The next generation of risk cost measurement
As alternative finance programmes such as captive insurance have become more common, owners and managers often struggle to know what is considered reasonably insurable within a captive. This is particularly true for enterprise risk captives (831[b] captives).
These small captives have started to insure risk exposures that are either not available or not economical in the commercial insurance market. Tax authorities have suggested, at least indirectly, that a company’s total cost of risk (TCOR) can be used to determine valid captive usage.
This article will demonstrate why that is inappropriate, and will lay out an alternative metric that could be used instead.
TCOR: an obsolete risk metric
As a concept, TCOR has been around for decades. Its first formalised use in a corporate environment was in 1962 by a Canadian risk manager, Douglas Barlow. He defined the TCOR as “the sum of insurance premiums, self-funded losses, and risk control expenditures, as well as internal and external administrative costs.”
Since the early 1990s, the Risk and Insurance Management Society (RIMS) has captured its members’ TCORs and published the consolidated results in an annual publication. In 1993, a new TCOR standard was created, narrowing the standard definition to exclude risk control costs, which has become the industry standard for TCOR ever since.
Today, TCOR is widely used by risk managers in all industries to keep track of their insurable costs and to benchmark themselves against peers and historical trends. However, despite its wide use, TCOR has some serious limitations.
First, TCOR fails to provide an understanding of a risk management programme’s complete costs, especially looking at a single year of TCOR. Consider this: if a risk manager cancelled all corporate insurance, that company’s TCOR would suddenly look remarkably low. Risk management theory suggests that over time, such a situation would level out as increased self-insured losses are captured, but it may be a long time, if ever, before such costs would show the true impact of such a drastic move.
While it’s extremely rare that a risk manager can convince stakeholders that cancelling all insurance is a solid strategy, avoiding certain lines of coverage and reducing limits or sub-limits to save fixed costs is very common. Such costs not currently included in TCOR include:
- Opportunity costs of non-purchased coverage types/limits;
- Complete agent/broker compensation (including indirect commissions);
- Management time/costs devoted to risk;
- Self-insured loss volatility costs; and
- Loss control costs.
Because TCOR is a snapshot of costs at a moment in time, and isn’t flexible enough to capture avoided fixed costs or the cost of volatility over time, it fails to give a true apples-to-apples picture of risk expenditures.
As a professional risk manager for over 20 years, the lack of a modern metric that expands the definition of TCOR to fill in these gaps is very frustrating. In 2013, two analysts from the insurance broker Marsh proposed a new, expanded cost of risk, called Economic Cost of Risk (ECOR).
It was good to see this effort to fix TCOR, but unfortunately ECOR didn’t go far enough. ECOR started with basic TCOR but added an “implied risk charge”, which is simply cost of capital multiplied by an annualised loss overage (losses above average losses).
For most companies, such cost is immaterial, meaning ECOR isn’t enough of an improvement over TCOR to make a difference. ECOR’s method to capture loss volatility is sound but it doesn’t account for TCOR’s other four failings, the largest of which is the opportunity cost of not purchasing economically affordable limits or types of commercial insurance.
TCOR to OCOR: closing the gaps
There needed to be a way to solve all five TCOR limitations. In attempting to develop it, I established the Opportunity Cost of Risk (OCOR) metric, defined as TCOR + ECOR + opportunity cost of risk appetite.
The components of OCOR are:
- Insurance premiums (including implied costs of non-purchased, optimal coverage/limits: solves the first failing);
- Internal and external administrative costs (including broker fees/commissions, contingent commissions, internal risk department costs, and a percentage of executive management costs to cover the time they spend on risk management issues: solves failings two and three); and
- Fully-developed self-insured losses (including an implied risk charge, and loss prevention/reduction costs: solves problems four and five).
In discussions with industry peers, some have argued that determining “optimal” purchases of insurance, in trying to solve failing number one, is too arbitrary.
But in reviewing hundreds of insurance programmes from all industries over the course of my career, I have developed a rule of thumb for insurability that seems to address this concern: if there is a management consensus of a commercially-insurable risk exposure with a likelihood between 1 in 250 (or 0.4 percent) and 1 in 500 (or 0.2 percent), then it can be considered optimal to insure such exposure.
How can one easily determine this likelihood without hiring an actuary? One can take advantage of insurance companies’ actuarial departments by applying publicly available loss ratios to quoted insurance programmes.
For example, consider the following simplified scenario (Table 1): a company buys $75 million of umbrella/ excess liability insurance. Management clearly believes there is risk exposure here, or such insurance would not be purchased. But why did they buy $75 million? Looking at a recent insurance quote, management could have purchased an additional $25 million for $75,000, or an additional $10 million for $40,000.
Using my rule of thumb, it would not be unreasonable to select option two ($10 million), but would not be a good use of risk capital to buy option one ($25 million). If I were calculating this company’s OCOR, and if the company continued to only buy $75 million, I could reasonably add the expected annual opportunity cost of option two ($20,000) to my sum of premiums.
Table 1: Comparison of opportunity costs
Even though OCOR provides a much better understanding of cost of risk than TCOR, it is also limited to traditionally insurable risk exposures. An even broader metric is needed.
OCOR to XCOR: expanding the scope
The other key failing of TCOR is its limited scope. Modern risk managers are struggling to capture the true costs of risk they have recently been asked to oversee. These risks go well beyond the traditionally insurable hazard risks, to include operational, strategic, and financial risks.
These enterprise risks and their associated costs have not been formally captured in any reliable benchmark survey. Various organisations around the world have tried to define the scope of enterprise risks, the best of which is ISO 31000.
Despite the great work of ISO and others, no risk management standard has been able to effectively capture the true costs of enterprise risk.
Considering this problem led me to create another new cost of risk metric, called Existential Cost of Risk (XCOR). XCOR is defined as the premium that would be charged by an insurer if a company insured all of its risk exposures.
Of course, not all of a company’s risks could be reasonably called insurable, but there is a proxy for such a cost. Using the maths behind finance’s capital asset pricing model (CAPM), one can calculate the amount it costs all stakeholders to insure their interests in a company. XCOR is equivalent to the opportunity cost of stakeholders’ next best investment alternative, and is therefore also equal to a company’s overall cost of capital multiplied by its equity.
Consider the following example (Table 2). Using the miracle of the internet it is possible to calculate the XCOR of Pepsico.
Table 2: The XCOR of Pepsico
If a company like Pepsi were to insure its total enterprise risk it would cost it 13 percent of revenue. Compare that to the average TCOR for all industries as of 2019 of 1 percent of sales or $10 per $1000 of sales.
Applying XCOR to captives
How is this useful to a captive owner or manager? One additional step is needed to answer that question. XCOR can be used as the bright line beyond which no enterprise risk management (ERM) captive premiums can go. It can be the ultimate safe harbour for tax scrutiny of the captive. It doesn’t, however, reveal what captive premium is reasonable.
To answer the reasonableness question, I reviewed the individual daily returns from the last 30 years of the Russell 3000. I identified those stocks that had significant moves (>25 percent) relative to the market that were sustained longer than 30 days, and then determined the primary cause of the move. Finally, I categorised these causes as insurable or not, based on those perils typically included in both commercial and ERM captive insurance.
The result revealed that approximately 60 percent of the causes could have been reasonably insured. Thus, one could use XCOR x 60 percent, or XCORr, as a reasonableness test for captive programmes. In the case of the Pepsico that would give us an XCORr of 7.8 percent of sales.
A new spectrum of cost of risk
With these new metrics, risk managers can apply whichever metric is needed to help them support their risk management objectives. Traditional TCOR is likely to continue as the foundational industry metric, at least in the near term, but I hope ECOR and OCOR will ultimately be absorbed into TCOR as the industry recognises and addresses TCOR’s failings.
Figure 1: The cost of risk spectrum NB cropping
XCOR will be useful for those needing a comprehensive metric for tracking and managing ERM costs as well as captive insurance viability (Figure 1).
Randall Davis is managing partner at Delphi Risk Management. He can be contacted at: firstname.lastname@example.org