Comment on Forni's "Using Stationarity Tests in Antitrust Market Definition"

Economic analyses of prospective mergers must typically be undertakenwith limited data. Often prices are the only available data,so it is tempting to use their time series alone. Certainly,simple and robust measures of the extent of the market basedon limited data would aid merger analysis immensely. However,a couple of examples show how Forni's suggested methodologycan lead one astray. Consider a differentiated products market with demand systemln D_i(p_i, p_j) = 1 – ln p_i + ln p_j, i, j = 1, 2, i j,with associated constant marginal costs c₁ = _t c, c₂