The long-run pattern of economic growth and recession. According to the Center for International Business Cycle Research at Columbia University, between 1854 and 1945 the average expansion lasted 29 months and the average contraction 21 months. Since the Second World War, however, expansions have lasted almost twice as long, an average of 50 months, and contractions have shortened to an average of only 11 months. Over the years, economists have produced numerous theories of why economic activity fluctuates so much, none of them particularly convincing. A Kitchin cycle supposedly lasted 39 months and was due to fluctuations in companies' inventories. The Juglar cycle would last 8—9 years as a result of changes in investment in plant and machinery. Then there was the 20-year Kuznets cycle, allegedly driven by house-building, and, perhaps the best-known theory of them all, the 50-year kondratieff wave. Hayek tangled with Keynes over what caused the business cycle, and won the Nobel Prize for economics for his theory that variations in an economy's output depended on the sort of capital it had. Taking a quite different tack, in the late 1960s Arthur Okun, an economic adviser to presidents Kennedy and Johnson, proclaimed that the business cycle was "obsolete". A year later, the American economy was in recession. Again, in the late 1990s, some economists claimed that technological innovation and globalization meant that the business cycle was a thing of the past. Alas, they were soon proved wrong.