Managing different risks together

3.2.1. CENTRAL PROVIDER

There may be different ideas relating to a central provider with functions that do not deal with selection effects but rather with other systemic problems.⁴⁸ Naturally, it cannot manage every type of selection issue (e.g. it cannot man-age problems relating to choosing amongst different products or choosing the annuity’s starting point). Central providers manage selection risk by not allow-ing the risk community to choose between different providers, so that both good and bad risks present themselves at the same provider.

Theoretically, the central provider may be equally established and owned by the state or have a private owner assigned by the state. However, the monopo-listic position in the case of a single, central, privately-owned (and therefore for-profit) organisation would be difficult to justify so in reality only a state-owned provider could be considered here.

At the same time, one can envisage a central system but with a number of providers handling all selection issues. These providers may be for-profit pri-vately owned companies. Providers would have a monopolistic position not in respect of the total portfolio but only from the perspective of having a certain stake within the portfolio of insured people, e.g. people buying an annuity in one particular year, and this would solve the problem of selection.

An opportunity for its realisation is that the government invites tenders for a central provider for a given year. The selected provider would furnish every-one with an annuity (naturally based on mortality projections included in the given groups, and via this would deal with demographic upturns and down-turns, e.g. a baby-boom period). This right would be assigned to a provider in respect of every insured person who retires in a given year, until the death of such people. In follow-up years other providers would receive the same as-signment. (Possibly a restriction could be made, i.e. one provider could apply for this tender only at certain times thus avoiding the development of a mo-nopoly situation). Using this approach, the composition of a portfolio in the

48 Only as an example: this may be the “final provider” should nobody be providing annuities on the market, e.g. as a result of rigid regulations, and a provider to whom the savings of “disappeared members” could be transferred, etc.

hands of a provider would reflect the market average in addition to being ho-mogeneous (since everybody retired in the same year).

This is another example where it is not the entire portfolio of annuities that make up a risk community, yet the majority of selection problems are elimi-nated. This is one advantage of this possibility. The disadvantage is that people retiring in a particular year may find themselves in a worse situation compared to others if they end up with an underperforming provider. An additional dis-advantage is that this can pave the way for corruption, for example by collabo-ration among providers if they conspire among themselves with regard to “who will win” and when. A defence against this would be the strengthening of the Competition Office and an enhancement of competition by allowing providers from outside the country (although from within the EU) to apply for the tender.

Beyond the above facts, in respect to the central provider it is conceivable that centralisation might be implemented, not as regards the entire annuity system, but only in relation to certain elements of it e.g. annuity payments, management of mortality risk (where a central pool would be established), etc. In this case, the function of centralisation will not necessarily be the management of selection effects (with the exception of the central pool) but would have some other, im-portant aim such as cost saving, convenience for clients, etc.

3.2.2. PORTFOLIO UNIFICATION (POOL) AND PREMIUM EQUALISATION AMONGST PROVIDERS

Portfolio unification may be voluntary and therefore (usually) partial; this need not concern every provider although it may happen on a voluntary basis that the market unifies the portfolio. It may also be mandatory and cover the entire market (although theoretically it can also be partial, and only some groups in the market may be forced to unify the portfolio, although this would require very specific reasons).

In the case of voluntary portfolio unification, risk equalisation may include the joint management of investment and mortality risks, yet in this situation it is not worth maintaining separate organisations and instead of unifying the portfolio it becomes more reasonable to merge the organisations.

However, mandatory portfolio unification – basically for these reasons – may only include the management of mortality risk because if investment risk is jointly managed as well, then we are already speaking about a central pro-vider. From another point of view, this means that a mandatory pool may be considered in relation to investment and indexation strategy where risk is in-herent in the investment. Namely, if investment in bonds that are indexed to inflation is mandatory (as in Chile), if need arises a central provider is more

reasonable than a mandatory pool, as in this case the “competing” providers can only compete with each other with respect to cost, and this will probably be a priori lower with a central provider due to the scaled return than via dif-ferent organisations within a system composed of small organisations.

In summary, with respect to mandatory annuities mandatory portfolio unifi-cation may only cover mortality risk and it is worth implementing with a clas-sical yield-return investment and indexation strategy – otherwise it baclas-sically points to the central provider solution.

Portfolio unification refers to joint management of the mortality risk which can happen in different ways. In its minimal form, it means only the (mortali-ty) profit and loss equalisation generated at the provider when concluding annuity insurance; in its maximal form it means a continuous distribution and equalisation of the profit and loss generated during the entire period of opera-tion.

So the minimal form is a premium equalisation mechanism designed to equalise the initial profit and loss generated at different providers that occurs as a result of prohibiting differentiation, i.e. the compulsory use of the unisex premium table,⁴⁹ since it is almost certain that the gender composition of new contracts will differ depending on the provider within a given period, e.g. one year. With respect to equalisation, it is also conceivable that one might take other factors such as level of education into consideration. Equalisation will give a correct result if the otherwise projected unisex mortality table is correct.

As there is a need for a uniform scale for equalisation, with this mechanism it is necessary that the market has available a centrally prepared, unisex, project-ed mortality table. Premium equalisation may technically happen such that the different providers transfer the surplus of the actually collected premium (ac-cording to the unisex premium table) and the premium necessary for reserves according to the differentiated mortality table, and the deficit thus generated will be received from this organisation. This also means that, with this mecha-nism, the differentiated mortality tables necessary for reserving must also be unified for the whole market, so these must be centrally defined and imple-mented by mandate.

One problem with relation to premium equalisation is that profit and loss generated by different providers within a specific period of time do not

49In the EU, differentiation according to gender is prohibited, and this is probably al-so the case in many countries outside the European Union. In this respect, my message is restricted to those countries where such a prohibition is in force. With a lack of such a prohibition, selection problems can be solved by the oldest and probably the best insurance technique: differentiation of the insured according to the risk involved. An example of this solution is Chile, where these problems arise infrequently, and their method of operation cannot be adapted well to Europe.

sarily offset each other, as it is not certain that insured people with favourable and unfavourable characteristics from the insurer’s point of view will take out annuity insurance at the same rate and at all points of time throughout the market. If central mortality projections are correct, the balance of profit and loss will be zero in the long term, but it may not be so in the short term. This would cause a problem in practice if there is a deficit at the central organisa-tion in charge of premium equalisaorganisa-tion (i.e. initial mortality losses generated for the entire market, in other words more people with less favourable condi-tions from the insurer’s point of view conclude annuity contracts at a certain point in time). Such a transitory deficit may theoretically be covered by the state by repaying it from transitory surpluses. This may force the state to pre-pare mortality projections in a suitable manner. However, a problem occurs if a deficit or surplus is systematically generated in the central organisation and a clear problem in connection with the surplus is that insured people may be deprived of it in a questionable manner.

In its minimal form, providers carry the burden of mortality losses that emerge later, or they transfer this (or a part of it) to clients in the form of in-dexation. It is only logical to distinguish between initial and later mortality losses if the latter can be transferred to the client via indexation (which as-sumes a return-refund indexation technique). Otherwise one might note that if a pool is operated, it may be done in a maximal rather than minimal form.

In practice, the maximal form includes the above-mentioned minimal form as the mortality loss of the first year may be largely due to risk composition at the time of concluding the contract. In order to be able to continually redistrib-ute mortality losses and gains amongst providers, reliable and uniform evalua-tions are required. As with the premium blending mechanism, this requires centrally-designed standardised reserving rules and differentiated mortality tables along with a centrally and uniformly defined technical interest rate.

Based on these reserving rules, different providers will define the mortality profit and loss in each financial year and will summarise them in total at the market level. If the balance is positive or at least breaks even, then they will define the magnitude of the uniform mortality profit using a kind of baseline, e.g. according to required reserving, and where there is a surplus this will be transferred to where there is a deficit. If the balance is negative, clearly the surplus is not enough to cover it. Then the deficit would be distributed against the return (or if the given year was also bad with regard to return, then against the reserve, i.e. the benefit size of the annuity) or against the solvency capital function against whoever will cover the final loss (i.e. the insured person or the provider). Clearly, this will relate to the ownership and capital conditions of the provider. If the insured persons themselves are the owners and the provider and thus have no own capital, the loss shall be accounted for by the yield and if

there is an actual owner, meaning there is solvency capital, then the loss may be accounted for via this, while amounts may be regrouped from where the loss is below the market level (including the profit) to where losses exceed it.

It is worth mentioning that if a central organisation is set up on the market for management of the pool, this would most likely imply centralisation of other elements of annuity provision if this also allows for cost saving. Eventu-ally, only the investment will remain with the separate providers, and they will compete over the yield if the indexation rules permit this. This is exactly the same system as used by the Swedish private pension system.