In this paper, we examine the nature of risk and return over the past five decades in Canada (1961 to 2010) and also shed light on some conventional folktales about investment duration. The motivation for the paper is the recent financial crisis, which has led many to question the wisdom of investment in the stock market. We believe that a historical analysis of risk and returns from the stock market and a risk free security (T-bills) may provide the necessary evidence on the investment experience in Canada.
It is well known that investment in stock markets is associated with volatility, for which investors require compensation. Merton (1980) said it best when he noted that “to induce risk-averse investors to bear more risk, the expected return must be higher” (pp. 328). In fact, modern portfolio theory is based on maximizing expected returns for a given level of risk. However, the empirical evidence on the relationship between risk and return has been mixed at best (see Guo and Whitelaw (2006) for a synopsis). In this paper, we document the risk-return tradeoff in the Canadian stock market over the last fifty years.
The data used in this study is the monthly price returns and total returns (price plus dividends) data for the TSX 300 Index from 1961 to 2010, downloaded from the CFMRC database and supplemented with 2010 monthly total return data from the TSX e-Review. The T-Bill rates are obtained from Bank of Canada.
Wealth Creation
We first analyse the wealth implications of a dollar invested in T-Bills (zero risk) and in the stock market (higher risk). We analyse both price returns and total returns to provide insights into the relative importance of capital gains and dividends. As can be seen in Figure 1, the importance of reinvestment of dividends and of compounding is striking. For an investor reinvesting dividends back into the portfolio, a dollar invested in January 1, 1961 would have grown to $120 in 2010 on a total return basis, and $25 if only capital gains or price returns are considered. On the other hand, a dollar invested in T-Bills in 1961 would have grown to $21 by the end of 2010.
Figure 1: Value of $1 invested in stocks (price and total return) and in T-Bills in 1961 (click on the chart to enlarge)
What is perhaps surprising is that if we ignore dividends, the price returns and T-Bills returns result in almost similar levels of wealth. While this may not be obvious from Figure 1, Figure 2 shows the underlying reason for this observation.
Figure 2: Monthly TSX 300 Price Return vs. T-Bills (click on the chart to enlarge)
Monthly T-Bills returns have been relatively smooth and positive, averaging 0.50% (median of 0.4%, maximum of 1.6%, minimum of .01%, standard deviation of 0.3%) – even if in some years or even for decades it has been close to zero. In comparison, the TSX 300 Price Return series shows considerable volatility from month to month; out of 600 months over the five decades, about 246 (41%) months’ price returns are negative. This means that one has about a 41% chance of losing money in any one particular month. The minimum monthly price return on the TSX 300 is -22.6% (realized in October 1987) while the maximum return is 16% (realized in January 1975) with an average return of 0.6%, median of 1% and a standard deviation of 4.5%. These results also indicate that the annual 3% dividend returns have been the main source of wealth creation in Canada. In the remainder of the paper, we focus on total return series.
Volatility Patterns across Decades
To understand changes in the volatility across decades, we present information about total returns and volatility on the Canadian stock market over the past five decades (Figure 3).
Figure 3: Average, Median and Standard deviation of TSX 300 Total Returns (click on the chart to enlarge)
As shown in Figure 3, the 1970s and 1980s were actually more volatile than the most recent decade, having higher standard deviations (5.2% and 5%, respectively) than the average standard deviation for all years and that of the most recent decade (4.5% in both cases). However, the 1970s and 1980s provided higher average returns (approximately 1.2% and 0.7% per month respectively) compared to 0.6% per month in the most recent decade. Thus the risk-return tradeoff has worsened.
Impact of Investment Horizon
While the results shown above focus mostly on monthly or annual returns, they do not allow insights about returns and risk (volatility) associated with increasing holding period horizons. Table 1 presents the risk and return profile for three investment horizons: one year, five years and ten years over the entire fifty year period.
Table 1: Risk and Return descriptive statistics for one year, five years and ten years investment horizons (click on the chart to enlarge)
A few stylized facts emerge from Table 1. With respect to T-Bills, not surprisingly, the average yearly return is about 7% across the different investment horizons. Although similar across the different horizons (i.e. from 3.8% for the one year horizon to 2.7% for the ten-year horizon), the volatility of T-Bills appears to be displaying an inverse relationship with investment horizon. Given the average return on stocks of about 10% to 12%, the risk premium on the Canadian market over the five decades of our study is about 3% to 5%. It should be noted, however, that the 7% average returns for T-Bills is greatly influenced by the high interest rate regimes in the 1980s and early 1990s. In the recent decade of our study (2001 to 2010), T-Bills are yielding about 4% (for the ten-year investment horizon) while stocks are yielding 9% for the same investment horizon – leading to a premium of about 5% for investing in stocks.
With respect to stocks, first, there is an inverse relationship between volatility and the length of investment horizon. Investors with only a one-year investment horizon face higher volatility (standard deviation of 19.5%) than those having a five-year investment horizon (standard deviation of 6.1%). Investors with a 10-year investment horizon experience the lowest volatility (standard deviation of 3.3%). Second, with respect to average stock (geometric) returns, the one-year investment horizon has slightly higher yearly returns (11.8%) than the five-year investment horizon (10.2%) and the ten-year investment horizon (10%).1 Third, the average returns for the different investment horizons provide clear evidence that investment in the stock market is best suited for investors with long term horizons and that the risk-return tradeoffs improve monotonically with investment horizon. Fourth, the range between the minimum and maximum holding period returns also decreases as the investment horizon increases. Figure 4 presents further evidence of the volatility of returns as the investment horizon increases.
Figure 4: Frequency distributions of yearly returns for investment horizons of one year, five years and ten years (click on the chart to enlarge)
This analysis of risk and return data from the last five decades shows that investors who maintain a long term investment horizon realize respectable (10% per year) investment returns at highly reduced volatility. However, these returns are not without risk. For example, there have been specific time periods when investors have received a minimum return of 2.8% (year ending August 2010) or maximum return of 19.5% (year ending August 1987). In contrast, an investor with a one-year investment horizon would have realized a slightly higher return of 11.8%, but would have had a likelihood of experiencing a minimum return of -39.2% (year ending June 1982) or a potential of realizing the maximum return of 86.9% (year ending June 1983).
Conclusions
In this paper, we provide evidence on risk (volatility) and return in the Canadian stock market over the past five decades. Our results provide support for a positive relationship between risk and return in Canada as well as a positive relationship with the time horizon of investment. The paper also reinforces the fact that stock markets are meant for long term investors and that dividends do matter.
1. However, when one considers only the last decade (2001 to 2010), the ten-year horizon yields an average return of 9% while the five-year horizon yields an average return of 8% and the one-year horizon yields an average return of 7%.