Last week, the Minerals Council Australia (MCA) came up with a KPMG report (download here) that suggested that the newly introduced Resource Super-Profit Tax (RSPT) would lead to many future mining projects being non-viable. This is of course a cornerstone in their scare-campaign about this tax and I had a look at the report to see what they did.
A preamble to this is that the timing and source of this report raises an eye-brow. I first presumed that KPMG must have some very fast modelers in order to be able to come up with a whole report on the effects of a new tax on new mines within weeks of the budget announcement of this new tax. It would be a testimony to how fast markets can generate research if there is a quick dollar on offer for it. However, this appears not to be a case of fast modeling. The research was commissioned way in advance of the budget. This is somewhat extra surprising if you reflect on the fact that KPMG also modeled the long-term consequences of the RSPT for the Treasury, as part of the Henry Review! At the same time, in the last months of 2009, they were apparently already modeling extensions of their RSPT models on behest of the Minerals Council Australia. Large sections of the report are hence devoted to explaining the differences which read a little forced to me. You can smell the guilt.
I doubt any formal codes of conduct were broken in terms of conflicts of interest, but I find it a little dubious that the same modelers are able to sell the advise to the Treasury that the RSPT will have no adverse long-term consequences whilst simultaneously selling advise to the MCA that there might be some negative short-run consequences. It is hard to see how you can conscientiously serve two masters at once.
I have the following quick comments to make about the content before turning to the main issue:
– The report nowhere gives you the actual models and codes used for the calculations. I could not find the code books either on the KPMG website, linked to within the report, or on the website of the MCA. It is hence very much a ‘trust me, I know what I am doing’ piece. Since these were the same guys as that did the calculations for the Treasury, they probably did know what they were doing, but it would be nice to have independent access to the data and models, and I gather from the introduction that the MCA could give this information out if it wanted to.
– The report is highly selective in terms of what it chooses to calculate and highlight. It doesn’t tell us what the RSPT would do to the Net Present Values (NPVs) of all possible future projects, but only talks about the NPV of the second quartile of profitable projects. This is of course because the first quartile will go ahead anyway and the third quartile will probably see increases in NPVs due to the cost-sharing in the RSPT. It loads the dice towards the negative to focus on only 25% of all considered future projects.
– The report leaves out the effects of all existing projects, at least in its headline treatment. As Chris Richardson pointed out in his presentation on June 3rd to the Minerals Council, existing projects will probably start to see more intensive mining activity because of the reduction in output-taxes associated with the RSPT increases the incentives to produce more at existing mines. This is also implicit in the KPMG report, but the finding that in the next 2 years mining activity should increase is not highlighted at all.
– Trying to make the argument that the RSPT leads to high overall taxes by international standards, the report compares effective tax rates under the new regime with existing tax rates elsewhere. The crucial questions are of course which commodities the comparisons are made to. In choosing comparison commodities, the report leaves out oil which is more heavily taxed than the mining industry would be under these changes (but where production and investment in oil exploration haven’t suffered in the slightest despite these high taxes!), but leaves it to a side-note on page 30 to mention this. Since mineral production is becoming more profitable, it is not at all strange to compare the future of mining with the present treatment of oil, since both involve exploration, investment, and production phases.
– The report itself mentions that long-term effects of the RSPT should be positive for mining activity (a similar point is made by Richardson).
– Computed internal rates of return look very healthy for all types of mining under the RSPT, something given little attention to.
– All these highly selective choices already make it clear that the report, and in particular the summary, is indeed not an objective appraisal but a piece of propaganda that was bought for a reason. The newspaper headlines ‘KPMG report shows miners are going to be ok the next 2 years and in the long-run’ clearly is not what the MCA wanted others to get from this report, even though such headlines would be warranted by it.
Then, to the true matters of substance.
The main thing the report does is to calculate how much profit future projects will make under the existing situation compared to with an envisaged implementation of the RSPT. New projects are presumed to consist of a certain amount of start-up costs, ongoing equity costs, and various yes-no decisions along the way as to whether a mine is taken into production. The key assumptions in coming up with Net Present Values of projects surround the costs of equity and what the RSPT will do to the cost of equity.
The main ‘trick’ employed in the report is to envisage new projects as having to borrow its equity at the same rate as the historical returns to equity. Concretely, for 4 out of the 6 minerals looked at, the presumed interest rate that has to be paid on equity is 15% (plus or minus a half percent), whilst the RSPT is presumed to pay back losses at the end of a 30-year old project using the long-term bond rate as an incremental factor.
It is crucial to see the basic sleight of hand involved here: because new projects are constrained to take 30 years (whether they fail quickly or not), the pay-back on losses implicit within the RSPT in their models is going to be almost nothing compared to the costs of equity, simply because one dollar spent in year 1 on equity needs 66 dollars of return after 30 years, whilst one dollar of RSPT off-sets (presuming a 4% long-term bond rate) is worth no more than 3.24 dollars in pay-backs after 30 years! Hence, by assuming huge discrepancies in the costs of borrowing equity versus the value of RSPT off-sets for losses, the report effectively presumes that the government pays back almost nothing in the case of losses, reducing the RSPT to higher taxes in case of positive profits. It is not hard to see how this might lead to large reductions in NPV if you focus on projects with some chance of going negative and some chance of making high profits.
How sensible is this assumption of 15% cost of equity and a 30 year wait until the government pays back some of the costs at very low levels of annual increments? They are completely unrealistic for two main reasons:
- The hypothesized cost of equity essentially derives (via a quick reference to CAPM calculations of the ‘Australian School of Business’, which is hardly an authoritative source on this since the empirical usefulness of CAPM is highly questionable. It tells you something if a report basis its crucial assumption on a nursery-rhyme version of the CAPM) from historical stock-market returns on equity, i.e. on what happened in the last 30 years or so. This is completely unrealistic because part of the reason for the tremendous growth in the stock market in the last 30 years is the huge relative increase in profits which cannot be repeated in the future (profits by design cant go above the threshold of 100% of output). Also, a lot of the stock-market return in recent decades came in terms of price increases, not dividend returns. No one expects the returns the next 30 years to be just as high. Even more importantly though, the stock-market return already contains a notion of return to risk and a sharing in potential high gains. This hence means the 15% is not the cost of borrowing at all, but the historical rewards for investments. For huge diversified companies like mining companies the valid cost of borrowing is not the stock market return, but rather the bond returns that large private companies have to pay in order to borrow money. Those costs are not even close to 15%. If we’d pick a more realistic cost of borrowing, say 8%, then the RSPT is going to look much better compared to existing taxation.
- Its the presumed 30-year wait in the report that makes the RSPT off-sets worthless. If these off-sets were immediate or even just took a few years, or could be transferred to other projects, they would be worth far more to companies. Hence the report’s compartmentalization of the RSPT to individual projects of fixed length is a pure rabbit-in-the hat conjuring trick (and I hedge my bets here by saying that this is how I interpreted the report. Since the code books are not to be found, I am best-guessing what the modelers actually did). As soon as you allow transferability across projects, which is entirely reasonable for any large mining company, then the RSPT is in many cases going to give you higher NPVs compared to the existing tax arrangements.
Hence, as far as I can ascertain, this report is guilty of double-dipping in terms of clientele, uses various assumptions that are far from realistic, and systematically highlights the worst case for the RSPT. In my own mind it carries a big sticker saying ‘some poor competent modeler was told to make up a set of assumptions that would help the cause of a rich client’. I am not at all surprised that the report carries the immortal phrase on page 2 that ‘neither KPMG nor any member or employee of KPMG undertakes responsibility arising in any way from reliance placed by a third party on this report’ which basically means KPMG washes its hands of the use of its report in the media and by opponent of the RSPT. I wouldn’t want to be held responsible for the content of this report either.