Home Bias

Scope. MarketClue publishes no target geographic allocation, does not describe any portfolio as under- or over-diversified internationally, and does not tell any reader where to invest. This article explains what home bias is, how large it has been measured to be, what the diversification arithmetic says, and the reasons it may be less irrational than it looks. All figures use the illustrative teaching parameters fixed on this pillar's hub and are not forecasts.

It is one of the most robustly documented patterns in international finance and one of the least settled.

How large it was measured to be

The finding that named the phenomenon appeared in 1991. French and Poterba estimated the equity holdings of investors in three countries and found more than 98% of Japanese investors' equity portfolios held domestically, 94% for the United States and 82% for Britain — at a time when the US market was less than half of global equity market value.

Their inference is the part worth carrying. Working backwards from the holdings, they showed that portfolios weighted that heavily toward home implied investors expected their domestic market to outperform foreign markets by several hundred basis points a year — an expectation nobody had stated and few would defend if asked directly. The allocation contained a forecast its holders had not consciously made.

Home bias has fallen since and has not gone away, and it appears in every market that has been examined rather than in any particular one.

What the arithmetic says the bias costs

Combining two equity markets of equal variability, on the hub parameters:

Worked example — the benefit is real and it is shrinking for a reason. At a correlation of 0.50 — roughly the level the 1991 work was written against — combining two markets removes 2.14 percentage points of variability. At 0.85 it removes 0.61. A rise in correlation from 0.50 to 0.85 removes 71% of the benefit, and a rise to 0.95 removes 91% of it. Since markets have become more integrated over the period in which home bias has been criticised, the size of the prize has been falling at the same time as the criticism has been made. That does not make the criticism wrong. A benefit of 0.61 percentage points is still a benefit and it costs nothing to obtain. But an argument that assumes the correlations of thirty years ago overstates what international diversification now buys, and this portal states the sensitivity rather than the conclusion. (Equal weights in two markets of equal variability: combined standard deviation = 16% × √((1 + ρ) ÷ 2).)

The overweight, expressed plainly

The size of the distortion depends entirely on how large the home market is.

The same behaviour is a rounding error for an investor in a very large market and an extreme concentration for an investor in a small one — so "home bias" describes one habit with wildly different consequences depending on where the person lives.

Five reasons it may not be irrational

Liabilities are denominated at home. Someone who will spend in their own currency has a real exposure that domestic assets match. This is the strongest of the five and it is frequently omitted, because it means a globally weighted portfolio is not automatically the neutral position for a person with local obligations.

Currency risk is added by foreign holdings, which is the subject of the next article and is a genuine cost rather than an excuse.

Costs, taxes and access differ, and tax is parked to Annex A throughout this portal.

Information and familiarity. Investors know more about companies operating around them. Whether that is a real informational advantage or a feeling of one is the contested part.

Domestic indices may already carry global exposure. A large domestic company earning most of its revenue abroad supplies foreign economic exposure without foreign listing — so measuring diversification by the listing location of holdings can overstate how concentrated a portfolio actually is.