RDP 9107: The Cost of Equity Capital in Australia: What can we Learn from International Equity Returns? 2. Data – Stockmarket Returns

The stock price data used in this study are the national stockmarket indices produced by Morgan Stanley Capital International (MSCI).[3] The indices include the largest companies in each of the national markets, and should be a very good proxy for the total market. The MSCI index for Australia contains around sixty stocks, with the composition changing over time with takeovers, delistings, and as market capitalisations change. They account for the reinvestment of dividends on a monthly basis, and thus are indices for total returns, not just price movements.[4]

The study uses end-month data from December 1969, when the MSCI database begins, until December 1990. Monthly returns data thus begin in January 1970. The countries included in the study are Australia, Canada, France, West Germany, Italy, Japan, the Netherlands, Sweden, Switzerland, the United Kingdom, and the United States. The MSCI series for the World Index is also used. Selection of countries was based largely on the size of the national markets, and on the availability in the MSCI database. The most notable omissions are probably South Africa and New Zealand, which are similar to Australia in their dependence on commodities, but were unavailable on the MSCI database.

Table 1 provides some summary measures of returns in different countries for the period December 1969 to December 1990.[5] To enable comparison across countries, the accumulation indices were converted into a common currency, the Special Drawing Right (SDR) of the International Monetary Fund. Since the SDR is a weighted average of five major currencies, the indices in SDR terms will be a reasonable approximation of the returns available to a representative world investor.

Table 1: Summary Statistics, SDR Returns, Dec 1969–Dec 1990
Country Standard Deviation of Monthly Returns[6]
(per cent)
Compound Annual Return[7]
(per cent)
Dec 1990
Index Value
(Dec 1969 = 100)
Australia 8.30 5.80 326.5
Canada 5.84 8.27 530.2
France 6.74 10.52 817.5
West Germany 5.74 10.37 794.8
Hong Kong 11.78 16.78 2,598.5
Italy 7.34 4.62 258.4
Japan 5.93 17.14 2,773.4
Netherlands 5.14 13.38 1,397.9
Sweden 6.26 13.55 1,442.4
Switzerland 5.22 9.69 698.1
United Kingdom 7.23 11.61 1,003.7
United States 5.01 8.22 524.9
World 4.13 9.76 707.5

The data indicate that there have been significant divergences in returns in different markets. The Japanese and Hong Kong markets have been star performers, while markets like Australia, Italy and the US have lagged. Australia's relative performance will be discussed in Section 4(a): however, it should be noted that the data period starts around the time of a metals-driven boom. Apart from Hong Kong, Australian returns show the greatest volatility. It is also apparent that the World Index has a lower variance than all individual national indices, reflecting the effects of diversification.

Because of the selective nature of the stocks included in the MSCI database, the MSCI series for Australia and the US were compared with other accumulation indices for these countries: for Australia, with the All Ordinaries and Fifty Leaders indices produced by the Australian Stock Exchange, and for the US, with the series produced by Ibbotson Associates (1990). In each case, there was extremely high correlation in the returns series, so it seems reasonable to conclude that the MSCI series are good proxies for individual national markets, and by extension, that they provide a good proxy for world equity market returns.


I am grateful to Morgan Stanley for providing the data. [3]

A description of the database reads as follows: “Indices with dividends reinvested constitute an estimate of the total return arrived at by reinvesting one twelfth of the trailing twelve month yield reported at every month end.” The dividend adjustment may be slightly imperfect, but this will not be a major factor over the long run, and will still be a fairly good short run approximation. Indeed, some studies that estimate CAPM-type models across countries use only price data and ignore the effect of dividends. [4]

Because exchange rate data were not available for Hong Kong, it could not be included in the analysis below, but the summary statistics are provided here for interest. [5]

Calculated as the standard deviation of the difference in logs of the indices. [6]

Calculated as the geometric growth rate from starting-values and end-values of the indices. [7]