AlphaLetters Weekly Paper Review #2 & #3

Below are two reviews from the guys at AlphaLetters:

Article 1.
Category: Mutual fund managers and corporate board member, education connection
Title: The Small World of Investing: Board Connections and Mutual Fund Returns
Author: Lauren Cohen, Andrea Frazzini and Christopher Malloy
Source: University of Chicago working paper

This paper finds that portfolio managers tend to overweight stocks whose corporate board members share education network (e.g., went to same universities), and in managers’ portfolio such stocks perform significantly better relative to non-connected firms.

A strategy that is long connected stocks held by fund managers, and short non-connected stocks generates 8.4% per year.

The extra return of this strategy is found concentrated around corporate news announcements.

AlphaLetters comments:

1. Why important
First of all, this paper seems to show that one can always expect extra alpha from less used database (education connection in this case).

On the other hand, we do not think the strategy is backed by a very convincing story (economic rationale), and may be a result of data mining: Would you/dare you offer more information to a portfolio manager simple because he/she went to the same college as you did? What’s the likelihood of a board member meeting/talking with a portfolio manager in private (since otherwise, other managers should get similar

Per the analysis below, we think an interesting extension (hopefully a more sound strategy) to be tested is:
· Alumni of influential schools are better managers, the companies they manage earn higher returns compared with others

2. Data
1990/01 – 2006/12 data on mutual fund holdings are from the CDA/Spectrum Mutual Funds database. Portfolio managers’ education information are from Morningstar. Board of directors and senior company officers’ biographical information are from BoardEx of Management Diagnostics Limited, a private research company. This study covers 1,648 US actively managed equity funds and 2,501 portfolio managers, and educational background on 42,269 board members and 14,122 senior officials for 7,660 CRSP stocks.

Stock return and accounting data is from CRSP/COMPUSTAT.

3. Discussions
We question whether this strategy makes economic sense. It’s interesting to note that although portfolio managers earn high returns on their connected holdings and purchases, they can not time the sells well. So these managers can get good news earlier than other, but not bad news.

This suggests to us that maybe connected stocks are better managed companies, and the reason may be that connected stocks are managed by people from more influential institutions. Consequently, these stocks earn better returns compared with non-connected stocks.

Indeed, if one looks at Table IV in the paper, connected stocks earns a 16.05% annually, while nonconnected stocks earn just 7.69% (which is close to the mean of all holding stocks (7.81%). This means that
1.) Connected stocks have better performance
2.) Connected stocks are just a fairly small sample in the stocks covered in the study.

Article 2.
Category: Strategy, analysts target price
Title: The Value of Equity Analysts’ Target Prices
Author: Zhi Da, Ernst Schaumburg
Source: HEC seminar paper

Strategy summary:
Long stocks with high “target-price implied return” (percentage difference between the analysts’ target price and traded stock price), and short stocks with low target-price implied return. The annualized profit is shown to be 22%.

AlphaLetters comments:
1. Why important
This paper makes innovative uses of the “target price” item from FirstCall database. The story is intuitively appealing, i.e., collectively analysts can discern the relative mis -pricing of stocks within sectors (albeit not on sector level). The higher the target price implied return, the higher the expected return.

2. Data
Target price, recommendation, and earning announcement data are from First Call. Prices and returns from CRSP/COMPUSTAT and TAQ.

3. Next steps
When looking into the correlation with other existing factor, we found it difficult to reconcile the two tests that compare performances within value and growth segments (table 21 and table 23). Table 21 shows similar profits for value and growth stocks, while table 23 indicates that the strategy does better in the former. This is important for practitioners that focus on different styles. Another caveat is that this looks to be a high-turnover strategy that needs monthly rebalance.

We are not fully convinced by the proposed explanation that the abnormal return can be due to liquidity. A high r-square with measures of liquidity (the bid-ask spread, price impact) does not necessarily prove a causal relationship. In our view, this finding shows that analysts can tell the difference between individual
firms but not relative value between at sector level.