Despite growing public skepticism over how useful financial statements are in providing information to investors, researchers at Stanford’s Graduate School of Business have found that the value of financial ratios for predicting bankruptcy has not declined significantly over time.
Professors Maureen McNichols and William Beaver and graduate student Jung-Wu Rhie have reexamined the usefulness for predicting bankruptcy of financial ratios such as return on assets (net income divided by total assets), cash flow to total liabilities (earnings before interest, depreciation, and taxes divided by both short- and long-term debt), and leverage (total liabilities to total assets). The study explored how three forces have influenced this predictive value over the past 40 years.
The first force is that standard-setting bodies such as the Financial Accounting Standards Board and the Securities and Exchange Commission have been trying to increase the usefulness of the information found in financial statements and to enhance the ability of such statements to convey the fair value of assets and liabilities.
“One prediction is that if standard-setters are successful at incorporating additional information about fair values into financial statements, then we might expect their predictive ability for bankruptcy to increase,” said McNichols, the Marriner S. Eccles Professor of Public and Private Management at the Business School.
On the other hand, two forces could work against an increase in the usefulness of financial statements. First, the range of activities that today’s companies engage in might be less well captured by traditional accounting methods.
“If we look back to the 1960s, intangible assets—as represented by investments in brands, research and development, and technology—were much less pervasive than they are today,” said McNichols. “These kinds of transactions are not well captured by our current accounting model.” So even if standard-setting bodies were successful in improving standards and increasing the informativeness of financial statements, the shift in economic activities of businesses might offset that.
Second, financial statements may be more “managed” today than in the past. “Certainly, there is much greater documentation of earnings management today than we’ve seen historically,” said McNichols.
To determine if the usefulness of financial statements for predicting bankruptcy has changed over time, the researchers examined a sample of bankrupt and non-bankrupt firms spanning 40 years, from 1962 to 2002. First, they divided the companies into two time periods: from 1962 to 1993, and from 1994 to 2002. They looked at a number of financial ratios that have been shown in prior literature to be predictive of bankruptcy: return on assets (ROA); cash flow to total liabilities; and leverage.
ROA is a measure of the profitability of the assets, and is a critical ratio to track since prior research has shown that capital markets are concerned about the ability of firms to repay debts—and profitability is a key indicator of ability to pay. Cash flow to total liabilities is a measure of the ability of the firm to use cash flow from operations to service the principal and interest payments. Leverage, or liabilities to total assets, is a measure of the debt to be repaid relative to the total assets of the firm available as a source for repaying the debt.
Based on these ratios, the researchers found that financial statements did a good job of predicting bankruptcy in both periods. Although the predictive ability of the ratios declined from Period 1 (1962 to 1993) to Period 2 (1994 to 2002), that decline was not statistically significant. Moreover, the predictive ability in both time periods is very high. And the “hazard rate” of a company, which reflects the risk of going bankrupt given all three of these ratios, predicted a higher risk not just in the year before bankruptcy but in a number of years prior to bankruptcy. “In fact, we see differences in the ratios of bankrupt and nonbankrupt firms up to five years prior to bankruptcy,” said William H. Beaver, the Joan E. Horngren Professor of Accounting Emeritus.
The next thing the researchers looked at was the ability of three “market-based variables” to predict bankruptcy: the variability of stock returns; the cumulative stock returns over a 12-month period; and the size of the firm as measured by the market capitalization (common stock price per share times the number of common shares outstanding).
These market-based variables also proved to be highly successful at predicting bankruptcy: 92 percent of bankrupt firms were in the highest three deciles of hazard rates for Period 1 and 93 percent for Period 2. In this case, the ability to predict bankruptcy actually rose over time. The slightly higher predictive power of the market-based variables occurs because market prices reflect a broader set of information than just the financial statements. The difference in the predictive power of the accounting-based and market-based models reflects the incremental importance of non-financial statement data.
Then, the researchers combined the two models for predicting bankruptcy—using both financial ratios and market-based variables—and came up with a 96 percent chance of predicting bankruptcy correctly for Period 1 and 93 percent for Period 2. According to Beaver, what this seems to suggest is that market prices appear to compensate for even slight deterioration in predictive ability of financial ratios, consistent with the market drawing upon additional information not included in the accounting ratios.
There’s further work ahead for the researchers to see if they can separate out whether increased use of intangibles and/or financial statement manipulation are offset by reporting improvements from the work of standard-setters.
“But it’s comforting to know that the behavior of the combined model over time is so stable,” said McNichols. It suggests capital market participants’ ability to predict a key financial event, bankruptcy, has not eroded.
