Uninsured deposits represent a theoretically appealing but relatively untested alternative to subordinated debt for incorporating market discipline into banking supervision. To make the deposit market a useful supervisory tool, it is necessary to know what types of risk are priced by depositors and in what proportions. Using a clustering technique to select from among a large set of potential regressors, as well as a carefully chosen set of control variables, we attempt to determine the types of risk that cause uninsured depositors to react in both the price and quantity dimensions. As a benchmark for economic significance, we estimate similar regressions on supervisory ratings. We find that, in contrast to government supervisors, depositors have not priced most types of risk since 1997. Indeed, the only risk variables that consistently come up as statistically significant are those that measure capital adequacy. Our interpretation of these results is that, because aggregate banking conditions are good, it is not worth depositors' effort to investigate individual bank quality very carefully. We conclude that, in the current economic and regulatory environment, the market is content to delegate most of its monitoring and discipline to the government. To the extent that it does monitor, it only monitors capital. The jumbo-CD market is thus not likely to be of much supervisory use, particularly given that examiners already have good information about capital levels. The depositor emphasis on capital also supports the conjecture that market discipline was responsible for much of the recent capital build-up.