Abstract
Selected Issues
Financial Stability Risks from Euro Area Insurance and Pensions Sector1
Insurance company and pensions fund (ICPF) assets have grown over time in the euro area and the sector has come under pressure due to the low interest rate environment. Given the dominance of insurers in this sector, the discussion in this note mostly covers the vulnerabilities of these institutions. Insurers in Germany, France and Austria are most vulnerable to market risks due to the prevalence of both guaranteed products and large asset-liability duration mismatches. But comfortable solvency buffers in France and Austria provide some protection. In some countries, the asset portfolios of insurers are concentrated in investments in their own sovereigns and banks, creating strong domestic interconnectedness. Even though the shortfall in buffers of a ¼ percent of euro area GDP arising from the severe downside scenario in the 2016 stress tests of the European Insurance and Occupational Pensions Authority (EIOPA) is modest, these shortfalls are higher for high-debt countries and could be even higher if shocks were amplified through domestic interconnectedness.
A. Background
1. Euro area ICPFs have grown since the global financial crisis and their combined asset amount to 90 percent of GDP. ICPFs have also grown as a share of bank assets. With a balance sheet size of €9.7 trillion, the ICPF sector is dominated by the four systemically important insurance (SII) groups AXA (France), Allianz (Germany), Generali (Italy) and Aegon (The Netherlands) (Appendix I). Consequently, Germany, France, Italy and Netherlands are the largest ICPF markets. Insurance companies dominate the sector, with about €7 trillion in assets, and while most insurance companies offer both life and nonlife insurance business, 65 percent of gross written premia are for the life insurance business (EIOPA, 2016a).
Size of Insurance and Pensions Sector in the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: European Central Bank (ECB); S&P Global Market Intelligence; IMF staff calculations.Note: Generali (ITA) was removed from the GSII list in 2015, but have been included in the list due to its large size. ICPF assets are based on unconsolidated date. The GSII assets are based on consolidated group data.Size of Insurance and Pensions Sector in the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: European Central Bank (ECB); S&P Global Market Intelligence; IMF staff calculations.Note: Generali (ITA) was removed from the GSII list in 2015, but have been included in the list due to its large size. ICPF assets are based on unconsolidated date. The GSII assets are based on consolidated group data.Size of Insurance and Pensions Sector in the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: European Central Bank (ECB); S&P Global Market Intelligence; IMF staff calculations.Note: Generali (ITA) was removed from the GSII list in 2015, but have been included in the list due to its large size. ICPF assets are based on unconsolidated date. The GSII assets are based on consolidated group data.2. The two main types of insurance companies are related to life and non-life business. Insurance companies are funded by long-term policy holders and have to hold technical provisions or reserves against these policies. The largest share of the technical provisions in Europe belongs to life insurers. A pure life insurer would provide a lump sum payment to beneficiaries upon the death of the insured, in exchange for premium payments. Within this model: “term” life insurance provides financial protection for a specific period; “universal” life insurance is a permanent type of coverage through life but flexibly allowing the raising or lowering of premia or coverage amounts; and, “whole” life insurance is like universal life, but have fixed premia and could have a cash value that functions like a savings product. Sometimes, life insurers offer investment products (unit-linked to stocks and bonds) that are like mutual funds and where the insured usually bears the risk. Nonlife insurance companies span the spectrum of health, workers’ compensation, property, automobile, fire, etc. About 55 percent of insurers in the EU are life, 43 percent offer both life and non-life, and only about 2 percent offer only non-life.
3. Insurance companies tend to be more exposed to market risk than banks. A large part of their balance sheet consists of marketable securities. For instance, 60 percent of Allianz assets (shown in Figure 2 as an example) consist of investments in government and corporate bonds, equities and other assets. For a typical bank, the share of such investments in total assets would be 15–20 percent. For insurers, since the duration (time-to-maturity, weighted by the present discounted value or PDV) of investments is usually less than that of liabilities, insurance companies profit from the duration mismatch when interest rates go up, as they can reinvest assets at higher rates.2 Also, with their higher duration, the value of liabilities falls more than the value of assets when interest rates go up. Companies offering unit-linked products (that follow broad indices) have to hold separate accounts for their clients (shown in Figure 2 as equal (gray) quantities on both sides of the balance sheet), who bear the risk of such investments.
Balance Sheets: A Large Insurance Company and Euro Area ICPF Sector
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Allianz from S&P Global Market Intelligence; ECB; IMF staff calculations.Balance Sheets: A Large Insurance Company and Euro Area ICPF Sector
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Allianz from S&P Global Market Intelligence; ECB; IMF staff calculations.Balance Sheets: A Large Insurance Company and Euro Area ICPF Sector
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Allianz from S&P Global Market Intelligence; ECB; IMF staff calculations.4. Pure insurance companies are usually a source of stability. With long-term liabilities, these entities provide stable long-term financing to the government, corporate, and infrastructure sectors. Since their investment strategies are opposite to banks’—funding long-term and investing short-term—they provide support to markets during distress (Appendix II). With diverse activities, the insurance sector as a whole had lower variability of stock market value than the banking sector during the global financial crisis and the European debt crisis (Figure 3).
Eurostoxx Indices: Overall, Banks, Insurance
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Bloomberg Finance L.P.; IMF staff calculations.Eurostoxx Indices: Overall, Banks, Insurance
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Bloomberg Finance L.P.; IMF staff calculations.Eurostoxx Indices: Overall, Banks, Insurance
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: Bloomberg Finance L.P.; IMF staff calculations.5. However, certain activities, connections and macroeconomic conditions could contribute to financial stability risks. Historically, insurance companies that failed or had to receive government help typically traced their distress to problems on the liabilities side of the balance sheet (Sugimoto, 2016). Insurers with large duration mismatches between assets and liabilities, and with high rates and numbers of guarantees provided to clients are most vulnerable to falling interest rates (Appendix II, section B). Moreover, through investments on their asset side, insurers could be strongly interconnected with banks, corporates and governments, creating one conduit for systemic risk (section C). Finally, insurers may “gamble for resurrection” during low-rate periods by investing in risky and illiquid assets to seek higher yields:
Guaranteed rates. Life insurance products with guaranteed rates of return above investment yields are especially vulnerable to a low-interest rate regime. There are instances in the United Kingdom (Equitable Life in 2000) and Japan (eight life insurers in 1997–2001) where a combination of high guaranteed rates and an environment of prolonged low interest rates ultimately led to fiscal transfers to the industry compensation schemes or to the policyholders’ protection funds to make up for the deficit.
Links with banks and other sectors. Conglomeration and intragroup transactions with banking entities can lead to spillovers from insurance to banks, and vice versa. Even without conglomeration, insurance companies have deposits in banks and invest in bank bonds, including convertible bonds. Insurance companies provide long-term funding to corporates and governments, and so stress in these companies can result in a funding shock. In fact, interconnectedness is an important consideration for designating insueres as global systemically important (GSII, Appendix I). Sectoral spillovers based on asset price movements have been shown to be high for banks and insurers in Europe (IMF, 2016).
Derivative trading and securities lending. Selling credit default swaps without hedging these exposures or setting aside capital and reserves; reinvesting cash collateral received in securities lending operations into collateralized debt obligations (AIG in 2008, which was rescued by the U.S. government) are examples of risky strategies.
Yield-seeking behavior. In low interest rate environments, ICPFs may venture into risky investments such as lending to customers, real estate (so that they benefit from illiquidity premia), or high-yield corporate bonds. For instance, euro area insurance portfolios are becoming riskier with a shift toward corporate and lower-rated bonds (ECB, 2016).
Lapse rates. A high degree of withdrawals or cancellations of life insurance policies can give rise to liquidity risks, especially if penalty rates for early policy cancelation are driven down to zero amid competition.
6. The Solvency II Directive, introduced in 2016, requires insurance and reinsurance companies to hold capital and adequate provisions against adverse market conditions. The new quantitative requirements include market-consistent valuation of assets and liabilities and risk-based capital requirements.3
Capital requirements. Insurance and reinsurance companies have to hold own funds, the difference between assets and liabilities, to cover two types of capital requirements. The solvency capital requirement (SCR) is risk-based and is the amount of own funds needed to withstand the worst annual loss expected to occur over the next 200 years. The SCR ratio is the ratio of eligible own funds to SCR, and this ratio should be at least 100 percent. The minimum capital requirements (MCR) is between 25 and 45 percent of the SCR; and the minimum MCR ratio is the ratio of eligible own funds and MCR. If an insurer or reinsurer is not complying with the SCR, it has to take measures (increasing capital or lowering risk) to meet the SCR again within six months. A breach of the MCR could result in a withdrawal of authorization unless it is covered again in three months (EIOPA, 2016a). The SCR can be calculated with a standard formula or with an internal model. The standard formula consists of modules for the different risks that an insurance company is exposed to (market, underwriting, counterparty default, and operational risks). Most companies use the standard formula (EIOPA, 2016a).
Long-term risk-free rate for calculating technical provisions. The main liabilities of insurance companies comprise technical provisions set up for the insurance and reinsurance obligations of the undertaking. Since the duration of these are longer than the availability of liquid risk free rate benchmark quotes in the market, there is technical guidance for the long-term risk free rate that insurers should use for high maturities. For instance, the EIOPA has determined that in certain countries, 20 years is the longest maturity with liquid markets and the valuation of liabilities up to 20 years could be based on the actual market interest rates. Beyond that, the EIOPA uses the ultimate forward rate (UFR)—defined as the long-term average of short-term real interest rates plus long-term inflation expectations—to extrapolate yields at longer maturities. The UFR is set at 4.2 percent until the end of 2016 for the euro area.4 A new methodology for the calculation of the UFR on an ongoing basis was just published in April 2017 with a decrease in the UFR to 3.65 percent, phased-in at 4.05 percent in 2018. A lower UFR would reduce yields on 20+ year maturities and increase the PDV of liabilities, without necessarily increasing the PDV of assets.
7. Since the risk-free rate is a crucial component in calculating technical provisions, insurance companies are allowed to make certain adjustments to the interest rate to prevent procyclical outcomes. Taken together, these measures are called long-term guarantee (LTG) measures that allow insurers mainly offering long-term guarantee products to adjust the risk-free rate for market volatility, for instance, to calculate technical provisions. These calculations would then affect own funds—a higher interest rate would decrease technical provisions and increase own funds, and hence increase the SCR and MCR ratios, amounting to capital relief under adverse market conditions.
B. Vulnerabilities from Business Models
8. The yield curve is a key determinant of financial stability in the ICPF sector. Lower risk-free rates increase the PDV of cash flows of both assets and liabilities. But because the duration of assets is typically lower (average of seven years) than the duration of liabilities (average of 12 years), liabilities increase more than assets, thus lowering own funds (assets minus liabilities). But falling short-term interest rates also lowers interest income (and hence profits) from bond yields in addition to the adverse impact on own-funds as the shorter maturity of bonds have to be reinvested at lower yields. Falling yields since the global financial crisis (Figure 4) have, therefore, created concerns for the ICPF sector. Appendix II provides a simple example to illustrate the impact of changes in interest rates (especially falling yields) on the market value of assets and liabilities of insurance companies.
Lower Yields Partly Explain Falling Profits
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Lower Yields Partly Explain Falling Profits
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Lower Yields Partly Explain Falling Profits
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
9. Insurers that provide high guaranteed rates of return are more vulnerable to a scenario where interest rates remain low for long. Although there is no legal definition of guarantees, over 75 percent of life companies offer such guarantees in Austria, Belgium, Germany, Spain, France, Greece, Italy, Latvia and Lithuania (EIOPA 2016b). The most common are minimum guaranteed rates every year for the term of the product, or a guaranteed sum assured until maturity. However, a few also offer guaranteed yields on unit-linked investments, which are usually managed for clients in separate accounts. For life insurers, guarantees form part of their technical provisions and have come down over time. But the legacy guarantees that are much higher than the current yield will remain part of the technical provisions. In the EIOPA stress test sample, almost a third of the contracts guarantee interest rates between 3–4 percent for the next 12 years on average; another 10 percent of the contracts promise rates above 4 percent (EIOPA 2016b).5 These rates contrast with the (early 2017) average 30-year AAA-rated yield-to-maturity of 1.3 percent and those on all bonds at 2.4 percent.
10. Countries with insurers that have large duration mismatches and a greater prevalence of guarantees are more vulnerable to prolonged low interest rates. Germany, Austria, France, and Latvia are among countries that fall under this category (Figure 5). The relatively low duration gap and the limited use of guaranteed products in Netherlands might explain increasing profits for Dutch insurers between 2012 and 2015 (Figure 4). The EIOPA stress tests showed that under a scenario where the yield curve moves down and flattens and the UFR falls from 4.2 percent to 2 percent, German, Austrian and Latvian insurers would have the largest differentials between changes in values of liabilities and assets, with the largest hit on own funds.
Asset-Liability Duration Mismatches and Prevalence of Guaranteed Products
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Macaulay duration of liabilities, and modified duration of assets. See Appendix II for pros and cons of using the Macaulay duration measure. Countries with 75–100 percent of insurers offering guaranteed products are marked in red. If guaranteed technical provisions as share of total technical provisions were used as a criterion, then Netherlands would be marked red.Asset-Liability Duration Mismatches and Prevalence of Guaranteed Products
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Macaulay duration of liabilities, and modified duration of assets. See Appendix II for pros and cons of using the Macaulay duration measure. Countries with 75–100 percent of insurers offering guaranteed products are marked in red. If guaranteed technical provisions as share of total technical provisions were used as a criterion, then Netherlands would be marked red.Asset-Liability Duration Mismatches and Prevalence of Guaranteed Products
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Macaulay duration of liabilities, and modified duration of assets. See Appendix II for pros and cons of using the Macaulay duration measure. Countries with 75–100 percent of insurers offering guaranteed products are marked in red. If guaranteed technical provisions as share of total technical provisions were used as a criterion, then Netherlands would be marked red.11. Solvency buffers such as the SCR look comfortable for now, but should be interpreted with caution. The baseline SCR ratio for the euro area in the EIOPA stress test was 200 percent (196 percent for the EU), which means that own funds covered almost twice the solvency capital requirements. However, life insurers can adjust the risk-free interest rate, used for discounting liabilities, for long-term guarantee and transitional measures.6 Insurance and reinsurance companies covering more than 60 percent of the technical provisions in the EA use volatility adjustments (VA), an adjustment to the term structure of the risk-free rate to mitigate the impact of unusual market conditions that could increase bond yields temporarily. The SCR ratio without LTGs is about 168 percent for the euro area (136 percent for the EU). Although all SCRs were above 100 percent in the baseline for the EIOPA stress tests, SCRs shorn of LTGs and transitionals were lower than 100 percent for Greece and Portugal, and were lower by more than 50 percentage points for Belgium, Germany, Greece, Spain, and Netherlands (Figure 6).
Solvency Capital Requirements Ratios
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: LTG measures refer to Long-Term Guarantees measures that were introduced in the Solvency II Directive to ensure an appropriate treatment of insurance products that include long-term guarantees.Solvency Capital Requirements Ratios
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: LTG measures refer to Long-Term Guarantees measures that were introduced in the Solvency II Directive to ensure an appropriate treatment of insurance products that include long-term guarantees.Solvency Capital Requirements Ratios
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: LTG measures refer to Long-Term Guarantees measures that were introduced in the Solvency II Directive to ensure an appropriate treatment of insurance products that include long-term guarantees.12. Insurers were more adversely affected by EIOPA’s “double hit” scenario. In this scenario, there is a sudden increase of risk premia combined with the low yields. Increases in government bond yields and credit spreads of corporate bonds coincide with fall in stock prices, property prices and commodity prices. While the low-for-long scenario mainly increases the PDV of insurers’ liabilities more than assets, the “double hit” scenario decreases the valuation of assets more than that of liabilities.7 The higher yields on government and corporate bonds, the main component of assets, lead to valuation losses. Lower asset values also allow those insurers without guaranteed payments to adjust the benefit payments (dependent on the performance of the assets), which reduces liabilities to some extent. For example, those with unit-linked business encounter erosion of values held in separate accounts on both the asset and liabilities sides. The LTG measures on the liabilities side provide a cushion against the asset volatilities; for instance, the volatility adjustment on risk-free rates begins to kick in during market turmoil, preventing fire sale of assets.
13. The maximum shortfall in the coverage of SCR by own funds implied by the EIOPA stress tests is about a ¼ percent of euro area GDP. The stress tests mainly identified Germany (and to some extent France) as having low buffers against the two specific shocks, with the highest impact for the “double hit” scenario. For countries with high guaranteed rates, duration mismatches and strong domestic interconnectedness, the government might want to step in (and refurbish policy protection schemes, for instance) to stave off spillovers to other sectors. Such contingent liabilities seem low for the overall euro area, but could be concentrated in a few countries. The maximum contingent liabilities coming out of the shortfall identified in the stress test would be about €14 billion for Germany (0.46 percent of GDP) and only about €310 million for France (0.01 percent of GDP) for the insurers to go back to an SCR ratio of 100 percent, without taking LTGs into account (Figure 7). The difference between France and Germany stems from lower baseline solvency buffers in the latter.8 Even though the Portuguese and Belgian insurance sectors are small in the euro area, their own funds shortfall under the severe scenario are 1.5 percent of GDP and 0.8 percent of GDP, respectively, owing mainly to thin baseline solvency buffers.9
Shortfall in Own Funds to Cover the SCR (“SCR-coverage”)
(Percent of GDP)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA, 2016a; IMF staff calculations.Note: Based on 2016 EIOPA Stress Test Results for the “Double Hit” Scenario. The SCR ratio of 100 percent, without LTGs, is taken as the threshold.Shortfall in Own Funds to Cover the SCR (“SCR-coverage”)
(Percent of GDP)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA, 2016a; IMF staff calculations.Note: Based on 2016 EIOPA Stress Test Results for the “Double Hit” Scenario. The SCR ratio of 100 percent, without LTGs, is taken as the threshold.Shortfall in Own Funds to Cover the SCR (“SCR-coverage”)
(Percent of GDP)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA, 2016a; IMF staff calculations.Note: Based on 2016 EIOPA Stress Test Results for the “Double Hit” Scenario. The SCR ratio of 100 percent, without LTGs, is taken as the threshold.14. EIOPA’s adjustment of the UFR could have a large impact on insurance technical provisions. At the current 4.2 percent, the UFR implies a yield of more than 1.5 percent for 30-year maturities (as of July 2016) (Figure 8). If the UFR were to be reduced to 3.65 percent (as was recently proposed), the 30-year yields would fall by about 50 bps, and 60 year yields by 75 bps. Countries where insurers have long duration liabilities—such as Germany, Netherlands, and Austria—would be the most adversely affected with lower SCR ratios. Indeed, EIOPA’s impact analysis (EIOPA, 2017) suggest that a 50 bps reduction in the UFR would reduce the EU-wide SCR ratio by about 4 percentage points, with much greater impact on Germany (16 pp), Netherlands (15 pp), and Austria (8 pp).
Implications for the Yield Curve for Different UFR Assumptions
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Used with permission of Bloomberg Finance L.P.Implications for the Yield Curve for Different UFR Assumptions
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Used with permission of Bloomberg Finance L.P.Implications for the Yield Curve for Different UFR Assumptions
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Used with permission of Bloomberg Finance L.P.C. Vulnerabilities Due to Interlinkages
15. Systemic risks could be higher for those ICPFs that are strongly interconnected with other sectors. ICPFs provide long-term funding to banks (deposits and bonds), corporates (equities and bonds) and governments (bonds). Thus, problems with the ICPFs could increase fire sale risks or withdrawals of funding from other sectors. Ownership linkages through conglomerates or bancassurance groups could increase funding costs for banks if the insurance arm is in trouble. On the other hand, problems in banks could trigger conversion of convertible bank debt (held by ICPFs) to equity, making ICPF’s investments more volatile. Corporate distress could increase the risk premia on ICPF investments, as in the “double-hit” scenario stress tested by EOIPA.
16. At the euro area level, ICPFs are less interconnected to other sectors, compared to banks, but could pose a funding risk to sovereigns. ICPFs are most connected to sovereigns, investing 19 percent of GDP or about €2 trillion (Figure 9). This is still less than the amount MFIs (including banks, ECB and NCBs) and the Rest of the World (ROW)’s invest in euro area sovereigns. But a funding shock from ICPFs would have to be absorbed by other players (possibly the ECB, subject to issuer limits and capital key). The ICPFs holdings of ROW assets and bank deposits and bonds are another 10–12 percent of GDP (as shown by the vertical column on ICPFs in Figure 9). Comparatively and unlike banks, ICPFs do not borrow heavily from the rest of the sectors (as shown by the ICPF row in Figure 9). The ECB’s holdings of ICPF bonds through the Corporate Sector Purchase Program (CSPP) is included in the 1.3 percent of GDP borrowings from the MFIs.
Flow of Funds for the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.Note: MFI: Monetary Financial Institutions (banks and central banks); OFC: Other Financial Corporations; Non MMF: Non-Money Market Funds; ICPF: Insurance and Pension Funds Sector; Govt: Government; NFC: Nonfinancial Corporates; HH: Households; and, ROW: Rest of the World.Flow of Funds for the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.Note: MFI: Monetary Financial Institutions (banks and central banks); OFC: Other Financial Corporations; Non MMF: Non-Money Market Funds; ICPF: Insurance and Pension Funds Sector; Govt: Government; NFC: Nonfinancial Corporates; HH: Households; and, ROW: Rest of the World.Flow of Funds for the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.Note: MFI: Monetary Financial Institutions (banks and central banks); OFC: Other Financial Corporations; Non MMF: Non-Money Market Funds; ICPF: Insurance and Pension Funds Sector; Govt: Government; NFC: Nonfinancial Corporates; HH: Households; and, ROW: Rest of the World.Flow of Funds for the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.Note: MFI: Monetary Financial Institutions (banks and central banks); OFC: Other Financial Corporations; Non MMF: Non-Money Market Funds; ICPF: Insurance and Pension Funds Sector; Govt: Government; NFC: Nonfinancial Corporates; HH: Households; and, ROW: Rest of the World.Flow of Funds for the Euro Area
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.Note: MFI: Monetary Financial Institutions (banks and central banks); OFC: Other Financial Corporations; Non MMF: Non-Money Market Funds; ICPF: Insurance and Pension Funds Sector; Govt: Government; NFC: Nonfinancial Corporates; HH: Households; and, ROW: Rest of the World.17. Systemic risks are greater in countries where ICPFs invest mainly in domestic sovereign and banking sectors. Problems with ICPFs could easily spill over to domestic banks and sovereigns if insurers do not diversify into non-domestic banks and sovereigns. For instance, more than 70 percent of the sovereign bond portfolios of French insurance companies, among the largest in the sample, are concentrated in French sovereign bonds, and 55 percent of their bank debt holdings are in French banks (mainly through bancassurance linkages) (Figure 10). More than half of the bank bonds held by German insurers are for domestic banks. In fact, German insurers allocated 68 percent of their assets in domestic financial sector in 2014, compared to a euro area average of 49 percent (European Commission, 2016). Countries where insurers are large and domestically interconnected contribute to and are conduits of systemic risk.
Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country.Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country.Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country.18. ICPFs have been withdrawing their bank deposits and, within their remaining bank deposits, shifting to shorter term deposits since the global financial crisis (Figure 11). At the euro area level, these withdrawals are small, but are concentrated in Germany, Cyprus, Spain, Greece, and Slovakia. In some countries, there is a shift from banks to money market mutual funds. Bank stress tests in these countries need to specifically look at the liquidity risks of banks from ICPF deposit outflows against the liquidity coverage ratio (LCR). LCR requires that banks hold enough high-quality liquid assets to withstand 30-day run of liabilities. The run-rates differ by type of liability. If the run rate for insurers is taken to be the same as that for unsecured wholesale funding from large corporates and sovereigns, it is 40 percent (if not fully covered by deposit insurance). That is, an LCR of 100 percent (that banks are required to meet) needs to take account of a 40 percent run rate of ICPF deposits. So far, the deposit reduction at the euro area level has been only 7 percent for 2016Q3 over 2015Q3, meaning that the monthly run-rate is considerably lower than the 40 percent rate.
ICPF Deposits with Banks
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.ICPF Deposits with Banks
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.ICPF Deposits with Banks
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: ECB; IMF staff calculations.19. Sovereign-ICPF linkages could add to vulnerabilities, especially in high-debt countries. This is especially so in countries where insurers are heavily invested in the domestic sovereign debt market. More than 60 percent of insurers’ sovereign debt exposures are in domestic sovereigns in Italy, Slovakia, France, Belgium and Portugal (Figure 12). Among these countries, public debt in France, Belgium and Italy is above the EA-average of 90 percent of GDP (shown as red bubbles in Figure 12). Problems in ICPFs, for example those struggling with the low-interest environment or market volatility with high guaranteed rates amid low solvency buffers, could trigger a fire sale of government securities and bring a stop to rolling over government funding. Moreover, there are limits to the ECB’s ability to step in to substitute for insurers’ sovereign debt holdings, because its Asset Purchase Program is subject to various constraints, including the capital key.
Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country; red bubbles represent countries with government debt higher than 90 percent of GDP, the euro area average in 2016.Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country; red bubbles represent countries with government debt higher than 90 percent of GDP, the euro area average in 2016.Domestic Systemic Risk from Linkages
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: The bubble size is proportionate to the size of the insurance sector in the country; red bubbles represent countries with government debt higher than 90 percent of GDP, the euro area average in 2016.D. Summary Indicator of Vulnerabilities
20. Countries can be ranked based on a summary indicator of ICPF vulnerabilities. Four sets of vulnerabilities are considered: interest rate risk due to duration mismatches (z-scores for duration mismatch) and prevalence of guaranteed products (1 or 0); domestic interconnectedness with own sovereigns and banks (z-scores for the fraction of bond-holdings exposed to own sovereigns and banks); low solvency buffers (z-scores for the 2016 EIOPA stress test baseline SCR ratios without LTGs); and finally, size of the insurance sector (z-scores based on the size of own funds). Thus, the z-scores based on the four sources of vulnerabilities can be summed up to create an overall index of systemic vulnerabilities from ICPFs (Figure 13). In addition, the EIOPA stress test results are captured by red bars in Figure 13, showing countries where insurers fell below the 100 percent SCR ratio in at least one of the two stress scenarios. Countries can also be ranked by their z-scores on each of the four sources of vulnerabilities (Figure 14).
Summary of Vulnerabilities
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.1/ There is no data on interconnectedness for Spain.Note: Countries are ranked based on the following indicators: duration gap, prevalence of guaranteed products (together “interest rate risk”), relative size of net asset positions (“large size”), relative SCR ratios based on reported SCR without LTGs (“low solvency buffers”), and exposures to domestic banks and sovereigns (“domestic systemic risk”). Positive (negative) numbers are “higher (lower) than average” on each risk. The red horizontal markers identify countries in which assets/liabilities fell below 1 for at least one of the two 2016 EIOPA stress scenarios.Summary of Vulnerabilities
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.1/ There is no data on interconnectedness for Spain.Note: Countries are ranked based on the following indicators: duration gap, prevalence of guaranteed products (together “interest rate risk”), relative size of net asset positions (“large size”), relative SCR ratios based on reported SCR without LTGs (“low solvency buffers”), and exposures to domestic banks and sovereigns (“domestic systemic risk”). Positive (negative) numbers are “higher (lower) than average” on each risk. The red horizontal markers identify countries in which assets/liabilities fell below 1 for at least one of the two 2016 EIOPA stress scenarios.Summary of Vulnerabilities
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.1/ There is no data on interconnectedness for Spain.Note: Countries are ranked based on the following indicators: duration gap, prevalence of guaranteed products (together “interest rate risk”), relative size of net asset positions (“large size”), relative SCR ratios based on reported SCR without LTGs (“low solvency buffers”), and exposures to domestic banks and sovereigns (“domestic systemic risk”). Positive (negative) numbers are “higher (lower) than average” on each risk. The red horizontal markers identify countries in which assets/liabilities fell below 1 for at least one of the two 2016 EIOPA stress scenarios.
Vulnerabilities: Top-3 Countries
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Bubbles are proportional to the size of the insurance sector. The placements of the bubbles under each category of vulnerability is based on the z-scores used to depict Figure 13.Vulnerabilities: Top-3 Countries
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Bubbles are proportional to the size of the insurance sector. The placements of the bubbles under each category of vulnerability is based on the z-scores used to depict Figure 13.Vulnerabilities: Top-3 Countries
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA; IMF staff calculations.Note: Bubbles are proportional to the size of the insurance sector. The placements of the bubbles under each category of vulnerability is based on the z-scores used to depict Figure 13.21. Germany and France stand out as being especially vulnerable, followed by Austria (Figure 13). The biggest source of risks for these countries come from interest rate risk (Germany and Austria) and large size (Germany and France). The more disaggregated scores (Figure 14) show that Germany, Austria and Latvia are the most vulnerable to interest rate risk from both duration mismatches and prevalence of guaranteed products. Greece, Portugal and Spain had the least baseline SCR-coverage buffers (for the EIOPA stress tests), and so would have the least ability to withstand shocks. Slovakia, France, and Germany score high on domestic interconnectedness, although Italy and Portugal come close behind.
22. The results of the EIOPA stress tests can be seen to test the insurers on the combination of interest rate risk and low solvency buffers. The overall SCR-coverage shortfall of a ¼ percent of GDP for the euro area for the “double hit” scenario assumes a generalized market turmoil. Some of the countries with higher-than-average SCR-coverage shortfalls—Portugal (1.5 percent of GDP), Belgium (0.8 percent of GDP), and Spain-Greece (around 0.45–0.48 percent of GDP)—also have high sovereign debt that are partly funded by the ICPF sectors.
23. The EIOPA stress tests do not address the additional SCR-coverage needs potentially arising from interconnectedness of the domestic sovereign-bank-insurance markets. Problems in insurance sectors could increase sovereign bond yields, which could affect wholesale funding cost of banks (sovereign-bank linkages). Intense problems in banks could lead to bail-ins, which could, in turn, lead to haircuts on insurers’ holdings of bank-bonds, reducing the SCR ratio even further. Thus, the impact of amplification of shocks could see even higher SCR shortfalls, especially in countries that already have high debt.
E. Summary and Policy Implications
24. Countries with large insurance sectors and those with high government debt are also exposed to the risk of erosion of SCR-buffers due to the low interest rate environment. Germany, France and Austria stand out with a combination of interest rate duration mismatches on assets and liabilities and the prevalence of guaranteed products. Among these countries, the 2016 EIOPA stress tests revealed capital shortfalls from a severe version of the scenarios primarily in Germany, at about 0.46 percent of GDP. These shortfalls are higher in some countries with above-average government debt-to-GDP ratios: Portugal (1.5 percent of GDP), Belgium (0.8 percent of GDP), and Spain (0.48 percent of GDP).
25. The ability of insurers to adjust guarantees provides them with a policy lever against interest rate risk and protects policy holders. As shown in Appendix II, keeping constant the long-term cash flows or liabilities would only erode buffers when the interest rate falls. Guaranteed rates of return at 3–4 percent for the next 12 years far exceed rates of return on assets of 1.3–2.5 percent on 30-year products. As a rule of thumb, guaranteed rates should be 2 percentage points below market rates (Sugimoto, 2016). While it is difficult to adjust guarantees on legacy products, there should be rules against promising guaranteed returns on new products. Importantly, there should not be any guaranteed returns on unit-linked products, which were originally designed to pass on market risks to clients.
26. Stress tests should assess risks that can arise from domestic interconnectedness of the insurance sectors. The EIOPA stress tests currently do not address the additional shortfalls in SCR-coverage buffers that could be coming from interconnectedness of the domestic sovereign-bank-insurance markets. For instance, some countries with a higher-than-average buffer shortfall also have high sovereign debt that are partly funded by the ICPF sectors. The impact of amplification of shocks through sovereign-financial linkages could lead to higher SCR-coverage shortfalls.
27. Insurers’ deposits in banks have been falling, especially in Germany, Cyprus, Spain, Greece, and Slovakia. Liquidity risk stress tests of these banking systems should assess the ability of banks to withstand the continued withdrawal of such deposits. While the overall, run-rate of these deposits was only 7 percent for the year in 2016Q3, these rates could be higher in specific banking systems.
28. Insurers should diversify their businesses and consolidate to reduce vulnerabilities. It would be difficult for insurers to adjust to low rates by changing the asset mix alone. Changing the asset portfolios to cover solvency margins in a low for long environment entails taking unacceptable levels of risk (IMF, 2017). Solvency II risk weights on risky assets are higher than those required for the U.S. and Japan (Appendix I). For instance, investments in real estate (and earning illiquidity premia) carry a risk weight of 25 percent under Solvency II, unlike in the U.S. (15 percent) and in Japan (10 percent). Competition is growing for assets such as infrastructure debt that combine relatively attractive returns with low capital charges. But liquidity and issuance is low and it is tougher for smaller insurers to access the market. While some insurers are expanding their unit-linked businesses (ECB, 2016), there is stiff competition from ETFs. Large insurance companies are adjusting through cross-border mergers, as seen by increased M&A activities in 2015 and 2016. This trend needs to continue especially in countries with SCR-coverage shortfalls, with M&As for smaller insurers as well.
References
European Central Bank (ECB), 2016, Euro Area Financial Institutions, in Financial Stability Review, November.
European Commission, 2016, Country Report: Germany, Staff Working Document.
European Insurance and Occupational Pensions Authority (EIOPA), 2016a, Report on Long-Term Guarantees Measures and Measures on Equity Risk, December 16.
European Insurance and Occupational Pensions Authority (EIOPA), 2016b, 2016 EIOPA Insurance Stress Test, December.
European Insurance and Occupational Pensions Authority (EIOPA), 2017, Results of the Impact Analysis of Changes to the UFR, April 2017.
International Monetary Fund (IMF), 2016, The Insurance Sector—Trends and Systemic Risk Implications, Chapter 3 of Global Financial Stability Report, April.
International Monetary Fund (IMF), 2017, Low Growth, Low Interest Rates, and Financial Intermediation, Chapter 2 of Global Financial Stability Report, April.
Sugimoto, Nobuyasu, 2016, How to Assess the Insurance Sector in Surveillance: A Few Pointers, internal IMF presentation, November.
Appendix I. Designation of GSIIs and Comparison of Risk Weights
Attributes of being globally systemically important institutions (GSIIs): differences between insurers and banks (IMF, 2016)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Sources: Basel Committee on Banking Supervision 2011; International Association of Insurance Supervisors 2013a: and IMF staff calculations.Note: GSIBs = global systemically important banks; GSIIs = global systemically important insurers: NTNI = nontraditional non-insurers.Attributes of being globally systemically important institutions (GSIIs): differences between insurers and banks (IMF, 2016)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Sources: Basel Committee on Banking Supervision 2011; International Association of Insurance Supervisors 2013a: and IMF staff calculations.Note: GSIBs = global systemically important banks; GSIIs = global systemically important insurers: NTNI = nontraditional non-insurers.Attributes of being globally systemically important institutions (GSIIs): differences between insurers and banks (IMF, 2016)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Sources: Basel Committee on Banking Supervision 2011; International Association of Insurance Supervisors 2013a: and IMF staff calculations.Note: GSIBs = global systemically important banks; GSIIs = global systemically important insurers: NTNI = nontraditional non-insurers.Capital Charges for Risky Investment of Insurers, (IMF, 2017)
Capital Charges for Risky Investment of Insurers, (IMF, 2017)
| Solvency II (Standard Approach) | U.S. Risk-Based Capital Requirement | Japanese Solvency Margin Ratio | |
|---|---|---|---|
| Listed Equity | 22% | 15% | 20% |
| Private Equity | 49% | 30% | 20% |
| Non-investment grade corporate bonds | Up to 37.5% (5 year) | 30% (Class 6) | 30% |
| Real Estate | 25% | 15% | 10% |
Capital Charges for Risky Investment of Insurers, (IMF, 2017)
| Solvency II (Standard Approach) | U.S. Risk-Based Capital Requirement | Japanese Solvency Margin Ratio | |
|---|---|---|---|
| Listed Equity | 22% | 15% | 20% |
| Private Equity | 49% | 30% | 20% |
| Non-investment grade corporate bonds | Up to 37.5% (5 year) | 30% (Class 6) | 30% |
| Real Estate | 25% | 15% | 10% |
Appendix II. Impact of Interest Rate Changes on Financial Institutions’ Balance Sheets: An Example
The impact of interest rate changes can be illustrated with hypothetical balance sheets of a life insurance company and a bank. As discussed in Section I, the example insurance company’s liabilities are longer-dated than its assets, and a bank’s liabilities are shorter-dated than its assets. For illustrative purposes, both have zero capital. They earn income from fair value (or the present discounted value) changes in their assets and liabilities. The yield curve as of April 2012 is taken as a benchmark and it is assumed that the yield curve has now shifted down in a parallel way by 2 percentage points. This is not an unreasonable assumption, given that the 20-year yield-to-maturity on AAA zero coupon bonds (as estimated by the ECB) has indeed shifted down by 2 percentage points.
Based on ECB’s Zero-coupon yield curve for AAA bonds in the euro area, April 4, 2012.
| Life Insurance | Bank | Yield to Maturity 1/ | ||||
| Years to maturity (t) | Asset Cash Flow | Liability Cash Flow | Asset Cash Flow | Liability Cash Flow | In April 2012 | Parallel Shift in Yields = −2% |
| 1 | 200 | 10 | 10 | 200 | 0.149 | −1.851 |
| 2 | 200 | 10 | 10 | 200 | 0.389 | −1.631 |
| 3 | 200 | 20 | 20 | 200 | 0.668 | −1.332 |
| 4 | 200 | 30 | 30 | 200 | 0.996 | −1.004 |
| 5 | 200 | 100 | 100 | 200 | 1.32 | −0.68 |
| 7 | 50 | 180 | 180 | 50 | 1.892 | −0.108 |
| 10 | 50 | 200 | 200 | 50 | 2.486 | 0.486 |
| 15 | 50 | 250 | 250 | 50 | 2.924 | 0.924 |
| 20 | 50 | 400 | 40O | 50 | 3.013 | 1.013 |
| 1200 | 1200 | 1200 | 1200 | |||
Based on ECB’s Zero-coupon yield curve for AAA bonds in the euro area, April 4, 2012.
| Life Insurance | Bank | Yield to Maturity 1/ | ||||
| Years to maturity (t) | Asset Cash Flow | Liability Cash Flow | Asset Cash Flow | Liability Cash Flow | In April 2012 | Parallel Shift in Yields = −2% |
| 1 | 200 | 10 | 10 | 200 | 0.149 | −1.851 |
| 2 | 200 | 10 | 10 | 200 | 0.389 | −1.631 |
| 3 | 200 | 20 | 20 | 200 | 0.668 | −1.332 |
| 4 | 200 | 30 | 30 | 200 | 0.996 | −1.004 |
| 5 | 200 | 100 | 100 | 200 | 1.32 | −0.68 |
| 7 | 50 | 180 | 180 | 50 | 1.892 | −0.108 |
| 10 | 50 | 200 | 200 | 50 | 2.486 | 0.486 |
| 15 | 50 | 250 | 250 | 50 | 2.924 | 0.924 |
| 20 | 50 | 400 | 40O | 50 | 3.013 | 1.013 |
| 1200 | 1200 | 1200 | 1200 | |||
Based on ECB’s Zero-coupon yield curve for AAA bonds in the euro area, April 4, 2012.
A useful concept of capturing the impact of interest rate changes on assets and liabilities, and the net impact on income through fair value changes, is “duration”. Duration measures the maturity of assets or liabilities by taking into account the size and timing of payments between now and maturity. What matters is the time remaining to maturity rather than the original maturity. It is a direct measure of the interest rate sensitivity or elasticity of an asset or liability to changes in interest rate. Macaulay Duration, a simple measure of duration, is given by:
The larger the numerical value of duration (in years), the more sensitive the market value of the assets or the liabilities is to the changes in interest rates.1
The duration of assets and liabilities for the hypothetical life insurance company is given in Table 1 below. For a positively sloped yield curve, the PDVs of liabilities get smaller the longer the maturity. In contrast, shorter dated assets are discounted much more.
Duration of Assets and Liabilities of the Insurance Company
Duration of Assets and Liabilities of the Insurance Company
| Interest Rates: as of April 2012 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Assets | Liabilities | ||||||||||
| Years to maturity (t) | Asset Cash Flow (CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow* time tomaturity | Years to maturity (t) | Liability Cash Flow(CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow * time to maturity |
| [1] | [2] | [3] | [4]=[3]*[1] | [5]=[4]*t | [1] | [2] | [3] | [4]=[3]*[1] | [5]=[4]*t | ||
| 1 | 200 | 0.149 | 0.999 | 199.7 | 199.7 | 1 | 10 | 0.149 | 0.999 | 10.0 | 10.0 |
| 2 | 200 | 0.369 | 0.993 | 198.5 | 397.1 | 2 | 10 | 0.369 | 0.993 | 9.9 | 19.9 |
| 3 | 200 | 0.888 | 0.980 | 196.0 | 588.1 | 3 | 20 | 0.668 | 0.980 | 19.6 | 58.8 |
| 4 | 200 | 0.996 | 0.961 | 192.2 | 768.9 | 4 | 30 | 0.996 | 0.961 | 28.8 | 115.3 |
| 5 | 200 | 1.32 | 0.937 | 187.3 | 936.5 | 5 | 100 | 1.32 | 0.937 | 93.7 | 468.3 |
| 7 | 50 | 1.892 | 0.877 | 43.9 | 307.0 | 7 | 180 | 1.892 | 0.877 | 157.9 | 1105.1 |
| 10 | 50 | 2.486 | 0.782 | 39.1 | 391.1 | 10 | 200 | 2.486 | 0.782 | 156.5 | 1564.5 |
| 15 | 50 | 2.924 | 0.649 | 32.5 | 486.8 | 15 | 250 | 2.924 | 0.649 | 162.3 | 2433.8 |
| 20 | 50 | 3.013 | 0.552 | 27.6 | 552.3 | 20 | 400 | 3.013 | 0.552 | 220.9 | 4418.2 |
| 1116.8 | 4627.5 | 859.5 | 10193.9 | ||||||||
| Duration DA = Σ[5]/Σ[4] years | 4.1 | Duration DL = Σ[5]/Σ[4] years | 11.9 | ||||||||
Duration of Assets and Liabilities of the Insurance Company
| Interest Rates: as of April 2012 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Assets | Liabilities | ||||||||||
| Years to maturity (t) | Asset Cash Flow (CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow* time tomaturity | Years to maturity (t) | Liability Cash Flow(CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow * time to maturity |
| [1] | [2] | [3] | [4]=[3]*[1] | [5]=[4]*t | [1] | [2] | [3] | [4]=[3]*[1] | [5]=[4]*t | ||
| 1 | 200 | 0.149 | 0.999 | 199.7 | 199.7 | 1 | 10 | 0.149 | 0.999 | 10.0 | 10.0 |
| 2 | 200 | 0.369 | 0.993 | 198.5 | 397.1 | 2 | 10 | 0.369 | 0.993 | 9.9 | 19.9 |
| 3 | 200 | 0.888 | 0.980 | 196.0 | 588.1 | 3 | 20 | 0.668 | 0.980 | 19.6 | 58.8 |
| 4 | 200 | 0.996 | 0.961 | 192.2 | 768.9 | 4 | 30 | 0.996 | 0.961 | 28.8 | 115.3 |
| 5 | 200 | 1.32 | 0.937 | 187.3 | 936.5 | 5 | 100 | 1.32 | 0.937 | 93.7 | 468.3 |
| 7 | 50 | 1.892 | 0.877 | 43.9 | 307.0 | 7 | 180 | 1.892 | 0.877 | 157.9 | 1105.1 |
| 10 | 50 | 2.486 | 0.782 | 39.1 | 391.1 | 10 | 200 | 2.486 | 0.782 | 156.5 | 1564.5 |
| 15 | 50 | 2.924 | 0.649 | 32.5 | 486.8 | 15 | 250 | 2.924 | 0.649 | 162.3 | 2433.8 |
| 20 | 50 | 3.013 | 0.552 | 27.6 | 552.3 | 20 | 400 | 3.013 | 0.552 | 220.9 | 4418.2 |
| 1116.8 | 4627.5 | 859.5 | 10193.9 | ||||||||
| Duration DA = Σ[5]/Σ[4] years | 4.1 | Duration DL = Σ[5]/Σ[4] years | 11.9 | ||||||||
When interest rates fall (Table 2), the PDV of liabilities rises more than assets. In this example, the PDV of liabilities increases by 233 [=1092.7 – 859.5], whereas, the PDV of assets increases by only 99 [=1216.3 – 1116.8]. So, the insurance company makes a loss.
Duration of Assets and Liabilities of the Insurance Company, Lower Yields
Duration of Assets and Liabilities of the Insurance Company, Lower Yields
| Interest Rates: Parallel Shift down by 2 percentage points | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Assets | Liabilities | ||||||||||
| Years to maturity (t) | Asset Cash Flow (CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow* time to maturity | Years to maturity (t) | Liability Cash Flow(CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow * time to maturity |
| [1] | [2] | [3] | [4] = [3]* [1] | [5] = [4]*t | [1] | [2] | [3] | [4] = [3]*[1] | [5] = [4]*t | ||
| 1 | 200 | −1.851 | 1.019 | 203.8 | 203.8 | 1 | 10 | −1.851 | 1.019 | 10.2 | 10.2 |
| 2 | 200 | −1.631 | 1.033 | 206.7 | 413.4 | 2 | 10 | −1.631 | 1.033 | 10.3 | 20.7 |
| 3 | 200 | −1.332 | 1.041 | 208.2 | 624.6 | 3 | 20 | −1.332 | 1.041 | 20.8 | 62.5 |
| 4 | 200 | −1.004 | 1.041 | 208.2 | 833.0 | 4 | 30 | −1.004 | 1.041 | 31.2 | 124.9 |
| 5 | 200 | −0.68 | 1.035 | 206.9 | 1034.7 | 5 | 100 | −0.68 | 1.035 | 103.5 | 517.4 |
| 7 | 50 | −0.108 | 1.008 | 50.4 | 352.7 | 7 | 180 | −0.108 | 1.008 | 181.4 | 1269.6 |
| 10 | 50 | 0.486 | 0.953 | 47.6 | 476.3 | 10 | 200 | 0.486 | 0.953 | 190.5 | 1905.3 |
| 15 | 50 | 0.924 | 0.871 | 43.6 | 653.3 | 15 | 250 | 0.924 | 0.871 | 217.8 | 3266.7 |
| 20 | 50 | 1.013 | 0.817 | 40.9 | 817.4 | 20 | 400 | 1.013 | 0.817 | 327.0 | 6539.5 |
| 1216.3 | 5409.2 | 1092.7 | 13716.8 | ||||||||
| Duration DA = Σ[5]/Σ[4] years | 4.4 | Duration DL = Σ[5]/Σ[4] years | 12.6 | ||||||||
Duration of Assets and Liabilities of the Insurance Company, Lower Yields
| Interest Rates: Parallel Shift down by 2 percentage points | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Assets | Liabilities | ||||||||||
| Years to maturity (t) | Asset Cash Flow (CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow* time to maturity | Years to maturity (t) | Liability Cash Flow(CF) | Yield Curve = R (%) | Discount factor (DF) = 1/(1+R/100)t | Discounted Cash Flow | Discounted Cash Flow * time to maturity |
| [1] | [2] | [3] | [4] = [3]* [1] | [5] = [4]*t | [1] | [2] | [3] | [4] = [3]*[1] | [5] = [4]*t | ||
| 1 | 200 | −1.851 | 1.019 | 203.8 | 203.8 | 1 | 10 | −1.851 | 1.019 | 10.2 | 10.2 |
| 2 | 200 | −1.631 | 1.033 | 206.7 | 413.4 | 2 | 10 | −1.631 | 1.033 | 10.3 | 20.7 |
| 3 | 200 | −1.332 | 1.041 | 208.2 | 624.6 | 3 | 20 | −1.332 | 1.041 | 20.8 | 62.5 |
| 4 | 200 | −1.004 | 1.041 | 208.2 | 833.0 | 4 | 30 | −1.004 | 1.041 | 31.2 | 124.9 |
| 5 | 200 | −0.68 | 1.035 | 206.9 | 1034.7 | 5 | 100 | −0.68 | 1.035 | 103.5 | 517.4 |
| 7 | 50 | −0.108 | 1.008 | 50.4 | 352.7 | 7 | 180 | −0.108 | 1.008 | 181.4 | 1269.6 |
| 10 | 50 | 0.486 | 0.953 | 47.6 | 476.3 | 10 | 200 | 0.486 | 0.953 | 190.5 | 1905.3 |
| 15 | 50 | 0.924 | 0.871 | 43.6 | 653.3 | 15 | 250 | 0.924 | 0.871 | 217.8 | 3266.7 |
| 20 | 50 | 1.013 | 0.817 | 40.9 | 817.4 | 20 | 400 | 1.013 | 0.817 | 327.0 | 6539.5 |
| 1216.3 | 5409.2 | 1092.7 | 13716.8 | ||||||||
| Duration DA = Σ[5]/Σ[4] years | 4.4 | Duration DL = Σ[5]/Σ[4] years | 12.6 | ||||||||
The larger impact on liabilities is also captured by the change in duration. For the same balance sheet structure, the duration of assets increases by [4.4 – 4.1] 0.3 year and the duration of liabilities increases by [12.6 – 11.9] 0.7 year. Thus, the larger the (negative) duration mismatch between assets and liabilities, the greater the impact when interest rates fall.
The approximate impact of changes in capital or own funds can be captured by a simple formula, where L is liabilities, A is Assets, DL is duration of liabilities, DA is duration of assets, R is interest rate in percent, ΔR is change in interest rate in percentage points:
Thus, larger the duration mismatch between liabilities and assets, the larger is the impact on capital.
The hypothetical bank, with a profile of assets and liabilities that is mirror-opposite of the insurers’, reaps gains with an interest rate fall. Its duration of assets increases by 0.7 and that of its liabilities increases by 0.4. Thus, the bank gains from the positive duration mismatch between assets and liabilities and its profits increase.
When interest rates fall, the insurance company can protect itself by being flexible on the payouts of its long-dated liabilities. The higher the prevalence of guaranteed returns to clients, which makes the long-term cash flow constant, the insurance company loses from the increase in duration mismatch.
Appendix III. EIOPA 2016 Assessment of Risks Under the “Low for Long” and “Double Hit” Scenarios
EIOPA Stress Test Results: Impact on Assets/Liabilities (Ratio)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA.Note: Shaded areas show the effect of the LTG and transitional measures.EIOPA Stress Test Results: Impact on Assets/Liabilities (Ratio)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA.Note: Shaded areas show the effect of the LTG and transitional measures.EIOPA Stress Test Results: Impact on Assets/Liabilities (Ratio)
Citation: IMF Staff Country Reports 2017, 236; 10.5089/9781484312353.002.A005
Source: EIOPA.Note: Shaded areas show the effect of the LTG and transitional measures.Prepared by Srobona Mitra.
See Appendix II for stylized calculations of duration and the sensitivity of the PDV of assets and liabilities to interest rate declines.
Insurance and pensions sectors are supervised by national authorities. EIOPA is an independent advisory body to the European Commission and one of the three European Supervisory Authorities that ensure an effective and consistent level of regulation and supervision across member states. EIOPA’s powers include issuing guidelines and recommendations and developing draft regulatory and implementing technical standards. It conducts stress tests for EU insurers, in cooperation with ESRB, ECB and EBA on scenarios.
Gourinchas and Rey (2016) predict that the global average real short risk-free rate for 2015–25 will be only −2 percent. With an inflation target of 2 percent, that implies that the nominal long-term rate would be about 0 percent for 2015–25. Moreover, the authors show that the average real rate for the U.S. over more than a hundred years has been around zero percent.
Other forms include traditional life insurance with profit contracts, savings products, endowment policies; health insurance with guarantees of wage maintenance in the event of illness, etc. Maximum guaranteed rates are under the capacity of the national regulatory authorities. Although these rates are adjusted in some countries, there is no fixed mechanism of such adjustments set at the EU level so far.
The discussion on LTG measures in this note refer both to the long-term guarantee (LTG) package and transitionals. LTG measures aim to mitigate artificial volatility in balance sheets that does not reflect changes in the financial position or risk exposure of an insurer. These measures include volatility and matching adjustments to discount rates, the extrapolation of the long-term risk-free interest rate, transitional measures for the calculation of liabilities and the possibility for an extension of the recovery period under exceptional market conditions (EIOPA, 2016b, ECB, 2015). The transitional measures relate to the transition from Solvency I capital requirements to Solvency II requirements. Insurers apply a transitional adjustment to the risk-free rate for the valuation of insurance and reinsurance obligations based on the difference between discount rates of Solvency I and the risk-free rates. Over the 16-year transition period, these transitional measures would be reduced to zero.
Austrian, Dutch and German insurers were more impacted by the low-for-long scenario owing to large duration mismatches or long duration of liabilities, or both.
Regulators, such as in Germany, recognize the insufficiency of the SCR-coverage and require insurers to present plans about how they intend to achieve a sufficient SCR-coverage at the end of the 16-year transitional period for the full adoption of Solvency II (European Commission, 2016).
It should be noted that the 2016 EIOPA stress test results did not specify a pass-fail criterion. The calculations on SCR-coverage shortfall are based on the results of the two scenarios—changes in assets and liabilities—published by EIOPA. The more conservative threshold for the SCR ratio—the one without adjustments for LTGs—is used to calculate the SCR-coverage shortfall in this note.
See IMF’s Financial Soundness Indicators’ Compilation Guide at https://www.imf.org/external/pubs/ft/fsi/guide/2006/ Appendix VI, for a more discussion of duration. The Macaulay duration concept is usually used for fixed income assets and may not be fully suitable for ICPFs. This is because this (or other duration measures) do not account for the fact that some cash flows can be adjusted with interest rates—such as, with-profit guarantees—and the duration of some classes of assets, like equities, is difficult to define.
