1
Appendix H:Australian Government Actuary Assessment of KPMG’s Analysis and Conclusions – August 2012
APPENDIX H | NDRRA PHASE 2 REPORT
H:\Agencies\comcover\ndrra review\local govt\Let-pm-review-final.docx
1
Copyright Notice
ISBN: 978-1-922096-15-9 (print)
ISBN: 978-1-922096-16-6(online)
With the exception of the Commonwealth Coat of Arms, this work is licensed under a Creative Commons Attribution 3.0 Australia licence (CC BY 3.0) (
This work must be attributed as: “Commonwealth of Australia, Department of Finance and Deregulation, Review of the Insurance Arrangements of State and Territory Governments under the Natural Disaster Relief and Recovery Arrangements Determination 2011”
H:\Agencies\comcover\ndrra review\local govt\Let-pm-review-final.docx
1
Ref No.: Telephone: 02 6263 4127
Facsimile: 02 6263 2677
E-mail:
21 August, 2012
Mr David Tune PSM
Secretary
Department of Finance and Deregulation
John Gorton Building
King Edward Terrace
PARKES ACT 2600
Dear Mr Tune
KPMG review of state insurance arrangements under the NDRRA
You have asked me to review a report prepared by KPMG as part of the NDRRA review process (KPMG overarching report). I have reviewed a final draft of this paper, dated 3 August 2012.
The KPMG paper sets out the results of its review of states’ and territories’ (including local governments) insurance arrangements in respect of their essential public assets (EPA). The KPMG review was commissioned by Finance. My high level conclusions are set out below.
The submission process was intended to provide the Commonwealth with comfort that states and territories(jurisdictions) have insurance arrangements in place which result in a fair NDRRA deal for both jurisdictional and Commonwealth taxpayers.
Spin-off benefits from the submission process would ideally have included the provision of useful information for both the Commonwealth and jurisdictions around their relative risk exposures under the NDRRA. Some useful information has indeed become available.
In practice, though, the submission process has fallen short of delivering on its objectives.
It will be both important and instructive to consider carefully the reasons for this outcome. This consideration would inform the most fruitful path forward. On the other hand, without this consideration, it would be easy to proceed in a way that was unlikely to yield good results.
In my view, part of the reason relates to less than complete understanding of risk exposure.
KPMG has made a number of recommendations in this regard and, to the extent that future submission processes follow a similar path to the one just completed, I generally agree with these recommendations.
However, in my view, it is also arguable that part of the reason for the less than ideal outcome relates to the submission process itself.
A particular challenge for jurisdictions concerned the related requirements to demonstrate that their insurance decisions and arrangements are cost-effective to both the jurisdiction and the Commonwealth and also that the financial exposure for both jurisdictional and Commonwealth taxpayers has been minimised as a result of their insurance decisions.
In my view, careful thought needs to be given to these requirements in order to ensure that practical progress is not impeded.
Since the NDRRA operates on a net-loss basis and since it does not differentiate between insurable and uninsurable assets, it does not provide a strong natural incentive to jurisdictions to seek to obtain insurance cover for their insurable EPA.
The revisions to the Determination sought to address this deficiency by requiring jurisdictions to consider cost-effectiveness and financial exposure minimisation for both jurisdictional and Commonwealth taxpayers in making their insurance decisions.
However, the submission process has highlighted the very significant challenges involved in giving effect to, and monitoring compliance with that requirement.
Among a number of sound recommendations made by KPMG around thresholds and rates, I strongly encourage examination of the recommendation for differential thresholds for insured and uninsured assets.
In my mind, this recommendation represents a specific example of a family of models where the treatment of insurable assets differs from the treatment of uninsurable assets.
Models that differentiate between insurable and uninsurable assets can, in theory, deliver a number of benefits. Importantly, these models have the potential for improved administrative efficiency as there could be less need for jurisdictions to demonstrate, and for the Commonwealth to test, compliance with the requirements around cost-effectiveness and financial exposure minimisation. At the same time, it is likely to be easier to achieve these objectives under this sort of model than under the current structure. Other potential benefits include:
•A natural incentive to jurisdictions to explore a range of insurance options for their insurable assets
•More transparency and certainty for jurisdictions and the Commonwealth
•Financial efficiency
Finally, for completeness, there are two attachments to this letter. The first gives a very brief summary of the relevant NDRRA provisions and the second summarises KPMG’s main findings and my brief comments on those findings.
Should you wish to discuss, please call me on 6263 4127.
Yours sincerely
Yours sincerely
Peter Martin
Australian Government Actuary
Attachment 1
Background – the NDRRA Determination
The NDRRA Determination sets out the rules for Commonwealth assistance in the event of natural disaster.
The Determination covers assistance to households, businesses, community groups and states and territories (jurisdictions).
The relevant assistance considered here is to jurisdictions and relates to the restoration of Essential Public Assets (EPA) following a natural disaster.
Key elements of the Determination include:
•The program to provide assistance for the restoration of EPA operates on an expenditure-year basis. An expenditure year is any of:
–the financial year in which a natural disaster occurs; and
–the following two financial years.
•Eligible expenditure refers to expenditure incurred by a state in respect of eligible measures (restoration of EPA and other measures). Expenditure applied to the restoration of EPA is eligible expenditure to the extent that it is net of any third party recoveries (eg insurance recoveries)
•As the amount of eligible expenditure incurred by a jurisdiction during an expenditure year increases, so does the level of Commonwealth assistance that is provided. The table below illustrates the thresholds for assistance and the rates at which is assistance is provided in relation to the restoration of EPA:
Amount of jurisdiction’s eligible expenditure / Commonwealth contribution< 0.225% of jurisdiction’s revenue / 0
> 0.225% but less than 0.394% of jurisdiction’s revenue / 50% of the part above 0.225% of revenue
>0.394% of jurisdiction’s revenue / 0.169% of revenue plus 75% of the part above 0.394% of revenue
Following the Queensland floods of 2011, the Determination was revised. The main elements of the revision include:
•Commonwealth assistance for the restoration of EPA is now, in effect, conditional on a jurisdiction having in place ‘appropriate’ insurance arrangements in respect of its EPA
•The Commonwealth will review jurisdiction’s insurance arrangements from time to time to assess appropriateness
•If a jurisdiction’s insurance arrangements are found to be inappropriate, then the level of assistance available to the jurisdiction will be reduced (by adjusting either or both of thresholds and rates)
In order to facilitate the Commonwealth review jurisdictions have been required to:
•Self-assess their own insurance arrangements
•collate certain information, including in respect of their local governments
•provide this information to the Commonwealth
This process took a number of months.
Attachment 2
KPMG’s findings
KPMG was asked to:
•advise on suitable benchmarks for assessing the appropriateness of jurisdictional insurance arrangements
•assess the appropriateness of jurisdictional insurance arrangements against the benchmarks
•make recommendations around thresholds and rates in light of that assessment
Benchmarks
Quantitative benchmarks
KPMG’s stated preference would have been to seek to establish one or more quantitativebenchmarks for each jurisdiction.
KPMG have highlighted a number of quantitative metrics that should ideally be available for each jurisdiction. Most notably, KPMG identifies two measures that should be available:
- expected annual costs
- maximum event retention
In the circumstances, insufficient data was available in this respect and KPMG has indicated that this is something that should be pursued in future.
A central message from the KPMG report is that, in order to be able to manage risk, it is first necessary to understand it.
Comment
The two measures identified by KPMG (expected annual costs and maximum event retention) would provide important information:
•firstly, for the jurisdiction around its exposure to natural disaster risk; and
•secondly, for the Commonwealth around its NDRRA exposure to the jurisdiction.
Indeed, it is difficult to see how a jurisdiction could properly understand and, therefore, manage its natural disaster risk exposure without this sort of information.
In order to establish a quantitative benchmark, against which to assess the appropriateness of insurance arrangements, it is likely that more information would be required in addition to the two metrics presented above. In particular, information would also be needed around the cost of insurance and the impact of insurance on the maximum event retention.
In other words, while I agree with KPMG’s suggestion that information around expected annual costs and maximum event retention should be provided, in my view, this information should be provided under a range of insurance scenarios. Firstly, it should be provided assuming no insurance. Secondly, it should be provided assuming current insurance arrangements are in place.
This sort of information would assist in the establishment of a quantitative benchmark.
Importantly, to establish a benchmark, a degree of subjective judgement will inevitably be needed to identify the appropriate trade-off between increased expected annual costs (which arise with the purchase of insurance) and a reduced event retention (which is the benefit of insurance). The required judgement is subjective because the costs and benefits of insurance cannot be measured in the same units. The cost of insurance is measured in annual expenditure dollars while the benefit of insurance is a reduction in volatility.
In summary, KPMG’s report correctly points to the desirability of seeking to ensure that certain quantitative metrics are available. KPMG has nominated expected annual costs and maximum event retention in this regard. In my view, there would be merit in seeking to ensure that these metrics are available, both assuming no insurance is in place and also under current insurance arrangements.
There would be costs involved for jurisdictions, which would need to commission financial modelling from an appropriately qualified expert.
However, this information would provide useful information for both the jurisdiction and the Commonwealth.
Qualitative benchmarks
KPMG then considered a qualitativebenchmark related to the insurance decision-process undertaken by a jurisdiction.
The qualitative benchmark seeks to test whether or not a jurisdiction has undertaken a suitable process in arriving at an insurance decision.
At a minimum, KPMG has argued that jurisdictions should:
•seek to understand their risk exposures
•obtain market quotes for insurance
•monitor the performance of their insurance arrangements over time
In summary, KPMG argue that the insurance-decision should be predicated upon an ongoing, systematic, evidence-based process.
Comment
KPMG’s approach appears consistent with the requirements of the Determination which provide that ‘the onus is on states and territories to explore a range of insurance options in the marketplace and assess available options on a cost-benefit basis’.
More broadly, KPMG’s approach is consistent with a reasonable expectation that jurisdictions adopt and implement sound risk management principles.
Appropriateness of State insurance arrangements
Preliminary
As well as remarking on the appropriateness of jurisdictional insurance arrangements, KPMG has made a number of observations around the nature of the arrangements themselves.
Firstly, KPMG has broadly split EPA into two categories, referred to in its report as civic assets and civil assets. Civic assets tend to include buildings such as schools, hospitals, libraries etc and their contents while civil assets include roads, bridges, tunnels as well as underground pipes, power distribution infrastructure etc. KPMG notes that insurance arrangements are in place in most jurisdictions for civic assets. Cover is available for natural disaster risk[1] although sub-limits might be in place on flood, for example. Relatively little insurance is in place for civil assets.
Secondly, KPMG has described the interaction between the NDRRA and the arrangements that jurisdictions have in place with their local governments for financing the restoration of local government assets following natural disaster. Related to this, KPMG observed that it did not appear that all jurisdictions have an understanding of their consolidated asset exposure, taking into account local government assets as well as jurisdictional assets.
KPMG noted that the interpretation of the term EPA appears to vary between jurisdictions, meaning that it is difficult to establish a reliable register of EPA which might be eligible under the NDRRA.
KPMG Assessment
KPMG highlighted a number of deficiencies in jurisdictional submissions. For example, jurisdictions are required under the Determination to ensure that they put in place insurance arrangements which are cost effective for both the jurisdiction and the Commonwealth. KPMG report that no jurisdiction has demonstrated (provided supporting analysis) that this requirement has been met.
Similarly, jurisdictions are required to ensure that the financial exposure borne by jurisdictional and Commonwealth taxpayers is minimised. Again, jurisdictions have not provided evidence that this requirement has been met.
Thus, jurisdictions which do not have commercial insurance in place (NT and Tasmania for all assets and most jurisdictions for civil assets) have not provided analysis in respect of the requirements around cost-effectiveness and financial exposure minimisation to support the no-insurance decision.
Where jurisdictions do have commercial insurance in place (all but NT and Tasmania for civic assets and ACT and Victoria, along with some local governments, for civil assets) the cost-effectiveness and the extent to which financial exposure is minimised as a result can only be inferred. Thus, although it is clear that some decision-process has been undertaken, jurisdictions have not provided supporting analysis in respect of these requirements.
KPMG reports that Queensland has recently secured a reinsurance program over its civic assets but not its civil assets (in particular, roads). This program was not in place prior to the revision to the Determination. However, since the majority of the NDRRA expenditure following the Queensland floods has been in respect of roads, the extent to which these new insurance arrangements will result in reduced NDRRA outlays is unclear. KPMG has, therefore, suggested that Queensland be asked to provide annual submissions for the next few years. Further, KPMG has suggested that these submissions should include the type of quantitative information set out above.
Comment
KPMG has highlighted the significant effort that has gone into preparing these submissions, particularly when local government profiles are consolidated within the relevant jurisdictional submission.
Despite this, KPMG has been unable to strongly endorse any submission.
In my view, the reasons that KPMG has provided for reaching its position are valid.
The onus is on jurisdictions to demonstrate that they have met the requirements of the Determination. Notably jurisdictions are required to put in place insurance arrangements which are cost-effective to both the jurisdiction and the Commonwealth and which minimise the financial exposure to jurisdictional and Commonwealth taxpayers. To the extent that jurisdictions have not done this, it is difficult for KPMG to conclude that they have fulfilled their obligations under the Determination. Note that this comment is in no way intended to diminish the practical difficulty presented by these requirements.
Thresholds
The Determination implies that, should a jurisdiction’s insurance arrangements be found to be inappropriate, then an adjustment should be made to the threshold and/or rate of assistance that is available to that jurisdiction under the NDRRA. Put bluntly, if a jurisdiction is found not to have enough insurance in place, then the Determination provides for the level of its NDRRA support to be reduced. The mechanism available to give effect to this reduction is adjustment to either or both the threshold and the rate(s).
KPMG was asked to make recommendations around thresholds and rates as part of its scope of work.
Although KPMG has been unable to strongly endorse any jurisdictional submission, it has not made specific recommendations in respect of the thresholds or rates applying to individual jurisdictions.
KPMG has, however, made a number of suggestions around the calibration of the NDRRA assistance formulae:
•that thresholds and rates that apply to assistance for the restoration of EPA be de-coupled from those applying elsewhere
•that the basis of assistance (for the restoration of EPA) be changed from an expenditure-year basis to a per-event basis
•separate thresholds should be established for insured and uninsured assets
•the basis for the threshold should take account of both the value of EPA and the amount of jurisdictional revenue rather than simply being based on jurisdictional revenue
• review of the small disaster criterion so that only significant disasters qualify for assistance in respect of EPA
Comment
Although KPMG has been unable to strongly endorse any state submission, KPMG has not, at this stage, considered whether differential thresholds and rates should apply to particular jurisdictions.
In my view, this response is sensible at this time.
It is clear that sufficient information is not yet available to support an informed recommendation around differential thresholds and rates.
KPMG has emphasised the importance of improving understanding as an essential prerequisite for improving risk management. This submission process might be seen as a first step along the path to better understanding.
In relation to KPMG’s general recommendations around thresholds:
•I agree that de-coupling the EPA parameters from the other parameters is needed in order to increase the likelihood that the right incentives are in place for jurisdictions to adopt and implement appropriate risk management practices in respect of EPA;