Fair value measurement by listed private equity funds
International Financial Reporting Standards and US GAAP require certain categories of assets and liabilities to be measured at fair value. Fair value measurements should reflect the price at which an orderly transaction to sell an asset or transfer a liability would occur.
Fair value measurements may use Level 1, Level 2 or Level 3 inputs:
- Level 1 inputs are quoted prices for identical assets and liabilities for which there is an observable actively traded market.
- Level 2 inputs are observable market data other than Level 1 inputs, including quoted market prices for similar assets, quoted prices in inactive markets, and other observable inputs relevant to pricing or derived from or corroborated by other observable market data.
- Level 3 inputs are unobservable inputs used when observable (Level 1 or Level 2) inputs are not available.
Level 3 inputs include estimates and assumptions necessary for the valuation process. Prior research examining the relevance of fair value measurements reported by banks suggests that the stock market valuation of Level 3 assets is at a discount.1,2 A possible reason for this is that investors perceive Level 3 fair value asset valuations to be less reliable than either Level 1 or 2 fair value asset valuations. This view is in line with one of the main criticisms levelled against fair value accounting: that it allows potentially unreliable fair value measurement when the asset (or liability) in question is not traded in a liquid market. In contrast, recent research based on the fair value measurements reported by closed-end funds3 suggests that Level 3 assets are priced similarly to Level 1 and Level 2 assets.
We examine whether reported fair values of individual investment assets held by listed private equity (LPE) funds reflect the economic fundamentals of the investees. We exploit the requirement for private companies in the UK to file financial statements. Under the assumption that company financial statements capture relevant information about economic fundamentals, we examine the extent to which the fair value measurements of individual investments recognised and disclosed by LPE funds reflect the fundamentals of investees.
Initially, we use LPE fund annual reports to identify 26,602 investee-year observations for 96 unique LPE funds over the period 2006-2015, with on average approximately 41 investments per fund year. We then eliminate observations relating to:
- investments in debt instruments or preferred stock, because the fixed cash flows for these instruments imply that different valuation models focused on credit risk are applicable;
- funds that do not disclose the carrying values or the size of ownership stakes of individual investments; and
- investees whose net income or equity book value is not provided in Bureau van Dijk’s Financial Analysis Made Easy (FAME) database, or whose equity book value is negative.
Our final sample comprises 7,802 individual investee-years for which fair value measurements and ownership stakes are available along with matching investee fundamentals data. When viewed at the LPE fund level, our sample comprises 405 investor-year observations. When we classify the available fair value measurements relating to individual LPE investments, we find that 5,131 (65.8%) use Level 1 inputs and 2,671 (34.2%) use Level 3 inputs.
We estimate linear regression models expressing fair value measurements as a weighted combination of equity book value and net income of investees. This enables us to evaluate the extent to which book value and net income capture the information used by LPE managers in making their fair value estimates. We examine for differences in the weights on book value and net income across Level 1 and Level 3 fair value assets. Our findings indicate that Level 3 assets' equity book value reflects information used by LPE fund managers when valuing the Level 3 assets, but net income does not. However, both equity book value and net income of Level 1 assets capture information used by LPE fund managers when valuing Level 1 assets. Additional analysis suggests that one reason for this difference might be that earnings of Level 3 assets are less persistent than those of Level 1 assets.
Next, we estimate the discretion used by LPE fund managers in determining the fair values. Discretion is defined as the extent to which information used by LPE fund managers is not captured by investee book value and net income. If LPE fund managers use discretion opportunistically when valuing investees, we would expect investors in LPE funds to discount investee fair values recognised by LPE funds. However, we find no evidence that investors in LPE funds discount the fair values recognised by LPE funds when the level of discretion exercised is relatively high.
Taken together, our results suggest that Level 1 and Level 3 assets reported by LPE funds reflect the financial statement information reported by investees, but to different degrees. Further, our evidence suggests that investors in LPE funds do not discount the discretion used by the LPE funds when valuing investees. Our results suggest that in fair value measurement LPE fund managers use information extending beyond the book value and net income of investee companies, and that the resulting fair value measurements are seen as credible by investors in LPE funds.