The theory is that the insurer isn’t then exposed to changes in the term structure on interest rates, only exposed to illiqudity/reinvestment risk to the extent of mortality fluctuations, isn’t exposed to currency risk and certainly isn’t exposed to credit risk. Without complex margining requirements like some swaps and without the need to roll cash investments over, government bonds should allow ALM teams to sleep well.

Now, Solvency II is likely to adopt a swap yield curve rather than bond yield curve. There are some good reasons here, including arguably fewer distortions from temporary supply and demand imbalances, improved liquidity and so on. The same yield curve is used for liquid liabilities so the allowance for an illiquidity premium over and above the swap curve at some times, in some ways and for some products is still under debate.

But what should Greek insurers do in the meantime?

Frankly, Greek government bonds don’t remove credit risk and the huge credit spreads on these instruments will create huge funding gaps and variability in earnings unless a Greek govi yield curve is used to value liabilities as well. It’s not clear at all that Greece will stay part of the Euro, so German government bonds don’t remove currency risk. German government bonds in any case are show signs of nervousness as yields creep up.

The swap market is exposed to the same Euro break-up risks as bonds. Which banks will survive, what happens to currencies in the meantime and what does that do to long-term Euro swaps? What about Euro-Sterling swaps issued by Greek banks (I’m not sure if these even exist though).

All in all, it’s good to be involved in ALM in South Africa, and even the Middle East just at the moment.

I wouldn’t want anything to do with any bank that has much at all to do with European banks or European sovereign debt. The old South African Rand is seeming like a safer relative bet than at pretty much any other time in the last decade.

As an aside, the use of IFRS balance sheet figures for regulatory reporting is also an unusual idea. There is no particular reason to believe that a shareholder financial reporting basis is appropriate as a regulatory measure. It can be, with specific capital rules perhaps, but it’s not automatically so.

Why the 10% of OCR rule for IBNR liabilities is so bad:

1. For very long-tailed business with no or low claims reported in the first year, the IBNR will be massively understated
2. as claims are reported (and before they are paid), the OCR will increase. The IBNR should decrease as the claims have now been reported, but given the 10% rule it will actually increase.
3. The reconciliation of opening to closing IBNR and the comparison of actual vs expected IBNR claims over time is not useful since there are no explicit expectations built into the methodology
4. Clearly the method is not sensitive to risks and delays of product lines or processes.

So what’s better? Well aside from the range of standard but fairly complex techniques (including Ultimate Loss methods, Basic Chain Ladder, Bornhuetter-Fergusson, Average Cost Per Claim and a whole range of stochastic methods) there are better simpler measures.

A starting point, although also very far from ideal, is the current (soon to be changed) South African statutory requirement of 7% of net written premiums. It also isn’t sensitive to different delay patterns and will give poor results if net written premium is growing or shrinking rapidly.

Really, the ideal simplification requires a little more complexity, but as a reward for this effort is a far more robust, more accurate measure that behaves sensibly in a far wider set of scenarios.

For each line of business for each delay year, we use a specified percentage of gross earned premium for the gross IBNR. Reinsurers’ share can be calculated similarly. The information relating to earned premium per line of business going back several years should be trivial to obtain and ensures we get a sensible pattern taking into account the growth in the business, the mix of business as well as change in mix of business. The method works well for start-up, mature or declining books.

The fundamental drawback of not reflecting a particular insurer’s patterns remains, but aside from using actual delay data this is about as good as one can hope for.

Frankly, why more regulators don’t prescribe this method is a mystery. The information is available, it’s trivial to calculate and verify and the results are robust.

This is not a new idea, and has historically been resisted by the major banks since they feel, generally rightly, that they are less likely to have a problem of confidence and therefore less likely to benefit from deposit insurance. Smaller banks, on the other hand, are certainly more at risk of a run on the bank arising due to perceptions thus causing a liquidity problem when a solvency problem doesn’t exist.

Determining the appropriate mechanism to charge for deposit insurance is not straightforward. Clearly the charge can’t be the same fixed amount for all banks as the exposure will be very different between banks. The large banks would lobby hard to pay a lower rate (even if a higher overall amount) for the insurance given that they should be less subject to those confidence issues.

But how to determine that difference? One could take cue from the market by looking at credit ratings and, even more market-oriented, the spreads on debt issued by the banks. Three immediate problems come to mind:

1. Credit quality of long-term debt isn’t the same as protection for depositors
2. Probability of default is an input into confidence issues but certainly isn’t the entire story
3. Does anybody seriously still believe in the Efficient Market Hypothesis and trust that we can believe what the market offers as an impartial, objective and balanced view of reality?

So there are some fascinating technical problems to solve when implementing deposit insurance, not least of which is deciding how much gets protected.

What caught my eye was Moneyweb columnist, Phakamisa Ndzamela, demonstrating his disbelief at how deposit insurance could create a moral hazard and ultimately increase risk within the banking system.

some senior bank executives have cautioned that this could push the cost of banking higher and somehow encourage risky lending. [emphasis added]

Obviously Ndzamela hasn’t heard of the Savings and Loan crisis in the US in the 1980s, or, I don’t know, the Global Financial Crisis that we are still in, which was massively exacerbated (if perhaps not quite caused) through risk being accepted without due care because it was being passed off immediately to someone else.

Nevertheless, S&P did downgrade. Why? I don’t think it is primarily to do with a materially increased estimated probability of default. It has more to do with a change in the payoffs in a ‘game’ (as in game theory) S&P is playíng.

Consider the quadrant of options. S&P downgrades or doesn’t and the US defaults or doesn’t. I’ve constructed totally hypothetically, but perhaps plausible scenarios below, for the S&P’s potential assessment of losses under each possibility given their views and external perceptions of them before and after 2008.

Before 2008, the fallout that would come from downgrading the US and the US not defaulting would be significant and cries of “un-American” might be heard again. Even if the US were downgraded, default would still be a blow for S&P since anything above a BBB rating really shouldn’t ever default if there models are “correct”. I’ve thrown in another hypothetical, a 0.01% probability of default – in other words very low, and as you’ll see in the next scenario, not necessarily higher now for S&P to change their view.

Now, either on a traditional minimax (minimizing the maximum cost) or an expected value basis, before 20008 S&P wouldn’t downgrade the US. This is an important calibration, since S&P didn’t downgrade the US.

After 2008, even if we leave the assessed probability of default unchanged, the world is different and therefore we have different costs.  If S&P doesn’t downgrade the US – even if the US doesn’t default, there will be a cost to S&P since might share the view that the US could default now. The dent in credibility since 2008 means that S&P has to try harder to convince the skeptics that they don’t rate risky instruments as AAA. Along with this goes a massive hit if the US does default and S&P hasn’t downgraded the US. The good news is that at least now a downgrade is viewed more with more understanding even if the US doesn’t default (although be sure Obama’s White House is not happy at the moment).

After 2008, even if the assessed probability of default is unchanged, the minimax and expected value rules both suggested a downgrade is the better option for S&P.

It’s a good piece of legislation, even if at times some aspects of it may result in greater costs than benefits.

TimesLive has a story about the alleged noncompliance of medical schemes with the CPA.

Some of the issues may have merit, but this struck me as particularly troubling:

According to the act, it is unfair when a consumer is discriminated against on the grounds of age.

Our constitution explicitly allows discrimination on actuarially sound rating factors that have both a statistical and causal link. This is how insurance is South Africa still uses underwriting to select homogenous groups of risks and to limit anti-selection by policyholders. If widespread anti-selection were to occur, then life insurance would not be viable.

Medical Schemes in South Africa have only very limited underwriting options in order to provide as many citizens as possible with fair health coverage. “Late joiners” are charged a premium since they haven’t contributed to the societal risk pool since they were most healthy and therefore haven’t paid “their fair share”. This has to do with a specifically identified risk rather than general discrimination based on age. These restrictions are important to maintain the solvency and viability of medical schemes.

Some schemes prevent women who fall pregnant within nine months of joining the scheme from claiming for the pregnancy even though they pay full premiums

This point is more tricky, but it does again reflect a misunderstanding. “Full premiums” on an actuarial sound basis have probably not been paid, since the fair premium for a member who joins just to get pregnancy benefits and hasn’t contributed at other times would be much higher than the premium that is charged. This one is a little more grey and while I feel the rules are entirely fair, they may not be viewed that way by a particular judge on a particular day.

Some schemes require that members give three months’ notice when terminating their membership, whereas the act deems 20 business days to be reasonable

This might reflect the desire to not have members leave a scheme immediately after having utilized the maximum benefit available to them before joining another scheme. I don’t know how much of this behavior would ever happen, so this might also ultimately be changed.

Many schemes don’t enforce the allowed waiting periods for members joining. If some of these other changes were to be made, I would expect these provisions would be more regularly used. Of course, that is another of the problems cited with medical schemes arising from the CPA.

All in all, we may see some changes, but by and large these comments reflect a lack of appreciation for the actuarial realities of managing a health scheme with community rating.

New Business Margin, which is the Value of New Business (VNB) as a percentage of the Present Value of New Business Premiums (PVNBP) is a common measure of profitability of that news business.

But it’s a flawed measure, especially when it comes to comparing product lines and insurers or even to understand the change in profitability from one period to the next. It uses and unequal yardstick to measure business.

New Business Margin on Revenue (NBMR) provides a significantly improved measure of profitability that can be used to compare margins across products, across insurers and across time. Further, it leads easily to a component analysis of the margin, adding additional insights to shareholders, brokers and regulators.

If you haven’t read my introductory post on New Business Margin on Revenue, it would be worthwhile doing so now – this post is going to illustrate the sort of results it provides in a practical, numerical example.

Example 1 considers how NBMR clarifies distortions from a change in mix of business.

Example 2 shows how more complex dynamics can be understood through a component analysis of NBMR. The spreadsheet showing the underlying calcs is attached at the end of this post.

Example 1 – A change in mix of business

AGGREGATE PICTURE

Let’s look at the sort of aggregate information you’ll typically see in an EV report.

What one might take from this analysis is that VNB is constant, but margins are declining to below 2.0%. If this is the only information on which to base our analysis, this company might be a clear “sell” and products and pricing need to be updated by management.

Let’s see how this would look using NBMR

From this table we see again that VNB has been constant, but that the profitability of the business has actually been slightly increasing.  So, although we should consider attending to the zero-growth VNB, the actual margin we’re achieving on our business on this measure has improved slightly from 25.5% to 26.1%.

This is a more accurate picture, because as you can see from the following tables, all that has changed is our mix of business – and the investment business that we’re writing more of now actually has a higher NBMR than the risk business we’re selling less of.  The traditional New Business Margin measure is distorted because it treats the entire premium paid by the policyholder as “revenue” when in fact only a small share of it is fees and charges and the rest is more like a deposit.  Again, banks measure profitability and performance through RoE, Cost to Income Ratios and Net Interest Rate Margin and only very much behind those the return on total assets.

Analysis of NBMR and components via product line

So the New Business Margin is constant on a product view, but when compared it appears as if there is a declining trend.  Also, at 26.7% of revenue taken as profit for investment business, this is profitable business, a fact not obvious from the superficially low New Business Margin of 1.3% (which is actually a perfectly good profit margin on that measure, just difficult to understand and compare).

The Revenue Per Premium (RPP) is 100% for risk business and 5% for investment business.  Depending on policy size, a 5% deduction from every premium might be a little on the high side and from a Treating Customers Fairly and sustainability perspective, this area may require some attention.  It’s also fair to ask “why are we losing market share in risk products and what can be done about it”, but at least we see this as a mix of business and market share issue and not a business margin issue.

Example 2 – Complex changes in volume, profitability and other components

Aggregate picture

Example 2 is more complex.  Let’s look at the aggregate information on a traditional presentation first.

In this case it looks like we have a severe margin squeeze problem in spite of increased business volumes and increased VNB. By now it should be clear that it’s dangerous drawing conclusions from this information.

A more coherent aggregate view can be obtained using NBMR:

Here we see a strong 2011 increase in all of VNB, Business Volumes and New Business Margin on Revenue after a poor year in 2010.  On the whole, we are hanging to customers longer than before (Discounted Premium Term or DPT up form 5.4 in 2010 to 5.6 in 2011, but still not at the levels of 2009. This is worth investigating.

The share of each premium we get as revenue has dropped sharply – clearly suggesting a change in mix of business as this sort of change wouldn’t typically be seen otherwise. Clearly we need to dig further, but the previously bleak picture is already looking better – and as we’ll see this is a more accurate reflection of business reality.

ANALYSIS OF NBMR AND COMPONENTS VIA PRODUCT LINE

Immediately we see a huge amount of new information.  Risk business has been declining in profitability significantly and has also had a dramatic increase in lapse rates (since the Discounted Premium Term has dropped to 3.2, suggesting major problems with persistency).

At the same time, although 2010 was a step backwards in terms of DPT for Investment business, the increase in volumes of business (API), allied with a restoration of the DPT to close to 2009 levels, a reduction in RPP (suggesting better value for policyholders, which should give rise to better future sales, lower persistency and less regulatory intervention) and a strong growth in NBMR, driven off efficiencies, expense reductions and economies of scale through greater sales.

Our risk business is in trouble and requires attention, but we are building a solid, profitable and sustainable investment business that should provide good returns to shareholders.

Conclusion

These are stylised examples filled with hidden good news. The reality is that many insurers are struggling in several business units.  The analysis and tools outlined here can help make better informed decisions around product strategy and pricing, and for analysts wanting to better understand the current and potential future financial performance of these stocks.

For short-term business (group risk, health insurance, health administration, general insurance etc.) it is something different since these lines don’t have long-term prudent provisions. In these cases it reflects the present value of future profits expected to be earned out of the existing business.

The inclusion of these short term types of business within EV is widespread. It seemingly increases consistency between different types of contracts since we are considering the long-term expected profitability in all cases. More on this apparent consistency in a moment.

What do we include in the EV and VIF?

EV is not a complete economic measure of the value of an insurer, since it ignores future profits arising from future new business. This is by design. An Appraisal Value incorporates the Value of Future New Business (VFNB) as well, although there is always subjectivity over how many years of future new business should be included. (More on this in a separate post.)

Existing Business vs Future Business

The idea of “existing business” and “future new business” is clear in the individual life context. It’s existing business if you have a contract, and future new business if not. Premium increases and slight benefit modifications can usually be accommodated within the existing contract and so should ideally be included in the Value of In Force (VIF) based on the expected probabilities of these changes.

As such, the expected future premium increases should also be factored into the Value of New Business (VNB) when the business is originally sold. (Don’t confuse VNB with VFNB.  VNB is the value of business written over some past historical period, typically the last month, 6 months or year. VFNB is the expected value to owners of the business from business that will be sold at some point in the future.)

What counts as In Force for short term contracts?

The position is less clear for Group Risk or Motor policies. These are typically annual contracts, but with high expectations of renewal. In this case the renewals are typically included in the VIF at some assumed level of premium inflation reflecting wage growth rather than new business. Premium inflation is an important consideration in countries with non-negligible inflation. New employees on an existing Group Risk policy are also usually not considered new business since no new sales activity was performed or purchasing decision was made.

The rules applied then are quite well understood. However, we are valuing two different things between short term and long term contracts, and also a little bit of something we shouldn’t.

The source of value for long-term and short-term contracts in the EV

For long term contracts, we are valuing expected future surpluses arising out of a long-term contractual right and prudent regulatory provisions.

For short term contracts we are valuing the existing contract (a small portion of the total value) and the customer relationship that will give rise to future renewals.

There is also a component of customer relationship in the long-term contractual rights since poicyholders can lapse the contract, incur some penalties possibly, but ultimately not be required by law to continue paying premiums. It’s difficult to separate these components though.

Unintended inclusion of brand value

Now the interesting part: I maintain that by using a best estimate future renewal rate for short term contracts, we are actually valuing a portion of the brand over and above the customer relationship. This is best demonstrated by an example.

Example of confounding of customer relationships and brand value

Take a company with a 20pc market share in the group risk market in which it operates. Let’s also say that best estimate future renewal rate is 80pc. If we assume a constant market share percentage (arising because of the value of the brand, but may be also broker networks etc.) then 1 in 5 employer groups looking for group risk cover will choose our hypothetical company – ignoring customer relationships.

Some of our customers will be very pleased with our service and will renew for that reason. The good customer relationship is a significant source of value. However, some of our customer won’t particularly value our service or relationship. Some of these will leave, but others will stay anyway – because of the value of our brand.

In fact, we should expect 20pc of our existing customers to renew even if there is no customer relationship at all. Why should existing customers be less likely to choose us again even if there is no customer relationship than a brand new customer?

Some of our customers may be so put out by an unfortunate incident, poor service or a repudiated claim that they may deliberately not renew. A customer relationship doesn’t have to have positive value.

So of our 80pc renewals, 20pc renew for reasons independent of our customer relationship. So the correct renewal rate to apply when measuring customer relationships is actually 60pc.

The impact can be significant

This makes a significant difference to the VIF. For a 20 year projection and 20pc market share, the difference is a reduction in VIF of 46pc. For a 10 year projection with a 5pc market share, the difference is still 14pc.

You can test the impact using this Renewal Test model under Example Models.

What does this mean practically?

Now we haven’t changed the true value of the business; we’ve changed how much of it we recognise as a customer relationship. With a lower VIF, we would expect to see positive renewal experience variances (if we compare actual renewal rates to modelled) or higher volumes of new business (if we treat the portion of renewals related to overall market share as new business) athough the new business margin on revenue will be lower since we are modelling a lower Discounted Premium Term (DPT).

Neither of these results is very satisfactory. Perhaps the answer lies in separating renewal into a change in market share impact and the remaining renewal variance so that the VNB is still objectively measured as new sales, but the persistency profits are separately allocated between the known adjustment for brand and the actual deviations from renewals being further different than expected.

Merger and Acquisition pricing and accounting

Specifically though, this has important implications for accounting for insurance mergers and acquisitions and the types of intangibles created. It also feeds into the purchase price decision and valuation adopted for an acquisition. Ultimately, views of the Appraisal Value of annually renewable business must incorporate these important differences when compared to individual life businesses.

Bahrain has a population of around 1.2 million but has a well developed insurance sector. Economies of scale limit domestic growth and regional expansion faces tough competition and increased complexity. Will be interesting to see more.

Richard Thaler presents some of the common perception problems with annuities in this article in the NY Times. The basic message is still as it has been for decades. Individuals are reluctant to pay a large portion (often the majority) of their life savings to an insurer with the risk that they will die in a few years and “not have got their money back”. The peace of mind that should come to the policyholder turns into a matter of stress.

The bequeath motive is strong – and amplified by a lack of understanding of exactly how long we’re likely to live in retirement these days and how much money will be required. Those to whom many plan to bequeath may ultimately become the source of support when the income draw-down products are depleted with no longevity guarantee to boost the funds available.

It’s a good explanation although he doesn’t break much new ground. He also doesn’t talk about the concerns some potential policyholders have, in some countries at least, of whether the insurance company who sells the annuity will definitely be around over the next 40 years come what may.  This is more common in developing markets with weaker regulation (probably a good reason to have concerns) and less history of annuities (a cultural bias that will probably disappear over time).

Mr Thaler doesn’t propose any solutions for the insurers in boosting sales – a common “fix” is to combine a traditional pay-until-death annuity with a guaranteed minimum period or a death benefit (either for a limited term or at any point).  These adjustments reduce the “risk” of “making the wrong decision but purchasing an annuity but only living for a short period”.

There’s no free lunch.  In the same way that cash-back bonuses on short-term insurance products actually increase the average cost of insurance and reduce the risk-transfer from insured to insurer, these guarantee periods increase the cost of annuities.