CHAIR —Today the committee continues its inquiry into the proposed changes to the Australian taxation system. The Senate referred the inquiry to the committee on 25 November last year. This hearing relates to the second stage of the inquiry. In this stage, the committee is examining the broad economic effects of the government's taxation legislation proposals. It will have regard to the fairness of the tax system; the living standards of Australian households, especially those on low incomes; the efficiency of the economy; and future public revenues. The committee will report on the second stage of its inquiry in just a few weeks, by 19 April this year. This committee called for submissions to be lodged on 29 January 1999—in fact, the committee, I am advised, is still accepting submissions—and so far has received over 1,400.

This is the final public hearing to be held by the committee in Canberra in the course of the inquiry and, I think, the final public hearing for the whole of the inquiry. I note for the record that the committee has agreed to release all of the submissions received in the course of the inquiry, except for those submissions marked `Confidential'.

For the record, this is a public hearing and, as such, members of the public are welcome to attend. Before we commence taking evidence, let me place on record that all witnesses are protected by parliamentary privilege with respect to the submissions made to the committee and evidence given before it. Parliamentary privilege means special rights and immunities attached to the parliament or its members and others necessary for the discharge of the functions of the parliament without obstruction and without fear of prosecution. Any act by any person who operates to the disadvantage of a witness on account of evidence given by him or her before the Senate or any committee of the Senate is treated as a breach of privilege.

I now welcome Professor Neil Warren, Professor Ann Harding and Mr Martin Robinson. I invite you to make an opening statement. The normal format that we have adopted is that we have, of course, received your written research—that is before us and we have had a chance to become familiar with it—and we would invite you to provide a short overview of what that contains and then be available for questions from the committee. Do you have any comments to make on the capacity in which you appear?

Prof. Warren —I am appearing here as one of the co-authors of the study commissioned by the Senate inquiry.

Prof. Harding —I am appearing here as a co-author of the report, as is my colleague Martin Robinson.

CHAIR —I have a question for the committee. This is the document that has been supplied by our witnesses. Is it the will of the committee to release this document publicly, after the fact it would seem? There being no objection, it is so ordered, and the document is now published. With those remarks, I invite you to begin.

Prof. Harding —We were going to divide this presentation into six parts. First, I was going to give you an overview of the options that we have modelled. Second, Professor Warren was going to describe how the indirect tax changes were modelled in the STATEX component of the STINMOD-STATEX model and how the price effects were calculated. Third, Martin Robinson, from NATSEM, was going to give you an overview of how the income tax and cash transfer payment changes were modelled in STINMOD. Fourth, Professor Warren was going to compare some of the results for options 1 to 6. Fifth, I was going to compare some of the results for options 3B, 5B and 7. Finally, we were going to make some concluding comments about what is an extremely long and complicated report.

There are a couple of points I want to make first about the methodology. We have modelled 10 separate options for the Senate committee. We are looking at the medium-term price effect, so we are looking effectively at the price effect in year 2, which is 2001-02. As you know, it is expected that price effects will be higher in the first year following the introduction of the GST from July 2000. But the committee asked us to model the medium-term impact because that gives you an indication of whether households are going to make an ongoing loss or gain from the impact of the tax reform package.

As you know already from evidence before this committee, Treasury had estimated that the year 1 price effect would be 2.5 per cent and the year 2 price effect would be 1.9 per cent. The reason why prices are expected to be higher in the short term is that some of the removal of indirect taxes does not occur until the second year following the start of the GST.

So it is quite hard to think through conceptually the world that we are modelling, but with only one exception. At the committee's request, we have modelled the forecast July 2000 world for population benchmarks, for earnings, for the income tax system and for the family benefits system proposed by the government, and we have melded onto that the forecast July 2001 indirect tax effects, which is year 2, for the longer-term impact of the indirect tax effects. We have also used the government's intended longer-term pension increase. So, although the government had said in year 1 that it would increase pensions by four per cent, we have used their longer term estimate of 3.4 per cent, which is where they intended pensioners to end up in year 2. It is a bit of a mix, but the intent is to give you a feel for what the medium-term impact of this package is likely to be, after a lot of transitional effects in the first few months have settled down.

Senator CONROY —What is your definition of medium term?

Prof. Harding —Really it is year 2. It is 2001-02, but the income tax system that we modelled was that proposed by the government for July 2000. If we had not done that, we

would have had to index every social security and income tax threshold and so on just for that short one-year period between July 2000 and July 2001.

Senator CONROY —I just wanted to clarify what you meant by that.

Prof. Harding —We have modelled the effect of the ANTS package on hypothetical or cameo households. There is an alternative methodology that has been used, for example, in the ACOSS-Melbourne Institute reports, where you use the expenditures and incomes of households as actually shown in the ABS household expenditure survey sample. As this committee is aware from papers that have been put before you, there are some problems associated with using the HES data to look at savings rates and dissavings rates and to look at what price changes are likely for small subgroups of the population. Because of these problems, the committee asked us to use the HES data to simulate the average price effects facing particular types of households and to impose hypothetical dissavings rates just to see how sensitive the results were to different assumptions about dissaving.

I will just go through the options that we modelled. Option 1 is the government's option using the government's methodology. So when you are saying how much extra indirect tax a household is going to have to pay, the measure of the price effect of the indirect tax changes is assumed to be the CPI. The CPI measure excludes the impact of the tax changes on housing and tobacco prices, and that is for reasons that Professor Warren will explain in more detail. The assumed increase in pensions is 3.4 per cent. This is the same as the figure the government actually used in the cameo households published in the ANTS document. So it is their intended ultimate outcome for pensioners.

Senator MURRAY —It is 1.9 plus 1.5 per cent, isn't it?

Prof. Harding —Yes. There is no dissaving, so when you are answering the question of how much extra money a household needs in their pocket to be fully compensated for the price rises they are facing at the supermarket checkout the Treasury assumption was that you do not take account of the fact that they might be spending some of their savings. The assumption was that, when looking at what compensation was required, you just assume that the amount of money people were spending was exactly the same as their disposable income. So there was no dissaving occurring and no decline in the purchasing power of the dissaving component of people's expenditure. These were the two marginal tax rates assumed to apply in those tax bands from $20,000 to $50,000 and from $50,000 to $75,000.

Neil Warren will go into this in more detail, but on the whole, when we replicate the Treasury's approach and use their methodology, we have matched their numbers very well. So we get very slight differences due to our getting a very slightly different CPI estimate in the second year from that derived by the Treasury, and we have slight differences because we are using more recent CPI and AWE inflators. There are some slight differences in how we have modelled some of the programs where we believe there were some problems in the original Treasury analysis and we have done it slightly differently. On the whole, our results match the Treasury results, if you use their methodology. This was the package as run by the government.

Option 2 was what happened if you just varied one assumption, and that was that tobacco and housing should not be excluded from the price impact of the indirect tax reforms. The result of that, as Neil will discuss, is a higher measured price effect from the indirect tax reform package. That is the only difference between those two options.

Option 3 starts moving further down the chain in that it varies the two key assumptions made by Treasury in their modelling which have attracted the most criticism. The first assumption that Treasury made in their modelling was that the consumer price index would provide an adequate indicator of the impact of the indirect tax reforms on the prices facing all households, irrespective of their income or their family composition. So we varied that assumption in option 3 and in subsequent options to say that the price effect facing households is the average price effect for that particular cameo household type. So we have 29 different price effects because we have more than 29 different types of households.

I should emphasise again that this is the average price effect for all households within that cameo. So, say, for the single person results, the price effect we get for the single person household is the average for all single person households. In fact, some of them would face below average and some would face above average price increases depending on their particular expenditure patterns.

The second key assumption that the committee asked us to vary was the assumption that no-one was dissaving or, if they were dissaving, that no additional compensation was required. So, given the difficulties in trying to estimate exactly how much people are dissaving, we instead just completely arbitrarily imposed a maximum 10 per cent dissavings rate amongst households who had no private income, and that was just assumed to decrease lineally to reach zero dissaving at $20,000 of private income. I should add that we pulled those figures out, effectively, from the 1985 draft white paper on tax reform, written by the Treasury, just with the dollars inflated to year 2000 dollars.

One important point is that you should be aware that we have assumed that households who are saving do need to be compensated for the decline in the purchasing power of their savings. So the only area where we have varied the Treasury assumption is for dissavings. For dissavers we have said you need slightly more compensation than the Treasury had estimated, but for people who are saving or spending all of their income the approach is the same as used by the Treasury. It is also important to note that tobacco and housing are included within the price effect in option 3.

So option 3 was effectively what happens when you move to own price effects and you start looking at dissaving and option 4, which the committee asked us to model, was what happens if you assume there is not 100 per cent shifting of the indirect tax cuts—the removal of stamp duties, wholesale sales tax and some of the other indirect taxes. Option 4 looks at what happens if businesses fail to pass on via lower prices to consumers the full impact of the cuts in their existing indirect taxes. In this case we looked at wholesale sales tax and stamp duties. It was assumed that they pass on only 70 per cent of the reduction in those two taxes in that first year.

Options 5 and 6 are the two food out options that the committee asked us to look at, and I note that the committee specified that food was to exclude expenditure on restaurants and

takeaway food. Option 5 shows the outcome when food is excluded from the GST revenue base and the lost $5 billion of revenue is recouped by reducing the generosity of the income tax cuts. You can see the way we have done it here. We essentially followed an option put forward by ACOSS, the peak welfare group, where the government's proposed 30 per cent marginal tax rate between $20,000 and $50,000 was increased to 33[half ] per cent and the 40 per cent marginal tax rate between $50,000 and $75,000 was increased to 43 per cent. In essence, we financed food out by reducing the generosity of the income tax cuts.

Option 6 is an alternative way of looking at it, and it says what happens if you take food out but instead increase the GST rate upon all those other items that are still within the base of the GST. We estimated that that would require a 12 per cent GST rate. In that case, the income tax cuts stay as they were originally planned by the government.

Things vary quite a lot from this point on. In all of these first six cases, the assumed increase in social security pensions has stayed at 3.4 per cent, which was the government's intended long-term position. Except for option 1, which was the government's methodology, tobacco and housing price changes have been included in all of those cases within the estimate of the price increases facing households. But both of these assumptions are varied in option 7 and in the 3B options which follow.

Option 7 shows the distributional impact when pensions and allowances are increased by a more generous six per cent but food remains within the base and the slightly higher compensation costs of the slightly higher pension increases are funded through slightly higher than planned income tax cuts. The government had planned 30 per cent and 40 per cent, but to fund more generous compensation and to keep the package revenue neutral they had to go up to 30.5 per cent and 43 per cent marginal tax rates.

In doing this, the impact upon prices is still household specific, so we have not used the CPI, and we still have dissavings within the picture. But we have excluded tobacco and housing from the measured impact of this tax reform package, and that is also true for all of the B options because we believe, for reasons that Professor Warren will explain in more detail, that there is a case for excluding tobacco and housing from the measured effect of the tax reform package. So option 7 was essentially provided just to give the Senate committee a feel for the likely distributional impact of a food out/income taxes up package versus a food in/compensation up package.

Finally, when modelling the first lot of options which the committee asked us to model, it was not always clear to us what the most appropriate benchmark or counterfactual should be or what the fairest assumptions were to make. For example, under option 5 we estimated that, if food were excluded from the GST base, the impact of the indirect tax reforms on the CPI would fall by 1.1 per cent. So if food is out, prices will increase by less. But if that happened, would it still be fair to assume that the government would still increase pensions by the planned 3.4 per cent. So the original 3.4 per cent equalled their expected 1.9 per cent CPI effect, plus the 1.5 per cent buffer for pensioners.

So in these B options we have tried to give you a feel for what happens if you maintain the conceptual integrity of the government's package in the sense that the increase in pensions in each of these B options is the estimated increase in the CPI, plus the

government's intended 1.5 per cent buffer. In each case tobacco and housing are out of that CPI estimate. In one case, in option 5B, if you increase pensions by the CPI movement, plus 1.5 per cent, you have only a 2.3 per cent increase in pensions. That is obviously lower than the 3.4 per cent increase in pensions that had been originally envisaged based on the earlier higher CPI effects. Because that saves money, we just restored some of that money to taxpayers by giving them a tax cut relative to option 5. So there is just a half a per cent difference in the marginal tax rate between $20,000 and $50,000. I will now hand over to Professor Warren to explain the modelling of indirect tax reforms.

CHAIR —Just before you do, why did you not do all of the work that was commissioned?

Prof. Harding —To my knowledge, we did. But you feel that is not the case?

CHAIR —No. The terms of reference are on page 42 of the document you have submitted and they ask for the first year impact from July 2000. You have reported to us that you did not do that. I am just wondering why you did not do that.

Prof. Harding —Neil might want to come in here. I think it is pretty hard to work out conceptually what it is that you are modelling. Although we were asked to model it in July 2000, the terms of reference did also say that we should assume that all of the indirect tax increases were there and that all of the indirect tax cuts were there. Some of the indirect tax cuts are not scheduled to come in until 2001-02. So, in a sense, we felt that the brief asked us to abstract from the phasing of the tax reforms issue and try to give you a feel for what it was going to look like once the dust had settled.

CHAIR —If you had done that, there would have been more losers, wouldn't there?

Prof. Warren —The answer to that is no. The reason is that we investigated that particular option. The thing that concerned me about it was that, in the first instance, we had to ask what our benchmark was, and we chose to replicate and to test the sensitivity of the cameo results in the ANTS document. I discussed with the secretariat of the committee that we should take a benchmark which is a common, accepted one, and that is the one that is in the ANTS document.

Senator CONROY —Whom did you speak to?

Prof. Warren —To Tas Luttrell, during the preparation of the terms of reference.

Senator CONROY —But when you said you spoke to the secretariat, I just wanted to know to whom.

CHAIR —He said that it was Tas Luttrell.

Prof. Warren —The point is that, in looking at that year, that 2001-02 focus, that allows us to test the sensitivity of the results that are reported in the ANTS document and to take a longer term—replicating the Treasury methodology. The question is: what difference does it

make if you move to July 2000? Subsequent to our presenting this report—I heard those concerns that you are now expressing—

CHAIR —It is just that the terms of reference make it very clear. It is very odd that, when the terms of reference were debated extensively in the committee before they were handed out, the work was not done.

Senator FERGUSON —Let him finish.

Prof. Warren —Can I tell you what the implications of the results are if you do go to July 2000 and that everything we have done does not significantly change?

Senator CONROY —What is the first year inflation rate?

Prof. Warren —The first year inflation rate is higher by 0.7 of a per cent.

Senator CONROY —So that is a 2.7 per cent first year inflation rate.

Prof. Warren —If you are going to add on the tobacco—

Senator CONROY —That is before you add on the tobacco.

Senator FERGUSON —Mr Chairman, I want to raise a procedural matter here. Are we going to hear the presentation?

CHAIR —We are going to hear the presentation, but this is a quite serious threshold point.

Senator FERGUSON —Let him answer the question.

CHAIR —Excuse me, I am the chairman here, thank you very much! This is a quite serious threshold point. The terms of reference are quite explicit. This work was not done. I understand that, if it had been done, it would have shown that there are more losers than in fact we now have before us. I am very interested in having an explanation for that. So please proceed, Professor Warren.

Prof. Harding —I would like to make one quick point. I want to point out that, even if you used the first year price effect, which Neil is saying is 0.7 per cent higher, it would not have been appropriate for us to model the 3.4 per cent increase in social security, which is effectively where the government intends to end up.

Senator CONROY —We did not ask you to make any counterfactuals; you decided to help us.

Prof. Harding —But if we had modelled it, we would have modelled a four per cent pension increase, which is what they say they also intend to do right on July 2000. So both of those things would have gone up.

Prof. Warren —Can I tell you why it makes no difference, which I hope will allay your concerns? It makes no difference because the CPI goes up and so does that July 2000 compensation. The July 2000 compensation is four per cent. So that is 0.6 of a per cent higher than the 3.4. The price effect of going to July 2000 is 0.7 higher. The issue you are talking about is about 0.1 of a percentage point. I subsequently looked at that myself, off my own bat, to address that concern because I saw that issue being raised. If everything ratchets up, the difficulty for us in undertaking that approach is that the benchmark we have taken we now model on July 2000 with four per cent compensation, a particular income tax schedule, and all the results are no longer comparable to those that are in the ANTS document. As to the conclusions that we make—sure, the inflation rate will be higher, but the losers you are looking for will not be there because everything will have moved up by the same amount.

Senator CONROY —That depends on the assumption. Professor Harding, you said that somebody at the committee asked you to average out for income. That is not in the terms of reference, is it?

Prof. Harding —Do you mean for the price effects?

Senator CONROY —Yes.

Prof. Harding —It is in the sense that we were asked to use cameo specific price effects. We understood that to mean that within each cameo group—so within each of the 29 household groupings—households would be assumed to face the same price increase as the average increase for people within that household grouping.

Senator CONROY —It is just that that is not written in the terms of reference.

Prof. Harding —The cameo specific is in there, isn't it?

Senator CONROY —Not the averaging question. You actually said in your introduction that the committee asked you to average out for income. I am just trying to find out who asked you to average out for income because that is not in the terms of reference.

Prof. Harding —If it is specific to the cameo, then by definition I think it is the average price effect for that cameo.

Senator CONROY —When you did the modelling for ACOSS last year and you produced five decile groupings within each, did they ask you in the same way or were they much more specific about it?

Prof. Warren —I did that. The point is that, if you look at the sample depth we are working with when we move to the 29 cameos, and as we admit in the paper, we are at risk of problems with the sample. If you went to a decile or to a quintile approach, you would simply have unreliable cells, not an adequate sample. So in that piece of work where there was a deep enough sample you could go to quintiles, but if the sample is not adequate there is no way in the world you can rely on going to HES and making judgments on the basis of five or 10 people.

Senator CONROY —We are getting into the substance of the debate, which is not what I am actually trying to get to. I am happy to come back and discuss that part. I am trying to ascertain who it was who said to you to average it out, because you said that the committee asked you to average it out for income. I am just trying to understand who it was who asked you to do that, because that is not in the terms of reference.

Prof. Harding —That is what we thought cameo specific CPIs meant.

Prof. Warren —Also there is an issue here about how detailed the terms of reference are before you actually ask the person undertaking the study whether they should come back to you on a dozen different points. I would be very disappointed today if we spent a considerable amount of time going through all the little details about the specific assumptions that we made that might err in a particular direction or err in another direction.

There is a possibility that those price inflation rates will vary across the groups, but you have to rely, when we are looking at those groups, on the professionalism of the researchers who are undertaking it to make some judgments. We have to make some judgments about what are appropriate procedures. If in the light of day you do not like those judgments, then you raise those judgments and we respond to them in a forum like this. But in undertaking a study you would not expect the terms of reference to be at that level of detail nor to get into that kind of discussion with the researchers, who rightly have an authority to make some judgments as the professionals that they are.

You are now questioning some very specific things that I would have some difficulty with, and we can discuss those, but you to have to defer to us on some things when you commission a study like this. I have been doing it for 20 years and Ann has been doing it for a similar amount of time. If I had a deeper sample, we could look at those issues that you are talking about.

Senator CONROY —We will come back to that issue. I am trying to ascertain the position in terms of the work that we thought we commissioned.

Senator FERGUSON —Mr Chairman, I presume we are into questions. Senator Conroy has had five minutes, so I presume that we can have five minutes as well.

CHAIR —I led off with a question about the threshold and the fact that the terms of reference were not performed by what we have had produced.

Senator FERGUSON —Senator Conroy has gone on with a whole range of other things.

Senator CONROY —No. It is on the same point.

Senator FERGUSON —I presume we are into questions, that is all.

CHAIR —We can have some questions. I just beat Senator Harradine to the punch and I am interested in giving him the call, but I am happy to give you the call as well.

Senator FERGUSON —I think we need to establish what the procedure is going to be. I understood that Professor Harding said when she started that they wanted to give us a presentation, that they would go through the presentation and then we would have questions afterwards. It seems as though that is not going to be the procedure. I am just wondering, if we have questions on the way through, whether in fact we might not even get to the end of their presentation.

CHAIR —I do not want people coming back to me later and saying that we did not raise this question in the Hansard at this point and, therefore, your argument fails. I have raised it in the Hansard at this point.

Senator FERGUSON —Mine is not an argument. I am trying to ask what the procedure is, because when we have had other people here sometimes you have allowed questions and sometimes said that we should wait until the presentations have finished.

CHAIR —My preference is to wait until the presentation finishes, but it seemed to me that Professor Harding—to take your point, Mr Deputy Chairman—had presented to us the cameo work that she had done. It seemed to me that at that point it was relevant to ask why all the work had not been done.

Senator FERGUSON —I am happy for the presentation to go on, but if there are going to be further questions I think the call should go to this side before it goes to the other side.

CHAIR —Do any of you have any questions?

Senator FERGUSON —I am prepared to let them keep going.

Senator MURRAY —I think they should carry on with the presentation.

Senator HARRADINE —I can ask it later.

CHAIR —Please proceed.

Prof. Warren —In this segment I want to walk through some of the assumptions we made. I would hope that in the process of doing that some of the concerns that have been raised will be addressed and can be focused on later on in question time. What I want to focus on in the first instance is the period that we addressed which is exactly that question that Senator Conroy just raised.

We chose that period essentially because it forms a useful benchmark against which to evaluate the sensitivity of other information that is in the public domain on this issue, in particular, the ANTS document, the government's cameos. We could have chosen July 2000, but you introduce when you do that a serious of transitional types of issues which relate to four per cent and the adjustments to the CPI as you move to 3.4 per cent over that period of the six to 12 months to when those other indirect taxes come in in July 2001 by the time they are all in.

In looking at the fact, as I raised earlier, that there is a 3.4 per cent compensation for a 1.9 per cent medium-term price effect, I felt—and I have not discussed this with Ann—that the issue of going to 2001-02 seemed to me to be more reasonable, because in the light of day you would find that it made only a marginal difference because you are dealing with 3.4, 1.9, 2.5 and four. I have investigated that subsequently and what I said earlier stands. The general pattern of all the results stands. The inflationary effects will be higher, but the number of losers and the nature of those losers remains largely unchanged. No doubt we will discuss that further later.

In terms of the indirect taxes, I looked at the 2001-02 reforms and the assumption that we made in the first instance was a 100 per cent full shifting forward. Obviously that is a controversial assumption, and I suppose a qualifier to that is the existence of the ACCC and its particular brief. But to pursue that particular issue we looked at a 70 per cent shifting assumption for the wholesale sales tax and stamp duties. There was the possibility of looking at the partial shifting of a range of other indirect taxes which obviously would have also influenced that number.

You might rightly ask, `Why those two taxes? Why not more?' Again, I defer to our particular judgment about what we saw as taxes that would make sense to look at in terms of a 70 per cent shifting assumption. I did not feel that that would be the case in relation to FID. It obviously would be a case in relation to stamp duties and potentially a case in relation to wholesale sales tax, because there is so much of those taxes which are embedded in business inputs. In that sense that assumption more reasonably fits with wholesale sales tax and stamp duties. There is obviously an issue there.

There is equally an issue with diesel fuel rebate. We could keep expanding all those ones coming off. The question is: where do we keep coming back and seeking clarification and where do we make a judgment? It is quite likely that at the end of day we will have a list of things which you would like us to look at in relation to the assumptions that we made.

In terms of the savings ratios, as Ann said earlier, what we did was to walk through and have a look at the options available to us. Do we try to use HES with all the criticisms that have been levelled by Treasury and others about using household income versus expenditure? We decided instead to demonstrate the importance of this particular issue by adopting a procedure used by Treasury in their draft white paper. It was not subsequently used by them, but it would throw adequate light on the particular issue.

In terms of the CPIs, the approach we took to measuring household price index, there are two ways you can go about it. Senator Conroy has raised the issue of why we did not do household specific price indexes. We chose to go with household group specific price indexes rather than household quintile specific price indexes. The issues there relate to sample size. If you look at the table in the material, you will see that at times for the aggregate group we ourselves had difficulty with the sample size by itself without dividing it up into quintiles, let alone think about it in terms of deciles.

That is why we chose for modelling and sample reliability reasons to go to the household groups own price index. The evidence would indicate—and I have no problem in flagging this because of the work that I have done for ACOSS also—that the price index will vary

across those quintiles, but it was not us hiding anything; it was just really a sample size issue and adopting a common methodology across the different approaches that we took.

I am raising here issues of contention in modelling that need to be flagged; some have been flagged already. The other issue is in relation to the price index and how it is estimated. We talked about it being a group specific price index. There is also the issue about using the household expenditure survey for that purpose when the ABS, in competing its own price index, does a fair amount of reweighting to take account of problems relating to tobacco and alcohol, for example, and underreporting in HES. We have adopted a similar methodology to that of Treasury. That means that at the end of the day our price index is similar to that which would be in the 13th CPI, which is the new one which has subsequently come out.

I will touch on the issue of tobacco and housing. In modelling these various scenarios, you always want to prepare—and we have gone to great lengths—an independent report. I am a little concerned about the targeting of some of the assumptions, but we have tried to give balance. In a sense, that is why option 7 is there—to make it informative. It is interesting to see tobacco and new house prices in, but it is also interesting to take them out. So we have presented a range of results which have tobacco and new house prices in and out so that we can provide a more informed picture.

The reason for tobacco out is related to history. It is pertinent to think about history in this particular case. In the case of Fightback, tobacco was out. In the case of John Dawkins' budget of 1993—some of us who have long memories remember that there was a series of increases to tobacco excise during the period—on 17 August 1993 it went up by three per cent, in February 1994 by five per cent, in February 1995 by five per cent and in May 1995 by 10 per cent. Did the government ever consider compensating people for those excise increases? No. Why not? The excise was meant as a tax that people should pay and that should alter their behaviour. If that is the case on both sides—we have both Labor and the ANTS document taking tobacco out—why are we now wanting to put tobacco in? Yes, it is an issue of interest, but should it be out? We opted to present both options of seeing it in and out because history is on the side of taking it out.

What troubled us in relation to new house prices was that we were only telling half the story. Half the story was: this is what can potentially happen to new house prices. But we were not telling the other side of the story about the First Home Owners Scheme and the wealth effect of increases in house prices. So again we erred on the side of doing what we as independent researchers should do to balance off and present to the committee. That is part of what I call being a responsible, independent consultant—to provide additional information that provides light, which is important, and complements what is already there. That is why we presented tobacco and house prices both in and out.

A lot of attention has been given to the credibility of the household expenditure survey. We have gone to fairly significant lengths to try to distance ourselves from those kinds of criticisms. That is why we have created the hypothetical savings ratios and why we have gone to group CPIs rather than across the different quintiles or deciles.

So that is background to why we did what we did—why we took the 2001-02, why we gave the standard and the B case in a number of the options and why we used the price index approach that we took.

I will defer to Martin and then come back to some of the price and distributional effects. Martin can talk about the STINMOD side of it.

Mr Robinson —I will give a very brief overview of how the ANTS package or the cameos were modelled in STINMOD. The version of STINMOD that we were using is an amended version—STINMOD release 97A. We use this version because it is based on the household expenditure survey for 1993-94. This version of STINMOD was given to Treasury, in particular to the Retirement Income Modelling Unit, which made modifications to the model. The main modification that was made was the adjustment of household incomes to bring the 1993-94 HES up to the period of analysis, in particular to 2000-01 and 2001-02. The 2000-01 file was used for modelling some of the options such as 5, 5B and 7 where we had to compensate for food out through increased taxes and we needed to do some estimates of the budgetary impact. The 2001-02 was for Professor Warren to estimate the price effects in that medium term. Ageing of the household weights was also made by RIM. Again it took the household weights to be representative of those two particular years.

The other major change to the model made by RIM was to implement the ANTS policy changes that required structural changes to the model. Some of the changes proposed under the ANTS system, such as increases in payment rates, are just simple parameter changes in the model; other changes, such as the introduction or replacement of family tax benefit with the family allowance and so forth requires structural change to the code of the model. These changes were made by RIM and subsequently checked thoroughly by NATSEM, as was the adjustment of incomes and household weights.

We also made some changes to the hypothetical component of the model. The hypothetical component of the model allows us to model the cameos in such a way that we can specify hypothetical families and at different levels of private income to estimate their entitlements to government cash transfers and calculate the income tax liability, as was required for the cameo cases. As I said before, the forecast household weights for 2001-02 and the incomes for each household were supplied to Professor Warren so that he could estimate the indirect price effects.

RIM also provided NATSEM with a list of their projected tax and government cash transfer payment rates on the basis of forecast CPI and AWE. Since the modelling of the ANTS package by Treasury, the Australian Bureau of Statistics has released more recent average weekly earnings and CPI figures which have subsequently been incorporated into their projected payment rates yielding some small discrepancies between the payment rates used in the ANTS document and the payment rates used in the modelling of cameos by NATSEM. I will leave it there.

Prof. Warren —I will go through some of the price effect issues and then give an overview of the various results we have.

CHAIR —How much longer do you think you will take?

Prof. Warren —About 15 to 20 minutes, all up, I would hope.

CHAIR —The purpose of this inquiry is for the committee to inquire into the work you have done. We have a specified amount of time. That means that the ability of the committee to examine you on your report is reduced. That is why I said at the beginning that I hoped we could have a brief coverage. We have now been going for about three-quarters of an hour and we are going to take another 20 minutes. Let us proceed and see if we can fit in all the needs of the committee.

Prof. Warren —I will truncate my presentation. Going through the price effects, you will see from the information in the report that, under the various options, we get close to but not exactly the same outcome as the one in the ANTS document, and this is with option 3, which is the government's package as at 2001-02. We have a price estimate of two per cent with tobacco and new homes out and 2.7 with them in. If you then go to option 4, which is with the stamp duties and wholesale sales tax 70 per cent shifted forward to consumers, then that jumps by about another 0.9 per cent to 3.6 per cent with tobacco and housing in. I will come back in a minute to how that impacts on different groups.

With option 5, which is taking food out, it is not surprising that that has a fairly significant impact on the consumer price index, taking it down from 2.7, with tobacco and housing in, to 1.6 per cent, and I will talk about who the beneficiaries are in that case. The alternative package is option 6 which—instead of taking food out and using income tax changes to fund that—simply raises the GST rate to 12 per cent. The price effect of that was a little bit higher than the outcome around 2.9 per cent up from the 2.7 per cent.

It is probably worth looking at the general patterns. We found that prices do not appear to be too different depending on whether or not couples had children. The age of the child appears important but only marginally so. The age of the heads seems to be quite important. What you find, particularly in relation to option 3, which is the government's package, is that the groups that are particularly impacted upon are the age pensioners, the self-funded retirees and those on disability support pension.

Looking at the price impact if food comes out of the GST base—which is option 5 versus option 3—you find that sole parents gain more but that the single income couples gain more than they do with a reduction of 1.5 to 1.6 per cent in their price index. For those on pensions, it becomes quite significant. This should come as no surprise because what we are looking at here is the distribution of food and the importance of food within people's budgets—with pensioners, the singles 1.9 per cent reduction going to a 10 per cent GST with food out and 2.1 per cent for married couple pensioners. Self-funded retirees, again above average, gain 1.4 and disability pensioners around 1.9 per cent. It is quite clear the group that such a change is meant to target are beneficiaries.

The beneficiaries of having tobacco and housing in or the people who are impacted most by having it out are those who seem to be under stress—people on disability pensions, sole parents and the unemployed, typically low income individuals. That is obviously an important point also.

When you look at option 4, which is with 70 per cent wholesale sales tax and stamp duties with a CPI effective 3.6, it is not surprising that a large number of people are losers from that scenario, obviously considerably more than option 3. The key losers in option 3 are single age pensioners, the self-funded retirees and those on disability support. They are general observations that we make, given the time constraint which we confront. Maybe it is worth leaving that and going to the 3Bs, 5Bs and 7s because the price effects really do mirror the distributional impact which arises from those different scenarios.

Prof. Harding —I will briefly show you four overheads.

Overhead transparencies were then shown —

Prof. Harding —This is the impact of three of the options that we modelled, in this case on single age pensioner households. In each of these cases, tobacco and housing prices are excluded from the price effect. These three packages are all revenue neutral, so they are directly comparable. Option 3B is the government's package but with household specific price effects and saving; so the two key things that the modelling has been criticised for are corrected in this version. Option 5B is the food out income tax up package and option 7 is the food in compensation up package, which is the six per cent pension increase.

We did not get any losers on average in any of these three cameos. There was one apparent exception of self-funded retiree couples in option 7, but that was an error on our part and I have distributed the new tables to you today. One of the problems is that for some of the cameo groups the apparent gains are relatively low—one per cent or half a per cent. You have to recall that, when we produce the average price effect for a particular type of household and income grouping, there would be some people who face a price effect that is above that average and some who would face a price effect that is below in order to get the average. Therefore, where there are very marginal gains, that is a cause for concern.

In this transparency, you can see with option 3B, which is the government's package, that there are marginal gains for single age pensioners. The food out option helps age pensioners greatly. Option 7, which is the higher six per cent increase in age pension, delivers more assistance right at the bottom end.

I will show you what it looks like for three other types of families. The single person one essentially gives you a feel for the key flavour of all of the results where you compare food out income taxes up to food in compensation up. The essential thing is that at the bottom end low income people fair better under option 7, which is compensation up, than they do under food out. There are higher gains at the top end under option 7 than under food out, but still the gains are not as high as they are under the government's original package, which is 3B. There is a zone in the middle where it appears that food out delivers higher benefits than option 7 which is the compensation up food in package.

There is one other feature as well which is that sometimes at very high income levels food out can be seen to deliver higher benefits. So what you see here above about $100,000, in this case for a single income couple with two children aged five to 12 is that option 7 delivers lower gains than does food out at this very top end, but the three effects that I spoke about are also still apparent. Option 7 delivers more assistance at the very bottom end than

does food out. In the middle, food out is often better. In the upper middle, option 7 usually delivers higher gains than does the food out option.

I will make three concluding comments, one being about the adequacy of the compensation package. The worst case package that we were asked to model was option 4, which had much higher CPI increases and where we assumed that pensions would still only increase by 3.4 per cent. In that case, there were a reasonable number of cameo households who were expected to make losses. In most cases, those people had little or no private income, but there is another zone for concern which is two-income couples who are in the middle range of $30,000 to $50,000. Some would argue that the assumptions underlying option 4 are too strong, that there is a case for excluding housing and tobacco and that it is not clear that pensions would still be increased by only 3.4 per cent if there was a higher than expected CPI result. Nonetheless, even if you look at options 3B, 5B and 7, then there are gains of less than one per cent for some categories of people, particularly age and disability support pensioners, some dual income couples and some self-funded retirees.

I think this raises some concern because the average result disguises the fact that, within each of these cameo household types, some will do better than the average and some will do worse than the average as a result of their different expenditure patterns. I think overall there are some question marks about the adequacy of the compensation package. On the food in food out issue, it is one of those usual extremely difficult social policy trade-offs. Delivering high compensation under something like option 7 delivers more assistance right at the bottom end, but it looks as though it delivers less assistance at the middle than does food out, and it delivers more gains at the middle to higher income levels.

Just a final comment, it is not possible for us to model the number of winners and losers—to say how many winners and losers there are from the package—because of the hypothetical way in which we have modelled these results. We are looking at typical types of families rather than how many of those families there are in Australia.

Prof. Warren —One of the things that became clear to us is that there are some parameters that are important and some that are not important. We feel the issue of tobacco and housing is one that has to be considered and taken into account, because it becomes an issue determining the sensitivity of the results, particularly on that line between losers—option 3 versus option 3B. It is important to address that issue, and we have sought to do that.

We saw the savings issue as important but not critically important. We saw that the own price issue is particularly important. It very much influenced the results that we got. The phasing of the indirect tax, whether you take 2001-02 or July 2000, is not important because it is only a matter of ratcheting the whole thing up. So they are the critical issues; the ones that we saw as important.

CHAIR —Thank you very much. If there are no questions at this stage, we will take a break. If there are any questions on just the presentation before we go to the broad questions, I am prepared to allow a couple. If not, thank you very much for that presentation.

Proceedings suspended from 10.15 a.m. to 10.33 a.m.

CHAIR —The time set down in the program for this part of the hearing is 1[half ] hours. I intend to divide the questioning as follows: half an hour to the coalition, half an hour to the Labor Party and half an hour between the Democrats and Senator Harradine. If at the end of that we still have demands for questioning, I am prepared to run over a short time but not a very long time. I hope we do not have excessive demands for questioning so that we do not have to run over at all. I ask that the questions be reasonably succinct, and I would appreciate if the answers can be to the point.

Senator FERGUSON —I want to go over a few of the key points you have made, particularly in relation to the options as they have been presented. You were asked to do a variety of things. Firstly, option 1 was to do modelling in relation to the government's ANTS package. In the initial part of your report you say:

The estimates of the distributional impact of the package are extremely close to the Treasury's in almost all cases.

Is it fair to say that as far as option 1 is concerned you confirm the Treasury's estimates as they have been laid out in the ANTS document?

Prof. Harding —I think on the whole that is the case. Option 1 is not the first option that the committee asked us to model in the sense that they wanted tobacco and housing included in the first option they asked us to model, but we tried to do an independent audit of the way Treasury had done it with tobacco and housing excluded, which was our option 1. We felt that a few minor errors had been made in the modelling, but no doubt we probably suffer from the same thing. But on the whole the numbers were reasonably close. We had a slightly different price effect, which Neil might talk about. We had a 2.0 per cent price effect versus the Treasury's 1.9 per cent. We had very slightly different social security payment rates because there is now more recent estimates about what AWE and CPI movements will be and have been over the recent past. In a couple of areas there were some modelling differences. But on the whole the numbers were a reasonable match.

Senator FERGUSON —That 0.1 per cent could come about because of the different figures you have used because of more recent estimates that are available?

Prof. Warren —Not necessarily. It could just arise from different modelling methodology and so on at the margin. That kind of size is just an error that you are going to have.

Senator FERGUSON —There is a margin of error anyway, I presume, and 0.1 per cent does not seem to be very much. You have explained in detail why you do not think tobacco and new home prices should be included. You explained that in detail. Option 2 is just an inclusion of those on top of the variations of option 1. You have an option 3 and an option 3B, which does exclude tobacco and housing. I think you say on page 25 of your report that, for reasons you noted earlier, the methodology adopted in option 3B where tobacco and new home prices are excluded is probably the most appropriate. Can you expand a little on why you think option 3B is the most appropriate?

Prof. Warren —The two issues are in relation to tobacco and new homes. In relation to tobacco, it is really a question about what it is you are trying to achieve with tobacco taxation. If you look at what the government is doing in relation to what is causing the

tobacco price to increase, it is moving from a weight based excise to a per stick excise because of concerns about the industry developing technology to make extremely light cigarettes—for the same cigarette to have less tobacco—and therefore giving one company a price advantage over another. A per stick system is where a lot of other countries have gone, so there is rationale for that.

Also the issue about increasing the excise is one, for example, that the National Heart Foundation and other organisations have argued for for a long time and made various submissions. I think they made submissions to Brian Gibson's committee on moving to per stick and raising the level of excise. It is all an issue about the negative externalities associated with tobacco smoking.

It would be contradictory of that move to then say that, having moved to a per stick system and moving in the direction of the Heart Foundation and other organisations concerned with the impact of tobacco smoking, people should be compensated for that increase. That is not logical. If you told the Heart Foundation that that is what you are subsequently going to do, having accepted their case—a case that they put for a long time—it would just be anathema to the whole thing. That is in relation to tobacco. In the Dawkins changes back in 1993, stretching through to 1995, there was never really any consideration that people should be compensated for what were at that stage significant increases in tobacco taxation.

In relation to new homes, new homes is a complicated one because, as we move between the 12th series and the 13th series, CPI housing is treated quite differently between those two series. That in itself is an appendix. That in itself is a complicated issue because it is going to have implications for the price effects of this whole package anyway. There is an issue there in that if house prices go up those people who are currently home owners do gain a certain wealth effect as a consequence of that. The question then is whether you compensate them again for what those home owners already have. So it is an issue about compensating them twice. Obviously for those people who are not home owners there is an issue that they have missed out. The government has argued that it is introducing a first home owners scheme, and we were unable to model that. But, to do this properly, you would have to take account of the benefits for current owners in the first home owners scheme.

It is without doubt an issue for one group—that is, those people who will never own a home, the renters. That is, if you like, the area in which we do have some concern. You can start to address those kinds of issues by the way you design your compensation package with rent assistance and so on, but there clearly is an issue. That is why standards 3 and 3B are there. There clearly is an issue in relation to renters, and we will not deny that. While the majority of people are owner occupiers and the first home owners scheme exists, if you assume that everybody ultimately becomes a home owner, which is probably an unreasonable assumption, then we have erred one way. Putting it all in is erring we feel disproportionately the other way. So that is why.

Senator FERGUSON —Bearing in mind that you say that option 3B, where you exclude tobacco and new homes, is probably the most appropriate, is it not a fact that under the 3B scenario every group modelled will be better off under the government's tax plan?

Prof. Warren —It does result in no individuals confronting price effects which are greater than 3.4 per cent, which is the compensation in that long-term period. That is an outcome from that. As Ann said in her conclusion, there are obviously always issues about the price indexes across different income groups. So in a sense, yes, but there is a question of the adequacy of the compensation—that is, the adequacy issue. You can sail close to the wind. How close are you sailing? You are on the positive side but you are extremely close. What is clear from what we have done is that you can change your assumptions and you can push yourself over that line.

Senator FERGUSON —Rather than saying that everybody will be better off, would it be fair to say that, on average, at least under option 3B no-one will be worse off?

Prof. Warren —With the methodology that has been adopted and the assumptions made, yes.

Senator FERGUSON —On Wednesday in the Sydney Morning Herald Paul Cleary wrote an interesting article. He might have thought it was interesting because he must have got all of this information from somewhere. He makes a couple of observations which I want you to comment on. He said:

The most comprehensive study on the Government's GST—

he is talking about the most comprehensive study—

has found that up to 31 income groups will be made worse off by up to $7 a week, including pensioners, self-funded retirees and low- to middle-income families with children.

A summary of the key results of the study, obtained by the Herald , demolishes once and for all the Government's pivotal claim that no-one will be made worse off by tax reform.

If he had read the whole of this report, which includes option 3B, is that a fair assessment of your report to this committee?

Prof. Harding —I think what we have shown quite clearly is that the expected gains or losses depend critically upon the assumptions you make. The dissaving issue and the extent of apparent dissaving did not appear to have such a big impact on whether people were winners or losers, but the estimated increase in the prices facing them turned out to be an absolutely critical issue. Because there is some uncertainty attached to the estimation of the price increases facing each of the cameo household groups and because we are dealing with imperfect data in an imperfect world, this goes to Neil's point about sailing close to the wind—that is, some of the modelling assumptions we were asked to make tip people over the line from being apparent winners to apparent losers.

What we would emphasise is that you should not treat any of these numbers as other than broadly indicative. I do not think you can say, for example, that a household of this particular type will be better off by exactly $2.10 a week, which is what could be implied from a first reading of the tables. The science of the modelling is not accurate down to that sort of detail. What you are looking for is where there appears to be possible potential problem areas, in particular with regard to the adequacy of the compensation.

Senator FERGUSON —I have to make a confession that I have not read all of the second half of your report which has hundreds of pages of tables, but in any of those tables under the headings do you use the word `profiteering' at all?

Prof. Harding —No.

Senator FERGUSON —In this same article there is a set of tables which has `Source: NATSEM' on the bottom. The top of it says `Option 4. Option 3. Plus high inflation from profiteering'. It would appear as though this is one of your tables that has been put in the paper and yet you say you have never used that word.

Prof. Harding —No, we did not use the word `profiteering'.

Prof. Warren —Not in the tables. It is used in reference to those kinds of issues in the text in relation to option 4.

Senator FERGUSON —But it is not used as a reproduction of a table. This looks like a reproduction of a table headed `Howard's GST losers' which says `Source: NATSEM' on the bottom, as though it is a replicated table from your report. I guess the only thing I can do is show it to you and you can tell me whether or not it is a replicated table. If it is not, it is a misrepresentation of your work.

Prof. Harding —He was leaked the summary tables, but not by any of us on this side of the table. He obviously just extracted selected figures from that in order to produce that table in the Herald .

Senator FERGUSON —So it was sort of selected quoting.

Prof. Harding —Yes.

Senator FERGUSON —You also mention the work on dissaving where you say that because of the tight time frame it is not possible to make a comprehensive analysis of dissaving rates. So you just made an assumption of your own in relation to dissavings because you felt that that was the only way you could give a proper analysis.

Prof. Harding —The time that we had to undertake this study was extraordinarily short for a study of such magnitude. To look at dissaving you would probably want to do something like the Treasury had done when it prepared the earlier draft white paper estimates, and has done subsequently, where they adjust all the HES data in detail so that the aggregate matches the sorts of savings rates that are apparent in the national accounts. At the moment if you look at the average household in the household expenditure survey, they are dissaving. That contrasts quite sharply with the savings shown in the national accounts. There is effectively a discrepancy between the micro-data and the macro-data, and trying to correct for that will take a considerable period of time. In that event, we assumed that Treasury had undertaken all of that research when preparing the draft white paper and just borrowed their assumptions but increased their dollar values by the CPI to forecast 2000 dollar values.

Senator FERGUSON —I want to move to the 70 per cent scenario of pass-through. You were asked to model that as an assumption by the committee; it is not a choice of yours. It does seem as though in the body of your report that you do not believe that a 70 per cent scenario is a realistic one and that it should not be viewed as a standard case. Do you have any opinions on pass-through that you would care to share with the committee?

Prof. Warren —The issue of shifting is a highly controversial one. Various attempts have been made to try to see what happens when other countries have introduced reforms like GSTs. The difficulty is that every case is unique. The New Zealand case was unique. This case in Australia is unique. It is unique for a number of reasons. One of the most significant I suppose is the ACCC's brief in terms of companies and the potential they might have for not passing on price cuts as a result of indirect tax cuts. If you see that as a potential check and balance, which we did, obviously that is going to become an issue. Large corporations which will be driving the price setting will be extremely mindful because the ACCC would obviously be viewing the large players rather than the small players as the ones to prove certain things in terms of rules.

So our view was that it is important. There is an issue of the transition, and the 70 per cent is an issue you cannot overlook. But the period before that washes out is extremely difficult to know. It is really how effective the education program is. It is how effective the ACCC is. It is all of those kinds of issues. It is something you need to look at and something you need to be mindful of. It is another issue that does influence the adequacy of compensation. Where is the buffer if this does actually occur? Those kinds of issues are quite important. So it is, if you like, another one of those issues that adds to our concern about the adequacy. If you do have that, then it does push people over the line. If you do have these different assumptions you can create the losers that were said not to be there. It does need to be heeded in that case.

ACTING CHAIR (Senator Ferguson) —I will hand over to Senator Gibson. He has a couple of questions in relation to food and other things.

Senator GIBSON —I would like to explore the question of families—those with children versus those without—because there has been concern on the committee through our inquiries about how adequately families are compensated. I take you to page 26 of your report and quote the main paragraph:

Based on the estimates of price changes shown in Table 5, it does not seem possible to conclude that families with children generally face higher potential price increases than families without children.

It then goes on to say:

This is higher than the estimated price increases facing any of the other single income couples with children. Similarly, the estimated price increase from Option 3 for two income couples without children is 2.6 per cent. This is higher than the estimated impact for a two income couple with . . . children . . .

Then on the next page, on page 27, in the second paragraph you say:

Suppose for the moment we take the `worst-case' dissaving assumption . . .

And in the next paragraph:

Looking . . . at families with no private income, who are fully dependent on the social security system

—

the report actually says `Table 11', but I think you mean `Table 8'—

. . . indicates that most social security dependent households seem likely to make small gains from the tax reform package under the Option 3 assumptions.

And, again, it goes on to imply that those with children do better than those without children. On the next page, page 28, the final paragraph says:

What about families with children a little further up the income spectrum? Again, the figures presented in Table 9 provide a `worst-case' estimate, in the sense that tobacco and housing prices are included . . .

and the dissavings estimate is included. Would you take us through what is actually going on there?

Prof. Harding —Since we wrote the report, at such great speed, we have been having a further look at the issue of how families with children fare. The price effects do move around quite a bit, and I think not only is that a reflection of the fact that people's purchasing patterns do differ but it also comes back to the issue about the variances introduced because of the small sample size of the household expenditure survey and the small number of people within each of these groups. But there does seem to be a trend for couples with older children to face higher price increases than couples with younger children. So it looks as though the older children issue is going to be an important one. There were some concerns when you looked at the distributional outcomes about the narrowness of gain for some dual income couples in the middle income ranges, including those with older children.

Senator GIBSON —But the paragraphs I just quoted from the report are basically indicating that families with children are not at a disadvantage relative to families without children. Is that not essentially true?

Prof. Harding —There is particularly generous compensation for single income families with young children in the proposed family package reforms. So, on the whole, single income families with children did better than two-income families with children, and families with children generally did better than people without children.

Senator GIBSON —On to another topic, options 5 and 6 looked at the question of food, apart from restaurants and takeaways, being excluded. The questions of compliance costs for the businesses having to look after their tax returns and their administration of their accounts and running their businesses and, secondly, the costs for the government of running a more complicated system are not in here, are they? Would you care to comment on that?

Prof. Warren —That is a legitimate issue. And, no, there is no account at all of what could be the counteracting force here in terms of the impact on particular types of

businesses—small business and so on—and how that might affect their prices. We have just assumed that you can replace one instantly with another and that you do not get any distortions which might arise from an increase in administrative costs for government and compliance costs. And, in a sense, to the extent that those costs will exist, in our revenue neutral scenario you do need to raise those prices that people confront in the food out case to reflect that. Particularly in relation to food the costs will not go down that far because there simply will be administrative costs that to some extent will need to be passed forward and recovered, one would presume, which would offset the extent to which the food out case price might go down.

Senator GIBSON —So in those cases would you expect the CPI to be actually higher to take that into account?

Prof. Warren —To some extent; how far is a very difficult question. You would expect that it would move in the upward direction as a consequence because of that added cost, but how far I cannot tell you.

Senator MURRAY —Professor Harding, I want to talk to you first about housing. The figures I have for current housing costs are these. The bottom 20 per cent spend 72 per cent of their housing costs in the area of rent, insurance, repairs and body corporate rates, whereas the top 20 per cent spend 37 per cent in that area. The bottom 20 per cent spend 8.6 per cent of their housing costs on mortgage interest repayments, whereas the top 20 per cent spend 39 per cent of their housing costs on that. And, if you look at rent, insurance, repairs and so on as a percentage of total expenditure, for the bottom 20 per cent it is 10.5 per cent and for the top 20 per cent it is three per cent. Really what I am saying to you is: if we exclude housing, which is one of the assumptions being examined, we are excluding a price effect for low income earners which is in favour of a first home buyers benefit for higher income earners. Are you with me? It seems to me that if we go down that path we are excluding a major impact on low income earners.

Prof. Harding —I think when you are looking at these sorts of cross-sectional analyses, which are essentially a snapshot of what the population looks like at a particular point in time, it is more difficult to think about the life cycle forces that are occurring as well. So, in part, the bottom 20 per cent are more likely to be renting, but they are also more likely to be in different stages in their lives and not have yet got to that peak first home buyer age range of 25 to 35. I should add, though, that the exclusion of housing is by far a lesser component of the estimated 0.7 per cent change in the CPI. Neil may want to talk more about this, but of the 0.7 per cent addition to our estimated CPI, which was made by including tobacco and housing, 0.5 was tobacco and 0.2 was housing.

Senator MURRAY —Let me explore that a little more. The ABS income distribution shows that 80 per cent of home buyers are in the top two income quintiles, and only six per cent of home buyers are in the bottom two quintiles. When we are looking at compensation for the bottom two quintiles, the First Home Buyers Scheme really is quite irrelevant in how they actually live and spend their money, especially when housing makes up 19 per cent of the spending by low income earners. So, for me, that just reinforces the point that we should be pressing home: the need to examine housing and the compensation rates for those bottom

two quintiles. I think the response you are giving implies that it is less of a problem than it is, and I think it is much greater than you express it.

Prof. Warren —I would like to clarify your understanding of the word `housing', because in the study we are excluding new homes only; rents are in it. What we are doing is addressing the house price issue, not the rent issue. So rents and what happens to the price of rents is in the analysis. It is not housing per se; it is new houses, new homes. That is what has dropped out and that is why it is so insignificant. So, if you think about that and everything that you have said, that skews it. We are pulling off things that affect the people at the top, but the rent issue is still addressed. It is only the new homes issue that we have dropped out because that is the issue that is important. Rent is still there.

Senator MURRAY —Okay. So wherever you have said `exclude housing' you just mean new homes?

Prof. Warren —New homes.

Senator MURRAY —All right. According to the tax statistics, there are 650,000 taxpayers earning less than $10,000 a year, of which 64 per cent receive no government income assistance. There are also 1.6 million non-taxable people, of whom 75 per cent receive no government income assistance. Would you agree that, in those groups of people, we are looking at a large group of working poor who are largely outside the social security safety net?

Professor Harding, given the remarks you made earlier about being very careful about the assumptions and imperfect data, aren't we obliged as a parliament and as a committee to actually err on the side of compensating far more because these people are not going to be picked up—it is a real risk—in this package?

Prof. Harding —I think there is a problem for people who are not within the social security net but are on low incomes. There is a bit of a movement up in the income tax threshold under the proposed tax reform packages, and there is the drop in the marginal tax rate applicable between the $6,000 and $20,000 range by three cents in the dollar. Nonetheless, if you do your sums, I think for some people the gains are very marginal or non-existent if they have got very low incomes, if they are not getting money for their children and they are not being picked up by any of the other compensation measures.

Senator MURRAY —From a professional point of view you quite rightly said, `Look, we can't put numbers of people to these assumptions we're making.' But if you listen to those numbers I have given you—that is, 650,000 taxpayers earning less than $10,000 a year and 1.6 million others—that is a massive number of Australians. I think the point I am seeking for you to emphasise is in fact that any measure of adjustment for those people needs to err on the side of the generous because of the difficulties of understanding the impact on them and of reaching those people.

Prof. Warren —I would like to make a comment there. There is an issue here about who they are. Tax statistics report information on individuals; we are reporting information on income units. Therefore, a lot of these people are simply going to be in some sort of

dependency relationship within an income unit with, maybe, independent income from investment sources that puts them below that $5,400 threshold. So what you are saying is reasonable if these people are trying to live on it, and I would be very surprised if they are trying to live while not paying tax. I would presume that most of these people either are sitting within our income units and that they are the single income families or are, in our case, the 67-33 types of households where the person you are talking about earns 33 per cent of the household income. It is an issue about tax stats being persons versus our work being income units.

Senator MURRAY —But the fact is that, with the statistics and the modelling and the assumptions available to you, you cannot qualify or quantify how those people fit in this package. It is very difficult to estimate. That is right, isn't it?

Prof. Warren —We could, if we like, pull out individuals within those income units and look at them, but that is not particularly meaningful for us because we are looking at the impact of the tax system on those family structures. You can imagine our task of trying to attribute consumption tax to the individuals within the income unit. We would have to try to allocate some of the expenditure right down to the children and to each of the spouses, and all distribution analysis will either move to income unit or to households because that is the consumption unit, which is the unit you need to focus on. So, while it is an issue, what is important in these studies is the consumption unit—in that sense it is the income unit or the household.

We are mindful of those issues when we look at the single income families and the 33-67 income split, and that is where those people come into play, but I am not sure that we could even begin to un-tax single individuals, given that we are working with a household expenses survey and looking at indirect taxes and their impact on those various groups.

Senator MURRAY —Which simply is a longwinded way of saying we must be very careful of assuming no-one will be worse off!

Prof. Warren —But then the issue is: how can you say that some person within a household is more worse off than somebody else in the household? To do that you have to attribute certain benefits and burdens to the individuals within each of those households, and that is problematic. But the point about being careful and taking care is taken on board and accepted.

Senator MURRAY —Professor Harding, new migrants in their first two years, retrenched workers aged 55 to 60 with superannuation assets over $125,000, people on waiting lists for Newstart while working out leave periods or breaches of activity tests all fall outside the social security net. The social security budget in 1996 alone was cut by $1.6 billion dollars, which represents full payments to around 200,000 people. My reading of your option 7 is that those kinds of people cannot be adequately compensated because they are not picked up or identified. The whole compensation issue has to relate to how many people fall within the safety net and how many fall without. What I am driving at continually with this line of questions is that the data available to you really leaves you with a large hole of analysis—it is not your fault, but that is the way it is.

Prof. Harding —I think in part that is true. When we did the analysis we excluded, for example, multiple income unit households which have an instance of, say, an unemployed teenage son in his twenties who is not receiving social security and the parents are not in the social security system either but are just earning wages. It is hard to see how the son will be compensated under that sort of scenario. Although Neil's point is obviously valid about it depending on what sorts of families and households the people who might potentially fall through the cracks are living in and whether other people in those households are in some way picked up by the compensation measures, I accept your point that there is still a problem with people who potentially may fall through those cracks.

Senator MURRAY —But we do know—don't we?—with absolute certainty that these millions of people which I have categorised in various ways would all benefit from food that is GST free. That is absolutely certain because everybody eats.

Prof. Warren —But what does that mean?

Senator MURRAY —Do you mean some people don't eat?

Prof. Warren —No, but what does it mean? It is like saying that if I take out motor cars everyone who drives is going to benefit. That must be leading to some other observation about the superiority of taking food out.

Senator MURRAY —What it means is this: a compensation package is directed towards giving people money and, if they are not in the system to receive that money, they do not get the money. Whereas everybody eats—if you make food GST free, I would have thought the point self-evident.

Prof. Warren —The issue of the adequacy of compensation for those people who are in neither the tax system nor the welfare system has always been a problem when you move with the kind of change that is being considered with this package. That is why in other cases there are moves to things like GST credits and so on that are refundable, because you can get those benefits to people who are currently not in the system. That is why, in a sense, what the government is doing with its package is working with what it has already got. What we would like to say about option 7 is that it is rough and ready and was never intended to be the last word. It is, if you like, trying to indicate a point, but you can still have—

Senator CONROY —An ambit claim?

Senator MURRAY —I had better move on, if I may. The tax scale, Professor Harding, in option 5B raises $5 billion to fund food as GST free in 2001-02. It has a 33 per cent rate from $20,000 to $50,000. I might remind everybody that the average family income at the moment is about $47,000. It has a 43 per cent rate from $50,000 to $75,000. Is it possible to raise that money with a more progressive tax scale that pares back more of the tax cuts off high income earners than you have shown and leaves the 30 per cent tax rate on people below $35,000 and still fund the tax exemption? For example, a scale with a 30 per cent rate from $20,000 to $35,000, 36 per cent from $35,000 to $47,000, 43 per cent from $47,000 to $60,000 and 49 per cent above $60,000. Those are a lot of figures, but what I am saying to

you is that you can work the income tax scales in a different way to which you have to achieve that return essentially by hitting the richer harder.

Prof. Harding —Yes, you definitely can. That package that you outlined would raise about $5 billion in revenue, so it would have much the same impact as the tax scales that we proposed under 5B. I guess what you start coming into is all the dilemmas that have been raised in the income tax field on these very issues—the political problems created by the fact that people at average weekly earnings are seen to be facing very high effective marginal tax rates with a possible disincentive to work.

Senator CONROY —Bracket creep.

Prof. Harding —Yes, bracket creep, and the ability of higher income earners, if marginal tax rates are extremely high, to arrange their affairs so they do not face those marginal tax rates anyway. The other problem of course is that, because so many people are concentrated towards the lower to middle end of the taxable income spectrum, it costs an awful lot of money to lower marginal tax rates in that $20,000 to $50,000 range, but you get very little money from imposing higher marginal tax rates at the top end.

Senator MURRAY —I have a lot more for you, Professor Warren, but I suspect I will have to put those on notice. Professor Harding, the distributions provided do not change expenditure patterns for income levels, do they?

Prof. Harding —You mean our estimated price effects within each particular household type do not vary in line with income?

Senator MURRAY —That is right.

Prof. Harding — That is correct.

Senator MURRAY —And that is a real weakness because they actually do vary, don't they? The real world is that patterns do vary per household.

Prof. Harding —We were obviously very concerned about this issue, but again when we looked at the cameo specific results the evidence was not clear cut. For example, if you looked at the result for a single age pensioner, we estimated they would face a 3.7 per cent increase on average in the prices facing them, so that sounded quite high. A disability support pensioner couple faced a 3.4 per cent increase. That was above the average, but the sole parents were facing below average price increases, yet most sole parents are on lower incomes as well. So there was not an entirely clear-cut relationship between income level and the price effect.

Senator MURRAY —I just want to deal with the issue which has concerned the Democrats a great deal throughout this—that is, food. Everyone agrees that food constitutes a much larger proportion of spending for lower income people. The per head category is different, but this is as a proportion for low income people. It seems to me that your tables would therefore understate the GST impact on low income earners because of that. So

wherever on low incomes we have a marginal figure—say, a gain of 50c or a dollar—if you are understating it the effect is that those people will be losers.

Prof. Harding —I think that is a possibility. There is a possibility that the methodology we have used systematically understates the price effects not only for low income groups but for other sorts of people as well. What we expect is that expenditure patterns will vary greatly within each of these cameos. So while the average result may be a gain or a loss, some people will make below average gains and some will make above average gains.

Senator MURRAY —If you are looking at option 5B, the beneficial effects of zero rating food could be understated for low income earners and overstated for high income earners—on the basis of that answer—and that would reinforce your finding on page 32 that zero rating food would be a progressive measure.

Prof. Harding —Option 5 and 5B did look progressive relative to the original ANTS package, but whether we are understating any possible effect just for low income groups is not clear. For example, it also looked as though exempting food would be of particular benefit to families with older children because they were also more likely to spend money on food. There is not a clear-cut relationship between the income side and the impact of exempting food; it also depends on your family characteristics. Generally speaking, it looked as though the age pensioner couples and families with older children would be particular winners from exempting food.

Senator MURRAY —I will have to make this my last question, although I have a lot more.

ACTING CHAIR — We may be able to come back, Senator Murray. It depends on the chair but I should let Senator Harradine have a go.

Senator MURRAY —Certainly. I have one last point on this food issue. Would you be able to put any dollar amounts on the likely extent of the understating of the impact of food on low income households? If you cannot do it now, would you be able to come back and give us some indicative figures?

Prof. Harding —I am not sure. My feeling is generally that you should in any case regard these figures as only very broadly indicative. Plus or minus half a per cent either way is quite within the bounds of what we know. We start with not particularly good data and we have to model extraordinarily complex tax programs which include behavioural responses by firms and so on—all of these things are very difficult to capture. What you end up coming back to is your point that there needs to be a buffer; you need to err on the side of caution because you should not assume that, just because we said someone was 0.64 per cent, that is the exact answer. We are not able to predict those sorts of effects with that degree of certainty. This is to provide you with an indication of the sorts of people who appear to be relative winners and losers.

Senator MURRAY —So anyone negotiating with the Treasurer should surely, where the gains are marginal in both the Treasurer's and your modelling, be arguing for at least a half

per cent buffer or cushion wherever that marginal area appears simply because of this statistical difficulty you face.

Prof. Warren —I think that is a reasonable observation, and that is why option 7 sits there with six per cent being a scenario that gives you that adequate buffer. That is implied in the calibration of that—the level of margin that needs to be in this package is the sail close to the wind point I made. It is very close. If you look in the ANTS document, even with their assumptions, those people are extremely close in terms of just being very marginal beneficiaries. What this report is really saying by option 7 is that you need a greater buffer, and I do not think we have any problem with that.

Senator HARRADINE —I will be fairly brief. I would like to ask Professor Harding to go to page 4 of the summary document. It says that option 5 is the same as option 3 but with food out and income taxes up, so the drop in the CPI impact is from 2.7 per cent to 1.6 per cent. Is that solely because food is taken out, or is there a slight adjustment for income taxes going up?

Prof. Warren —It is just price effect.

Senator HARRADINE —It is solely price effect?

Prof. Warren —Yes.

Senator HARRADINE —I am intrigued. I am just comparing option 5 with option 6. Option 6 is the same as option 3. Option 5 and option 6—you have a 10 per cent GST rate but, if you shove that up to 12 per cent, you get the increase from 1.6 to 2.9 per cent in the CPI. Could you explain that?

Prof. Harding —I can. Option 6 is more comparable to option 3. The price effect from option 3 was 2.7 per cent, including tobacco and new houses. Under option 6, if you knock food out and you increase the GST rate on everything else to 12 per cent, there is a slightly higher price effect of 2.9 per cent which looks odd but it is just to do with the weights that things have in the construction of the consumer price index.

Senator HARRADINE —Why would you compare it with 3 rather than with 5? Food is in in option 3.

Prof. Harding —But you are raising a lot less money from consumption taxes in option 5 than you are in options 3 and 6. So because you are sucking $5 billion less out of people's spending on goods and services, the price effect is very much lower.

Senator HARRADINE —I must be reading this incorrectly. Option 5 is food out and option 3 is food in.

Prof. Harding —Yes.

Senator HARRADINE —Option 5 is food out and option 6 is food out. Would you not compare option 6 with option 5 when looking at the estimated CPI impact if it is only prices alone?

Prof. Warren —In option 5 with a 10 per cent GST and food out, you lose $5 billion—

Senator HARRADINE —And in option 6 it is 12 per cent if you take food out.

Prof. Warren —The reason Ann was comparing it with options 3 and 6 is that, in a sense, at first glance you might think it would go back up to 2.7 per cent because that is raising the comparable revenue to option 3. If you like, options 3 and 6 raise the same revenue. Option 5 raises $5 billion less from the GST, hence that is what is being reflected in the consumer price index. So option 6 is going back up to raising the same revenue as the government's package but this time leaving food out of the GST base and raising the rate to 12 per cent. We get that price effect, it appears, from the weights that are in the consumer price index for those goods that have gone up as a consequence of moving to 12 per cent.

Senator HARRADINE —Besides those groups mentioned by Senator Murray, self-funded retirees that you mentioned here are those of pensionable age, presumably.

Prof. Harding —That is true, and so there is this possible hole for earlier aged self-funded retirees who fall through the cracks.

Senator HARRADINE —There is a huge number of them. So there is nothing in the compensation at all, and naturally nothing in your figures, which would assist them at all except if food were out?

Prof. Warren —There is the savings bonuses that the government has considered for those people.

Senator HARRADINE —But they are only for aged—

Prof. Warren —This is one of the areas that did concern us. They stand to benefit from the potential different threshold and the marginal tax rate going from 20 to 17 per cent, but it is very marginal for that group. They have an advantage if they have a house, in the sense that that asset can appreciate, but there is not the kind of direct mechanism for assisting that particular group, and that was one group that did concern us. When we looked at the results, self-funded retirees were, again, extremely marginal in terms of their position.

Senator HARRADINE —I have read your document and I have read your reasons for taking tobacco out and I have listened to what you have said today. But in the real world, in Bent Street or Como Crescent or wherever, this is an expenditure; in some cases it is a relatively significant expenditure. If you take out tobacco, why not take out alcohol, for example?

Prof. Warren —I suppose you could say that with those kinds of price increases, the government is trying to create a behavioural effect. That, in a sense, does give people the opportunity to compensate themselves by, if you like, smoking less rather than giving up.

There are all sorts of explanations for this. It is clearly important and needs to be identified, but the question is: do you want to compensate for it and why are you compensating for it, given why you did it in the first case? Governments in the past have shown themselves not willing to compensate for what they have done. State governments have raised these taxes, as has the federal government—Labor and conservative.

Senator HARRADINE —Gambling will be a growing problem and a growing source of revenue for the states, particularly with virtual reality digital TV coming and the powers that be seeking to get control of the Internet gambling and so on. In your professional capacity, have you had a look at the question of the significance of gambling as an expenditure from homes and the likely effect if there is this Internet gambling with virtual reality in the home, the sounds of the casinos and so on?

Prof. Warren —This is one of these evolving issues that comes about with new technology. In a sense, we are working here with the household expenditure survey from 1993-94. So we obviously have a fundamental flaw here in addressing these current issues. Quite clearly this issue is one that has focused the minds of state treasuries because some of the world's largest Internet gamblers are based in Australia—in Alice Springs at Lassiters, et cetera.

How significant that is in terms of the potential is very difficult to know, but it is an issue in the short term about security and whether you get your payback. Once you get your security with your encryption and so on, there is no doubt they will not have the overheads and will have the payback rates, and it could become an issue. But we really have not addressed it; the Industry Commission is looking at these issues.

Senator HARRADINE —I think that is really a question for Treasury this afternoon.

Senator CONROY —The average earnings—that is, male average total weekly earnings—are projected to continue rising faster than inflation. How much of the so-called compensation for pensioners would actually accrue to them anyway and, therefore, should not be regarded as compensation at all?

Prof. Harding —The possible durability of any compensation package is a very important issue. Our suggestion would probably be that, if a specific percentage increase was given to pensioners and allowees as part of a compensation package, it could be a separately identified component of the pensioner allowance, in exactly the same way that the pharmaceutical allowance currently is—it is a separate block. It could stay there as a block and remain intact and be indexed to the CPI over time so its real value is maintained. Although overseas experience has raised concerns here about the possible durability of compensation packages, at least as a separately identified component of the payment it would be less easy to erode.

Senator CONROY —You would be aware of Mr Carmody of Access Economics, who has made a submission to this committee on this issue. Have either of you caught up with that at all?

Prof. Harding —No.

Prof. Warren —I have not seen the detail of that at all.

Senator CONROY —He argues that, because AWOTE increases faster than the CPI, the package actually mathematically erodes within three years. Are you with me?

Prof. Warren —Yes. I am aware of that argument.

Senator CONROY —You have not addressed it at all in terms of, if I can use the phrase, a lock-in mechanism so that the package does not erode? You see, the government is denying that it erodes.

Prof. Warren —The question you are asking us is one about the terms of reference for the project, and it is really about how you design your compensation package. The issue of a 25 per cent AWOTE is a genuine one.

Senator CONROY —Do you accept Mr Carmody's analysis?

Prof. Warren —I completely agree. The question is: do you go for 25 per cent or do you take some other number? Say, picking a number around 26 per cent, maybe you should pick around 26 per cent or 26[half ] per cent. The question is: how do you lock it in? One of the things that Ann was asking is: what is the mechanism? Twenty-five per cent of AWOTE does not lock it in. It has the scope, ultimately, for erosion.

Senator CONROY —Referring to your option 7, just because you put a higher compensation package, all that would do under Mr Carmody's formula is take two or three extra years to erode.

Prof. Harding —Not if you say that the pension remains pegged at 25 per cent of AWE but, in addition, you get this extra payment that at the moment would be worth six per cent of the pension as a separately identified component, and those two keep going into the future.

Senator CONROY —Mr Carmody argued it in probably a slightly different way. He suggested that the way to fix it was to increase the pension tied from 25 per cent of AWOTE to, say, 26[half ] per cent, which is the 1[half ] per cent.

Prof. Warren —We talked about this at one stage. We asked: how far should we go? We have got into trouble with you already this morning with option 7, and now you are asking us all sorts of things about option 7.

Senator CONROY —About option 7—3B, 5B.

Prof. Warren —We looked at 26[half ] per cent, and at one stage we thought of developing something. Then we concluded that that was beyond our brief in government issue.

Senator CONROY —You design a compensation package for the government but you do not bother to make it actually sustainable.

Prof. Warren —We have flagged that compensation and the sustainability of it is an issue. So we have addressed that.

Senator CONROY —I want to confirm that you do agree with Mr Carmody's analysis that, at the moment, under any of the government's proposals, the package automatically erodes; it just dissipates because of the way it is tied at the moment.

Prof. Harding —It is not clear to me how, legislatively, they would do it. We did debate the option—

Senator CONROY —I am just trying to confirm that you agree with Mr Carmody that it is automatic because of the way legislation is set at the moment.

Prof. Harding —It could be.

Senator CONROY —No, not in the future—right now. Because of what is being proposed by ANTS right now, the package automatically erodes. Does it require, as with the New Zealand government, a decision to withdraw the compensation package? It automatically erodes on a simple mathematical formula.

Prof. Warren —Yes.

Prof. Harding —Yes.

Senator CONROY —There are different ways to address it. Mr Carmody suggested one and you have an alternative one—I am not arguing about that. I just wanted to get on the record that you agree that that does happen. On behalf of Senator Harradine, what was your alternative as a lock-in mechanism?

Prof. Warren —We discussed a range of alternatives. The 26[half ] per cent of AWE was one issue. We put up the six per cent as a hypothetical, and when it came to whether we should then say six per cent and 26[half ] per cent—like the government's 25 per cent—to effectively lock it in, we toyed with the idea of targeted GST credits. But these were issues which we thought we could address in some other place at some other time. So we left that six per cent in there.

Prof. Harding —As someone who spends a lot of time tracking trends in incomes and income distribution over time, it worried me greatly that, if you perhaps reset that threshold from 25 per cent of average weekly earnings to 26[half ] per cent, over time people would forget why that was. Then we would all be concluding that age pensioners had done fantastically well compared with, say, people of working age, whereas the idea was that, at least if it were a separately itemised GST compensation payment, you would be less likely to forget that was what it was there for.

Senator CONROY —I want to move on to consumption patterns. I think you have already set this out, but I just want to clarify it. It appears you have assumed that every household in a given household type has the same consumption pattern, irrespective of their income?

Prof. Harding —That is effectively right, yes. It is the average.

Senator CONROY —Because you picked an average.

Prof. Warren —Within the groups.

Senator CONROY —Yes, within the group. So, for example, a sole parent earning $100,000 per annum consumes the same proportion of food as a sole parent with no private income. Or, a couple on $150,000 a year with three kids spend the same proportion of their income on food as a couple with three kids who earn $25,000. That is your assumption. This seems to me to be similar, conceptually, to what the government did with the 1.9 per cent of CPI for everybody anyway.

Prof. Warren —I have difficulty with that observation, conceptually, only in the sense that the Treasury took the whole population as the group, and we have gone down to very specific subgroups.

Senator CONROY —You have gone to a narrower group, but you still made, in the end, the same conceptual—

Prof. Warren —Then I can turn around and say, `If I go to deciles, conceptually I am adopting the same approach as the Treasury on your logic.'

Senator CONROY —We probably would not have minded if you had.

Prof. Warren —It is a matter of degree.

Senator CONROY —This is important because, as you have described it, when you get up towards those thresholds, when you are doing an average mathematically—and in any of the modelling and statistics I have done over the years, which is a long time ago, the average is always distorted higher—you can have a few high incomes within a group and that actually pulls your average up to a higher position. So an overwhelming majority of people can be below an average figure. Is that an unfair position?

Prof. Harding —That is the difference between the median and the mean, yes.

Prof. Warren —That is the problem with averages.

Senator CONROY —So the information that you have put up in your charts does not necessarily address the issue of the consumption patterns of the low income earners within that averaging situation. Therefore, if you are saying that 3.4 per cent was an average for one particular group, you can have an awful lot of people who have a much higher figure than the 3.4 because some rich people who happened to fall into that household group had a much lower figure. I think that, if you look at the work you have done with ACOSS, where you did break it down into deciles, you can see that there is significant variation between the figures. Is that a fair way to describe it?

Prof. Warren —Yes. That is the nature of an average across a group.

Prof. Harding —In essence, the committee faced a trade-off here between having analysis done by quintiles and having the diversity of the 29 income groups that you selected. So, for example, for a single income couple with one child aged between five to 12 years, there were only 28 such single income unit households in the household expenditure survey, so it would not have been possible to take that group and break them down into quintiles because there would be only three or four observations in each quintile and the results would be statistically completely invalid. So it was a trade-off between doing it by quintiles and income levels and having this diversity of household types. You cannot have both.

Senator CONROY —You, Professor Harding, said that the key is the price impact that the people are going to face. That is the single biggest issue that people have to face in terms of compensation and whether they are better off, worse off, and those sorts of things. So if you have averaged a figure and the majority of low income earners in that group have quite a significantly high figure—again, if I pick 3.4 as a figure—of, say, 4.2 or 4.5 but a few rich people down the bottom of that household survey had 1.2 or 1.8, to give you your average figure you can still have the bulk of your low income earners well above your average figure.

Prof. Harding —I think that is potentially possible, and that was why we have argued consistently for a good-sized buffer, because the modelling is not accurate to the nth degree.

Senator CONROY —Have you examined any consumption patterns from low income groups in particular?

Prof. Warren —Not in this study.

Senator CONROY —We have received evidence from, for instance, St Vincent de Paul, of the consumption pattern of the people whom they deal with, who are in the very bottom categories. Their expenditure on things like food and other necessities is anywhere up to 70 per cent, as opposed to the sorts of figures that have been bandied around by the Melbourne Institute, for example, which are just averaged out at 25 per cent. What sort of figure did you use? Were you able to make assumptions?

Prof. Warren —We were using the household expenditure survey as it reports those individuals' expenditure patterns. We have not found it necessary to go in there and interrogate that data source.

Senator CONROY —I appreciate it is hard. Part of the argument Treasury has used is that it is too hard. You have argued that, no, it is not too hard, that if you make sensible adjustments you can get there.

Prof. Warren —All things can be done, but you have to stop and ask, `What is the sample size like?' and those kinds of issues. It is also the methodology that you adopt and we opted not to go with quintiles or deciles but to replicate again comparable information to the ANTS cameos approach. Those questions that you raise can be addressed, but it is a different methodology and results in a different brief.

Senator CONROY —To quote you, Professor Harding, when you appeared before us last time, you said:

Although the average might be a small net gain, none the less you might have 50 per cent of the group making a loss based on what their particular expenditure patterns are.

You would still absolutely make that point, even after you have produced this for us?

Prof. Harding —I would absolutely still say that what we have given you is the average price effect for each of the households within 29 groupings of the population. There is absolutely no doubt at all that that average disguises the fact that some people will have higher than average price effects and some will have well below average price effects. So in cases where the apparent gain is extremely low, there is a very distinct possibility that some people fall below that threshold.

Senator MURRAY —Even a majority.

Senator CONROY —Yes, even a majority. You used the word `disguises'. I think that is a fair way to characterise this. I have been looking for a word and I think `disguises' captures it. Your study disguises that there could still be a significant—even on all of the assumptions that are favourable to the government and we will get into argument about them in a minute—range of people who can still be worse off, despite the average saying that they are actually better off?

Prof. Harding —They can be worse off than shown by the average effect.

Prof. Warren —If those people are on the borderline; if they are sitting there with an apparent no net gain, if the mean and the median are the same, they are going to fall each side.

Senator CONROY —I picked a 50 per cent figure there, but if you picked the mode, or the most commonly occurring, then that can be well below the average. A majority—as Senator Murray kindly interjected for me there—is still possibly far enough below what you are describing as the threshold, the zero. You can actually have people who are represented by maybe a couple of dollars above and a majority of people in that category could still be worse off, despite the average showing that this category is better off.

Prof. Warren —Which brings us back to the point that we continue to make that it is the adequacy of the compensation and the buffer that is the issue.

Senator CONROY —That depends on when you start playing with other assumptions.

CHAIR —It also depends on how long and permanent the compensation will be.

Senator CONROY —If I could move on to some of the assumptions. Have you assumed a rate of dissavings of 10 per cent for low income households? I think you said you did.

Prof. Harding —Yes, a maximum of 10 per cent when you have zero private income. That is households that are entirely dependent upon social security.

Senator CONROY —If the actual dissavings rate for low income pensioners and unemployed people is higher than this, would this mean that you have understated the losses to them?

Prof. Harding —Yes, it would—losses or the extent of the gains.

Senator CONROY —You indicated that you just picked 10 per cent. It was something they used in 1985?

Prof. Harding —That is correct.

Senator CONROY —On page 12 of your paper you gave the example of an unemployed person with $200 per week of disposable income spending $250 per week. That is someone with a 25 per cent dissaving rate?

Prof. Harding —Yes, it is. I had to try and make the maths easy.

Senator CONROY —I appreciate that you are pitching it to us. AC0SS and others would argue that 10 per cent is a conservative estimate for people in the low income categories, and that a more realistic figure is 20 per cent.

Prof. Harding —It is difficult to know. If you look at the household expenditure survey data, it probably does look a bit conservative. Then you come smack up against that issue that you know that the ABS has specifically told you that you should not take the difference between expenditure and income as shown in the household expenditure survey as a measure of the savings. They are telling you not to do it. Then you have to get into all those complex adjustments of what you do to the micro data to try and make it stack up more adequately against the macro picture.

Senator CONROY —Hopefully, AC0SS would have been talking to organisations like St Vincent de Paul, Good Shepherd, financial counsellors and a range of welfare groups who deal with people in this sort of category. They have to help them prepare a financial budget, so they actually have many nitty-gritty examples which they have supplied to us.

Those are the sorts of real world examples that they are drawing on to try and indicate a figure of 20 per cent rather than 10 per cent. If 20 per cent was a more realistic figure, based on Good Shepherd, other financial planners and St Vincent de Paul who deal with those things, that would either understate their gain or possibly move them into a loser category, depending on what assumption you make.

Prof. Warren —I am curious how a person in an extremely difficult financial position can maintain a 20 per cent dissaving on an indefinite basis.

Senator CONROY —I am sure they would welcome you to come down and sit with them for a day while they went through people's finances. It might be useful for everybody.

Prof. Warren —Hopefully, what you are talking about are people in transition and we are talking about buffering those people through that transition period. And, yes, we need to ensure that there is adequate compensation for those groups. The 20 per cent is not a lifetime scenario for those particular groups. Hopefully, they will come out of that significant dissaving position. While we focus on this, we would hope that those people would move out of that difficult position that they find themselves in. While it might be the case, one would hope that it was not an endurable case.

Senator CONROY —I am sure that everyone would benefit from you having a chat with them direct, rather than through an intermediary. I am confused about your treatment of the 70 per cent passing on of indirect tax reductions. On page 1 of your report you say:

. . . the Senate Committee asked us to abstract from the timing issues associated with the introduction of the various indirect tax reforms . . . In effect, this means that this report examines the expected longer-run distributional impact of the package.

On page 2 you say:

We consider this 70 per cent figure to be well at the upper end of likely medium term outcomes.

You repeat this on page 15. Yet the terms of reference reproduced at page 42 require an examination of four packages for July 2000, including a 70 per cent pass through. On page 6 you describe option 4 as being:

. . . with 70% pass through of indirect tax cuts in the first year . . .

And on page 30 you confirm that:

. . . the Senate Committee instructed us to model a scenario in which only 70 per cent of the expected price reductions were passed onto consumers in the first year.

Why did you reinterpret the Senate committee's instructions on the 70 per cent pass through as applying to the medium term, when you clearly acknowledge in your report that it was to apply to the first year?

Prof. Warren —What we were replicating was the methodology adopted in the government's report. That was our understanding.

Senator CONROY —We did not ask you to do that.

Prof. Warren —That is your wording in the terms of reference. It is the same as the wording that is given. That was our interpretation of that, and I clarified it.

Senator CONROY —This is your clarification with—

Prof. Warren —Yes. That is my understanding, that what we are doing is a sensitivity analysis for the cameos. Now, that is the basis on which we proceeded. That is the benchmark and we worked from that benchmark to look at the sensitivity of those various assumptions that we have looked at and how they have changed. In clarification of our terms

of reference, that is how I have in the end interpreted those and proceeded on. In ANTS it is the medium term in July 2000 prices.

Senator CONROY —We did not actually ask you to model ANTS. You did that for free.

Prof. Warren —No, option 1 is where you include tobacco, but if you are going to use it as your benchmark, you have to start from somewhere. We have to start from what is the beginning point.

Prof. Harding —It is also difficult to do the 70 per cent pass through when stamp duties do not come off until the second year.

Senator CONROY —I understand that aspect. I will quote back to you, Professor Harding, your evidence from 28 January, at page 217 of Hansard , where you say:

My understanding is that some of the indirect tax changes take a long time to travel through the system.

You do not seem to think that today.

Prof. Harding —To be honest, this is Professor Warren's area of the model. He made assumptions about the indirect tax side. Indirect taxes are not my area.

Senator CONROY —Professor Warren, do you stand by your assertion that, on Saturday 1 July 2000, every good and service embodying the wholesale sales tax will fall by the full extent of the abolished wholesale sales tax?

Prof. Warren —Instantly on that day?

Senator CONROY —That is what we asked you to model.

Prof. Warren —Is that 12 o'clock or 5 o'clock? That is being pedantic.

Senator CONROY —Twelve or five; take your pick.

Prof. Warren —In a sense, this comes back to the adequacy of the compensation. I have a problem, and I think members of the committee will too: you will not know the adequacy for several months of that four per cent that is given on that day.

Senator CONROY —Correct. You help the government out by retrospectively then counterfactually—to use your word—saying that they would of course respond. That is a very generous assumption on your part. What we asked you to model was what happens to a real person who gets his four per cent on day one.

Prof. Warren —Then you asked us to make some judgment about what would happen. Do you want us to look at six weeks or three months? How effective is the ACCC going to be? How effective is the education program going to be?

Senator CONROY —We will come to that in a tick.

Prof. Warren —All of those are issues which come to bear.

Senator CONROY —Let me give you a real world example and then seek your opinion. On Saturday 1 July, your local repair man comes around. You would be expecting him to pass on the savings he has obtained on his equipment when he charges you his new price including the 10 per cent GST. You are assuming that he will, of course, discount that a bit because he is going to pass on to you the savings he got from the gloves or the lawnmower he bought a few months ago.

Prof. Warren —There is always going to be a ramping down, a transition. There is no question about that. On that day, I would not expect it all to instantly happen.

Senator CONROY —The question we asked you to model for that first 12 months was how much.

Prof. Warren —If you are asking me about midnight on 1 July, when the GST comes in, that is very difficult. It is a function of the extent to which wholesale sales tax credits are given for various stock that businesses have. There are a raft of issues. The question is how quickly that flows through and the kind of buffer that government has given to those people in the transition to that new outcome. So it is an issue about the adequacy of that four per cent. The CPI that flows through on that day will be higher than the government's expected level.

Senator CONROY —I will read to you from a submission to us from Howard Smith on the question of the realisation of the benefits of tax reform. Howard Smith had been invited to have a chat with the ACCC about how the ACCC will implement what is going on. Howard Smith says:

However, until the benefits are realised by business, the prudent course of action will be to continue pricing based on their current costs. As these costs will fall due to the benefits of reform being realised, prices should also fall. In some capital intensive businesses this may take longer as costs will not fall until the capital costs which drive prices fall either through revaluation or replacement.

It is essential for the ACCC to acknowledge the need for businesses to set prices based on their cost structures at particular points in time. An onerous monitoring regime which does not allow for lags in receiving the benefits of reform will create additional pressures and costs which may undermine the reform process.

So a significant company is negotiating with the ACCC to make sure that the ACCC do not hypothecate cost savings and try to enforce them on day one. The ACCC have invited every industry in this country to come to talk to them about their specific problems on this front. The ACCC are not running around, as you hope, chasing everybody down to make sure those savings are immediately passed on.

We have also had evidence from the ACCC that they cannot name one single genuinely competitive sector in this country. They could, however, point to hundreds of examples of uncompetitive behaviour. They exist because uncompetitive behaviour exists. So you still

stand by your statement that 100 per cent can pass through and that 70 per cent in the first 12 months is unrealistic, unfair?

Prof. Warren —You have to ask the ACCC what kind of time frame they feel these taxes that are coming off are going to be passed through. It is problematic. One would presume, if you have an organisation like the ACCC being vigilant, that part of its task in doing what it is doing is an education process—to make people mindful of this. I have been to seminars where the focus is on ensuring that you get from your supplier the expected price reduction for the inputs that you have got. It is very difficult to make a judgment as to how fast that will flow through. Economics does not really tell you how quickly it will go through; it has got to do with the state of the economy, the competition, how vigilant the ACCC is, how effective it is at the end of the day—all those kinds of issues. It is not going to happen instantly. There will be a ramp down. The question is how fast that transition phase will be. That is very difficult.

Senator CONROY —This comes to the key question of the price rise that the people are actually going to face, which is what Professor Harding says is the critical question. You have been very strong in your paper in supporting the 100 per cent pass through.

Prof. Warren —We have looked at the medium-term position.

Senator CONROY —We did not ask you to look at the medium-term position; we asked you to look at the first year impact.

Prof. Warren —You are asking us a question that we, and I presume the Treasury, the ACCC and others, would really love to know the answer to: what will the position be on day one, week 3, week 6 and so on? That is problematic. Again it comes down to the issue of adequate compensation and adequate buffer. That is the critical issue. Ultimately my view is that the majority will simply flow through.

Senator CONROY —But 70 per cent is a majority.

Prof. Warren —The effects will flow through ultimately. It might take time; the question is how long.

Senator CONROY —Look at page 31—I am under pressure from time. Your report says:

The only potential long term downside for the consumer is if businesses never pass on the remaining 30 per cent, which is hopefully unlikely in competitive markets and with a vigilant ACCC.

Are you aware that there is a sunset clause of two years after the introduction for the ACCC to look at this issue?

Prof. Warren —Yes. One year in; two years out.

Senator CONROY —So that, in your medium term, in which you are hoping it is all passed through, they have actually stopped looking specifically at this issue.

Prof. Warren —One would expect that within two years the majority of these things will have flowed through.

Senator CONROY —I want to move on to the tobacco question. My questions mirror those of Senator Harradine's slightly. You say on page 1 that you do not endorse using tobacco price increases associated with the tax package as the appropriate indicator of the likely change in prices facing households. It is one thing for you to make a value judgment about the morality of the increase in the cost of smoking, which you have chosen to do, but it is another thing to claim as economists that the increase in cigarette prices is not the appropriate indicator of the likely change in prices facing households. By discounting fully for the increase in cigarette prices you are saying that smokers will not face these increased costs. How do you defend that?

Prof. Warren —You can argue, with the economic rationale, like the Pigouvian tax, that what they are doing is paying for adverse externalities associated with smoking.

Senator CONROY —But the prices are actually increasing. What you are arguing is that they should be compensated. The point is that the prices are actually going up.

Prof. Warren —Yes.

Senator CONROY —That is the real world effect. As Senator Harradine described, when a person walks into the shop the price has increased.

Prof. Warren —Yes.

Senator MURRAY —Perhaps I can assist you, because I think the point both of you have made is an important one. We received evidence from Professor Dixon that there was a danger of significant wage demands—and the ACTU confirmed that—if the price effects overall were regarded by workers as being very heavy. The point they are both making is that, regardless whether we think it is right or wrong that if tobacco is a major portion of their expenditure that will push the wage demands.

Senator CONROY —Some people will pay more in tobacco tax than income tax.

Senator FERGUSON —You don't.

Senator CONROY —I don't smoke.

Senator MURRAY —That is the issue, isn't it? It is the wage demand issue.

Senator CONROY —That is absolutely right. That is one element of it, definitely. It is just that I do not remember John Howard saying, `No-one will be worse off under this package, unless they smoke.' He said no-one is going to be worse off. You are now making a judgment to say, `That is okay then. If you smoke you are allowed to be worse off.' People, when they walk into that shop, are going to be worse off if they smoke. And John Howard said no-one is going to be worse off—or have you got a press release from the Prime Minister that I missed?

Prof. Warren —No, but you also need to understand the argument we are making in terms of tobacco.

Senator CONROY —What is the optimal Pigouvian tax on alcohol?

Prof. Harding —It is good for you.

CHAIR —No, we have evidence from the AMA that for a lot of low income earners it is not, in view of the way in which it is consumed.

Senator CONROY —That is a humorous answer but is avoiding the serious point. You are not advocating that there should be some increase in alcohol taxes, are you?

Prof. Warren —No, but by definition you are arguing that we should compensate people for something we have made a social judgment about not wanting them to do.

Senator CONROY —But you are making an assumption now that the current level of tax on alcohol is actually less than the Pigouvian rate, because you are saying that when it is going up—have you got any evidence on where it is?

Prof. Warren —The implied judgment is that it is suboptimal, and if you base that on work that the likes of the Heart Foundation and others have undertaken, then you would presume that in their research and thinking on this they are following some kind of standards to make that informed judgment.

Senator CONROY —I am not worried about what they are doing. I am worried about what you have based your judgment on—that it is suboptimal at the moment.

Prof. Warren —Firstly, the rationale that other governments—Labor as well as conservative—have taken out tobacco when thinking about this issue, and also the Heart Foundation, the Cancer Council and others have argued that tobacco taxation is not high enough. I can only defer to precedent in other places and to the literature rather than to a personal point of view. It is citing, basically, the economic position by others.

Senator CONROY —You are also aware of the press conference held by the AMA where they condemned the reduction in diesel excise on health grounds. Do you agree with the AMA that diesel prices should not fall? What is the optimal Pigouvian tax on diesel? You have a view in one area but not in others.

Senator FERGUSON —It is generally accepted that the AMA's position—

Senator CONROY —So the Heart Foundation's is and the AMA's isn't, all of a sudden, Senator Ferguson.

Senator FERGUSON —It is not.

Senator CONROY —I just find it a little unusual that you have got a particularly strong view on a Pigouvian tax in one area but not in others.

Prof. Warren —In a sense, I suppose we could replicate the same kind of conceptual framework in relation to diesel fuel and make an issue there about effectively what has happened and you can push that argument through a whole range of other issues as well.

Senator CONROY —Yes. It is just that you have a particularly strong view about the compensation in this one area but not in the others.

Prof. Warren —I suppose it is drawing on the precedent of what has been done in the past. That is why I have drawn on what they did in Fightback, what the government has done, and what the Cancer Council and other agencies have argued. That is far more important, in this particular example, than is the diesel fuel issue as it flows through. It is still a conceptual framework.

CHAIR —This is the problem that the chair faces. We are now over time, but we still have some time for the coalition to use up. And looking at the questions that Senator Conroy wishes to ask, he would need a lot more time if he were to be able to ask all of the questions he wants to.

Senator MURRAY —And me too.

CHAIR —And from the Australian Democrats, and I think Senator Harradine indicated a similar situation. I just point this out from the point of view of saying that time is going to be a problem for us in bringing this to an adequate conclusion. I flag that, over lunch, we might have to talk about the matter of a rescheduling for other witnesses today, in order to do justice to the type of inquiry we are conducting. I make that as a comment to the committee.

Senator FERGUSON —May I make a suggestion that perhaps we go until 12.45 and make that the cut-off point? That would give us three-quarters of an hour then.

CHAIR —Let us proceed and we will just see how far we go. I must say I am not very confident that we will actually acquit all of the questions that members are wanting to ask. You have the call, Senator Ferguson.

Senator FERGUSON —I just want to return to the tobacco issue, because it seems as though Senator Conroy has announced a change in Labor Party policy today. During the 13 years that Labor were in government, they regarded tobacco as a punitive tax. I am wondering, Professor Warren, whether you can recollect in that 13 years that they ever compensated for any of the tobacco tax increases that occurred during that period of time.

CHAIR —I thought we were a policy free zone.

Senator FERGUSON —Senator Conroy announced one this morning, I think. No longer do they regard tobacco taxation as punitive.

Prof. Warren —The precedent of the past, which is what I was drawing on in making those observations, has been not to seek compensation for something that is purposely designed to elicit a behavioural response. To raise the tax and then compensate questions the

whole purpose of raising it, because you took with one hand and gave with the other. So what is the purpose of what you have done? If that is what you are going to do, then do not change it. It is designed to get a response and, by definition, you are not going to get the money back, which could neutralise that response.

I know there is an issue here about the distribution of the impact of such a price change and the effectiveness of Quit campaigns and so on. The committee has not read the Heart Foundation stuff; it is quite an interesting read for developing the issues, but the precedent of the past has been not to compensate for those—

Senator FERGUSON —And this government has only followed precedent in not calculating increases in—

Prof. Warren —And state governments themselves, when they lost the tobacco franchise, were at a 100 per cent rate, and at no stage did people really cry from the hilltops that this deserved compensation by state governments. They continued to do it. There was some discontent but everyone accepted it. There was no cry about compensation when the 1993 Dawkins budget introduced these program changes, and now we have this concern about the need for compensation in this case.

Senator FERGUSON —It seems that follows on from the fact that there is almost universal agreement that smoking is harmful, and that is why the punitive taxes have been put on. I want to ask a question of both you and ACOSS, when they appear after lunch. Your NATSEM results do not seem to support ACOSS's model which they released recently, which I can only presume you have had a chance to either see or read about. Have you got any comments as to the reasons for the apparent discrepancy between ACOSS's modelling and your own results?

Prof. Harding —We do have views, but we thought it might be appropriate to hear ACOSS describe their methodology. On the part that we modelled, we felt that they had not modelled the family tax package.

Senator FERGUSON —You felt that they had not, or—

Prof. Harding —It is obviously difficult for us because—

Senator FERGUSON —It is pretty hard for me to ask them that question, because I am not going to get a chance to ask you again.

Prof. Harding —I suppose we hoped that we might be coming back to talk to you again after the presentations by ACOSS and Treasury and we could address any of those issues that it had raised for you. It is difficult for us to tell, in some cases, exactly what they have done, because the methodology is not fully spelt out in the document that they put out. But we believe that they did not model the benefits of the family tax package.

Senator FERGUSON —You believe, from your observations, that they have done their modelling without taking into account any of the effects of the family tax package?

Prof. Harding —We believe that is one issue. There are some others.

Senator FERGUSON —And you do not want to share those others with us?

Prof. Warren —There are some concerns about how they got to July 2000 from the Melbourne Institute numbers. Again, like Ann, I have a concern about the methodology, but that is really seeking an explanation more from ACOSS as to how they got to those of July 2000. They are quite important in terms of how the price effects move and then how people lie each side of the four per cent. It is from there that they generate the concern about the losers. Like Ann, I would prefer to defer.

Senator FERGUSON —I cannot remember the details of their results; I have only read the stuff that they have sent us. But you are suggesting that in fact they somehow or other devised a methodology which would be able to give them results on 1 July 2000, and you question that methodology.

Prof. Warren —They used a technique which essentially takes the 2001-02 numbers and adjusts them for the information that the Treasury indicated as being the expected July 2000 number versus the 2001-02. The 2001-02 number is 1.9 per cent; the July 2000 number is 2.5 per cent. So they adjusted the Melbourne Institute numbers by 2.5 over 1.9 times the 2001-02 numbers. That is the methodology they use to get to July 2000. We had difficulties with some of that, which begins to explain why we do not get the same order of magnitude of the losers as they get. But, in terms of the detail, it is best if you ask those questions of ACOSS.

Senator FERGUSON —So what you are doing is questioning the validity of the methodology that they used?

Prof. Warren —It needs further explanation, essentially, because there are very different outcomes between the two studies. It is a question that can legitimately be asked of us as to why we are different. In a sense, it would assist if those questions were asked of ACOSS and they explained how they see us differing from them.

Senator FERGUSON —We will ask those questions of ACOSS. The problem is that we may not get a chance to get you to respond once they have told us what their answers are. There are significant differences. One of the problems that we have had right from the start is that, with all models, it depends on what assumptions you put in and whether those assumptions are either valid or realistic.

Prof. Warren —I would prefer their explaining it.

Senator FERGUSON —We will ask them this afternoon. Are you remaining this afternoon to hear their evidence?

Prof. Warren —Yes.

Senator FERGUSON —The only other issue I would raise right now is that of, again, your confirmation. We have heard other people ask you questions about options 7, 6, 5 and

the whole lot, yet you still say in your package that, for all the reasons you noted earlier in your report, the methodology you adopted in option 3B, where tobacco and new home prices were excluded, is probably the most appropriate.

Senator CONROY —That is a moral judgment; it has nothing to do with economics.

Senator FERGUSON —Senator Conroy suggests that it is a moral judgment. I suggest that he change their policy, because they now think that punitive taxes on tobacco should be compensated for when in fact the reason they are put there is in order to discourage people from smoking.

Senator CONROY —The Prime Minister said that no-one would be worse off. It is just an accident that the low income groups contain the heaviest smokers.

Senator FERGUSON —I suppose what I am really putting to you involves taking into account all the things you have used in the other options. That includes the 70 per cent pass-throughs and your own judgment; Senator Conroy calls it your moral judgment, but it is the moral judgment that has been passed by both previous Labor and coalition governments as well. Taking into account all the options you have presented to us, option 3B is the one that is the most appropriate for us to base our assumptions on. Is that right?

Prof. Harding —We think 3B is an appropriate summary of what the government intends from the ANTS package. But we would argue that, given the uncertainty that attaches to the gains results—in that particular case everybody is a gainer—due to sampling error, data error and all sorts of things over which we do not have control, the margin of gain appears too slim for some groups to enable absolute reassurance about the adequacy of compensation. So we are not saying that 3B is the best package in that sense.

Senator FERGUSON —In that sense.

CHAIR —On the Pigouvian tax issue—

Senator MURRAY —That is a new one for me, I must say.

CHAIR —It is true, is it not, that low income earners have a higher propensity to buy cigarettes than high income earners? So, if you have their prices in, you are better reflecting their consumption pattern. Leave aside the moral argument. That is true, isn't it? Was that a nod?

Prof. Harding —We certainly did some research looking at who smoked. Lower income groups had higher smoking rates.

CHAIR —It is also true that there is far less disposable income available for lower income earners by definition than anyone else. They are mostly dissavers. Is that right?

Prof. Harding —I am not sure what you mean. Do you mean that they have lower disposable incomes than other people?

CHAIR —They have low discretionary income or no discretionary income. They are basically dissavers.

Senator MURRAY —What they have they spend.

CHAIR —Leaving aside all the other arguments, these smokers would need to be compensated for that, if they were going to be adequately taken care of in this package. That is the point. It is a cost they face?

Prof. Warren —You are assuming that they continue to smoke when the whole purpose of the tax change is to discourage them from smoking.

Senator CONROY —That is not the purpose of this tax change.

Prof. Warren —What is the purpose of it in that case?

Senator CONROY —It is about a broad based consumption tax. No-one said that we need to put up the price of tobacco because of the Pigouvian effect.

Prof. Warren —It is moving to a per stick system. You could argue about whether they need the rate of tobacco to go up.

Senator CONROY —Treasury have not argued that the reason the tobacco prices are going up is Pigouvian. It is because they want a broad base. They have not put forward a Pigouvian argument once. Maybe they have listened to you and they will this afternoon.

Prof. Warren —But there have been various arguments by the National Heart Foundation, the Cancer Council and others.

Senator CONROY —Treasury have not cited them as their reason for what they are doing.

CHAIR —There have been various arguments by the AMA about alcohol and the diesel fuel rebate, none of which are taken into account by the analysis.

Prof. Warren —One is always reluctant to throw it back to the committee. However, when Labor put in the Dawkins budget a series of much more significant increases in tobacco excise between 1993 and 1995—

Senator CONROY —And a CPI increase in pensions. You continue to try to make this assertion—

Prof. Warren —But the CPI index—

Senator CONROY —Included tobacco. So when tobacco went up, pensions and the dole were increased.

Prof. Warren —It was marginal.

Senator CONROY —It is a component of that increase. You are saying that it should not be a component because you are claiming there was no compensation previously. That is not actually true. I am not saying that it was the whole of it. Part of the indexation they got through the CPI was because tobacco was in it.

Prof. Warren —What percentage?

Senator CONROY —You have asserted that there was no compensation. That is not correct.

Prof. Harding —Although only for those within the social security net, not for other low income earners.

CHAIR —The poor people.

Prof. Harding —Some of them.

CHAIR —In this inquiry, we are bound to look in particular at the effects on low income earners. The other question is about this pass-through issue. We have quite a lot of evidence before this committee that, particularly in regional and remote Australia, there is no perfect competition at all. In fact, monopoly rents are extracted in the supply of goods and services. There are sole suppliers. Melville Island, for example, is served by one ferry. There is no competition and, therefore, no pass-through. The prices charged in stores in those regions of Australia are nothing like the prices charged in capital cities. They are much higher. You cannot account for it on the basis of fuel costs or transport costs only. There is a monopoly rent situation.

In those circumstances, those people in regional and rural Australia will be worse off, won't they, under this package? You cannot make comfortable assumptions about pass-through due to competition. They are always hard to nail down by the ACCC. There is a tradition of this type of behaviour. They have monopoly rents extracted.

Prof. Warren —I suppose the only thing you could argue would be the extent to which you could differentiate between businesses in a rural area and those in a city area. I have talked to some people in ATSIC who have sent me various data on exactly this issue. I fully understand and appreciate their particular concern with this situation. Again, you ask me to make a judgment, and the issue that you have raised is one of concern. But, again, we refer back to this being an issue of the adequacy—the buffer that is there to protect in those particular circumstances where this becomes an issue.

Senator CONROY —You are not being fair in the way you are describing that.

CHAIR —I will put it this way: we have had evidence that the cost of a lettuce in Perth is $1.50 and in Kalgoorlie $2.00. Both have the 10 per cent applied, so you pay a higher rate of tax on a fresh vegetable in Kalgoorlie than you do in Perth. A state government appointed regional development authority tells us that only one supplier undertakes transport radiating from Kalgoorlie to remote mining communities—and I am not talking about Aboriginal communities. This is not a contested market; there is only one supplier. Their regional

development concern is that monopoly rents are extracted by that one supplier. The lettuce is $2.50 to $3.00 in those remote communities.

The point is that, in terms of pass-through, for residents of that part of Australia and some other regions, there now is no perfect pass-through, and they have no expectation that there will be in the future. If the buffer is there, and if that is the way it is all compensated for, then you have to go to the analysis of whether the buffer is at all adequate, don't you?

Prof. Warren —I am sorry, would you please repeat that last part, as to whether the buffer is—

CHAIR —Whether it is at all adequate; whether or not the buffer does resolve the problem.

Prof. Warren —Our analysis cannot start to throw light on that because we have lost the regional identifier within HES; we no longer have the regional identifier we had in the previous HES. That is a very difficult one for us as researchers, because we used to address it with the previous HES.

Having looked at that issue, and having helped some of those people in ATSIC on it, I understand and appreciate it. In a sense, how effective the ACCC could be in that particular environment is a question for Treasury rather than for me. As you were saying, you can contest whether it would be. They would obviously raise the issue of the diesel fuel rebate and the potential benefits that would accrue to those with distance—

CHAIR —If it is passed on.

Prof. Warren —If it is passed on. But, in a sense, that is part of the brief of the ACCC. I can then only say that I would presume they would do their job and, with complaints, respond to them.

CHAIR —But the identikit portrait of the one who really gets it in the neck under this package is someone living in a remote region who is a low income earner.

Senator FERGUSON —But, if they are a monopoly, who sets the price today? Nobody is setting what their margin will be today. So when a GST comes in—

CHAIR —You are making out my case for me, Senator.

Senator FERGUSON —No. When a GST comes in, the ACCC will have the chance to make sure that the GST is passed on; they do not at present. Nobody sets the price today.

Senator CONROY —It has been raised with me that there was a nod of the head before that could not be recorded by Hansard. I would just go back to your assertion of there being no compensation for the 1993 increases. I put it to you that in actual fact there was. With tobacco being within the CPI and with that increase flowing through into the indexation, there is some compensation for that increase. I thought you nodded, but I have been told that Hansard cannot record a nod.

Prof. Warren —The explanation is, to the extent that wage and salary earners in the eight capital cities smoke and that is taken account of in the CPI, there is for people who are welfare beneficiaries.

Senator CONROY —That is a yes?

Prof. Warren —That is a yes in relation to welfare, but no in relation to non-welfare recipients.

Senator HARRADINE —Are you now dealing with the 13th series and taking that as the yardstick?

Prof. Warren —We have opted to stay with the 12. But there is an appendix which details the 13th series. Once the CPI used to include homes, then it went to mortgages, and now homes are back in it again.

Senator HARRADINE —Yes, I understand that.

Prof. Warren —As a consequence of the restructuring moving to this 13th series, there are implications for the price index which comes out of this package. So we have continued to work with the structure that comes with the 12.

Senator HARRADINE —I will cover that in questions to Treasury.

Senator MURRAY —Professor Harding, with the qualifications you have made, the options are of great interest to all of us. We are very interested in the food approach. Option 5B has food out and income tax adjusted but pensions up only 2.3 per cent. From my reading, option 5B still is more progressive, or produces more progressive outcomes overall, than option 7, particularly looking at the impact on families. I wonder whether you could comment on that.

If we accept that the government's package will be adjusted through either compensation tax cuts or compensation, food out and tax cuts or changes, then those two options have to be examined. My reading of 5B is that it is more progressive, particularly looking at the impact on families, than option 7.

Prof. Harding —I suppose I should flag first that option 7 might not necessarily be in any way what you ended up going to the table with. It was merely an illustrative option to give you a bit of an impression about what would happen if compensation were more generous. The package was still revenue neutral, but food was still in the base. That was just to assist you in making choices.

It looks to me to be really hard to pick which is more progressive out of 5B and option 7. Option 5B does deliver lower dollar gains and percentage gains for social security recipients, for people right down the bottom of the income spectrum, than does option 7. So 7 is better, I think, for the very, very poor who are on social security—and, as you have emphasised, the `who are on social security' qualification can be a significant one.

Senator MURRAY —Yes.

Prof. Harding —At middle income levels, although it varies greatly with the family type and the household composition, it does look as though food out delivers greater gains at the lower-middle to middle income levels—say, somewhere around $30,000 to $55,000—than does option 7, as it is presently constructed and configured.

Senator MURRAY —This is the seminal issue. Business groups are saying to this committee: if the committee finds that compensation is inadequate, then give greater compensation. Do not take food out of the GST because of the complexity issue. The social security groups—welfare, the churches, the bishops—are saying: if you take food out of the GST, you make sure that everybody gets the benefit and that is a better way of dealing with this problem than just relying on compensation. So they are saying compensation plus food and businesses are saying just compensation. That is why your answers and judgments on this are very important. I have understood your response.

Professor Warren, I want to go back to housing. We are obsessed with food and housing because food is 56 per cent of low income expenditure and housing is 19 per cent. That is why we emphasise it continually. On page 47 you have been good enough to give the changes in the CPI weights of the 12th and 13th series. In the middle of that list, for those who are referring to it, is the housing utilities and rents sections and how those weights have moved. Could you explain to us what is or what is not picked up in that list of things? Essentially I understood you to say that, in the 13th series, rates which have a weight of 1.18, home repairs of 1.72 and insurance of 0.21 are in the calculations but you are recommending that the price of new dwellings at 6.88 stay out. Is that right?

Prof. Warren —Yes, that is right. You will see that in the old series mortgage interest is in there. So when you are working with the 12th series obviously that is going to be in there. Where the house price issue comes in in the 12th series is as it relates to things like alterations and those kinds of components. If you went to the 13th series, because those weights have changed you are going to get different outcomes because now new house prices will go in there as a component which was not previously in there. So there is some discussion of this issue in the report, that the price index under the 13th series will be a different number with housing in because of the way the CPI weights are now determined.

Senator MURRAY —In the CPI determination in the 13th series—and of course the 13th series now picks up nearly two-thirds of the community whereas before it only picked up 30 per cent, so it is a major improvement—the total weight is 19.35. Are you telling us that all this is accounted for except the 6.88? Is that my understanding?

Prof. Warren —That is correct. What I have done is dropped out things that relate to houses—alterations and those kinds of components. But the rents—those issues—are still there as they are here. Home ownership is not rental.

Senator MURRAY —I will have to think more about what that means for us. Professor Harding, would you agree that, if the government went down the compensation route, increasing pensions by two per cent, I think, as presented in option 7, would not be enough

and that rebates for low income workers and additional family assistance would also be needed? I ask that question in terms of the qualifications you have put previously to us.

Prof. Harding —It was a six per cent increase. A group who makes relatively low gains is two-income couples at middle income levels. Today most children live in two-income families, not single income families, so there are a lot of them.

Senator MURRAY —It is an additional two per cent, is it not, because it is four per cent at present?

Prof. Harding —In effect, it is an additional 2.6 per cent because we modelled them as being where they intended to end up, which is also what they modelled in the cameos.

Senator MURRAY —Would you have a figure as to what that would cost?

Prof. Harding —Yes. Going from the 3.4 per cent to the six per cent would cost about $1.1 billion.

Senator MURRAY —But if greater rebates for low income workers and additional family assistance is needed, it would cost a lot more than that, surely?

Prof. Harding —Yes. The package is very generous towards younger children and towards single income families, but there are possibly some question marks about the extent of the increases for children aged five years and over. I guess in part, particularly given the fact that whatever other drawbacks they might have, as the other questions have illuminated, the household or the cameo specific CPIs have suggested that families with older children do face a relatively higher price effect. But it costs a lot of money.

Senator MURRAY —Yes, it does. You have stressed to us that you spend—it is your job—a great deal of time on distributional analysis and you have been at this for a long time. You can confirm to us, can you not, that the trend has been for tightening access to the social security safety net over time?

Prof. Harding —Absolutely. But to some extent it has been counterbalanced by the wholesale introduction of assistance to low income working families with children. At the beginning of the 1980s there were no government payments specifically for low income families in the labour force with children, only for social security recipients, whereas now there are very large amounts of money relative to average weekly earnings.

Senator MURRAY —But, if that trend continues and you keep shrinking the number of people who can access the social security net, it just reinforces the number who fall outside of that and therefore will not get compensation. You have to be in the net to get the compensation, haven't you?

Prof. Harding —I think you have alluded to some of the potential holes, and they are people who earn low incomes but are not unemployed. Technically they are defined as employed, whether they have part-time work or casual work or intermittent work, but they

do not have children. The key compensation mechanisms that we have to low income working people are only for those with children, not for those without.

Senator MURRAY —My difficulty is that no matter how generous the government are, whether they go to six per cent or eight per cent or 10 per cent compensation, which would be very nice for those people who will get it, there will always be a substantial number—and we think running into millions—of people who will fall outside that.

Prof. Harding —I am not quite sure why—we have not modelled it either—people did not look at things like just increasing the low income rebate or introducing a GST compensation credit through the tax system, which would have got to those people who are not current recipients of social security programs.

Senator MURRAY —So the tax credits idea, in other words?

Prof. Harding —Yes.

Senator MURRAY —Yes, we are interested in that. Would you concede though that, given the government's present policy settings and their refusal so far to look at the tax credits approach, zero rating food is a valid policy response to offset the equity concerns of the tax package as it affects those many millions who fall through the hole, as it were?

Prof. Harding —When you look at the distributional impact of package 5 or 5B relative to essentially the government's package, which is option 3, then there is no doubt that options 5 and 5B look much more progressive. In the government's original package there were higher percentage gains as one moved up the income scale and lower percentage gains as one moved down the income scale to low and middle incomes. The problem that you have, though, is that you do not know for sure that the very progressive changes that we modelled to the income tax scales as part of option 5 will actually be what happens. So it would be entirely possible, for example, to imagine a scenario where food was excluded, but the increases in income tax revenue required to finance that loss of $5 billion of revenue were not generated in the same way as we have generated them here. So we imposed a very progressive income tax clawback to get that $5 billion.

Senator MURRAY —As a person who really does understand these issues, because you deal with them all the time in your work, the Treasurer has on several occasions, as recently as last week in question time, described you personally, Professor Harding, as a strong supporter of taxing food. Is that an accurate reflection of your position? If it is, why would you, as a professional analyst and expert in this area, support taxing food when so many people are going to miss out on the compensation package?

Prof. Harding —We have obviously debated this extensively amongst ourselves as well. We would favour taxing food in the sense that it is more efficient, simple and so on to do that. If you did not tax it, then some of the costs that would arise we have not been able to take account of in this sort of modelling. With that said, nonetheless, if you assume that the $5 billion of income tax revenue would be generated in the way that we have done it here and food was out, then from an equity point of view that delivers more gains down the bottom than does the government's current package. Whether it is better than option 7 is a

decision that only you people can make. Option 7 gives more right at the bottom but gives less to the middle and then gives more to the upper middle.

Senator MURRAY —But I will happily give you great respect as a professional modeller in terms of distributional analysis. I will give you very little respect as an analyser of compliance costs because it is not your field. Frankly, we have had a lot of evidence which indicates that the compliance cost issue is overstated because of modern computerisation, because of how other countries have coped with it, et cetera. I do not mean to demean the definitional issues, because they are real issues, but the problem for me is this: you as a professional are making a professional judgment in terms of things you really do know about and it seems to me from your answer that you are also assuming that it is better to tax food because of notional compliance costs which you do not really know about. That is why I am very interested in why you would hold to a taxing food position. Can you respond to that?

Prof. Harding —I do not know if Neil has strong views on the compliance costs.

Senator MURRAY —My question is to you though, because the Treasurer has been picking you out as a great supporter, and that matters, because we all do respect you in your field.

Prof. Harding —If you are only concerned about the equity side, then the option with food out and income tax cuts structured in the way that we have them here looks more progressive. It gives more to the poor than does the current package, but I do not know whether that would be the only consideration you would use in making that decision.

Senator MURRAY —Yes, I understand that. Thank you.

Senator FERGUSON —I have one final question to you. We spent over an hour talking about tobacco and the effect of compensation for tobacco. Figures that I have suggest that in the lowest two quintiles the total amount spent on tobacco on average is two per cent or 2c in every dollar of household expenditure. I know that Senator Conroy has been very keen to defend the tobacco companies, and I am quite sure that Phillip Morris will ask him to the Grand Prix again next year and there is a fair chance that mine will get lost in the mail. Are you aware of those statistics that suggest that, in total expenditure, only 2c in every dollar is spent by the lowest two quintiles on tobacco in household expenditure?

Prof. Harding —I should be, but we wrote that report a few years ago now, sorry.

Senator FERGUSON —That is my only question, Mr Chairman.

Senator CONROY —The government claimed in its tax booklets which it posted out to every Australian that we are going to get all these benefits. It said that if you look at this cameo and that is you, you are going to get $30, $40 or whatever it was. Your analysis, even under all of the government's favourable assumptions such as the 100 per cent pass-through and excluding tobacco and housing, shows that the government has exaggerated all of its claims in terms of those benefits. As examples, it will be $5.58 a week for single income groups. That is option 3b, case 4. It will be $8.67 for two income couples with a kid. It will be $7.65 for two income couples with two kids. It will be $6.17 a week for two income couples with three kids and $10.45 for age pensioner couples, et cetera. That is under all of those favourable assumptions. A single age pensioner gets 37c a week from the package—
that is, option 3b, case 22—while a couple on $150,000 a year gets a tax cut of $145—that is, option 3b, case 11. And this is a fair package?

Prof. Harding —It is not my package. As I just commented, in the original package the apparent percentage gains from the package generally increase with income increases. It varies with type of family, whether or not you are single income, whether or not you have very young children and so on. But, on the whole, that is the impact of the package.

Senator CONROY —But even under your revised assumptions, which is your 3b, there is still that massive differential between somebody on $150,000 who is getting $145 a week and a pensioner getting 37c.

Prof. Harding —Yes.

Senator CONROY —That is under your revised package revised assumptions, not Treasury's?

Prof. Harding —Yes.

Senator CONROY —I want to go to compliance costs very quickly. At page 37 of your report you say:

These increased compliance and administrative costs—

you refer to the increase in compliance and administrative costs of exempting food—

must mean a greater burden on households than we have been able to estimate in this study. This will manifest itself as reduced GST revenue (net of administrative costs) and the prices of goods and services increasing by more than we have estimated in this report.

What is the estimate you have made in your report of the reduced GST revenue and the increased price of goods and services from the compliance of administrative costs associated with the GST itself?

Prof. Warren —In that analysis there is no attempt to identify compliance costs that might arise and to put those back in. In terms of the compliance cost issue, if you have a tax and you pull more items out of the base a la food, you are obviously going to increase your compliance costs.

Senator CONROY —That is an increase on top of a massive increase.

Prof. Warren —Yes, whatever the compliance cost is. That is why international comparisons are extremely difficult. But if you have a tax in a situation and you change the tax status of one of those items, you are simply going to increase your compliance costs. I think that is quite uncontroversial. There will be increased costs full stop. The point that we are making is that in this analysis we take no account of whatever those costs might be. I have talked to some of my colleagues where I work about this, and I know Mike Walpole has appeared before the committee. The point is, and I have raised this point with Mike, that there is no question that as you pull items out of the base you will add to your compliance

costs. The question is how significant that will be. That is extremely difficult to ascertain, but what it will mean is that the benefits that we have ascribed to these individuals as a consequence might not be realised because there will be the added costs relating to compliance.

Senator CONROY —I absolutely accept your point, but that is the tiny bit on top of the massive bit. If you are prepared to chance your arm that there are going to be increased prices because of this little bit on the top, where are you chancing your arm on the massive costs in increased prices from the GST as a whole? You have not modelled that, but you are prepared to have a slash at the Democrat position.

Prof. Warren —No, I thought this was just a factual observation, rather than wanting to make some observation that was party political. It is simply that there will be increased compliance costs. The Taxation Office has indicated from their analysis that 300,000 to 400,000 businesses will have added compliance costs—that is not going to be costless and needs to be acknowledged.

Senator CONROY —It is more than that; it is 1.6 million.

CHAIR —This concludes your formal evidence before us. We thank you for the work you have done and for your willingness to appear and answer questions. After lunch we will resume with ACOSS, but we might have some further questions for you, if you are able to remain, after we finish with ACOSS and Treasury. I cannot, at this stage, possibly give you any idea of when that might be. I appreciate that difficulty but, if you can remain, there may be other questions from the committee. But, for the moment, thank you very much.

Proceedings suspended from 12.50 p.m. to 1.38 p.m.