How much home-bias do people have and how it tilts the global market portfolio
The data and discussion of this article are largely based on the paper of Wallmeier et al., 2022, “Home bias and expected returns: A structural approach”, all data analysis is based on the cutoff date 31.12.2020.
Motivation for home-bias from an academic perspective
From the theory of capital asset pricing, the rational investor has an incentive for overweighting specific assets. And for this reason, the “home-bias” with overweighting of home country assets can be viewed from the home investor’s expectation of better risk-return characteristics, having an alpha, an edge, over the rest of the portfolio. This view about the same asset is different than what foreign investors think. The home investor’s outbidding of foreign investors leads to them needing to tilt their portfolio away from that specific country; this process must lead to negative mispricing, meaning higher prices for the same amount of risk, for assets of this country and thus lower expected returns.
The structural approach by Wallmeier et al. is based on assumption that investor’s portfolios can be subdivided into a two-fund solution with the home-country asset and a global portfolio. This in turn leads to the possible analysis of the home-bias we find in investor’s portfolios worldwide and the implied portfolio characteristics resulting from the expectations of investors by having chosen those assets. Especially, how the home-biases of all lead to a systematic tilt of the diversified part of the portfolio of global investors.
Measuring home-bias in the portfolio
In the literature there are two main definitions for home-bias that mean significantly different things. Firstly, the more classic measure HB is defined as being the allocation to the home-country in excess to the share in the market cap. weighted global portfolio (“market portfolio”). It is defined as follows:
HB = 1 – (foreign Invested capital / foreign share in market portfolio)
HB = 0 means no home-bias; HB = 1 means only home-investments
In addition, sH is the share of total home assets in the portfolio. If sH is 100%, HB is 1 and there is only home-investments and if sH = wM, being the market cap. weight of the country, HB would be 0. Both HB and sH are maybe not the most intuitive measures for looking at home allocation data. Therefore, in the following we will instead discuss the active home-bias wH.
Difference between share of total home-investment (sH) and the active home-bias (wH)
If investors from all countries just hold the market portfolio, nobody will have a home-bias. However, the measure HB does not account for the fact that already one country experiencing a home-bias of home investors will change the composition of the market portfolio and thus will impose a different global portfolio for investors without any home-bias of any other country (the tilting effect!).
The second popular measure wH accounts for the tilting and is the “active home investment”. It is defined as the home allocation in excess to an allocation wC to a common “global passive portfolio”. For this model it is assumed that each investor in each country has a two-fund portfolio consisting of a home asset and the passive global portfolio, with the weights for the active home investment and the global passive portfolio adding up to one: wH + wC = 1.
Then the share of wealth invested by home-bias into a specific country can be written as: wH*wI*WG; wI being the share of capital wealth invested in total by investors of this country and WG being the global market capital. Summing that up with the share of that country in the global passive portfolio (wC*WC) will give the market weight share (wM) of the global market cap (WG) for this country:
wH*wI*WG + wC*WC=wM*WG
To solve for wH in the latter equation, we need to first find the shares of wC of each country within the global passive portfolio.
The concept of the global passive portfolio
The global passive portfolio C is a very interesting concept because it is modeled as the portfolio that the average global investor would hold in addition to the home-asset as the optimal risk-adjusted global portfolio. As such, it should in theory have better expected risk-return characteristics than the market portfolio.
To find the passive portfolio, we need to clear the active home investment part from the market portfolio. To make this active part of the home-bias clearer, we can consider the example of an US investor. As the US market has a strong weight in the global market weighted portfolio and as such likely also in a home-bias cleared global passive portfolio, the active home-asset allocation of US investors will add up with the high portion of US stocks in a passive portfolio. A strong active home-bias of US investors will have a strong impact on the overall structure of the market portfolio, just because of the size of the US market and total capital wealth invested by US investors. On the other hand, only looking at the excess market share of the US in US investor’s portfolios would likely overestimate the home-bias.
“Clearing” the home-bias from the market portfolio.
The market weights wM and total capital WG are given from common capital market data, e.g. from MSCI. In addition, the paper utilizes data from the Coordinated Portfolio Investment Survey (CPIS) about how much total capital wealth investors from different countries invest at home and abroad. It is mentioned that the data set might have some systematic errors, e.g. with some assets systematically domiciled in countries like Ireland or Hongkong, however, the general trends should be well covered by the data.
The clearing process involves estimating the weighting relation of two countries at a time within the passive portfolio. It is assumed that investors from all other countries are holding the same ratio of assets with respect to the two countries, and the respective ratio can be calculated from the aggregated total amount of assets held by foreign investors for the two countries. This procedure is redone for all further pairs to yield all ratios. That in turn gives the weights in the passive portfolio because the weights sum up to one.
When all weights wC of the passive portfolio have been calculated, also the active home bias can be calculated from the previous equation.
How big is the home-bias in different countries?
Table 1 shows the active home investment share (wH) as well as the total share of home investment (sH) of investor’s portfolios in different countries. The two measures are almost exactly the same for most countries, but the US is a stronger outlier, confirming the common overestimation of the US home-bias, because it influences to such a strong degree the composition of the market portfolio.
European countries tend to have a lower home-bias with a declining trend in the last ten years. The US is somewhat within the global average of ca. 70%, Asian-pacific and emerging market countries tend to have a very strong home-bias with extremely high values in China and India with almost no foreign equity investments.
How starkly do the home-biases distort the market portfolio
Table 2 shows the weights of invested wealth wI, the market cap shares wM and the shares in the passive portfolio wC. In addition, I have also attempted to reduce possible influences from valuation levels by adjusting the weights in the passive portfolios with the CAPE10 valuations in 12/2020 relative to the median of the prior 10 years.
The structure of the passive portfolio is quite similar to the market portfolio, however, for many countries there are strong differences. European markets tend to have a higher weight (ca. 40 – 70% increase), with especially GBR, NLD having a much higher weight of up to over 2x the weight within the market portfolio.
The difference between passive portfolio and market portfolio is negatively related with the home bias for most countries – this means the countries with higher home bias (e.g. emerging markets) tend to have a lower weight in the passive portfolio than in the market portfolio and those with lower home-bias tend to have a higher weight. This strengthens the argument, that a strong home-bias might reduce expected returns for foreign investors, thus tilting their portfolios away from those countries.
I had the concern that unduly inflated evaluation levels measured by CAPE would have an influence on the market clearing. To account for that effect, I came up with the following idea: for a foreign investor without any home-bias just holding the passive portfolio, how would the composition of this portfolio change if the valuation levels were aligned with the 10-year median values for each country in the short term. The respective calculated weights can be found in the column wC:CAPEadj and differ only slightly.
Expected return differences between market and passive portfolios.
In the paper by Wallmeier et al., there is also an analysis of implied expected returns for a country-specific investor that is expected by holding a two-fund portfolio of the home asset and a global portfolio – either the market one or the passive one. The analysis involves estimating the “alpha” expected return share of the home asset for any country, which “motivates” the home-bias. The expected return of the second global fund can be calculated from that and is compared with the expected return of a global investor without home-bias. The results show that for all countries the differences between choosing the global passive portfolio or the market portfolio as the global portion is very small for any country. This is not very surprising, given the strong correlation between equity markets with only small differences in expected returns and the similar structure of the two global portfolios.
And why all this – conclusion
The paper and my analysis towards it, show that home-bias has a large impact on global stock markets and asset pricing. For investors in many countries increasing the global portion of the portfolio would likely reduce idiosyncratic risks attributed to their home-country. However, which allocation of global stocks they choose, as long as they are not deviating too far from the market or passive portfolio described here, should make no huge difference.
You might think, so what, nice exercise, but why the effort if it doesn’t really matter which international allocation I choose. The truth is, that any globally diversified portfolio will have very small differences in expected returns than any other, and that is a good thing for investors, as it is not important which allocation you choose if you can stick with a plan. And to draw up a plan you can stick with, I see educating yourself as very helpful.
In addition, in stock markets, the unexpected return can totally dominate the expected return, and that is why I like integrating e.g. some anticyclical elements, like equal-value weighting, to build a robust, timeless portfolio.