Long Waves, Long Cycles, and Long Swings
Tylecote, Andrew, Journal of Economic Issues
Have there been, since (say) 1800, "long wave" cycles of some 45-60 years in world rates of economic growth? If so, we are in a downswing. Some 60 years' data have accumulated since Kondratiev and Schumpeter's works, with much improvement in data for earlier periods and in econometric sophistication. Yet Solomou  shows that what looks like a regular long wave in twentieth century growth rates can be "dismantled" into a number of episodic traverses without obvious cyclical character. Reijnders , on the other hand, shows that for the nineteenth century, where no long wave is evident, one appears once allowance is made for perspectivistic distortion-the influence on the data of fluctuations of longer and shorter periods than the one looked for.(1) This controversy shows that empirical analysis cannot be conducted in isolation from theory. This paper emphasizes "low theory," which engages closely with empirical knowledge. It seeks to clarify what kind of long wave fluctuations we might expect to find in the data and how we might expect them to have altered over the period. It also asks what other cyclical movements we should look for and allow for to reduce "perspectivistic distortion" and when and how we should allow for "episodic traverses." Finally, and briefly, it considers implications for prediction and policy.
The only cycle of longer period than "long waves" that might be expected to affect economic growth is the "long cycle" in international relations [Wallerstein 1983; Modelski 1987]. This is a cycle of leadership or hegemony of some four generations--100-120 years. During the first generation, a global war takes place, ending with the victory of a new hegemon, or world power. This state thenceforth largely controls world trade and economic relations. In Modelski's scheme, the next generation is one of uncontested leadership, the third of delegitimation, and the fourth of deconcentration, with a serious challenger emerging; then global war returns. During our two centuries, the hegemons have been Britain (till 1914) and the United States (since 1945); the global wars have been the Revolutionary/Napoleonic (1792-1815) and the 1914-45 periods. Periods of global war presumably tend to slow growth. Maddison  and Solomou  show that the world economy did grow decidedly more slowly through 1914-45 than in the 30 years before, and grew far faster in the 30 years afterward. Likewise, we know that growth was slow from 1792 to 1815 in the United States and Britain, and for the decade afterward also in continental Europe [Tylecote 1993a]. This deceleration takes place not only because of the destruction, diversion of resources, and obstruction of trade that result from war, but also because of the vacuum of leadership that exists during such periods. Thus, in 1918-39, the United States behaved without any sense of the responsibilities that went with its new economic power. Maddison  argues very reasonably that world growth was blighted by a surge of protectionism at this time, and clearly this was connected with the effects of the 1914-18 war and with U.S. irresponsibility-for example, the large tariff increases in 1931.
The 1914-45 deceleration was less pronounced for Britain, the ex-hegemon, than for any other major economy, as was the subsequent acceleration for the United States, the next hegemon. Hegemonic responsibilities seem a burden.(2) (One might also expect a state newly arriving at world power status to be enjoying a surge of economic dynamism at the time.) Moreover, neither the old nor the new hegemon suffered nearly as much physical destruction or social disruption as most other contestants in the global war. Much the same appears to be true for the previous global war, except for the oddity that the "retiring" and new hegemon were one and the same, Britain. This explains why Reijnders , examining only Britain, found no "long cycle" output fluctuation. …