John Rentoul, chief political commentator at the Independent and biographer of Tony Blair, has achieved a cult following for his list, published online of “Questions to which the answer is no”, currently standing at number 730 at time of writing. Despite some postulating to the contrary in recent weeks, the title of this blog is a good contender for slot number 731.
A report, by Adrian Sargeant and Jen Shang of the Bristol Business School at the University of the West of England, has strongly suggested, however, that the scheme may not be “fit for purpose”, and that it may warrant scrapping, to be replaced by widespread solicitation of donations via direct debit in the workplace.
It is true that payroll giving is not without its flaws; Potter and Scales (2008) identify in their review of payroll giving a considerable number of areas for improvement, including the need to simplify the system of enrollment, better market the offering, and to speed up the process so that charities receive money given more quickly.
While nobody disputes that payroll giving could be improved, it is foolish to argue that it is without important merits, such as employer matches and tax effectiveness. Moreover, although payroll giving is unpopular among fundraisers, it is popular both with government (who ultimately will decide whether the system is to be scrapped) and with donors. If the retail idiom that “the customer is always right” holds for charity, this should by itself be enough to justify its continuation.
These advantages to payroll giving pale in comparison, however, to the strength of the assumptions made by the report’s authors.
The authors argue that while employer matches and some level of tax efficiency will be lost by moving from Payroll to Direct Debit giving, this will be compensated for by removing the need to contract with a Payroll Giving Agency, and, seemingly, through increased revenue per head (people tend to give more through direct debit than through payroll).
For this to be true, current direct debit givers would need to be identical to current payroll givers in every dimension except their means of giving. The evidence does not seem to bear this out, with a recent report by the Charities Aid Foundation, finding that payroll givers are on average young and more likely to be male than the general population of givers. Since these groups tend to give less to charity generally, it seems that selection, not the mechanism of giving, is driving the smaller donations.
If we don’t believe that donations will suddenly rise on conversion to a direct debit system, can we at least assume that donations per head will remain the same? Probably not. Donations through Payroll Giving are cheaper than through direct debit because of their tax effectiveness; for each unit of “good” done by the charity, the donor must forego less other their own consumption. This move would amount to a rise in the price of giving for an important element of society, and under reasonable assumptions could lead to a fall in donation amounts.
Nor is it clear that the number of donors would remain constant. Potter and Scales (2008) identify “widespread opinion that employer-matching schemes were very valuable in encouraging employees to sign up, and provided a real incentive” (p 53), suggesting that at least some donors are induced to donate by the match – if the match varies with their donation (as is common), this amounts to a further reduction of the “price” per unit of good done by the donor, and removing it would further discourage donation.
Some money would undoubtedly by recouped from the portability of direct debits, which is estimated (Jenkins (2007)), cost charities around £7million a year; whether the full amount of this would be recouped is uncertain, as some employees leaving a firm for unemployment or retirement may wish to curb their spending and hence cancel their direct debits, despite incurring a guilt cost.
How sensitive might giving be to some of these factors? Without rigorous analysis, it is impossible to say with much confidence, but for illustration, we can look at two model cases of payroll giving compared with the average. Mean donation size through payroll giving is around £10, with only 6% of those eligible taking part. Royal Mail and British Telecom, both of whom offer an employer match, raise £2.5million apiece, Royal Mail through high levels of enrollment (25%), and BT through large donations (around £20 a month), and both attribute their success in large part to the match. Both firms give around 4 times the average for a firm of their size, and between them contribute 5% of all payroll giving donations – were enrollment to fall to average levels, this would make a marked difference.
Comparing the two worlds as best we can, we see that £4.4million (Guardian, 2011), of revenue could be saved through reduced administrative cost, and £7million through portability in the best case scenario, for a saving of £11.4million. Leaving aside possible changes to amounts donated and to individuals’ likelihood of donating, Potter and Scales (2008) estimate that employer matches amount to 10% of payroll giving. Since payroll giving is valued at £114million in 2011, the cost of scrapping it to charities would be £11.4million – exactly what it would save under the best case. In short, the case for scrapping payroll giving just doesn’t add up.
 Potter and Scales (2008): “Review of Payroll Giving” Institute of Fundraising
 CAF (2011) “Payroll Giving in the UK”
 Jenkins (2007): “Report to support the proposal for portability of Payroll Giving” Institute of Fundraising
A recent poll by Workplace Giving, a payroll giving company, indicated that of 1000 respondents, 68% would favour the introduction of an automatic-enrolment system for payroll-giving. This kind of system would make the default for all workers in participating firms to be enrolled in a the scheme, giving to a given charity at a given level – they could then, if they so wished, opt-out of the scheme, or change the amount they give, and/or the charity to which they are giving (64% of respondents indicated a preference for having a hand in choosing the charity in the first place).
Why should this be the case? Only a minority of workers (only 6% of workers in participating firms, 2.4% of all workers), actually give through payroll giving at the moment. While it is natural to suspect some selection bias in the answers to these questions, the magnitude of the difference suggests that at least some respondents who answered positively about the automatic enrolment system are not currently enrolled in payroll giving schemes in their firms.
If they are so keen to give to charity that they would favour and opt-out arrangement, why do they simply not enrol themselves in existing schemes, or set up a direct debit of their own accord? Work from across the field of Behavioural Economics appears to suggest some answers.
The first, and most obvious suggestion, is that individuals have a self-control problem, and continue to put off the act of enrolling. In these cases, a self-aware individual may prefer a “commitment device” such as an automatic-enrolment system, which compels (or in this case forces) them to partake. Evidence from the pensions literature in the united states suggests that automatic enrolment can have powerful effects on individuals’ behaviour in overcoming their self-control problems.
Respondents to the survey may also be demonstrating systematic overconfidence. Although the existence of this phenomenon is observed in previous research, for example finding that 93% of all drivers believe they are above the median ability, it is particularly interesting when combined with self-control problems. Despite previous evidence to the contrary, for example not yet having signed up to Payroll Giving, individuals strongly believe that they are more likely to overcome their self-control issues in the future. Given their expectation that they will sign up to payroll giving…tomorrow, and that their colleagues will not, automatic enrolment will make no difference to them in the long run, but will have a larger impact on others’ giving.
Finally, the institution of an opt-out is likely to induce more homogeneity of charity choice (among those passively remaining in the scheme), compared with softer policies such as “active enrolment”, whereby individuals are compelled to make a decision for themselves (often through a mandatory form). If the respondents were motivated by a particular cause (heart disease, for example), and believe themselves a more active participant in decision making, a prescriptive policy may be better for them, albeit at the cost of social welfare.
These phenomena, and others like them, may pose significant obstacles to the Government’s plan to induce a ‘step change’ in giving in the UK – something which, as this blog has previously discussed, will be very hard to do. They also point to a more nuanced problem with giving; instituting auto-enrolment may increase the level of giving, but if workers are heterogeneous in their charity preferences, it may decrease social welfare, or be a second best policy compared to a softer intervention. Although the results of this poll are positive for those who support payroll giving schemes, it is clear that more research must take place before we are able to accurately predict its effects.
 Potter & Scales (2008) “Review of Payroll Giving” Strategy Complete Ltd. Commissioned by Institute of Fundraising
 Choi, Laibson, Madrian, Metrick (2006) “Saving for Retirement on the Path of Least Resistance”. Behavioral Public Finance: Toward a New Agenda. Ed McCaffrey and Joel Slemrod, eds. New York: Russell Sage Foundation pp304-51.
 Svenson (1981): “Are we all less risky and more skilful than our fellow drivers?” Acta Psychologica 47: 143-148.
 Carroll, Choi, Laibson, Madrian, & Andrew Metrick (Forthcoming) “Optimal Defaults and Active Decisions.” Quarterly Journal of Economics