Equity Home Bias

Table of contents

List of abbreviations

List of tables and figures

1 Introduction

2 Definition of equity home bias

3 The deadweight cost of equity home bias

4 Institutional explanations of equity home bias
4.1 Overview of institutional explanations
4.2 Asymmetric information

5 Behavioural explanations of equity home bias
5.1 Optimistic expectations for the domestic market
5.2 The competence hypothesis
5.3 The influence of equity characteristics on equity home bias
5.3.1 Investing in the familiar
5.3.2 The influence of advertising and product consumption on equity choices
5.3.3 Familiarity of equity: a combination of distance, product usage, salience, language and culture
5.4 The influence of investor characteristics on equity home bias
5.4.1 Financial sophistication
5.4.2 Motives and strategies for investing
5.4.3 Cultural Aspects

6 Conclusion

References

List of abbreviations

Abbildung in dieser Leseprobe nicht enthalten

List of tables and figures

Table 1: Differences in expectation across nations

Table 2: US and Japanese investors' expectations on the Nikkei 225 and the Dow Jones Industrial average

Table 3: Balance of bullish minus bearish fund managers for four global markets

Table 4: Competence level and ranking for US and German stocks among US and German subjects

Table 5: Median measures of imputed expected returns

Table 6: Comparison of local RBOCs with their out-of-state counterparts

Table 7: The effect of advertising on number of shareholders

Figure 1: Influence of equity characteristics on familiarity

Figure 2: Influence of investor characteristics on the equity home bias

1 Introduction

Every investor faces the challenge of making the right investment decisions. Upon analysing the allocation of wealth among countries, it becomes evident that investors do not invest their financial wealth internationally, but tend to invest the majority of their wealth in domestic equity. Financial theory deems this behaviour irrational, since holding a domestic portfolio is considered to be suboptimal due to the foregone benefits of international diversification.

Assuming that the financial theory is right in this prediction, the question as to what are the causes for this irrational behaviour comes to mind and forms the focal point of this work.

One the one hand, investors may be well aware of the costs connected with holding a domestic portfolio. Market restrictions, however, do not allow investors to attain the optimal international portfolio. On the other hand, investors may be unaware of the benefits of international diversification, and instead have a preference for domestic equity and fail to perceive the domestic portfolio as suboptimal. The traditional financial theory for this behaviour provides the institutional explanations with the focus on market imperfections and the behavioural financial theory provides explanations with the focus on investor irrationality. Following this classification of both theories, this work briefly reviews institutional explanations, as many of them lack empirical evidence and concentrates mainly on the behavioural explanations, as they are the focal point of current research and find wide empirical support.

After defining equity home bias and related concepts in Chapter 2, the costs of equity home bias are discussed in Chapter 3.

In Chapter 4, institutional explanations are considered. Section 4.1 reviews briefly a number of older institutional explanations, such as direct investment barriers, transactions costs and taxes, as they do not find much empirical support. Section 4.2 explores in more detail an explanation based on information asymmetry, as it may at least partially contribute to the solution of the home bias problem.

With the emergence and acceptance of behavioural finance new explanations based on irrationality of investors were advanced and are presented in Chapter 5. Section 5.1 explores optimistic expectations about domestic markets as one of the early behavioural explanations. Section 5.2 deals with the competence hypothesis and creates a foundation for the familiarity hypothesis. Through the concept of equity familiarity, section 5.3 shows, how one can determine the characteristics of which equity is likely to be held internationally and domestically. Section 5.4 describes investor characteristics that may affect the probability of holding foreign or domestic equity.

Chapter 6 presents the conclusions derived from the preceding chapters.

2 Definition of equity home bias

Equity home bias is, generally speaking, a tendency to invest large portions of portfolio wealth in domestic equity.^[1] The corollary is a loss of the benefits of international diversification. The widely used benchmark for optimal domestic and foreign holdings is the portion of country's market capitalisation relative to the world market capitalisation.^[2] Where the actual portfolio weight of domestic equity exceeds the benchmark portfolio weight, this is considered to be home bias. Where the actual portfolio weight for a given foreign country differs from the benchmark, this is considered to be foreign bias.^[3] The home bias can also be understood as a divergence away from the optimal internationally diversified portfolio towards domestic equity.

Not all equity shares are, however, available to all investors on an equal basis. Large number of stocks may be withheld by controlling investors, who prefer to hold a either a majority or else a large portion of a company's stocks to maintain control and exercise decision rights over the company.^[4] Consequently, Dahlquist, Pinkowitz, Stulz and Williamson (2003) propose that the relevant benchmark portfolio should be based on a world float portfolio, i.e. a portfolio based on the equity available to all investors, not withheld by a certain groups of investors. Such an approach would reduce the observed home bias, but not eliminate it altogether.^[5]

The tendency for an investor to prefer equity from his own geographical vicinity is known as local bias.^[6] Even within national markets investors prefer to hold equity in companies geographically close to them rather drawn from the national market portfolio as a whole.

The home bias is well documented for a large number of countries and historically a long period.^[7] While the degree of equity home bias has steadily fallen over time, nonetheless it is still present.^[8] The home bias is evident from the micro data of equity held directly by individual and institutional investors, mutual funds and an aggregated level of countries' holdings.^[9] In view of this, the equity home bias is an enduring and stable phenomenon that is not predicted by the standard financial theory and therefore it is often called a home bias puzzle. The reason, why the standard theory considers the home bias irrational, is discussed in the next chapter.

3 The deadweight cost of equity home bias

It has been generally known for a long time that holding an ill diversified portfolio is costly. Common sense and the old adage teach us not to put all eggs into one basket. This is sadly illustrated by the recent earthquake followed by tsunami and nuclear crisis in Japan that has not only claimed many human lives and caused untold distress, but also by the immediate impact the utter destruction has had on the country's economic prospects. Consequently, the home biased Japanese investor has lost a substantial portion of wealth, as a result of the Nikkei 225 index falling from a level of 10,500 to 9,500 points, a loss of almost 10%. However, not only extreme situations like this give reasons for international diversification. The financial theory provides analytical, mathematical models that enable one to compute the benefits of international diversification, and respectively in case of lack of diversification, its deadweight costs.

Within the standard financial theory, rational investors base their financial decisions exclusively on risk and expected returns in order to reach an optimal portfolio by way of mean-variance optimisation, as developed by Markowitz (1952). The crux of this method lies in combining individual assets into a portfolio by considering not only the expected return and its risk measured by return variance of the single assets, but also the interdependencies between them measured by covariances.^[10] Since the returns of assets are not perfectly correlated, the risks of individual assets combined in a portfolio partially cancel each other out.^[11] For any given individual assets, it is possible to construct portfolios by appropriate asset weighting that can be characterised by an overall portfolio expected return and variance. From the quantity of portfolios attainable, a rational investor considers only the efficient portfolios – i.e. the ones with the highest expected return for a given variance or, alternatively, the ones with the lowest variance for a given expected return.^[12]

Home biased investors forego the opportunity to implicate foreign equity into the mean-variance optimisation and thus forfeit a potential improvement in portfolio performance. For a US investor in the period 1959 – 1966, Grubel (1968) estimates the benefit of international diversification as an increase in annual return from 7.5% to 12.6%, with an identical standard deviation.^[13] Similarly, for the period from 1951‑1967, Levy and Sarnat (1970) find a significant portfolio improvement with higher returns and lower risks.^[14] According to Lewis (1999), the deadweight cost of home bias for a US investor calculated on the basis of a simple, crude two-country model based on historical data from 1970 - 1996 approximately amounted to 0.6% reduction in annual returns for the given variance of portfolio actually held.^[15] For aGerman investor, the benefit of holding an international portfolio was calculated as a 5.08% return increase for the given standard deviation of the benchmark domestic portfolio or as 8.01% standard deviation decrease for the given return.^[16] Since historic data is a bad predictor of future returns and volatilities, and the mean-variance optimisation is highly sensitive to data inputs and often provides results with null portfolio weights for many countries, the data-based mean-variance approach has substantial drawbacks.^[17]

By incorporating the mean-variance approach into a capital market equilibrium analysis, the capital asset pricing model (CAPM) is constructed. Extending the scope of CAPM into a global perspective, Solnik (1974) derives the international investment rule to weight the international assets according to their market capitalisation relative to the value of the world portfolio.^[18] The immense advantage of this model based approach lies in its simplicity and independence of historical data.

Using this method to construct the benchmark portfolio, Black and Litterman (1992) show that for a given risk level, the return on a world equity portfolio is 0.76% [0.84] higher than return on a domestic US portfolio with no currency hedging [with 80% currency hedging].^[19] Analysing the portfolios of G7 countries and the world portfolio by taking the currency exchange rate risks into consideration, vonNitzsch and Stotz (2006) observe that substantial deadweight costs exist as a result of home bias. Measured by the Sharpe ratio, the average domestic portfolio across all G7 countries is 0.12 worse off than the world portfolio with an optimal hedge of the currency risks.^[20] Measured by the return, the average domestic portfolio is 0.6% [1.2% and 1.8%] worse off than the world portfolio at a given volatility of 5% [10% and 15%] with an optimal hedge of currency risks.^[21]

The benefits of international diversification for different countries vary with the capitalisation of domestic markets.^[22] Solnik (1974b) shows that investors can hedge a bigger portion of risks on larger financial markets, as the multitude of industries and assets provides greater diversification potential. In the case of a small number of assets, the benefit of diversification for each additional asset is substantial. However, with an increasing number of assets the diversification potential decreases and converges towards null.^[23] For domestic portfolios with approximately 20 or more assets, the domestic risk is completely hedged.^[24] In order to reduce the risk level beyond the domestic threshold even further, international diversification is therefore necessary.^[25]

Within the behavioural financial theory, in contrast to rational investors, behavioural investors are not exclusively concerned about expected returns and variances.^[26] Irrational investors may derive expressive utility from the mere holding of a certain portfolio.^[27] One possible explanation is that a home biased portfolio may induce positive emotions (e.g. because of patriotism) and thus offset its economic deadweight costs of portfolio underperformance.^[28] Alternatively, the benefits of international diversification remain unknown or unappreciated, leaving the investor with his idiosyncratic investment strategies to make financial decisions.

From the normative point of view, both the traditional and the behavioural financial theories concur that, since international diversification is desirable, the equity home bias comes at a cost. However, the explanations of the equity home bias vary considerably depending on the underlying theory. In broadest terms, the traditional theory focuses on market imperfections, whereas the behavioural theory enlarges its scope also on investor imperfections. The traditional and behavioural explanations are the main topic of this work and are explained in detail in the following chapters.

4 Institutional explanations of equity home bias

4.1 Overview of institutional explanations

Early explanations of the equity home bias within the realm of traditional theory concern market imperfections. Following World War II numerous countries imposed various barriers to international financial trade.^[29] As foreigners were restricted to owning domestic equity, currencies were not easily convertible and additional taxes on foreign equity were introduced, the international diversification was indeed complicated.^[30] Black (1974) and Stulz (1981) incorporate into the international capital market equilibrium model barriers to international investment in the form of a tax on foreign equity holdings.^[31] Although international barriers have since been substantially reduced and the markets are today largely integrated, the equity home bias still pertains.^[32] Thus, explicit barriers to international capital flows cannot explain the equity home bias.

Potentially, investors may hold domestic equity to hedge specific domestic risks. Domestic equity may provide a hedge against domestic inflation risk and against wealth not traded in capital markets (especially human capital).^[33] International capital market models assume that all investors assess the real equity returns uniformly, since the purchasing power parity (PPP) equalises the inflation rates through currency exchange rate adjustments. However, as the PPP hypothesis does not seem to hold empirically, investors view real returns in fact differently and may be willing to hold different portfolios.^[34] Adler and Dumas (1983) derive an optimal portfolio as a combination of a portfolio common for all investors and of a national portfolio hedging against inflation risk.^[35] Based on this relation, Cooper and Kaplanis (1994) test empirically whether the hedging of domestic inflation risk through domestic equity explains the equity home bias and reject this hypothesis.^[36]

Human capital (i.e. labour) constitutes a large proportion (60%) of economic wealth.^[37] As human capital is not tradable on financial markets, rational investors may deviate from the optimal portfolio predicted by capital market theory to adjust for this shortcoming. Investors have to decide whether to hedge the risk of labour income loss with domestic or foreign equity. When returns on human capital are less correlated with domestic equity returns than with foreign equity returns, domestic equity provides a better risk hedge than foreign equity.^[38]

In a simple analysis, Bottazzi, Pesenti and van Wincoop (1996) find negative correlations for twelve out of sixteen countries.^[39] Applying various approaches to different data sets, they find evidence for the equity home bias resulting from hedging the human capital return risk.^[40] The results of Baxter and Jermann (1996), however, are in stark contrast. They find strong positive correlations between human capital and capital returns for four countries based on a relatively long data period from 1960 - 1993.^[41] This implies that the internationally diversified portfolio accounting for the human capital return hedge comprises a substantial short position in domestic equity and long positions in foreign equity^[42] Baxter and Jermann argue that this result holds for all countries.^[43]

As the imputed correlations are problematic because of a dependency upon statistical method employed, Glassman and Riddick (2001) use an alternative method to test the relevance of human capital on the equity home bias. They include an omitted asset in the optimisation model of Adler and Dumas (1983) so that the actual portfolio weights of six countries match the model prediction.^[44] However, the imputed characteristics of the omitted asset are not consistent with the expected characteristic of human capital.^[45] Consequently, hedging against risks arising from labour income does not seem to explain the home bias.

Investing in the equity of domestic multinational corporations may provide the benefits of international diversification even without investing in foreign equity. The idea is that the returns of corporations having business operations in different countries are not solely exposed to domestic risk but provide a natural risk hedge. However, Jacquillat and Solnik (1978) find that the returns of multinational companies correlate strongly with the domestic market indices and conclude that multinationals do not provide the same benefits of international diversification as foreign equity.^[46] In a more recent study, Rowland and Tesar (2004) compare the performance of domestic portfolios containing multinational companies first with purely domestic portfolios excluding multinational companies and second with portfolios containing domestic and foreign equity. The inclusion of multinationals in purely domestic portfolios [domestic portfolios and foreign indices] statistically (at 0.02% and 0.06% [2%, 0.03% and 0.1%] significance levels) shifts the efficient portfolio frontier in two [three] out of seven G7 countries, indicating potential diversification gains from multinationals in specific countries.^[47] Despite this finding, diversification through multinationals seems an unlikely explanation for the equity home bias since the performance of portfolios containing equity of domestic multinationals improves after the inclusion of foreign market indices.^[48]

The additional variable costs of cross-border trading may have a prohibitive influence on foreign equity investment and lead to the preference for domestic equity.^[49] When these costs are substantial, investors are not expected to trade heavily abroad but rather to follow the buy and hold strategy. Tesar and Werner (1995) compare the turnover rates on international equity with the turnover rates on domestic equity. In 1989, the turnover rate (defined as the ratio of the sum of purchases and sales over the market capitalisation) varied between 0.61 and 1.74 for domestic equity, with an average of 1.09.^[50] The turnover rate in foreign equity held by domestic residents was 7.7 for Canadian investors and 2.5 for US investors.^[51] The turnover rate in domestic equity held by foreign residents was 2.2 for Canadian investors, 1.37 for UK investors and 1.6 US investors.^[52] Both perspectives provide evidence that investors trade more heavily on foreign than on domestic equity. This finding rejects the hypothesis that variable costs on foreign equity transactions have a prohibitive influence on foreign equity investment leading to home bias.^[53] Warnock (2002) shows that, due to an estimation error in the official data, the results of turnover rates in foreign equity held by domestic residents are flawed and recomputes the values based on updated data as 0.83 for Canadian and 1.18 for US investors, being very similar to turnover rates in domestic equity.^[54] However, the computed rates for US investors in 1997 (based on the best available data) reveal that investors turn over their foreign holdings more often than their domestic ones.^[55] Thus, the notion that transaction costs on foreign equity do not cause the home bias seems to be valid. Further evidence is given by Domowitz, Glen and Madhavan (2001), who show that after consideration of transaction costs, North-American investors should transfer a substantial portion of wealth invested in domestic equity into countries with lower transaction costs.^[56]

Explicit barriers to international equity investments, hedging specific domestic risks, diversification through multinationals and higher variable costs on foreign trades appear not to account for the equity home bias. Implicit barriers, however, in form of informational disadvantages of domestic investors towards foreign markets may largely contribute to the explanation of the puzzle and are therefore discussed in more detail in the following chapter.

4.2 Asymmetric information

The most prominent traditional explanation of the equity home bias is the informational advantage of domestic investors over foreign investors. Domestic investors may have more precise expectations about the future performance of home companies than that of foreign companies because of better access to generally more up-to-date and concise information, or because of a better understanding of accounting methods allowing them to gauge a clearer picture of company's condition and its future prospects.

Realising that various explanations fail to square the equity home bias with traditional theory,^[57] in an early paper Gehrig (1993) proposes a theoretical model with heterogeneously informed investors. This leads to differing return expectations among domestic and foreign investors due to the higher precision of domestic signals and subsequently to the home bias. Focusing on international investment flows and asymmetric distribution of information, Brennan and Cao (1997) construct a model that expects the sales of foreign equity to be linearly dependable upon the returns of domestic and foreign equity markets. They find, however, only little empirical evidence for the proposed model.^[58]

Whether domestic investors actually possess an informational advantage over foreign investors is empirically controversial. Informational advantage is difficult to measure. Therefore the realised portfolio returns are in many cases treated as a proxy, since the informational advantage should guarantee a higher portfolio performance.

Although the study of Coval and Moskowitz (2001) is limited to the US market, it helps contribute to the explanation of the equity home bias. For the period 1975‑1994 Coval and Moskowitz analyse returns on equity portfolios of US mutual fund managers and find higher raw and risk adjusted returns on local than on distant equity relative to the fund manager's headquarters. Local equity, defined as being situated within 100 km of the fund headquarters, yielded an annual raw [risk adjusted] return surplus of 2.67% [1.18%] over distant equity.^[59] Moreover, the risk adjusted return of local equity managers choose to invest in outperforms the risk adjusted return of local equity that is not included into the portfolio by 3.01%.^[60] These results suggest that mutual fund managers are better informed about local stocks and use this advantage in their favour.

Analysing the portfolios of individual US investors, Zhu (2002) does not find statistically different raw and risk adjusted returns on portfolios containing local and on portfolios containing distant equity.^[61] Interestingly, using the same data set as Zhu(2002), both Ivkovic and Weisbenner (2005), and Seasholes and Zhu (2010) analyse the potential informational advantage of individual investors and come to opposite conclusions.^[62] Whereas the former argue that on average investors earn on local stocks a premium of 3.2% over distant stocks,^[63] the latter find no statistically or economically significant difference.^[64]

Another means to assess the informational advantage hypothesis is by analysing the accuracy of advisor forecasts. Bae, Stulz and Tan (2008) find on a sample of 32 countries that domestic analysts make more precise earnings forecasts than foreign analysts.^[65] The average local analyst accuracy advantage relative to foreign analysts is $0.023 per average equity share worth of $15.^[66] The informational advantage of local analysts seems to be larger in countries with higher barriers to information dissemination.^[67] Concentrating solely on the US market, Malloy (2005) finds similar results. Local analysts' accuracy advantage is on average $0.025 per share, but in remote areas it rises to $0.054 for all stocks and to $0.141 for small stocks.^[68]

The informational asymmetry can be reduced in two ways. Either investors intensify their research on equities or companies alleviate the information gathering process by cross listing of their equity on foreign exchanges (thus accepting foreign investor protection regulations), by issuing their financial reports in multiple languages and according to international accounting standards.^[69]

From the perspective of US investors, Ahearne, Griever and Warnock (2004) compare holdings of foreign equity listed on US exchanges with holdings of foreign equity listed exclusively abroad.^[70] The listing of foreign equity on US markets has a statistically (at 1% significance level) and economically significant effect on the US bias against foreign countries.^[71] If all foreign companies were listed in the US and maintained their weight in the world portfolio, US investors would increase their foreign holdings from 10% to 25% and the home bias would be reduced from 80% to 50%.^[72]

Covrig, Defond and Hung (2007) explore the level of foreign equity ownership of companies using the International Accounting Standards (IAS).^[73] Descriptive statistics reveals that companies using international accounting standards are held more intensely by foreign mutual funds than by domestic mutual funds.^[74] On average, equity of companies using IAS [local accounting standards] amounts to 6% [2%] of foreign mutual funds holdings and to 3% [4%] of domestic mutual funds holdings.^[75] These differences are statistically significant (at 1% significance level). Although the number of companies using IAS is higher for domestic and foreign mutual funds than the number of companies using local accounting standards, foreign funds hold a substantially higher number of IAS reporting companies than domestic funds do.^[76] These results are supported by a multivariate regression controlling for a variety of variables.^[77] Similarly, Bradshaw, Bushee and Miller (2004) find that US institutional investors prefer the foreign equity of companies using accounting methods that conform to the US Generally Accepted Accounting Principles (US GAAP). Although the usage of internationally recognised accounting methods may reduce informational asymmetries, its effect on the equity home bias is presumably very small.^[78]

Issuing financial reports in multiple languages may also attract more foreign investors. Grinblatt and Keloharju (2001) demonstrate that Finnish companies reporting in Swedish have a large investor base from Sweden and state that US companies reporting in Spanish or Japanese may improve their investor breadth accordingly.^[79] Jeanjean, Stolowy and Erkens (2010) analyse the effect of introducing reports in English in addition to the mother language on the reduction of information asymmetry, the level of analysts following and number of foreign investors among 166companies from 16 European countries. After adopting reports in the English language, companies reduce the information asymmetry, enjoy a 10.8% higher analysts following and attract more foreign investors.^[80]

The potential informational advantage of domestic investors arising from their natural proximity to the market can be reduced through diverse measures taken by companies to alleviate foreign investors of the information gathering processes. However, the character of the information has to be taken into account. Under the traditional framework only valuation-relevant information can assure an advantage and subsequently better investment performance. The introduction of reporting according to international accounting standards or in multiple languages diminishes the asymmetry of valuation-relevant information, but the effect on foreign investment is rather modest.

Simple awareness of a particular equity would not induce a rational investor to consider it as a preferable investment opportunity. Despite this traditional prediction, investors seem to hold equity with which they are familiar, but about which they do not possess any informational advantage. Kang and Stulz (1997) analyse Japanese foreign equity holdings and find that foreigners prefer to hold the equity of large, export-oriented Japanese companies, although the chosen equity portfolio underperforms the market index.^[81] One explanation is that foreign investors are better informed about large, better known Japanese companies, relative to the smaller ones. Alternatively, they do not possess any advantageous information at all and choose the equity they know. This, from the perspective of traditional theory, irrational behaviour will be discussed in depth in the following chapter.

[...]

^[1] Excellent literature reviews of home bias are provided by Lewis (1999), Karolyi and Stulz (2003), and Sercu and Vanpee (2007).

^[2] Cf. e.g. Lütje and Menkhoff (2007), p. 24. The data based approach is another means to estimate the benchmark portfolio weight. For details cf. the following chapter The deadweight cost of equity home bias.

^[3] Whereas there is no evidence for the underweighting of domestic equity (negative home bias), countries do over- and underweight the portfolio weights of other countries. E.g., Expect for a home bias, Germany has also a European bias and underweights the equity of US and Japan. Cf.Oeheler et al. (2008), pp. 154-156.

^[4] Cf. Dahlquist et al. (2003), pp. 88-89.

^[5] Cf. ibid., pp. 98-100.

^[6] Cf. Coval and Moskowitz (1999), and Grinblatt and Keloharju (2001).

^[7] For example, Chan et al. (2005) find a home bias in the mutual funds from 26 countries. Cf. Chan et al. (2005), pp. 1503-1509. Cf. also Sercu and Vanpee (2007), p. 8, Table 1, for the home bias of 42 countries in 2005.

^[8] For the historical development of home bias in Germany cf. Oehler et al. (2008), p. 152, in USA cfKyrychenko and Shum (2009), pp. 2-3, Figures 1 and 2. Cf. also Sercu and Vanpee (2007), p. 7, Figure 1 for nine developed countries.

^[9] Cf. e.g. Grinblatt and Kelohaju (2001), Karlsson and Norden (2007), Oehler et al. (2008), and French and Poterba (1991).

^[10] Cf. Markowitz (1952), pp. 79-81.

^[11] The lower the correlation, the higher the gains from diversification. Cf. Lewis (1999), p. 577.

^[12] This relationship holds only for efficient portfolios. For a given expected return [variance] lower [higher] than the minimal expected return [the maximal variance] of an efficient portfolio, all portfolios are inefficient because of domination. Cf. Markowitz (1952), p. 82.

^[13] A domestic portfolio is compared to an international portfolio of 11 (8) countries. Cf. Grubel (1968), pp. 1036-1038.

^[14] Cf. Levy and Sarnat (1970), pp. 672-673.

^[15] Cf. Lewis (1999), pp. 572 and 573, Figure 1. The calculation is based on the S&P 500 index, as an approximation for the domestic asset and on EAFE index, as an approximation for the foreign asset. Assuming that US investors held 8% of financial wealth in foreign equity during the observed period, this meant that they held a portfolio with an annual mean return of 11.2% and standard deviation of 14.5. Increasing the portfolio weight on foreign equity would result in portfolio with the same standard deviations but a mean return of 11.8%.

^[16] Cf. Gerke et al. (2005), pp. 56-57. Interestingly, the optimal strategy for German investors in this case would be not to hold any German equity at all. The benefits of international diversification for German investors are further analysed by Maurer and Mertz (2000), and for German and Hungarian investors by Bugar and Maurer (2002). Wilcox and Cavaglia (1997) analyse the benefits of international diversification for Dutch institutional investors, and Haavisto and Hansson (1992) for Nordic investors.

^[17] Cf. Black and Litterman (1992), p. 28 and Jeske (2001), p. 34. Ideally, an investor should construct efficient portfolios by considering all available assets, provided the knowledge of expected returns and volatility is available. Results also suggest large short positions that are not attainable in practice. The data based approach refers to the mean-variance optimisation and the model based approach to the International CAPM approach. Cf. Baele et al. (2007), p. 607.

^[18] Cf. Solnik (1974a), p. 521. The model is based on several traditional assumptions, e.g. market perfection and efficiency, investor rationality. Cf. ibid., pp. 502-503.

^[19] The currency hedging is even more profitable for bonds. Cf. Black and Litterman (1992), pp.39‑40, Table 18.

^[20] Cf. von Nitzsch and Stolz (2006), p. 110, Table 2. In the case of no currency risk hedging the loss amounts to 0.064 and in the case of complete hedging to 0.118.

^[21] Cf. ibid., p. 111, Table 3. For complete currency risk hedging the results are very similar, for no hedging the return decrease is approximately one half of the returns with optimal hedging.

^[22] Cf. ibid., p. 112.

^[23] Cf. Solnik (1974b), pp. 48-52.

^[24] Cf. ibid., p. 49 and pp. 50-51, Tables 1-8.

^[25] Cf. ibid., pp. 49-52 and Tables 9-11.

^[26] The representative economic agent of the traditional theory is an ideal, rational investor. The behavioural theory assumes that economic agents make systematically irrational choices based on wrong reasoning or emotions and are thus cognitively and emotionally biased as opposed to the rational ideal. Statman (2005) describes the behaviourally biased investor as a normal investor and argues against the traditional notion of investor rationality.

^[27] Mean and variance capture the utilitarian benefits. E.g. status, patriotism, social responsibility and fairness determine the expressive benefits. Rational investors consider only the utilitarian benefits, irrational (behavioural, normal, bounded rational) investors consider both utilitarian and expressive benefits. Cf. Statman (2005).

^[28] Cf. ibid., p. 157.

^[29] Karolyi and Stulz (2003), pp. 997-998.

^[30] Cf. ibid. and Black (1974), p. 338. Financial theory refers to international markets with barriers as segmented markets and to markets without barriers as integrated markets.

^[31] The tax represents the aggregated costs of foreign investment: the risk of expropriation, direct controls on capital trades, reserve requirements, restrictions on holding equity by foreigners and unfamiliarity with foreign markets. Cf. ibid.

^[32] Cf. Karolyi and Stulz (2003), p. 999.

^[33] Cf. Lewis (1999), p. 579.

^[34] Cf. Adler and Dumas (1983), pp. 929-933.

^[35] Cf. ibid., pp. 939-944 and Cooper and Kaplanis (1994), pp. 48-49. The optimal portfolio x for a national investor l is, where α denotes the risk tolerance, Ω is the covariance matrix of the nominal return rates, μ is the expected index return, r is the risk free rate and w the vector of covariances of the risky asset returns and the rate of inflation. The first part of the equation is identical for all investors (universal logarithmic portfolio). The second part (representing the hedge portfolio) differs for every national investor because of differences in covariance of risky assets and of investors' inflation rates.

^[36] The statistical test is derived from the relation, where e denotes the proportion of country's l market capitalisation to the world market capitalisation and v the proportion of country's wealth to the world wealth. The data considers eight countries and spans the period 1978 - 1987. The results reject the hypothesis for all eight countries. Cf. Cooper and Kaplanis (1994), pp. 49-52 and the preceding footnote.

^[37] Cf. Baxter and Jermann (1997), p. 170. and Bottazzi et al. (1996), p.220.

^[38] Cf. Bottazzi et al. (1996), p. 221.

^[39] Cf. ibid., p. 223 and p. 224, Table 1.

^[40] However, the results vary across countries, the measure and the method used. E.g., the first data set displays an average home bias of 31% with the first measure and an average foreign bias of 26% with the second measure. Cf. ibid., pp. 242-244 and p. 241, Table 3. The second data set displays an average home bias of 3% [24%] for the first method and 7% [35%] for the second method. Cf.ibid., p. 247 and p. 248, Table 5.

^[41] Cf. Baxter and Jermann (1997), p. 174, Table 1 and p. 173, Figure 1.

^[42] Cf. ibid., p. 178, Table 2.

^[43] Cf. ibid., p. 179.

^[44] Cf. Glassman and Riddick (2001), p. 44, p. 39, Table 1, p. 37 and footnote 35.

^[45] The correlations are predominantly positive, yet in many cases higher for foreign than for domestic equity. Cf. Glassman and Riddick (2001), p. 45-46, Table 4.

^[46] The market betas of multinationals are in most of the cases very near to one (perfect correlation with a market index). Cf. Jacquillat and Solnik (1978), p. 10, Table 2.

^[47] Cf. Rowland and Tesar (2004), pp. 805-806, Table 5.1 and 5.2.

^[48] Cf. ibid., pp. 806-812.

^[49] The cost on cross-border trading can be divided into variable and fix. The language and cultural barriers, institutional and regulatory differences, informational costs about foreign markets build the bulk of fixed costs and are discussed in more detail in the next chapter on informational advantage. Stamp taxes, commissions and higher spreads are variable costs. Cf. Tesar and Werner (1995), p. 479.

^[50] Cf. ibid., p. 480 and p. 481, Table 5, Panel A.

^[51] Cf. ibid., Panel B.

^[52] Cf. ibid., Panel C.

^[53] Cf. ibid., p. 485.

^[54] The main problem of the data is the inaccuracy of data updates and the incorrectly attributed source of the traded assets. E.g. US equity traded in UK appears in the statistic as UK equity. Cf.Warnock (2002), pp. 799-800 and Tesar and Werner (1995) p. 474, who realise the problem.

^[55] Cf. Warnock (2002), pp. 801-802, Table 2.

^[56] To the low cost countries belong European and Latin American countries. Cf. Domowitz (2001), p.243, Table 7. Further, Ahearne et al. (2004) do not find an direct impact of transaction costs on foreign bias of US investors. However, equity from countries with high transaction costs listed on US exchanges reduces the US foreign bias, indicating that foreign listing reduces except for indirect infotmational barriers (cf. next chapter) also the variable transaction costs. Cf. Ahearne et al. (2004), pp 329-330 and p. 328, Table 3A.

^[57] Cf. previous chapter on institutional explanations.

^[58] Cf. Brennan and Cao (1997), pp. 1869, 1872 and p. 1876.

^[59] Cf. Coval and Moskowitz (2001), pp. 817-820, Table 1.

^[60] Cf. ibid., p. 820 and p. 817, Table 1.

^[61] The raw returns on portfolios with the highest and the lowest local bias (all portfolios are divided into quintiles) are not statistically different. Using the Fama-French three factors regression, the risk adjusted returns do not differ significantly for the lowest and highest local bias quintile. Cf.Zhu (2002), pp. 16-17 and p. 41, Table 6.

^[62] The differences are based on different econometric approaches and data inclusion. Cf. Ivkovic and Weisbenner (2005), pp. 268-269 and Seasholes and Zhu (2010), pp. 1989-1991.

^[63] Cf. Ivkovic and Weisbenner (2005), p. 285, Table 5.

^[64] Cf. Seasholes and Zhu (2010), p. 1997-2000.

^[65] They believe that domestic analysts exploit an informational advantage for domestic equity. However, they do not reject an alternative, demand driven explanation that the informational asymmetry could be eliminated, when investors demand more foreign equity and analysts' advice on it, leading to greater research effort. Cf. Bae et al. (2008), pp. 600-605.

^[66] Cf. ibid., p. 587 and 590, Table 3. However, this advantage does not appear to be economically significant.

^[67] Cf. ibid., p. 597-600, Table 7. These are countries with lower disclosure levels, with higher earnings manipulation, less efficiently functioning markets and less active institutional investors.

^[68] Cf. Malloy (2005), p. 731 and p. 735. Similar to the reasoning in footnote 65, the advantage could be eliminated through additional effort of remote advisors. Most probably, the advantage of local advisors arises naturally from social contacts with local business leaders and the possibility to asses the business activity on the spot. Cf. ibid., p. 738.

^[69] Cf. Ahearne et al. (2004), p. 316.

^[70] Foreign investors can list their equity directly, as American Depositary Receipt (ADR) or by issuing bonds in the US. Cf. ibid., p. 323.

^[71] Cf. ibid., pp. 327-329, Table 3A.

^[72] Cf. ibid., p. 329.

^[73] The International Accounting Standards (IAS) and the successive International Financial Reporting Standards (IFRS) are hereby considered equally. Cf. Covrig et al. (2007), p. 42.

^[74] Cf. ibid., pp. 52-55.

^[75] Cf. ibid., especially Table 2.

^[76] Cf. ibid.

^[77] Cf. ibid., pp. 56-60.

^[78] Cf. Beneish and Yohn (2008), p. 441.

^[79] Cf. Grinblatt and Keloharju (2001), p. 1071 and also the following chapter Investing in the familiar.

^[80] Cf. Jeanjean et al. (2010), pp. 21-23 and p. 36, Table 4.

^[81] The correlation between foreign holdings and company size (measured as the market value) is very high (the average Spearman rank correlation is 0.455). Cf. Kang and Stulz (1997), p. 18. For the effects of size and export intensity cf. ibid., pp. 19-20, Table 5. In the period analysed, the portfolio actually held had the sema returns as the market portfolio, but a higher standard deviation. Cf.ibid., pp. 23-25.