Two essays in asset pricing : PEAD anomaly in equity market and arbitrage strategies in fixed income market

Abstract

The first essay examines whether customers' or suppliers' earning announcements can help reduce firms' own post-earnings announcement drift (PEAD). On one hand, customers' or suppliers' earnings contain information helpful to resolve earning uncertainty, causing investors to respond more to the news. On the other hand, announcements by a customer or supplier would draw more attention to a firm, and the increased attention would lead to more informed trading and price movement after announcement. Both these two channels imply that the PEAD drift will be smaller for firms with customers' or suppliers' earning announcements before their own. With a novel database of supply chain information, I find that firms with customers' or suppliers' earning announcements prior their own earning announcements have lower drift in the post-announcement period, but higher cumulative abnormal return (CAR) during the three-day event window. This effect is robust and more pronounced in the small-size, negative earning surprise firms. Moreover, I find that more than half of the firms either have customers' or suppliers' earning announcements in all quarters or not have it in any quarters, and the additional information from supply chain can reduce analysts' forecast error. The second essay investigates two arbitrage strategies in fixed income market: treasury on/off spread and CDS-Bond Basis. For the treasury on/off spread, I test whether the spread convergence speed is higher when market liquidity is abundant and vice versa. There are 35 ideal convergence cases among the 97 issuing cycles: the on/off spread is high at the beginning and gradually converges to almost zero near the cycle end. However, the on/off spread in other cycles is volatile and it is difficult to find a valid model to describe the convergence process. At current stage, there is merely anecdotal evidence that the convergence speed is much lower during the crisis than in normal times. For the CDS-Bond Basis, I test whether the negative basis trading is a potential explanation to the extremely negative basis during the 2007 financial crisis. For firms in the most negative basis group, there is a significant higher customer net buy before the sorting time and a significant higher customer net sell afterwards when they are sorted in Dec2007 or Jun2008 (three months before the bankruptcy of Bear Stearn/Lehman Brothers). However, due to data limitation, the results are only robust at these two data points. First, there is no identification to separate basis arbitrage trades from others; second, small sample size makes it difficult to be statistically significant; last but not the least, the timing of basis arbitrage is unknown, the effect of customer net buy or net sell might be wiped out when numbers are averaged across a period of time.