X hits on this document

PDF document

DFA Insurance Company Case Study, Part I: - page 9 / 40





9 / 40

artificial ones introduced by the fictitious public data created for DFAIC. Artificial data problems are those leading to questions that could be resolved with additional public information7. The information provided for DFAIC is more reflective of the limitations one might experience dealing with non-U.S, companies, where publicly available data is much more limited.

Despite the fact that this DFAIC case study does not require a formal data collection process, there are still the important chores of data evaluation and reconciliation of the DFAIC data. We have undertaken such processes and found a number of areas where additional data and research would be warranted. Since no additional research could be undertaken, we will note a few of these irregularities here and then make reasonable assumptions to correct for these inconsistencies.

A solid data evaluation process requires the reconciliation of the provided information to the company's last reported financial statements. In putting together this reconciliation, we found that the investment income earned on the portfolio is inconsistent with the stated asset allocation. The asset allocation for DFAIC is purported to be 70% fixed- income, four-fifths of which is in tax-exempt bonds. DFAIC's large allocation to tax- exempt bonds, given the lower yields on these securities, is inconsistent with its reported above-average investment income (7% versus 5% industry averageS). In order to reconcile the stated asset allocation to the reported investment income, the tax-exempt bonds would have to be earning a book yield in excess of 7.5%. Since market yields on tax-exempt municipal bonds were in the range of 3%-6% during 1999, the tax-exempt holdings of DFAIC would have to contain a large unrealized capital gain. However, the reported book and market values on these holdings are very close to each other: $3,327M and $3,478M, respectively, it is unlikely that the tax-exempt bonds could be providing such a contribution to investment income.

This inconsistency raises some serious concerns regarding data quality of the DFAIC asset portfolio. Examination of Schedule D reveals that over half of the bond holdings of DFAIC are classified in the "Special Rev & Assessment" category. This category typically contains taxable structured bonds (mortgage-backed and asset-backed securities). Since tax-exempt holdings are not specifically categorized in Schedule D, it is likely that some or all of the bond holdings of DFAIC that were reported to be tax- exempt bonds are actually taxable bonds.

Further evidence to support this theory lies in the fact that if all bonds were assumed to be taxable, the calculated investment income would reconcile with the reported investment income results. If complete statutory records for DFAIC were available, analysis of the Schedule D details would resolve any doubt about this inconsistency. Since no such details were available, professional judgment must guide us on how we should model DFAIC's current asset portfolio. Thus, given that the investment income, and book and market values of the asset portfolio can be directly traced to the provided

Because DFAIC is not a real company and it would have been impractical for the CAS to provide a complete DFAIC annual statement, there is no such additional information (e.g., prior years' annual statements, annual statement schedules).

e 1999 industry excluding state funds investment yield. "Best's Aggregates & Averages, Property- Casualty U.S.", 2000 Edition, p. 117.


Document info
Document views65
Page views65
Page last viewedThu Oct 27 15:43:48 UTC 2016