My suggestion of a new bank transaction tax aimed at discouraging conspicuous consumption because of its progressive nature has attracted some interest. As I make very clear in chapter nine of The Joy of Tax, I know that replacing national insurance with such a tax would take time, and suggest trialling of this new tax before it could be relied upon to collect the £115 billion that national insurance does now. So, let me make clear, everything I am suggesting here is exploratory. That's what a book of ideas is about, after all.

The logic of my proposal is pretty straightforward. All deposits into and payments out of bank accounts would in principle be subject to this transactions tax. The rate would be low: in Brazil when such a tax was charged the rate was 0.38%. But that was a flat rate: I am not keen on a flat rate because this tax is quite deliberately intended to change behaviour. That requires it to be progressive. That means the rate has to increase for both individuals and businesses, and in that case the issues of avoidance and evasion have to be considered.

The aim then is threefold. First, there is a need to identify all the accounts under a taxpayer's control or accounts would be split to avoid higher rates of tax. This process of identification would not be especially difficult. That's firstly because all banks, anywhere, have to identify the beneficial owners of all bank accounts. The data on who controls accounts is, then, already available. There is a problem with data consistency, I admit. Too many accounts may use passport identification. That's useful but national insurance numbers (yes, I get the paradox) would be better. They are not perfect, I know: duplicates exist and fraud too, but that's the real world. And we already use them for tax when it comes to bank accounts: ISAs are already regulated using them, so the precedent exists. It would also be remarkably easy to collect NIC data for all privately controlled bank accounts: freezing those where data had not been supplied would likely produce responses pretty quickly. The result would be quite quickly produced data on accounts under common control.

Two consequences would then flow from the availability of this data. First, transfers between these accounts under common control would be tax free. This is important: there would be no reason to penalise diversified account holding. Second, it also means that whilst charges on each account would be made at source (logically monthly for the advantage of the taxpayer and government) at each year end data for related accounts would be aggregated and any additional tax due at higher rates would then be assessed, and collected. Logically this would be by direct charge to what may be considered the dominant account, nominated by the taxpayer.

There are, of course, issues arising from doing this. The first is joint accounts. I do not see this as particularly hard  to manage. At any time it seems quite appropriate for a person to be able to nominate one other person with whom they can share a bank account with the deposits and payments in that account being split between the two. Of course anti-avoidance provisions would be needed to prevent, for example, holding joint accounts with children for such purposes. But the principle seems relatively easy to achieve, as does aggregation.

The idea that people will offshore accounts, which several people have suggested likely, can be dismissed pretty easily. First, try to open a sterling account outside the UK now: beyond the Crown Dependencies that is hard. It's getting much harder the world over. Information exchange is closing down the mass market for overseas accounts. And information exchange from the Crown Dependencies is improving. So, in practice this risk is low: those who try to move accounts are likely yo be found out. If banks aiding them are fined as a result the incentive for them to assist tax evaders will be low. I think the tax could be enforced without many difficulties in the UK.

So, what about business, which obviously pays national insurance? Small self employments need to be charged at low rates, not least because they pay little national insurance now. So, these businesses need to be identified as to their ownership. Those under common control would have to be aggregated. Artificial disaggregation would need to be addressed: it is already for VAT, but within this constraint partnerships would for example, be treated separately from its owners. The fact that all self employed accounts would have to be notified to HMRC might have a considerable benefit for the tax gap: I leave that idea hanging for the moment, but all those who have said that such a tax will increase the tax gap are, I think seriously wrong for precisely this reason.

The rate to be paid by the self employed would grow with turnover. I stress, rates would be progressive i.e. banded. This will require year end adjustment but the aim would be to ensure that micro businesses paid little or no such tax, and those like large partnerships would pay a fair rate to compensate for the ending of employer's national insurance.

Limited companies would be treated in the same way, with those under common control (whether by substantially shared ownership or by being formally grouped) being aggregated for assessment purposes. Intra-group payments could be eliminated within the UK. Those outside would be taxed.

Transactions now wholly outside, for example, VAT and stamp duty would fall within this tax, such as house purchases, but I suggest the abolition of stamp duty in The Joy of Tax: I am not proposing double taxation in these cases. Many capital transactions would also be taxed: designing an appropriate system to ensure banks were appropriately taxed having allowed for the possibility of offset of liabilities and assets would be a challenge, because you can be sure some such arrangements would be contemplated otherwise. I do not see them as insurmountable in any way though: there is a default position that everything is taxed in the case of abuse.

Why do it? First because NIC is out of date, regressive and discourages employment.

Second, because merging NIC with another tax would make rates too high.

Third, because this tax will be progressive and we need new progressive taxes.

Fourth, because this tax will reach into areas like banking that are undertaxed because VAT does not impact significantly on them at present.

Fifth, this tax can replace stamp duty.

Sixth, because this tax is likely to aid efforts to reduce the tax gap (there is not that much case in the economy that does not get either banked or stashed out of use eventually). I think this is a major win.

Seventh, because the collection is relatively simple, and for most people will be very straightforward.

Eighth, because I do want to see conspicuous consumption reduced and this tax could help achieve that goal with significantly rising rates at very high levels of spending. This is essential if we are to move to a more sustainable society.

Ninth, this tax is equitable when many are not.

Tenth, because this tax can develop as payment methods do as well.

And finally, this tax would allow for rapid fiscal adjustment within the economy when monetary tools are running out. You could even, theoretically, run negative rates for some (now there's an idea....)

I am open to comments.