VoxEU Column Financial Markets Institutions and economics

The invisible service: The economics, regulation, and systemic risk of insurance markets

Despite the importance of insurance, discussions about the macroeconomic role and the risks of insurance markets have been surprisingly limited. This column explores some of the key theoretical and conceptual questions still unanswered in this field, and suggests that a two-fold approach combining a focus on individual firms and an activity-based approach across the sector is needed to tackle systemic risk within the insurance industry.

The insurance sector in the Eurozone manages €7.3 trillion in assets, employs about one million people directly, plus many outsourced employees and independent intermediaries, and it has virtually every household and firm as a client (ECB 2016). By managing risks across clients and investors, insurance companies enable individual and collective social, economic, and financial activities. In addition, given the long-term nature of insurance, insurance companies are important long-term investors in global financial markets.

Discussions about systemic risk gained attention following the failure of AIG in 2008 (McDonald and Paulson 2015), and the stress in the sector during the financial crisis more broadly (Koijen and Yogo 2015), as well as the subsequent low interest rate environment. In response to these events, nine large global companies have been qualified as ‘systemically important’ by the Financial Stability Board and are facing new supervisory and prudential measures and standards, in addition to their long-standing national prudential regulatory frameworks. While these new regulatory frameworks and qualifications are all nascent, many conceptual and empirical questions remain unanswered regarding the nature and intensity of systemic risk in the insurance sector.

With a recently published book (Hufeld et al. 2016), we aim to contribute to this debate. Here, we summarise key findings and highlight questions going forward for academics, policymakers, and the insurance sector.

Insurers and other financial actors – similarities and differences

In designing regulatory frameworks, it is important to understand the unique features of the modern insurance sector. One key feature is that insurance liabilities are mostly long-term and relatively illiquid. While liquidity risk is inherent in banking, it is not in insurance. Bank liabilities are predominantly short-term and withdrawable at will. Insurance liabilities are less fugitive. The liabilities for insurance of general protection, property, casualty, and health are not callable at will. They mostly relate to exogenous events that policyholders do not easily influence, or are explicitly devoted to long-term savings. If liabilities are callable, then there are often penalties for early withdrawal, and tax benefits might vanish. This implies that one cannot translate bank regulation one-for-one to insurance companies. Second, it is important to distinguish idiosyncratic risks that are unique to the insurance sector, such as property, health, and life risks, and aggregate financial risks coming from modern insurance products (for instance, variable annuities) and modern tools to manage capital (for instance, securities lending, shadow insurance, and derivatives). Aggregate risk mismatch can expose the insurance sector to common shocks, even if insurance companies are not directly connected to each other.

Insurance contributes to economic growth, stabilisation, and distribution

The size and macroeconomic role of the insurance sector is often measured by insurance premiums relative to GDP and the number of employees in the sector. Over time and across countries, GDP and the size of the sector are positively correlated (Figure 1).

Figure 1. GDP and insurance

Note: The top panel plots the correlation between fire insurance and growth in the UK. The bottom panel plots the increasing insurance penetration versus economic growth.

However, insurance is a rather ‘invisible’ service affecting virtually all individuals and many investments and innovative activities. Although there are some estimates of the welfare benefits of various insurance markets, more research is needed to quantify the macroeconomic role of the insurance. Kessler et al. (2016) discuss measuring the impact of the insurance sector on economic growth (by optimising savings and investments, lowering interest rates, and easing access to credit) and on improved risk sharing (across firms and households, and between generations). Given the important role insurance companies play in promoting innovative activities, insurance regulation needs to keep pace with the risks covered by insurance companies. Examples are the accumulation of cyber risk and the need for a new liability regime in a world growingly exposing humans to robots (starting with driverless cars).

Insurance companies use a rich set of tools for risk management purposes

Modern insurance companies pool risks from consumers and firms. Duverne and Hele (2016) discuss the risk management tools modern insurers use to manage these pools of risks, summarised in Figure 2. To manage aggregate financial market risks, particularly interest rate and equity risks, insurance companies increasingly use derivatives. In addition, risks are managed via product design or by transferring risk to reinsurance companies and broader capital markets (e.g., via insurance-linked securities).  

Figure 2. Overview of risks in the insurance sector

Low nominal long-term interest rates raise new challenges of risk management for insurers

Traditional sources of risk for life insurers are uncertainty in interest rates, aggregate longevity or mortality, and policyholder behaviour. Although insurance companies have extensive experience managing interest rate risk, traditional hedging techniques such as duration matching are less accurate when interest rates change in largely unanticipated ways, such as the extended period of very low rates following the financial crisis.

Hartley et al. (2016) propose a top-down approach to measure the interest rate exposure of life insurers before and after the financial crisis. They examine the sensitivity of insurer stock prices to changes in interest rates while controlling for changes in the overall stock market.  This approach allows them to examine hedging at the corporate group-level, as well as across countries, even in the absence of detailed balance sheet data.

Figure 3. US insurers’ equity return to 10-year bond returns

Note: The “Coeff on Share Life * 10 Yr Return” measures the differential exposure of US insurers’ equity return to 10-year bond returns for insurance companies with and without life insurance policies.

While equity prices of insurance companies in the US were insensitive to interest rates before the financial crisis, and this is still the case for non-life insurance companies, equity prices of life insurers now decrease when bond returns are positive (and hence when yields fall) (see Figure 3).  However, in the UK, stock prices of life insurers are relatively insensitive to interest rates, both before and after the financial crisis. Hartley et al. (2016) provide evidence suggesting that the difference in interest rate exposure across the US and the UK is, in part, due to guaranteed-return products that make up a significant portion of life insurers’ liabilities in the US but not in the UK.

The risk profile of insurance companies has changed during the last two decades

In addition to these traditional risks, the risk profile of insurance companies changed significantly during the last 20 years, due to product innovation (variable annuities) and new tools that insurance companies use to manage their capital (securities lending, new reinsurance schemes between affiliated companies – shadow insurance – and derivatives). Koijen and Yogo (2016b) measure the trends in these activities from 2002 to 2014 in the US and use the financial crisis as a case study to quantify the risks (see Koijen and Yogo 2016a for a more detailed discussion on shadow insurance).

In particular, variable annuities and securities lending caused large losses during the financial crisis. Figure 4 shows the losses from annuities, both in billions and as a share of capital and surplus (insurance companies’ equity). Variable annuities are long-term savings products that combine mutual funds with long-term minimum-return guarantees. With equity prices and bond yields falling, these products experienced large losses. Variable annuities account for 40% of the liabilities of US life insurers, which are often subsidiaries of European insurers. There is an ongoing discussion how to value, hedge, and regulate variable annuities.

Figure 4. Losses from variable annuities

Measuring systemic risk for insurance companies

The changing risk profile affected the systemic risk of insurance companies. Acharya et al. (2017) build a framework to measure systemic risk in the insurance sector. The model incorporates externalities arising from an aggregate capital shortfall, which leads to a reduction in intermediation activity (going-concern externality), and from fire sales caused by the degree to which liabilities are liquid and under the threat of potential runs (which is a smaller concern for insurers than for banks, as discussed before).

While measuring the cost of going-concern externalities and fire sales externalities is non-trivial, relative measures of systemic risk can be computed from asset prices. The NYU Stern Systemic Risk Rankings (SRISK) approximate how much a firm’s market value of equity falls below a fraction of the firm’s total assets when a crisis hits. Figure 5 plots the dynamics of SRISK for Metlife, Prudential, and Lincoln National to measure the going concern externality. Quite strikingly, and different than for banks, SRISK did not decline following the financial crisis.

Figure 5. The dynamics of SRISK for Metlife, Prudential, and Lincoln National

Capital regulatory frameworks can have unintended consequences

Risk regulation prevents insurance companies from taking too much risk on the asset side. In the US, the capital requirements of corporate bonds are tied to a bond’s rating. However, within each rating category, the regulation is insensitive to risk and insurance companies have an incentive to take on more risk and mostly in relatively good times (Becker and Ivashina 2015, Becker 2016) (see Figure 6). This evidence highlights the importance of adequate risk regulation as risk insensitivities may lead to excessive risk taking; see also Becker and Opp (2014) in the context of mortgage-backed securities.

Figure 6. Holdings of insurance companies and yield spreads

Note:  The top panel shows the holdings of insurance companies in new issuances across rating categories. The bottom panel shows the yield spreads, or CDS spreads, within rating categories.

Interactions between the accounting and regulatory frameworks

The current risk regulation may lead to fire sales by coordinating sales of downgraded assets. Once a security is downgraded, an insurer needs to hold more capital or sell it. This provides an incentive for constrained insurers to sell the security (Ellul et al. 2011).

The risk regulation interacts in important ways with the accounting regime. Property and casualty (P&C) insurers in the US must mark assets to market, while US life insurers can record securities at historical cost. Thus, when a bond gets downgraded, a P&C insurer may be indifferent between selling the bond and holding it. A life company, by contrast, may prefer not to sell the bond to avoid realising the loss. To make up for the additional capital that is required, a life company instead has an incentive to sell a security that trades at a gain. Ellul et al. (2016) summarise this behaviour for asset-backed securities during the financial crisis (see Figure 7). Taken together, the interaction between the accounting regime and risk regulations has important implications for insurers’ trading incentives, asset prices, and potentially systemic risk.

Figure 7. Fraction of downgraded ABS that were kept at historical cost (HCA), kept but revalued, or sold

A regulatory framework for systemic risk in the insurance industry

The existence of systemic risk is a key lesson from the financial crisis of 2007-8. This is true for all sectors of the financial industry, including the insurance and reinsurance industry. In terms of systemic risk in the insurance industry, a hybrid approach – combining a focus on individual firms and an activity-based approach across the sector – must be applied. It would combine two conceptually different regulatory frameworks that are largely driven by the distinction between direct and indirect systemic risk. Direct systemic risk represents the potential consequences on the financial system from a single insurer, while indirect systemic risk refers to the potential negative consequences based on the collective behaviour of many insurance undertakings. General insurance regulation has so far been neglected as an important potential tool to address indirect systemic risk and financial stability concerns in addition to traditional policyholder protection. It needs to be sharpened with a view as to how to consider financial stability issues as well. Only the combination of both approaches into a unified hybrid approach will ensure a comprehensive regulatory response to systemic risk in the insurance industry.


Acharya, V, L Pedersen, T Philippon, and M Richardson (2017), “Measuring Systemic Risk”, The Review of Financial Studies, 30, 2-47. 

Becker, B (2016), “How the Insurance Industry’s Asset Portfolio Responds to Regulation”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

Becker, B, and V Ivashina (2015), “Reaching for Yield in the Bond Market”, Journal of Finance, 70, 1863-1902.

Becker, B, and M Opp (2014), “Regulatory Reform and Risk-taking: Replacing Ratings”, Working paper, Stockholm School of Economics.

Duverne D and J Hele (2016), “How the Insurance Industry Manages Risk”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

ECB (2016), “Insurance corporations and pension funds online”, Statistical Data Warehouse, September.

Ellul, A,  C Jotikasthira, and C Lundblad (2011), “Regulatory Pressure and Fire Sales in the Corporate Bond Market”, Journal of Financial Economics, 101, 596-620.

Ellul, A, C Jotikasthira, C Lundblad, and Y Wang (2015), “Is Historical Cost a Panacea? Market Stress, Incentive Distortions and Gains Trading”, Journal of Finance, 2489-2537.

Ellul, A,  C Jotikasthira, and C Lundblad (2016), “Spillover Effects of Risk Regulation on the Asset Side to Asset Markets”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

Hartley, D, A Paulson and R Rosen (2016), “Measuring Interest Rate Risk in the Life Insurance Sector: the U.S and the U.K.”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

Hufield, F, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

Kessler, D, A de Montchalin and C Thimann (2016), “The Macroeconomic Role of Insurance”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

Koijen, R S J, and M Yogo (2015), “The Cost of Financial Frictions for Life Insurers”, American Economic Review, 105, 445-475.

Koijen, R S J, and M Yogo (2016a), “Shadow Insurance”, Econometrica, 84, 1265-1287.

Koijen, R S J, and M Yogo (2016b), “Risk of Life Insurers: Recent Trends and Transmission Mechanisms”, in F Hufield, R S J Koijen, and C Thimann (eds) (2016), The Economics, Regulation, and Systemic Risk of Insurance Markets, Oxford University Press.

McDonald, R, and A Paulson (2015), “AIG in Hindsight”, Journal of Economic Perspectives, 29, 81-106.

4,409 Reads