We demonstrate that transients in subsidence rate in the Tokai area of central Japan correlate with changes in the rate and size distribution of earthquakes. Three largely independent seismic catalogs cover this region: National Research Institute for Earth Science and Disaster Prevention (NIED), Japanese Meteorological Agency (JMA), and Japan University Network Catalog (JUNEC). Comparing seismicity patterns between these three data sets offers a unique opportunity to verify seismicity anomalies based on independent sources. A study of data quality, magnitude reporting consistency, and consistency across the three data sources reveals that the NIED data for the period 1981--2002 are best suited for seismicity analysis. We spatially and temporally map seismicity rates, finding that a significant decrease in the earthquake rate of M ≥ 2.0 events coincides with an accelerated subsidence period; however, this anomaly disappears when including smaller magnitudes in the analysis. This relative quiescence of larger events can readily be explained when interpreting the transient in seismicity in the framework of a change in the earthquake size distribution, or b value. The background b value of about 0.8 increased in the period 1987.5--1989.5 to a value of b = 1.2, a highly significant change that can be confirmed in all three data sets. The correlation between subsidence and b value change is significant. Stress tensor inversions of focal mechanism data for the anomalous period show an increase in thrusting-type earthquakes. While a unique interpretation of the relationship between subsidence and b value is not possible, we propose that an increase in the locking stress, a slow stick event, is consistent with all observations.