Signaling through Timing of Stock Splits

We develop a dynamic structural model of stock splits, in which managers can signal their private information though the timing of the split decisions. Our approach is consistent with the empirical evidence that shows that the majority of stock splits have 2:1 ratio but are announced at various initial price levels. The model allows us to estimate the preferences of investors about nominal share price levels from stock split data. In addition, we can decompose the split announcement return into the value of new information and the signalling costs. Our estimates show the signalling cost is around 2%.