The idea of gross premium valuations (GPV) is to obtain a best estimate of the amounts required in the future for a particular policy. In simple terms this is the future outgo less future premium income. We add to this some sort of explicit conservatism, which can be called risk adjustment, risk margin (RM), padding, buffer or some other name, and we have the final GPV. This calculation is useful in determining the market value of liabilities and ensuring consistency between insurers. The drawback of the GPV calculated in this manner is that future profits are released once the policy is incepted.
Consider the following example:
- Premium: USD1,000, payable annually
- Policy term: 5 years
- Total expenses and benefits (over the 5 year term): USD3,000
- Risk adjustment: USD100
Ignoring discounting, the total premiums is USD5,000. Thus, when the policy is first incepted, the policyholder pays an initial premium of USD1,000 and sets aside a GPV reserve of –USD 1,000. (i.e. USD3,000 – USD 4,000). We would need to add the RM of USD100, which still gives a negative reserve of –USD 900 along with the initial premium of USD1,000 which is now listed as an asset. Thus, once this policy first starts, the first premium of USD1,000 is received and becomes an asset and the GPV reserve is –USD900, again an asset (since expected income is greater than expected outgo). This USD1,000 plus USD900 represents the future profit of the policy (along with the RM of USD100). Whilst this is a fair and true description of the market value of this liability, it feels risky to allow shareholders to take this future surplus as a dividend as it as of now, doesn’t actually fully exist yet (only USD1,000 initial premium was received). This is why there are Risk Based Capital (RBC) guidelines which determine the potential risks such as liability, asset, ALM, operational and whatnot. We calculate the value of these risks and call this the risk capital requirement (RCR). This is, compared to the assets which are available, the Total Capital Available (TAC), to get the Capital Adequacy Ratio (CAR). Based on the RBC Forms provided by the regulator (IRCSL), the CAR is computed based on the following formula:
In the example above, when the policy is first incepted, assets of USD1,900 are generated. We would then determine the RCR to get the CAR. For instance, if the RCR is USD500, then the CAR = USD1,900 / USD500 = 380%. Depending on what CAR we are targeting, this determines if there can be a dividend to shareholders and how much is acceptable.
Using Sri Lanka as an example, a very popular product in Sri Lanka is universal lifestyle products, whereby a fund value is built up from the premiums paid, and amounts are taken from this fund to cover expenses and benefits.
The fund value is paid out at maturity as well as upon surrender, less a surrender charge. The open question is whether the GPV should be calculated as per normal, or if there should be some sort of minimum reserve equal to the fund value (on a policy-by-policy basis, or on a plan basis or on a fund basis)?
Reserving under GPV versus Fund Values for Universal Life Plans
Although there are several technical reasons for the GPV to be different than the fund value, at a basic level, the issue is the discount rate used versus the assumed investment returns in the fund (assuming the fund charges match the expense assumptions). If the discount rate is higher than the assumed investment return, then the GPV will be lower than the fund value. Does this make sense? Should the insurer be forced to ‘top up’ the deficit in reserves between the actual investment returns and the GPV discount rate? One argument is that if there are mass surrenders, this will open up a risk to the insurer (since surrender values are usually set as the fund value), as the GPV is lower than the fund value, but we have a specific RBC charge for this, called Surrender Value Capital Charge (SVCC).
So, would it be better to reserve the fund value instead of GPV? Reserving the fund value would definitely be more conservative and mitigate the risk of mass surrenders, but on the other hand, this would have an impact on the solvency of the insurer. As discussed above, the solvency would depend on two things, the TAC and the RCR.
Under a GPV valuation, new business policies with regular premium payments would usually have negative reserves, since for profitable policies, future premiums would be higher than future benefits and expenses (ignoring the RM). This would reduce the overall reserves on a portfolio level. These negative reserves are ‘eliminated’ to a great extent through the solvency requirements. However, depending on the product structure, fund values are usually positive from the day the policy is incepted as premiums are received. Hence, especially for new business policies, reserves would be much higher should we consider the fund values instead of GPV. Furthermore, part of the TAC is actually the surplus arising from the difference between the technical life fund and the policy liabilities based on the valuation results. Reserving under fund values would reduce the TAC significantly (and even a possible deficit!) as per example in illustration below.
On the other hand, under an RBC regime, one may claim that the risk charges (under the RCR) would decrease as well since there is no Liability Risk Capital Charge (LRCC) and Market Risk Capital Charge (MRCC) when we reserve the fund values. The LRCC is calculated by taking the difference between the liabilities with a higher risk margin (e.g., 99.5th percentile reserves) and valuation reserves, which may not be required if we were to reserve fund values (fund values might be higher for both valuation and 99.5th percentile reserves). Similarly, the MRCC is calculated by discounting the assets and liabilities using stressed interest rates, which are no longer applicable now since there are no liability cash flows to discount. Depending on the business mix, for insurers with significant reserves from universal life plans, the TAC would reduce much more significantly when compared to the reduction in RCR. This would worsen the solvency position of the insurer.
To wrap this up, by right we need to follow regulatory rules for GPV TOGETHER WITH RBC RULES. Thus, if we take the fund value as the reserves, we are breaking the connection between GPV and RBC calculations, where in reality the difference between the GPV and fund value can be possibly considered as a risk charge under RBC (e.g., SVCC).