Greenbackd

Empirical Finance Research Blog has a review of a new paper, Repurchases, Reputation, and Returns, which finds that long-run stock returns are higher for companies announcing buybacks that had substantially completed a previous buyback. In other words, companies with a track record for following through on announced buybacks enjoy higher returns following a subsequent buyback announcement than companies that did not follow through on a previously announced buyback. While that might seem obvious, the paper makes two observations that we find particularly interesting in the context of our investment strategy:

1. Past buyback completion rates are predictive of future buyback completion rates.
2. Stocks with high completion rates but low stock returns following previous buybacks enjoy abnormally large returns following a subsequent buyback announcement.

It’s worth remembering that a buyback announcement does not bind a company to undertake a buyback, a situation we encountered recently: RACK suspends buyback and enters agreement to acquire Silicon Graphics; Greenbackd exits position. Companies frequently fail to follow through on announced repurchase plans. Empirical Finance Research cites a 1998 study by Stephens and Weisbach that found that firms on average repurchase only about 80% of the sum announced.

Empirical Finance Research summarizes the paper as follows:

This author measures the level of completion of previous buybacks, as measured by the shares bought as a fraction of the amount specified in the announcement, then uses this to explain how well various stocks do after subsequent buyback announcements. What she finds is that companies that had low completion rates on a previous share buyback experience much lower returns upon the announcement of another buyback. She interprets this as evidence of the company’s credibility, that investors don’t really believe a company about a share buyback when the company has failed to complete one in the past.

First the author confirms that past buyback completion rates are predictive of future buyback completion rates. Next she shows that the stock returns to a company making a buyback announcement are much higher for those with high past completion rates. Companies in the 90th percentile of past completion rate see returns 2.5% higher than those in the 10th percentile of past completion rate in the three days after a the new announcement.

Despite the size of these returns, this isn’t a very good trading strategy, because buyback announcements are clearly unexpected. However, in the next part of the paper the author finds that long-run returns are also reliably higher for repeat buyback companies with high past completion rates. Two year returns are 13.64% for those companies with above-median past completion rates versus 7.43% for those below the median. We should be leery of these results, however as they are not statistically significant.

Next the author splits her sample of repeat buyback companies into quintiles based on the return to the stock during the previous buyback. The two-year abnormal returns to companies in the lowest quintile (those with the lowest returns after their last buyback) are 17.33% after their subsequent buyback. This result is very statistically significant too.

But if we split this quintile in half based on past completion rates, and buy only those stocks with above-median past completion rate, the returns explode to 27.13% for the two year period.

Empirical Research Blog’s takeaway?

The value of this paper is not necessarily in a specific investment strategy but rather in the insight it provides in to a trading strategy we already knew about. I would be hesitant to implement this stand-alone for the practical difficulty in doing so. But any trading strategy that already uses share buybacks as a signaling factor might benefit from an augmentation that accounts for past buyback completion rate.

(Emphasis added)

Quite.