Charitable bequests and wealth at death: Institutions matter

Tony Atkinson, Peter Backus, John Micklewright 24 March 2018



Giving to charity at death is an age-old phenomenon and its importance is likely to grow. For charities, bequests are a major income source – in the UK, they make up about a quarter of donated income for the top 500 fundraising charities (CAF 2004). Many governments seek to encourage charitable giving, both in life and upon death, via favourable treatment in the tax code. The tax treatment of estates continues to be debated in the US and the UK, and understanding the effects of changes to these treatments on the willingness to leave something to charity is of considerable policy relevance.

In our recent paper (Atkinson et al. 2017), we consider the decision of whom to leave one’s wealth to – a decision unlike all others. To better understand how a charitable bequest is realised, we propose a multi-stage framework for the decision to leave a charitable bequest and examine evidence about charitable giving at death provided by new data on the population of estates passing through probate (those worth at least £5,000) in a 12-month period in Britain. These data allow for many new insights, not least that about half of the population die with very little to leave to anyone. In Scotland, for example, only about 45% of people die with wealth of at least £5,000.

The novelty of our framework is its distinction between the different steps leading to a charitable bequest. For a charity to receive a bequest from someone, a number of steps must be taken. The individual must have wealth to leave, he or she must make a will, the will has to include a charitable bequest, the bequest may be conditional rather than absolute, and the bequest may be a cash amount or a residual share. In analysing the different steps, it is important that our new data cover the whole population that are leaving non-trivial estates. In contrast, much of the US literature based on estate tax data relates only to the upper tail of the distribution of estate size at death. For example the study by Joulfaian (2000) covers only 3% of decedents in the US, Kopczuk (2007) considers only estate tax payers in the US, and Aldous (2005) considers only 911 estates in the UK.

Moreover, in contrast to studies based on tax data, our data have fuller information about donor intentions available from the accompanying wills so that we observe all charitable bequests made, including those with conditions that may not be realised. This is an important feature as it allows us to study charitable intent as opposed to just the realisation of that intent. These are more likely to diverge in the case of bequeathing as opposed to living donations, given the length of time between when a will is written and one’s death (on average, five years in our data) and subsequent uncertainty about the final estate size, debts owed, and inheritance tax rules. A shortcoming of our data is that although we have very rich information on the estate size distribution and people’s charitable intent we rarely observe the value of the bequest. As such our analysis focuses on people’s willingness to leave a bequest, rather than the amounts that are bequeathed.

Using these data, we consider several aspects of charitable bequeathing, but will focus on two in this column. The first is the association of estate size with the probability of making a charitable bequest. Wealth is likely to affect the various steps in making a charitable bequest differently. For example, we find the propensity to make a will may be expected to rise with wealth and then level off, whereas the propensity to make a charitable bequest may rise more steadily. There is at present little empirical evidence about the patterns that prevail. For the UK, Wedgwood (1929) documented charitable bequests in wills published in The Times newspaper, but the nature of his sample raises obvious questions concerning the representativeness of the data. Dawson et al. (2003) studied all estates passing through probate in Northern Ireland in 1937, 1967, and 1997, but did not consider the estate values. Even in the US, where the literature is far more extensive, the restriction of the great bulk of studies to tax data means that knowledge is limited to the variation in realised charitable bequests among just large estates.

Our results, shown in Figure 1, suggest that testacy (i.e. having a will) increases rapidly, rising from about half for the smallest estate in our sample to over 90% for estates worth around £150,000 (just above the median), and up to 99% for estates larger than £3 million (not shown in the figure). 

Overall, 16% of wills include a charitable bequest, though this rises non-linearly with estate size. The rise is modest over the bottom half of the estate size distribution from 11% for the smallest estates to 14% for estates worth around £150,000. This propensity for charitable bequeathing rises much more rapidly over the upper tail of the estate size distribution, however, increasing to 22% for estates worth around £300,000 and up to 51% for estates worth more than £3 million – meaning half of the very wealthiest decedents not only leave nothing to charity, but express no intent to do so.

Figure 1 Testacy, bequeathing to charity conditional on testacy, and absolute bequeathing conditional on charitable bequeathing, by estate size up to £500,000

Notes. The figure shows, for estates worth less than £500,000, the percentage of decedents who are testate, the percentage of testators who leave a charitable bequest and the percentage of those leaving a charitable bequest who leave at least one absolute charitable bequest. The vertical dotted line marks the overall median estate size of £146,000. We only show the patterns for estates up to £500,000 but show these results for the full estate size distribution in our paper.

We also consider the likelihood of a person to leave an absolute bequest, i.e. one that is made without conditions and will be realised provided there are sufficient assets in the estate at the time of death. This propensity rises from 57% for the smallest estates to 71% for those around the median, to 88% for the largest estates. Note this is just the propensity to leave at least one absolute bequest. People often leave multiple bequests, on average people leave about 2.8 charitable bequests, and these can include some absolute and some conditional ones. 

The second aspect we will discuss here is the effect of the inheritance tax threshold. One reason why charitable bequeathing may rise with wealth is progressive estate taxation combined with tax-deductibility of charitable bequests. Under the UK Inheritance Tax (IHT), the excess of an estate above a threshold (£300,000 for most of the period in question) is subject to a 40% marginal tax rate. The ‘price of giving’ to charity (the amount by which inheritance of other beneficiaries is reduced) falls from 1 to 0.6 when the threshold is exceeded. The US literature on charitable bequests has been particularly concerned with identifying a price effect (e.g. Boski 1976, Joulfaian 2000). 

The model developed in our paper suggests a jump in the proportion of decedents making a charitable bequest when the estate reaches the IHT threshold – £300,000 for most of our sample. The percentage rose from 17% to 33% between estates in the range £250,000 to £299,999 and £500,000 to £999,999. But it is unclear how much of the increase is due to the rise in estate value and how much to any IHT effect. Disentangling the two is complicated as the impact of IHT is unlikely to appear as a discrete jump. Instead we expect the shift in the propensity to take place over a range of wealth for two reasons.  First, the point at which the tax applies for a particular estate may differ from the value of the IHT threshold. Second, charitable bequests are often determined some time before death. The testator has to form a view about likely future estate value and the future tax regime. Both may change substantially.

Using a piecewise linear specification, we estimate an implied price elasticity of the probability of bequeathing to charity of about -0.3. We also use a more flexible approach that more generously allows non-linearities in the relationship, and plot the estimated marginal effects in Figure 2, which illustrates the significant increase in the marginal effect around the IHT threshold relative to its value at both higher and lower estate values. The marginal effect increases as the IHT threshold is approached and falls beyond it, eventually returning to parity with that for smaller estates. We make no claim to have firmly identified a causal effect, but the results are consistent with the IHT leading to an increase in charitable intent as predicted by our theoretical model.

Figure 2 The marginal effect of estate size on the probability of making a charitable bequest by estate size (estates of £200,000 to £599,999)

Notes: The figure plots the marginal effect of estate size on the probability of making a charitable bequest (conditional on testacy) by estate size (in £ millions). The marginal effects are calculated from a linear probability model with a restricted cubic spline functional form with knots located at £275,000 and £325,000. A total of 70,165 estates fall within the employed range.  The vertical dotted line shows the IHT threshold in April 2007-April 2008.

Many of our results would be unobtainable using datasets of the type used in much of the existing literature. Our data provide a fuller understanding of the estate size distribution and about donors’ charitable intentions, as the estates under study are not restricted to the small fraction of the population covered by estate tax returns for large estates that detail only realised charitable bequests.


Aldous, J (2005), “Analysis of legacies to charities – destination at death sample”, HMRC, unpublished.

Atkinson, A B, P G Backus, and J Micklewright (2017), “Charitable bequests and wealth at death in Great Britain”, Economic Journal, 127, F1-F23.

Boskin, M J (1976), “Estate taxation and charitable bequests”, Journal of Public Economics, 5 (1), 27-56.

Charities Aid Foundation (2004), Charity Trends 2004: 25th Anniversary Edition, CAF and CaritasData.

Dawson, N, S Grattan, L Lundy, R Glenn, and G Cran (2003), Dying to Give? Trends in Charitable Giving by Will, London: The Stationery Office.

Joulfaian, D (2000), “Estate taxes and charitable bequests by the wealthy”, National Tax Journal, 53 (3), 743-63.

Kopczuk, W (2007), “Bequest and Tax Planning: Evidence from Estate Tax Returns”, The Quarterly Journal of Economics, 122 (4), 1801-54.

Wedgwood, J (1929), The Economics of Inheritance, London: Routledge.



Topics:  Taxation

Tags:  charitable bequests, wealth, estates, inheritance tax

Former Centennial Professor at the London School of Economics and Fellow of Nuffield College, Oxford

Senior Lecturer in Economics, University of Manchester

Professor of Social Statistics and Policy Analysis, University of Southampton and CEPR Research Fellow


CEPR Policy Research