360 likes | 559 Views
Asset Securitizations and Audit Effort. Yu Yu Zhang a , Gary Monroe a , Dominic Gasbarro b , Grant Cullen b , Greg Shailer c a Australian School of Business, UNSW b Murdoch Business School C The Australian National University. Research Questions.
E N D
Asset Securitizations and Audit Effort • Yu Yu Zhanga, Gary Monroea, Dominic Gasbarrob, • Grant Cullenb, Greg Shailerc • a Australian School of Business, UNSW • b Murdoch Business School • C The Australian National University
Research Questions Is audit effort associated with asset securitization risks from 2003 to 2009? If associated, does this association change after the occurrence of the GFC?
Motivations “Where were the auditors in asset securitization”: perceived audit failures in asset securitizations; No prior research on auditors’ role in asset securitizations. Limited studies on bank audit;
Why were asset securitizations so problematic to auditors? • FAS 140 (FAS 125 before 2000, FAS 156 after 2006): allows for sale accounting if the transaction could satisfy certain criteria. • FIN 46R: (Exemption of QSPE from consolidation); • FAS 157: Fair value measurements Accounting benefits of sales accounting and non-consolidation: • A lower leverage ratio; • A better liquidity ratio; • A higher profitability; • A flexible and favourable risk-based capital ratio.
Empirical Evidence on Asset Securitization The economic and accounting benefits are gained at the expense of information asymmetry (Cheng et al. 2011). Information uncertainty and asymmetry in asset securitizations: -the true risk status is veiled (Minton et al. 2004) -RI creates another information asymmetry on the private information with regard to the retained tranches and the underlying assets, discount rate assumptions, and other assumptions for the fair value estimation. Evidence on EM in securitizations: -timing (Rosenblatt et al. 2005; Dechow and Shakespeare 2009); -classification of sales versus borrowings (Karaoglu 2005); -selection of loans to be securitized (Ambrose et al. 2005; Minton et al. 2004); -valuation of retained interests (see FAS 157; Dechow et al. 2009).
Hypotheses Development: The level of securitized assets and audit fees Complexity and lack of transparency in asset securitization transactions (Barth et al. 2003; Schwzrcz 2004; Ryan 2007) , increase audit risk; Sale accounting and non-consolidation treatments veil the true economic substance and financial risk status of the originating banks (Shipper and Yohn 2007; Barth et al. 2010; Chen et al. 2008; Landsman et al. 2008 etc.), increasing audit risk ; Securitizations create opportunities for earnings and capital management, increasing audit risk. Information asymmetry between originators and subsequent investors (Ryan 2007); managers have incentives to engage in earnings and capital management via securitizations (Matsumoto 2002; Karaoglu 2005 etc.); approaches include timing, valuation, selection of loans, sale/borrowing classification (Karaoglu 2005; Ambrose et al. 2004; Minton et al. 2004; Dechow et al. 2009 etc.); auditors consider earnings management and aggressive reporting in audit planning and pricing (Houston 1999; Gul et al. 2003; Bedard and Johnstone 2004 etc.) Increased audit risk leads to additional audit effort and then increases audit fees. H1: Audit fees are positively associated with the amount of securitized assets.
Hypotheses Development: Audit Fees and Retained Interests Auditors also consider the risk retained by the originator when evaluating the inherent risk. -Regulators and standard setters believe the risk retained by the originator is restricted to the extent of retained interests (FAS 140; FIN46R); -empirical evidence reveals that the financial market treat the risk retained to the extent of total amount of the securitized assets, especially under unfavourable market conditions (Niu and Richardson 2006; Landsman et al. 2008; Barth et al. 2010 etc.). If auditors adopt market participants’ point of view, they will focus more on the securitized assets, supporting H1. If auditors adopt regulators and standard setters’ point of view, their risk consideration will focus more on the components of retained interests, supporting H2. H2: Audit fees are positively associated with the amount of retained interests.
Pre- and GFC Comparison The GFC: 2007 The impact of the GFC to auditors: -Changing economic environment leads to the variation of audit risk factors on asset securitization (constraints on finance availability; going concern and liquidity problems; the discretion and flexibility in using OBS financing; estimation and valuation uncertainties); -Auditors address the changed audit risk factors by modifying audit effort or charging a fee premium. H3: There is a stronger association between audit fees and asset securitization risks after the GFC compared to before the GFC. H3a: audit fees –securitized assets; H3b: audit fees –retained interests.
The impact of asset securitization on audit pricing to on-balance-sheet credit risk factors There is a positive association between audit fees and bank credit risk factors (Fields et al. 2004). Asset securitizations dress up on-balance-sheet credit risks by removing loans from the balance sheet. The awareness of the risks embedded in asset securitizations should raise auditors’ concerns about on-balance-sheet credit risks, resulting in more audit effort on credit risk factors. H4: As asset securitization risks increase, audit fees increase relative to credit risks. H4a: As SA increase, audit fees increase relative to credit risks. H4b: As RI increase, audit fees increase relative to credit risks.
Model Basic Model: Modified Fields et al. (2004) bank audit fee model Model (1) The importance of interest sensitive assets should be linked with the magnitude and the direction of interest rate changes: SENSITIVE*ΔINT to replace SENSITIVE INTDERIV, to proxy for off-balance-sheet interest rate risks Test variables: (from Y9C report from BHCs) Test Variables: SARATIO= total outstanding securitized assets, deflated by total assets (Barth et al. 2011); RIRATIO= total retained interests, including retained interest only strips, retained credit enhancements, and unused commitments to provide liquidity (service advances), deflated by total assets (Barth et al. 2011); GOSRATIO= relative gains on securitization, calculated as the net securitization income divided by net income; PGFC = 1 for the years 2007, 2008 and 2009, and 0 otherwise.
Models to Test the Hypotheses H1, H2 and H3 Model : LNAF=f (CONTROLS, SA, RI, GOS, PGFC, SA*PGFC, RI*PGFC) • H1: a positively significant SA; (the risk rests with total amount of the securitized assets) • H2: a positively significant RI; (the risk rests with RI portion only) • H3: PGFC, SA*PGFC, RI*PGFC A positive PGFC is expected. H3a: Positively significant SA*PGFC H3b: Positively significant RI*PGFC H4: interaction terms of securitization factors with Asset structure or asset quality proxies • Asset structure: COMMLOAN, MTGLOAN • Asset Quality: NONPERFORM, CHGOFF • E.g., COMMLOAN*SA, COMMLOAN*RI, COMMLOAN*SA*PGFC, and COMMLOAN*RI*PGFC
Data and Sample Data Audit fee, auditor change: Audit-Analytics; Financial statement data, asset securitization & derivative details: BHC Y9C reports from Bank Regulatory Database; Volatility of stock return: CRSP database; Interest rate: U.S. Treasury official website. Sample U.S. bank holding companies (BHCs); Period: 2003-2009 (securitization details are available from 2003); N=2,424 (N=2,423 in regression testing as we exclude JPM2009 from testing for an unbelievable increase in RIAMOUNT by $91bn in 2009)
Table 5: The impact of asset securitization on audit pricing to On-Balance-Sheet Credit risks
Problems with the preliminary results Insignificant results; Some results are weird. Multicollinearity problem among control variables and test variables. The test variables (SA, RI and GOS) tend to measure the same thing. Solution Test variables: using component factors from PCA to measure securitization uncertainty (Cheng et al. 2011); PCA on control variables; Using residuals model: LNAF=f (Control Variables)+ Residual Residual = f (Test variables)
Conclusion Auditors price audit risks associated with securitization uncertainties for the period 2003-2009; Audit pricing to securitization risks are higher for the pre-GFC period; For the pre-GFC period, auditors seem to price all five items of the securitization risk measures; For the during-GFC period, auditors only focus on retained interests and securitization incomes; With the presence of higher level of securitization risks, auditors give higher pricing to mortgage loans, indicating an influence of securitization risks on audit pricing to on-balance-sheet credit risks.
Robustness Tests Matched Pair Sample Tests Securitizer Only Sample Tests Tests with CAMELS Model Further controlling for: • Changes in Macroeconomic Conditions • Excluding Year 2006 • Excluding Year 2007 • Auditor change and Auditor Independence • Investment in MBS and ABS
Suggestions?Comments? Thank you!
Discussion Although the main tests and subsequent robustness tests and additional tests give us fairly consistent results, our results are, thus far, inconclusive. • There is strong evidence that retained interests affect audit effort (fees) after 2007. • The relevance of total securitized assets is still unclear. We need to further consider the following: (1) Size effects: Securitizers are clustered in large BHCs, and most of the control variables in the model are highly correlated with size measure. Size effect might have distorted the true association between audit pricing with the securitization factors and other control variables; (2) Self-selection on Big N Auditors; (3) Characteristics of Securitizers: BHCs with securitization activities are structurally different from non-securitizers as shown in Table 1. We are considering two-stage modeling (to account for the possible bias in the propensity to be a securitizer).
Implication The main tests and additional analyses are generally consistent on the association between audit fees and asset securitization factors in several dimensions. Retained interests are significantly associated with audit fees, and this positive association strengthens for the period GFC period. Our results suggest that auditors focused on the risks associated with the retained portion of the securitized assets in their audits, rather than securitization levels, especially after the GFC. This is consistent with auditors persisting in treating asset securitizations as sales, even though bank failures during the GFC demonstrate that it might be wiser to treat them as secured borrowings.
Other Findings Off-balance-sheet risks (INTDERIV) are also priced in audit fees; Big N auditors might be better at pricing asset securitization risks.
Contribution It relates to several points addressed in Basel (2008); It provides insights on auditors’ behaviours around the financial crisis, particularly focusing on asset securitizations; The relatively stable regulatory environment during the study period creates an ideal situation to examine the inter-temporal variation of the association between audit fees and asset securitization under a changing economic environment from prosperity to recession; It helps explain the audit failures on asset securitizations. The results can be useful for regulators, standard setters and the audit profession.