Asset Allocation and International Investments

The economic downturn during 2000 left many investors with burnt fingers and weary of investing in equities. Since then, there has been a search for alternative asset classes to fulfill the need to preserve returns, while not involving too high a risk. Arising from the media’s continued concern about a potential bubble in the housing market, many investors are showing an increasing interest in alternative asset classes that are not so highly correlated with equities, and provide hedging potential as part of a diversified portfolio of investments. One such innovative alternative asset class to stocks, bonds and real estate is art, which is seen increasingly as not merely items with aesthetic value, but also as attractive investments with a potential capital gain. The planned launch of a Fund of Art Funds by ABN Amro in 2005, aiming to channel money into some existing (and some yet to be launched) independent art funds, serves to highlight this point.

Despite the vast literature on optimal currency hedging, there still is considerable disagreement about how international investors should hedge their currency risk. One argument is that investors should fully hedge, since exchange-rate changes in excess of the forward discount rate average out. Therefore, hedging decreases the risk of foreign investment, but does not reduce its expected returns. In the words of Perold and Schulman (1988), currency hedging is a free lunch. However, there is a large branch of literature that does not agree with this viewpoint. As an early example, Froot (1993) argues that the free-lunch argument does not hold in the long run. If exchange rates and asset prices display mean reversion, the optimal hedging policy becomes time-varying. In particular, real exchange rates revert to their means according to the theory of purchasing power parity, and investors should maintain an unhedged foreign currency position. Therefore, for an investor with a long investment horizon, it becomes optimal not to hedge at all. Froot argues that real-exchange rates may deviate from their theoretical fair value over shorter horizons, and currency hedging in this context may become beneficial.

The international capital asset pricing model (ICAPM), based on traditional portfolio theory developed by Sharpe (1964) and Lintner (1965), suggests that, to maximize risk-adjusted returns, investors should hold a world market portfolio of risky assets. However, domestic assets are heavily weighted in investors’ portfolios even after the relaxing of capital control after 1980. For example, in 1997, 89.9 percent of US investors’ equity portfolios were domestic equities, while the size of the USA in world market capitalization was about 48.3 percent (Ahearne et al., 2004). The wide disparity between actual and recommended international equity portfolio weights constitutes the equity home bias, one of the unresolved puzzles in international finance literature.¹

Assume that there are information costs and asymmetric information in the marketplace. These conditions have been associated with segmented markets, with investors having a preferred habitat. Therefore, investors will be searching for local minima and maxima in their preferred habitat rather than a global market optimization based on equilibrium market asset pricing. Investors will have unique utility functions depending on their preferred habitat, and attempt to maximize their utility through a localized solution.

Recent studies have shown evidence of asset market price correlation and contagion. For example, Karolyi and Stulz (1996) show evidence of co-movements of US and Japanese stock returns, and find large shocks to broad-based market indices positively affect both the magnitude and persistence of return correlations. Masson (1998) identifies several sources for contagion. In the so-called “monsoonal effect”, pressures common to affected assets are the source of the contagion. In the spillover effect, changes in fundamentals affecting one set of assets cause fundamental changes in other asset groups, leading to contagion. However, some observers (see, for example, Eichengreen and Mody, 1998) argue that these economic effects cannot fully explain contagion. In fact, a change in one set of asset prices may trigger changes elsewhere, for reasons unexplained by economic fundamentals, perhaps because there are shifts in the market’s attitude towards risk, leading to the notion of pure contagion.

Research on emerging market bonds has been growing quickly (see, for example, Min, 1998).¹ This can be attributed to two main factors. First, while emerging market bond capitalization is relatively small compared to the size of the fixed-income market, it has still attracted the attention of investors. There have been times (for example, in the summer of 1997) when the average performance of the Emerging Market Bond Index is better than that of the S&P 500, and is considerably better than the US high-yield index. Among the debt instruments in the emerging markets, Brady bonds are the most important. Overall, there are about US $200bn of Brady bonds outstanding. According to the Emerging Market Traders Association, the secondary market turnover for Brady bonds represents the majority of all trading in emerging market debt instruments. Second, the sovereign bond yield spreads over the yield of similar issues from the US Treasury has become a market-based measure of sovereign credit worthiness. It has been argued that the credit worthiness of economies as measured by agencies such as Institutional Investor, Moody s and Standard and Poor’s are only ex post indicators and may not be useful measures for those who are more interested in the future performance of an economy. As a result, recent literature on sovereign credit worthiness has focused on market-determined indicators, such as the Brady bond stripped yield spread, as the purest form of market-based sovereign risk.

For decades, both practitioners and academics have been searching for the ideal asset-pricing model. The birth of the capital asset pricing model (CAPM) as a single-factor model in the 1960s revolutionized the concept of asset pricing as it enabled the quantification of the risk-return relationship. For practitioners, asset-pricing models may be important to identify whether stocks are over- or undervalued, which could influence their trading decisions. Academics, however, are particularly interested in finding an accurate asset-pricing model to facilitate the testing of the efficient market hypothesis (EMH).

With the appearance of assets reaching beyond the risk-return characteristics of conventional assets, performance measurement have also had to answer new questions. Option-like structures, dynamic strategies, leverage, the introduction of asset classes such as hedge funds, CTAs, credit funds and so on, produced novel shapes of risk-return.

On October 28, 1999, the Floor Members of the Sydney Futures Exchange (SFE) voted to transfer SFE’s floor-based futures and options contracts to an electronic trading system. On March 4, 1999, the SPI ceased floor-trading and became screen-traded. Later, on Friday, November 12, 1999 the remainder of the floor-trading ceased, and on Monday, November 15, 1999 the Sydney Futures Exchange (SFE) became a completely automated system, with all trading conducted by SYCOM IV.

Systematic risk is known to affect the market prices of traded financial assets (Stulz, 1999a, 1999b, 1999c). Indeed, the capital asset pricing model (CAPM) theory argues that each financial asset bears an undiversifiable risk known as systematic or market risk, as introduced by Sharpe (1963, 1964, 1970) and Treynor (1961) among others.¹ Such a risk can be estimated through a well-diversified portfolio so far as this portfolio presents as low as possible an idiosyncratic risk (French and Poterba, 1991). Recent literature focuses mainly on a sound assessment of the influence of systematic risk on financial assets, along with the beta coefficient in a CAPM framework. Koutmos and Knif (2002) estimate the influence of systematic risk while employing time-varying distributions (for example, conditional distributions depending on past innovations). Using market stock indices of the financial markets under consideration, they find that financial assets’ betas are stationary mean-reverting processes with an average degree of persistence equal to four days. Gençay, Selçuk and Whitcher (2003) use wavelet techniques to assess the influence of systematic risk on any asset, or equivalently to compute its beta in a CAPM model. These authors use the S&P 500 index as a systematic risk benchmark. Therefore, common practice resorts to available stock indices as proxies for a well-diversified market portfolio, and pays little attention to the sound assessment of systematic risk itself.²

The assessment of portfolio performance is fundamental for both investors and fund managers, and this applies also to Canadian hedge funds. Traditional portfolio measures present some limitations when applied to hedge funds. For example, the Sharpe ratio uses the excess reward per unit of risk as a measure of performance, with risk represented by the standard deviation.