• Accountants and analysts as financial intermediaries Volume I, Part I • Measuring corporate earnings and profitability Volume I, Part II • Financial ratios, the risk of failure, and stock returns Volume I, Part III • Equity valuation Volume I, Part IV • Price-earnings ratios, market-to-book ratios, and stock returns Volume II, Part I • Earnings and stock returns Volume II, Part II • Fundamental analysis and stock returns Volume II, Part III

We use confidential, U.S. Census Bureau, plant-level data to investigate aggregation in external reporting. We compare firms’ plant-level data to their published segment reports by grouping a firm’s plants that share the same four-digit SIC code into a “pseudo-segment.” We then determine whether each pseudo-segment is disclosed as an external segment, or whether it is subsumed into a different business unit for external reporting purposes. We show that a pseudo-segment is more likely to be aggregated when the agency and proprietary costs of separately reporting the pseudo-segment are higher and when firm and pseudo-segment characteristics allow for more discretion in the application of segment reporting rules. For firms reporting multiple external segments, aggregation of pseudo-segments is driven by both agency and proprietary costs. For firms reporting a single external segment, we find no evidence of an agency cost motive for aggregation.

We develop an empirical method that allows us to evaluate the reliability of an expected return proxy via its association with realized returns even if realized returns are biased and noisy measures of expected returns. We use our approach to examine seven accounting-based proxies that are imputed from prices and contemporaneous analysts' earnings forecasts. Our results suggest that, for the entire crosssection of firms, these proxies are unreliable. None of them has a positive association with realized returns, even after controlling for the bias and noise in realized returns attributable to contemporaneous information surprises. Moreover, the simplest proxy, which is based on the least reasonable assumptions, contains no more measurement error than the remaining proxies. These results remain even after we attempt to purge the proxies of their measurement error via the use of instrumental variables and grouping. We provide additional evidence, however, that demonstrates that some proxies are reliable when the consensus long-term growth forecasts are low and/or when analysts' forecast accuracy is high.

We evaluate whether reported accounting numbers are informative about earnings uncertainty and whether earnings uncertainty is priced. We use quantile regressions to forecast the standard deviation, skewness, and kurtosis of future earnings. These three moments are important measures of earnings uncertainty because they reflect the size of the average deviation from expected earnings and the amount of extreme upside potential, extreme downside risk, or both. We develop a novel approach for evaluating the reliability of our forecasts and we show that they are reliable. We also document that: (1) equity prices are increasing (decreasing) in the standard deviation and skewness (kurtosis) of lead return on equity and (2) credit spreads are increasing (decreasing) in the standard deviation and kurtosis (skewness) of lead return on assets. Our results indicate that historical financial statements are informative about earnings uncertainty and that earnings uncertainty is priced. Data Availability: Data are available from the public sources cited in the text.

We use a simple k-nearest neighbors algorithm (hereafter, k-NN*) to forecast earnings. k-NN* forecasts of one-, two-, and three-year-ahead earnings are more accurate than those generated by popular extant forecasting approaches. k-NN* forecasts of two- and three-year (one-year)-ahead EPS and aggregate three-year EPS are more (less) accurate than those generated by analysts. The association between the unexpected earnings implied by k-NN* and the contemporaneous market-adjusted return (i.e., the earnings association coefficient (EAC)) is positive and exceeds the EAC on unexpected earnings implied by alternate approaches. A trading strategy that is long (short) firms for which k-NN* predicts positive (negative) earnings growth earns positive risk-adjusted returns that exceed those earned by similar trading strategies that are based on alternate forecasts. The k-NN* algorithm generates an empirically reliable ex ante indicator of forecast accuracy that identifies situations when the k-NN* EAC is larger and the k-NN* trading strategy is more profitable. Data Availability: Data are available from the public sources described in the text.