4.4.5.2 Portfolio Firm Nationality

Survey evidence suggests that institutional investors accord more weight to voting recommendations with respect to foreign portfolio firms than to voting recommendations with respect to domestic portfolio firms (see e.g. McCahery et al. 2010). Accordingly, we expect the funds to deviate more often from voting recommendations in respect of domestic firms than from voting recommendations in respect of foreign firms. To test this hypothesis, we divide the sample into three subsamples of votes in domestic firms, votes in foreign firms (excluding the US) and votes in US firms, and we measure the degree of deviation for each subsample. The results, reported in Table 4 (Panel D) and visualized below, show that the median fund deviates from voting recommendations regarding domestic portfolio firms more often than from recommendations regarding foreign firms (4% for domestic firms vs. 1.8% for foreign firms excluding the US and 1.3% for US firms).

For the full sample, Panel E shows that the average deviation rate for domestic portfolio firms is 15.5% versus 4.5% for foreign firms (excluding the US) and 5.5% for US firms, and that the difference between the means is statistically significant (p-value<1%). The results are thus consistent with the results obtained for the median fund.

A possible explanation for these findings could be that investors tend to have more firm-specific knowledge with respect to domestic firms than with respect to foreign firms. The funds in our sample have informally stated their belief that they possess such knowledge, partly as a result of the relatively long history of their investment in local firms, their access to management and their familiarity with the local economy. Indeed, one of the funds in our sample has indicated that it aims to leverage its expertise on domestic portfolio firms by publicly and privately acting as an opinion leader among these firms' major shareholders, most of whom are foreign.

An alternative explanation could be that funds invest more resources in deciding how to vote shares in domestic firms because they are under pressure to do so. For example, local funds tend to be held accountable by local media and by management of local firms for how the funds vote their shares in these firms, even if they only hold small stakes. Funds are also subject to local codes of conduct, such as the UK Stewardship Code, that encourage them to actively monitor portfolio firms. Funds' compliance with these codes is monitored by their beneficiaries, who may be inclined to focus on how the fund has voted in domestic firms. After all, votes in domestic firms are easily observable, and beneficiaries may also have direct stakes in these firms, for example because they are employees. The funds in our sample have informally confirmed that these are all relevant factors, albeit to varying degrees. Of course, business ties could also form part of the explanation: the earlier finding that funds deviate mostly to vote FOR management also applies at a domestic level.

Still another explanation could be that the funds simply hold larger stakes in domestic portfolio firms, and that stake size is driving the result that the funds deviate more often in domestic portfolio firms. We cannot exclude this possibility, since the median fund's average stake in domestic portfolio firms is indeed larger (0.25% for the domestic sample vs. 0.07% for the entire sample), and has a higher market value (USD 6.6m for the domestic sample vs. USD 2.7m for the entire sample). The reverse causality story, that portfolio firm nationality is driving the results for stake size (reported in section 4.1), is unlikely. When we restrict the sample to domestic firms, we still see that on average, funds deviate more often in portfolio firms in which they hold a relatively large stake. This suggests that stake size matters independently of portfolio firm nationality.