One of those things that tend to get missed in political debate is that how we tax is and should be a different question from how much we tax.
The first, the correct structure of a reasonable and effective tax system is subject to some basic and well-known rules. Well known among economists at least.
The second is an entirely political question, how much government do we want and what do we want the government to be doing?
Sadiq Ahmed of the Policy Research Institute of Bangladesh discusses some of the details of how the Bangladeshi tax system should change in this newspaper.
His suggestions are entirely and wholly reasonable.
It is worth though taking that step back and looking at the true basics here.
In terms of how we tax -- the how being a different question from the how much? -- we have two basic considerations: deadweights and distortions.
A deadweight is just the flip side of the obvious truth that we get less of whatever it is that we tax.
If we tax cigarettes then fewer people smoke fewer.
If we tax incomes then at some margin fewer people work less to make incomes and so on.
How much less of something we get for the revenue we raise is that deadweight cost.
As it happens, we are rather happy for people to smoke fewer cigarettes so we can tax them heavily.
It is also true that even as we raise the price the number consumed does not go down very much -- this is called the elasticity of demand.
That is just the jargon for how much does behaviour change when prices do?
We want to tax the things that have a low elasticity with respect to price for two reasons.
Firstly, because we lose less of whatever it is because people still do it even when the price changes.
Secondly, this means we can impose heavy taxes upon those things -- if people still do it even when taxed then we do not see the revenue slipping away as people stop doing that thing.
So, given that people like sinning -- tobacco, alcohol although this is less relevant in a Muslim nation, and so on -- we can and should have high “sin” taxes. Just because that is a good place to get money from.
After that, we know that we have a spectrum of taxes when measured by their deadweights.
The lowest is land value taxation -- this is why Mr Ahmed recommends a better property tax system. They are not making land anymore -- taxing it does not stop it from existing, so it is a great place to gain tax revenue from.
After that, with a higher deadweight cost, is a consumption taxation. So, something like a VAT. Then income taxes, then higher still corporate and capital taxes.
Finally, at the top of the range are transactions taxes like import tariffs and financial transactions taxes.
Note what this deadweight really is: an economic activity that does not happen just because we are imposing a tax.
It is entirely true that we might do lovely things with the tax money when we spend it but it is still true that we lose just by taxing.
Less economic activity is also known as “being poorer”.
So, we would like to get our necessary taxes by killing less economic activity. That means no to import taxes and things at that end of our spectrum and yes to land and consumption taxes at the other.
Distortion is a closely allied concept -- the changes in behaviour brought about by what we tax. But here, the emphasis is on exactly and specifically rather than in those general terms above.
So, if we say that income from farming is not taxed but income from wages is then we are going to find an awful lot of people claiming to be farmers rather than being employees.
Or, if capital gains from housing do not pay tax and those from shares do then there will be much more speculation in housing than in the stock market.
The perils of distortion mean that once we have decided upon out basic theoretical structure then all items within a group -- consumption, income and so on -- should face the same tax rates.
That means few exemptions, which means that we can have lower rates for the same revenue. This means that, obviously, we also have fewer of those deadweight costs as our tax rates are lower.
All of this is entirely independent of how much tax we think we should be trying to get.
Or, the other side of that same question, how much government are we trying to pay for? That is an entirely political question.
I have my prejudices here, as will you, and mine is that we desire a low tax and small government society.
It is entirely possible to run a country on 15 per cent of GDP -- this being more than Bangladesh currently raises in tax.
This is enough to pay for the things that government must do and only government can do.
It also leaves enough -- just! -- for a very basic welfare system and so on.
It does not pay for very much of what an awful lot of people would like the government to do but then to me, that is one of the benefits. But your prejudices may well be different from mine on this point.
Once we get past about 20 per cent of GDP then we are not really paying for government any more.
Greater sums tend to be redistribution, taking money off the more heavily taxed richer to give to the poorer.
Whether we do that or not is a political, perhaps an ethical, question -- or as I would put it, a result of our prejudices about what government should be doing.
That question of how much to tax in order to do what does not change those points about the efficient design of the tax system though.
It is still true that we want to tax land, consumption, incomes, capital and corporations, and in that order.
More revenue from those things at the beginning of the list, less to none from those at the end.
How we tax is different from how much we tax and our world will be very much better when we design our system to acknowledge that.
The author is a senior fellow at the Adam Smith Institute in London



