Wednesday, 29 March 2017

The deadweight loss of a pillow tax

For those of you who have missed the news over the last couple of weeks, Auckland mayor Phil Goff has proposed a targeted rate for accommodation providers in Auckland, to be used to fund Auckland Tourism, Events and Economic Development (ATEED) in its role of promoting the city to tourists. At first glance, this might seem like a useful user-pays charge. After all, the accommodation industry receives benefits from the increased tourism, so why shouldn't they pay the costs rather than the ratepayers at large?

The New Zealand Herald has several stories on this topic, including this one from Tuesday:
As part of its annual budget, Auckland Council wants to shift the funding of Ateed from ratepayers to 330 accommodation providers, ranging from backpackers to camping grounds to big hotels.
Auckland mayor Phil Goff says the accommodation sector has profited from the boom in tourism and increased room rates so it was fair it should pay for Ateed rather than ratepayers.
Money saved could be redeployed to fund infrastructure such as roads, which also benefited the tourism sector.
But the accommodation sector says it has been unfairly singled out and should not be the only sector of businesses to pay for the cost of Ateed.
In its submission it says that it gets 9 per cent of visitor spend in Auckland but is being asked to fund 100 per cent of council efforts through Ateed to increase this spend.
On average rates will increase 150 per cent for the affected accommodation providers and in some cases by more than 300 per cent.
The accommodation sector is right in its criticism of this targeted rate. There are two lines of criticism though, and the sector has struck on only the first: fairness. If a commercial business benefits from the activities of ATEED, they should be paying towards its costs. To be equitable, all businesses would pay proportionally to the benefits received, but that is clearly infeasible. Paying proportional to current property value might be a second-best option. Clearly, some sectors (e.g. finance, law firms) would pay vastly more than necessary due to high property values and low benefits received, while others would pay less relative to benefits received. I would argue that including residential ratepayers in this system would be inequitable (but others might argue the opposite, since the activities of ATEED creates jobs that benefit ratepayers).

The second line of criticism, which is missed by the accommodation sector, is efficiency. If the government has a targeted amount of funding they want to raise (e.g. $27.8 million to fund ATEED), then a small tax (like rates) on a wide number of taxpayers generates a much smaller deadweight loss than a larger tax on a smaller number of taxpayers.

To see why, consider the market diagrams below. There are two sectors (A, on the left; and B, on the right). Without the tax, both markets operate at equilibrium (prices P0 and Pa, quantities Q0 and Qa), and total welfare (a measure of benefits to buyers and sellers in these markets, combined) is the area AED in Sector A, and the area FKJ in Sector B. If the government taxes the firms in both sectors a similar amount, we represent this by a new curve (S+tax) [*]. The per-unit amount of the tax is equal to the vertical distance between the S and S+tax curves (the distance BC in Sector A, or GH in sector B). The quantities traded fall to Q1 and Qb. The prices paid by customers in the two sectors increase to P1 and Pb, while the effective prices for the sellers (the price after paying tax to the government) fall to P2 and Pc. Total welfare falls to ABCD in Sector A (with a deadweight loss, or lost total welfare, equal to the area BEC) and FGHJ in Sector B (with a deadweight loss equal to the area GKH).

Now consider what happens if the government wants to raise the same tax revenue, but by taxing only one sector instead of both sectors. This is shown in the diagrams below. Notice that there is no tax in Sector A, and total welfare is maximised at AED. However, in order to earn double tax revenue from Sector B, the government must more than double the tax rate, to the distance LM. [**] The quantity traded in Sector B falls to Qc (instead of Qb). The total welfare in Sector B falls to FLMJ, and the deadweight loss increases to LKM.

Now compare the size of the deadweight losses in the first pair of diagrams (BEC+GKH) to the deadweight loss in the second pair of diagrams (LKM). It should be clear that the size of the deadweight loss in Sector B is more than four times bigger when it is the only sector that is taxed, than when both sectors are taxed. The combined deadweight loss (when you factor in that there would no longer be any deadweight loss in Sector A in the second pair of diagrams) is more than doubled. Total welfare is therefore much lower if the tax is targeted on only one sector.

So, there are both equity and efficiency arguments against the proposed targeted rate for accommodation providers in Auckland. It probably needs a careful re-think.


[*] Strictly speaking, the targeted rate is different to the specific excise tax that is shown in these diagrams. The difference is that the S+tax curve should be curved in towards the supply curve (but not ever quite touch the supply curve) to represent that the average cost of the targeted rate would reduce, the greater number of accommodation nights provided by the industry. However, I have kept the diagrams simple, as this fact makes no qualitative difference to the discussion.

[**] The government needs to more than double the tax rate, because as the tax increases the quantity sold in the market decreases, so simply doubling the tax would not raise enough tax revenue.

Monday, 27 March 2017

Evaluating the fiscal cost of tax cuts

As reported in the New Zealand Herald today, the Taxpayers' Union released a new report today titled "5 Options for Tax Relief in 2017". I haven't had a chance to read the report in great detail, but here's what the Herald had to say:
The average New Zealand worker is paying $483 a year more in tax because income brackets have not been adjusted with inflation, a new report by a right-wing lobby group claims...
In an effort to put pressure on the National-led Government ahead of May's Budget, the report assumes $3 billion is available for tax relief in 2017/18.
The report outlines five different options for how those tax cuts could be divvied out.
They are:
• A tax-free threshold of up to $13,000. The report states this would save taxpayers $1295 a year, for those earning more than $13,000.
• Cutting the marginal tax rate for earnings between $48,001 and $70,000 to 17.5 per cent, and increasing the 10.5 per cent threshold from $12,000 to $24,000.
The report says this would particularly benefit middle income earners, giving the example of a $4000 a year saving for a dual-income household with a combined income of $100,000.
• Cutting tax rates for high income earners by eliminating the top tax bracket and reducing the rate above $48,001 to 26 per cent, and slashing company and trust tax rates to 26 per cent.
Those measures would save a person earning a $120,000 salary more than $4300 a year. A low income earner would get no benefit, while the average earner would save just $360 a year.
• Increasing the income thresholds of each tax bracket without adjusting any of the tax rates.
• Reducing the company tax rate from 28 per cent to 13 per cent, at a cost of $2.88 billion.
The report stated that an "average worker" on $57,000 is paying $483 a year more in tax than they would had income tax thresholds been adjusted for inflation since 2010.
Back in February in another article about tax cuts, Brian Fallow pointed us to this handy tool on the Treasury website: The Personal Income Tax Revenue Estimate Tool. It's an Excel-based tool (it would be even cooler if it were fully online) that allows us to evaluate the impact of changes in the tax brackets or marginal income tax rates (the proportion of the next dollar that would be paid in income tax) on total tax revenue (from personal income tax).

I had a bit of a play with the tool and the assumptions above, and here's what it came up with:

  • The tax-free threshold of up to $13,000 would reduce personal income tax revenue by $3.35 billion
  • Cutting the marginal tax rate for earnings between $48,001 and $70,000 to 17.5 per cent, and increasing the 10.5 per cent threshold from $12,000 to $24,000 would reduce personal income tax revenue by $3.46 billion
  • Eliminating the top tax bracket and reducing the rate above $48,001 to 26 per cent would reduce personal income tax revenue by $2.49 billion (and reducing company and trust tax rates to 26 percent would reduce taxes paid by those entities as well) 
  • Increasing the income thresholds of each tax bracket without adjusting any of the tax rates (as per their report) would reduce personal income tax revenue by $3.46 billion
  • Reducing the company tax rate can't be evaluated using the PITRE tool.
I leave it up to you to decide whether these different tax changes, each costing around $3-3.5 billion according to the PITRE tool, are affordable and worthwhile. The estimates should be taken with some caution however (and the PITRE tool even warns against making large changes to the tax rates and bracket thresholds). Whenever marginal tax rates change there are a raft of effective marginal tax rate changes that affect incentives to work (or not work), that cannot easily be evaluated with a simple tool such as this. However, I do encourage you to download and play around with the tool, especially if you want to see how marginal and average tax rates work, and the effects of changing them (slightly!).

Wednesday, 22 March 2017

Personality traits and field of study choice in university

I found this new paper by Martin Humburg (Maastricht University) published in the journal Education Economics very interesting (sorry I don't see an ungated version anywhere online). In the paper, Humburg used data on over 14,000 Dutch students, and looked at whether cognitive skills (maths ability, verbal ability, information processing ability) and the Big Five personality traits (measured at age 14) affected whether students later went to university, and (for those that did go on to university) their choice of field of study. Measuring these traits at age 14 is important, because it avoided any issues of reverse causality (field of study affecting personality traits). The fields of study are fairly coarse (limited to six categories).

In terms of the first part of the paper, Humburg found that:
all three measures of cognitive skills as well as extraversion, conscientiousness, and openness to experience increase the probability of going to university... Cognitive skills seem to be the primary driver of educational attainment. For example, a one standard deviation increase in math ability is associated with an increase in the probability of entering a university of 7.7 percentage points. These effects are very large, given that only around 15% of individuals in our sample go to university. Of the personality traits that influence individuals’ probability of going to university, conscientiousness has the largest effect. A one standard deviation increase in conscientiousness is associated with an increase in probability of entering university of 1.9 percentage points. While much smaller than the impact of cognitive skills, this effect is substantial and amounts to a relative increase of the probability of entering university of 12%.
So, university students (in the Netherlands) are more conscientious and open to experience that non-university-students, and have greater verbal, maths, and information processing abilities. No surprises there. I note that the extraversion result is only marginally significant (and negative - university students are less extraverted than non-university-students).

In terms of the second part of the paper, here is the key table that summarises the results:

Extraverted students are more likely to choose to study law, or business and economics, and avoid science, technology, engineering and mathematics (STEM). Agreeable students are more likely to study social sciences (excluding business and economics). Conscientious students are more likely to study medical studies (probably due to the entry requirements that keep out low-performers) and less likely to study social sciences (that probably confirms some people's priors). Emotionally stable (least neurotic) students are more likely to study STEM and less likely to study the humanities (more on that in a moment). Finally, students who are most open to new experiences are more likely to study law and less likely to study social sciences (which the author finds surprising - so do I!). The results for gender are unsurprising too, with women more likely to study social sciences, and least likely to study STEM or business and economics.

Importantly, the effects of the personality traits on field of study choice are similar in size to the effects of the cognitive skills (whereas university vs. not-university was mostly related to cognitive skills). The results are fairly robust to the inclusion of additional control variables (parental education, income, father's occupation, migrant status). You might worry about selection bias (since only the field of study choices of students who actually went to university are observed), but I doubt that is a big factor.

On that emotional stability and humanities result, Humburg concludes:
Another explanation can be derived from Tokar, Fischer, and Subich’s (1998) finding that emotionally instable (sic) individuals exhibit higher career indecision. It may therefore be the case that less emotionally stable individuals are more likely to choose the Humanities as they have a weaker link to particular occupations than STEM and Law programmes, which enables these young people to postpone their final career decision.
I'd be interested as to whether our conjoint degree students (who often seem to be studying joint degrees because they couldn't settle on one or the other) have similar personality traits. It would also be good to know, for those students who are non-conforming (in terms of their personality traits compared with others in their field of study), how well they performed academically (and in terms of getting a job, etc.). That additional work would certainly provide some valuable data for careers advisors, which is sorely lacking.

Tuesday, 21 March 2017

The Maxim Institute on dealing with population decline

Last week the Maxim Institute released a new report on regional development in New Zealand. Radio New Zealand reported on it here, but note that they say the report:
predicts populations in many regions will drop or stagnate within three decades.
Actually, that's based on work that Natalie Jackson and I have done, which is available in this working paper (forthcoming in the Journal of Population Ageing, and with an update based on stochastic projections methodology due in the journal Policy Quarterly later this year - I'll talk about that in a later post).

Anyway, the Maxim report (written by Julian Wood) doesn't contribute anything new research-wise, but does do a good job of collating important research on regional development with a particular focus on New Zealand. There are some parts of the report that should be required reading for local council planners, particularly those in rural and peripheral areas where populations are declining. As one example:
When looking at the age composition of population growth this broad-based regional decline is accelerated by the fact that “only 16 TAs will not see all their growth to 2043 at the 65+ years [age group].” In short, in 10 national election cycles (thirty years), the majority of local governments will not only be experiencing population stagnation, but the vast majority will be experiencing far older populations with far fewer people in their prime working age (aged 15-64). This reality means that the vast majority of rural New Zealand shouldn’t be planning for, or counting on population growth as a driver of economic growth.
Rather, as a rural community’s population ages and or declines it will likely come under increasing economic, financial, and social pressure. Fewer people of working age can mean less employment income in a community and less consumer spending and hence less business income. Local government income can also decline as there are fewer people and businesses paying rates.
And this quote from the report pretty much sums it up:
There is a need for a “growth everywhere” reality check.
The reality that population growth is not a given for most of the country has not dawned on many councils (based on many discussions I have had over the years). Many councils still refer to population projections as 'growth projections', which makes no sense whatsoever if your population has been declining for two decades or more!

"But won't Big Project XXXX lead to expanded population growth?" I've heard that one before, and in fact in one of the early population projection projects I was involved in we quantified and accounted for 'big projects', but it turned out later that if we took the projected populations excluding the big project effects we weren't too far off. 'Big projects' are simply business-as-usual for a growing city or district like Hamilton City or Waikato District, but for declining peripheral areas the money would probably be better spent elsewhere. The Maxim report notes:
The overall picture remains, however, that building physical infrastructure alone will be insufficient to economically “restart” a rural economy in long-term population stagnation and decline.
The Maxim report recommends three 're-thinks', which are definitely worth considering:
  • Rethink #1: All regional development goals must be explicitly and clearly stated to enable clarity, transparency, scrutiny and co-ordination. As part of this “regional wellbeing indicators” should be explicitly developed and included in these regional development goals.
  • Rethink #2: Regional development goals need to be ranked and prioritised with tensions, trade-offs, or the subservient relationships between the goals explicitly outlined and prioritised so as to enable evaluation.
  • Rethink #3: New Zealand needs to rethink its sole focus on economic growth, shifting to a framework that also empowers communities to meet both the economic and social needs of their populations in the midst of “no growth or even decline.”
The first two should be obvious for any decision-making, not just in terms of regional development. The third needs policy makers to face up to the reality that population growth is not the destiny of every part of the country. Which is a point I have made before.

Read more:

[HT: Natalie Jackson]