Do beneficiaries' beliefs affect funds' decisions?

That the investment decisions of such a large group of investors with enormous market power might be affected by the political attitudes of their members could have wide effects on financial markets: considering the largest of these funds have assets under management of US$200bn-US$300bn, changes in a state’s political climate could lead to a great channelling of money in or out of ESG investing, affecting firms’ cost of capital and the prominence of ESG investing itself.

Results

Schopohl found that pension funds with Democratic-leaning members tilt their portfolios more strongly toward companies that perform well on ESG issues, and this effect is even more marked when the state government is predominantly affiliated to the Democratic Party. The political leaning of funds’ members, rather than potential pressures by state politicians, appears the dominant force behind the preferences: funds from states with Republican-leaning members seem to engage less strongly in ESG investing, even if the state government is predominantly composed of Democrats.

The author also found that funds dynamically adjust their ESG investment approach in response to changes in their members’ political leanings. When changing from Republican to Democrat, funds systematically increased ESG investment and vice versa.

Schopohl concludes that at this institutional level, ESG investment preferences may be driven strongly by investors’ attitudes towards the social aims of firms, as opposed to pure financial considerations of risk and return. The funds studied tended to slightly overweight underperforming stocks and underweight outperforming stocks, but these tendencies do not seem to be consistently related to their ESG preferences and/or the political climate.

The study suggests that these ESG practices are not financially detrimental to beneficiaries and might provide them with indirect value as their pensions are invested in accordance with their own political beliefs.

Methodology

Schopohl studied the funds’ full public equity holdings, where available, from 1997 to 2013. The sample consists of 31 pension funds based across 23 states, whose holdings were internally managed.

She connected funds’ portfolio holdings to the ESG performance of the companies being held, in particular, relating a company’s weight in a fund’s portfolio to the ESG performance of the company. She then analyses whether this relation changes, depending on the political leaning of fund members.

Without direct information on the political affiliations of individual fund members, Schopohl defines the political interests of pension fund members as the political leaning of the state the fund is located in (based on the percentage of the state’s votes received for either the Democrats or Republicans in the most recent presidential election). She judges this a viable proxy for the members’ political values as members of state pension funds represent a considerable share of the state’s population and even taxpayers that are not employed in the public sector have a stake in how these pension plans are managed because the responsibility for funding these plans ultimately lies with the sponsoring government.

The study also tests whether the likelihood for pension funds to engage in ESG investing depends on the degree of pressures on the fund by state politicians. Their proxy for such political pressures is based on the proportion of Democrats and Republicans in the state government. As state pension funds’ ESG investment preferences might be subject to a variety of other factors, Schopohl controls for company characteristics, fund and state characteristics and the overall market conditions in her analyses.

To check whether a change in political leaning affects the level of ESG investing, Schopohl identifies pension funds that experienced a change in the political leaning of the state population after the four elections that took place over the sample period.

A set of robustness tests control for alternative explanations of the observed associations, including industry effects, effects of indexing, and alternative specifications of the main variables.