Insurance modelling

Insurance modelling – The estimates of future cash flows should incorporate all reasonable and supportable information available without undue cost or effort about amount, timing and uncertainty of those future cash flows. To accomplish this, an entity should estimate the expected value of the full range of possible outcomes. Estimates and assumptions should be unbiased (that is, neither conservative nor optimistic). Insurance modelling

The objective of considering the full range of all possible outcomes is to incorporate all reasonable and supportable information. An insurer is not required to identify every possible scenario. Explicit scenarios are not required if the result meets the objective. However, a single scenario based on the most likely outcome or the more-likely-than-not outcome would not meet the objective where there is a non-linear relationship between the different scenarios and the associated changes in measurement. Judgement is required to determine the appropriate number of scenarios that will capture material non-linearity. This will depend on facts and circumstances and should be periodically reassessed.

Incorporating different possible outcomes

The requirement that estimates incorporate all reasonable and supportable information without undue cost or effort about the amount, timing and uncertainty of future cash flows is achieved by estimating the expected value of the full range of possible outcomes – i.e. the probability-weighted mean. The risk adjustment for non-financial risk is included explicitly as a separate component of the measurement. [IFRS 17 33, IFRS 17 BC150] Insurance modelling

The expected present value of future cash flows is determined by: [IFRS 17 B38] InsuranceInsurance modelling modelling

  • developing a range of scenarios that reflects the full range of possible outcomes, in which each scenario specifies: Insurance modelling
    • the amount and timing of the cash flows for a particular outcome; and  Insurance modelling
    • the estimated probability of the outcome; and Insurance modelling
  • applying to each scenario: Insurance modelling
    • a discount factor to determine the present value; and Insurance modelling
    • a weighting based on the estimated probability of the outcome. Insurance modelling

The objective is not to develop a most likely outcome or a more-likely-than-not outcome for future cash flows. Insurance modelling

The scenarios developed exclude possible claims under possible future contracts and include unbiased estimates of the probability of catastrophic losses under existing contracts. [IFRS 17 B40]

When considering the full range of possible outcomes, the objective is to incorporate all reasonable and supportable information without undue cost or effort in an unbiased way, rather than to identify every possible scenario. It is not necessary in practice to generate explicit scenarios when determining the mean, if the resulting estimate is consistent with this objective. [IFRS 17 B39]

Therefore, it could be appropriate to use a small number of parameters, or relatively simple modelling, when the measurement result is within an acceptable range of precision. However, more sophisticated, stochastic modelling is likely to be needed when the cash flows and their probabilities are driven by complex underlying factors – e.g. for cash flows generated by options inter-related with the insurance coverage. Insurance modelling

Information that is available from an entity’s own information system is considered to be available without undue cost or effort. [IFRS 17 B37]

Something else -   Direct participating contracts
Implications of using the expected present value model

Insurers build up cash flow projections for different products in different ways, which could be driven by several factors including: [IFRS 17 B62, IFRS 17 BC152]

  • the complexity and diversity of the underlying factors;
  • the diversity of valuation systems and models used; and
  • whether the products were acquired in a business combination or a portfolio transfer.

An entity may need to review whether it has projections of cash flows that meet the objectives set out in estimating future cash flows. Significant resources might be required to develop and implement new methodologies to develop cash flow projections or to modify existing projections to meet the objectives.

Model updates that may be required
If, for example, an entity currently uses a valuation model that attributes no value to:Insurance modelling

  • embedded options; or
  • guarantees that have no ‘intrinsic value’ because they are currently out of the money (from the perspective of the policyholder),

then the entity would need to adapt its model to address both the intrinsic value and the time value of these options or guarantees. This is because the expected present value model considers all possible scenarios, which includes the possibility that the option will have intrinsic value in the future.

Another example is a model that assumes a 100 percent probability that a policyholder will exercise a surrender option when the surrender value is higher than the present value of expected benefits. This model would need to be updated to reflect the possibility that the policyholder will not exercise the option.

Property and casualty contracts
Estimates of future payments on property and casualty contracts are currently based mainly on the projection of historical claims data. Although the goal of these estimates is to determine the loss provision and potentially a range of outcomes, they may not give the same results as calculating a mean using estimates of probabilities.

The use of these approaches might still be appropriate under IFRS 17 as long as the resulting estimate is consistent with the measurement objective. If such a method is used, then an entity will have to show that the measurement results in an answer that is within an acceptable range of precision. However, these approaches would be unlikely to meet the measurement objective if they include conservatism aimed at a most likely or a more-likely-than-not outcome, or ignore some uncertain future events covered by the contracts – e.g. significant natural catastrophes.

Example – Stochastic and deterministic modelling

The table below describes an insurance contract under a range of scenarios that reflect all possible outcomes. The table summarises information about net cash inflows and the probability of each scenario:

Scenario

Net cash inflows/(outflows), CU

Probability

Probability-weighted outcome, CU

1

-10,000,000

5.00%

-500,000

2

15.00%

3

5,000,000

7.00%

350,000

4

15,000,000

73.00%

10,950,000

Total

100.00%

10,800,000

Under IFRS 4, entities use either stochastic or deterministic modelling for measurement of insurance liabilities. Stochastic modelling requires considering various scenarios in determining the value of the insurance liabilities. Deterministic modelling usually identifies the most likely outcome or more-likely-than-not outcome and is not based on a range of all possible outcomes. For this example, the value of the insurance liability determined using stochastic modelling is CU10,800,000 (that is, probability-weighted outcome), while using deterministic modelling the value is CU15,000,000 (that is, most likely outcome).

Unlike many current accounting models that develop a single ‘best estimate’, under IFRS 17 all scenarios and their associated probabilities (even remote ones) should be considered and weighted. However, not all cases will require the development of explicit scenarios. In cases where there are complex underlying factors that behave in a non-linear fashion, sophisticated stochastic modelling might be needed. This could happen, for example, if the cash flows reflect a series of interrelated options. The objective is to incorporate all of the relevant information and not ignore any information that is difficult to obtain.

Stochastic modelling can be complicated, both to initially implement and to maintain. This may be an additional IFRS 17 implementation complexity for entities that do not use stochastic modelling currently under IFRS 4.

Reasonable and supportable information is defined as information reasonably available at the reporting date without undue cost or effort. Uncertainty and judgement associated with available information does not necessarily mean that information is not reasonable and supportable. Information available without undue cost and effort will include an entity’s own internal information, such as historical claims, benefits and lapse data and any forecasts of potential future claims, benefits and lapses, as well as externally available information such as economist forecasts and statistics (for example, mortality information) for a country where the entity operates. Insurance modelling

The following are examples of possible sources of information about probabilities, amounts and timing of future payments:Insurance modelling

  • actual information available about policyholders, such as claims already reported; Insurance modelling
  • an entity’s own historical experience, such as claims previously reported for similar contracts; Insurance modelling
  • country or industry information about historical experience, such as country mortality rates; Insurance modelling
  • information about emerging trends or changes in economic, demographic and other conditions, such as development of a treatment for diseases that impact mortality rates; and Insurance modelling Insurance modelling
  • changes in an entity’s own procedures that might affect the way in which information is gathered and presented, such as gathering sufficient statistically credible data for new products that enable an entity to measure liabilities using its own statistics while previously that was not possible. Insurance modelling

Market and Non-market Variables

The estimates of future cash flows should reflect the perspective of the entity, provided that the estimates of any relevant market variables are consistent with observable market prices for those variables.

Something else -   Variable consideration in transaction price and 4 best examples

An entity is required to maximise the use of relevant observable inputs and minimise the use of unobservable inputs, except for the following circumstances:

  • alternative pricing methods are acceptable where an entity holds a large number of similar assets or liabilities, but the market price for each asset or liability is not readily accessible;
  • market price does not represent fair value at the measurement date; or
  • available market price should be adjusted where a significant adjustment is needed to reflect the characteristics of the asset or liability.

Unobservable inputs should be as close as possible to the observable inputs if observable inputs cannot be used without adjustment.

Market variables can be observed, or derived directly, from the market. For insurers, market estimates and assumptions can include interest rates, quoted prices of debt and equity securities for participating contracts, inflation rates and prices of embedded derivatives that are not separated, such as options and guarantees.

For some contracts, some cash flows from the liability will exactly match cash flows of a theoretical portfolio of assets in all scenarios (replicating portfolio). In this case, the value of the replicating portfolio of assets and cash flows arising from the liability would be identical. An entity can use the market value of the replicating assets portfolio as an observable input to measure cash flows from the liabilities. This is referred to as ‘replicating portfolio technique’. If an entity chooses not to use a replicating portfolio technique, it must satisfy itself that its approach will not lead to a materially different measurement. Insurance modelling

IFRS 17 does not require the use of any specific modelling techniques. Insurers should exercise judgement to identify the technique that best meets the objective of maximising the use of observable market inputs. In particular, the technique used should result in the measurement of any options and guarantees included in the insurance contracts being consistent with observable market prices for such options and guarantees. Insurance modelling Insurance modelling

Entities are not required to measure cash flows from insurance contracts separately (for example, separate measurement of cash flows related to participation features, options and guarantees, claims and expenses). Entities should use discount rates appropriate for a contract as a whole if cash flows are not measured separately based on their characteristics. This could be achieved by using stochastic modelling or risk-neutral measurement techniques. Insurance modelling Insurance modelling

Market consistent measurement of options and guarantees

IFRS 17 will require stochastic modelling of financial options and guarantees (such as a guaranteed maturity value), which might not be a common practice in certain territories, as discussed in ‘Example – Stochastic and deterministic modelling’ above. Options and guarantees should be recognised and measured on a current, market consistent basis. All cash flows, including fixed, guaranteed and cash flows variable with underlying items, should be measured on a probability-weighted basis using market variables, where relevant, and considering all possible scenarios.

The measurement of options and guarantees will, in many cases, involve stochastic modelling or using a deterministic model, run multiple times, to reflect a range of scenarios because of the non-symmetric distribution of outcomes for those features. A single deterministic approach might, for example, omit valuing the scenarios where the expected investment return is less than a guaranteed return. For certain simple options and guarantees, a formula (such as ‘Black Scholes’) might exist which could be equivalent to stochastic modelling.

The most common methods for measuring financial options and guarantees on a market consistent, stochastic basis are the ‘risk neutral’ and ‘real world/deflator’ methods. In these methods, the financial options and guarantees are measured consistently with the cost of hedging the obligation (where observable) at the balance sheet date. This is achieved through the modelling of the interactions between cash flows that vary with underlying items and the discount rate for the contract as a whole.

There are alternative ‘real world’ stochastic methods, used today in certain territories, where some asset classes (such as equity instruments and real estate) are assumed, based on historical market averages, to outperform fixed income asset classes. These ‘real world’ methods are not permitted under IFRS 17, because financial options and guarantees would not then be measured consistently with observable current market prices.

Something else -   Probability of default

Non-market variables include all variables that cannot be observed, or derived directly, from the market. For insurers, non-market estimates and assumptions can include information about amounts, timing and uncertainty of incurred and future claims, lapse rates, mortality and morbidity rates, and expectations about how the insurer will exercise discretion in the future.

Entities can use both internal and external sources of non-market variables. Judgement is required to identify the most relevant information where both internal and external information is available. For example, mortality information is usually available both internally (from an entity’s accumulated data about mortality experience) and externally (such as mortality statistics of the country where the entity operates). Insurance modelling Insurance modelling

Mortality statistics of a country might be irrelevant if an entity issues policies only in one region of the country. On the other hand, if such a company decideInsurance modellings to expand its business from a single region to the whole country, its internally accumulated mortality experience might be irrelevant for the new portfolio, and country statistics or other external sources of information might be more relevant. Insurance modelling Insurance modelling Insurance modelling

In some cases, non-market variables might correlate with market variables. For example, for a participating contract with an embedded guarantee of minimum returns, the lapse rate might correlate with market interest rates. That is, the probability of lapse decreases with a decrease in market interest rates. In such cases, entities should ensure in relevant scenarios that probabilities associated with non-market variables are consistent with observable market information. Insurance modelling Insurance modelling

Market variables are often associated with financial risk, and non-market variables with non-financial risks, but this will not always be the case. For example, debt and equity instrument prices and interest rates always represent financial risk but they are not always observable in the market. Non-market variables should be as consistent as possible with available market information.

Insurance modelling

Annualreporting provides financial reporting narratives using IFRS keywords and terminology for free to students and others interested in financial reporting. The information provided on this website is for general information and educational purposes only and should not be used as a substitute for professional advice. Use at your own risk. Annualreporting is an independent website and it is not affiliated with, endorsed by, or in any other way associated with the IFRS Foundation. For official information concerning IFRS Standards, visit IFRS.org or the local representative in your jurisdiction.

Something else -   Insurance contract

Leave a comment