Following up on the snippet below, I’ve done an estimates of the impact of the proposed Free Trade Agreement between Australia and the US, as it affects merchandise trade. My bottom line is that Australia loses about $1 billion per year, of which $800 million is a net transfer of tariff revenue to the US and $200 million is the deadweight cost of replacing the lost tariff revenue, net of reductions in the domestic deadweight cost of tariffs.
This estimate may be contrasted with the most optimistic one publicised before the FTA was signed, a benefit of $4 billion over 10 years or $400 million per year, IIRC.
I’d like to thank everyone who’s contributed to the debate so far, notably Harry Clark, Uncle Milton, John Humphreys and PM Lawrence. The discussion has helped me to sharpen up my own thoughts. Obviously what I’ve done is very preliminary, and I’d welcome more comments.
Here’s a sketch of how I worked this out
h5. Assumptions
Total Aust imports subject to tariff $100 billion
US share $20 billion
Total Aust exports to US $10 billion
Tariffs removed on 50 per cent of Australian exports
Average tariff rate 6 per cent in both countries
US elasticity of supply to Australia 2
Australia elasticity of supply to US 2
Rest of World elasticity of supply to Australia 2
Demand elasticity Australia -0.5
Marginal deadweight cost of taxation (including tariffs) 25 per cent
h5. Equilibrium price changes
No change in world price
No change in domestic US prices
US and world goods not perfect substitutes, hence elasticity of ROW supply is relevant as modelled
Change in Australian price can be approximated by change in tariffs on US goods multiplied by US share, giving a 1 per cent reduction.
Price received by US suppliers rises by 5 per cent
h5. Equilibrium quantity changes
US exports to Australia increase by 10 per cent (= 2 per cent of Australian consumption) or $2 billion
Australian consumption of imports increases by 0.5 per cent ($500 million)
Rest of world exports decline by 2 per cent (=1.5 per cent of Australian consumption) = $2 billion
Australian exports to US increase by 6 per cent ($600 million)
h5. Revenue effects
Australian tariff revenue reduced by $1.3 billion of which
– $1.1 billion accrues to US exporters
– $0.2 billion to Australian consumers
Tariffs paid by Australian exporters reduced by $0.3 billion
Net transfer from Australia to US is $800 million
Resource allocation welfare effects (net of transfers)
Gain in consumer surplus from reduction in tariffs paid by Australians is 25 per cent of $200 million or $50 million. Additional deadweight loss from standard trade theory is bounded above by $50 million. Welfare cost of replacing tariff revenue is 25 per cent of $1.3 billion or about $300 million. Net loss of welfare $200 million.
h3. Total impacts
h3. Transfer to US = $800 million
h3. Resource allocation effect = $200 million
h3. Total welfare loss = $1 billion.
John Quiggin 
Sorry to respond without cross checking my facts.(
Got to race off somewhere soon)
So I’ll just cut and paste and ask where the last sentence tits into the model?
“Andrew Stoler, executive director of the Institute for International Business, Economics and Law at the University of Adelaide and former deputy director-general of the World Trade Organisation, on the Australia-US FTA:
Having just concluded FTAs with Singapore and Thailand, and embarked on the study of an FTA with China, is there really any reason to think that Australia would be punished because it negotiated the AUSFTA?
Just recently, an Indonesian minister, hearing of the FTA results, suggested that his country might be next in line for an agreement with Australia. China appears to be serious about an FTA with Australia. Many speculate that Beijing sees it as a testbed for other bilateral deals, including one with the US.
The idea that the AUSFTA has distracted Australia from the Doha round, and that this is why the round is in trouble, is almost too silly an argument to consider.
I have spent most of my adult life in Geneva, and I can confidently say that anyone trying to use what happened in the AUSFTA on sugar as an indication that Australia and the US are willing to settle for no meaningful results on agriculture at the World Trade Organisation is someone who purposefully, or through ignorance, projects a misunderstanding of Australia’s position in Geneva…
Far from being left out of the deal, most Australian agricultural sectors should do quite well under the AUSFTA. Far too many people are quick to forget that in this modern Australian economy, nearly three-quarters of people work in the services sector where the AUSFTA clearly promises more competition and cost savings in Australia, and enhanced access for services exporters to the US.”
Oppps sorry, just noticed your qualifier “as it affects merchandise trade”
John
2 questions.
If only half of US exports to Australia are subject to tariff cuts, then that is $10b of exports. Doesn’t that mean the tariff revenue loss is 6% of $10b = $600m, not $1.3b?
With demand inelastic and supply elastic, as you have assumed, doesn’t that mean that most of the benefit should go to consumers rather than producers? (In the small country case, with perfectly elastic supply and fixed world prices, Australian consumers bear the burden of the tariff dollar for dollar.)
Take these points together,
Aust tariff revenue down by $600m of which
$100m accrues to US exporters
$500m accrues to Aust consumers
tariffs paid by Aust exporters reduced by $300m
Net transfer to Aust from US = $200m
Gain in consumer surplus 25% of $500m = $125m
Welfare cost of replacing tariff revenue = 25% of $600m = $150m
Total impacts
Transfer to Australia +$200m
Resource allocation effect -$25m
Total welfare gain $175m
Obviously, I’m not allowed to comment in detail right now. Though I might say that you’ve highlighted the importance of using a proper CGE model. 🙂
I suggest you wait for our numbers, and then change them as appropriate (given your preferred set of assumptions).
Milton, I meant to say half of all *Australian* exports benefit from tariff reductions. Fixed now, thanks. Also, your analysis has the distribution of benefits between Aust consumers and US exporters the wrong way around. If you switch it, you still get a net negative.
John, will your model results be published in a way that allows for the kind of adjustment you suggest? This doesn’t seem to be possible with the earlier report to DFAT or with the competing models results published by others. As I noted earlier, this is one reason I am not convinced that CGE models are an appropriate tool when the main concern is to estimate the aggregate impact.
For info, the main difference between our report and the ACIL report (written by Greg Cutbush) was that Greg dismissed any benefit from services liberalisation and had different assumptions about the armington elasticity. Cutbush basically assumed that Australia had significant control over world prices – and I believe he has since backed away from those assumptions. The reasons for the difference between the CIE and ACIL answers was outlined in the CIE’s March 2003 report ‘comments on the ACIL report’, which is available at the CIE webpage (http://www.thecie.com.au). I appreciate that this isn’t common knowledge – but that’s not due to lack of effort on our behalf (I guess some people just find us trade economists boring) 🙂
btw, I think your $1.3 billion tariff revenue loss number came from a Treasury report. Remember, Tsy generally does the old trick of reporting the aggregate from the forward estimates (ie sum of next 4 years).
I think you’ll find our report provides you with the info you need to make an informed evaluation, even with differing assumptions.
btw, I think your $1.3 billion tariff revenue loss number came from a Treasury report. Remember, Tsy generally does the old trick of reporting the aggregate from the forward estimates (ie sum of next 4 years).
I just applied my estimated average tariff rate to the value of US imports, though I did take some comfort from what I recalled of the Treasury report ($1.8 billion, I think). But if that’s over four years, there’s clearly something wrong.
According to the Budget papers, total revenue from customs duties is about $6 billion per year. Allocating 20 per cent of that to the US would give a number very close to mine.
But about two-thirds of the revenue comes from motor vehicles, TCF, petroleum products, liquor and tobacco, so the general tariff revenue is only about $2 billion, which would give a US share closer to $400 million. Presumably there must be lots of imports that aren’t affected by tariffs.
I was going to read the ACIL comparison, which I saw on your website and will get on to this soon.
John,
You surely know that the merhcandise trade gains were NEVER going to be big (e.g in GDP terms) given the low world-import weighted barriers in both countries (and our long-standing effective discrimination in favour of the USA). So I take it that this back of envelope estimate of merchandise only is mostly a provocation.
Leaving out services is sort-of excusable (given the difficulty of estimation) but, as you’d probably acknowlege, very damaging to credibility. Have we done better in mercantilist terms than the USA on the services balance sheet? Probably.
Also, you’ve omitted the gains from government procurement liberalization (both the mercantilist and own-efficiency gains). Also excuseable, but …
Although even the CIE use of the GTAP model found import prices on both sides fell by about 5% as a result of full liberalization (not what we have as it turns out), the result of the increase in exports in Australia and the USA is GROWTH in production and incomes. Where is this in your back-of-envelope? Your welfare changes seem limited to estimates of tariff transfers. What impact does this growth have on the costs of tax revenue replacements?
BTW, the CIE report was constrained by the DFAT terms-of-reference to UNDER-estimate the merchandise trade gains e.g. they were directed to consider the US beef import quotas as non-binding. In my view they also under-estimated the potential dairy gains by at least half. And I think, for a variety of reasons, they underestimated peanuts, horticulture and wine beneifts, too.
Of course, the dairy gains are now likely to be much smaller (perhaps a PV of 20% of my estimate of $500m).
Best wishes,
Peter
Peter, I said at the outset that this was an analysis of the FTA, as it affects merchandise trade. I can scarcely cover the whole FTA in a single post, and I think it’s unreasonable to attack me for not doing so.
Also the reference to mercantilism is gratuitous abuse. I haven’t said anything about impacts on the balance of trade, which is the main concern of mercantilists, and have focused entirely on standard neoclassical welfare analysis.
It’s pretty clear, if you read my earlier writing on this topic that my main concerns with the FTA are about issues other than the impact on merchandise trade. But I can scarcely write about the FTA without discussing merchandise trade at all. So I thought I’d tackle this relatively minor issue first before moving on to the main game.
John, I can only go on memory, but I heard the Treasury to say that they would lose $1.4 billion per annum in tariffs from US imports. They were explicit in saying that this was based on the past levels of imports. That is, it took no account of trade diversion, the notion that under the FTA we would import more from the US and less from other countries (motor vehicles for example).
Also the benefit from the FTA cited in the earlier report was, I think, US$2 billion with the $A at 50c, hence the $4 billion. I think this was the per annum figure after 10 years.
John,