Summary :- I explore why people feel that tech company taxation is unfair, look at how current structures work, and describe possible alternative models for taxing digital services.
A key issue that will be on the agenda of the incoming Biden administration is the taxation of digital services.
I hesitate to wade into the tax debate as it is a complex and technical area, but it is so important for our perception of tech companies that I am going to attempt to tease out some of the major questions in this post.
I need to start by setting out some definitions as there is a risk in this area that we are not all talking about the same things.
Let’s start with the ideas that tax payments may be either or both of ‘correct’ and ‘fair’.
Correct here means that taxes are being paid in accordance with all applicable laws, and international agreements, as they stand when the payments are made.
Correctness is an objective standard that can be tested by looking at legal codes, protocols and treaties.
There may be different interpretations of laws and treaties which can create doubt about whether particular behaviours are correct, but any disputes may be litigated and this will clarify what is the correct position as a matter of fact.
Fairness is a subjective test reflecting each person’s view on the amount of tax that is being paid in a particular set of circumstances.
I will describe some yardsticks by which we might judge whether a company is making a fair tax contribution later in the post but recognise that there will be quite divergent views on what fairness looks like amongst people reading this.
To illustrate the distinction between these concepts we might rework a phrase used to discuss misinformation – everyone is entitled to their own opinion on tax fairness, but they are not entitled to their own facts about tax correctness.
I am not going to offer a treatise on tax correctness in this post both because this is best left to legal experts and the courts, and because it is fairness that is the main concern at the political level, and is at the heart of the EU-US debate.
Before leaving correctness, I would note that, in my experience, the tax experts employed by publicly listed companies are people of professional integrity who would not advise doing things that are legally incorrect.
There will of course be some bad apples, but I think it is more accurate and useful to start from the assumption that companies have good reasons to think their tax arrangements are legal than that they knowingly set out to break tax laws.
If you think the laws are good and the problem is just that companies break them then you might see the path to fairness as being through getting better at enforcement of existing regimes and plugging any loopholes.
But if you rather accept that companies are generally applying the law correctly, as it is currently drafted, and you still don’t like the outcome, then you will want to focus on where laws need to be changed to make matters fairer – this is, I believe, the more fruitful path.
What Is Fair Tax for a Tech Business?
When we talk about companies paying fair taxes, we are generally interested in how much corporate income tax they pay.
Most companies will generate all sorts of revenue for governments, eg in sales (VAT) taxes, employee payroll taxes and property taxes, but these are not seen as the business making its own direct contribution to society.
We measure a company’s direct contribution by the extent to which it pays a healthy slice of its profits to government in the form of corporate income tax.
This is comparable with our expectations of individuals – that they should contribute to the state proportionately to their own income.
Rates of corporate income tax vary around the world, just as rates of personal tax vary, but most countries sit in a range between 15% and 30%, with the global average for 2020 being 23.79% according to KPMG.
If our test of fairness is whether tech companies are ‘paying taxes like everyone else’ then we might reasonably expect to see them being liable to pay 20-25% of their profits in corporate income tax based on this status quo.
We should note that there are often allowances that can quite legally reduce corporate income tax liabilities from the headline rate, so the final figures for tax paid may be reduced and yet still be both correct and fair in the sense that all companies benefit from the same allowances.
[NB These rates remain politically contentious with some on the left believing they have been taken too low, while others on the right would like to reduce them further, but these questions are obviously not specific to tech companies and beyond the scope of this post.]
Global vs Local
So the first check we want to make is the overall rate at which tech companies are paying taxes on any profits they make.
These rates are typically declared when public companies make periodic reports to stock exchanges so we have a lot of visibility into how much tax they are paying even if it can often be difficult to understand how they arrived at these numbers.
These ‘effective tax rates’ can vary widely for all sorts of factors including market conditions, accounting rules, investment allowances, one-off events etc so we cannot generalise from individual reports.
But the overall pattern appears to be one of tech companies being largely ‘in the pack’ in terms of their global tax rates when compared with other similarly situated businesses.
This may seem at variance with stories of profitable tech companies paying miniscule rates of less than 1%, but these generally relate to tax paid in particular (non-HQ) countries rather than representing the full global picture.
Looking at the overall global rate at which companies are paying corporate income tax is important for one dimension of fairness, but may not satisfy concerns about whether taxes are being paid fairly in any particular country.
We can think of this in terms of countries being ‘exporters’ and ‘importers’, where corporate income tax is largely paid to the government of the country where an exporting company is based.
There may be other tax benefits accruing to the government of an importing country in the form of sales taxes, tax paid by local distributors of the imported products and other support services etc, but the bulk of the profits head back to the company HQ for taxation by the government there.
This division of taxable profits between exporting and importing countries, which is not unique to tech and may be entirely correct under current tax law, is at the heart of the dispute about tech tax fairness.
In simple terms, the fact that a US-based tech company making significant profits on its global operations is paying corporate income tax at a rate comparable with everyone else in the US does not help people in the countries where much of that revenue is sourced to feel this is fair.
Place and Value
The rationale for the current system reflects the principle that business profits should be allocated (and therefore available to be taxed) to the places where the value is being created.
This formula will favour exporter countries over importers where companies are doing something clever that commands a high profit margin.
Things may be more balanced for companies producing low margin commodity products where there is more equality of cleverness between the creation and the distribution of the product.
For example, a manufacturer of generic widgets might only be making a 5% profit margin on its products, while its distributors in each country are taking a similar margin from any sales they make.
In this scenario, an importer government will receive corporate income tax on the 5% profit made by a local distributor for sales in their country, and they may feel this is fair when they know that the exporter government is only looking at a similar taxable profit on the same products.
The model for tech companies may be structurally similar to this widget model – profits accrue to local entities in importer countries as well as flowing back to the exporter country – but the distribution can look very different and herein lies the rub.
For example, a tech company might be making a healthy 25% profit margin on some very clever products it is selling in a country and taking most of this back to their HQ, while the work of the local operation is deemed to be responsible for only 2% of the profit as what it does is more mundane.
When a company in this scenario exports £1bn worth of goods, the importing government can levy tax at the local rate on the £20m of value that was added locally, while the exporting government takes their slice from the other £230m profit that was brought home.
It understandably causes frustration when a government sees another country being able to extract a lot more tax revenue, in this example more than 10 times as much, from sales made within its own jurisdiction.
This uneven distribution may apply even where the trade is in physical goods that have been manufactured locally in the importing country.
We can look at the value distribution when someone buys an iPhone in China as an example of how this works.
Both the customer and manufacturer may be in China where the transaction takes place but because the main cleverness – the design and software etc – happened in the US, this is treated as an exported not a local product.
The manufacturing company used by Apple will be a good source of revenue for the Chinese government – margins may be small but volume is massive.
Local retailers, telecoms companies and other associated services will also be making taxable profits from distributing and supporting the devices.
But the fact that much of the value was deemed to have been created in the US means that a chunk of the revenue for each device sold must be allocated as potentially taxable profit in the US.
Even if Apple, for some reason, wanted to keep all of the value in China and have it taxed by the Chinese government, this would not align with international tax principles as it would effectively be cheating the US government out of their slice of the pie.
The same principles apply in the non-tech sector, so a customer in the US might buy a BMW car that was ‘Made in the USA’ at BMW’s plant in North Carolina.
While the car never leaves the US, it uses design and technology that was developed in Germany, and so some of the value from each unit sold correctly accrues to the parent company creating taxable profits in Germany.
[NB A company might also sell the rights to the ‘clever stuff’, its intellectual property, to an entity outside its HQ country.
This is another complex and contentious area of tax law but is not specific to the tech sector as it is used by other companies with intangible assets like patents and valuable brands.
A key point to note is that these deals generate one-off income events that are liable to be taxed by the exporter government though at the cost of future income streams.]
Is Tech Exceptional?
So if these methods for assigning the value that exporter and importer countries can each tax are well-established practice for all businesses, then why is there such a fuss about tech?
We can look at 4 factors that make these arrangements feel unfair when applied to tech companies even if they have are accepted for other sectors.
- Feast or famine profitability.
- Little or no investment in importing countries.
- Significant societal impact.
- Data, data, data.
And I will use 3 hypothetical tech businesses to help illustrate how each of these factors can play out.
Ads Inc sells £1bn of digital ads each year. The business has costs and allowances of £750m so its profit is £250m. If corporate income tax is applied at a 20% rate we might expect it to pay £50m in tax.
Boxes Inc delivers £1bn worth of goods bought through its online platform each year. The cost of the goods and platform amount to £980m, leaving it with a profit of only £20m. Even though the sales volume is huge, it would only be due to pay £4m in corporate income tax at a 20% rate [NB it also makes money promoting products at margins similar to those enjoyed by Ads Inc but this is less visible to the public than its box-shipping].
Cars Inc takes £1bn of fares through its online ride service. Paying drivers and platform costs amount to £1.1bn so it makes a £100m loss. As long as it is loss-making, it does pay any corporate income tax but instead it accrues credits against future tax liabilities.
Feast or Famine Profitability
When thinking about tax fairness and tech companies we often have in mind the profit margins being made by companies that look like Ads Inc.
These are businesses running platforms where the attention of their users is significantly more valuable to advertisers than the cost of winning and holding that attention.
Many companies try to get into this place where they can make high margins showing ads to a large and loyal user base, and they will often run at a loss for several years before winning, folding, or being acquired.
For those that do make it there appear to be healthy profits to be made with relatively little risk of these drying up once you are established.
So for Ads Inc businesses, if they can survive an initial phase of famine, funded by willing investors, then there may then be long periods of profit feast,
[NB As a note of caution, local newspapers were making similarly high margins on their ad products for many years through to the 1980s and may have felt this would continue forever.]
There is a risk that we take these supersized ads profits and generalise them to the whole sector when the real picture is more complex.
For Boxes Inc, a key selling point is that they can provide goods at lower prices than others so they cannot look for an outsized margin on the physical goods they deliver.
While they are not necessarily in a state of famine (again, after an initial loss-making growth phase) their box-shifting business may have to settle for a meagre diet of profits.
But where they can run digital-only services, like advertising and marketplaces, or offer sophisticated technology services like cloud computing, then they may be able to secure margins much closer to those of Ads Inc.
For Cars Inc, the question may be whether they can ever make a profit at all.
Getting into this business means spending investor funds (which may the product of feasts generated by other tech businesses) to get into people’s consciousness, and daily lives, over several years of profit famine.
This is still playing out and possible end states include an Ads Inc-like feast, which is what investors hope for, or a Boxes Inc-like high volume low margin diet, or even that it proves not to be viable at all over the long term.
If Ads Inc offers a globally accessible portal for buying ads, it is able to make money in countries even where it has no staff and makes no investments at all.
Ads Inc may open offices in some places if it has significant business opportunities and it calculates that local staff will help realise these but it has considerable discretion about where and when to scale up locally.
This option of having little or no local footprint is generally available to businesses where people are paying for something that only exists in the digital realm.
As well as businesses selling ads, those selling entertainment products like music, video and games can follow a similar model and expand into countries without necessarily opening offices or hiring staff there.
For some forms of entertainment there are local regulatory and licensing conditions that may make some investment a condition of market entry, and technical constraints may mean they have to invest in local infrastructure, especially for high bandwidth services like video.
There are some comparators in sectors like news media and financial services which can, with the right technology investment, similarly deliver their products to people anywhere in the world over the internet with little or no local presence.
Any definition of digital-only businesses for a new taxation system would be likely to catch this wider group of businesses than just those following the Ads Inc model, unless it jumped through extraordinary logical loops to avoid this.
Boxes Inc by contrast necessarily has to make significant local investments for its core business of delivering boxes (though not for its ads business on the side).
These investments may be in more out of the way locations than existing retailers and, even where the investment is sizeable, Boxes Inc may not get the same credit for this as a more traditional business.
Boxes Inc may also seek government support when making investments and this may provoke a negative reaction (even if this is support that is seen as a good thing when claimed by other types of business).
It is hard to sustain an argument that the core elements of Boxes Inc’s business should be treated differently for tax purposes than many other local online retailers as the elements of its service are very similar.
Where you might try and define it as special is in its access to and use of data and we shall come onto this later.
There may also be issues related to its market impact if Boxes Inc is operating at a scale way beyond other online retailers, but this is more a matter for competition law than tax law.
Cars Inc does not invest directly into buying taxis and hiring drivers but rather claims to act as an intermediary between many self-employed local drivers and a pool of passengers using its service.
The actual investment in local Cars Inc employees, offices, technology etc is likely to be very light as the profitability of its model depends on the delta between the fees it can charge passengers and all of its costs, including both fees paid to drivers and operating expenses.
The model used by Cars Inc is an evolution of models that have long been used in the taxi industry with networks of self-employed drivers receiving jobs from a central despatch office.
The dispatcher is now an algorithm and the communications are over apps rather than radios, but the whole industry is making this shift not just the ‘tech’ businesses like Cars Inc.
This may though feel different because of the sheer scale of Cars Inc and we shall look at this now as we consider societal impact.
What all of these businesses have in common, whether locally invested or not, whether highly profitable or loss-making, is that they each have significant impact on the societies where they operate.
It is this aspect that perhaps most shapes our feelings about whether they are paying their fair share in taxes.
We instinctively expect ‘big companies to pay big taxes’ in part because we assume they are making ‘big bucks’, but even if they are not this does not get them off the hook.
On the usual profit-based income tax model, Cars Inc is unlikely to pay tax for years as it first makes a loss and then legitimately claims tax relief on those historic losses as and when it moves into profit.
Yet it has a very strong brand, and it is having a massive impact on the local transport ecosystem, and they are clearly doing a lot of business now, so surely they should already be paying up?
Likewise, we look for big contributions from someone who is shifting as many goods as Boxes Inc especially when we see this activity as replacing many other local (taxpaying) businesses.
The sheer scale of Boxes Inc – we see them in our streets every day – does not tally with miserly tax receipts, even if they can demonstrate that the tax isn’t there because they only make thin profits from those boxes.
With the Ads Inc businesses, our concerns may focus on what we see to be novel harms that they are causing to society, and we may be attracted to the ‘polluter pays’ principle.
If we feel that these businesses are creating significant social costs that have to be picked up by public agencies then fairness requires them to make correspondingly large payments to government in the form of taxes.
The perception may be that Ads Inc businesses are spiriting away all of the value while leaving others to pick up the costs and this feels unfair (if you are outside the country where the value ends up being taxed).
Data, Data, Data
A common thread that does bind many tech companies together is that we see them as creating value from data.
We may accept that profits should accrue to an exporter country where we agree that the clever stuff that created the value happened in that country.
When tech companies take most of their profits home this is based on their analysis that the lion’s share of the value they have created is due to the design and software activities that happen at and/or are led from their HQ.
An alternative theory has been developing to the effect that new value is being created when users generate the data which is exploited by the companies, and that this should be acknowledged by allocating more value to countries where the data-producing users live.
The question is essentially whether you believe it is the algorithms (developed in the exporter country) or the data (generated in the importer country) that creates the value and whether tax rules properly reflect this division.
Under this model, a country with a large user base for Ads Inc services would argue that a chunk of the profit that Ads Inc is making should stay in country as it is ‘economic activity’ by those local users that made the revenue possible.
The same argument may be applied to Boxes Inc and Cars Inc where their success depends on being smart about how they use the streams of data they get from their many local users.
If these services are more successful than local retailers and taxi companies then this reflects their ability to exploit local user data, and this should be seen as local value creation not something entirely due to clever stuff done by their HQs.
There is an appeal to this line of argument but we should also recognise that there are other sectors where similar arguments could be made, and that the trend is towards ever more businesses seeking value in data.
In the old model for automobile production, a strong case could be made that all the development and design value should be allocated to the HQ where smart engineers created new products.
When automobile companies are rolling out smart cars that generate data from which the company also hopes to derive value then these start to become more blurred.
If BMW HQ, for example, is entitled to all the value of the designs in cars it sells in the US, does this also apply to any value it extracts from US driver data, or should more of this be allocated to the US for tax purposes in future?
This is not meant to be a “20 reasons why not” argument but rather an example of the implications if we start to move to a ‘local value from local data is taxed locally’ model.
That was a lot of text on whether tech companies are or are not special and why many people feel that they are not paying their fair share of taxes.
It is time to come back to where I started and consider how the EU and US might try to change things to address these concerns.
There are broadly two different routes that could be pursued – corporate income tax redistribution and levies.
Corporate Income Tax Redistribution
If we believe that the distribution of corporate income tax is too heavily weighted in favour of exporting country governments for a particular sector, then an obvious way to address this is to redistribute any existing tax revenue.
Making a change would require there to be agreement on the definition of the types of businesses that would be subject to a new regime, and on any new formulae for allocating taxes from these businesses.
I will assume that a definition can be found that includes my example companies and use these to illustrate how different formulae might play out.
The simplest model is for the the exporter government to calculate the corporate income tax bill for its tech companies just as it does now.
It would work out each company’s liability according to its domestic corporate income tax rates and allowances.
Once it has done that it would divide up the income according to a formula that represents each countries notional contribution to those global profits.
This formula could just look at where the revenue came from: so the exporter government would keep 50% of the tax take if half its sales were made in the home market, a large importer market might get 10%, and a country with a less developed digital market might only be in line for a fraction of 1%.
An alternative model would recognise that more value was created in the exporter country and so give this extra weighting while still moving away from the ‘everything’ allocation that holds at present.
For example, you might decide that half of the value is in the algorithms, to be allocated to the HQ, and half in the data, to be allocated locally.
The exporter government would now get the 50% for its own market plus half of the 50% for goods sold elsewhere, leaving it with 75% of the total tax take while importer countries get to share out 25% between them.
Importer governments may want local tax rates to be applied to revenue in their countries which would mean allocating a share of the profits to each country under these formulae rather than an amount of tax revenue.
This might increase perceived fairness but at the cost of significantly more complexity if it results in comprehensive calculations under local tax rules including factoring in relevant allowances for each country.
Under all variants of this model, a reasonable amount of tax revenue would be likely to accrue to governments where Ads Inc companies have significant business and this should increase perceived fairness.
It would similarly create a healthy income from the marketing and services side Boxes Inc’s of business but would not produce much new revenue from its low margin core box-shifting activities, so people may feel things remain unfair due to a gap between the company’s perceived and actual profitability.
There would be no new revenue from a Cars Inc business unless and until it moves into profitable territory, and we may instead be looking at questions of how tax credits should be allocated for loss-making businesses in the new regime – this would be correct but will not appear fair to those expecting to see new tax income.
The big question is why any country that is a significant exporter in a sector would ever agree to shifting to a model where it has to give away a slice of the tax revenue it is legitimately entitled to under the status quo?
And this is precisely the challenge for governments seeking to move the US towards applying a new model to the tech sector.
The US might be overcome with a mood of extraordinary generosity and international solidarity and decide it is willing to give up revenue unilaterally.
More likely is that the US would not entertain any new regime unless it was confident that it would be getting something in return, and the signs to date are that they have not been sold on the idea that this is anything but a win (everyone else) – lose (the US) situation.
As an illustration of how difficult this is, we can look at efforts to introduce the redistribution of corporate income tax within the EU, a much more tightly integrated community than ‘the world’ or ‘the OECD’.
The EU launched its proposals for a ‘Common Consolidated Corporate Tax Base’ in 2011, then ‘re-launched’ them in 2016, and, oh, it’s now 2020 and there’s not much sign of movement.
The EU proposals are really interesting as a worked-up example of how the kinds of formulae I have described might work in real life, and the lack of progress tells us a lot about how politically challenging this issue can be.
Plan B – Levies
If importer governments feel they are not making progress getting international agreement around new distribution models for corporate income tax, then they may be tempted to take a shortcut by imposing unilateral levies on the activities of digital businesses.
A typical model would ask companies to declare all of the revenue that they have received in a particular country from specific types of activities and then order them to hand over a %age of that revenue as a tax payment.
The activities in scope may only be a subset of everything a tech business does if the intent is to capture the special ‘data-driven’ value rather than other activities which are also common to non-tech businesses.
We would expect most of Ads Inc’s revenue to be caught as it is purely digital and based on the use of consumer data for targeting.
For Boxes Inc and Cars Inc, their full turnover for retail products and taxi rides respectively might not be caught but only the portion of their revenue that they take in as service fees to the people using their platforms.
The fact that these kinds of levies may extract some income from all of the major digital businesses in a country will make them seem like a step forward for fairness where you believe that all large companies must pay.
They do not help us with the core idea that governments should be taking a slice of the company’s profits as they are charged whether or not the company is profitable, and bear no relation to actual profits made.
The hope of those creating the levies is that companies will absorb these new costs at the expense of their own profitability, so that they can at least claim these represent a share of profits even if raised via an indirect mechanism.
Companies may indeed choose to maintain pre-levy prices but it seems rather more likely that they will pass these new costs on to their customers, ie the cost of the imported goods will increase rather than the exporter picking up the tab.
From the exporter country’s point of view, it is preferable that the levies are passed on to customers in other countries as they do not want to see their companies becoming less profitable.
Passing on the levies may reduce the direct impact of these taxes on an exporting government’s tax income, but there may still be negative effects if the newly increased costs of their exported goods have a dampening effect on demand, as you would expect.
The US has threatened France with retaliatory tariffs in response to their proposed levy on digital companies for just this reason – if France is to increase the cost of some US-made products making them less competitive, then the US feels entitled to increase the cost of some French products.
Any legislative moves by importer countries to force companies to pay levies out of their profits rather than passing them on to customers would only increase the losses felt by exporter countries and invite more retaliation.
Left unchecked, you might imagine conflicts of law arising where exporter governments require companies to pass the costs of new levies on to their customer (to protect the flow of incoming profits), while importer countries seek to prohibit them from doing this (so they can say they are taxing the companies not local consumers).
Jaw Jaw or War War?
We are now in a high stakes poker game between the US and countries like the UK and France that have introduced these new levies that are explicitly and intentionally aimed at US tech companies.
The stated rationale of the UK and France for pushing their levies is to force the US to agree to a better long-term distribution of corporate income tax receipts – the levies are a means to an end rather than an end in itself.
They want to keep the scope of the levy to large US tech companies and will be keen to avoid changes that would impact on other sectors where they receive significant tax revenues from global exports by their local companies.
From a US perspective, they might be interested in a new system if there were other areas in the deal where they might see some benefits, ie they would get a new share of corporate income tax from non-US companies with a large footprint in their country.
Absent such a deal, we can expect to see more tit-for-tat moves.
As both the tech companies and the US government (especially under its outgoing management) are relatively unsympathetic complainants, they are unlikely to have many friends in the countries raising the levies, so there is little to hold more governments back from imposing them.
But from the point of view of having an orderly principles-based international tax system, tit-for-tat levies are a poor outcome and we should hope they can be headed off at the pass, or if implemented that they will be dropped in favour of new rules for distributing corporate income tax ASAP.
It is clear that the EU expects the incoming US administration to be more internationalist and this creates an opportunity to improve the tone of the dialogue, but the substance of what needs to be agreed is complex and in an area that is very sensitive for all the governments concerned.
The fact that the EU has not achieved agreement internally on similar questions is a salutary lesson, so, while I am as ever an optimist, we might expect things to get worse before they get better – taxing times ahead.