Introduction
A company deciding to retain its risk faces the decision between self-insuring or forming a single-parent captive insurance company. Self-insurance (including large deductible programs) and captive insurance are very similar in nature. In fact, for a single-parent captive writing solely first-party business, captive insurance is just a formal funding vehicle for retaining risk. However, as a formal funding vehicle, a captive has additional expenses that the typical self-insurer would not incur, such as the costs related to captive management, premium taxes, legal services, and board meetings. Also, there is a cost associated with tying up capital in a captive, either via actual hard assets or with letters of credit. Finally, a captive is regulated by its domicile and generally has less flexibility in terms of investing its assets compared to its parent company/owner.
To outweigh these additional expenses, captive insurance must provide several benefits. According to The 2020 Captive Landscape Report from Marsh McLennan, in a survey of single-parent captive owners, the three most popular reasons to pursue captive insurance were the following:
- Serve as a formal medium to fund and insure risk that the parent has decided to retain
- Draft new policy forms
- Provide access to reinsurance markets
These three drivers, while important for the business decisions of the company, are for the most part all non-economic and can’t be quantified. In terms of the Marsh survey, federal tax benefits, arguably considered the most popular economic benefit, was fourth, with 30% of captive owners listing it as a driver for pursuing captive insurance.
For a captive insurer that qualifies, the federal tax benefit is related to the timing of deductions. A captive is allowed to take a federal tax deduction for unpaid amounts on retained risks (also called reserves), whereas a self-insurer can only take deductions for paid amounts on retained risks. While eventually the total amount of federal tax deductions for the claims of the captive and the self-insurer will be the same (assuming the same federal tax rates), the accelerated tax deduction allows a captive to earn investment income on the taxes temporarily saved.
So, how exactly can one determine the economic benefits of a captive as compared to self-insurance? We’ll look at two approaches. The first is a rather well-documented approach comparing the federal tax benefit to the additional expenses incurred by a captive. The second is a new approach that incorporates the opportunity cost of capital required to own and operate a captive.
Approach 1: Comparing the federal tax benefit to additional expenses
In a Milliman study performed for the Washington State Office of Insurance Commissioner,1 a model was built to quantify the federal tax savings of using a captive as compared to self-insuring. This model requires five key inputs:
- IRS discount factors
- A federal marginal tax rate
- An estimate of ultimate losses
- An interest rate for calculating present values
- A payment pattern
We assumed that losses are paid in the middle of each year and that federal taxes2 are paid at the end of each year.
For a simplified scenario, let’s assume that a captive wrote one workers’ compensation insurance policy. The IRS publishes discount factors on an annual basis that vary by coverage. Using the year-end 2021 workers’ compensation discount factors, as well as the current corporate federal marginal tax rate of 21%, we can look at the impact of varying the three other inputs in the model.
First, let’s look at a standard workers’ compensation payment pattern, which is derived from a Milliman review of publicly available insurance industry data. In varying expected ultimate losses on this one policy between $5 million, $10 million, and $25 million, as well as interest rates of 3%, 5%, and 7%, we derive the following results.
Figure 1: Federal tax benefit at standard payment pattern—before expenses
| Interest rate | |||
|---|---|---|---|
| Losses | 3% | 5% | 7% |
| 5,000,000 | 79,757 | 120,998 | 155,316 |
| 10,000,000 | 159,513 | 241,995 | 310,632 |
| 25,000,000 | 398,783 | 604,988 | 776,579 |
Let’s also make a simplified assumption that the captive incurs an additional one-time $125,000 in expenses above what a self-insurer would incur. Figure 2 displays the federal tax benefit after accounting for these additional expenses.
Figure 2: Federal tax benefit at standard payment pattern—after expenses
| Interest rate | |||
|---|---|---|---|
| Losses | 3% | 5% | 7% |
| 5,000,000 | (18,993) | 22,248 | 56,566 |
| 10,000,000 | 60,763 | 143,245 | 211,882 |
| 25,000,000 | 300,033 | 506,238 | 677,829 |
Figures in Figure 1 less $98,750 [= $125,000 x (100% - a 21% tax rate]
As shown in Figure 2, the top left scenario suggests that, economically, the company would be better off self-insuring rather than forming a captive, as the federal tax benefit is less than the additional expenses. At higher interest rates, the company can earn more investment income on the accelerated deduction of losses, implying a higher federal tax benefit. Additionally, there is the impact of economies of scale. The additional expenses to own and operate a captive are largely fixed and typically do not vary based on the size of the program. Thus, all else equal, as a percentage of ultimate losses, a larger captive insurer would generate a higher amount of savings than a smaller captive.
Now, let’s look at the federal tax benefits for a slower payment pattern, where the captive writes excess workers’ compensation. Using a Milliman estimated payment pattern for workers’ compensation losses in excess of $500,000 per claim, Figure 3 displays the federal tax savings over the self-insurance scenario at the same loss amounts and interest rates.
Figure 3: Federal tax benefit at excess payment pattern—after expenses
| Interest rate | |||
|---|---|---|---|
| Losses | 3% | 5% | 7% |
| 5,000,000 | 299,239 | 423,075 | 495,471 |
| 10,000,000 | 697,229 | 944,900 | 1,089,692 |
| 25,000,000 | 1,891,197 | 2,510,376 | 2,872,354 |
Excess payment patterns are typically slower than standard payment patterns, due to the longer lag of claims to pierce the higher layer, as well as the complexity of larger claims. Thus, for a captive writing excess coverage, it can earn more investment income on the taxes temporarily saved than a standard, first-dollar coverage. For all loss amounts and interest rates displayed in Figure 3, the company would be economically better off using a captive.
This model, while helpful in performing a cost-benefit analysis between using a captive or self-insuring, represents the amount of savings after all losses have been paid out. For workers’ compensation insurance, which has a long tail, claim payments for a single policy year can extend out beyond 30 years. As such, the captive must hold assets underlying the reserves, as well as capital and surplus. The parent company could potentially receive a higher return by investing these assets in its business instead. Therefore, there is an opportunity cost of capital that the parent company needs to consider.
To recognize the opportunity cost of capital, we built a model to look at all of the cash flow differences between using a captive and self-insuring.
Approach 2: Modeling cash flows and rates of return
In modeling the cash flows of captives and self-insurers, the same assumptions used in the federal tax savings model in Approach 1 are used. Several additional assumptions are also made, which include the following:
- Initial capital and surplus is set equal to 50% of ultimate losses.
- For each subsequent year, the captive is required to hold capital and surplus of 50% of reserves.
- The captive is 100% owned by the parent.
- The captive dividends3 out its retained earnings to the parent each year.
- The parent company can earn investment income at a different (typically higher) rate than the captive.
The last assumption is extremely important and a key input into the cash flow model. A captive’s assets are ultimately needed in order to pay claims. Thus, they should be invested in fairly liquid assets, which will typically provide a return less than the return that the parent company could earn by investing those funds into their business. If a parent company has funds that are currently not being used to grow its business, then it could be economically beneficial to put those funds in a captive, as these funds would not be providing a return elsewhere.
Many companies rely on a cost of capital approach in contemplating the use of those free funds. Cost of capital is defined as “the minimum return that would be necessary in order to justify undertaking a new project.”4 The cost of capital often varies by industry and company and can be derived by methods such as the capital asset pricing model. With the concept of cost of capital in mind, the primary goal of the cash flow model was to determine the annual rate of return that the parent company is required to achieve, such that the cash balance between the captive scenario and the self-insurer scenario are the same once all losses have been paid. Another way of thinking about this is asking how much extra investment income the parent has to earn relative to the captive to offset the federal tax benefit described above. Before diving into the results of solving for these rates of return, let’s first look at the cash flows in both scenarios.
Self-insurer cash flows
For both the self-insurer and captive scenarios, expenses that would generally be the same, such as claims handling expenses, were ignored. In both scenarios, it is also assumed that the premium paid for the captive policy is equal to the ultimate losses (100% loss ratio).
Figures 4 through 6 below are based on a scenario using $10 million of ultimate losses, a captive annual rate of return on assets of 5%, the standard workers’ compensation payment pattern, and a calculated annual rate of return required to be achieved by the parent company that leads to the ending cash balances between both scenarios being equal.
A self-insurer’s cash flows are straightforward. The self-insurer would allocate an amount equal to the premium and initial capital and surplus (cash inflows)—think of this as a separate bank account from which losses can be paid and from which free cash can be invested. Losses would be paid out in the middle of each year (cash outflow), based on the assumed payment pattern. The self-insurer would also earn investment income on the cash balance (cash inflow), which is earned at the same rate as the parent company mentioned above. The final cash flow item is federal income tax (cash outflow), which is derived from the taxable income associated with the premium, paid losses, and investment income. Figure 4 displays the all-year total cash flows using the assumptions mentioned above.
Figure 4: Cash flows — self-insurer
| Cash flow | Value |
|---|---|
| Cash to support losses (= Captive premium) | 10,000,000 |
| Additional free cash (= Captive initial surplus) | 5,000,000 |
| Paid losses | (10,000,000) |
| Investment income | 20,002,708 |
| Federal income taxes | (4,200,569) |
| Ending cash balance | 20,802,139 |
Investment income is earned at an annual rate such that the self-insurer and captive scenarios have the same ending balance.
Captive cash flows
In the captive scenario, there are the cash flows of the captive, as well as the cash flows of the parent company. For the captive, the cash flows are fairly similar to those of the self-insurer.
The captive has the same premium and initial capital (cash inflows), as well as paid losses (cash outflow). The captive also earns investment income (cash inflow) on the cash balance but incurs additional fixed expenses (cash outflow). The captive similarly pays federal income taxes (cash outflow), but this is derived from the discounted incurred losses as allowed by the IRS for insurers, rather than the undiscounted paid losses in the self-insurer scenario. The final cash flow item would be the distribution/dividend from the captive to the parent (cash outflow), which consists of shareholder dividends and the return of the initial capital and surplus, neither of which are federally taxable to the parent company. Since the parent company may earn investment income at a different rate than the captive, any after-tax income, as well as reductions in required capital and surplus, will be distributed back to the parent.
As for the parent, it first receives the distribution of income and the reduction in surplus from the captive (cash inflow). The parent can then earn investment income on these distributions (cash inflow) and subsequently pays federal income taxes on the investment income earned (cash outflow). So, in the captive scenario, the total cash flows would be the sum of the captive cash flows and the parent cash flows. Figure 5 displays the captive scenario all-year total cash flows using the assumptions mentioned above.
Figure 5: Cash flows — captive
| Cash flow | Value |
|---|---|
| Premium | 10,000,000 |
| Initial capital and surplus | 5,000,000 |
| Paid losses | (10,000,000) |
| Additional captive expenses | (125,000) |
| Captive investment income | 3,088,593 |
| Federal income taxes | (622,355) |
| Distribution to parent | (7,341,239) |
| Captive ending balance | 0 |
| Parent distribution from captive | 7,341,239 |
| Parent investment income | 13,460,900 |
| Ending cash balance | 20,802,139 |
Distribution to the parent consists of $5 million in initial capital and surplus and $2,341,239 in shareholder dividends.
As one can see from Figures 4 and 5, the ending cash balances are the same. However, this does not mean they are the same for each year. Figure 6 below displays the cash balances at the end of each year for years 1 through 10.
Figure 6: Ending cash balances — self-insurer vs. captive — years 1 through 10
As mentioned previously, the timing of the cash flows is different, which results in the captive scenario having a higher cash balance at the end of year 1 than the self-insurer. This is largely due to the accelerated tax deductions for the captive, less the additional expenses. The spread in ending cash balances eventually converges to zero as losses are paid out (over 30 years) at the derived parent company investment rate of return. In this scenario, were the captive and the parent company to earn investment income at the same rate of return, the captive would have a higher ending balance. However, as mentioned above, the scenario in Figures 4 through 6 is such that the parent’s rate of return is derived to have the ending cash balances equal. So, exactly how much higher of an annual rate of return does the parent need for the self-insurer and captive scenarios to break even?
Rate spreads
The scenario of using $10 million of ultimate losses, a captive annual return on assets of 5%, and the standard workers’ compensation payment pattern displayed in Figures 4 through 6 shows a cash flow advantage for the captive in year 1 of over $1.3 million, primarily due to the impact of the accelerated tax deduction for loss reserves. In order for the self-insurer to make up this deficit in subsequent years, it must earn investment income at a higher rate of return than the captive. The deficit of over $1.3 million at first seems daunting to overcome in subsequent years. However, due to the self-insurer being able to take deductions for paid losses, combined with the effect of compounding investment income, the self-insurer only needs to earn investment income at an annual rate of 5.3%, a spread from what the captive is earning of only 0.3%! Figure 7 below displays the results at various loss amounts and captive annual rates of return.
Figure 7: Rate spread for parent company — standard payment pattern
| Interest rate | ||||||||
|---|---|---|---|---|---|---|---|---|
| Losses | 1% | 2% | 3% | 4% | 5% | 6% | 7% | 8% |
| 5,000,000 | (0.3%) | (0.3%) | (0.2%) | (0.1%) | 0.0% | 0.1% | 0.2% | 0.3% |
| 10,000,000 | (0.1%) | (0.0%) | 0.1% | 0.2% | 0.3% | 0.4% | 0.4% | 0.5% |
| 25,000,000 | 0.0% | 0.1% | 0.2% | 0.3% | 0.4% | 0.5% | 0.6% | 0.7% |
The red percentages in the top left corner of Figure 7 are negative, signifying that the parent company can earn investment income at a rate less than what the captive may earn and still break even with an investment in a captive. This occurs in scenarios where the additional captive expenses of $125,000 (after accounting for federal income taxes) are greater than the federal tax benefit. Even in the most bottom right scenario, the parent company would only have to earn investment income at a rate that is 0.7% higher than what the captive earns.
Like the federal tax benefit model in Approach 1, let’s look at the rate spreads using the excess workers’ compensation payment pattern.
Figure 8: Rate spread for parent company — excess payment pattern
| Interest rate | ||||||||
|---|---|---|---|---|---|---|---|---|
| Losses | 1% | 2% | 3% | 4% | 5% | 6% | 7% | 8% |
| 5,000,000 | 0.0% | 0.2% | 0.3% | 0.4% | 0.5% | 0.6% | 0.7% | 0.8% |
| 10,000,000 | 0.1% | 0.2% | 0.3% | 0.5% | 0.6% | 0.7% | 0.8% | 0.9% |
| 25,000,000 | 0.1% | 0.2% | 0.4% | 0.5% | 0.6% | 0.7% | 0.9% | 1.0% |
While in all scenarios in Figure 8 the federal tax benefit is greater than the additional captive expenses, the investment rate spreads are still extremely small, with the spread in the bottom right of the figure only 1%.
Is a captive worth it?
Based on the rate spreads above, unless a company decides to write third-party business in its captive, or has some other specific economic benefit, a parent company may be better off economically by self-insuring if the rate spread between what the parent company can earn by investing in its business (cost of capital) and what the parent company can earn in the captive is greater than 1%. Based on a review of target rates of return, it is common for businesses to target rates of return of 10% (or more). So, based on our assumption of a 5% rate of return for the captive, the actual spread is 5% (or greater). Through the results of the cash flow model, we see that the opportunity cost of capital being tied up in the captive is higher than the federal tax benefit.
With actual rate spreads over 5%, the investment income earned on the accelerated tax deductions, while positive, likely won’t outweigh the potential economic benefits of reinvesting funds back into the company. If a company has free funds sitting around that aren’t being used other than investing in liquid assets like a captive might, then investing those funds in a captive may be beneficial. However, assuming the parent company is trying to maximize its return on investment, if the federal tax benefit is the only economic benefit of the captive, more economic benefits could be provided through self-insurance and not tying up funds in a captive.
The results of the cash flow model reiterate the survey results from The 2020 Captive Landscape Report. Captive insurance is worth investing in because of its non-economic benefits, not solely due to federal tax benefits. So, if it’s just federal tax savings they’re after, it’s most likely not a good deal for the parent company, and federal taxes cannot (or should not) be the driver behind setting up a captive. If they are, the parent company either has free cash in low investment assets, has a very low cost of capital, or hasn’t done its homework! But as is almost always the case, federal tax savings don’t drive the formation of captives.
1 Chansky, J.S., Brophy, C.R., Kennerud, D.R., & Holahan, J.T. (January 18, 2021). Captive insurance study: Washington State Department of Revenue and Office of Insurance Commissioner. Retrieved February 28, 2023, from https://www.insurance.wa.gov/sites/default/files/documents/captive-insurance-study.pdf.
2 The focus of this paper is on federal taxes. State taxes were ignored.
3 Dividends are distributed to the parent as shareholder dividends and a return of initial capital and surplus, which are exempt from federal taxes.
4 Hayes, A. (June 13, 2022). Cost of capital: What it is, why it matters, formula, and example. Investopedia. Retrieved February 28, 2023, from https://www.investopedia.com/terms/c/costofcapital.asp.