Learning and Information Transmission within Multinational Corporations

We propose that multinational firms learn about their profitability in a particular market by observing their performance in nearby markets. We first develop a model of firm expectations formation with noisy signals from multiple markets and derive predictions on expectations formation and market entries. Using a dataset of Japanese multinational corporations that includes sales expectations of each affiliate, we provide evidence supporting the model's predictions. We find that a positive signal about demand inferred from nearby markets raises the probability of entry into a new market, or raises the firm's sales expectation in an existing (focal) market. The latter effect is stronger when (1) the firm is less experienced in the focal market (2) the signals from the focal market are noisier and (3) the firm is more experienced in markets where signals are extracted