Abstract

This thesis consists of three essays. The first essay (chapter two) examines short- and long-term persistence and predictability of earnings and cash-flow growth rates. Aggregate results show that the mean firm exhibits persistence in short-term earnings growth rates, but little persistence in growth at the long-run. More consistent with economic intuition, aggregate cash-earnings growth rates show some short-term mean-reversion and little persistence in the long-run. Consistent with the earnings smoothing theory, cross-sectional results show short-term persistence in earnings growth rates. Cash-earnings, which suffer less manipulation than accrual earnings, show short-term cross-sectional mean-reversion. Long-term growth rates show evidence of mean-reversion in both cases. Our most interesting findings are related to various examined attributes such as age, industry characteristics, listing exchange, and the number of analysts following the stock which are shown to have an impact on the growth persistence of individual firms. The second essay (chapter three) examines the increasing trend in idiosyncratic volatility of stock returns, which has been documented by Campbell et al. (2001). Using the Campbell (1991) return decomposition framework, we relate idiosyncratic volatility of returns to the volatility of changes in expected ROEs for one-, two- and three-year horizon forecasts (i.e., to the volatility of cash-flow news). The upward trend in idiosyncratic volatility documented by Campbell et al. (2001) is associated with a similar increasing trend in the volatilities of cash-flow news for the three forecast horizons. This relationship is persistent after correcting for analysts' forecast biases and is consistent for newly-listed and mature firms and for earnings (non-) announcement periods. Our findings support an informational explanation to the increasing trend in idiosyncratic volatility, and are consistent with the market efficiency hypothesis which implies that stock return volatility is caused by an increase in the uncertainty of fundamental variables. In the third essay (chapter four) we estimate the internal rates of return (IRR) for domestic and cross-listed Canadian firms, nine GICS sectors, and cap-based portfolios. Using the Fama and French (1999) methodology, we distinguish between the IRR on value or the cost of capital as expected by stakeholders and the IRR on cost that represents the implied real cost of investments. The real cost of capital [equity] over the period 1960-2003 is 5.90 [6.58] percent for the entire nonfinancial sector. The real return on such cost is 6.64 [7.78] percent indicating that, similar to the U.S., the Canadian corporate sector creates value. As expected, cross-listed firms enjoy, on average, lower costs of capital and equity than domestic firms. Although the cost of capital varies considerably across GICS sectors, value is created almost equally by all GICS sectors. IRR values on cap-based portfolios are consistent with the negative relationship between costs of capital and firm size. While both types of IRRs decrease on average after we correct for replacement costs and include extraordinary items, the IRR on cost remains higher than the IRR on value confirming the positive value creation hypothesis. Allowing for inflation illusion, our findings suggest that equity was undervalued during most of the sample period and that the misvaluation was lower for cross-listed firms