Gordon Moore propounded in 1965 his Law, that the number of transistors in an integrated circuit doubles every two years. That Law has been gradually ceasing to apply over the last few years. Meanwhile, Eroom’s Law, that the cost of discovering a new drug doubles every nine years, appears to continue in full strength. Our future thus hangs in the balance; in the Manichean struggle between Moore and Eroom for dominance of our economy. Regrettably, my money is on Eroom.
Gordon Moore was the founder of Intel; his Law has been a staple of the semiconductor industry and the tech sector in general since he propounded it. Since about 2005, semiconductor and device manufacturers have used tricks, such as massively parallel processing, to keep the capacity and functionality of electronic devices moving ahead in line with Moore’s Law, but the tricks available have become more and more marginal. That’s why Apple has failed to produce a product that is a significant step forward since the Apple 5S in 2012; it is not fair to claim that the late Steve Jobs (1955-2011) was the only decent designer in the place. Physics, not mortality has caught up with the company.
Eroom’s Law is as valid as Moore’s, and as time-tested. In the drug business, it has four causes. First, if a drug such as Lipitor is effective against a major problem such as hypertension, it is difficult to find new drugs that are significantly more effective. If every rock group had to be “better than the Beatles” there would be no new rock groups – and if the market had demanded that the Beatles be better than Mozart we would have been spared half a century of ear-splitting discordance.
Second, regulators become more cautious over time, sometimes with the encouragement of regulation-loving politicians like President Obama. Third, companies throw money at research and development, creating huge teams that get in each other’s way. Finally, there is a bias towards using “brute force” screening methods that are increasingly ineffective as complexity increases.
Moore’s biography is well known, and explains his insight; so too does that of Eroom. Tomislav Nepomuk Eroom (1892-1967) worked under Franz Kafka at the Bohemian Workers Compensation Bureau, where he gained an understanding of Austro-Hungarian bureaucratic procedure. After military service in World War I and graduation from the University of Vienna he moved to the United States in 1923, becoming a pharmacologist at Merck & Co. He then acquired a PhD in Biochemical Philosophy and undertook various academic posts, including a period at Princeton’s Institute of Advanced Study with Albert Einstein and John von Neumann. His deep insight into both pharmacology and state bureaucracy led to his 1955 masterpiece “Bürokratie und das tempo der Entdeckung” which quickly became known as Eroom’s Law and formed an inspiration to Moore’s countervailing insight a decade later.
There are many examples of Moore’s Law-like behavior outside the tech sector, and many examples of Eroom’s Law-like behavior outside the pharmaceutical sector. The earliest Moore-like example was the steam engine, which became exponentially more efficient following James Watt’s patent in 1782. As each new advance in steam technology was made, from then until around 1870, new uses for the steam engine became possible, and the steam engine became progressively more ubiquitous throughout 19th Century life. Similarly, following Karl Benz’s first internal combustion engine automobile in 1885, improvements followed thick and fast, and by 1930, at least in the United States, the automobile had revolutionized individual existence. Electric power, also, was ubiquitous in the urban United States within well under half a century from Edison’s first generating station of 1882.
Eroom’s Law also applied in the automobile sector. Even early on, the British law before 1896 requiring all automobiles to be preceded by a man carrying a red flag slowed development significantly. Then the 1910 “horsepower tax” forced British cars to be hopelessly underpowered for the next half century and blighted their chances in the post-war U.S. market (automobile engines built by British manufacturers according to tax-optimizing formulae fell apart when cruising at high speeds on U.S. roads.) In the United States, automobile advances slowed to a halt between 1950 and 1990, as manufactures strained to cope with the demands of the regulators, environmental and otherwise.
Eroom’s Law has applied to an increasing number of sectors in recent decades, not all of them obviously affected by environmental constraints. In aircraft design, the speed record was set as long ago as 1976, even though the first man to break the sound barrier in 1947, Chuck Yeager, is still with us (test piloting is clearly a healthy occupation!) In rocketry, we are now 45 years from the last Moon landing, and have still not made a manned landing on Mars. The household robots that we all expected in the 1950s are still primitive and far from ubiquitous.
Above all, Eroom’s Law has come to apply in infrastructure. The tunnel under the Hudson that cost $700 million (in today’s dollars) in 1927 costs $10 billion today and takes a couple of decades to plan, let alone to complete. California’s High-Speed Rail project should be christened the “Eroom Line” for its cost and snail-like pace of development; as Prime Minister Shinzo Abe said to President Donald Trump last weekend, the Japanese will be happy to show us how to apply a technology they have enjoyed since 1964.
The greatest question for the next generation is thus: where can sectors be found in which a Moore’s Law will apply, at least for a generation or two, and thereby overwhelm all the sectors in which the much more universal Eroom’s Law is doomed to rule. Clearly in Eroom-dominated sectors progress is not altogether impossible, but it is ludicrously time-consuming and costly. Professor Robert Gordon has postulated that we shall have no productivity growth at all by 2100; judging by the economy of 2015-16 productivity growth has already slowed to an infinitesimal pace, not only in the U.S. but throughout the rich world. Eroom’s Law appears to be the central organizing principle of our life.
Of course, battles against Eroom’s Law are possible. Margaret Thatcher took over a pretty Eroom-dominated economy in Britain in 1979, and made at least modest progress reversing it (except in financial services, where the world’s most innovative operations were legislated out of existence, and Eroom’s Law has subsequently ruled.) President Trump, if he de-regulates energetically and takes a construction magnate’s cleaver to the excess costs and delays in infrastructure projects, may at least reverse Eroom’s Law a bit. However, it is not yet clear how big any such reversal will be or how long it will last.
As we have seen in the pathetic economic recovery from the 2008 financial crisis, the forces Eroom-izing the world economy are very strong indeed. The way the world’s politicians have gone about handling the global warming issue has inserted the maximum possible delay and distortion into the world economy, for the minimum possible effect on global warming. Similar tendencies are notable in other areas; financial regulators and international bureaucrats in general, both of whose powers have enormously increased, are both massive generators of the economic sludge that Eroom so presciently identified.
There are however a number of sectors where Moore could make a comeback. Driverless transportation and robotics in general, once they get to a critical mass, should enjoy the types of rapid technological improvement that spreads their usage throughout the system with remarkable speed. Genetic engineering, to the extent it is not thwarted by regulations similar to the pre-1896 British red flag requirements, also has the potential to increase its benefits exponentially over several decades.
In genetic engineering, the Abrahamic religions are likely to be a serious societal handicap in the West and the Middle East, but much less so in areas where Confucianism or Hinduism are the dominant religious environments. (In this context, I have every expectation that if genetic engineering takes off, Pope Francis will issue several anathemas, which with luck will be remembered by future generations as unfortunate lacunae like the 17th Century denunciations of Galileo.)
There are thus two actions we can take in the great battle between Moore and Eroom. First, we can fight hard for the removal of regulatory and institutional barnacles from our economy, that have made the last decade such a miserable one. Second, we can nurture robotics, genetic engineering and other Moore-potential technologies with all our ability, seeking above all to block regulations equivalent to Britain’s pre-1896 “red-flag” laws that would block their development.
Tomislav Nepomuk Eroom was one of the 20th Century’s greatest thinkers, but his victory in the 21st Century is not inevitable, only very likely.
(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)