Dear Robert,
Thankyou for your charming letter. It's always a pleasure to debate with someone so measured and courteous.
I have had an interest in the comparative merits of reduced rates and raising the income tax threshold for a long time. My first published piece on the topic was the income tax and NI section I contributed to in the 1993 IFS Green Budget (p80ff) where I was arguing against John Major's introduction of the 20p reduced rate. In the 2000 Bow Group Ideas Book (p20ff) I proposed a target that, within a Parliament less than half of the population should pay income tax. I have also argued repeatedly that the basic rate should rise to 25p - that the reduction to 20p was a mistake.
I share this history with you to emphasize that this is not simply a debate that is contingent upon time and place - all about what one should do now, given current resources, for those today earning between £10,000 and £12,500. This debate concerns a long-running issue of principle.
The issue of principle has two key components. The first is this: relative to a rise in the threshold, a reduced rate always means the poorest paying more tax, for the same exchequer cost of the tax cut. Those that advocate reduce rates of income tax at the bottom often present their proposal as if it meant the poorest paying less tax, but let us be crystal clear: relative to a rise in thresholds, a reduced rate always means the affected poorest paying more tax. If the same amount is spent raising thresholds or introducing a reduced rate, there will always be someone that pays no income tax if thresholds are raised that would pay some income tax if there were a reduced rate.
I emphasize this point - I bash it and bang it - because there is so little comprehension of it. Saying "let's introduce a reduced rate" sounds like a pro-poor measure, but it never provides as much tax reduction for the poorest as does a rise in thresholds for the same money. It is not dishonest or distortionary to point this out. And in Labour's case the matter is quite explicit. Labour says its priority would, from here, be to introduce a 10p rate band, whilst the Coalition says its priority will be to raise thresholds. So Labour's policy priority involves more tax for the poorest than the Coalition's policy priority.
That brings us to the second point of principle: is it better if the poorest pay some income tax than none, so that they have a stake in public spending and tax rates? According to the advocates of reduced rates, raising thresholds involves a kind of trade-off - with higher thresholds we trade off having the poorest pay no tax (and hence having more money) for having the poorest pay no tax (and hence having no stake - on this story).
Robert is admirably straightforward in this - he argues in terms that it's better for the poorest to pay some tax than for them to pay none. I agree with what he's getting at, but think he frames it wrong and comes to the wrong conclusion. Let me explain.
First, Robert suggests that in my Telegraph article on the 10p rate I offer no principled basis on which to set personal allowances. That's not right. I stated:
"The right way to think about fair income tax is the following. We should not
impose income tax on the necessary minimum income to live. Above that
income, each of us should pay in proportion, a percentage of our income
above that minimum necessary to live. At some point, our incomes become very
high, such that further income services only luxuries.
"Perhaps it could be considered fair to have a higher rate of tax apply once we
are luxury consumers. But even then, tax rates should not be so high that we
work mainly for the state rather than ourselves; no one should pay more than
half their income as tax."
Here I had in mind my discussion of the application of proportionality to taxation in my 2011 Policy Exchange Paper (p25ff) On Fairness (which I am immodest enough to say everyone in current politics should read closely). Note something about this, though: I don't claim we ought to set the personal allowance no higher than the minimum amount necessary to leave; just no lower. But let's assume that we are setting it at that level for now.
We could unpack the principle of the "minimum necessary to live" in a number of ways. Some people might, for example, think the minimum necessary to live implied no-one should be receiving income-related benefits. But one fairly natural interpretation would be the income one derives from full-time work at the minimum wage, or at present £12,875 as per the latest Lib Dem personal allowance proposal.
Robert says that if we exclude those below some level from paying tax, they will not feel a stake in income tax decisions (as per his Nigel Lawson quote) or spending decisions (as per his free-riding meal). On this I would note two initial points. First, on the face of it this would appear to be an argument for having no personal allowance at all. Second, it would appear to be the claim that the poor cannot be relied upon to take public-spirited decisions about the higher rate (the 40p rate), and thus would appear to be an argument against that, also. Indeed, we can go further. Since the poorest may only pay the 10p rate, why would they have a stake in the 20p basic rate, and hence can they be relied upon to vote even on that? It seems like an argument for one flat-tax.
More fundamentally, I don't see why the stake of the poor in spending decisions must rest upon their paying income tax. For one thing they pay VAT and excise duties, and see the consequences of spending rises driving tax rises there. More importantly, they pay National Insurance - and that is key.
In my 2000 Bow Group Ideas Book contribution I proposed the following:
"The government could replace the current NI system with a series of State-Owned Enterprises (SOEs) for each existing area of welfare, from which individuals could purchase specific contractual provision of, e.g., pensions and unemployment insurance, by making contributions to separate central funds administered by each relevant SOE. This scheme could eventually embrace almost all areas of welfare provision [i.e. including Health]. Individuals or (in certain cases) the State would fund a compulsory Minimum Level of Provision (MLP). Individuals could purchase provision above the MLP from the SOE if they so chose. Entitlements would be secure with a public service-provider...Those not paying income tax would still feel involved in public spending through the MLP under the SOE scheme."
The original Beveridge concept of National Insurance rightly understood it as entailing not simply charitable provision by the state but, rather, earned entitlement by contribution. The stake the poor should feel in public spending decisions is that public spending increases should mean the social insurance payments deducted direct from their salaries (like PAYE and NI today) would increase. That is how to make them see consequences. Higher health spending should mean their health insurance premia rise. Higher unemployment benefits should mean their unemployment insurance premia rise. Of course, higher premia would also mean higher income-related benefits paid to them, but under Robert's scheme higher income tax means higher benefits to the poor anyway.
Income tax isn't the right way to involve the poor in public spending. They shouldn't pay at all. Income tax is a general payment into a general pot and should create no sense of entitlement since it buys nothing. The poor generally regard income tax as a straightforward impost. Insofar as it has a societal aspect they think of it as creating a general entitlement to help - a damaging concept.
Income tax should be (at a maximum) restricted to those that earn the minimum amount we as a society consider necessary to live on. That means a personal allowance of at least £12,875 and perhaps much higher. Above that threshold, we should have at most two rates: a rate on ordinary levels of consumption and a rate on luxury levels of consumption, and neither of these should mean anyone paying more than half of income as tax. The poorest should be involved in public spending decisions through National Insurance or (better) more specific social insurance schemes.