Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

The Opaqueness of Structured Bonds: Evidence from the U.S. Insurance Industry

  • 114 Accesses


It has been argued that the opaqueness of structured bonds, such as mortgage-backed securities, asset-backed securities and collateral debt obligations, was one of the major causes of the recent financial crisis that started in late 2007. We analyse the evolving nature of information asymmetry inherent in various types of structured bonds by examining the U.S. insurers’ assets. We show that, prior to 2004, structured bonds were not associated with greater information asymmetry; however, holding more multi-class structured bonds, especially privately placed bonds, increased the information asymmetry when evaluating insurers’ assets post-2004. The effect of information asymmetry was more significant with life insurers than with non-life insurers. In addition, by investigating the rating grades of such structured bonds, we find that the market views higher-grade, privately placed, multi-class structured bonds as having the highest information asymmetry among all types of structured bonds post 2004, an effect which is, again, more significant with life insurers. This result shows that structuring complexities and unreliable ratings make structured bonds more opaque than just securitisation itself.

This is a preview of subscription content, log in to check access.

Figure 1

Note: This figure shows the global CDO issuance from 2000 to 2010. Values are in millions. “Total CDOs” represents total CDO issuance, and “Structured Finance in CDOs” is the total structured finance included in the CDO asset collateral pools. Structured finance includes ABS/MBS, CDOs of CDOs, structured finance indices, and other securitised/structured finance.

Figure 2

Note: This graph shows the percentage of the U.S. non-agency mortgage-related securities issuance to the total U.S. mortgage-related security issuance from 1996 to 2010. Agency issuance includes mortgage issuance by FNMA, FHLMC, and GNMA.


  1. 1.

    See, for example, Dionne (2009); Harrington (2009); Scott and Taylor (2009); Caprio et al. (2010); Ryan (2008); Cheng et al. (2011).

  2. 2.

    See, for example, Murphy (2008); Hellwig (2009); Schwarcz (2009).

  3. 3.

    The loss in subprime MBSs was estimated at about $500 billion dollars, less than the losses from the technology bubble of 2000 (Hellwig, 2009).

  4. 4.

    Cheng et al. (2011).

  5. 5.

    Information asymmetry in insurance literature often refers to the information asymmetry between policyholders and insurance companies. The information asymmetry in this study, however, refers to the information asymmetry among capital market participants in assessing the value of insurance companies’ assets and liabilities.

  6. 6.

    Flannery et al. (2004); Zhang et al. (2009); Cheng et al. (2011).

  7. 7.

    Cheng et al. (2011) and Flannery et al. (2004) use the terms “information uncertainty” and “opaqueness,” respectively, to describe similar concepts for bank holding companies. Rigorously speaking, a lack of information may not always increase information asymmetry, but high information asymmetry has to be associated with opaqueness/lack of information disclosure.

  8. 8.

    Bagehot (1971); Kyle (1985); Glosten and Milgrom (1985).

  9. 9.

    Some insurers do participate in the securitisation market as issuers, but to a much lesser extent than banks. Life insurers conduct mortality and longevity risk securitisations to transfer risk to capital markets, and some life insurers conduct XXX or AXXX redundant reserve securitisations to fund the extra reserves required by regulations XXX and Actuarial Guideline AXXX. Some life and P&C insurers conduct embedded value securitisation to monetise the embedded value of a particular book of business in order to fund acquisition expenses or demutualisation costs. property/casualty (P&C) insurers often issue catastrophe bonds, industry loss warranties (ILWs), and sidecars to transfer risk to the capital markets. According to Cowley and Cummins (2005), insurance securitisation is actually “monetisation” in that there is no “true sale” of the asset to a special purpose vehicle, so such transactions are typically “on balance sheet.” As shown in Powers (2012), the size of insurance securitisation is small. For example, the value of annual issuances was $15.5 billion in 2007 and was $4.1 billion in 2008.

  10. 10.

    AIG and monoline insurers are exceptions because they were also involved in the crisis by “insuring” mortgage-related securities. We conduct robustness tests by excluding these insurers, and the main results still hold.

  11. 11.

    Liebenberg et al. (2010); IMF (2008); Manconi et al. (2012).

  12. 12.

    For example, Scottish Re (U.S.) Inc. was put under supervision in January 2009 by the Delaware Insurance Commissioner mainly because of its declining asset values from residential mortgage-backed securities. Before running into financial trouble, as of 31 December 2007, MBSs and CMOs (collateralised mortgage obligations) constituted approximately 19.3 per cent of the firm’s invested assets.

  13. 13.

    Mason and Rosner (2007).

  14. 14.

    A type of U.S. mortgage with a risk level that lies between prime loans and subprime loans.

  15. 15.

    Gorton (2009).

  16. 16.

    Official data on total subprime and Alt-A MBS exposures in CDOs are not available because all CDOs are privately placed.

  17. 17.

    More information both on how credit rating agencies evaluate structured bonds and on issues associated with the rating process is available in the Technical Committee of the International Organization of Securities Commissions’ 2008 final report on The Role of Credit Rating Agencies in Structured Finance Markets.

  18. 18.

    Caprio et al. (2010).

  19. 19.

    According to Vink and Thibeault (2008), the average MBS matures in just over 27.7 years. They use non-U.S. MBS data only.

  20. 20.

    The typical coupon rate for AAA-rated CDOs is about LIBOR + 26 bps, LIBOR + 75 bps for A-rated CDOs, LIBOR + 180 for BBB-rated CDOs, and LIBOR + 475 for BB-rated CDOs Lucas et al. (2006).

  21. 21.

    MBSs can be more attractive to life insurers than mortgage loans. Mortgages held directly by insurers receive a higher risk factor in the life insurance risk-based capital calculation than mortgages held as bonds with good ratings. As a result, the (re)packaging of mortgages as bonds may have facilitated misrepresentation of the asset risk of insurers because the mortgage risk held in the MBS may not be adequately reflected in the risk-based capital.

  22. 22.

    IMF (2008).

  23. 23.

    These numbers are in constant 2005 dollars.

  24. 24.

    Liebenberg et al. (2010).

  25. 25.

    Baranoff and Sager (2009).

  26. 26.

    Publicly traded companies may disclose more information on private bonds to their investors in their SEC (Securities and Exchange Commission) reports, but the details are not standardised, and the disclosure level is up to those individual companies.

  27. 27.

    Kwan and Carleton (2010).

  28. 28.

    Moody’s AAA/AA/A and S&P AAA/AA/A correspond to an NAIC-1 rating. Moody’s BAA and S&P BBB and lower ratings correspond to an NAIC-2 rating and lower.

  29. 29.

    On 11 February 2009, the NAIC formed a working group to conduct a comprehensive evaluation of state insurance regulatory use of the credit ratings on structured securities and municipal bonds provided by nationally recognised rating agencies, including S&P, Moody’s, Fitch, DBRS, A.M. Best, and Realpoint. This working group recommends developing alternative methodologies for assessing structured security risks and reducing regulators’ reliance on credit ratings, which the group views to be less than reliable for various reasons. Additionally, the group recommends the development of standards, greater standardisation of definitions, greater consistency in the agreements used for structured securities, and so forth (Hunt, 2011).

  30. 30.

    These are firms traded on the NYSE, AMEX or NASDAQ with an SIC code identifying them as insurance companies (SIC code: 6310-6399, except for agencies and brokers). COMPUSTAT and CRSP have slightly different SIC code records. When either COMPUSTAT or CRSP does not report an insurance SIC code, we check business information manually from the 10-K report available in the SEC’s EDGAR database for that year.

  31. 31.

    COMPUSTAT and NAIC asset is not directly comparable because COMPUSTAT uses the generally accepted accounting principles (GAAP) asset, whereas NAIC data are based on the statutory accounting principles (SAP). The GAAP asset includes non-liquid and intangible assets such as tax credits, goodwill, supplies, etc. The GAAP asset is generally greater than the SAP asset. In addition, the NAIC asset only includes the U.S. insurance subsidiaries, whereas COMPUSTAT asset also includes international insurance operations. We exclude those firms whose COMPUSTAT asset is much greater than the NAIC asset, because this difference could be due to large non-insurance (underwriting) operations, despite the fact that these firms have insurance industry SIC codes, or due to significant foreign businesses for which we cannot observe the asset mixture. We also run our analyses using 30 per cent criteria instead of 10 per cent, and our main conclusions still hold. The results are available from the authors upon request.

  32. 32.

    We also use the I/B/E/S database to obtain the number of analysts when doing a robustness check with analyst dispersion.

  33. 33.

    We also divided the sample at 2005 or 2006, instead of 2004. No major qualitative difference is found between these cuts.

  34. 34.

    Harrington (2009).

  35. 35.

    As a robustness check, we also calculate an average spread in January and link this with the previous year-end’s firm characteristics and balance sheet data. Our main conclusions still hold.

  36. 36.

    Flannery et al. (2004).

  37. 37.

    See, George et al. (1991); Lin et al. (1995); Madhavan et al. (1997).

  38. 38.

    George et al. (1991).

  39. 39.

    Zhang et al. (2009).

  40. 40.

    Zhang et al. (2009); Flannery et al. (2004).

  41. 41.

    The estimation of the GKN adverse selection component of bid–ask spreads is shown in Appendix 1.

  42. 42.

    Flannery et al. (2004); Van Ness et al. (2001).

  43. 43.

    Because of the limited number of publicly traded insurers, we do not use analyst earnings forecast dispersion as a major information asymmetry measure because the dispersion can only be calculated when more than one analyst follows the firm. This reduces the sample size almost by 50 per cent.

  44. 44.

    We apply this rule to correct the so-called keypunch error. Other studies also use similar rules. See Flannery et al. (2004); Lin et al. (1995); Hasbrouck (1991); Madhaven et al. (1997); Chorida et al. (2001).

  45. 45.

    We could observe private pass-throughs and public pass-throughs separately, but we combine these two because about 90 per cent of insurers do not hold any private pass-throughs, and for those insurers who do, the holding amount is very small, averaging less than 0.1 per cent of assets.

  46. 46.

    Cummins et al. (2001); Kwan and Carleton (2010).

  47. 47.

    We follow the definition of Phillips et al. (1998), which defines the following as long-tail lines: farmowners multiple peril, homeowners multiple peril, commercial multiple peril, ocean marine, medical malpractice, workers’ compensation, other liability, product liability, auto liability, aircraft, international, and reinsurance.

  48. 48.

    Baranoff and Sager (2002).

  49. 49.

    Egginton et al. (2010).

  50. 50.

    Detailed information on the security lending activities of insurance companies and related reporting changes can be found at http://www.naic.org/capital_markets_archive/110708.htm.

  51. 51.

    Van Ness et al. (2001); Zhang et al. (2009).

  52. 52.

    Brennan and Subrahmanyam (1995); Lin et al. (1995).

  53. 53.

    Easley and O’Hara (1992); Harris and Raviv (1993).

  54. 54.

    Previous research sometimes includes the number of analysts in the bid–ask spread regression because more analysts may reduce the information asymmetry. Out of the concern that the number of analysts and firm size are highly correlated, we ran regressions with and without the number of analysts. Our main results hold in both conditions. Zhang et al. (2009) adopt a 3SLS regression because the number of analysts is also determined simultaneously with the bid–ask spreads and may cause the problem of endogeneity. We follow their practice and also run a 3SLS regression when including an analyst variable. The results are available from the authors upon request.

  55. 55.

    We greatly thank one referee for this suggestion.

  56. 56.

    The giant insurance company AIG may represent an outlier. Given the special circumstances of AIG, we conduct robust check analyses by removing AIG from the regression sample. This removal does not change the significance of the key variables.

  57. 57.

    Neal and Wheatley (1998).

  58. 58.

    Lee and Ready (1991).


  1. Bagehot, W. (1971) ‘The only game in town’, Financial Analysts Journal 27(2): 12–14.

  2. Baranoff, E.G. and Sager, T.W. (2002) ‘The relations among asset risk, product risk, and capital in the life insurance industry’, Journal of Banking & Finance 26(6): 1181–1197.

  3. Baranoff, E.G. and Sager, T.W. (2009) ‘The impact of mortgage backed securities on capital requirements of life insurers in the financial crisis of 2007–2008’, Geneva Papers on Risk and InsuranceIssues and Practice 34(1): 100–118.

  4. Brennan, M.J. and Subrahmanyam, A. (1995) ‘Investment analysis and price formation in securities markets’, Journal of Financial Economics 38(3): 361–381.

  5. Caprio, G., Demirgüç-Kunt, A. and Kane, E.J. (2010) ‘The 2007 meltdown in structured securitization—searching for lessons, not scapegoats’, The World Bank Research Observer 25(1): 125–155.

  6. Cheng, M., Dhaliwal, D.S. and Neamtiu, M. (2011) ‘Asset securitization, securitization recourse, and information uncertainty’, The Accounting Review 86(2): 541–568.

  7. Chordia, T., Roll, R. and Subrahmanyam, A. (2001) ‘Market liquidity and trading activity’, The Journal of Finance 56(2): 501–530.

  8. Cowley, A. and Cummins, J.D. (2005) ‘Securitization of life insurance assets and liabilities’, The Journal of Risk and Insurance 72(2): 193–226.

  9. Cummins, J.D., Phillips, R.D. and Smith, S.D. (2001) ‘Derivatives and corporate risk management: participation and volume decisions in the insurance industry’, The Journal of Risk and Insurance 68(1): 51–90.

  10. Dionne, G. 2009. Structured finance, risk management, and the recent financial crisis. Working paper, from http://ssrn.com/abstract=1488767.

  11. Easley D. and O’Hara, M. (1992) ‘Time and the process of security price adjustment’, The Journal of Finance 47(2): 577–605.

  12. Egginton, J.F., Hilliard, J.I., Liebenberg, A.P. and Liebenberg, I.A. (2010) ‘What effect did AIG’s bailout, and the preceding events, have on its competitors’, Risk Management and Insurance Review 13(2): 225–249.

  13. Flannery, M.J., Kwan, S.H. and Nimalendran, M. (2004) ‘Market evidence on the opaqueness of banking firms’ assets’, Journal of Financial Economics 71(3): 419–460.

  14. George, T.J., Kaul, G. and Nimalendran, M. (1991) ‘Estimation of the bid–ask spread and its components: a new approach’, Review of Financial Studies 4(4): 623–656.

  15. Glosten, L.R., and Milgrom, P.R. (1985) ‘Bid, ask and transaction prices in a specialist market with heterogeneously informed traders’, Journal of Financial Economics 14(1): 71–100.

  16. Gorton, G. (2009) ‘The subprime panic’, European Financial Management 15(1): 10–46.

  17. Harrington, S.E. (2009) ‘The financial crisis, systemic risk, and the future of insurance regulation’, The Journal of Risk and Insurance 76(4): 785–819.

  18. Harris, M. and Raviv, A. (1993) ‘Differences of opinion make a horse race’, The Review of Financial Studies 6(3): 473–506.

  19. Hasbrouck, J. (1991) ‘The summary informativeness of stock trades: An econometric analysis’, The Review of Financial Studies 4(3): 571–595.

  20. Hellwig, M.F. (2009) ‘Systemic risk in the financial sector: An analysis of the subprime-mortgage financial crisis’, De Economist 157(2): 129–207.

  21. Hunt, J.P. (2011) ‘Credit ratings in insurance regulation: The missing piece of financial reform’, Washington and Lee Law Review 68(4): 1667–1697.

  22. IMF (2008) Global Financial Stability Report: Containing Systemic Risks and Restoring Financial Soundness. Washington, DC: International Monetary Fund.

  23. Kwan, S.H., and Carleton, W.T. (2010) ‘Financial contracting and the choice between private placement and publicly offered bonds’, Journal of Money, Credit and Banking 42(5): 907–929.

  24. Kyle, A.S. (1985) ‘Continuous auctions and insider trading’, Econometrica 53(6): 1315–1335.

  25. Lee, C.M.C. and Ready, M.J. (1991) ‘Inferring trade direction from intraday data’, The Journal of Finance 46(2): 733–746.

  26. Liebenberg, A. P., Colquitt, L.L. and Hollans, H. (2010) ‘The determinants and potential effects of life insurer mortgage backed securities exposure’, Journal of Insurance Issues 33(1): 1–30.

  27. Lin, J.-C., Sanger, G.C. and Booth, G.G. (1995) ‘Trade size and components of the bid-ask spread’, The Review of Financial Studies 8(4): 1153–1183.

  28. Lucas, D.J., Goodman, L.S. and Fabozzi, F.J. (2006) Collateralized Debt Obligations: Structures and Analysis, 2nd edn. Hoboken, NJ: Wiley.

  29. Madhavan, A., Richardson, M. and Roomans, M. (1997) ‘Why do security prices change? A transaction-level analysis of NYSE stocks’, The Review of Financial Studies 10(4): 1035–1064.

  30. Manconi, A., Massa, M. and Yasuda, A. (2012) ‘The role of institutional investors in propagating the crisis of 2007–2008’, Journal of Financial Economics 104(3): 491–518.

  31. Mason, J.R., and Rosner, J. (2007) How Resilient are Mortgage Backed Securities to Collateralized Debt Obligation Market Disruptions?, working paper, from http://ssrn.com/abstract=1027472.

  32. Murphy, A. (2008) An Analysis of the Financial Crisis of 2008: Causes and Solutions, from http://ssrn.com/abstract=1295344.

  33. Neal, R. and Wheatley, S.M. (1998) ‘Adverse selection and bid–ask spreads: Evidence from closed-end funds’, Journal of Financial Markets 1(1): 121–149.

  34. Phillips, R.D., Cummins, J.D. and Allen, F. (1998) ‘Financial pricing of insurance in the multiple-line insurance company’, The Journal of Risk and Insurance 65(4): 597–636.

  35. Powers, H. (2012) ‘Insurance securitization: a ripe market?’, CPCU eJournal 65(10): 1–10.

  36. Ryan, S.G. (2008) ‘Accounting in and for the subprime crisis’, The Accounting Review 83(6): 1605–1638.

  37. Schwarcz, S.L. (2009) ‘Keynote address: Understanding the subprime financial crisis’, South Carolina Law Review 60(3): 549–572.

  38. Scott, K.E. and Taylor, J.B. (2009) ‘Why toxic assets are so hard to clean up’, The Wall Street Journal 254: A13

  39. Van Ness, B.F., Van Ness, R.A. and Warr, R.S. (2001) ‘How well do adverse selection components measure adverse selection?’, Financial Management 30(3): 77–98.

  40. Vink, D. and Thibeault, A.E. (2008) ‘ABS, MBS and CDO compared: An empirical analysis’, The Journal of Structured Finance 14(2): 27–45.

  41. Zhang, T., Cox, L.A. and Van Ness, R.A. (2009) ‘Adverse selection and the opaqueness of insurers’, The Journal of Risk and Insurance 76(2): 295–321.

Download references


Sojung C. Park appreciates support from the Institute of Management Research and the Institute of Finance and Banking at Seoul National University. Xiaoying Xie would like to thank the research grant support from the Center for Insurance Studies, Mihaylo College of Business and Economics, California State University, Fullerton. We appreciate the helpful comments from two anonymous referees and the editor.

Author information

Correspondence to Sojung Carol Park.

Appendix 1: Bid–ask spreads

Appendix 1: Bid–ask spreads

Effective spreads and quoted spreads

The definitions of effective spreads and quoted spreads are as follows. For the sake of this paper, we have used the first quarter average spreads:

$$ESPREAD_{t} = {\text{avg}}\left( {\left| {\frac{{P_{t} - MP_{t} }}{{MP_{t} }}} \right|} \right) ,$$

where P t is the actual traded price at time t and MP t is the quoted midpoint associated with trading at time t, and

$$QSPREAD_{t} = {\text{avg}}\left( {\frac{{Ask_{t} - Bid_{t} }}{{MP_{t} }}} \right),$$

where QSPREAD is the quoted spread, Ask t is the ask price, Bid t is the bid price, and MP t is the quoted midpoint at time t.

The George, Kaul, and Nimalendran (GKN) model

The GKN adverse selection component is defined as the market maker’s expected stock value revision resulting from the submission of an order. This model assumes a unit trading size and equal ex ante probability of trades at bid and ask prices and is modelled as follows:

$$P_{t} = M_{t} + \pi \left( {\frac{QSPREAD}{2}} \right)Q_{t} ,$$
$$M_{t} = E_{t} + M_{t - 1} + \left( {1 - \pi } \right)\left( {\frac{QSPREAD}{2}} \right)Q_{t} + U_{t} ,$$

where P t is the observed transaction price, M t is the true price, U t is an innovation term in true price, E t is the expected return between transaction at time t −1 and transaction at time t based on public information, and π is the order-processing component proportion of the quoted spread. Q t is an indicator of a buying or selling transaction; Q t  = 1 when it is buying, and Q t  = −1 when it is selling. Neal and WheatleyFootnote 57 relax the constant quoted spread condition and suggest a modified regression estimation method:

$$2RD_{t} = \pi_{0} + \pi_{1} \left( {QSPREAD_{t} Q_{it} - QSPREAD_{t - 1} Q_{it - 1} } \right) + \varepsilon_{t} ,$$

where RD t is the difference of the transaction price return \(R_{t}^{T}\) and the quoted mid-price return \(R_{t}^{Q}\). Q t is not observable, so Lee and ReadyFootnote 58 tick test is used to infer Q t . \(1 - \hat{\pi }_{1}\) is the estimated adverse selection component proportion of the quoted spread. Therefore, (\(1 - \hat{\pi }_{1}\)) \(\overline{QSPREAD}\) is used as the adverse selection component proportion of bid–ask spread in this study, where \(\overline{QSPREAD}\) is an average of quoted spreads.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Park, S.C., Lemaire, J. & Xie, X. The Opaqueness of Structured Bonds: Evidence from the U.S. Insurance Industry. Geneva Pap Risk Insur Issues Pract 41, 650–676 (2016). https://doi.org/10.1057/s41288-016-0021-4

Download citation


  • structured bonds
  • information asymmetry
  • insurance
  • bid–ask spread
  • financial crisis

JEL Code

  • G01
  • G22
  • G24
  • G28