corporation tax is paid on profits, not sales, for a start.
That KPMG’s Chris Morgan felt the need to clarify this point in Tax Journal got me thinking. It seems that, in the wake of the recent spate of corporate tax scandals, there’s a confluence of opinion between the general public and many tax scholars. The first group assume, intuitively, that a company that makes a lot of sales in the UK should have a corresponding tax liability here, and are shocked to find that it doesn’t. Many in the second group (see, for example, the Mirrlees review [pdf]), have long argued that, in Michael Devereux’s words [£], “a good case could also be made for a reform that allocates taxes on profit according to where sales are generated.”
There are two ways in which a move to sales-based taxation might work. One would be, as the Mirrlees review authors advocate, to stop using profit as the basis of taxation altogether, replacing it with a VAT-style sales tax. The other would be to keep using profits, but through a unitary approach, apportioning a firm’s global profits between countries in proportion to its sales. This specific proposal (to rely only on sales) was mooted by some speakers at a Tax Justice Network conference in Helsinki earlier this year (although I can’t find it in the presentations).
Here are some arguments for and against such reforms – some familiar, some my own invention. I’m going to exercise my academic’s right to make a few observations, but not to come down on one side or another. [Edit: I tweaked the list a little bit to make clear that I’m summarising viewpoints, not expressing my own opinion]
The arguments in favour
- Corporate income tax is economically the worst kind of tax. It distorts incentives, discourages economic growth, and has a deadweight loss. Tim Worstall made this case in the comments here on Tuesday. If this is true, the corporate income tax at present is one of those trade-offs between economic efficiency, administrative practicality and fairness…
- …but it’s not as fair as you might think. Companies are not natural people, so the incidence of corporate income tax ultimately falls on their workers, consumers and shareholders. By taxing profits, a government leaves the firm’s management (and ultimately the market) to decide how to allocate that incidence, rather than choosing itself, and some evidence suggests that the incidence falls more on workers and consumers than shareholders (although the evidence is mixed [pdf])…
- …and it’s not that practical either. Because of the complexity of multinational companies today, and the growing porosity of geographical borders to services and assets, corporate income tax on the basis of transfer pricing is too complex to enforce. It would be better to use something less mobile and more tangible to divide up the tax base between countries and make it harder to avoid. One basis to do this would be sales.
- Because of tax competition, corporate income tax rates are falling, and tax bases are shrinking, around the world (except perhaps in countries like India and China?). If that trend continues, then, like it or not, this tax’s days are numbered.
- How about this, based on China’s approach to transfer pricing? Profits aren’t allocated very well by the global economy, so they aren’t a good basis for distributing taxing rights. Currently, China takes into account ‘location specific advantages’ to correct for what it sees as an under-valuation of China’s contribution to a firm’s value-added.
The arguments against
- The public don’t really want a sales-based allocation of tax. Right now, there’s outrage because of the lack of correspondence between some companies’ sales and tax payments. That’s because consumer brands have the biggest consumer profile, so they make the best stories. But what of companies with a large manufacturing or product development base in the UK? Intuitively, public opinion would expect a lot of tax revenue from this too. To survive the ‘intuitive outcomes’ test, a tax on companies needs to reflect the contribution of different parts of the business, including sales, assets and workforce (that argument tends to support a three-factor formulary apportionment).
- Corporations are citizens, legal entities in their own right, not just clusters of individuals. As a result, they have certain rights and privileges, not least those resulting from limited liability. Conceptually, it’s important that they pay taxes on their income in the same way that other citizens do. I remember a letter in the Financial Times a couple of years ago, which argued that the share of a person’s income that is ‘taken’ as tax is better seen as a contribution to the state in return for its role in generating that income – a fee for the education, infrastructure, rule of law and suchlike without which companies and individuals would not be able to generate that income.
- Taxing on the basis of sales will hurt developing countries, because most multinational companies make most of their sales in developed countries, where the consumer market is biggest. Shifting to a sales-based approach would shift the balance of taxing rights away from developing countries, in which many companies have their factories, mines and plantations, but not their sales. (It might be possible to mitigate for this by using a ‘global Value-Added Tax’ levied on export sales from developing countries at the same rate as domestic sales, but this would mean transfer pricing was still with us!)
- Even if corporation tax were eliminated, it would be politically unfeasible to raise other taxes on workers, consumers and shareholders enough to compensate. Similarly, if there were a shift to a sales-based allocation of taxing rights, developing countries would struggle to levy new taxes on businesses to make up for the revenue they lost.
What do you think? Any arguments I’ve missed or mischaracterised? Which do you find most persuasive?
Martin Hearson 
All interesting stuff and thought provoking. But none of these ideas are a panacea and I think they would create as many problems as they seem to solve. There is the interaction with current VAT/sales tax regimes. There will be just as many potential arguments over where sales are “generated”, especially in a world of multinationals, and the increasing economic importance of intangible goods and services (apart from advertising,what and where does Google sell?), intellectual property, brands, licenses and all the rest. (How would the Harry Potter books get taxed?) What role would be played by that old favourite “cost of sales” ? Would they matter? What about businesses that were selling at a loss?
No doubt there could be protocols, formulae and so on to address these sorts of things. (The prospect of having to read all of Mirrlees is not alluring!) I’m not yet convinced it will be a real improvement.
But it’s certainly a discussion worth having. Interested to see what others think.
“Apart from advertising,what and where does Google sell?”
This is a good point! Google sells adverts, to advertisers, but how do you define where the sale takes place? I suppose in the jurisdiction of the purchaser? Or could it be on a server in Luxembourg?
I think what the advocates want is essentially an increase in VAT to replace corporation tax revenues, rather than an additional VAT-like thing.
“a government leaves the firm’s management (and ultimately the market) to decide how to allocate that incidence,”
Err, no. Incidence is decided by the ability of capital to move in or out of the taxing jurisdiction. In an autarky corporate tax incidence will be solely on shareholders. In an entirely free market purely on workers (not that we’ll ever have such perfect capital mobility for as Adam Smith noted, even if the profits of the foreign trade are higher than the domestic some will still prefer to invest at home “as if led by an invisible hand” to favour their countrymen’s interests. Only use of invisible hand in Wealth of Nations BTW).
Please note, incidence does not work at the firm level. It works at the economy wide level. It is not that if we tax BP less then BP will pay higher wages. It is that by taxing the corporate profits of any company at all we lower the wages of all workers in that economy.
This is generally compressed down into the smaller the economy with respect to the world economy and the more open the economy is to foreign capital then the more the corporate tax falls on the workers, not shareholders. Which is where I start getting very angry indeed about some of you tax campaigners. Because you’re arguing for higher corporate taxes on companies investing in Third World or poor countries.
Which are by definition very small compared to the world economy. And given that we’re talking about foreigners investing they must be open to foreign capital.
So you’re all actually arguing for higher corporate taxes in exactly those places where we know that the burden is heaviest on the workers, not the shareholders.
Can’t you all see how insanely stupid this is? Especially when we consider the Atkinson and Stiglitz result, that incidence can be over 100%?
Ah hello Tim, I had noticed the spike in page views from Portugal.
Thanks for this clarification. What I have read on incidence suggests that the practical evidence is more mixed than the models predict, but it was a while ago that I did that reading and I could be wrong. I will come back to the incidence question in a few months and post something then, but one thought for now is this:
I haven’t seen any practical evidence that considers whether the incidence modeling actually works in developing countries. I’m intrigued to see how the dynamics would shake down in a low-income economy where a) most large businesses aren’t producing for consumption in that country, so the impact of corporation tax on prices is external, b) foreign investors rarely pay more than the legal minimum wage to most of their workforce anyway, so there’s little room for a negative impact there, and c) formal employment is a very small proportion of the workforce.
Note that I don’t dismiss tax incidence as a relevant factor for policymakers to consider, I just question some of the assumptions about how, and how much, it should influence their decisions.
“I’m intrigued to see how the dynamics would shake down in a low-income economy where a) most large businesses aren’t producing for consumption in that country, so the impact of corporation tax on prices is external, b) foreign investors rarely pay more than the legal minimum wage to most of their workforce anyway, so there’s little room for a negative impact there, and c) formal employment is a very small proportion of the workforce.”
None of these are relevant.
Taxing profits leads to less capital being invested in the country. Less capital leads to lower productivity of labour. Average wages are determined by average productivity of labour. Thus taxing profits leads to lower average wages.
We can see it in the actual points you make. The minimum wage for formal jobs is of course vastly higher than the informal wages earnt in the rest of the economy. More people building more factories will thus increase the number of formal jobs and thus average wages. Especially as we know very well that external investors usually pay well above prevailing wage rates.
Our problem of course is always Bastiat’s: we cannot see those who do not invest and do not build factories.
We can put it another way of course. Taxing capital in an economy which we actually define as capital poor does seem a little odd really.
Oh come on Tim, you can do better than that.
First, the paragraph of mine that you quote from starts, “I haven’t seen any practical evidence that considers whether the incidence modeling actually works in developing countries,” and your response does nothing to address this. Economic modeling results need to be tested against the real world.
The only practical studies I’ve seen (e.g. Arulampalam, Devereux & Maffini, or Felix & Hines) use data from firms in developed countries, and practical differences in labour markets such as those that I’ve described of course mean that the results of these studies can’t be assumed to hold in developing countries.
Second, I think your summary of the mechanisms through which the incidence might fall on wages is partial. A cursory glance through the above couple of papers shows that they do too.
This all starts from me attempting to summarise a viewpoint that you hold, and I’m grateful for your clarification of that. But the way I want this blog to work is that I’d prefer you to point me to evidence to support any assertions that you make.