Using a unique dataset linking investors’ cross-country GDP growth expectations to their investments into mutual funds and to the mutual funds’ cross-country allocation,we show that, while the flows into the funds are sensitive to the investors’ fund-specific aggregate expectations (computed using the fund’s portfolio shares), the funds’ allocation barely reacts to the country-level expectations.
- Speaker
- Date
- Tuesday 24 Feb 2026, 11:30 - 12:30
- Type
- Seminar
- Room
- 1.09
- Building
- Langeveld Building
(joint with Elio Bolliger and Margaret Davenport)
Using a unique dataset linking investors’ cross-country GDP growth expectations to their investments into mutual funds and to the mutual funds’ cross-country allocation,we show that, while the flows into the funds are sensitive to the investors’ fund-specific aggregate expectations (computed using the fund’s portfolio shares), the funds’ allocation barely reacts to the country-level expectations.
This gives rise to “co-ownership spillovers”, whereby negative expectations about a country in which a fund invests can adversely affect capital flows to the other countries that are part of the fund’s portfolio. Using a portfolio choice model with delegated investment, we show that these results arise naturally from a sticky portfolio friction.
These spillovers matter in the aggregate only if the portfolio shares are granular. Finally, using our data-based estimates and our model, we quantify the aggregate implications of these spillovers and find that co-ownership spillovers account for 90% of the expectation-driven capital flow reallocations.
Small countries are subject to larger co-ownership spillovers, while large countries are the biggest contributors to these spillovers.
Registration for bilateral, lunch or dinner
Lunch will be provided. If you would like to meet the guest speaker for a bilateral, join for lunch or dinner, then please register by filling in the registration form.
