Chapter 12: Actuarial investigations Flashcards
1
Q
Steps involved in a rating analysis
A
- estimating ultimate claims
- estimating profitability of existing rates
- projection forward to a new rating period
- reviewing the suitability of the existing rating structure
- comparing rates with those of competitors
- analysing the profitability of old years on new rates
2
Q
Rating analysis:
Estimating ultimate claims
A
The new premium rate for a policy will be based on the expected cost of claims for that policy.
We estimate this cost using recent claims experience from existing portfolio.
We then project claims to their ultimate levels to get realistic view of recent past claims experience
3
Q
Rating analysis:
Estimating the profitability of existing rates
A
Estimate the profitability of the existing premium rates by reference to the recent claims experience, adjusted if necessary for any abnormal features.
We will look at the claims corresponding to the premium and compare actual and expected experience. Analyse the reasons for differences between the two.
4
Q
Rating analysis:
Projection forward to the new rating period
A
Project forward to the period over which the new rates will be charged and the corresponding claims will be settled.
In doing this, we make assumptions about future claims trends and inflation. Introducing a further element of uncertainty into investigation.
5
Q
Rating structure
A
Refers to the relative levels of premium charged to different policyholders, depending on their particular risk profile.
6
Q
Rating analysis:
Comparing rates with those of competitors
A
In making the final pricing decision, we will be strongly influenced by what the market is doing.
Our premium rates must make allowances for our competitors’ rates.
7
Q
What shoud be done if an insurer is entering a new market or introducing a new type of policy with no suitable data from its own experience?
A
It may be able to use some relevant external data or internal data from related accounts, but in general it must adhere closely to existing market practice
As alternative options, we could:
- ask reinsurers
- partner with another insurer on quota -share basis and use their data
- use market data such as Lloyds
8
Q
Possible sources of “relevant” external data
A
- reinsurers’ data
- industry data
- other insurers’ data
- relevant organisations or government bodies
9
Q
Steps involved in an expense analysis
A
- split between direct and indirect expenses
- split between types of expenses (e.g. initial, admin, renewal, claims, investment)
- allocate by class and rating group
- express the expenses as a proportion of number of policies or claims, or of amounts of premium, sum insured or claim.
10
Q
Direct expenses
A
Expenses that can be directly allocated to a class of business
11
Q
Indirect expenses
A
AKA overheads
Expenses that relate to support functions and therefore cannot be directly attricuted to any one class of business or policy.
12
Q
What are non-commission expenses split into?
A
- initial expenses
- admininstration expenses
- renewal expenses
- claims expenses
- investment expenses
13
Q
Initial expenses
A
Arise when business is being aquired and written into the books of the insurer
14
Q
Renewal expenses
A
Incurred in the recosting, renewing or lapsing of a policy at the end of the policy period; these are often markedly different to initial expenses
15
Q
Claims expenses
A
Incurred in the assessment and payment of policyholder compensation
16
Q
Investment expenses
A
Relate to the management of the company’s assets
17
Q
Initial, admin and renewal expenses can be split according to:
A
Whether the expense is proportional to:
- the number of contracts being aquired or on the books
- the amount of sum insured or EML (size of the risks)
- the amount of premiums being aquired or on the books
18
Q
Most expenses are proportional to the number of contracts in force. Exceptions include:
A
- underwriting expenses - mainly related to the amounts of new business premium
- claims expenses - typically related to the level of claim payouts, with higher expenses for larger claims
19
Q
Fixed expenses
A
The expenses that remain relatively fixed, regardless of how many policies we sell
20
Q
Variable expenses
A
Expenses that vary according to the amount of insurance business being handled. These expenses may be linked to the number of policies or the number of claims or amount of premiums or claims.
21
Q
Main items of expense of general insurers
A
- salaries and salary-related expenses
- property costs (rent, property taxes, heating, lighting and cleaning)
- computer costs
- investment costs (investment department, stamp duty, commission, etc.)
- one-off capital costs
- claim handling costs
22
Q
Process of allocating expenses
A
Expenses are split down and analysed into the required “cells”.
Typically the cells may be:
- the whole business of the insurer
- the whole business of a particular accounting field
- each main product line of the insurer
23
Q
Why would a general insurer want to allocate expenses to different product lines?
A
The insurer will want to allocate the expenses as accurately as possible to ensure that:
- the premium charged to each class (and therefore to each policy) is correct
- the profitability of each class is assessed correctly.
24
Q
Salaries and salary-related expenses can be split in:
A
- staff whose work falls entirely within a single cell of the analysis
- staff whose work falls within more than one cell
- other staff
25
If a company owns the property they occupy, why would we charge a notional rent rather than allowing for the actual purchase cost?
It wouldn't be practical to charge a premium that includes an allowance for the actual purchase cost of the property, because the premium in the year of purchase would be excessively high. A notional rent is charged instead. This provides an approximate annual cost of the property, which can then be allocated by product line
26
Allocating one-off capital costs (other than the purchase of a new computer)
We should amortise a large one-off capital cost over the expected useful lifetime of the item purchased, We may then treat the amortised cost as part of the overheads.
We would exclude exceptional items (which are not likely to recur) from the analysis
27
Types of reinsurance investigations
Include the analysis of:
- the required risk retention allowing for the solvency position
- the extent of exposure to accumulations of risk
- the need for catastrophe reinsurance
- the need for reinstatements
- the value for money of existing reinsurance
- the appropriateness of existing cover
- the profitability of layers
- the effects on capital
- the cost of commutation
- to check on the reinsurer's solvency levels and assess the need for bad debt provision
28
Commuting a reinsurace agreement
Commuting a reinsurance agreement means that the insurer accepts a fixed premium now and the reinsurer does not make any future payments to the insurer, i.e. the insurer swaps future recoveries for a fixed sum.
29
Why are commutations used?
An insurer may wish to commute a reinsurance arrangement if it helps with the insurer's cashflow or if it is concerned about the solvency of the reinsurer.
The reinsurer may be interested in commutation because it removes future uncertainty, i.e. the risk is passed back to the insurer.
Commuting might also be preferred as a tidying up exercise between companies, for admin savings and for strategic reasons.
30
The consequences for an insurer of taking out too little or too much reinsurance
Too little reinsurance, and experience is worse than expected:
- risk of sharply reduced profit
- lower published solvency
- insolvency
Too much reinsurance:
- the reinsurer will be profiting at the insurer's expense in the long run
31
Main reasons for monitoring business written
To:
- assess performance against goals
- manage risk
- gain market intelligence
- satisfy the regulators
- influence the market
- assist with reserving
- validate assumptions as part of the actuarial control cycle
32
Key factors monitored when monitoring business written
- premium rate changes
- portfolio movements
- volumes of quotations
- persistency and profitability by source
33
When monitoring portfolio movements, what are the movements monitored in respect of?
- lapses or renewals
- new business
- quotes that result in written business (strike/conversion rate)
- mid-term cancellations
- policy endorsements
- mix of business
34
Reasons for monitoring business written:
Assessing performance against the organisation's goals
Ultimate goal for most general insurers is to exceed a minimum level of profit or return on equity for a given level of risk. Companies usually break down this goal into more specific targets.
Breaking down the target into specific targets enables different parts of business to be monitored more effectively.
If one target isn't met, this can be investigated and rectified, whereas an overall profit analysis would hide any underperforming areas behind other areas where targets are being exceeded.
35
Reasons for monitoring business written:
Managing risk
Allows the company to assess how much risk is inherent in the portfolio (e.g. accumilations).
The amount of risk will be a factor in determining how much capital the company should hold and what its reinsurance purchasing strategy should be.
36
Reasons for monitoring business written:
Gaining market intelligence
Monitoring business can provide useful information about competitors' strategies. It can also allow the company to compare itself to the market and assess the underwriting cycle.
37
Reasons for monitoring business written:
Satisfying regulators
Market regulators may require periodic monitoring and reporting of written business.
38
Reasons for monitoring business written:
Influencing the market
A company may be able to influence the market by publishing the results of its monitoring exercise.
39
Give an example of how an insurer's results might have an impact on the market behaviour
A company might choose to publish its persistency rates. If the company has been retaining its business, this may attract prospective policyholder in the belief that they must be getting "a good deal" since the company is popular with existing policyholders.
40
Reasons for monitoring business written:
Reserving
The outputs of any monitoring exercise can be used as inputs into the reserving process. Most common example is the use of a priori loss ratios in Bornhuetter-Ferguson reserving methods.
Choice of a priori loss ratio is important, and it may be that a premium index can assist with this assumption.
41
Premium rate definition
There are many definitions of premium rate used in the general insurance market.
Examples are:
- premium income per unit of expected loss
- premium per unit of limit (e.g. per peril limits)
- premium per unit of exposure (sum insured, EML, etc)
- premium per unit of risk-adjusted exposure
42
Change in premium rate from t1 to t2
Rate change (t1->t2) =
(Premium Rate t2)/(Premium Rate t1) -1
43
Ways to calculate rate changes
- direct calculation of the premium rate
- price a standard risk
- measure rate changes on individual renewals
- underwriter's view of the rate change
44
Ways to calculate rate changes:
Direct calculation of the premium rate
Simply calculating the premium rate for every policy, then comparing the result from one time period to the next.
Advantage:
- There are many factors affecting the expected loss can be considered
- The absolute premium rate is calculated in addition to the rate change (not just the relative change)
45
Ways to calculate rate changes:
Price a standard risk
The premium rate is calculated for a specific "standard" risk or sample of risks which reflect the business mix of the portfolio as a whole.
Advantages:
- simpler, quicker and less data onerous than calculating expected loss for every risk written
Disadvantage:
- requires the additional assumption that the rate of change for the standard risk is equal to the rate of change for the entire portfolio, which may not be the case in reality
46
Ways to calculate rate changes:
Measure rate changes on individual renewals
For some segments it is very difficult to calculate the absolute level of the premium rate - problems in assessing expected loss for heterogenous book - lack of credible data
In this method we:
- consider only the renewing policies and estimate the relative change in expected losses for these policies (without determining the expected losses themselves)
- compare this to the change in premium to assess the change in premium rate
Disadvantage:
- impact of new and lost business written at different premium rates is ignored
47
Ways to calculate rate changes:
Underwriter's view of the rate change
Involves recording how the underwriters perceive premium rates to be changing.
Advantage:
- can allow for more soft factors that are otherwise unquantifiable
Disadvantage
- very subjective and therefore difficult to ensure consistency
- difficult to assess across companies and verify in detail analytically
48
In studying the movements of the business, and their trends, we can:
- measure the extent to which different parts of the portfolio are growing/contracting
- get an early indicationof undue losses/gains in the business that might indicate that rates are out of line with the market (possibly leading to anti-selection)
- assess the effects of a new set of rates or marketing campaign on the business and hence the sensitivity of the portfolio to market influences
49
Lapse rate
= (no. of lapses for period x)/(no. of renewals invited for period x)
50
New business rate
= (no. of new policies for period x)/(no. of renewals invited for period x)
51
Strike rate
AKA conversion rate
The number of written policies divided by the number of quoted policies in a given period (although it can be adjusted for declinatures)
52
Not-taken-up (NTU) rate
The complement of the strike rate
53
Declinature
Where the insurer refuses to provide cover
54
Cancellation rate
= (no. of cancellations during period x)/(no. of policies exposed for period x)
OR
= (no. of cancellations during period x)/(no. of renewals invited for period x)
- this can be seen as an approximation for the first equation
55
Endorsement
Some change to the policy wording, usually following a change in the risk covered, that takes effect during the original period of insurance. It is usually, but not necessarily, accompanied by an alteration in the original premium
56
Endorsement rate
= (no. of endorsements during period x)/(no. of policies exposed for period x)
57
Deferred Aquisition Costs (DAC)
Aquisition costs relating to contracts in force at the balance sheet date
They are carried forward as an asset from one accounting period to the subsequent accounting period in the expectation that they will be recoverable out of future margins within insurance contracts, after providing for future liabilities
58
Features of a good system for monitoring business
- tailored output
- accurate and validated data/results
- easy to use and well documented
- consistent over time and with other data sources and analyses
- minimal delay between data cut-off and production of results
59
Issues to consider when analysing lapse rates
- we may do the analysis at a total level and on specific subgroups
- normally monitor lapse rates on a monthly basis - allow for the time lag between the renewal date and point at which it becomes apparent that a policy has lapsed
- if there are processing delays, we may not know the final number of lapses stemming from a particular month of renewal until some months later - impose deadline for acceptance of renewals
- Can obtain lapse rates by projecting the number of lapses processed to date for each month's renewal cohort to their ultimate value
60
Problems when analysing the profitability of business from different sources
- data needs to be standardised to remove distortions caused by the mix of business to form sensible comparisons between two sales outlets
- premium rating structure will usually result in some policy types being more profitable than others
- different intermediaries could be paid different levels of commission - calculate the total net expense (including commission) to itself of business from a given source
- take long-term view - initial expenses are often much higher than renewal expenses
- high persistency = higher profitability
61
Key features of a good system for monitoring business:
Tailored output
Output should be consise (information not data) and tailored to the strategic goals of the organisation. Output should aid decision making
62
Key features of a good system for monitoring business:
Accuracy
- data used should be reliable and validated
- calculations should take into account all key drivers
- results should be validated
63
Key features of a good system for monitoring business:
Ease of use
- data should be easy to collect
- calculations should not be overly complex
- system should be documented, extendable and low maintenance
64
Key features of a good system for monitoring business:
Consistent
- output should be consistent over time
- inputs should be consistent with other data sources. If two quantities are the same then it should be the same regardless of data source. Otherwise the reader will lose confidence
- outputs should be consistent with other analyses (e.g. results should be shown split into the same business segments)
65
Key features of a good system for monitoring business:
Timely production of results
The delay between the data cut-off date and the production of results should be kept to a minimum
66
Common investigations around estimating reserves
- determine liabilities in published, solvency and internal management accounts
- estimate the cost of claims incurred in recent periods as an intermediate step in the rating process
- value an insurer for purchase or sale
- assess the accuracy of previous reserve estimates
- provide management information on performance and profitability
- comparison of best estimates against held reserves to understand the extent of the margins held in the reserves
- calculation of ranges of results to understand the potential for reserves to be insufficient
- transforming an underwriting year to an accounting year
- calculating movements in reserves from one period to the next and analysing reasons for these
67
Investigations relating to the investments of a general insurer
- evaluation of the existing portfolio
- asset-liability modelling to determine the optimal investment policy considering the risk/uncertainty inherent in the business
- allocation of investment income (and hence capital) between different classes of business
- determination of return on capital
68
Investigations relating to the investments of a general insurer:
Evaluation of the existing portfolio
- see what return has been made
- how it compares to what was expected
- what the market as a whole has done
- compare the value of assets to value of liabilities to assess solvency position & compare to regulatory capital requirement
69
Investigations relating to the investments of a general insurer:
To assess capital requirements, risk and investment policy
- asset-liability modelling used to determine optimal investment strategy
- ensure solvency requirement met in future
- consider how much to invest in liquid assets to ensure cashflow remains positive at all times
- analysis could be stochastic/deterministic
70
Investigations relating to the investments of a general insurer:
Allocation of investment income between different classes of business
- requires that capital is allocated to different classes because investment income is earned on capital
- possibly hypothecate assets to different classes - may come from analysis of capital requirements
71
Analysis of experience may cover
- pricing and sales of policies
- claims reserving estimation and changes to claims experience/environment
- exposure & aggregation of risk
- expense analysis and allocation
- policyholder behaviour, e.g. renewal and cancellation rates
- estimation of claim trends
- anything else the management require
72
Analysis of claims experience may be done at 3 levels:
- at an overall company level
- at a class of business level
- at a more granular level, e.g. by risk/rating factor
73
Key points to consider when deciding on the extent to which the analysis should be split by rating factor
- what is the purpose of the investigation?
- do we need to split into different risk groups? For some analyses, the different rating factors don't have much impact on the results.
- have we enough reliable data to support a detailed analysis by risk group?
74
Analysis of claims experience:
Changing frequency and severity of reported and settled claims
Useful to split the effects of:
- frequency
- severity
Identify reasons for better or worse than anticipated performance
Provides more info on the underlying experience of an account that seems to be stable
75
Analysis of claims experience:
Impact and incidence of large claims
Useful to determine the drivers of such claims, and sepcific areas of the account that are more prone to certain types of large claims.
Helps us decide which reinsurance coverages are appropriate and will be a key driver of the capital which the business needs.
When estimating reserves, it may be useful to analyse "attritional" and large claims experience seperately
76
Attritional claims
Normal, non-large claims
77
Analysis of claims experience:
Concentrations of claims and aggregations of risk
- identify concentrations of claims so that we can deal with them seperately from the main body of the data
- may be able to do this from the individual claim records by grouping by factors
- set up a catastrophe code for a catastrophe and place on claim records for all claims arising from the event
- also assits with reinsurance recovery
78
Analysis of claims experience:
Types of claims reported
Analyse the types of claims reported to identify:
- new types of claim emerging
- the effectiveness of exclusions or changes to policy excesses in removing certain types of claim from the experience
79
Salvage
Amounts recovered by insurers from the sale of insured items that had become the property of the insurer by virtue of the settling of a claim
80
Main types of recovery on an insurer's gross claims experience
From:
- outward reinsurance protections
- recoveries from salvage
- third party involved in the claim incident
81
When presenting results, we must communicate:
- the source of the data, its preparation and verification
- our reliance on assumptions and the uncertainty of the results
- methodologies and definitions used
- key features of the results and why they have happened
- how actual compares with expected
82
Main claims analyses
- changing frequency and severity of claims
- impact and incidence of large claims
- assessing concentration of claims and risk
- splitting indemnity costs from expenses
- new types of claims
- recoveries on gross claims
- nil claims
- partial payments
- re-opened claims
83
Uses of actuarial investigations
- carry out profit testing
- estimate price elasticity curves
- create lifetime pricing models
- redesign rating tariffs
- help with other pricing and reinsurance decisions
- feed into other processes, e.g. capital modelling and reserve estimation
- carry out financial projections for budgeting, strategy and solvency purposes
- value general insurer for sale, purchase or merger
84
Profit testing
Involves projecting future income (from premiums and investments) and future outgo (expenses and claims) to give the expected profitability of a set of premium rates.
Can also be used to assess the premium rates required to reach a certain profit criterion.
May allow for any statutory solvency margin required and for the strength of the statutory reserving basis
85
Uses of price elasticity curves
- estimate changes in business volumes as a result of changes in premium rates
- useful for highly competitive classes of business
86
Customer lifetime value (CLV)
The inherent added worth of an existing customer to whom a policy has already been sold.
Takes account of the policyholder's likelihood to renew, the policyholder's acceptance of future rate increases and potential for cross-selling other products
