Comments on: Responses to AFL-CIO’s Critique of the Agrawal Study http://blogs.law.harvard.edu/corpgov/2008/03/18/responses-to-afl-cios-critique-of-the-agrawal-study/ Sponsored by the HLS Corporate Governance Program Mon, 30 Nov 2009 21:08:06 -0500 http://wordpress.org/?v=2.8.4 hourly 1 By: S.M. Jacoby http://blogs.law.harvard.edu/corpgov/2008/03/18/responses-to-afl-cios-critique-of-the-agrawal-study/comment-page-1/#comment-20115 S.M. Jacoby Wed, 23 Jul 2008 18:04:52 +0000 http://blogs.law.harvard.edu/corpgov/2008/03/18/responses-to-afl-cios-critique-of-the#comment-20115 In my humble opinion, there are flaws in the paper arising from its failure to recognize institutional factors that undermine the results. 1. The paper has a timing problem. Its sampling periods are designed to capture the consequences for proxy voting of the breaking away of several unions from the AFL-CIO into a new federation called Change To Win (CTW) in 2005. The time period of the study is 2003-2006, leaving 14 months to identify a shift in voting behavior. But the Carpenters disaffiliated from the AFL-CIO in 2001, not when they joined CTW in 2005. There are no controls for this in the paper. The question then is: assuming that the paper’s methodology is sound and the Carpenters were voting like the AFL-CIO staff fund between 2003 and 2005, why did they do so? That is, assuming, as the paper does, that a union departs from its federation’s pattern after it leaves a federation, why did the Carpenters wait until 2005 -- a full four years after leaving the AFL-CIO -- to change their voting behavior to better represent their members’ interests (vs. those of their retirees)? Note that when the Carpenters left the AFL-CIO in 2001, it was cause for great acrimony and condemnation by the AFL-CIO. So, we may conversely ask: why did the AFL-CIO wait until 2005 to change its voting behavior against the Carpenters? One can always rationalize these problems but another possibility is that the study’s findings might be fluky. 2. Taft-Hartley pension plans are joint union-management funds. They are not easily controlled by the union as could be the case with union staff funds. Employers get an equal vote. Why would employers support a union-initiated shift in proxy voting? In most instances, employer fiduciaries of Taft-Hartley funds resist any attempts by union fiduciaries to “politicize” a fund’s behavior. 3. Gerald Davis and E. Han Kim published a 2007 article in Journal of Financial Economics that showed non-fiduciary voting behavior by mutual funds. They found that when a mutual fund was part of a larger financial entity that supplied financial services to client corporations, the fund was more likely to vote its shares in favor of the client than other funds. Assuming that the Agrawal study is sound, it would appear that union pension funds are no different than mutual funds. Their proxy managers -- who vote fund shares (the union does not usually vote shares directly) -- seek to please their clients. In the grand scheme of things, this is no big deal. Yet powerful individuals sought to make the Agrawal study a big deal. The Wall Street Journal published two editorial pieces citing the paper, including an angrily anti-union op-ed essay by Eugene Scalia. Two prominent economists -- Steve Levitt and Steve Kaplan -- prominently featured the paper in their postings to websites at the prestigious New York Times and Harvard Law School (this blog). Rarely has an unvetted job market behavior received so much publicity, in contrast to silence on the Davis-Kim article. Why? Is it because the study is a methodological tour de force? No. Is it because the study breaks new ground in our understanding of institutional investors? No. By now, most readers will have guessed the reason. Rather than Agrawal being David to the AFL-CIO’s Goliath (Steve Levitt’s unfortunate metaphor), it’s more accurate to say that the working paper is being used as part of an attack on unions waged by libertarian partisans. Has anything changed since the days of Henry Simons? In my humble opinion, there are flaws in the paper arising from its failure to recognize institutional factors that undermine the results.

1. The paper has a timing problem. Its sampling periods are designed to capture the consequences for proxy voting of the breaking away of several unions from the AFL-CIO into a new federation called Change To Win (CTW) in 2005. The time period of the study is 2003-2006, leaving 14 months to identify a shift in voting behavior. But the Carpenters disaffiliated from the AFL-CIO in 2001, not when they joined CTW in 2005. There are no controls for this in the paper. The question then is: assuming that the paper’s methodology is sound and the Carpenters were voting like the AFL-CIO staff fund between 2003 and 2005, why did they do so? That is, assuming, as the paper does, that a union departs from its federation’s pattern after it leaves a federation, why did the Carpenters wait until 2005 — a full four years after leaving the AFL-CIO — to change their voting behavior to better represent their members’ interests (vs. those of their retirees)?

Note that when the Carpenters left the AFL-CIO in 2001, it was cause for great acrimony and condemnation by the AFL-CIO. So, we may conversely ask: why did the AFL-CIO wait until 2005 to change its voting behavior against the Carpenters? One can always rationalize these problems but another possibility is that the study’s findings might be fluky.

2. Taft-Hartley pension plans are joint union-management funds. They are not easily controlled by the union as could be the case with union staff funds. Employers get an equal vote. Why would employers support a union-initiated shift in proxy voting? In most instances, employer fiduciaries of Taft-Hartley funds resist any attempts by union fiduciaries to “politicize” a fund’s behavior.

3. Gerald Davis and E. Han Kim published a 2007 article in Journal of Financial Economics that showed non-fiduciary voting behavior by mutual funds. They found that when a mutual fund was part of a larger financial entity that supplied financial services to client corporations, the fund was more likely to vote its shares in favor of the client than other funds. Assuming that the Agrawal study is sound, it would appear that union pension funds are no different than mutual funds. Their proxy managers — who vote fund shares (the union does not usually vote shares directly) — seek to please their clients. In the grand scheme of things, this is no big deal.

Yet powerful individuals sought to make the Agrawal study a big deal. The Wall Street Journal published two editorial pieces citing the paper, including an angrily anti-union op-ed essay by Eugene Scalia. Two prominent economists — Steve Levitt and Steve Kaplan — prominently featured the paper in their postings to websites at the prestigious New York Times and Harvard Law School (this blog). Rarely has an unvetted job market behavior received so much publicity, in contrast to silence on the Davis-Kim article.

Why? Is it because the study is a methodological tour de force? No. Is it because the study breaks new ground in our understanding of institutional investors? No. By now, most readers will have guessed the reason. Rather than Agrawal being David to the AFL-CIO’s Goliath (Steve Levitt’s unfortunate metaphor), it’s more accurate to say that the working paper is being used as part of an attack on unions waged by libertarian partisans. Has anything changed since the days of Henry Simons?

]]>