I will read both policy reports in the future. I know both the policy teams well, they are smart, have integrity, and provide genuinely useful insights. And my interactions with OliverShaw have always been reasonable, so I’m sure that report is of interest also. As a result, I don’t have much interest in giving a knee-jerk reaction to anything until I’ve had a chance to read the work and to educate myself a bit.

The headline results from all the reports sound pretty plausible – my key concern is that people who aren’t the researchers might start talking about them without understanding what the numbers mean. And man, I don’t want to be one of those people!

How can I say all these different results sound plausible when they are all quite different?

Well, how about we chat about effective tax rates a little bit first to discuss how there are different measures – and why they are different!

I’m pretty into the topic (i.e. my studies, my hobbies) and for those who know me I have a more respectable brother who arguably gets even more excited when he hears about this topic. So it is something I enjoy thinking about and chatting about. Lets have a go.

Note: I have not read any of the reports that are online yet – so please don’t read these as comments about any of the work, as it will be out of context. They are comments about me being a nerd.

What is an effective tax rate?

What is a tax rate? People in the media will talk about how I’ll pay a tax rate of 39% if I’m earning over $180,000 per year – but is that the right number? What about tax I pay when I spend at the shops? What about when I pay local govt rates? What am I even paying 39% of?

Well, this 39% is called a marginal tax rate as it refers to the proportion of an additional dollar of income I would pay directly in income tax.

So it doesn’t count other types of taxes I’d pay. And it doesn’t tell me how much tax I’d pay on the first $180,000 earned during that year.

Now what does it look like when I consider the tax I paid during a year from some measure of the income I earned – that tax paid divided by the income earned is the average tax rate. If I was earning $180,000pa at work my average tax rate would be way lower than 39% – in fact using this neat tool it looks like it looks like it would be 28%.

An effective tax rate takes these two concepts and runs with them – it takes into account other bits and bobs of tax, and importantly it will count “government transfers” in some way – either as income or as a negative tax. Look we’ll come back to this point.

What is income, what is tax?

So what is income? Well that’s a pretty hard topic – income is, in principle, going to refer to what “people can contribute from” on the basis of the means they have. There are legitimate debates about people’s ability to contribute from some things that are called income – and personally I find it easier just to think about these topics in terms of consumption. But we have an income tax system, and so these debates are key!

Lets ignore them!

So we are thinking about contributions to government from some amount of income. There are three types of broad income concepts I want to think about here.

• Market income: The income people generate before they receive anything directly from government, or get taxed.
• Gross income: The income people are sitting around with before they get taxed.
• Disposable income: The income people are left sitting with after they’ve paid tax.

There are other terms, taxable income, adjusted gross income, capital income, labour income, imputed income. But for our purposes, who cares.

The first two (market and gross) are a base that we can build tax measures from. One thing that people get minced up on is that they hear “tax rate” and they think there is one measure to rule them all. But for different questions the relevant rate is completely different!

• Intensive work margin: Hey, I am thinking about work a slightly longer shift this week – what is my tax rate?
• Extensive work margin: I’ve been out of the workforce looking after my young child and I’m thinking about jumping into work – what is my tax rate?
• Redistribution of income: I’m worried about inequality in our society, and I wanted to understand how taxes and government spending is reducing this? Can I do this with a distribution of tax rates?
• Earnings period: I’m taking a big risk going out and building up my hairdressing business – and with COVID coming in and out my profits are going to jump around a lot – how does my tax rate compare to one of the hairdressers I hire over the 30 years I run the business?

Effective rates and the choice to do something … usually work

The first two are “financial disincentives to work” – the first one is a “marginal” tax rate (the effective marginal tax rate, or EMTR), the second is an average or participation tax rate (the EATR, or ETR).

For this we want to start with what people are receiving in the starting circumstance (their gross income) and talk about how much is taken away as a contribution (the difference between gross and disposable income). In the first case the person is working and works a tiny bit more, in the second case they aren’t working and jump to work a number of hours. We care about the two measures as they are refering to two different “choices” related to work – the decision to work a little bit more (intensive margin) vs the decision to move from not working to working (the extensive or participation margin).

The tax here refers to more than just the tangible tax contribution. It also includes any abatement of government support – or in english, the amount your benefits are reduced as you earn more income.

Here the tax rate is the ratio of (tax + benefit removed) divided by gross income. This is because this is the proportion of gross income that is sacrificed when you decide to undertake this other choice – or in english, if your boss pays you a given amount more, how much of that amount goes back to the government.

These types of disincentives can also be considered for a much wider range of behaviours – saving, investment, reporting of income – but work incentives are just a great example of how we’d use such a measure!

Effective tax rates and “fairness” – redistribution

The third is a wildly different rate – we aren’t asking about people’s incentives to work, but instead how taxes and transfers transform market income to disposable income. This one is more about “fairness” more generally. It doesn’t tell us what is fair (redistribution is a value-laden issue), but if we coherently want to talk about people paying their “fair share” we want to be building this measure – not the measures of financial disincentives to work.

This third one involves subtracting benefits from tax, and not including benefits in income to start with – so the “tax” is the net payment by the individual to government (tax – benefits) while the “income” base should be the “initial” market income in the absense of taxes and benefits.

The rationale is that this measure of the tax rate is representing the way the tax and transfer system transforms your market income which relates to how much is changes relative to that income.

The difference between disposable and market income is benefit minus taxes, and so the system has generated this by adding benefits to your income, and taking away tax. You “contribute” in a fiscal sense when taxes paid exceed benefits received, and you are “contributing” from the private/market income that is generated from transactions without the imposition of the tax-transfer system.

In terms of “contribution” it may be that many people are not net contributors, but that is by design – “market” incomes are not the demarcation of fairness, and the real question is then whether policy does enough (or too much) to address inequities that do exist in this distribution. In other words, there is no substitute for actually talking about the actual incomes and opportunities that people face and the trade-offs involved.

Furthermore there is a huge wrinkle in these definitions – behaviour. The market income distribution would look very different without taxes and transfers, and in this definition we’ve assumed it wouldn’t change. As a result, it is normally a bit more accurate to say that “the individuals involved in the transactions that lead to market income are contributing” rather than the sole individual who is paying the tax (to capture the idea of tax incidence) – but even then there are shortcomings related to behaviour and counterfactual activities. So we gotta be careful!

How do we get to this fairness measure?

Actually, this is a fun topic – so I’m going to dive into it a little bit.

Now there is a lot of debate about fairness and ETRs in the literature – and the literature often goes in a different direction. By definition the progressivity of the tax system is just a measure of tax paid from an income base, and gross income is a legitimate base to ask how progressive taxes are.

But this isn’t fairness.

If our questions are about whether government policy is “fair” we should be looking at all government policies together, and the way this redistributes income – any half measure of this just doesn’t really tell us anything useful. Yes you’ve stopped a bunch of people having a negative infinitey tax rate when you draw a graph – but you’ve also made your measure treat a dollar paid to someone and a dollar taken from the same person not equivalent!

Ok I need to explain this. Imagine we have a person earning $10. We then tax them $5 and given it back to them. What is their tax rate? If we only looked at taxes on market income it is 50%. If we added benefits in to get gross income it is now $5/$15, or 33%. But their situation – and any view about their contribution – is exactly the same as if there is no tax!!

Fairness is about asking about the change in an individuals situation due to the redistribution tax-transfer system – we then debate if this redistribution meets our “principles of fairness”. Looking at only one part of the system and ignoring the rest to make fairness claims is incoherent.

A “fair” system will redistribute between people based on a set of values – and so if we are using the tax rate to describe this we would want to treat circumstances where the individuals situation is unchanged in an equivalent way. This measure does not and so will give us different numbers for two policies that are exactly the same!

Now, if the goal is to try to look at taxes and benefits in isolation and talk about relative effects, then playing with a variety of income bases is sort of interesting! And if the goal is to comment on distortions in behaviour, and the $5 payment is happening regardless, then this is the damned right measure (hence why I’m using a similar definition here).

I’d even concede that we might have fairness concerns about sharp benefit abatement and poverty traps- which may be easier to understand by looking at the financial disincentive measure.

But if the goal is to comment comprehensively about fariness of the overall tax-transfer system through redistribution, we want to use market income as the denominator and subtract transfers from tax, to ensure we compare policies consistently – by looking at the net amount redistributed to the individual/family as a proportion of the initial (market) income. Anyway, if you are interested read this cool paper (Herault and Azpitarte 2015).

And before anyone swings to the extreme of saying “over 50% of people aren’t net contributors, that is unfair, lets cut taxes” let me give a clear response – yawn. In a world where significant value is generated from the technology and the knowledge we share, it could easily be the case that the “fair” distribution involves even more people being net contributors – and the argument you are making is as incoherent as the situation described above.

To do a fairness comparison we really need to ask “how does the observed contribution compare with how it should be according to our fairness principles”? For this it makes more sense to actually report the distribution of income, and counterfactual distribution if policy was different – what are the outcomes and opportunities for people on the basis of a tangible policy change.

Lifetime contributions

The fourth could be about either incentives or fairness as the person didn’t tell us – but we would want to think about it in a lifetime basis. In New Zealand we don’t tax capital gains (meaning there is income that doesn’t get taxed for the owner in so far as they spend their own time building up the business and goodwill) but we also have a progressive tax scale which taxes “variable” incomes more.

So here, should we be taxing capital gains? If so what part, do we tax inflation? Do we tax the “risk-free return”? If profits are taxed in the future are we double taxing now? Or are we just taxing the bit that is hidden labour income and economic rents?

Hells if I know – the main thing is that if we have a lifetime perspective of the net contribution to government from market income, some of this will wash out, and some important parts will not.

As a result, any discussion on effective tax rates – especially at the top end with a lot of “capital income” – will need to have a discussion of this time dimension. This is where stuff gets super fun – and where we’ll leave todays yarn!

The authors then used this data to ascertain if there was a consistent relationship between the taxes and regulations that businesses in each country face and the amount of investment taking place in each country. There was: the countries with lower tax rates and less onerous regulations tended to have more investment and more foreign investment. The data were considered so useful that the exercise is now repeated annually. One of the original papers by this group of authors, “The effect of corporate taxes on investment and entrepreneurship” (published in 2010) has been cited more than 750 times. 

New Zealand has low levels of capital for a country of its income level and quite high corporate taxes. 

‘Quite high’ is actually an understatement: in 2004 (that year the data used in the 2010 paper were collected) New Zealand was third highest in the OECD (if social security tax payments made on behalf of workers are excluded), and in 2016 it was the highest. At the same time taxes paid by average workers were the second or third lowest in the OECD, slightly higher than Chile but lower than just about everywhere else. 

The basic reason why New Zealand has relatively high taxes on capital incomes and relatively low taxes on labour incomes is that New Zealand has much smaller social security taxes than almost all other OECD countries, and does not have a compulsory contributory superannuation system. Social Security taxes are only levied on labour income, not capital income, and this forces a wedge between the average tax rates on these two forms of income. On average, OECD countries raise about 9 percent of GDP in these social security taxes. New Zealand raises about 2 percent of GDP, from ACC contributions. 

I have discussed the potential consequences of this in more detail here.  There are at least three theoretical reasons to think that social security taxes – or a compulsory saving scheme – may be a good idea. 

1. First, social security taxes should create fewer disincentives than other taxes, for in most countries the more money a person pays in taxes the larger the pension they receive. (This does not have to be the case, however: Ireland has a social security tax system, but pension entitlement depends on time in the country not the amount a person has paid in taxes.) 
2. Secondly, a compulsory saving scheme may reduce long run wealth inequality. 
3. And thirdly, social security taxes are a very easy way of letting the tax rates on labour incomes and capital incomes diverge.

The argument that there is no good reason for the tax rates on capital and labour income to be the same was formalised by the Nobel Prize winning economists James Mirrlees and Peter Diamond in the early 1970s (in two articles: on the non-taxation of intermediate goods and on an optimal linear income tax). There has been almost universal acceptance of the basic principle by tax economists ever since (see the discussion in Chapter 6 here or here). 

They argued an optimal tax system should result in a trade-off between redistributive equity objectives and production-enhancing efficiency objectives. Since capital incomes are unevenly distributed, the former motive suggests higher taxes on capital incomes. Since capital is supplied more elastically than labour (meaning the supply of labour does not change much when taxes change, but the supply of capital can change a lot or can be directed to sectors with low tax rates), this provides a reason for low taxes on capital incomes. 

This trade-off suggests labour income taxes could be higher or lower than capital income taxes, depending on both preferences for redistribution and the technical responsiveness of labour and capital to tax rates. 

Most countries have decided that tax rates on capital incomes should be lower than tax rates on labour incomes. These countries include Norway, Sweden, Finland and Denmark, – countries famous for their low income inequality and redistributive policies – which have invented the Nordic tax system (the Nordic model). 

The Nordic model has the same initial basic tax rate for capital and labour incomes, but labour incomes are subject to steeply rising marginal tax rates whereas capital incomes are not. The reason: the Scandinavian countries do not want to jeopardise the high levels of capital investment that help generate high incomes in their countries, fearing that high taxes on capital incomes would generate capital flight. 

Furthermore, wages are likely to decline if capital stocks decline in response to taxes, so that the burden of capital income taxation partially falls on workers as lower wages. (This is a widely accepted result, and there is recent empirical evidence from Germany and the United States supporting this idea.) The appropriate tax on capital incomes is not zero, as some theorists have argued, but there are few who argue it should be higher than the tax on labour incomes either. 

So why has New Zealand adopted a tax system that is significantly different from those in most OECD countries? Do these differences create incentives that lead to a low capital, low productivity economy? I suspect that they do, and it is a conversation – with evidence – that New Zealand needs to have.

At least two possibilities, and the assumptions that they require, come to mind.

The first is that there may be fundamental technical reasons why it is efficient for New Zealand to have low taxes on labour incomes and high taxes on capital incomes. The responsiveness of labour to tax rates could be much higher in New Zealand than in other countries, or the responsiveness of capital to tax could be much lower. This could mean it is not optimal to have higher taxes on labour incomes than capital incomes.

Secondly, it could be that New Zealand has a higher preference for income redistribution than other countries, so New Zealanders are prepared to have inefficiently high taxes on capital incomes because of equity considerations. (There is not much evidence that this is true, as our tax system is not particularly redistributive – New Zealand has a low top marginal tax rate, for example).

Either way, it would be useful to understand why New Zealand’s tax system looks so different from those in the rest of the world. We can hope the new Secretary is just the person to ensure the analysis takes place.

…distributional analysis is helpful. It helps inform the debate, and … shows how money is allocated by Government around the different income quintiles of society.

HM Treasury’s draft results have now been published. They show that low income households suffered the smallest pre-tax fall in income through the recession.

That is not entirely representative of low-income households’ experience because they were disproportionately exposed to unemployment, but it is interesting to see that upper-middle-income households saw the greatest fall in wages during the recession.

The impact of changes to taxes and welfare payments over the last Parliament has attempted to reverse those costs by penalising those on low incomes with sharp cuts to benefit rates. However, reductions in direct taxation have largely benefited middle-income households.

The Treasury’s draft analysis highlights that:

• The majority of the tax burden has fallen on the top decile of income earners.
• The cuts to tax credits and benefits have more than offset the cuts to direct taxation for low-income households and left them worse off
• Director’s Law held over this Parliament.

Unfortunately, Osborne announced alongside this release that, despite his earlier enthusiasm:

The Treasury will not be producing analysis of this kind for future fiscal events.

Why austerity?
Osborne claimed that his rapid deficit reduction improved confidence across the economy, which caused demand to recover and growth to return.

Why a fiscal rule requiring an overall surplus in every year?

• Paying down debt in time of growth makes sense, so a surplus is required.
• Rules based on cyclically-adjusted measures and forecast targets—such as the present, five-year rolling structural current balance requirement—have huge measurement problems, which makes them a less effective constraint on Government policy.
• He felt that the debt targets, which were anchored to a particular year, were the harder constraint on his actions in the past Parliament. The new requirement for a surplus each year is a similarly hard constraint and is intended to be so to effectively constrain future Chancellors.
• The overall surplus is used instead of the current balance for two reasons: splitting current from capital expenditure allows gaming of the rules, and also leads to an undesirably negative framing of current expenditure.

Ring fences
Osborne was unapologetic about using ring-fences to protect particular areas of Government spending. He characterised them as simple heuristics that clearly set out the spending priorities of the Government.

The review considers whether a numerical fiscal rule might help and makes the point that

..it might weaken the Government’s ‘ownership’ of the debt target, and its preparedness to save revenue windfalls…  It might create incentives for governments to comply with the rule through policies that would weaken other parts of the balance sheet.

In other words, once there is a rule then the game is compliance with the letter, not the spirit, and that can actually weaken fiscal governance.

The review is also lukewarm on the idea of a NZ fiscal council, pointing to

…the operational independence of the Treasury in the preparation of the forecasts and other documents of its responsibility, its well-established non-partisan reputation, its increased openness to outside inputs, and its strong record of relatively (compared to other national forecasters) accurate and unbiased macroeconomic and fiscal forecasts.

How does HM Treasury stack up against that assessment and should we be thinking about its role in fiscal sustainability rather than allocating the entire burden to the OBR?

Fiscal adjustments based upon cuts in spending appear to have been much less costly, in terms of output losses, than those based upon tax increases. The difference between the two types of adjustment is very large. Our results, however, are mute on the question whether the countries we have studied did the right thing implementing fiscal austerity at the time they did, that is 2009-13.

If the ideas about secular stagnation turn out to hold water then we may be hitting the ZLB far more often in the coming decades. Even if that doesn’t eventuate, real interest rates have been trending down for a long time, which has led some people to recommend a higher inflation target to avoid the ZLB in future. The upshot is that, unless there are changes to the standard monetary policy regime–flexible targeting of 2 per cent inflation–high fiscal multipliers and fiscal activism may become a regular occurrence. If it does then the government’s fiscal rule had better work during those times, too.

That has important consequences for the way surpluses and deficits are dealt with. It means that governments tend not to save surpluses beyond their term because they reap little benefit from it. They also attempt to close deficits within the term, which can be too rapid when the deficit is large. The recent recession in the UK is a textbook example of the latter problem. Faced with a deficit exceeding 10% of GDP, the Government sensibly altered their plans to reduce it. However, the five-year planning horizon of their term in office induced them to attempt to close it far more rapidly than would be ideal. The immediate cost was a reduction in aggregate demand just as monetary policy was losing steam. The chart shows two estimates of the cost that imposed on the UK’s output. Over the three years charted that probably amounts to a cumulative £3,500 per household.

The real costs of poor, short-term fiscal rules are twofold: they enable governments to avoid the challenge of long-term fiscal sustainability, and they encourage them to pursue damagingly rapid fiscal adjustments. An effective rule needs to solve both of these problems. It must bring the long-term challenges into focus for the current administration, but it must give the government the leeway to solve them over decades, not years.

I am a big fan of extending the OBR’s responsibilities but there are solid reasons for the Government to demur. Overseas experience shows that giving fiscal councils too much responsibility too soon can jeopardise their existence. The major risks to a young fiscal council, such as the OBR, tend to be changes in the ruling party and changes in the council’s leadership. The Hungarian council, for example, was established by one Government then swiftly neutered by the next when it was critical of the new Government’s plans.

The OBR faces the possibility of changes in the ruling party as early as May, and its charismatic Chairman, Robert Chote, will then have been in the role for nearly five years. Thankfully, the Labour party’s shadow Chancellor, Ed Balls, has been a strong supporter of the OBR in the past Parliament. Nonetheless, the organisation will need to work hard over the coming year to maintain the credibility that it has established, particularly if Chote departs.

That alone might be enough to postpone changes in the mandate until next year but the risks are multiplied by the massive expansion of the OBR that would be required if it were to assess Opposition plans. Civil servants do not advise the Opposition but they are heavily utilised by the OBR in preparing its reports. For the preparation of its reports the OBR uses an equivalent of 125 full-time staff each year and directly employs only 6 of them. The remainder are drawn largely from HMRC and DWP, with a few from HM Treasury. The Dutch CPB, which costs Opposition policies, employs about that many itself and the American CBO employs closer to 250 staff. For the OBR to cost Opposition policy it might need to grow to around ten times the current size. That expansion would seriously challenge its culture and quality; not ideal prior to a general election where its work is being scrutinised with ever greater interest by economists, journalists and politicians alike.

Together, those factors led an independent review of the OBR to recommend that

…caution be exercised in considering the expansion of the OBR’s mandate (e.g. costing certification of opposition manifestos). The OBR may not have the organisational capacity to expand its remit without further drawing on the resources of other government departments. In addition, the particularly narrow legal framework of the OBR and its interdependencies with the executive branch may risk creating perceptions of conflicts-of-interest.

An expanded role for the OBR is a very good idea in principle, but the institution needs to endure beyond one Parliament and one leader before it is ready to tackle such an enormous and politically sensitive task.

Update: For more details on overseas implementation check out Robin Munro’s post at the Institute for Government’s blog.

Getting the debt to GDP ratio to fall at some stage is a good idea, but having a target for a specific year is silly. It is not optimal because if some shock hits the economy before 2016/7 which means debt tends to rise relative to GDP, it is crazy to try and counteract that to meet the target in such a short space of time. It is not effective because it can be gamed by the government fiddling the timing of expenditures.

…

Having a five year rolling target for the deficit allows fiscal policy plenty of time to adjust to shocks. We saw this in action over the last few years, as the Chancellor was able to reduce the pace of fiscal consolidation from 2012 when the economy failed to recover as quickly as he had hoped. Changing this mandate from five to three years gives any Chancellor less time to adjust, which is why it is a backward step.