Myung Soo Cha, Yeungnam University

Both correlation coefficients and coefficients of variation are commonly calculated using regional prices to see whether markets are integrated. Other things being equal, and as markets become more closely linked, correlation coefficients are likely to rise, and coefficients of variation to fall. Over an extended period, however, other things rarely remain the same in the real world, which often makes the two coefficients imperfect and misleading indices of market integration. I first show that price correlations tended to be lower during the pre-1914 belle Žpoque than from 1914-38. Given the abundance of evidence indicating a disintegrating world economy in the latter period, the increase in the degree of price parallelism after 1913 is most likely to be a consequence of global shocks (e.g. simultaneous expansion of money supply to finance war efforts) or international transmission of aggregate demand shocks (e.g. deflationary policies to return to and stay on the gold standard) via trade and capital flows. These can cause most prices in different parts of the world to move in the same direction (generating high correlation coefficients) in the absence of market integration. I also show that coefficients of variation tended to be higher under a floating than under a fixed exchange rate regime. If a shift to a floating regime allows exchange rates to respond sensitively to shocks, with prices remaining sticky and adjusting sluggishly to remove the resulting international price gaps, a consequence of the regime shift will be a rise in coefficients of variation. I set up and estimate structural vector autoregression models to filter the impact of global shocks, internationally transmitted aggregate demand shocks, and exchange rate shocks out of regional rice and wheat price fluctuations. Correlation coefficients are then calculated using thus "purified" price series to determine whether various pairs of major rice or wheat markets are integrated in three different periods, 1877-1913, 1914-36, and 1961-94. "Pure" correlation evidence demonstrates more than anything else that "raw" price correlations substantially exaggerate the actual extent of market integration, particularly after the First World War, when massive aggregate demand shocks were generated and spread around the globe. Pre-1914 rice markets were not so well integrated internationally either as raw price correlations suggest or as pre-1914 wheat markets. If wheat markets seemed to become somewhat dislocated in 1914-36 in the wake of hostilities and the Great Depression, Asian rice market became better integrated during this period, probably because they were less closely linked with industrialized countries and therefore relatively insulated from the political and economic dislocations. Although peace returned and protectionism receded after 1945, rice markets grew more fragmented, and wheat markets failed to return to pre-1914 level of integration. I interpret this as a consequence of government interventions, as exemplified by policies pursuing rice self-sufficiency, industrialization through import substitution, and protection of domestic agriculture.