A Bayesian spatio-temporal error correction analysis of markets during the Finnish 1860s famine

We present a Bayesian spatio-temporal error correction model and use it to analyze grain market integration in Finland during the 1860s famine. Compared with the existing error correction methodology, our approach allows simultaneous modeling of multiple interdependent time series without cumbersome statistical testing needed to predetermine the point of reference -- the market leader. Furthermore, introducing a spatio-temporal structure enables a flexible analysis of the regional and intertemporal variations in the market mechanism. We detected spatially asymmetric ``price ripples'' that spread out from the shock origin. We corroborated the existing literature on the speedier adjustment to emerging price differentials during the famine, but we detected this principally in urban markets. This hastened return to long-run equilibrium means faster and longer travel of price shocks. This, in turn, implies prolonged out-of-equilibrium dynamics, proliferated influence of market shocks, and, importantly, a wider spread of famine conditions.