Abstract
Using a panel data set of more than 600 Dutch pension funds (PFs) between 1992 and 2006, we investigate asset allocation behavior of Dutch PFs across multiple asset classes. We find that domestic investments, also known as home bias, in portfolio choices of Dutch institutional investors have fallen. We also find that the introduction of the euro, the dot-com crisis (1999–2001) and individual PF's characteristics are significant determinants of home bias. Overall, mature PFs’ portfolios are diversified internationally, whereas large PFs seem to prefer to only scale up their foreign, less-risky positions at the expense of domestic fixed-income positions. The effect of the dot-com crisis is more pronounced for domestic bonds, whereas the introduction of the euro was more important for domestic equities.
Acknowledgements
We gratefully acknowledge helpful comments from participants at the 2010 Australasian Finance and Banking Conference meetings in Sydney, the 2010 Portuguese Finance Network meetings in Azores and the 2010 Netspar International Pension Workshop meetings in Zurich. The opinions are our own and do not necessarily reflect the opinion of De Nederlandsche Bank.
Notes
1. Bikker and De Dreu Citation(2009) also investigate home bias in Dutch PFs but at the level of euro zone, where home bias is defined as percentage investments (17) in the EMU without segregating investments into asset classes. A related paper investigating Dutch PFs’ investment strategies is De Haan and Kakes Citation(2012). They find that funds buy past losers and sell past winners.
2. Weighted averages have been calculated by taking ‘total assets’ and ‘number of participants’ as weights, but the later seems less useful so we focus only on the ‘total assets’ average.
3. Dutch institutional investors hold about 25% to 30% of foreign exposure in their portfolio mix (Wilcox and Cavaglia Citation1997).
4. DNB annual reports from 1992 to 1997.
5. The foreign investment policy of Dutch PFs is regulated by DNB who applied certain restrictions on capital mobility by the PFs. These restrictions were loosened around beginning of our observation period.
6. Dutch PFs held 28.7% (41%) domestic (foreign) holdings before the dot-com crisis, which decreased to 15.8% (66.9%) after the crisis on an asset-weighed assets basis.
7. Dutch PFs hold 13% bonds in domestic markets and 29.4% equity in foreign markets on an asset-weighed assets basis.
8. Higher returns to PFs that offer DB schemes can help to reduce sponsor contributions to the PF.
9. It is defined as ratio of total assets to the liabilities of a PF. For every PF statutory, requirement by the prudential supervisor is to maintain threshold cover ratio at 105%, and if it falls below this threshold necessary actions under the guidelines of prudent supervisor should be taken to regain the threshold within next three years.
10. A very small economically significant correlation does not affect the results in our regression models; nevertheless, its statistical significance explains us the direction of co-movement of size and cover ratio.
11. For a DB scheme, benefits are defined and higher returns actually reduce the matching contribution of the sponsor, the employer, to the PF. Bikker and De Dreu Citation(2009) document that small PFs have a high tendency to hold domestic assets.
12. Though international diversification reduces risk, PFs seem to prefer to diversify their international portfolio using less-risky securities, for instance bonds, while domestically focusing more risky securities, for instance stocks, to avoid cost issues. Another reason to focus on bonds for their international portfolio diversification is mismatch risk (Bikker and De Dreu Citation2009).
13. These regulations were changed significantly at the beginning or our period and then remained unchanged, especially with regard to the cover ratio.