| Abstract: | The ‘magnet’ or ‘gravitational’ effect hypothesis asserts that, when trading halts are rule‐based, investors concerned with a likely impediment to trade advance trades in time. This behaviour actually pushes prices further towards the limit. Empirical studies about the magnet effect are scarce, most likely because of the unavailability of data on rule‐based halts. In this paper, we use a large database from the Spanish Stock Exchange (SSE), which combines intraday stock specific price limits and short‐lived rule‐based call auctions to stabilise prices, to test this hypothesis. The SSE is particularly well suited to test the magnet effect hypothesis since trading halts are price‐triggered and, therefore, predictable to some extent. Still, the SSE microstructure presents two particularities: (i) a limit‐hit triggers an automatic switch to an alternative trading mechanism, a call auction, rather than a pure halt; (ii) the trading halt only lasts 5 minutes. We find that, even when prices are within a very short distance to the price limits, the probability of observing a limit‐hit is unexpectedly low. Additionally, prices either initiate reversion (non limit‐hit days) or slow down gradually (limit‐hit days) as they come near the intraday limits. Finally, the most aggressive traders progressively become more patient as prices approach the limits. Therefore, both the price patterns and the trading behaviour reported near the limits do not agree with the price limits acting as magnetic fields. Consequently, we conclude that the switching mechanism implemented in the SSE does not induce traders to advance their trading programs in time. |
