Zheng Liu

Date of Award


Document Type


Degree Name

PhD in Accountancy


Department of Accountancy

First Advisor

Ari Yezegel

Second Advisor

Rani Hoitash

Third Advisor

Dan Palmon


This first paper (co-authored) tests the effect of Markets in Financial Instruments Directive II (MiFID II), which entered into force on 3 January 2018, on European-based sell-side financial analysts. We find that MiFID II, in part, unbundles brokers’ execution and research services. We assess the effectiveness of MiFID II in unbundling and find that the trading volume generated by the brokers that issued recommendation revisions declined significantly after the enactment of MiFID II. Further, in addition to a smaller price reaction to EU broker analysts’ recommendation revisions in the post period than in the pre period, these broker analysts appear to provide less accurate forecast and cover fewer firms, indicating a “chilling effect” of MiFID II. However, a silver lining is that these broker analysts forecast bias decreases and forecast frequency stays the same.

The second paper (solo-authored) tests how US state-level social distancing polices, effected since mid-March 2020, are related with US-based sell-side financial analysts’ research, in the period ends in June 2020. Preliminary forecast accuracy related tests show that analysts gain “non-local advantage” after the effective of the policy and this is especially for the case when firms do not issue earnings guidance. Further results show that the seriousness of COVID-19 and the county-level political attitudes does not replace the results of the stay home-work safe policy, and that nonlocal analysts gain more advantage for firms in industries that are less sensitive to COVID-19.

The third paper (solo-authored) tests whether and how sell-side financial analysts in the United States react to the holidays of their country of origin. Results show that that during the one to three days period before these holidays, analysts’ forecasts are less biased; but during the one to three days period after these holidays, analysts’ forecasts are more accurate, and recommendations are related with more market reactions. Results stay when excluding US holidays, Mondays and Fridays, January or December, and stay when controlling country-level trust or hierarchy or individualism.