Senator LUDLAM: I will change the subject a bit to talk about the NATSEM modelling that was released I think earlier in the week and the relationship between some of the results that fell out of their model with what Treasury put to government. Can you tell us whether the model and some of their findings look reasonably plausible to you and maybe just explain why you think there was some departure in some of the results from what you put forward?
Mr Robinson : The National Centre for Social and Economic Modelling released some analysis that they had undertaken of the impact of the carbon price on households.
Mr Robinson : The National Centre for Social and Economic Modelling released analysis they had undertaken of the impact of the carbon price on households. The modelling uses a broadly similar approach to the approach that we have undertaken, although there are some slight methodological differences in the approaches. In particular, the price impacts that NATSEM have used have been taken from the Treasury modelling. They took, for example, the 10 per cent impact on electricity that we estimated and other price impacts for other goods and services and applied those to the household expenditure survey.
The results in the report look reasonably plausible to us, having looked at it only for a day and not being across the full detail of the methodology and the actual way it was undertaken. I guess some of the broad results would appear to have been driven by some differences in the data that NATSEM used in their modelling. At the time the modelling was undertaken in Treasury the most recent, comprehensive data available on household expenditures in Australia was from the 2003-04 household expenditure survey. We undertook quite a comprehensive process of updating that data to make it reflective of what we thought the world might look like in 2012-13. We updated for changes in prices for different types of goods and services from 2003-04-electricity prices, for example-and we took account of changes in incomes and demographic change. NATSEM used the 2009-10 household expenditure survey, which was released recently by the Australian Bureau of Statistics.
We think that some of the differences in the results are due to the use of the updated data. In particular, the result that NATSEM got on the average price impact faced by households was a little lower than the Treasury estimate. NATSEM estimated a weekly impact of about $8.50 per household, compared to Treasury estimates of $9.90. We think part of the reason for that result is driven by a reduction in the emissions intensity of consumption over time. For example, although electricity prices have increased substantially over the period-and our modelling updates took account of those increases-at the same time electricity consumption had been falling a little bit. That seems to be borne out in those figures. There seems to be also some substitution away from emissions intensity in terms of electricity consumption. That seems to have driven that result.
Senator LUDLAM: Which is the whole point of bringing this scheme into being in the first place.
Mr Robinson : That is correct.
Senator LUDLAM: So they have actually folded back the beneficial impacts of a slightly more energy efficient economy as a result of the scheme coming into being. Have you been asked to feed the 2009-10 household expenditure survey results back into your model or will you just let your work rest in that area?
Mr Robinson : At this stage we have not been asked to update that, although in the ordinary course of events of updating models that we regularly maintain within the department we would incorporate updated data as it became available.
Senator LUDLAM: Is that a big task?
Mr Robinson : In this case it is quite a large task. Not only are there aspects around the actual household expenditure data but also as a flow-on from the household expenditure data there are changes to the consumer price index. The construction of the index is updated. Effectively the 15th series CPI, which has been in place since 2005-06, has just been updated to the 16th series. There are other sources of data that fall as part of that. In addition to the household expenditure data an important input to the modelling that we undertake is data on household incomes. The ABS also released the 2009-10 survey of income and housing at the same time. That is a dataset which is incorporated into the STINMOD model, which would also be required as part of the update.