Automotive Futures: High Tech or High Dreams?

In the online NBR Japan Forum (and in a chat with a retired vice miniser of METI – the Japanese Ministry of Economy, Trade and Industry [経済産業省]) – I hear about how electronics, hybrids, and yes, fuel cell vehicles will turn around the [domestic] Japanese auto industry. (The US, too, has its advocates of industrial policy.) After all, the Prius has been a huge success.

This is wrong on three counts.

First, it relies upon a mistaken definition of the industry. As Klier and Rubenstein show in their 2008 book, Who Really Made Your Car?, the vast majority of employment – 3 jobs to 1 – is in parts manufacturers, not the car companies. Twenty-five years ago the Detroit Three sharply reversed their vertical integration, with GM and Ford spinning off Delphi and Visteon. In Japan, though, that process of dis-integration began circa 1950, when Toyota as part of its de facto bankruptcy spun off Denso. According to its Census of Manufactures [工業統計表] in Japan the employment ratio is on the order of 4:1.

Furthermore, while high-tech is sexy, it is not the name of the employment game. Mundane-seeming items such as wire harnesses and seats are the most expensive individual items purchased, and while I've sat through engineering presentations on new technologies they incorporate, they remain comparatively labor intensive in their manufacture. As bulky items seats are almost universally made within a short truck drive of an assembly plant, but harnesses come from the Philippines (where they are the biggest single source of manufacturing jobs), China and Mexico. In addition, even if Japanese parts firms were to dominate the high-tech end of the industry (they don't at present), it wouldn't halt the shrinkage of the domestic car market, thanks to demographics, as outlined in my August post, The Decline of the Japanese Auto Industry. Success in high-tech parts won't stem the erosion of domestic jobs.

...high tech won't stem the erosion of domestic [Japanese] automotive employment

Second, to me it is not clear that Japanese suppliers are ahead of the game. There are certainly individual suppliers (and particularly specialized sub-suppliers) that dominate their particular niche. When it comes to transmissions, or headlights, or turbos, or many other items, Japanese firms may be major players, but in the [confidential] "benchmarking" competitive analyses that I see as a PACE judge, they seldom appear on the leading edge. One reason may be that, despite 30 years of Japanese assembly operations in the US, suppliers remain parochial: when I visit firms I am the only non-Japanese in the room, unlike in the US or Europe. Yes, the better firms have R&D outposts around the globe, but careers are made or broken at headquarters, recruitment from the outside is highly unusual, and a "not invented here" attitude is all too common. Let me caution that this is based on a small set of personal observations. I hope I'm wrong: there are many talented people in Japan, but their value is best realized when they're embedded in the global network of automotive engineers.

Third, while eventually electric cars or fuel cell vehicles may be a big share of the market, it's unclear that they warrant effort today. That's not the perspective of Toyota Chairman Uchiyamada, as conveyed for example in Bloomberg. Yes, Toyota has sold 3 million Prius (is the plural Prii??) over 15 years. But with 75 million vehicles on the road in Japan and 250 million in the US, strict hybrids are a scant 1% of the fleet. The share is growing, and if we include vehicles that aren't marketed as hybrids (such as those with start/stop alternators robust enough to boost acceleration and recapture braking energy), the share is larger. Nevertheless, the reality remains that hybrids have not been a market success.

Quite simply, putting two complete, interconnected drivetrains in one vehicle drives up cost and weight. They must be sold at a premium to make a profit, while offering only a modest performance improvement. They also face a moving target, as "standard" gasoline and diesel engines continue to get better – thanks for example to turbos that permit downsizing. Hybrids don't offer sufficient benefits to be a strong value proposition. In some markets, such as Europe, almost none are sold; efficient, tried-and-true diesels dominate. In Brazil, with the wide availability of ethanol and CNG, new cars are designed and assembled to use any mixture of gasohol at minimal incremental cost, and to switch seamlessly to/from CNG and gasohol. Who needs a hybrid to save money? Hybrids likewise fare poorly in China. So far, however, global markets are diverging, not converging, undermining the business case for hybrids, battery electrics and fuel cells.

markets are diverging, not converging

I'm left with a feeling of deja vu from encounters with the "magic bullet" mindset that prevailed at GM 25 years ago. At that time executives, including the visionary Roger Smith who brought us $19 billion in robot-ridden factories that promptly closed, and the Saturn venture that neither made money nor drove the Japanese from our shores, were searching for the "secret" of Japanese management. Once they had that miracle cure in hand, GM with its high-powered MBAs and PhD engineers would in short order leapfrog the competition.

Japan has lots of great engineers and competent managers. I hope they aren't forced to devote themselves to the Sisyphian task of making money from fuel cells.

mike smitka

Japan has a monthly survey of manufacturing [生産動態統計] that provides output data on 42 separate components. As is inevitable, definitions are about what was made, not what will be made, and so do not highlight "high tech" from commoditized parts. The only new items included are airbag modules and electronic braking systems, but plain vanilla frong airbags and antiskid braking systems are now comparatively mature technologies. No electronics per se are listed, and while car audio, car navigation systems, and automotive lighting appear under electrical equipment, semiconductors and sensors are broken down by type and not end use, so there are no data specific to (for example) engine and stability system electronics.