Save It For A Rainy Day: Setting Charity Reserve Levels

What's so hot about reserves?

Financial reserves for charities seems to be a hot topic at the moment. The press has been baying for blood over the seemingly high level of reserves held by some of the larger charities, while other charities have gone to the wall when hit by unexpected drops in income (e.g. those charities worst affected by the National Lottery). The Charity Commission has recently published a consultation draft of comprehensive guidelines (cc19 Charities and the Retention of Income Reserves Exposure Draft) for the sector on the level of reserves which charities should hold. Many individuals and organisations have already contributed to the debate. In particular, the policy and discussion paper "Charities and Reserves" (Charity & Reserves - Policy discussion paper for the Chief Executives of National Voluntary Organisations (31/33 College Road, Harrow, Middlesex, HA1 1EJ)) published by the Association of Chief Executives of National Voluntary Organisations (ACENVO) makes a thorough and well-reasoned contribution which we used and which we do not intend to repeat. In particular, the discussion in the ACENVO paper on different types of reserves (e.g. designated and restricted funds) is of great value to charities that have various restrictions on their reserves. The type of analysis shown below is appropriate for charities with restricted funds, but the case study used only has general reserves. We have used data on Broadcasting Support Services as a case study from which to draw some further lessons on the appropriate analyses and conclusions which charities can make when considering financial reserves. We expect that this case study will be of interest to many charities, as most charities share some or all of BSS' characteristics with regard to reserves. We have submitted the study to the Charity Commission; we hope that it provides useful input to the consultation process by illustrating an alternative approach to for analysing a charity's reserves position.

Broadcasting Support Services (BSS)

BSS is a charity which provides follow up services to the public on behalf of broadcasters and other communicators in order to advance education, promote health and relieve poverty. It started its life in 1975 supporting the adult literacy campaign "On The Move" for the BBC and has since grown into a multi-faceted charity working with all broadcasters and running long-term helplines. BSS has grown steadily over the past fifteen years (last year's income was over £4M, see Table Two) and is currently experiencing further growth through the BBC Radio Helpline. BSS has no buildings or large tangible assets and its streams of revenue (and therefore expenditure) are largely based on contracts for which BSS provides services.

BSS and Z/Yen have been working together for the last few months on a strategic planning exercise for BSS. As part of that process, BSS wanted to understand the extent to which it can invest in opportunities while maintaining a sensible level of reserves. Until recently BSS has worked on the hypothesis that it needed to retain its modest surpluses as reserves (see Table Three) in order to provide prudent cover and to enable BSS to cope with fluctuating levels of income and expenditure. This article draws on some of the BSS data used in the strategic planning exercise.

The reserves continuum

We believe that the following table summarises the principal reasons for reserves in charities in the form of a continuum.

Table 1: The Reserves Continuum

A charity which has reserves below the prudence level has insufficient reserves. It is probably difficult to administer the charity in an efficient way, so the trustees might be in breach of their duties (see CC19 Charities and the Retention of Income Reserves (Exposure Draft)); such a charity's finance people probably cannot sleep well at night. A charity with reserves in excess of the capital projects level will have trouble explaining itself to the Charity Commissioners (and possibly also the press), especially if it is still fundraising. Many charities grapple with these problems, but most charities are more concerned with the middle ranges; e.g. what level of reserves is prudent? to what extent should we allow for unexplained variability? are we now able to finance special projects and capital programmes?

Unexplained variability

Many charities find unexplained variability difficult to explain [sic]. Basic statements of prudence (e.g. "we require reserves which cover three months of annual expenditure") are helpful yardsticks for minimum cover. The problem comes when deciding the point at which cover is sufficient to enable the charity to invest its reserves in special projects and capital funding. Consider the BSS data below:

Table 2: BSS Budgets, Actuals and Variants

BSS experiences significant income variances for which it cannot easily plan (83%, 69% and 53% in the most extreme years). Who would have predicted the scale of Live Aid before it happened in 1985? Should BSS have anticipated that the BBC Children in Need Appeal would grow some 2000% in less than fifteen years? BSS' role often needs to be reactive to broadcasters' requests. These unexplained variabilities can be looked on as special projects which have not been planned by the charity. Most charities experience unexplained variabilities to some extent. BSS experiences them perhaps more than most charities.

Despite the large unexplained variabilities BSS seems well able to plan expenditure (partly because BSS undertakes well-defined, discrete projects and partly because BSS seems to be good at budgeting for such projects, even at short notice). How then can we measure the potential impact of these expenditure variabilities on the level of reserves which BSS should hold? In a situation such as BSS, where we have a strong hypothesis on the cause of relationships between income and expenditure, correlation and regression analysis is a logical first step. The regression statistic known as the R-squared statistic is a standard feature with most spreadsheets (e.g. Excel or, as used in this case, Lotus 1-2-3) and shows the strength of relationship between two or more factors, such as actual expenditure, actual income and budgeted income. A figure close to 0% would mean that there is no statistical relationship between factors. A figure close to 100% means that the factors are highly correlated. (A basic textbook such as Statistics Without Tears by Derek Rowntree (Penguin) or Teach Yourself Statistics by Alan Graham (Hodder & Stoughton) should provide sufficient additional information on regression analysis and related statistics). In BSS' case 94.3% of actual expenditure can be explained by budgeted income. We also ran multiple regressions (in this case running both budgeted income and actual income against actual expenditure). 99.8% of actual expenditure can be explained by budgeted income and actual income, so a further 5.5% is explained by the actual income (see Table 2, actual expenditure against actual income). This shows that BSS' unexplained variability of income and expenditure is not having a great affect on reserves. The next set of data bears out this analysis.

* excludes expenditure for BBC Radio Helpline which is funded differently

Table 3: Net current assets, reserves expenditure cover

Actual expenditure has a 95.9% fit with reserves (see Table 3). The unexpected variability is not affecting reserves much. In plain English, BSS has been able to expand and contract its operations in response to demand for its services without a material effect on its level of reserves cover.

BSS conclusions

BSS is able to take on new opportunities and expand its operations with less risk than it previously thought, especially in areas of discrete, contract-based projects. BSS has learned three other main lessons from this exercise: working capital and reserves are not the same thing: as BSS has few tangible assets, BSS' reserves and net current assets have tended to be similar amounts, so BSS has tended to regard working capital and reserves as more or less the same thing. However, BSS can fund the working capital requirements of its activities other than self-funding through reserves. For example, the working capital for BSS' large new contract, the BBC Radio Helpline, is primarily funded by the BBC rather than BSS, thus releasing BSS from most of the working capital obligation unexplained variability is not as significant in BSS' case as the fluctuations in income and expenditure would infer. This appears to be partly due to tight budgetary control by BSS and partly due to the nature of BSS' contracts which have tended to be service oriented and place the variation burden on the other party BSS is now past the prudence and variation stages on the reserves continuum and can invest in some special projects in line with its new strategic plan. BSS has therefore established a modest development fund with which to so do. Sound application of this development money should enable BSS to meet its charitable objectives even more effectively. BSS has undertaken to be more proactive in identifying needs and ensuring that those needs are met through its efforts and the efforts of its working partners.

More complex cases

Statistics enthusiast readers are probably frustrated at the simplicity of the statistics used in this case study. "What about non-linear multi-variable regressions, cointegration and tests for spurious correlation?" we hear them cry. The reality is that the BSS example does not warrant more complex analysis; the statistics used, together with some common sense analysis, provides sufficient information upon which BSS can base its decisions.

In more complex cases (and many of the larger charities have more complex reserves analysis requirements due to multiple funding sources and various types of restrictions), we would advocate using more sophisticated analysis using the above statistics (e.g. multiple regression) and/or risk assessment using Monte Carlo simulation modelling (using tools such as Predict, @Risk or Crystal Ball) to test and model different hypotheses. The interested reader can obtain further information on these statistics from text books or, in cases where the reader is using statistical software, from the software manual.

Who should undertake this type of analysis

To quote the Charity Commission, "contingency funding can normally only be justified as provision for identified and quantified expected demands rather than mere uncalculated possibilities, fears and rumours". We believe that the form of analysis shown in this case study can help charities to quantify appropriate levels of funding.

Larger charities (say £10M per annum or more) are increasingly under the microscope on the issue of reserve levels and most would do well to perform such a reserves analysis in order to plan a change in policy or to silence the critics. Medium-sized charities (say £1M to £10M) with significant timing and/or unexplained variations are likely to benefit from such an analysis. In smaller charities and straightforward cases, anyone with a spreadsheet and some basic statistics can improve their understanding of their organisation's financial position.

[A version of this article originally appeared as "Save it for a Rainy Day: Setting Charity Reserve Levels", NGO Finance, (April 1996) pages 62-65.]