Tear down those walls

Jess Irvine wrote a piece on the weekend about what would happen to the Australian economy, and to Australian society, if immigration were shut off. This seemed to enrage a bunch of people on Twitter, provided further confirmation of my surprisingly robust theory that Andrew Bolt is illiterate, and provoked this detailed response from Leith van Onselen at MacroBusiness. The crux of my disagreement with van Onselen is in this paragraph:

While it is true that population growth has boosted overall economic growth (more inputs equals more outputs), the data suggests that it has made individual living standards worse. As shown in the next chart, GDP per capita has plummeted over the past 12 years as population growth has surged. In fact, the 10-year annualised rate of growth has plummeted to levels not seen since the early-1980s and early-1990s recessions:

The second sentence contains what I hope is a typographical mistake: van Onselen’s graph shows that (an average of) the first derivative of per capita income has ‘plummeted’, not that per capita income itself has plummeted. Living standards, if you accept that these can be represented by per capita GDP, have gotten better at a slower rate; they have not actually declined.

First of all, it’s worth pointing out that a long-term decline in the growth rate of per capita income is more or less what you’d expect in a rich country like Australia where the low-hanging fruit have mostly been plucked. Indeed, it’s accelerations in the growth rate that require special explanation; slowing growth is more or less to be expected from a reading of either growth theory or economic history.

 In addition, van Onselen’s causal reasoning is not particularly watertight. Although he concedes that immigration is unlikely to explain all of the decline in the growth rate, he nonetheless asserts that the GDP growth rate figures makes it ‘hard to argue that mass immigration is making us better-off’. This is very crude bivariate causal reasoning: for all we know, the growth rate might have declined even further without the increase in immigration. Simply comparing growth before to growth after provides no indication of the impact of immigration on growth; and it doesn’t, strictly speaking, provide any indication about the sign of the impact on economic growth, either (i.e., whether it’s positive or negative).

What is worse, though, is that real per capita GDP is not really a meaningful measure of the welfare impacts of increased immigration. Immigration could cause real per capita GDP to decrease (in absolute terms, not just in the growth rate) and still make everyone better off. Consider the following hypothetical. It’s unrealistic, but it highlights an important conceptual distinction.

There is an economy, Utopia*, with ten people in it. Each of these ten citizens earns $100,000 a year, every year. Utopian GDP is therefore $1 million (10 x $100,000); per capita GDP is, unsurprisingly, $100,000 ($1 million divided between 10 citizens). A few kilometres across the sea, on the island of Dystopia, is a potential immigrant to Utopia. In Dystopia, she earns more or less nothing. She has few skills, so when she migrates to Utopia, she gets a menial job and earns only $50,000—half of what the other Utopians earn. However, because of positive externalities—larger market size, maybe, leading to scale economies in production, or the fact that the migrant is visited by her family and therefore tourism exports go up—let’s say that native Utopians each now earn $101,000 a year each. Total GDP in the economy increases to $1.06 million, due to the increase in native workers’ wages and the addition of the immigrant’s wages. Per capita GDP, however, declines: dividing $1.06 million between 11 citizens gives us a per capita GDP of $96,364. And yet every single member of the economy, native and migrant, is better off than they were prior to the increase in immigration.

If a Sudanese refugee comes to Australia and gets a minimum wage job in a supermarket (and we assume that job didn’t exist before), then the refugee’s income will have dramatically improved, but Australian per capita GDP will very likely be slightly lower than it otherwise would have. This by itself does not mean that the Sudanese refugee has caused a single Australian citizen any economic harm. In general, you have to assess the welfare effects on the native population and the welfare effects on the immigrant population separately: the correct measure of the increase in economic welfare would be to examine every person’s income and compare it to their pre-immigration income, and add these up.

 Of course, a crucial moral question arises at this point: should you consider the Sudanese refugee’s (dramatically increased) welfare in this calculation? I would say that you ought to, at least to some extent, but some people—I suspect van Onselen would be among them, since he seems to endorse the idea of ‘a frank and honest national conversation about population policy, which focuses on raising the living standards of the existing population’—would rather you ignored the welfare of immigrants. But regardless of your answer to this question, it does not absolve you from the necessity of comparing apples with apples: ‘native’ income without immigration with ‘native’ income with immigration, and migrant income with immigration to migrant income without immigration.

This is quite tricky to do practically, but one thing that has been done quite frequently by economists is to assess the impact of immigration on the wages and employment levels of non-immigrants. The general consensus of the empirical literature on this topic is that the effects on native-worker wages and employment is likely to be only barely positive or only barely negative, if they exist at all.

Of course, van Onselen might respond, I am ignoring the other problems with mass immigration that he mentions in his piece: the strain placed upon limited resources, and other externalities. This is again partly a moral question and partly a practical one.

It is a matter of moral preference whether you think that Australia’s stocks of clean air and clean water (and its stock of high-quality infrastructure, for that matter, much of it built and paid for by people who are now dead)—like our stocks of iron ore and coking coal—are the property of existing citizens by right or whether you think that the unequal division of resources among the peoples of the world is largely arbitrary and often the result of historical patterns of state violence and conquest. If you believe the latter, as I do, then barring immigrants from enjoying these stocks of natural and infrastructural wealth, even at the cost of their partial degradation, is hard to justify morally.

However, even if you believe that Australian citizens have the moral right to prioritise maintaining clean air, clean water and uncongested roads that is ‘theirs’ merely by an accident of birth, then it is not obvious that you need to restrict immigration in order to do this. Part of the reason that Australia faces environmental degradation and infrastructural deficits is that we don’t make people pay for the damage they do to when they use our rivers, lakes, atmosphere, roads and so on. Comprehensive externality pricing—that is to say, pollution taxes, smarter water pricing, congestion pricing on roads—can lift a lot of the pressure on scarce natural and infrastructural resources. They would, of course, still be a good idea even if we didn’t have any immigration. But to those of us who favour high immigration rates, they should be seen as a necessary condition of high immigration into the future, just as a muscular welfare state is more or less a necessary political condition of free trade and investment, even though many people, like me, would support a welfare state in autarky as well.

Would comprehensive externality pricing fix all of the problems that migrants impose upon ‘native’ citizens? Maybe not, but it seems immoral to me to reduce immigration levels by, say, 100,000 people a year in order to reduce the pressure on roads and water supplies if you could price these things in such a way as to achieve the same reduction in externalities as you could by lowering immigration intakes.

With good policy in other areas, there is little reason to suppose that immigration has to have much of an impact on the welfare of ‘native’ citizens; and since revealed preference would tend to suggest at least some positive welfare effects on the migrants themselves, it makes sense to maintain a policy bias towards higher immigration.

***

(* I pick the name because in Thomas More’s fable, it is almost impossible to travel to the island without a Utopian captain to steer your vessel through the treacherously narrow entrance to the bay. I expect this rather alleviated Utopia’s need for an organism with a name like Border Force staffed exclusively by people who failed the police exam, but More is silent on this topic.)


Class warfare, payroll tax edition

I habitually mock journalists who descant upon the theme of ‘reform’ without specifying what exactly they want amended or why—so due credit to Peter van Onselen, who had a go at saying, in moderate detail, what he thought should be done to the Australian tax system.

On the other hand, much of what he suggests is based on some pretty basic misunderstanding of the economics of tax. I don’t have the time to go through it all, but I was particularly amused by his suggestions on payroll tax (boring disclosure: years ago, I did a summer internship in the Tax Economics unit at Treasury and was assigned the thankless task of completing the once-a-decade updating of the payroll tax briefing note). Van Onselen says:

Upping the GST, which is distributed to the states, should be incorporated into wider federation reform, which would see the commonwealth take over payroll taxes, putting them up at the same time as reducing the corporate tax rate. While payroll taxes are seen as a barrier to businesses employing more people, that’s not the case if they go up in line with company tax reductions.

A Commonwealth takeover of payroll tax is actually not the worst idea in the world—it was in fact originally a Commonwealth tax, handed over to the states to allow them to levy their own . Since the 1970s, though, the States have turned what was originally a reasonably efficient tax into a reasonably inefficient one. Instead of levying a low rate of tax on a lot of businesses, the States have historically demanded higher statutory rates of tax from increasingly few businesses, by granting thresholds that exempt many firms with small payrolls from paying anything. As with the ludicrous abolition of death taxes (an act of fiscal vandalism for which you may blame Joh Bjelke Petersen), competition between the states has actually had rather deleterious impacts on their tax systems.

The existence of higher thresholds has some positive economic effects—it reduces the compliance cost of raising tax, since fewer firms have to do any administrative work to pay it. If it’s true that small or young businesses are inherently valuable (because they grow into big firms, for example, or they employ people who otherwise might not be employed, or they increase competition in markets) then the thresholds could also have positive effects. On the other side of the ledger, thresholds distort the labour market by decreasing wages in larger firms relative to wages in smaller firms. (This will probably be levelled out by workers moving from large to small firms, but this is another kind of distortion.) Larger firms are also likely to invest more in R&D and to export (which can be a mechanism of innovation), so taxing large firms more than small ones can have negative effects on productivity growth, too. Most of the economists I’ve talked to about it think that the efficiency cost of currently high thresholds probably outweight the benefit.

A Commonwealth takeover of the payroll tax could help with the problem of states competing to raise their thresholds in order to attract firms to their state. Interestingly, Van Onselen, while he recommends ‘putting…up’ payroll taxes, doesn’t specify whether he means lowering the thresholds, thereby increasing the tax base, or raising the tax rates. There’s a big difference between the two.

But the real problem I have with Van Onselen’s column is the suggestion that payroll taxes replace company taxes.

Company taxes are levied on the profits enjoyed by firms; payroll taxes on the wages that firms pay. So aren’t they just different ways of taxing companies? No, not exactly. As anyone who’s ever studied first year microeconomics knows, the legal requirement to pay a tax doesn’t mean that you actually pay the tax, in the sense that your income or welfare declines because of the tax. Whether you pay a tax in this economic sense depends on the particular characteristics of the market that you’re in. In general, we tend to think that payroll tax is largely passed on to workers through lower wages and possibly to consumers via higher prices.

Who pays the corporate income tax is a much more contested issue, although in a small open economy like Australia’s it’s generally thought that because capital is quite mobile, at least some—possibly quite a lot—is, as with the payroll tax. paid by workers and consumers. However, a lot of us suspect that the Australian economy is also riddled with location-specific rents (for example, the existence of valuable minerals that many companies obtain access to quite cheaply, and the rents enjoyed by the big banks who get privileged market positions because of government policy). If that’s the case, then capital will be less mobile, and the incidence of the corporate income tax will fall more on capitalists.

What’s weird about the article is that Van Onselen seems to understand why payroll taxes are more efficient than corporate income taxes—because they fall on labour that, unlike capital, finds it difficult to relocate:

If major parties are serious about multinational tax avoidance (more politely described as offshore profit-shifting), this reform is the answer. Businesses can’t get around payroll taxes, and the lowering of company taxes may even see profit-shifting work in Australia’s favour, with more companies choosing to base themselves here because of a low company tax rate in a stable democracy.

Why can’t businesses avoid payroll taxes while they can avoid profit-based taxes? Well, because labour is much less mobile than capital. But this very fact is a good reason to believe that the incidence of these taxes falls more heavily on workers. The upshot of all this is that, if payrolls taxes work like we think they do (and are eventually paid by workers, because of lower wages) and if capital-owners assume some of the burden for corporate income tax, then increasing payroll taxes while decreasing corporate income tax means that you’re actually just hitting workers more and capital owners less.

And because capital tends to be owned by people who are richer, the effect of replacing corporate income tax with payroll taxes is regressive  in the same way that Van Onselen’s desire for higher GST in exchange for lower income tax will likely be regressive. Van Onselen’s grand plan for federation and tax reform seems to be: make the poor pay more. That’s a defensible position, I guess, but I he should make that clear, rather than just pretend that he wants to tax companies differently.


Tuesday graphs: Australians work a lot, given the wages on offer

Given that there’s been a bit of debate in Victoria and the Australian Capital Territory about gazetting new public holidays, I thought I’d take a look at how many hours the average Australian works a year compared to the rest of the world. As you can see, in the countries for which the OECD supplies data, there’s a strongly negative relationship between the average wage rate in a country and the number of hours worked a year in that country. Given the average wage in Australia, though, we tend to work more hours than you’d predict (about 86 hours a year over what you’d predict based on our wages).

Screen Shot 2015-08-25 at 12.27.34 pm

The Germans work much less than you’d expect; the Mexicans much more:

Screen Shot 2015-08-25 at 12.31.05 pm

This doesn’t mean that we should or shouldn’t have more public holidays, of course—but it does give us an idea of how hard Australians work compared to other advanced economies.


Links to some recent writing

Marriage & inequality: On Medium (because the graph display is better than WordPress), I respond to Jennifer Oriel’s call for more ‘compassionate conservatism’ centred around encouraging people to get married. I say she got the numbers wrong, and the welfare state is better at tackling disadvantage anyway.

Free trade agreement modelling: In Inside Story, I take a look at the government’s commissioned modelling for the China-Australia free trade agreement and say ‘meh’.

The metaphors of money: In Inside Story, I give a mixed review of John Lanchester’s new dictionary of money-related terms.

Foreign investment in Australia: On East Asia Forum,Shiro Armstrong, Sam Reinhardt and I argue for deregulating Australia’s foreign investment regime for all countries, not just the big ones.


Carbon tax doomsday cult reassembles! (The world ends in 2030 now.)

It turns out they just got the calculations a little wrong.

Four years or so ago, we went through what Australia’s fanatical centrists usually insist on calling a national conversation about carbon abatement policy. National conversations, for some reason, always seem to consist mainly of an extra 500 word allowance on top of Paul Kelly’s weekend column—that, and comically bad reporting from the Daily Telegraph. Anyway, four years ago, there were going to be three million families suffering under the carbon tax. It was going to cripple us, leaving Australia with ‘no one to trade with‘.

It was a tax that would ‘destroy the very foundations of the economy‘, ‘the single most destructive thing’ any government in our history had ever done. And it came into being in 2012—the very year the Mayans had warned of!

The tax was introduced and…alas, the world continued, stubbornly, to turn. The Australian economy continued to grow. Did our prophets renounce their faith and explain their errors to their followers? They did not so renounce. No, you see, they had made a small, regrettable error in their reckoning. Someone had forgotten to carry the one, or something, and instead of the world ending in 2012, we just have to wait until 2030. That’s when the four horsemen are scheduled.

The Daily Telegraph were received new confirmation of the apocalypse yesterday via the infallible medium of Greg Hunt’s office, which supplied details of economic modelling undertaken for the previous government regarding ‘medium’, ‘low’ and ‘high’ carbon prices and their economic impact. Let us too examine what the signs portend. Here is how you will recognise the end of days:

“Economic growth will plummet” Screen Shot 2015-08-10 at 11.10.56 pmI am a trained economist and I initially had no idea what this graph means, but I have studied up on denialist iconography and I shall attempt to expose its secrets for you.

What that graph is seeming to want to tell you is that the economy will shrink by 3% in 2030. (You can’t tell, because they didn’t label the y-axis, presumably to heighten the sense mystery.) What I think the graph is actually trying to tell you is that with a high carbon price, the economy will be 3% smaller in 2030 than it would be in the absence of the carbon price. (Economic growth refers to the change in one year’s output compared to the previous year’s output. It does not refer to the level of output itself, but why let that stop us? The end of days cometh, and our wrathful god hath no need for terminological exactitude.)

I have attempted to translate the Tele’s mysterious symbols into normal human ones for you. Here is the growth rate of GDP with a high carbon price, and without any carbon price at all: Screen Shot 2015-08-10 at 9.46.19 pm Let him that hath understanding count the number of the beast: for it is the number of a man; and his number is Six hundred threescore and six.

“A loss of $4900 in nominal GNI terms per person.”

Assuming you know what GNI means (gross national income), this certainly sounds apocalyptic. I don’t know if I could find $4900 down the back of my couch. How has this number been devised? The Telegraph have added together, for each year between now and 2030, the difference between GNI per capita with a high carbon tax in the modelling and GNI per capita without a carbon tax. It’s a bill spread over 15 years, in other words.

Now, the Telegraph have used nominal GNI, which means unadjusted for inflation; I’ve decided to translate this into actual meaningful numbers by using real GNI. Let’s assume I’m the average Australian, with an average share of real gross national income from now to 2030. What will this cumulative $4900 look like to my pile of money? Screen Shot 2015-08-10 at 8.29.37 pm Behold, he cometh with clouds; and every eye shall see him, and they also which pierced him: and all kindreds of the earth shall wail because of him.

It will cost $600 billion dollars!Screen Shot 2015-08-11 at 8.25.31 am Or, if you read the actual article, $455 billion. Close enough. What does a ‘$600 billion bill’ mean? Well, if you look at Treasury’s GDP figures, and if, as we did with GNI, we add up all the differences over 15 years between GDP with a high carbon price and GDP without a carbon price, we get something like $455 billion. (The $600 billion is uncorrected for inflation, and, hence, meaningless.) This is what that bill looks like compared to the total cumulative GDP between 2016 and 2030? Screen Shot 2015-08-11 at 8.28.35 am Repent ye: for the kingdom of heaven is at hand.


Electricity is $550 a year cheaper. Or $60,000. Or $200. Or whatever.

Pick a number between 1 and 10. Now multiply it by ten. Now multiply it by ten again. Now add some non-negative integer x. 

Write it down.

Congratulations! You’ve just officially modelled the impact of the removal of the carbon tax on household electricity prices. Please send the answer you’ve just derived (justifying all assumptions and showing all working) to Parliament House, where the Treasurer of the Commonwealth of Australia will file it away for later use in an interview with the national broadcaster.

This at least seems to have been the method by which the Treasurer obtained an estimate of the impact of the carbon tax repeal for an average household’s electricity bill as he walked into the studio where he was booked to have a cosy little chat with Fran Kelly, broadcast free of charge to taxpayers by the AB of C. When the topic of carbon pricing came up, Mr Hockey said:

Under Freedom of Information, Treasury released documents last week that showed the electricity prices have come down $550 per household as a result of us abolishing the carbon tax.

The number $550 did appear on a little sheet the Treasurer had in his hand, one that had come from his department. To be strictly accurate, it appeared next to the words “average household costs” — that is, the cost of all goods and services rather than just “average electricity bills”, which the document said ‘would be around $200 lower’ without the carbon tax — and, to be sure, the 9% decrease in electricity bills also mentioned by the document implies, together with the $550, that the annual household electricity bill is now about $5500, instead of the ~$2000 that the average household actually spends, and of course, the $550 seems wildly out of step with the Australian Energy Market Commission’s price impact estimates from last year:

Screen Shot 2015-07-28 at 12.25.12 am

But this is the carbon tax debate, baby! Don’t like the unapocalyptic abatement cost numbers the economists give you? Select your ownlike Greg Hunt did. Inconvenienced by the fact that our trading partners are implementing emissions trading schemes? Evolve an elaborate medieval casuistry like Chris Kenny’s, according to which, if you exclude the countries whose emissions trading schemes are new, and countries whose emissions trading schemes are not connected to the European Union’s scheme, and emissions trading schemes that have been announced but not implemented, and emissions trading schemes that operate in sub-national regions rather than national ones, and emissions trading schemes that operate in gay, atheist Europe, and emissions trading schemes that operate in countries that have been satirised by Sacha Baron Cohen, no ‘serious’ countries are implementing emissions trading schemes. Reckon the price of wind-generated electricity sounds too low? Literally pick any big number you like and use that, like Alan Jones did last week on Q&A. Whatever works for you.

(It’s worth pointing out just quickly that the $200 — the one that would have been the correct figure for Joe Hockey to use in his interview with Fran Kelly — is just the result of a modelling exercise; it’s probably in the ballpark, but it’s an ex ante estimate from a computable general equilibrium economic model, not an estimate derived from empirical observation of what actually happened in Australia. No diss on Treasury, who came up with the number—this sort of thing is very hard to do. The 9% counterfactual is about what Frank Jotzo and Marianna O’Gorman estimated to have been the actual price response in their post-repeal impact study.)

I’d have a little more to say about the slovenly journalistic culture that permits and encourages this sort of thing, but I have to go and file my tax return so I can do my part to prop up Mr Hockey’s budget. I’ve worked it out on a piece of butcher’s paper and I think I owe $500. Or $30,400. Or $13. Whatever.


Black Magic 2: Rumble in New South Wales

About 90 per cent of an iceberg lies beneath its surface. So it is with political commentary. It’s easy enough to open a national broadsheet newspaper after an election and scan the pages filled with phrases like “hostage to entrenched interests” and “politically fabled Labor heartland” and assume that all political commentary is equally and uniformly silly.

But that would be lazy. Each columnist is unique, and by careful examination of the underlying data, using only the most up-to-date and high-tech statistical techniques, we can actually explore the heterogeneity of hacks. Whose commentary is based on nothing more than vague impression and anecdote, and whose is instead based on dodgy mathematics? Today I propose to you that we undertake such analysis. And who better to begin with than an old friend of ours?

Readers of my blog and my Twitter feed will no doubt be aware that, of the hucksters and charlatans that enliven the Australian scene, John Black is my favourite. Perhaps, just possibly, on a good day—maybe depending on which way the wind is blowing through the turbines?—he can be beaten by Graham Lloyd. But on average, in expectation, John Black is the man for me. Read the rest of this entry »