Chips As Usual

In Praise of Industrial Policy

CHIPS As Usual: A Defense of US Industrial Policy

Clyde Prestowitz

(I am grateful to the Hinrich Foundation for also publishing this article)

Recent articles about the semiconductor industry and the Biden administration’s CHIPS act to provide it with government support have described it as having evolved as a result of entrepreneurial exploitation of free markets, free trade, and globalization.

In fact, as a long time advisor to and U.S. negotiator for the industry, I must say that it is closer to a perfect example of industrial policy and mercantilism than of free international markets.

The transistor--predecessor of today’s microchips--came out of Bell Laboratories, a unit of the then government-regulated AT&T telecommunications monopoly. It was developed into integrated circuits and today’s memory and processor chips by the entrepreneurs like physicist William Shockley and Intel founders Robert Noyce and Gordon Moore.

But it was the US government who paid much of the way. Intel was established in part on a loan from the US government’s Small Business Administration. The Defense Advanced Research Projects Agency (DARPA) was a major semiconductor industry investor, and the US Defense Department was a major customer.  

By the early 1980s, the US semiconductor industry produced 70%-80% of the world’s chips domestically. However, there was a rapidly growing challenger–the Japanese semiconductor industry. Since the end of the Second World War, Japan used industrial policy to rebuild its economy. As Vice Minister for International Trade and Industry Naohiro Amaya once told me: “We did the opposite of what the American economists advised.”

Soon after the invention of the transistor, Japan’s Sony Corp. paid $25,000 for a license and MITI was off to the races to help the Japanese chip industry catch up to the US. By 1985, the Japanese producers not only had gained about 40% of the global market, but had also beaten the US makers to the new 256K random access memory chip.

How had they done this so rapidly?

The real keys to Japanese success: MITI sent written instructions to Japan’s major chip users telling them to buy Japanese. Japan’s banks were directed by MITI to make cheap capital available for investment in semiconductors. Japan’s Ministry of Finance intervened in international currency markets to maintain a weak yen versus the dollar, reducing the price of Japanese exports and vice versa.

The chief executive of Intel Japan used to have what he called a “waterfall map”, which showed Intel sales in Japan climbing dramatically whenever Intel introduced a new chip, but fall dramatically as soon as there was a Japanese-made version of the product.

As counselor to the Secretary of Commerce in the 1980s, I was one of the lead US negotiators on trade with Japan including on semiconductors. Our experience was that Japanese producers systematically dumped (sold below cost of production and/or the home market price) their products in the US market. The success of Japan’s semiconductor industry was due much less to free trade and globalization than to industrial policy, protectionism, and mercantilism.

To compete better with the Japanese, US producers began searching for cost reductions. In the mid-1980s, it occurred to American makers and the governments of countries like Singapore that tasks such as testing and packaging might be performed at much lower cost by inexpensive female labor offshore. This would help makers compete better with Japan. To encourage this, Singapore, Malaysia, and others provided land and utility subsidies, low-interest loans, capital grants, and 20-year tax breaks.

Watching all this was the government of Taiwan. For its self-preservation, it needed to stay ahead of economic development in China. Taipei approached my friend Morris Chang who had recently retired in the late 1980s from a career making chips at Texas Instruments. Why not, thought Morris, have a single foundry that could produce chips designed by individual makers? In this way, competitors could avoid creating excess production capacity and boom and bust market prices as the result of creating excess capacity if two makers built new factories when the market really needed only one.

It was a game-changing insight, but it required one important thing: gobs of money made available by the Taiwanese government. The success of Taiwan Semiconductor Manufacturing Company (TSMC) had everything to do with industrial policy and managed trade.  

This brings us to China and eventually the new US CHIPS and Science Act of 2022. Beijing carefully studied Japan and the Asian Tigers as they pursued managed trade, undervalued currencies, and industrial policy. It aimed to entice US producers to put fabs in China. I served for some time on Intel’s Policy advisory board and sometimes traveled with Intel executives to China. I remember in one meeting that the Chinese asked how much a new fab would cost. The Intel people at that time said about $6 billion. The Chinese response was to offer free land, $1 billion in capital grants, subsidized utilities, and 10-year tax breaks. Intel eventually put a fab in China. But it hedged. The fab was not Intel’s most cutting-edge model.

That the hedge was wise became evident five years later when China adopted its Made in China 2025 policy, aiming for high-tech self-sufficiency.

This policy made clear that China does not believe in globalization, or plan to practice Western, neo-classical, comparative advantage-based trade or even fully free domestic markets.

It also makes clear that concerns of CHIPS’ potential restraints on US corporations’ China operations are misplaced. These companies are going to lose market share in China regardless of what the US government does. The Chinese Communist Party is dedicated to maximum self-sufficiency and control of cutting-edge industries.

In view of all this, it may well be wise for the US to go back to its own future. From 1816 until 1948, America pursued trade and industrial development essentially as China, Japan, and the Asian Tigers have done. It protected key domestic markets and was a champion of industrial policy.

At the height of the American Civil War, when Britain was the low-cost steel producer, President Abraham Lincoln imposed a high tariff on steel imports. He said he did not pretend to know much about the political economy, but understood enough to know that “when an American paid $20 for steel to an English manufacturer, America had the steel and England had the $20. But when he paid $20 for the steel to an American manufacturer, America had both the steel and the $20.”

America became rich by using industrial policy to spur development of the railroads, the telegraph, the airplane, the semiconductor industry, and the internet. It all worked out pretty well. Maybe it will again.

---

 [CWY1]Or some?