Ask a director what their corporation tax bill will be and the honest answer is usually a shrug and a guess. The accounts get filed some months after the year end, the figure appears, and the only decision left is how to pay it.

By then every lever has gone. Pension contributions had to be paid before the year end. Salary and dividend levels had to be set while the year was still running. Capital purchases had to be made, not planned. Finding out the number after the year has closed turns a planning exercise into an administrative one.

That is the gap this article is about, and it starts with the fact that the rate itself is not fixed.

For the 2026/27 tax year, the lower limit for marginal relief is £50,000 and the upper limit is £250,000, according to HMRC’s corporation tax rates and allowances. Profit under the lower limit is taxed at 19%. Profit over the upper limit is taxed at 25%.

Between the two, something less obvious happens. The main rate applies and marginal relief claws part of it back, which means each additional pound of profit in that band carries more tax than either headline rate suggests. Plenty of owners are paying that rate right now and have never had it explained to them, because nobody sat down and worked out where their profit lands.

Until April 2023 there was a single rate and the arithmetic was simple. Since then, corporation tax has had a structure rather than a rate, and the structure has not changed for 2026/27.

There is the small profits rate of 19% for taxable profits under £50,000. There is the main rate of 25% for profits over £250,000. And there is the band between those two limits, where the main rate applies with marginal relief deducted. The marginal relief fraction for 2026/27 is 3/200.

Three positions, and only one of them is the one your company meets. The first job in any corporation tax forecast is establishing which.

The most common position we see is an owner who set money aside at 19% because that is what the company paid in its early years, and who has never revisited it.

It works perfectly well until the year profit climbs past £50,000. Growth rarely announces itself at the year end. A good contract, a quiet year for capital spending, a decision to leave more in the company, and suddenly the profit figure sits in the middle band. The reserve set aside at 19% is short, and nobody notices until the computation is done.

The assumption is reasonable. It is the absence of a check that causes the problem.

Marginal relief is not a separate rate and there is no line on the return that says 26.5%. The company is charged at the main rate of 25% and then relief is deducted using HMRC’s formula, (F x (U – A)) x (N ÷ A), where F is the marginal relief fraction of 3/200, U is the upper limit, A is augmented profits and N is taxable total profits.

The effect of that calculation is that every extra pound of profit between the limits attracts 26.5% tax. HMRC’s own company taxation manual describes 26.5% as the marginal rate for exactly this reason. It is the slope between 19% and 25%, and it is steeper than both.

Now take a position we see often among consultants and contractors, and among trades businesses that have grown a second arm. The owner runs two limited companies. Perhaps one holds the trading work and one holds a separate service line, or a second venture that started small and kept going.

Both companies are profitable. Neither is anywhere near £250,000. Looked at on its own, each one reads as a comfortable small profits rate company, and the accounts for each get prepared, signed and filed on that basis.

Read separately, those two sets of figures tell you nothing useful. The limits do not work company by company.

The £50,000 and £250,000 limits are proportionately reduced by the total number of associated companies, including the company itself. With two associated companies, the lower limit becomes £25,000 each and the upper limit £125,000 each. HMRC’s marginal relief guidance gives the same principle with four companies, where the limits divide by four.

So the company that looked like a straightforward 19% case because it made £40,000 is now a marginal relief case, taxed at the main rate with relief deducted. Nothing about the trading changed. The structure around it did.

Short accounting periods reduce the limits in the same way, which catches companies in the year they change their year end.

The reason this gets missed is procedural rather than technical. Each company is a separate filing job with its own deadline and its own file. In a practice running volume, those jobs often sit with different people, or with the same person at different times of year.

The association question is asked on the return, but asking it at the point of filing is far too late to be useful. The number that matters is the one the owner needed twelve months earlier, when the profit split between the two companies could still be influenced.

Nobody sits the two companies side by side and asks what the group position means for the limits. That is the work.

The cost of finding out late is rarely the tax itself, because the tax was always going to be due. The cost is cash flow and lost choices.

A director who reserved at 19% and faces a computation built on the main rate with marginal relief has to find the difference from somewhere, usually from money already committed to a hire, a vehicle or a dividend that has been drawn. The nine months and a day after the year end sounds generous until you remember the accounts are often not finalised until well into that window.

There is also the planning that never happened. Pension contributions, timing of capital spend and the salary and dividend mix all had to be decided inside the accounting period, not after it.

The way we handle this at Nexus is to work the question in the opposite order. Before anything is filed, we establish the number of associated companies, the length of the accounting period and the adjusted lower and upper limits that follow from both. Then we look at projected profit against those limits and work out which rate the company is going to meet.

That gives a forecast rather than a result. We tell the client the corporation tax number about a month before it is due, and where the figure sits in the marginal band, we say so plainly, because that is the position where decisions during the year make the biggest difference.

When the rate is known early, the conversation changes. A director sitting just above the lower limit can look at an employer pension contribution and see exactly what it does to the effective rate on that slice of profit. A director with two companies can look at where the work is being invoiced and whether the current split still makes sense.

None of that is exotic. It is ordinary planning, and it only works if the numbers arrive while the year is still open. Salary and dividend levels, timing of equipment purchases, whether to leave profit in the company or take it out: all of them are decisions, and all of them need a rate attached to be made properly.

A perfectly accurate corporation tax computation delivered after the year end tells you what happened. A reasonable forecast delivered while the year is still running tells you what you can still change. The second is worth more, even though it is the less precise of the two.

So the practical takeaway for 2026/27 is short. Find your taxable profit, count your associated companies, adjust the £50,000 and £250,000 limits accordingly, and see which of the three positions you meet. If you land between the limits, the effective rate on that profit is 26.5%, and that is a number worth knowing in advance.