While much attention has been paid to the impact of COVID-19 on banks (for examples of prominent early work, see Beck 2020, Cecchetti and Schoenholtz 2020), insurers have received less attention. Insurers are important financial intermediaries, holding more than $30 trillion in assets globally (FSB 2020), and operating on a comparable scale to banks in some countries.
COVID-19 will affect insurers both directly, via health shocks (increases in mortality and morbidity), and indirectly, via financial shocks (higher credit spreads, potentially widespread downgrades, lower interest rates, and lower equity prices). The financial impact of pandemics is often much larger than the direct health impact, as even comparatively mild outbreaks can have a large economic footprint (McKibbin and Fernando 2020).
Insurers’ exposure to COVID-19 varies by their line of business. For example, life insurers’ liabilities commit them to payouts decades ahead, and are very sensitive to changes in interest rates.1 Although these long-term liabilities permit longer investment horizons, potentially supporting financial stability (Chodorow-Reich et al. 2020, IAIS 2019), life insurers are more directly affected by pandemics through their liabilities. If insurers are hit hard, risk appetite may be impaired—as was the case following the global financial crisis (Kirti 2019)—affecting the availability of credit for riskier borrowers. Policymakers working to preserve credit supply to the economy should account for any changes to insurers’ risk appetite.
Impact on US life insurers
We present estimates of the impact of COVID-19 on capitalisation levels for US life insurers, considering both the direct pandemic shock and the indirect financial shock.
Life insurers provide protection against both mortality and longevity risk. Joint provision of both types of protection can help insurers manage risk. In a pandemic with significant mortality impact, life insurance claims rise immediately, whereas expected future payments on life-contingent annuities fall. These risks can therefore offset each other from a long-term solvency perspective, although large life insurance claims may have a meaningful short-term impact. Moreover, for annuities in the accumulation phase, as is the case for a majority of US annuities, account values are typically payable upon death.2
We focus on the impact of potentially larger mortality on life insurance claims. We base key assumptions (share of population with coverage and average size of policy by age group, and the extent to which mortality may be lower conditional on coverage) on SOA (2007) and apply case fatality rates by age group as experienced in Korea (a more benign scenario) or Italy (a more severe scenario with limited health care capacity) as of early June.3
Although the pandemic appears to be tapering in Europe and even in the US as of early June 2020, it is difficult to exclude further waves of infections. Figure 1 shows the impact on US life insurers as a share of capital given the share of the population affected by COVID-19. Insurers, regulators, and industry observers typically use an excess death ratio of 1.5 per 1,000 as a standard pandemic stress test scenario. In the US, this would imply about 500,000 deaths (about 8% of the population affected with case fatality rates in Korea, and about 4% with case fatality rates in Italy), larger than current estimates of the potential death toll released by US authorities. In such a scenario, losses would be about $15 billion, or about 3% of pre-shock industry capital.4 Should the pandemic prove challenging to control (with treatment and vaccines remaining elusive), an extreme scenario more comparable to the 1918 Spanish flu (closer to 30% of the population affected with case fatality rates in Korea) could become relevant. In such a scenario, losses would be about $45-55 billion—about 10% of pre-shock capital. Such a scenario would likely unfold over multiple years, giving insurers time to accumulate additional net premium income.
Figure 1 Impact of potential mortality claims on US life insurer regulatory capital
Sources: S&P Global Market Intelligence and IMF staff analysis. Note: This figure shows the US life insurance industry’s aggregate capital net of pre-tax excess claims as a function of the share of population infected. Estimates are based on the age-group-specific number of policyholders and average face amount in force used by SOA (2007), adjusted to match the current US population and stock of life insurance in force net of reinsurance as of end 2019 (close to $20 trillion). Case fatality rates by age group as of 3 June 2020 are from the Italian National Institute of Health and the Korean Ministry of Health and Welfare. Case fatality rates are assumed to be 23% lower for the insured population as in the severe scenario used by SOA (2007).
US life insurers are generally not required to mark bonds to market for the purpose of determining regulatory capital.5 Capital requirements for US insurers are tied to regulatory risk categorisation of assets. For most asset classes, regulatory risk categories are a function of credit ratings.6 US life insurers’ holdings are somewhat skewed toward bonds only one or two rating notches above regulatory thresholds for significantly higher risk weights (Kirti and Shin 2020, Becker and Ivashina 2015).[vii] Risk weights for junk bonds are more than triple those for investment grade bonds. Downgraded bonds tend to be sold by insurers facing capital pressure (Ellul et al. 2011).7 We focus on the impact of bond downgrades on required capital.8
Despite unprecedented policy responses from monetary and fiscal authorities, many corporate issuers may face rating downgrades due to the widespread economic fallout from COVID-19; the extent to which some sectors will face serious solvency concerns in any ‘new normal’ is not yet clear. If downgrades move assets into higher-risk regulatory categories, capital requirements would increase. Downgraded bonds may see fire sales, particularly if downgrades are widespread, although selling distressed assets would potentially trigger recognition of losses that permanently impair regulatory capital for US insurers, which would disincentivise fire sales.
We use bond-level holdings data for all US life insurers as of the fourth quarter of 2019, along with bond ratings (and downgrades) as of end-May. More than $60 billion in bond holdings have already been downgraded across categories, while $210 billion have been downgraded at least one rating notch. We consider two further scenarios: a moderate scenario in which 50% of bonds (by value) one notch away from regulatory thresholds are downgraded into the next category, along with 25% of bonds two notches away; and a severe scenario with double these downgrade rates.9
Figure 2 shows how risk-based capital (RBC) ratios (the ratio of actual capital held to required capital) would be affected under different scenarios. US insurers must submit corrective action plans if their RBC ratio is below 150% and has a negative trend, or falls below 100%, and can be placed under regulatory control if their RBC ratio falls below 50%.10 The median RBC ratio is above 400%, with some insurers already below 200%. The severe downgrade scenario would significantly lower RBC ratios and bring some insurers close to or below company action thresholds.
Figure 2 Impact of potential bond downgrades on US life insurer RBC ratios
Sources: S&P Global Market Intelligence and IMF staff analysis.
Note: This figure shows the distribution of RBC ratios for (1) the current level (as reported in 2019), (2) the actual level (accounting for actual downgrades to date), (3) the moderate scenario (half of one-rating-notch-above and a quarter of two-rating-notches-above corporate bonds are downgraded into the next NAIC category), and (4) the severe scenario (all of one-rating-notch-above and half of two-rating-notches-above corporate bonds are downgraded into the next NAIC category). The RBC ratio is total adjusted capital divided by two times authorised control level (required) capital. The figure is restricted to insurers with assets of more than $5 billion. Observations outside the 10th and 90th percentiles are not shown.
Joint impact of health and financial shocks
Figure 3 illustrates the joint impact of the health and financial shocks. Panel A shows an indicative aggregate RBC ratio in three increasingly severe scenarios in addition to the pre-shock level: (1) the upper end of a mortality scenario presented by US authorities in late March (240,000 deaths) along with rating downgrades as of end May; (2) deaths in a standard pandemic stress test scenario (1.5 excess deaths per 1,000, or about 500,000 deaths) combined with the moderate financial shock scenario; and (3) deaths in line with Spanish flu (using excess deaths in the US reported by Barro et al. 2020) combined with the severe financial shock scenario (1.7 million deaths).
Panel B shows the impact for the largest US life insurers by net life insurance in force for the same set of scenarios. Under severe joint shock scenarios, capital levels for several large insurers would approach, and in the case of one insurer breach, regulatory thresholds.
Figure 3 Joint impact of health and financial shock on US life insurer RBC ratios
Panel A Aggregate
Panel B Large life insurers
Sources: S&P Global Market Intelligence and IMF staff analysis.
Note: Panel A shows the current aggregate risk-based capital (RBC) ratio and the joint impact under three scenarios: (1) US authorities’ upper end March estimate of 240,000 deaths combined with the actual downgrades in 2020; (2) excess deaths of 1.5 per 1,000 as in a standard pandemic stress test scenario; (3) excess deaths of 5.2 per 1,000 as experienced in the US during the Spanish flu (Barro et al. 2020) combined with either the moderate or the severe financial shock scenario. For each scenario, age-group-specific case fatality rates in Korea are used to convert excess deaths to the share of population infected, assuming infection rates are constant across age groups. The share of population that would need to be affected to reach these scenarios is about 4%, 8%, and 28%, respectively. Panel B shows RBC ratios for the largest life insurers by net life insurance in force with assets of more than $50 billion for the same scenarios. The RBC ratio is total adjusted capital divided by two times authorised control level (required) capital.
Risk appetite may be curtailed well before capital reaches regulatory thresholds (Keeley 1990, Kirti 2019). Insurers rely on strong financial strength ratings to attract new business. In normal market conditions, RBC ratios below 300% are associated with drastically lower ratings (Figure 4).
Figure 4 Interquartile range of RBC ratios for US life insurers by AM best rating
Sources: S&P Global Market Intelligence and IMF staff analysis.
Note: Risk-based capital (RBC) ratios for this figure are capped at 600%. The RBC ratio is total adjusted capital divided by two times authorised control level (required) capital.
COVID-19 could have meaningful impact on insurers due to extensive financial dislocations across asset classes and, in an extreme scenario, potentially large increases in morbidity and mortality. Life insurers with high exposures to morbidity and mortality risk could be hit particularly hard if the pandemic proves difficult to control. Mortality rates in severe scenarios could trigger meaningful payouts relative to capital. Widespread asset-rating downgrades and persistently low interest rates would add to the difficult environment. Authorities looking to preserve credit supply should account for changes in insurer risk appetite. In a scenario with widespread bond rating downgrades, regulators should closely monitor and, as appropriate after all supervisory measures have been taken, reassess linkages to rating actions within supervisory frameworks, while enhancing supervision for insurers with risky holdings. Financial stability assessments should examine the implications of the pandemic for insurers.
Authors’ note: The views expressed in these notes are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.
Barro, R J, J F Ursúa and J Weng (2020), “The Coronavirus and the Great Influenza Pandemic: Lessons from the ‘Spanish Flu’ for the Coronavirus’s Potential Effects on Mortality and Economic Activity”, NBER Working Paper No. 26866.
Chodorow-Reich, G, A Ghent and V Haddad (2020), “Asset Insulators”, The Review of Financial Studies, 22 May.
Domanski, D, H S Shin and V Sushko (2017), “The Hunt for Duration: Not Waving But Drowning?”, IMF Economic Review 65(1): 113–53.
Ellul, A, C Jotikasthira and C T Lundblad (2011), “Regulatory Pressure and Fire Sales in the Corporate Bond Market”, Journal of Financial Economics, 101(3): 596–620.
Ellul, A, C Jotikasthira, C T Lundblad and Y Wang (2015), “Is Historical Cost Accounting a Panacea? Market Stress, Incentive Distortions, And Gains Trading”, The Journal of Finance, 70(6): 2489–538.
Financial Stability Board (FSB) (2020), Global Monitoring Report on Non-Bank Financial Intermediation 2019, Basel.
Foley-Fisher, N, B Narajabad and S Verani (2019), “Assessing the Size of the Risks Posed by Life Insurers’ Nontraditional Liabilities”, FEDS Notes, Board of Governors of the Federal Reserve System, Washington, D.C., 21 May.
Girardi, G, K W Hanley, S S Nikolova, L Pelizzon and M Sherman (2020), “Portfolio Similarity and Asset Liquidation in the Insurance Industry”, SSRN.
Hanley, K W and S S Nikolova (2020), “Rethinking the Use of Credit Ratings in Capital Regulations: Evidence from the Insurance Industry”, Review of Corporate Finance Studies, forthcoming.
International Association of Insurance Supervisors (IAIS) (2019), Holistic Framework for Systemic Risk in the Insurance Sector, Basel.
International Monetary Fund (IMF) (2015), “United States, Financial Sector Assessment Program, Stress Testing—Technical Note”, IMF Country Report 15/173, Washington, D.C.
International Monetary Fund (IMF) (2020), “Markets in the Time of COVID-19”, IMF Global Financial Stability Report, Washington, D.C.
Keeley, M C (1990), “Deposit Insurance, Risk, and Market Power in Banking”, The American Economic Review, 80(5): 1183–200.
Kirti, D (2019), “When Gambling for Resurrection is Too Risky”, IMF Working Paper 17/180, International Monetary Fund, Washington, D.C.
Kirti, D and N Sarin (2020), “Private Equity Value Creation in Finance: Evidence from Life Insurance”, Institute for Law & Economics Research Paper 20-17, University of Pennsylvania, Philadelphia.
Kirti, D and M Y Shin (2020), “Impact of COVID-19 on Insurers”, IMF Special Series on COVID-19, Washington, D.C.
Koijen, R S and M Yogo (2016), “Shadow Insurance”, Econometrica, 84(3): 1265–87.
Koijen, R S and M Yogo (2020), “The Fragility of Market Risk Insurance”, Becker Friedman Institute Working Paper No. 2018-09, University of Chicago, IL.
Koijen, R S, F Koulischer, B Nguyen and M Yogo (2020), “Inspecting the Mechanism of Quantitative Easing in the Euro Area”, Becker Friedman Institute for Economics Working Paper No. 2019-100, University of Chicago, IL.
Paulson, A, T Plestis, R Rosen, R McMenamin and Z Mohey‐Deen (2014), “Assessing the Vulnerability of the US Life Insurance Industry”, Modernizing Insurance Regulation 61–83, Hoboken, NJ: John Wiley & Sons.
Society of Actuaries (SOA) (2007), “Potential Impact of Pandemic Influenza on the U.S. Life Insurance Industry”, Research Projects—Life Insurance, Schaumburg, IL.
Society of Actuaries (SOA) (2020), “Impact of COVID-19”, Society of Actuaries Research Brief, Schaumburg, IL.
1 Lower interest rates across tenors will also have a significant impact on life insurers. Long-term liabilities mean that life insurers have high liability duration. Life insurers may face a trade-off between obtaining asset duration to offset this liability exposure and boosting returns. Lower long-term rates may therefore impact insurers’ willingness to take on credit risk. More broadly, insurers’ portfolio choices tend to amplify changes to long-term rates, including changes induced through quantitative easing (Domanski et al. 2017, Koijen et al. 2020).
2 Run risk is generally perceived to be lower than for banks; most withdrawals trigger large surrender charges, repricing of mortality risk, or adverse tax consequences (Kirti and Shin 2020, Paulson et al. 2014), although non-traditional liabilities such as institutional borrowing, borrowing from Federal Home Loan Banks, and securities lending have increased in recent years (Foley-Fisher et al. 2019).
3 We make several simplifying assumptions: mortality rates are assumed to be fixed within age group, reinsurance (including affiliate reinsurance) is assumed to be fully paid, and participating life contracts are not analysed. Annuities are also not analysed: this ignores both death benefits payable for annuities in the build-up phase (80% of US annuity reserves) and reductions in ongoing payments on immediate and annuitized annuities. Uncertainty about the mortality rate from COVID-19 remains very large. Current case fatality rates may overestimate mortality, as all cases may not have been detected. Mortality rates may also be even lower within age groups for holders of life insurance policies than is assumed here—mortality rates from flu and pneumonia are lower in US counties with higher incomes (SOA 2020). On the other hand, some deaths due to COVID-19 may have been incorrectly attributed to other causes. Mortality rates from other causes may increase with stress and loss of access to medical treatment. The US Centers for Disease Control and Prevention report that excess deaths from all causes have risen markedly since late March 2020. Affiliate reinsurance may not be fully paid (Koijen and Yogo 2016). Media reports indicate that some life insurers have significantly raised prices for mortality coverage due to the mortality impact of COVID-19 and the new market environment.
4 This note provides estimated nets of reinsurance, assuming reinsurance claims are fully paid. The 2015 US Financial Sector Assessment Program (IMF 2015) reported estimated losses of $20–25 billion, gross of reinsurance, in a standard pandemic scenario.
5 Market valuation of insurers’ own equity securities does account for lower market valuations of insurer assets. Insurer equity prices fell sharply in late February and March 2020, before beginning to partially recover in April 2020 (IMF 2020).
[vi] During the GFC, mortgage-backed securities (MBS) suffered mass rating downgrades. Risk classification of MBS was therefore modified in a manner that incentivised US insurers to mark bonds to market (reducing both actual and required capital) and likely prevented fire sales (Hanley and Nikolova 2020). However, the change in regulatory treatment was permanent, not temporary, and dramatically altered the capital treatment of MBS (Becker et al. 2020). It also applied to new purchases of MBS. US insurers have been disproportionate purchasers of newly issued junk MBS in recent years (Becker et al. 2020), and private equity-backed insurers have actively added risky MBS to their portfolios (Kirti and Sarin 2020).
6 Insurers also tend to hold similar assets. Overlapping holdings can also lead to common sales, amplifying market stress (Girardi et al. 2020).
7 European insurers are required to mark assets and liabilities to market under Solvency II. For levered balance sheets, marking assets to market has an amplified impact on equity. For example, suppose that fixed-income with observable market prices accounts for 70% of assets, with equity financing 10% of assets. Even a 5% reduction in the value of the fixed-income assets then reduces equity by one-third.
8 Variable annuities expose some life insurers to equity markets. Although the majority of variable annuities do not offer guaranteed returns, some insurers do provide guarantees and face pressure from falling equity prices and higher volatility. Variable annuity providers have seen their stock prices fall by as much as 50–70% in the first quarter of 2020 (Koijen and Yogo 2020). The cost of hedging via equity derivatives rise with market volatility.
9 US capital requirements are a nonlinear function of several components, one of which moves linearly with the ratings-based composition of bond holdings. We estimate a lower bound for the impact on total capital requirements based on Kirti and Sarin (2020).
10 These thresholds apply to an RBC ratio defined as total adjusted capital (TAC) divided by company action level (CAL) RBC. CAL RBC is two times authorised control level (ACL) RBC.