Evidence is scarce and inconclusive on the announcement effect of dividend changes on bondholders due to poor quality and availability of bond price data. This paper fills this gap using daily bond transaction data from the over-the-counter market. Most of my results are more consistent with the signaling hypothesis than the wealth transfer hypothesis. I find that abnormal bond returns over a three-day event window surrounding an increased (omitted) dividend announcement are positive (negative) and statistically significant. Consistent with the signaling hypothesis, the bond market reaction to the dividend increases announcement is more positive for larger percentage dividend increases, speculative grade bonds, and the period from 2008 to 2010. While most of my results are consistent with the signaling hypothesis, my findings of insignificant (insignificant) bond market price reaction and significantly negative (positive) stock market price reaction to dividend decreases (initiations) announcements suggest that there is also a wealth transfer effect.