How many women go on statutory maternity pay in the UK each year?

Go on, have a guess. Or maybe you know the number?

Perhaps you do. Or maybe you only think you do.

According to news reports last month in the Independent newspaper – headline: “Parental leave system is broken: number of fathers taking paternity leave plunges to 10-year low” – and the supposedly specialist journals Personnel Today and People Management, it’s about 650,000:

Only 27% of eligible fathers took time off last year; 176,000 men took paternity leave and claimed statutory paternity pay in the 12 months to 31 March 2021 compared with 652,000 women who took maternity leave over the same period.

The data was released by HM Revenue & Customs following a Freedom of Information request by law firm EMW.

All three news reports were based on a press release by law firm EMW. Indeed, in recent years, EMW have been admirably effective at getting press coverage out of pretty much the same story, based on the same (updated) set of HMRC data: here they are in the Independent and Management Today in July 2019, and the Telegraph and HR News in August 2020. One year it’s ‘Only one third of fathers are taking paternity leave!’, and the next it’s ‘Two thirds of fathers are not taking paternity leave!’ Well done, law firm EMW.

We can only guess how EMW would have framed this unchanging story last month, had the Covid19 pandemic not come along in 2020. This provided them with the somewhat over-dramatic “paternity leave take-up has hit a 10-year low, with only approximately a quarter of eligible fathers (27%) taking paternity leave after the birth of their child” hook, to the excitement of sub-editors at the Independent and the usual suspects such as Pregnant Then Screwed. But then it is hardly surprising, given how much of the workforce was on furlough for much of 2020/21, that some 30,000 new fathers decided to continue at home on furlough, on 80-100% of their normal wages, rather than take one or two weeks of statutory paternity leave on just £150 per week.

Whatever, here’s that HMRC data in full, showing that – *checks notes* – the paternity leave system is broken:

However, there are several problems with this narrative, and the HMRC data on which it is based.

The first problem is the assumption by EMW that the number of fathers (and other second parents) who are eligible to claim statutory paternity pay (i.e. the denominator for their paternity leave take-up rate) is the same as the number of women who started statutory paid maternity leave. There are a number of reasons why this is not the case, but to be fair it’s probably the best proxy available, so we can probably let this one pass.

The second problem is that, if we do assume that the number of fathers (and other second parents) who are eligible to claim statutory paternity pay is the same as the number of women who started statutory paid maternity leave, then we probably need to include in that latter number at least some of the 60,000 women who go on Maternity Allowance, rather than SMP, each year. The legal eagles at EMW seem to have forgotten about Maternity Allowance, but adding the 40,000 employed new mothers who get Maternity Allowance to their base figure of (about) 650,000 would make the paternity leave system look even more “broken”: it would, for example, indicate a paternity leave take-up rate of just 30% in 2019/20, not 32% as suggested by EMW.

However, the third – and biggest – problem with EMW’s narrative is that the number of women who start statutory maternity pay each year is not (about) 650,000, or anywhere near that number.

For a start, according to ONS data, in 2018 there were only 649,626 maternities in England and Wales, plus about 60,000 in Scotland. And a significant proportion of those 710,000 mothers (plus those in Northern Ireland) will not have been in employment (in March 2019, the Department for Business, Energy & Industrial Strategy estimated the employment rate of ‘women of child bearing age’ to be 73%). So it’s just not credible that pretty much all of the 710,000 went onto either SMP or Maternity Allowance.

More to the point, every year the DWP publishes data for benefits expenditure and caseload, including for SMP and Maternity Allowance. This data is published under the terms of a Memorandum of Understanding with the Office for Budget Responsibility (OBR), and reflects the Government’s financial delivery plans. So, if it’s wrong, the Budget is a pile of pants. And, according to this data, between 2012/13 and 2020/21, the average number of women going on to SMP was just 268,000 (within a range of 262,000 to 275,000).

The DWP also publishes quarterly data for Maternity Allowance starts, including a breakdown by employment status (’employed’ or ‘self-employed’), so we can add the some 40,000 employed women who started on Maternity Allowance each year to the OBR-approved number of women who started on SMP. (Note that, in 2020/21, there was also a pandemic-related dip in the number of employed women who started on Maternity Allowance, from 40,000 to 30,000. Believe it or not, neither law firm EMW nor Pregnant Then Screwed have yet made a fuss about this.)

This combined data gives us a significantly different denominator (for calculating the take-up rate of paternity leave) to the HMRC data relied upon by law firm EMW and their friendly (but somewhat uninformed) journalists, and so a rather different picture of the take-up of statutory paternity leave:

Now you might still argue that even this data shows the paternity leave system to be ‘broken’. But I would suggest that a take-up rate of about 70% is not that bad, really, given the stupidly low rate at which such leave is paid and the other barriers to take-up, not to mention the fact that a good chunk of the ‘missing’ 30% will be new fathers who are self-employed, so are not even entitled to statutory paid paternity leave. Indeed, this is one of the reasons why, as mentioned above, law firm EMW are wrong to assume that the number of new fathers who are eligible to take statutory paternity leave is the same as the number of new mothers who start statutory paid maternity leave.

So, which set of data on SMP is correct, and why the discrepancy between the two? Well, at the time of writing, there’s no definitive answer to that, but this recent Answer by Treasury minister Jesse Norman to a Parliamentary Question by Kirsten Oswald MP strongly suggests (a) that the Government favours the OBR-certified data in the DWP’s annual Benefit Expenditure & Caseload tables; and (b) that one reason for the discrepancy is that the HMRC data on SMP provided to law firm EMW (and others) via Freedom of Information requests includes a lot of double counting, because it “includes claimants in each year in which they received statutory payments”.

In other words, a woman who took nine months of statutory maternity leave on SMP from October 2019 to June 2020, say, will have been counted twice by HMRC, first in its figure for 2019/20, and then in its figure for 2020/21. With women taking nine months of statutory paid maternity leave, on average, this means the HMRC data overstates the true figure by up to 75%. (There may well be such double counting in HMRC’s figures for statutory paternity pay, too, but as claimants take only one or two weeks of such pay, the inflationary effect will be negligible.)

Allowing for this double counting would reduce the HMRC figure of some 650,000 SMP starts a year to as little as 370,000 (which would increase EMW’s paternity leave take-up rate for 2019/20, from 32% to 56%). But that still leaves a significant discrepancy between the HMRC data and that published by the DWP as part of the OBR-certified Benefit Expenditure & Caseload tables.

I’m awaiting a response from HMRC to a Freedom of Information request seeking an explanation of the discrepancy, so maybe everything will yet become clear. Or, more likely, it won’t. But watch this space.

(For the record, none of the above is to dispute that our parental leave system is broken. It is very broken, and I have written extensively about just how broken it is. But paternity leave is the least important part of that system and, while it should be better paid and be a Day One right for all workers, including the self-employed, there are much bigger fish to fry in this policy area.)

Update (27 September): In its response to my FoI request, HMRC has confirmed that the figures provided to the law firm EMW (and others) include lot of double-counting:

The number of claimants is the total number of individuals in receipt during that year, irrespective of when the payment first started. Where a given spell of [statutory shared parental leave pay, statutory maternity pay or statutory paternity pay] extends across two years, the claimant will be included in both years’ figures.

So, EMW’s narrative about the take-up rate of statutory paternity leave is indeed ‘broken’.