As the Chancellor rushes to try and make her sums add up, a number of tax rises are being mooted and tested out in the press in advance of the Budget. Some of these – e.g. capital gains tax and inheritance tax – were quite predictable, with Labour seemingly happy to let discussion gather pace by making pledges specifically not to raise taxes on ‘working people.’ Capital gains arise from sales of assets and so are a return on capital and not on labour; and inheritance tax is due on wealth passed from previous generations, and so not a return on the inheritor’s labour. Both clearly are not earned income for their beneficiary, and therefore – according to the Labour manifesto – fair game for tax rises.
Who pays a tax and who sends the money to the Government are different things
But National Insurance Contributions (NICs)[1] on the employer side are a different matter – even though UK Government is seemingly happy to go on the airwaves to argue that they comply with both letter and spirit of the manifesto.
Some of this relies on a distinction that often puts economists on one side of the divide and legal experts and tax officials on the other. It is one thing to establish legal liability for compliance with tax – the formal remittance of the funds and who the Government will pursue if it doesn’t receive the money – and another altogether to establish on whom the burden of the cost of the tax falls.
The former is a legally important but economically uninteresting question. If the legal question were the most important criterion, we would be arguing that VAT is a business tax because it is the responsibility of business who sell goods and services to pay over the tax to HMRC to make sure that the right amount is transferred. And in fact that’s how it is organised within HMRC for tax compliance purposes.
If the Government really finds this legalistic argument persuasive, I look forward to it increasing VAT on the basis that it is a tax on businesses and not ‘working people.’ There clearly is little chance of that. VAT is obviously a tax on consumption – it’s the purchase of a good or service by the final consumer that creates a liability. The business collects the tax on behalf of the Government, and then remits it to them, essentially becoming a middleperson in this.
In the case of employer NICs, the formal liability clearly falls with the employer – hence the name. But the liability arises from the employment of a person, and it is based on the employee’s income. So there is more than a passing link with the employee and their personal circumstances.
Tax compliance burdens are often decided on the basis of whom it is easiest to pursue for liability or on whom the least burden will fall. The Government could very well ask every consumer to keep a record of all transactions in a year and ask them to submit a self-assessment of their VAT liability. Of course, if it did so, not only would it be very burdensome to everyone in the country, but it would also probably lead to large under-declaring of liabilities. Instead, the Government piggybacks on the fact that businesses have to keep records of transactions for their own purposes, and instead asks them to keep track of just another line.
But this underlines the fact that there are multiple ways of sharing legal burdens for compliance for a particular tax, and those are operational and legislative decisions rather than economic ones. Once they are done, however, they are taken as given. But we also understand that someone has to bear the cost of the tax that’s raised, and who does that – and to what extent – is a much more interesting question.
For taxes on goods and services, it’s fairly straightforward to determine on whom the economic burden falls
And it all depends on whether the seller can increase prices or not.
In technical terms, the key parameter is known as the price elasticity of demand – essentially how responsive consumers are to prices. The intuition is pretty simple: once a tax is put on a product, and in the absence of a price increase, the portion flowing to the seller is reduced. The seller will then look to increase the price by as much as it can to compensate for that.
If consumers are very price responsive, increasing prices might actually reduce overall profits because quantities sold will fall significantly. In that case, the producer’s optimal behaviour is to absorb most of the cost. On the other hand, if consumers don’t respond to price in a significant way, sellers can increase the price to almost or even fully match the tax put on the product.
This is slightly different for a tax like employer NICs
Employer NICs are paid on top of an employee’s gross salary, and form the difference in the National Accounts between wages and salaries (going to employees before personal income taxes are deducted) and total compensation of employees. Of course, for a firm what matters is compensation of employees (including employer NICs) as that is the reflection of the actual direct cost of an employee.
The effect of employer contributions to social security – also called payroll taxes in the literature – is an area which has seen plenty of studies in the circa 80 years of the modern welfare state. The link between employee costs and profits is much less direct, but there clearly is one. But the effects on employees’ wages and on jobs differs across time.
In the short run, firms are quite restricted in how they can react to an increase in employer NICs. Once a contract is in place and a salary has been negotiated, the firm is unlikely to be able to reduce wages to account for higher costs on top of that. Nor can it suddenly change the number of people it needs to operate while maintaining output. So in the very near term, firms are very likely to absorb the higher costs.
However, the story is very different over the long run. Firms have more power to decide as a whole on labour costs – subject to labour market conditions, of course – and they also have the option to adjust how labour intensive their production processes are – again, depending on the exact type of goods or services they provide. One of the effects we might expect to see is a tendency would be for wage growth to be slower than it would otherwise have been as more of firms’ spending on employee costs goes in taxes than before.
These lower wages than workers were expecting have two effects on their decisions as to how they decide to work. On the one hand, workers are made poorer – what economists call the income effect – and therefore are incentivised to increase the hours they work to increase their income. This effect would raise the labour supply.
On the other hand, the returns to employment have decreased relative to other activities (e.g. leisure), which is called the substitution effect and would tend to decrease the labour supply. Both of these effects exist in pretty much any market, and the question then which one dominates.
But those are not the only effects we might see. Firms might choose to pass this increase in costs onto prices if the products they sell allow them to and they can’t otherwise recruit the people they need. If they can’t increase prices, they might have to absorb the increase in costs through lower profits.
Over an even longer time horizon, the higher cost of using employees to produce goods and services makes capital relatively cheaper. Of course not all firms will be able to substitute capital for labour to a great extent, but especially if they are significantly affected by the increase in taxes, they might choose invest in labour-saving technologies. Over the long run, the main effect of this would be to reduce the number of jobs available.
There is also an increased incentive for employees to move into self-employment. This is because self-employed persons don’t attract employer NICs (as they have no employer), and therefore this further increases the wedge between employment and self-employment taxation. However much one can quibble with exact changes in tax policy in recent years, the move towards a lower wedge between the two has been welcome, and reduces labour market distortions. Increasing employer NICs would reverse some of that.
All these effects are real, and therefore the key question and judgement is how significant each of them is.
What does the evidence say, and how would the OBR incorporate these changes into the economic and fiscal forecasts?
The exact effect of a policy change like this depends not only on its design and the characteristics of the labour market on which it is imposed, but also on when it is introduced. For example, if the labour market is very tight – meaning unemployment is low and vacancies are hard to fill – firms have less leeway to reduce wage growth, and therefore are more likely to have to absorb the increased cost, at least in the first instance.
The structure of the labour market matters a lot too. The UK economy is largely service-based, and our capital stock is relatively low compared with our peers. This means that production costs are highly affected by staff costs, and therefore we might expect firms to try to pass these costs onto prices more aggressively than in other jurisdictions.
Economists have historically though that most of the burden of these employer-side contributions would fall on workers.[2] Recent evidence supports a high degree of that burden being in fact borne by employees,[3] although there is some heterogeneity across types of workers. In particular, workers with lower qualifications and with more manual occupations tended to see a larger drop in demand for their labour, and therefore are disproportionately affected by increases in this type of taxes[4].
This is broadly consistent with what the OBR assumed the last time employer NICs (or an equivalent) were changed. When the Health and Social Care Levy was announced in 2021, the OBR assumed that it would leave the labour supply unchanged, but that in the long run it 80% of the additional costs would come through lower wages, while the additional 20% would come in the form of higher consumer prices. It also assumed that firms would bear about a fifth of the first-year cost in lower profits, but only for that period – households bearing the full cost subsequently, be it through earnings or prices.[5]
When the Health and Social Care Levy was then withdrawn in 2022, the OBR made some slightly different but broadly consistent assumptions. It assumed that the substitution effect – that is, the increased incentive to work from lower taxation – would have outweighed the income effect – that is, the increase in wealth from lower taxes[6]. This is a slightly more generous assessment for a tax cut, but obviously would mean a slightly more pessimistic assessment for a tax increase like the one mooted for the upcoming budget.
Of course, it is in the OBR’s gift to reassess these assumptions, and it may also put forward a view that increases and decreases could have asymmetric effects. But regardless of that, the direction of travel is pretty clear: the ultimate burden of higher employer-side NICs is on workers, be it through lower wages, fewer jobs or higher prices. It’ll be for others to pass judgement on whether it was covered by the manifesto pledge or not, but it would be pretty disingenuous for the Government to claim it is not a tax on ‘working people’.
_____________________________________________________________________________________________________________________
[1] NICs are in National Accounts parlance not taxes strictly speaking – they are classified as compulsory social security contributions – but they are for all intents and purposes a tax, and the OBR even include it in their ‘National Accounts taxes’ aggregate. In any case, it’s total receipts (both taxes and other forms of receipts) that matter for the Government and that form the totality of its revenues.
[2] Brittain, John A. 1971. “The Incidence of Social Security Payroll Taxes.” American Economic Review, Vol. 61, No. 1 (March 1971), pages 110-125.
[3] Carloni, Dorian. 2021. “Revisiting the Extent to Which Payroll Taxes Are Passed Through to Employees.” Congressional Budget Office Working Paper 2021-06.
[4] Guo, Audrey. 2024. “Payroll tax incidence: Evidence from unemployment insurance.” Journal of Public Economics, Volume 239, November 2024, 105209.
[5] https://obr.uk/box/the-economic-effects-of-policy-measures-17/ (Accessed 21 October 2024).
[6] https://obr.uk/box/the-demand-and-supply-side-effects-of-policy-measures/ (Accessed 21 October 2024).
Authors
João is Deputy Director and Senior Knowledge Exchange Fellow at the Fraser of Allander Institute. Previously, he was a Senior Fiscal Analyst at the Office for Budget Responsibility, where he led on analysis of long-term sustainability of the UK's public finances and on the effect of economic developments and fiscal policy on the UK's medium-term outlook.