The influence of political bias in state pension funds

https://doi.org/10.1016/j.jfineco.2015.08.017Get rights and content

Abstract

Using a sample of state pension funds’ equity holdings, we find evidence of not only local bias, but also bias towards politically-connected stocks. Political bias is detrimental to fund performance. State pension funds have longer holding durations of politically-connected local firms and display disposition behavior in these positions. Political bias is positively related to the percentage of politically-affiliated trustees on the board and Congressional connections. The more politically-affiliated trustees on the board, the more the fund shifts toward risky asset allocations. Overall, our results imply that political bias is likely costly to taxpayers and pension beneficiaries.

Introduction

Investment decisions are often subject to local bias. Ivkovic and Weisbenner (2005), Massa and Simonov (2006), and Seasholes and Zhu (2010) find that individual investors tend to invest more in stocks that are close to home. Coval and Moskowitz (2001), Hong, Kubik, and Stein (2005), and Baik, Kang, and Kim (2010) find evidence that local bias also transcends institutional investors’ behavior.1 Local bias also exists in common equity (Brown, Pollet, and Weisbenner, 2012) and private equity portfolios (Hochberg and Rauh, 2013) of state public pension funds. The most common explanations for local bias suggest that local investors stick close to home because they are more familiar with local firms (Brown, Pollet, and Weisbenner, 2012) or because they can exploit their informational advantage of geographically proximate firms (Baik et al., 2010, Coval and Moskowitz, 2001).2

In this paper, we examine factors that could contribute to local bias in state pension funds from a political perspective and the impact politics can have on fund performance.3 In our sample of internally managed state pension funds over the 1999–2009 period, we first analyze whether corporate political strategies influence local (state) public pension funds’ portfolio investments. Consistent with previous studies that show evidence of local bias in various settings, we find that pension funds overweight local firms by 26% relative to the market portfolio. More important, we estimate that state pension funds overweight local firms that make political contributions to local (state) politicians or have significant lobbying expenditures by 23% and 17%, respectively.

After demonstrating that pensions overweight politically active local stocks, we examine if this political bias impacts performance. We offer three non-mutually exclusive hypotheses highlighting the reasons that politically connected equity investments could influence fund performance. The information advantage hypothesis implies that fund performance should improve when the fund invests in local firms because of superior information available to fund managers about local firms. If political connections lead to better information flow, this effect should be exacerbated in the case of politically-connected firms. The familiarity hypothesis predicts that fund managers overweight local firms simply because they are more familiar with these firms. Familiarity alone, however, should not influence fund performance. Finally, the political bias hypothesis posits that if investment decisions are dominated by conflicted political motivations, then investments made under these conditions are likely to be detrimental to fund performance.

Our evidence is most consistent with the political bias hypothesis. When estimated independently, our baseline results show that local bias in general has a positive albeit insignificant impact on fund performance, whereas local political bias has a pronounced negative effect on it. For instance, a one standard deviation increase in local political bias results in about a 0.25% to 0.28% decline in quarterly equity performance.4 Given that the equity assets of state pension funds are on average $21 billion, this implies an annual decline in fund performance in the neighborhood of $225 million. When we run a horse race between local and political biases, we find that they largely offset each other. This implies that any potential benefits to fund performance from superior local information are countered by the detrimental effects of political bias.

To obtain a clearer picture of these effects, we also consider the holding durations until equity positions are liquidated. We find that politically connected local firms have significantly longer expected holding durations. For instance, depending on the specification, the hazard of a complete liquidation by state pension funds of local politically connected firms is between 0.61 and 0.76 times the hazard of a complete liquidation for local firms that are not politically active. Funds could be optimally holding politically connected firms longer and excessively trading in their nonpolitical counterparts. However, we find that funds display disposition behavior for politically active stocks. That is, they sell winners too soon and ride losses too long, which can be costly to fund beneficiaries (Odean, 1998, Frazzini, 2006). This disposition effect is not present for non–politically active stocks.

Given our evidence that state pension funds’ overweighting of politically connected stocks has negative implications, we attempt to explain this phenomenon from a fund governance perspective. A key difference between state pensions and actively managed mutual funds is that trustees of state pension funds can be active or they can be former state legislators, members of Congress, ex officio members with official positions in the state’s public sector, or appointed by the governor. By design, this governance structure creates variation in how politically infused the fund’s board of trustees likely is. We exploit such variation and examine if it is related to political bias.

We find that state funds having boards with a larger percentage of politically affiliated trustees invest more in politically connected local firms and those having boards with more financial experts invest less in such firms. Next, we consider the political atmosphere and the power of local congressional politicians, which we measure by the degree of their influence in the congressional bills’ cosponsoring network.5 Our findings suggest that the existence of more powerful politicians in a state is positively related to political bias in funds of the same state. Using the Bipartisan Campaign Reform Act (BCRA), which became effective on November 6, 2002 and banned unregulated soft money contributions to political parties, serving as an exogenous shock to firms’ political activities landscape, we find a decrease in political bias after the act and, in particular, for states with stronger ties in Congress. These results suggest that powerful politicians can impose more political pressure on state pensions to invest in politically connected local firms. Moreover, we find that pensions with a higher proportion of politically affiliated trustees invest in riskier assets.

Our baseline analyses could suffer from omitted variables bias due to the possibility that dimensions of governance quality we did not control for are correlated with funds’ tendency to engage in politically connected investments, which can also affect fund performance and funding levels. To help establish causality and address this potential endogeneity problem, we estimate a two stage least squares (2SLS) instrumental variable (IV) model using the predicted political bias measures generated from the first stage model. We find these predicted values are negative and highly significant in the second stage estimation of fund performance, confirming our baseline results. Furthermore, we exploit a plausibly exogenous shock to fund governance when the board of trustees shifts to a more politically affiliated structure. We use a difference-in-differences (DID) approach by comparing treatment funds with like controls not experiencing such turnover. Following the transition, we find that political bias measures significantly increase and fund performance deteriorates for treatment funds compared with controls.

Our study adds to the extant literature on local bias and expands on the developing literature focused on the interplay between politics and investment behavior. For example, Bonaparte, Kumar, and Page (2012) suggest that investors are more optimistic and willing to invest in riskier assets when the president belongs to the party they support and that they become more conservative and tend to invest more in local stocks when the opposition party is in power. Sinclair (2011) and Hochberg and Rauh (2013) imply that political pressure could explain local bias in private equity holdings by state pension funds, but they do not provide direct empirical evidence.6 Aabo, Pantzalis, and Park (2014) suggest that political interference with markets can induce geographic segmentation in the domestic (US) stock market and cause stock prices to exhibit a local component. Our paper extends this line of research and provides further insights into how political factors can cause local investors to make suboptimal portfolio investment decisions.

Our paper also has important policy implications. Our evidence of politically influenced investment decisions by state pension fund managers that are detrimental to fund performance suggests that at least some managers (trustees) are not upholding their fiduciary duty to act solely on behalf of the plan’s beneficiaries.

The rest of this paper proceeds as follows. Section 2 presents our hypotheses and data. Section 3 investigates the impact of local and political biases on equity pension fund performance. Section 4 takes a closer look at pension fund governance. Section 5 provides identification, and Section 6 concludes.

Section snippets

Hypothesis development and data

In this section we develop our hypothesis and describe the data. Section 2.1 reviews the literature and motivates our hypotheses. Section 2.2 describes the data and provides summary statistics.

Impact of local and political bias on pension fund equity performance

This section provides empirical results on pension fund equity performance. Section 3.1 examines the pre- and post-buy performance of local and nonlocal stocks held by pensions. Section 3.2 considers multivariate performance models. Section 3.3 examines pension fund holding durations of their stock positions. Section 3.4 focuses on the disposition behavior of pensions.

Pension fund governance

Given our findings that political bias negatively impacts fund performance, we turn to the factors that influence such bias. Policy and governance profiles of state pension funds vary considerably across the United States. Several studies show that some of these characteristics are related to local bias. For instance, Brown, Pollet, and Weisbenner (2012) and Hochberg and Rauh (2013) show that the magnitude of corruption in a state is positively related to local bias. In addition to state-level

Identification and robustness

Our baseline performance results suggest that political bias negatively impacts pension fund performance. However, our estimation on the impact of political bias on fund performance could suffer from endogeneity problems. First, some unobserved factors or omitted variables such as governance quality in state pension funds could be correlated with our political bias measures, which also affect fund performance. Second, reverse causality problems could exist in our analysis as fund performance

Conclusions

Local bias in state pension fund investments has been observed in several studies, but evidence on its impact on performance is mixed. In this study, we focus on an important aspect of local bias: bias stemming from local firms’ political activities. Many firms engage in political contributions to local politicians or engage in lobbying. We find that state pension funds overweight these politically active firms and doing so is detrimental to fund equity performance. We show that holding

References (41)

  • J. Bennett et al.

    Greener pastures and the impact of dynamic institutional preferences

    Review of Financial Studies

    (2003)
  • Y. Bonaparte et al.

    Political climate, optimism, and investment decisions

    Social Science Research Network

    (2012)
  • J. Brown et al.

    The investment behavior of state pension plans

    (2012)
  • A. Cameron et al.

    A practitioner’s guide to cluster–robust inference

    Journal of Human Resources

    (2015)
  • M. Carhart

    On persistence in mutual fund performance

    Journal of Finance

    (1997)
  • K. Chan et al.

    What determines the domestic bias and foreign bias? Evidence from mutual fund equity allocations worldwide

    Journal of Finance

    (2005)
  • H. Chen et al.

    The value of active mutual fund management: an examination of the stockholdings and trades of fund managers

    Journal of Financial and Quantitative Analysis

    (2000)
  • L. Cohen et al.

    The small world of investing: board connections and mutual fund returns

    Journal of Political Economy

    (2008)
  • J. Coval et al.

    Home bias at home: local equity preference in domestic portfolios

    Journal of Finance

    (1999)
  • J. Coval et al.

    The geography of investment: informed trading and asset prices

    Journal of Political Economy

    (2001)
  • Cited by (69)

    • The transitional impact of state pension reform

      2024, Journal of Economic Dynamics and Control
    • Geographic clustering of institutional investors

      2022, Journal of Financial Economics
      Citation Excerpt :

      First, because local institutions are better endowed with local information and geographic distance increases costs of information acquisition and monitoring, a stronger local institutional presence can be associated with a more informed investor base (see, e.g., Coval and Moskowitz, 1999; Baik et al., 2010; Chen et al., 2020). Second, institutional investors could exhibit local preference for non-information reasons because of behavioral bias and local political considerations (see Pool et al., 2012; Bradley et al., 2016). Hong et al. (2008) shows the demand of local individual investors affects firm valuation due to such local bias.

    View all citing articles on Scopus

    We thank Mikael Bergbrant, Honghui Chen, Cejergji Cici, Vladimir Gatchev, Matthew Gustafson, Delroy Hunter, Francis Laatsch, Tunde Kovacs, Mitchell Petersen, Hieu Phan, Jianping Qi, Dahlia Robinson, John Stowe, Robert Suban, Ninon Sutton, Lei Wedge, Marno Verbeek, an anonymous referee and seminar participants at the University of South Florida, the University of Southern Mississippi, Ohio University, the University of Massachusetts at Lowell, the University of Central Florida, the 2013 Eastern Finance Association annual meeting, the 2013 Financial Management Association annual meetings, and the 2013 Southern Finance Association annual meeting for useful comments and suggestions.

    View full text