DISCLOSURE QUALITY AND VALUATION RISK IN LEVEL 3 INVESTMENTS: A DISCLOSURE-IMPLIED RISK FRAMEWORK FOR U.S. BUSINESS DEVELOPMENT COMPANIES
Keywords:
Business development companies, Fair value hierarchy, Level 3 inputs, Disclosure quality, Valuation risk, Double machine learning, Conformal prediction, Audit samplingAbstract
Business development companies (BDCs) report most of their investments as Level 3 fair value measurements, so the net asset value (NAV) relied on by investors and auditors rests on unobservable inputs. This paper asks how much of that valuation risk can be recovered from what BDCs publish. We propose a disclosure-implied valuation risk (DIVR) measure that reconstructs each row of the ASC 820 unobservable-input table from its range and weighted average with a mean-matched Beta-PERT law, calibrates ranges for the number of holdings with order statistics, propagates rows to NAV through a Gaussian copula, and adds a benchmark-gap estimate of overstatement from stale or optimistic marks. DIVR is paired with a nine-item Level 3 Disclosure Quality Index (L3-DQI). Both are evaluated on a disclosure-calibrated panel of 48 BDCs over fiscal years 2016–2023 (384 fund-years, 54,618 positions) in which hidden portfolios are fixed while disclosure formats vary counterfactually. Granular input tables raise the out-of-fund R² for latent valuation dispersion from 0.71 to 0.80, whereas weighted averages are what make overstatement detectable (rank correlation 0.54 versus 0.09). In price-to-NAV regressions estimated by panel OLS and double machine learning, disclosure-implied risk carries a discount and the benchmark gap is priced only after weighted averages became mandatory, while the price effect of disclosure quality is positive but small relative to price noise in a market of 48 listed funds. For auditors, a back-testing score captures 44% of absolute misstatement when 20% of Level 3 fair value is tested, against 22% for size-based selection. Recoverability and audit results hold in all 10 independent panels.References
[1] Chernenko S, Erel I, Prilmeier R. Why do firms borrow directly from nonbanks? Review of Financial Studies, 2022, 35(11): 4902-4947.
[2] Financial Accounting Standards Board. Fair value measurement (Topic 820): Disclosure framework—Changes to the disclosure requirements for fair value measurement. Accounting Standards Update No. 2018-13. Norwalk: FASB, 2018.
[3] U.S. Securities and Exchange Commission. Good faith determinations of fair value. Release No. IC-34128. Washington: SEC, 2020.
[4] Public Company Accounting Oversight Board. Auditing accounting estimates, including fair value measurements, and amendments to PCAOB auditing standards. PCAOB Release No. 2018-005. Washington: PCAOB, 2018.
[5] Song C J, Thomas W B, Yi H. Value relevance of FAS No. 157 fair value hierarchy information and the impact of corporate governance mechanisms. The Accounting Review, 2010, 85(4): 1375-1410.
[6] Riedl E J, Serafeim G. Information risk and fair values: An examination of equity betas. Journal of Accounting Research, 2011, 49(4): 1083-1122.
[7] Hanley K W, Jagolinzer A D, Nikolova S. Strategic estimation of asset fair values. Journal of Accounting and Economics, 2018, 66(1): 25-45.
[8] Christensen B E, Glover S M, Wood D A. Extreme estimation uncertainty in fair value estimates: Implications for audit assurance. Auditing: A Journal of Practice & Theory, 2012, 31(1): 127-146.
[9] Cannon N H, Bedard J C. Auditing challenging fair value measurements: Evidence from the field. The Accounting Review, 2017, 92(4): 81-114.
[10] Glover S M, Taylor M H, Wu Y J. Current practices and challenges in auditing fair value measurements and complex estimates: Implications for auditing standards and the academy. Auditing: A Journal of Practice & Theory, 2017, 36(1): 63-84.
[11] Chernozhukov V, Chetverikov D, Demirer M, et al. Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal, 2018, 21(1): C1-C68.
[12] Bach P, Chernozhukov V, Kurz M S, et al. DoubleML—An object-oriented implementation of double machine learning in Python. Journal of Machine Learning Research, 2022, 23(53): 1-6.
[13] Barber R F, Candès E J, Ramdas A, et al. Conformal prediction beyond exchangeability. The Annals of Statistics, 2023, 51(2): 816-845.
[14] Angelopoulos A N, Bates S. Conformal prediction: A gentle introduction. Foundations and Trends in Machine Learning, 2023, 16(4): 494-591.
[15] Grinsztajn L, Oyallon E, Varoquaux G. Why do tree-based models still outperform deep learning on typical tabular data? In: Advances in Neural Information Processing Systems 35 (NeurIPS 2022), Datasets and Benchmarks Track, 2022.
[16] Ke G, Meng Q, Finley T, et al. LightGBM: A highly efficient gradient boosting decision tree. In: Advances in Neural Information Processing Systems 30, 2017: 3146-3154.
[17] Lambert R, Leuz C, Verrecchia R E. Accounting information, disclosure, and the cost of capital. Journal of Accounting Research, 2007, 45(2): 385-420.
[18] Easley D, O’Hara M. Information and the cost of capital. The Journal of Finance, 2004, 59(4): 1553-1583.
[19] Laux C, Leuz C. The crisis of fair-value accounting: Making sense of the recent debate. Accounting, Organizations and Society, 2009, 34(6-7): 826-834.
[20] Bens D A, Cheng M, Neamtiu M. The impact of SEC disclosure monitoring on the uncertainty of fair value estimates. The Accounting Review, 2016, 91(2): 349-375.
[21] Chung S G, Goh B W, Ng J, et al. Voluntary fair value disclosures beyond SFAS 157’s three-level estimates. Review of Accounting Studies, 2017, 22(1): 430-468.
[22] Hope O K, Hu D, Lu H. The benefits of specific risk-factor disclosures. Review of Accounting Studies, 2016, 21(4): 1005-1045.
[23] Bloomfield R J. The ‘incomplete revelation hypothesis’ and financial reporting. Accounting Horizons, 2002, 16(3): 233-243.
[24] Cassar G, Gerakos J. Hedge funds: Pricing controls and the smoothing of self-reported returns. Review of Financial Studies, 2011, 24(5): 1698-1734.
[25] Brown G W, Gredil O R, Kaplan S N. Do private equity funds manipulate reported returns? Journal of Financial Economics, 2019, 132(2): 267-297.
[26] Getmansky M, Lo A W, Makarov I. An econometric model of serial correlation and illiquidity in hedge fund returns. Journal of Financial Economics, 2004, 74(3): 529-609.
[27] Lee C M C, Shleifer A, Thaler R H. Investor sentiment and the closed-end fund puzzle. The Journal of Finance, 1991, 46(1): 75-109.
[28] Berk J B, Stanton R. Managerial ability, compensation, and the closed-end fund discount. The Journal of Finance, 2007, 62(2): 529-556.
[29] Ahn J, Hoitash R, Hoitash U. Auditor task-specific expertise: The case of fair value accounting. The Accounting Review, 2020, 95(3): 1-32.
[30] Malcolm D G, Roseboom J H, Clark C E, et al. Application of a technique for research and development program evaluation. Operations Research, 1959, 7(5): 646-669.