GST and labour income tax

In a recent podcast episode Steven Hamilton and I were taking as given that a consumption tax is a tax on labour income. The logic given by Steven is that the value of work is the amount you can buy, so whether you tax the income when you receive it or when you spend it this is the same thing - nice explanation.

I have had a lot of questions about this since though - specifically why we focus on “labour income” rather than income in general. So I thought it would be fun to discuss this below.

tl;dr

1. Do we want to reduce tax on the normal rate of return?

2. Do we want a one-off tax on wealth?

3. Do we see sufficient benefits in compliance and reduced evasion from a consumption rather than income base?

If so, we want to switch part of your income tax or a payroll tax into a GST.

Note: Steven wrote a great article in the AFR outlining some reform principles and practices for the tax system - I’ll probably blog about a few of the specifics on this in the future, and if you want me to try to twist his arm to talk about one of the topics comment below!

By your logic a GST is a tax on all income

Good point, I can see that if we take the narrative Steven outlined this seems like the natural way of viewing this.

However, a GST does not tax the normal return on capital and in that case it is often termed a “tax on labour” rather than a “tax on income”. In fact, for a tax economist a consumption tax is often viewed as *roughly* equivalent to a payroll tax - but with much better administration and compliance properties.

So how do we get here? A nice academic piece is this by Auerbach, the main tax review on this is this Mirelees Review (specifically the chapter on the taxation of household savings - rather than the consumption tax chapter), and the IFS Green Budget from 2009 is also pretty nice.

But I’m keen to see if I can explain it in my own way - so I hope this is useful!

Lets start with an example.

Imagine you earn $100 gross and you can either consume it now or in the future. You have a discount rate of 5% and can receive bank interest at 5% to compensate - we’ll call this the normal rate of return.

Furthermore, we will assume that the incidence 100% falls onto us for both taxes - so we end up paying the payroll tax (the gross wage doesn’t change) and we end up paying the consumption tax (the goods price rises by the tax). Given the equivalence we will find, we will expect prices and wages to adjust in a consistent way for both taxes - so this assumption just buys us some clarity!

If we tax gross payroll at 10%, then in the initial period you receive $90 and pay $10 in tax. You pop the $90 in the bank and earn 5% interest ($4.50) and end up with $94.50 to spend on consumption.

If we instead tax the gross spending on consumption at 10% (or a GST increase of 11.111% with lots more 1’s ongoing), then in the first period we have $100. We save it and get 5% ($5) giving us $105 to consume. But there is now a tax payment of $10.50, leaving us with $94.50 of consumption in prior prices … exactly the same!

So the key takeaway is that both taxes will end up with the individual consuming the same amount. A payroll tax will give us $10 of tax upfront, and lets the saver consumer $94.50 in the future, while a consumption tax gives us $10.50 later while allowing the same consumption. However, this higher tax payment for the consumption tax is simply the payroll tax with interest added - so in present value terms, the tax being paid and received by government is also the same!

What happens if the interest rate and the discount rate are different?

Nice question. In our example the discount rate didn’t matter until the end - the individual ends up with the same consumption in the second period in both cases. The discount rate we discussed is the social discount rate and captured when the government would prefer to have revenue.

But as soon as people are thinking about when to time their consumption this does matter - but the consumption or payroll tax treatment does not influence this!

At a discount rate of 5% an individual was indifferent between consuming now or in the future for both taxes.

With a payroll tax if you consume in the first period you get $90 of consumption. If you consume in the second you get a present value of $94.5 divided by 1 plus the discount rate - which is also $90.

With a consumption tax if you consume in the first period you get $90 of consumption. If you consume in the second you get a present value of $94.5 divided by 1 plus the discount rate - which is also $90.

If the discount rate declined then in both cases the person consumes in the future, if the discount rate rose then in both cases the person consumes now.

However, what happens if the interest rate is higher than the discount rate in terms of the value of tax and consumption paid?

Lets increase the interest rate to 10%. Here consuming in the second period gets you $99.

With a payroll tax the tax paid is unchanged - it is $10 in the first period. And the final period consumption is $99.

With a consumption tax, the tax paid rises to $11, with has a present value of around $10.48 with our 5% discount rate. And the final period consumption is $99.

What has happened here is that the individual is able to earn supernormal returns from their initial labour income. The payroll tax takes away these funds from the person, and so means that these “supernormal” returns cannot be taxed.

This is why we often say that a benefit of a consumption tax is it only excludes the normal return on capital - and still taxes supernormal returns (unlike a payroll tax). Note: It gets a bit more complicated here in terms of what is generating the return and the ability to scale up and down investments - but we don’t need that right now.

So it taxes more than an income tax

No no no.

An income tax will tax labour income. It will tax the supernormal return on capital. And it taxes the normal return on capital. Due to this broader base an income tax can be applied at a “lower rate”. But it also means that we start taxing consumption in the future more than consumption now.

Lets go back to our example with a 5% interest rate. An income tax will take the initial labour income and reduce it to $90 raising $10 of tax. We then make $4.50 of capital income from savings, which generates $0.45 more tax. In the end, we have $94.05 to use for consumption in the second period - compared to $94.50 in the case of a consumption or payroll tax.

If someone consumed in the first period, then they would receive $90. Using our 5% discount rate, in the income tax case the present value of consuming in the second period is now $89.57, which is less than $90.

So the income tax - due to the taxation of the normal rate of return - will make this person switch to consume in period 1 instead of period 2.

The key distinguishing feature between a flat income tax and GST is that the normal rate of return on capital income is excluded with a GST when with an income tax it is not - in both cases labour income is indeed being taxed.

This can change the timing of when people spend, and may seem unfair when two people with the same earnings capacity end up paying different amounts of tax just because of when they would like to spend.

What about labour supply

At face value there is no difference between a payroll tax and a GST in terms of labour supply effects - in both cases the value of an additional earned dollar is being taxed.

It may be said that a GST taxes the market value of consumption - even when it isn’t financed by labour income. So it may encourage participation from those with other income sources (i.e. government transfers, capital endowments). However, we need to be a bit cautious with this argument.

Government transfers are set up to provide a certain minimum level of income to people, so the participation response to those is due to the target level of income - not the tax treatment itself. If participation was being encouraged it would be because the real value of benefits have been cut. This isn’t about the GST switch, it is about not thinking about the tax-transfer system as a system.

On capital endowments it is true that a GST tax is a one-off tax on existing wealth. So we need to make sure our view on the participation effects of a one-off wealth tax and a GST are equivalent - and to be clear these are one-off effects, not flow effects over time.

Wealth tax?

Noticed all our examples took a person and had them initially get some income from their labour. But what happens when people have some initial wealth?

Say you have a net stock of wealth of $1,000 dollars. The price level suddenly rises. With that $1,000 dollar you can’t buy as much stuff anymore.

That is the one-off wealth tax of a consumption tax, nothing fancy.

No fancy statements about appreciation or depreciation of asset prices - the level of consumer good prices has gone up, without any improvement in firm profitability or returns to labour. There is no liability due to “holding” the asset, you just can’t buy as much stuff from now and for forever.

This additional one-off tax on wealth from the introduction of a consumption tax is a boon for government and has distributional consequences. I realise it is still popular to bash boomers - a view I don’t agree with unlike the many others my age - but the introduction and increase in consumption taxes has specifically redistibutioned AWAY from them.

Often when you hear people talk about not taking workers and instead taxing consumers the discussion is actually about this one-off wealth tax - shifting a tax liability away from income generated by current workers towards prior workers.

And these aspects are not irrelevant given the specific economic changes occurring (i.e. aging) - hence this transition should be considered on its own merits.

What is your point

We’ve had a whirlwind tour of consumption tax equivalences! Hopefully the examples were useful, but what is our key takeaway.

1. Payroll and GST are equivalent when individuals only receive a normal rate of return on assets.

2. Payroll and GST taxes differ in their treatment of “fixed” supernormal returns on capital (GST taxes them).

3. Income and GST taxes differ in their treatment of normal returns on capital (income taxes them) - this means that income taxes tax future consumption more than current consumption.

4. None of these taxes are necessarily “favouring” the treatment of labour at a given rate - instead the argument is about the size of the base.

5. GST adds a one-off tax on wealth into the mix - meaning that there is an intergenerational difference at the time the tax change is introduced.

So there are three key things to think about if we want to do a “tax-mix switch” from a flat component of income tax to GST (noting that the tax-free threshold makes this a little more complicated in Australia than in New Zealand):

1. Do we want to reduce tax on the normal rate of return?

2. Do we want a one-off tax on wealth?

3. Do we see sufficient benefits in compliance and reduced evasion from a consumption rather than income base?

People who recommend the switch generally support 1 and 3, and those that talk about intergenerational equity tend to think about the 2nd.

If you have any questions shoot below! And also feel free to call me out for anything I’ve said that is dumb.