LTG Symposium: response to industry comment

Simˈpōzēəm

Enrico PerottiDr Enrico Perotti, University of Amsterdam, DNB and CEPR I welcome the chance to discuss with industry participants the key valuation choices introduced in Solvency II. The insurance industry embraced the move to market-based accounting in the heady days of the credit boom. Since the crisis, higher risk premia and lower rates have led to a rush to seek arguments to weaken market valuation as envisioned in Solvency II. Let us start with the economic case for correcting market valuation. Financial prices may be at times excessively volatile, implying that they may, within some time frame, partially revert. Excessive volatility – not volatility per se! – creates a case for smoothing reserve adjustment or own funds volatility for intermediaries with long-term liabilities. Such a deviation from market valuation would clearly have to be monitored, and would need to be transparent. The economic cause of the problem calls naturally for an asset side adjustment, as a measure of temporary impairment. However, the strategy taken by the industry has been to lobby for a nonstandard valuation approach, which is little short of reckless. The Matching Adjustment originated, according to a Morgan Stanley-Oliver Wyman report1 published this year, as an obscure UK accounting practice. It introduces an accounting adjustment to discount liability values to reduce required capital reserves technical provisions for insurers with long-term obligations. The argument is that illiquid liabilities are worth more (to insurers), so they  should be discounted to reflect the value they create for the insurance companies. The consensus among academics is (remarkably) unanimous: arbitrarily reducing the stated value of liabilities has no economic foundation whatsoever. A limited use of the Matching Adjustment in narrowly defined matched portfolio would do limited damage, as in such very specific cases the reserve or own funds calculations over a long time frame would not differ much. However, the procedure is flawed, and its use creates a dangerous precedent. Indeed, the latest proposed Solvency II adjustments further extend the use of discounting liabilities well beyond the limited context in which it was justified. The three contributions to this forum differ in their views, with the (personal) opinion from the ECB policymakers clearly more critical. The industry contributions offer some intuitive-sounding sentences defending the adjustment to liability value. The reasoning misses the springboard by a foot. The contribution by Insurance Europe discusses examples under the view states: “if the existing market-consistent framework methodology does not work for insurance business, it needs to be improved and adapted”. I would like to think that by “what works for insurance business” they mean what ensures safety of insurance promises that supports confidence in insurers. I also presume their argument is that insurers provide “risk mitigation” though matched portfolios. The key issue is then, how much do the matched portfolios eliminate excess volatility? The list of assumptions under which they do is impressively long (no redeemability, matched cash flows, no longevity risk, no missing markets for long assets, all credit risk fully provisioned, all spread volatility is fully resolved if assets are held to maturity, etc). So I agree very much with their leading sentence: “We should also not forget that market consistency is a theoretical framework designed to model how the real world works. As with all theoretical frameworks, in trying to model that world it makes a number of simplifying assumptions. It is vital to check whether these assumptions are fit for purpose and — if not — to make improvements.” I therefore, would welcome a public study by EIOPA on the underlying assumptions of the matched portfolio approach. But given this view it follows that the authors from Insurance Europe cannot support extending liability discounting to claims not in matched portfolio, since there is no risk mitigation there. The consequences of poor accounting choices hurt the industry as well as policyholders. A much cited University of Chicago study2 proves the point. The study shows how US life insurers “sold long-term insurance policies at deep discounts relative to actuarial value. In January 2009, the average markup was -25 percent for 30-year term annuities as well as life annuities and -52 percent for universal life insurance.” The findings are instructive for Solvency II. The paper goes on to state, “This extraordinary pricing behavior was a consequence of financial frictions and statutory reserve regulation that allowed life insurers to record far less than a dollar of reserve per dollar of future insurance liability.” Mr. Creedon from Groupe Consultatif shares the need for transitional arrangements, but he insists in justifying liability discounting as a structural measure. He argues that a long-term irredeemable liability is a great convenience for insurers, and is therefore worth more than their present value to them. I agree with the first part of this argument, but not the second. Insurers do indeed benefit from having long-term liabilities because this grants them a comparative advantage (relative to other intermediaries) to invest in long-term illiquid assets, which earn a liquidity premium. The gain is on the asset side, and has nothing to do with the value of their obligations, which are unconditional and thus should not bear any risk discounting (surely not for the sake of computing prudential ratios). It makes no sense for the insurer to claim a gain on the asset AND the liability side. Yes, being an insurer implies having the ability to invest for long-term value, but this is fair only on condition that the insurers properly backs the long-term claims for which policyholders have accepted nonredeemability. This is the whole point of insurance regulation, is it not? Mr Creedon believes in solutions which, “calibrate the adjustment by reference to a synthetic portfolio”. Solutions to reduce capital ratios based on risk-less hedges assumptions are well known from Basel II. The underlying assumptions proved tragically faulty, leading to a new Basel III approach. While Mr Creedon states, “actuaries are usually pragmatic”, I am sure he believes in building on solid foundations. The contribution of the authors from the ESRB Secretariat is most rigorous in its approach. Its call for counter-cyclical buffers in good and bad times correctly identifies how part of market valuation losses may be temporary and thus be allowed to be recovered over some time, rather than hidden from sight. I believe the contributors should go a step further and state clearly that even if matched portfolios may be tolerated as a very limited measure, liability discounting is still not justified as a principle, and surely not for unmatched liabilities. I will be happy to discuss this issue further. Balance sheet valuation in a market consistent regime is not a theoretical issue built on abstractions. Accounting choices need to be based on sound, general economic principles. Even for matched portfolios which satisfy all the many assumptions, the approach simply produces the same numerical value for the capital requirement, which is zero (except for the credit risk reserve). The reason is simple: by assumption, all other asset and liability risks vanish when assets are held to maturity. In banking, this is equivalent to claim a zero capital risk weight for mortgage back securities matched with a Credit Default Swap claim. But even if this mathematical equivalence may hold for some combinations of assets and liabilities with unique characteristics, it does not establish an economic principle justifying discounting unconditional promises for prudential regulation, ever. And clearly, it is impossible to extend the use of discounting for unmatched liabilities. The proper and transparent solution is a counter-cyclical buffer and a ladder of intervention graduated on the degree of long term liabilities. Accounting principles reflect economic principles, not rhetorical exercises. Whatever the underlying motivation, an accounting fiction to report lower liabilities than they really are is a misrepresentation. There are straightforward measures to deal with the issue of excess volatility in insurer reserves (not volatility per se, which reflects economic reality), built upon transparent economic foundations. A revaluation reserve is the natural adjustment to excess volatility in prices, both in booms and crises. Capital reserves should be set aside in good times, when risk premia may be excessively low, in order to be used in distressed times. This ensures that capital is raised when less expensive, and immediately available in bad times. A transitional regime would enable insurers with longer maturity of liabilities more time to adjust to phases of exceptional distress, reflecting their greater inter temporal risk bearing capacity. The time to adjustment should be reassessed regularly. The proper calibration of any corrective tool to market valuation is a challenge. Liability discounting is an arbitrary and opaque accounting adjustment, a dangerous channel to regulatory capture and forbearance. A revaluation reserve is an economically sound and transparent measure to adjust insurer capital ratios. This is the road Solvency II should take. The views expressed are the author’s own.

References 

1 <  Morgan Stanley and Oliver Wyman (2012), “Solvency 2. The Long and Winding Road”, report, March 2012 2  The Cost of Financial Frictions for Life Insurers, Ralph S.J. Koijen, Motohiro Yogo, NBER Working Paper No. 18321, Issued in August 2012 Link to Symposium index page.]]>