies because of what they called ‘trade frictions’. They did more than one thing in response.
Yancey – yes, building plants in North America was another way that Japan circumvented the quotas – I don’t know the extent to which they also outsourced assembly to S Korea etc also, but that was also starting to happen around that time.
At the same time, it’s important not to be too parochial and try to explain changes in Japanese auto manufacturers’ behaviour entirely in terms of the US market. The 1970s and 80s are the time of really astonishing growth in Japan. This meant, first, that there were big increases in consumers’ incomes. Japanese consumers are notoriously demanding, and this would have led manufacturers to search for way to improve quality. In terms of the Japanese market, recall too that the 1970s emissions controls essentially gave Japanese consumers huge incentives to retire their cars early and buy new ones (see my previous post on Japanese emission controls). On the wage side, too, there were changes happening in Japan in the 1980s – labour is no longer quite so so cheap and abundant.
Fran R Wooley: wages changes: exactly what happened to Germany who switched from the Beetle and BMW 2002 ( not to say the Heinkel Bubblecar…) to today’a Bimmer.
Chancey Yard: the effect was the same as the canadian trade barriers at the beginning of the 20th century. The number of blue-collar jobs was essentially unchanged (though not their distribution among firms)but the ownership and control passed to foreigners. Of course, if the Big 3 had been as technically, managerially and commercially efficient as the Japanese, they would have had no problems. But they stood in the same situation as say the canadian aut in 1910-1920omakers. The least-efficients folded while McLaughlin, even though it was succesful and well-managed,saw the wisdom of merging into GM.
Well, I have been a car nut since the 60s…
In the 70s, most American brand cars were custom ordered – either by the buyer,, who would wait a couple of months for his car, or by the dealers, who decided what suited the local market. The manufacturers also built some cars for inventory (Chrysler was worst). Each car was different – read teh book future shock, there were about 1 million different ways a mustang could be ordered from the factory. Only pickup trucks are like this.
The japanese were bringing cars in from across the world – so they came up with a diferent method – have a limited number of exterior/interior colours, and have a limited number of trim levels. it was cheaper to put a cigarette lighter in every car even if people didn’t want it than to give people a choice. for some options, like air conditioning, it might be a factory sourced kit that the dealers could install. I bought 1980, 1984 and 1988 Hondas.
Partof the reason for the shift from small imports to bigger ones was that the cars were primarily designed for the company’s domestic market – these countries were less prosperous and had higher gas prices and luxury taxes. As these countries became prosperous, the home market could support bigger models… and the US used to export big cars around the world – look at a french movie from the late 50s or early 60s and you see a fair number of US models.
As companies like toyota and honda increased their uSmarket share and dealership network, they could start to build models designed mainly for the US and Canada – which generally buy much bigger cars than Europe, japan etc because that is what people want and the infrastructure (streets, parking) was built for. The Honda Accord is quite different from the one sold in Europe or Japan.
There is the story that in motorcycles, Honda initially wanted to go head to head with Harley – they tried and failed, but when they started importing small bikes – under 400cc, their sales took off (The Beach Boys even sang about them!) but eventually their reputation etc. meant that they could go against Harley and the British bikes.
The other dodge was to build cars in canada, and export them to the US – but import the same model into canada from Japan so on paper imports from jsapan into the uS seemed low!