How many charity employees have been furloughed?

This blog post is a write-up of a twitter thread exploring data available. These are very rough estimates, based on data that's available at the time of writing.

It's frustratingly difficult to find out how many charities have used the government's Coronavirus Job Retention Scheme, or how many employees have been furloughed through it.

For the rest of the economy, HMRC has produced "experimental statistics" with estimates of the number of employers and employees that have used the scheme. These statistics are based on the data HMRC uses to administer the scheme. As well as overall totals, they've been able to produce estimates of how use of the scheme has varied across sectors of the economy (aka "industries", using SIC2007) and geography (local authorities and regions).

Unfortunately it's not possible to directly extract charities from this data - charity is a legal form, and charities work in lots of different industries as classified by SIC2007 - health, arts, social care, education, etc. So the industry breakdown doesn't allow charities to be identified.

Estimating using sector data

However, even though the industry data doesn't allow us to directly estimate the numbers of charity employers and employees using the scheme, we can use the industry breakdowns as a base for some rough estimates.

If we make the assumption that charities behave similarly to other organisations in their industry (which may not be true) then if we know the proportion of organisations in each industry that are charities, we can apply that proportion to the number of employers and employees to make and estimate of figures for charities.

But how to get an estimate of the proportion of companies that are charities? Here I've turned to Companies House and Charity Commission data. Using data extracts from these two regulators, it's possible to get a list of companies, their SIC codes (for their industry) and whether or not they are a registered charity. From this we can produce a list of industries and the proportion of organisations in each that are charities.

But is this a fair picture? Is the number of companies a proxy for the number of employees? Well possibly, but it's probably more complicated. We know that employees aren't evenly distributed in any sector of the economy - NCVO's data shows that 40-50% of employees in private & voluntary sector work in organisations with more than 50 employees.

There are lots of small charities without paid staff, and similarly there are hundreds of thousands of small registered companies that don't employ people, or are currently dormant. To strip these out, I used a field in the company data that describes the type of accounts the company last submitted - this is based on their size. But limiting it to accounts returned by active companies, we get a better sense of the proportion of "active" companies that are charities.

Results of these estimates

The two methods - using all companies and then using larger "active" companies, give two different figures for the proportion of registered companies that are charities. For all companies charities represent 0.7% of companies, compared to 2% of "active" companies. The latter figure perhaps feels instinctively more correct - we know from NCVO that charity employees are around 3% of the workforce.

The figures give the proportions for industries. These proportions feel correct - charities are concentrated in 4 SIC industries: education (20% of companies), health and social work (12% of companies), arts & recreation (12% of companies), trade union, religious, political and repair (8% of companies). They're also absent where you would expect - mining, manufacturing and construction show negligible numbers of charities (the SIC classification system is definitely not perfect in every case).

And using these proportions we can make estimates of the number of charities that are using the furlough scheme, and the number of employees. Using the proportions for all companies gives a figure of 8,000 charities and 69,000 employees, but using just larger active companies gives 22,000 charities and 188,000 employees.

While there's a number of assumptions going into these figures, I think they feel like sensible estimates of the boundaries of the actual number, though the larger one is perhaps closer to the truth.

Comparing with survey results

When I shared these results on Twitter, there were some helpful pointers to attempts to answer the same questions using surveys.

Nick Temple of the Social Investment Business estimates that 13% of their customer base's employees are on furlough.

Charity Tax Group & Charity Retail suggest that many of the 26,000 charity shop employees have been furloughed - the unique nature of these employees (similar private businesses do not have retail employees) means that they are potentially missed from my estimate.

A Charity Tax Group survey showed 70% of respondents had made a claim.

66% of respondents to Pro Bono Economics survey (9-10 June) said they had furloughed or used job retention scheme.

What next?

These estimates are probably wrong. They're based on imperfect data and assumptions that are unlikely to be correct. However, given the absence of better data, I think they give an indication of the scale we would expect to see in figures.

ONS and HMRC have done a fantastic job during this crisis of producing statistics to understand the environment and measure responses. But charities can often fall through the cracks in official statistics - due to lack of data, or because they're not thought about when the measures are designed.

I'd like to see better official data on charities in this crisis, to help the charity sector measure the impact and respond.