Approach
In estimating the required premium pool for 2006-07, SAWC’s actuaries would estimate the present value of all expected benefits associated with the underwriting period. In addition to these projected benefit costs, associated expenses, and an allowance for profit, we understand that SAWC also increases levies in order to improve the funding position of the scheme.
The methodology adopted may be regarded as a simplified actuarial approach. In essence, it is necessary to build a model which estimates the number and dollar amounts of weekly claims that are likely for the 2006-07 injury year. Such payments shall be made for many years, potentially until all workers injured in 2006-07 reach retirement age. In all cases we have assumed that changes to the amounts, caps, thresholds, and length of weekly payments shall only affect future underwriting periods – that is, changes shall not be retrospective and thus have no effect on existing claimants. Furthermore, while the analysis estimates the financial effect of legislative changes via changes to the distribution of weekly claims, it does not attempt to model any additional ‘behavioural’ effects. This is a conservative approach, as it is likely, as in the case of redemption policy, that changes to legislation or practice will likely induce marketplace changes which could amplify any ‘savings’. Using statistical information published by the WorkCover Corporation, estimates are made of the total amounts of weekly claims for a future accident year. A pattern of assumed continuance rates is also produced. This allows estimation of the pattern of claims for all future payment periods. The continuance rates are based on recent experience of the WorkCover Corporation. Different scenarios of future continuance rates have been modelled.
curves. While there is little difference between the curves in the early months post injury, it is the difference in the curves after six months which is relevant, as these are the periods for which the most significant legislative changes are proposed.
With regard to the dollar amounts of weekly benefits, we were unable to obtain data on the distribution of pre-injury earnings, or individual benefit amounts. In order to analyse the financial effect of a change to caps and thresholds, it is necessary to estimate the distribution of wages in South Australia. Data on household income was obtained from the Australian Bureau of Statistics, and a log-normal curve was fitted to the data. This curve closely represents the distribution of income, and allows a good estimation of changes to thresholds and caps. For convenience, income has been expressed as a multiple of AWE.
The distribution is shown below. The curve exhibits the feature where most earners earn lower amounts with fewer earners with higher amounts. The distribution is particularly important when considering caps for weekly payments. At present weekly benefits are effectively capped at 2 x AWE, so using the distribution shown, about 9% of injured workers are subject to the cap, and so are restricted to 2 x AWE for their benefit. If the cap is lowered to 1.5 x AWE, approximately 22% of injured workers are now subject to the cap.
$1,960 per week). As outlined above, the distribution of NWE mean that a cut in the cap does not lead to a pro- rata reduction in benefit costs.
If all other benefit designs were held the same but the cap was lowered, approximate savings are as follows:
Cap at 1.5 x AWE – saving of approx $9.5m, or 0.02% of levy rate.
Cap at 1.25 x AWE – saving of approx $20.4m, or 0.04% of levy rate.
Results
The combination of the continuance curves, distribution of pre-injury earnings, and recent experience then makes it possible to estimate the direct savings associated with particular legislative amendments. This has been done in each case by making comparisons of the amended situation with a base model representing the existing legislative package.
As a range of underlying assumptions have been adopted, particularly with respect to continuance curves, a range of outcomes is also produced. These are presented in Appendix D. Again, it is important to realise the limitations of the data underlying the models, and to view the results as indicative.
Financial effects of proposed legislative packages In Appendix C, the approach to modelling was outlined. The following scenarios give estimates of the savings under different legislative changes. It is important to realise the modelling is simplistic in the sense that it has relied on publicly available data, so relies on some subjective (yet sensible) assumptions. The results should be considered indicative. Better estimates and/or ‘tighter’ ranges may be achieved with access to better data.
For the continuance of claims, three possible underlying claim curves (median, optimistic and pessimistic) have been considered. Similarly, from knowledge of other jurisdictions, various scenarios regarding the continuance at 104 weeks have been assumed. It is important to understand that the experience of one jurisdiction may not be readily transferred to another jurisdiction. Benefit differences,
environment and culture are some of the factors that will affect claim experience. This is one of the reasons that a range of possibilities has been considered.
The variation of assumptions leads to ranges of financial effects. It is important to consider the financial estimates as indicative only. While the ranges can be quite wide, the modelling is imperfect, and the ‘true’ estimate could well lie outside the given range. This is particularly important when we consider the flow-on or behavioural effects of legislative change that has not been considered explicitly. For example, any reduction in income replacement benefits are likely to lessen the incentive for less meritorious claims to remain on benefits, amplifying the effect of legislative change beyond the direct effect on existing structures.
The primary assumptions made with regard to income distribution and continuance rates are contained in Appendix C.
Experience regarding partial payments, redemptions etc could not be considered explicitly, so with no reasonable alternative, have assumed ‘experience shall be similar to recent periods’.
Modelling the 104 week threshold, it is assumed that the three possibilities are that 45%, 55% or 65% of pre-104 week claims shall continue past 104 weeks. Requirements for the104 week mark are assumed to mirror the Victorian requirements.
The scenarios have also been modelled so no payments are received more than six years after accident. This is designed to mimic the situation in some jurisdictions where there are overall dollar caps on the amount of benefits. This modelling component is imperfect, but is reasonable given the data available.
Results are given for the scenarios below. In each case there is a ratio of NWEs paid for each post-injury period. Different caps (1.25 or 1.5 times AWE apply). In each case the assumptions re the 104 week thresholds apply. In the later scenarios, payments are limited to six years beyond accident. In each scenario the financial estimate of savings relates to one year only, so has also been expressed in a percentage of the levy saving (presently 3.0%). At present, part of the levy is designed to facilitate the return to full funding. The savings on levy are expressed assuming the savings are on
the entire levy amount rather than trying to isolate the components to meet future benefits as distinct from the component to return the scheme to full funding.
Scenario 1
Weeks post injury Multiple of Cap as Other (and 2 week NWEs multiple
deductible) payable of AWE 0-13 85% 1.25 13-52 75% 1.25 52-104 75% 1.25 104+ 75% 1.25 Victorian rules post 104 weeks No time limit on benefits Savings: from $117m to $150m, or from 0.65% to 0.83% of levy rate
Scenario 2
Weeks post injury Multiple of Cap as Other (and 2 week NWEs multiple
deductible) payable of AWE 0-13 100% 1.25 13-52 85% 1.25 52-104 70% 1.25 104+ 70% 1.25 Victorian rules post 104 weeks No time limit on benefits Savings: from $123m to $153m, or from 0.68% to 0.85% of levy rate
Scenario 3
Weeks post injury Multiple of Cap as Other (and 2 week NWEs multiple
deductible) payable of AWE 0-13 100% 1.5 13-52 85% 1.5 52-104 70% 1.5 104+ 70% 1.5 Victorian rules post 104 weeks No time limit on benefits Savings: from $117m to $147m, or from 0.65% to 0.82% of levy rate