Insurance Contracts Project: User Feedback

Transcription

1 Summary Overview Several months into redeliberations, the staff has gathered input from the user community on the tentative decisions made to date and sought feedback on the presentation alternatives brought to the Boards in March. On March 31, 2011, a targeted was sent to two dozen analysts and investors. Since that date, the staff has received feedback from those and other U.S. and global buy-and-sell-side analysts and others that focus on life, (property and casualty, or both. General Feedback The majority of users were concerned about the proposals for insurance contracts relating to volatility, presentation, and comparability. Some users also were concerned about due process and urged the FASB to continue to work toward a converged standard with the IASB, but emphasized the importance of field testing. Volatility Users were significantly concerned with the volatility in insurers financial statements that may result from the proposals. Most users were focused on the discount rate and, more specifically, how changes in the discount rate should be recognized in the financial statements. Volatility caused by changes in the market discount rate is not thought to be fully economic due to the long-duration nature of liabilities with relative stability in predicting payouts. Many users think that the income statement would be overwhelmed by including changes in market discount rates rather than providing a true indication that the underlying economics of the business (mortality, surrenders, lapses, etc.) have changed. Many users calculate an insurer s normalized earnings for their analyses and noted that they would need to adjust for changes in the insurance liabilities resulting from changes in the discount rate if those changes are recognized in current period earnings. For the majority of investors, which cover the financial services sector in general, the increased volatility and complexity compared to banks would deter them from investing in the insurance space. While many users agree that a current measure of the insurance contract liabilities would be useful, those users do not want changes in that amount resulting from a change in the discount rate to be recognized in current period earnings. Some users suggested locking in the discount rate but updating the measurement for other changes in assumptions. Other users suggested the use of other comprehensive income for changes in the discount rate so that the insurance contract liabilities would still be measured at a current rate, which would avoid what many users see as unnecessary noise in earnings. Some users supported presenting changes in cash flow estimates in earnings. However, most of those users did not distinguish between experience adjustments (backward looking) and changes in expectations (forward looking). One user preferred that (1) experience adjustments be recognized in earnings in the period incurred and (2) changes in expectations be used to offset or remeasure the margin to reduce the potential for an insurer to recognize large losses followed by subsequent income on the same portfolio. Page 1

2 Presentation Investors generally preferred current presentation and stressed the importance of volume information (premiums, investment income, benefits, losses incurred, etc.) on the primary financial statements; although some thought that the separation of underwriting results from investing results would be beneficial in their analyses. For those users, the current presentation appears intuitive and allows for more comparability with other industries. Generalists rely on the income statement and would not be able to perform deeper analyses, which could result in outflows of capital from the industry. Some users also noted that portfolio managers are not typically specialists, which could result in decreased interest in the industry. Other users indicated that margin information would be useful in the footnote disclosures. A few users supported a summarized margin approach but only if information on premiums, cash flows, and source of earnings were separately disclosed. Comparability Many users noted that the proposals would further isolate the insurance industry and create increased comparability issues with other sectors. Many investors think that this could lead to capital allocation moving away from the insurance sector. Users also were generally concerned about the comparability of insurers within the sector given the number of assumptions insurers would need to make under the proposals. Some users suggested disclosures to mitigate those concerns while others doubted insurers would be willing to provide disclosures that would be detailed enough to make them comparable and useful for investors. Many users questioned the usefulness and comparability of a risk adjustment. Some users suggested that if a risk adjustment is necessary, it should be disclosed instead of included in the measurement of insurance contract liabilities. A few users questioned whether disclosures would be comparable because of each insurer s individual aversion to risk in calculating a risk adjustment. Due Process Some users were concerned that the new guidance for accounting for insurance contracts is going to rely heavily on disclosures to resolve any issues that were not sufficiently deliberated. Many users suggested that the Boards allow sufficient time for field testing given the potential breadth of changes for insurers. P&C Many of the investors do not think that current P&C accounting is broken. They stated that P&C and life insurance are fundamentally different and that a single model is not necessary. Many users noted that the information they receive for nonlife contracts from insurers financial statements prepared under current U.S. generally accepted accounting principles (GAAP) is useful. Most users do not think that insurers should discount the preclaims liability. Additionally, many users do not think that insurers should discount the claims liability. Those users noted that discounting loss reserves adds additional complexity. Some users mentioned that discounting claim liabilities may have academic or conceptual merit, but would not be practical. Other Comments One user noted that an improvement could be made in the disclosure of loss triangles by insurers. That user also noted that the loss triangles for some multinational insurers are useless because of the effect of foreign currency adjustments. Page 2

3 Several users noted that it is important for the accounting for insurers to be consistent with the accounting for banks. Those users noted that the asset-liability management of insurers is similar to that of banks and that similar accounting would allow for investors to analyze both spaces together. Page 3

4 User Comment Details Insurance Contracts Project: User Feedback Discount Rate and Volatility Outreach with Small User Groups Low underlying volatility in the industry today Not much in individual life; a bit more in group benefit; variable annuity is the most volatile but the accounting has more underlying volatility than economics. The proposal will cause income statement volatility that will not be representative of normalized earnings. Risk-free discount rate is subject to less manipulation and is more representative (difference between asset/liability is risk premium). Generalists rely on the income statement and will not be able to perform deeper analyses, resulting in possible outflow of capital from the industry. The Canadian experience was cited as an example (it became less investable because of how volatile it became). There is some support for the proposals because of current mismatch as evidenced by the use of book value-based metrics. (P&C) Recognition of changes in interest rates/margins in the income statement will add more volatility and will need to be backed out to get true operating performance. Increased volatility is not an improvement and will lead investors to look elsewhere for places to put their money (which will lower the value of nonlife insurers). Asset/liability matching may be very important for life insurers, but it s not the primary concern for nonlife insurers. (P&C) Locked-in discount rate with unlocked nonfinancial assumptions: o I support the option of a mixed measurement model that would allow locking the initial discount rate for liabilities where management of assets and liabilities is closely linked. I believe this approach is consistent with the actual business model used for many life insurance products. o Merely moving the impact of discount rate changes from the P&L to OCI, as proposed by some, is not, in my view, a sufficient compromise, since it would still create significant differences and greater volatility vs. bank accounting. o I believe that non-interest assumptions should be maintained on a current best estimate basis and that changes in these assumptions should be reflected each period in the P&L. Issues with construction of discount rate: o Risk free plus liquidity has no bearing on the economics of the pricing and will likely produce day one losses that are not economically relevant. Rate should be closer to asset earned rate and the long-term yield assumption should be unlocked. Page 4

5 o Substantial changes in market rates should drive changes in lapse assumptions. Following a dynamic lapse model would better portray the true economic impact. Similar to how variable annuity companies currently model deferred acquisition costs. Discount rate proposals may represent more of a black box than current U.S. GAAP. The general account portfolio yield is something that can be empirically observed. Suggestion to include any mark-to-market changes in Accumulated Other Comprehensive Income to remain consistent with treatment of other noneconomic effects (pensions, foreign currency, etc.). Life insurance is not a volatile industry when properly managed, and changes in discount rate should not move when liabilities are properly matched. If you increase volatility while expecting to see stability, you will lose the generalist investor. Discount rate changes would lead to unintended consequences, citing Canadian insurers who experienced significant volatility under current U.S. GAAP when yields plummeted during crisis. Discounting in P&C is more interesting philosophically than it is practical. Project will increase volatility, particularly for life insurance. Marking liabilities to current value will not offset the mismatch with the assets backing the liabilities; however, this may not be apparent to the users. Concerned about unlocking the discount rates on a frequent basis. Project introduces volatility into what some investors view as a fairly steady business. Changes in current period assumptions (experience) other than the discount rate should be recognized in current period earnings and other changes in assumptions (forward looking) should be used to remeasure the margin. Tentatively prefer the use of (Other Comprehensive Income) for changes in insurance liabilities resulting from changes in the discount rate. Change in the discount rate should not flow through the income statement. Changes in the discount rate only should flow to OCI, not the profit and loss. Changes to the liability cash flows should flow through the P&L as proposed. Conferences and Generalists focus on results, therefore, the additional volatility generated will cause insurance to become less investable Recognizing changes in discount rate in P&L is lunacy. Page 5

6 Suggestion to use a rolling three-year average for the discount rate, similar to some embedded value calculations. Disclose adjustments in discount rate construction and sensitivity analysis. and Proposal will cause too much volatility. Some would prefer to see discount rate anchored to something market related so that there is very little discretion around what the rate would be. o Insurers would have very different liquidity premiums. Updating assumptions would be useful (for morbidity/morality) but disclosure also would be necessary to indicate where capital is needed. o Does not think Canadian GAAP provides useful information for life insurers. Suggestion to use a 20-year average discount rate that is not updated quarterly (maybe yearly). Suggestion to only unlock discount rate for long-term movements (liability adequacy test). If assets are recorded at amortized cost or fair value other comprehensive income, prefers change in liabilities due to change in discount rate to be in OCI. o That is, put everything unrealized in OCI. o Is okay with the fact that there would still be a credit difference. Discount rate: o If use a top-down approach, insurers may have different rates depending on whether they start from a high-quality fixed income portfolio or a junk bond portfolio. o Does not want investment decisions to impact the discount rate of the liability. Presentation of Volume Information and Performance Metrics Outreach with Small User Groups Use supplemental information included with earnings release. Do not rely heavily on statutory filings due to delay (blanks filed last day of February; statutory audited financial statements not due in many states until June 1). Companies may no longer qualify for Fortune 500 if income statement uses the summarized margin approach. Presenting the face of the income statement with a margin-based approach instead of traditional revenue would make it difficult to draw comparisons with other industries. Deferred acquisition costs (DAC) represent about 70 percent of current U.S. GAAP book value; if it is not part of new proposal, there should be a gross premium approach 1 that reduces the reserve liability. 1 Current U.S. GAAP requires the liability for future policy benefits to be measured as the present value of estimated future policy benefits to be paid to or on behalf of policyholders and related expenses less the present value of estimated future net premiums to be collected from policyholders. The net premiums represent the portion of the gross premium required to provide for all benefits and expenses or the portion of the gross premium required to support a reserve, adjusted or statutory. Presumably, a gross premium approach would replace net premiums with gross premiums. Page 6

7 Summarized margin approach provides less transparent information and it is unclear how many past years of financials will be prepared under new guidance. Important performance metrics mentioned included accident year, benefit ratio, investment income analysis, and non-gaap numbers. o Relies on statutory accounting schedules, 10-K investment portfolio, and statistical supplements when performing analysis. o Would like to see underwriting margin broken out. o Does not rely on embedded value. o Does not want to lose volume information in income statement (which would happen if use summarized margin approach). o Prefers to receive metrics that increase comparability within and outside the industry. Prefers current U.S. GAAP presentation (premium, benefits, loss incurred rate, etc.), whereas proposal seems very different from reporting standards of other companies (even banking). Thinks it will make it harder for portfolio managers, who typically are not specialists. May not get as much interest from those that are not focused solely on insurance sector. Volume information is crucial: o I am against an insurance company income statement that does not include premiums, claims, and benefits. o Expanded margin approach Better presentation than the Statement disclosure approach. However, when I read because it may not be clear what the numbers and captions represent. In particular, some may consider at least one of the numbers to be a plug and another comment, revenue was a number derived from expected or actual expenses rather than directly determined, I don't care whether the presentation is okay if the underlying accounting standard is based on this flawed logic. o Segmentation between new business written contracts and existing contracts provides investors with new, more understandable information. o Dual Statement Approach. My choice for best presentation. If you provide investors with Appendix A-6 (from Memo 59H/Agenda Paper 2H) and then included the Existing Contracts/Underwriting results section from Appendix-5 in a footnote, you would provide investors with an appropriate income statement and significant new information about an insurance company's asset-liability risks. Favor summarized margin approach out of all presentation approaches for building-block contracts. o Cash flows for new business written information in Appendix A-5 example are useful as supplemental disclosures. Page 7

8 o o Prefer current U.S. GAAP revenue recognition for short-duration contracts with disclosure of premiums received for long-duration contracts. Proposals will cause shift toward operating earnings metric to increase; prescribing a method for comparability is logical. It is important to show a traditional income statement with volume information; supports splitting underwriting from investing components. Credit Analyst Source of earnings approach is consistent in Canada due to small number of participants and extensive collaboration between actuaries and accountants; would not be as comparable in a global setting. Creating a revenue number and providing useful performance information is difficult: o Because revenue recognition is such a farce and subject to manipulation, I think either (or both) providing the cash for the year (which would seem to approximate premium due for a year in most cases maybe separate single premiums) or/and providing a written number (that is, what was actually agreed to in a particular year knowing that single premiums can completely skew the number but obviously that is something we would want to know and the reason for the annual premium equivalent measure). The juxtaposition of the cash number and the written number would provide evidence of timing (another useful question that could be asked). Conferences and Losing a lot of information with the margin presentation. Would at least like to see premiums because insurer actually received cash for that amount. Would like to use premiums/deposits to see growth without going to the footnotes. Gross information is needed for analysts to be able to forecast future performance. Risk Margin Outreach with Small User Groups Comparability concerns will be further exacerbated if risk adjustment is used; further noted risk may be entity specific and different depending on each insurer s aversion to risk. The only way to mitigate comparability concerns may be through disclosure of assumptions. Sees merit of understanding risk margin numbers (so you can almost perform a stress test), but does not like mixing capital requirements with what should be the best view of underlying economics of the business. Would like to see the information in disclosures. Introducing margins and discount rates for short-duration contracts will make it difficult for investors to reconcile financial reporting with how they think about those businesses. Page 8

9 Supports dual-margin approach with conditional tail expectation as the only option for risk adjustment; allowing for three methods will reduce comparability while the other two methods lack the robustness of the conditional tail approach. Does not understand the risk margins due to little data available. Concerned with comparability on Day 1 and subsequent measurement. Conferences and Risk margin turns accountants into regulators and does not reflect economics of the business. Buy-side and sell-side risk adjustments will lead to false precision. Not useful or comparable. Some users had no idea what was meant by a risk adjustment, what it was supposed to represent, or how it would be useful. General Concerns Outreach with Small User Groups Discounting nonlife reserves adds another layer of unnecessary complexity that will need to be adjusted for analysis. Lack of investor support: o I was also surprised to hear that IASB and FASB have been operating under the assumption that their proposals have the backing of analysts and investors. From my little circle of colleagues, nothing could be further from the truth. I haven t come across a single person, investor, or analyst that is supportive of the changes (at least as they relate to nonlife insurers). Nonlife proposals unnecessary: o My initial reaction when I first heard of what IASB and FASB were working on was U.S. GAAP seems to be working pretty well for nonlife insurers and is widely used globally, so why do we need such a radical change when it seems like we could just tweak the current accounting system if necessary? I was frankly stunned when I saw the current proposals and how far along they are in the approval process. o Life and nonlife are different businesses and should not be forced into the same accounting structure. Back testing against historical trends to recognize strengths and weaknesses in new guidance is necessary. Concerned that IASB finishing before FASB will make it difficult for FASB to adopt divergent approaches. Page 9

10 Consistency with banks: o Some would value greater global cross-border consistency, but all would value consistency between U.S. insurers and banks. o It is surprising and unfortunate, in my view, that FASB and IASB have proposed establishing a significant divergence between banks, with a system of cost-based accounting, and insurers, with a system of fair value accounting driven by changes in discount rates. In my view, this inconsistency, if established, would significantly diminish investors ability to make reliable comparisons and rationally allocate capital among financial institutions, particularly during times of market volatility. Create DAC comparability and transparency without focus on narrowing breadth of costs: o A narrow definition will result in a poor matching of expenses and associated revenues, potentially understating profits for rapidly growing firms while overstating profitability for slower growing or shrinking firms. Investors often have confusion and frustration related to current DAC accounting. However, this confusion and frustration relates to issues of consistency and transparency, rather than the size or breadth of acquisition cost. Take time to fully develop proposals that do not heavily rely on disclosures. Many take threat to industry cost of capital seriously, and volatility needs to be reconsidered. Accounting changes will impact capital markets: o Majority of investors in life insurance stocks are not experts in the industry or accounting. o Lack of comparability to banks will limit capital available, which could translate to higher insurance costs for customers. o The recurring theme I hear is that as long as the noneconomic elements of these proposals are isolated to AOCI and companies give disclosure that is consistent with current financial supplements, the changes are something they can work through. o To be clear, I believe these new accounting changes will turn off marginal investors in the space because they will not want to learn the new rules or deal with the uncertainty. But, if the changes can be accommodated by the Boards, I think the disruption can be significantly mitigated. DAC is a true asset. Pending guidance leads to inconsistency between direct marketing and agentacquired contracts. Investors are concerned because of the possibility of change and preparers are telling them that they don t want the changes being proposed. When the underlying products are difficult to understand, the accounting can be difficult to understand as well. Recognized life insurance accounting was not perfect, but doesn t think it needs entire overhaul. Page 10

11 Doesn t see anything wrong with P&C accounting. Of 100 investors at large insurer investor day, he felt only 20 to 30 were insurance specialists and the rest did not rely on details heavily; this made him think increased cost of capital to the industry from accounting changes and complications is a real possibility. His view is that any time you increase volatility while expecting to see stability, you will lose the generalist investors. Does not have preferences regarding DAC and margins as long as results are comparable. Current P&C accounting is very useful; do not discount preclaims liability. Conferences and One user commented that as a result of the proposals, insurance accounting would go from a black box to a black hole. That user thinks that the building blocks are more complex than current U.S. GAAP. Too many different assumptions in the proposals. Some users were primarily concerned with the potential effect on book values and earnings of insurers upon transition to the new guidance. Does not understand the insurers that currently are using IFRS and does not think that proposals will alleviate this issue: o Thinks that transparency is more important than convergence. Does not want to converge to something that is more flawed than current U.S. GAAP. Under the proposals, would like to see the following disclosed: o Discount rate for reserves o Liquidity premium o Discount rate for new business (pricing) o Equity return assumptions o Surrender assumptions. Page 11