This thesis investigates the effect of allocation weight constraints, estimation error and taxes on the potential benefits from international diversification. The empirical results presented in this thesis address gaps in the literature concerning the statistical significance of the potential gains from international diversification and whether these potential benefits can be realized ex-ante. Weight constraints on market allocations used with static in-sample optimized portfolios are found to reduce the potential benefits of international diversification to statistically insignificant levels versus the domestic portfolio for the majority of investors in developed markets for the 1993 to 2014 investment period. The naive global market capitalization weighted (1/M) portfolio, the most strongly constrained portfolio, provides statistically significant positive gains for only 2 of the 34 markets studied. Estimation error reduction strategies designed to improve ex-ante optimization are not found to outperform the naive domestic market portfolio or the naive 1/M port- folio for investors in the majority of countries measured. And the unequal taxation of overseas equity income is reported to reduce the potential benefits of international diversification and increase the efficiency of a home bias for New Zealand investors indirectly investing in equities held in a portfolio investment entity (PIE). These empirical findings suggest that a home bias may exist in part because allocation constraints, estimation error and taxes on overseas equity investments can make overweighting the domestic market a statistically efficient investment decision.