Accounting and Control RSS

  The Conservative Bias in Asset Fair Value Measurements 

Print Share


In recent years, accounting regulators have favored the use of fair values over historical costs in the measurement of financial assets for financial reporting. Their rationale is that fair value measurements are more useful for investors because they show the current market value of assets instead of their historical cost. Taking a different view, many investors, Warren Buffett among them, have long warned about the presence of optimistic biases in so-called fair values. "Errors will usually be honest, reflecting only the human tendency to take an optimistic view of one's commitments (...) (but) in extreme cases, mark-to-model degenerates into what I would call mark-to-myth (valuations)," Buffett told his shareholders back in 2003.

But current research by Marc Badia, Miguel Duro and Fernando Peñalva of IESE, with Stephen Ryan of NYU Stern School of Business, uncovers a different slant. The co-authors find evidence that companies measure the fair value of financial assets conservatively when there is more uncertainty in the inputs used to obtain the fair value. In this way, firms try to mitigate investors' discounting of these measurements. Applying this conservative approach is even more likely under pressure from investors or governance bodies, perhaps to gain their confidence and trust.

While fair values are meant to provide an accurate picture of a firm's financial assets, markets are not always liquid and market prices are not always available to reliably measure the fair values. Discretion must then be used. Understanding the discretionary biases at play can help empower executives and investors to make better decisions.

Fair Values and Bias
So how do companies estimate fair values? In liquid markets, with many buyers and sellers, it's as easy as finding the stock price of the financial assets involved. These "mark-to-market" or level 1 inputs are virtually free from bias, as they are based on quoted prices for identical instruments in active markets.

But in illiquid markets, companies are forced to estimate fair values of financial assets, relying on market information that is incomplete (level 2) or unobservable (level 3), as labeled by U.S. Generally Accepted Accounting Principles (GAAP) and International Financial Reporting Standards (IFRS). Everyone knows these level 2 or 3 inputs are discretionary, and, especially after the fallout caused by the recent financial crisis, investors are likely to take them with a grain of salt. In this context, conditionally conservative estimates of fair value try to mitigate investors' distrust.

The Test
The authors hypothesized that firms with knowledgeable, involved investors would be more likely to report financial assets conservatively. And, based on U.S. public financial data from 2007 to 2014, they were right: companies with knowledgeable investors were incentivized to provide conservative fair value measurements.

Knowing that financial reports will be verified by a third party further increases a firm's likelihood to report conservative fair values. The study shows that companies who employed top auditors were much more likely to report conservative fair values for their financial instruments.

But in certain circumstances, companies resist the urge to provide conservative figures. Specifically, the authors found that companies who either meet or beat their earnings targets by a narrow margin are less conservative in reporting fair values. "Narrow beaters" seem to have incentives that interfere with conditional conservatism, reporting more optimistically to meet their earnings targets. This is in contrast to firms that clearly miss their targets.

From Finance to Oil
The authors first measured the level of conditional conservatism displayed by companies in the context of financial instruments. But their conclusions also seemed to apply to non-financial assets. They tested their hypotheses with Canadian and U.S. oil and gas firms and concluded that these companies also report conservative fair values of their oil and gas reserves, under the right circumstances.

The message is clear: there is a tendency towards reporting fair values conservatively, when illiquid markets are involved, consistent with firms trying to mitigate investors' discounting of the measurements.

So, if we first accept that fair value estimates can be biased, next we should determine whether the biases firms show are positive or negative in the long run. As biases go, conditional conservatism may be the way to win back investor trust.

Methodology, Very Briefly
In the central study, the authors analyzed fair value reporting by companies trading in financial instruments to test whether firms report conditionally conservative fair value measurements in illiquid markets. Data spanned 2007 to 2014. As a placebo control, these data were compared with income not attributable to fair value measurements. The full sample was partitioned in order to determine which specific governance mechanisms incentivize conservative reporting. To demonstrate the generalizability of their conclusions, the authors also conducted a similar analysis on Canadian and U.S. oil and gas firms' disclosed measures of reserves, which are akin to certain fair value measurements.

Marc Badia and Miguel Duro acknowledge financial support from the Spanish Ministry of Science and Innovation, Grant ECO2010-19314, and from Columbia University CIBER, respectively. Marc Badia and Fernando Peñalva acknowledge financial support from the Spanish Ministry of Economics and Competitiveness, Grants ECO2013-48328-C3-1-P, and ECO2016-77579-C3-1-P.
This article is based on:  Conditionally Conservative Fair Value Measurements
Publisher:  Elsevier
Year:  2017
Language:  English

Related Material

Conservatism and Accounting Earnings Giner, Begoña; Peñalva, Fernando  Conservatism and Accounting Earnings
Information Consequences of Accounting Conservatism García Lara, J.M.; García Osma, B.; Peñalva, Fernando  Information Consequences of Accounting Conservatism