Over the last decade, economic capital has captured the risk management spotlight. Recognizing its merits, insurers have deployed economic capital for many uses. Regulators now rely on it, too — especially internationally — and have put it at the center of their prudential regulatory agenda.
Economic capital (defined as value at risk over a year) has two unique and extremely useful characteristics. First, the concept can be applied to any event with an uncertain outcome where a probability distribution of the outcomes can be postulated. Thus, insurers can value, in a consistent and comparable manner, very different risky events — such as mortality claims, credit losses and catastrophic property damages. Second, economic capital calculated for a portfolio of risks can be readily subdivided into the economic capital attributable to each risk in that portfolio. Or, alternatively, economic capital calculated at the individual risk level can be aggregated to economic capital at the portfolio level and beyond, across portfolios to the enterprise level.
However, there are four critical enterprise risk management (ERM) tasks for which economic capital is not an effective tool; unfortunately, because of this, we have observed a tendency for risk managers to de-emphasize those tasks and sometimes ignore them altogether. We believe this should change.
See also: How to Improve Stress Testing
In response to these shortcomings, insurers should take full advantage of stress testing, a valuable risk management tool that is on par with economic capital in terms of its potential to help solve problems and improve performance. And, because stress testing enables insurers to tackle many of the important tasks that economic capital cannot, it gives insurers the opportunity to double the size of their risk management tool kit and thereby double their ERM output.
By design, economic capital assumes assets and liabilities can be monetized at their formulaic values — that is, at the values derived from the probability distributions’ assumptions. But, as we saw in the credit crisis of 2008-09, credit markets can seize up under extreme stress. When that happens, many assets — regardless of their formulaic value — cannot be sold at any price. Because of this, economic capital is not an effective tool to understand and manage liquidity risk.
To address the risks posed by insufficient liquidity, insurers need to play out meaningful stress events and postulate how they might affect both the ability to monetize assets and the asset’s price if they can be monetized, as well as critically assess the ability to actually access pre-arranged credit in the event these stress events unfold. Then, with an understanding of the likely challenges these stresses may impose, insurers can test the effectiveness of the potential mitigating strategies that they can deploy immediately or when stress events begin to unfold. Selecting and documenting the most effective options can become the insurer’s liquidity risk management game plan.
Diversification is a cornerstone of effective insurance underwriting and risk management. The industry acknowledges the benefit of diversification across similar, independent risks and is able to apply considerable mathematical rigor to measuring this benefit. However, matters become less certain when attempting to quantify diversification across dissimilar risks such as mortality, credit and catastrophe. Extending the benefits of economic capital across risks requires that the capital amounts assigned to different risk types be combined.
Recognizing that extreme outcomes for each risk type are not likely to occur simultaneously, the combined capital requirement is typically calculated as the sum across risk types, with a credit given for diversification.
Deciding how much credit should be assigned for diversification is a critical question in establishing the enterprise’s total required capital. Unfortunately, historical information about the precise interaction of disparate extreme events is sparse.
Empirically, establishing diversification credits is difficult at best and is largely impossible for some combinations. For enterprise risk capital, a best guess may have to suffice. But, just because such a guess is sufficient for the purpose of ascribing required capital, it does not follow that it is sufficient for other purposes — particularly for charting a course of action across all risk types in the event of an extreme risk occurrence.
Stress testing is useful for this purpose. Playing out the series of interactions and events that could follow from a catastrophe such as an epidemic will yield much more actionable information than guessing the magnitude of the diversification credit. Constructing a future scenario that thoughtfully considers how an extreme event in one risk type will have an impact on others is key. These impacts occasionally are asymmetric and not easily accommodated in a standard diversification credit matrix. For example, we can be fairly certain that an extreme drop in equity values will not have significant impact on mortality rates. Conversely, it would seem imprudent to assume that an extreme pandemic would not have any impact on equity values.
In a survey of insurance company board members and CROs that PwC conducted in June, the area where board members felt more attention would be most beneficial was “searching for, understanding and finding ways to address new risks” — meaning risks outside of traditional insurance, credit and market. Upon further discussion with the survey respondents, it became clear that they are not as interested in esoteric dialogues on black swans or unknown unknowns as they are in addressing more practical questions about currently evident business risks. In particular, survey respondents want to understand how those risks could materialize in ways that have an impact on their companies and how to mitigate those impacts.
Using stress testing to map out the impact of these business risks will help insurers assess how serious the risks are. The stress projection can measure the impact on their future financial condition after a risk event. And if the impact is significant, they can further deploy stress testing to map out potential management actions to reduce the risk’s likelihood of impact or mitigate damage if the impact occurs. Having an effective course of action is far better than hoping black swans won’t materialize.
If insurers use only the economic capital tool, then there is a real risk that it will become a hammer, rendering everything in its path a nail. On discovering a new risk, the most likely reaction will be to call for more required capital. However, in the case of, for example, liquidity and business risk, a more effective approach is to use stress testing to create a plan for reducing or eliminating the risk’s impact.
Likewise, seeing economic capital as the sole means of addressing insurer insolvency can lead to an overly restrictive regulatory agenda that focuses only on the economic capital formula. This unfortunately appears to be the case in the development of some required capital standards. We think a more productive approach would be to recognize that no economic capital formula will ever be perfect, nor can one formula fit all business and regulatory needs around the world. Instead, a simpler formula augmented with stress testing can form a more effective, globally consistent solvency management framework.
Moving to the next level
In the paper we published earlier this year about the results of our stress testing survey, we noted that stress testing is well established in the insurance industry. Insurers use it for many purposes, and it has had significant impact. In fact, 36% of survey respondents indicated they have made key decisions markedly differently than prior to or without stress testing. A further 29% indicate stress testing has had a measurable influence (though no single key decision came to mind). The paper also identifies areas where only a little more effort can yield substantial benefit: through a clear definition of stress testing, through more thoughtful stress construction and through building a more robust stress testing platform.
See also: Risk Management: Off the Rails?
To get the most advantage from stress testing, we have two further suggestions: 1) Insurers should apply a governance framework commensurate with stress testing’s status, and 2) insurers should advocate its use in new areas.
A good governance framework should include policies and procedures, documentation, model validation and independent review, as well as review by internal audit. Board and senior management oversight is also important. While our survey report notes that boards usually receive stress testing results from management, we recommend that management engage the board more in the stress selection process.
While stress testing certainly can add additional insight to insurance, credit and market risk analysis, economic capital already provides a good foundation in these areas. We recommend that insurers use stress testing, in particular, to tackle business risks where economic capital is not an effective tool. This includes new threats like cyberterrorism and their reputational impact. Stress testing can also be useful for understanding the risk of missed business opportunities, such as the failure to address how emerging trends in technology and customer behavior may have an impact on future sales and earnings potential.
We believe that the scope for the application of stress testing is as significant as for economic capital. And as with economic capital, once an effective tool comes into use, many more useful risk and business management applications will ensue.