Lump sum payments on divorce Looking back at the Duxbury calculation 30 years on...

This article is reproduced from Financial Remedies Journal 2023 Issue 1, with the kind permission of the publishers, Class Legal. It was written by Megan Jenkins and Amy Gilbert of Saltus, and Michael Allum, solicitor and partner at The International Family Law Group LLP.

Overview

The Duxbury formula seeks to ascertain a capital amount which if invested to achieve capital growth and income yield (both at assumed rates and after tax on the yield and required gains) could be drawn down in equal inflation-proofed instalments over a period of time (often the recipient’s life expectancy) but would be completely exhausted at the end of the period.

The Duxbury concept originated from the case of Duxbury v Duxbury [1990] 2 All ER 77, CA, in which the wife – aided by an accountancy firm [1] she had instructed – put forward a projection of the capital sum she required to meet her future needs based on various assumptions. For the best part of a decade it ruled supreme in all co-called bigger money cases on the then law based on the reasonable requirements of the applicant. Many predicted the seminal decision of the House of Lords in White v White [2000] UKHL 54 in 2000 would signal an end to Duxbury but it has withstood the changing landscape and remains an important tool which is often used by family lawyers and judges.

Duxbury is now most frequently used to assist the court in quantifying a clean break in a non-sharing case and as a guide to capitalising existing periodical payments in variation proceedings. It can also be used as a cross-check on whether the application of the sharing principle is likely to meet the recipient’s needs.

The courts have acknowledged that Duxbury has been subjected to criticism, particularly in relation to the returns that it assumes will be made. It does, however, remain the tool most often used by family lawyers and judges when quantifying a clean break in a non-sharing case and capitalising a periodical payments order in variation proceedings.

This article explains some of the terminology in layperson’s terms, summarises the leading case-law then speculates (see the Appendix) as to how the Duxbury funds in previous reported cases may have performed if invested in more cautious risk portfolios than Duxbury assumes, based on a set of assumptions prepared by a financial planner modelled on historic market performance. Michael Allum is the primary author of the text as a financial remedies practitioner. Megan Jenkins and Amy Gilbert have undertaken the calculations based on the alternate set of assumptions as specialist financial advisers.

Jargon busting

Capital Growth is the increase in the value of an asset or investment over time. It is measured by the difference between the current value, and its purchase price. Duxbury presently assumes 3.75% per annum.

Income yield is earnings generated by an investment over a particular period of time. It is expressed as a percentage based on the value of the investment. Duxbury presently assumes 3% per annum (1.5% in the first year).

Inflation is the rise in prices, which relates to the decline in the purchasing power over time. Inflation means that a unit of currency effectively buys less than it did in prior periods. Duxbury presently assumes 3% per annum.

The key assumption is therefore that over a long period of time a fund will perform in actual gross terms by 6.75% annually but the owner will suffer inflation of 3% giving a real rate of return of 3.75% per annum.

Post-White case-law

The 2003 case of Pearce v Pearce [2003] EWCA Civ 1054 at [37]–[38] concerned an application by the husband to discharge/capitalise a joint lives periodical payments order together with a cross application by the wife to increase the quantum of the periodical payments order. Thorpe LJ held that a trial judge should not be put in a ‘straight jacket’ but that the discretion afforded to the court in departing from the Duxbury formula when deciding what capital sum should replace a periodical payments order was a ‘narrower one’.

In 2010 in Vaughan v Vaughan [2010] EWCA Civ 349, Wilson LJ (as he then was) endorsed the principles set out in Pearce. The learned judge commented that the decision in Pearce had ‘rightly received wide approbation’ because it identified what Thorpe LJ had described as ‘a relatively simple, certain and predictable method for the calculation of the capital sum’. Wilson LJ concluded by holding that:

‘The court has, thank goodness, only a narrow discretion to arrive at a capital sum otherwise than by application of Duxbury formula and it should exercise it in order only to reflect special factors.’

In 2014 in H v H [2014] EWCA Civ 1523 at [31] Ryder LJ held as follows:

‘In summary, it is not wise to assume that because the Duxbury Committee are of the opinion that in the context of their calculations 3.75% gross is achievable over the long term with a cautious investment strategy that the parties will agree that that rate is applicable to capital funds that are not to be amortised on the facts of a particular case. However, if they do agree or if the judge decides that assumption is valid on the facts of a case, I cannot for my part see how objection can be taken. If they do not, then the rate chosen by the court should be reasoned.’

The following year in JL v SL (No 3) [2015] EWHC 555 (Fam), in a supplemental judgment arising out of a request for amplification from the wife’s counsel, Mostyn J held (at [13]–[15]):

‘It is important to remember that a Duxbury fund is usually calculated over a long period. In this case from the start of phase 1 to the end of phase 3 over 30 years are covered. Generally speaking in most human fields the best prophet of the future is the past. The key assumption is that over a longish period it can be reasonably predicted that a fund will perform in actual gross terms by 6.75% annually (i.e. 3% income yield plus 3.75% capital growth) but the owner of the fund will suffer inflation of 3% thus giving a real rate of return of 3.75%. The key datum is however the predicted actual gross performance of 6.75%. Is that a reasonable guess?

Between January 1985 and January 2015 the FTSE 100 index rose from 1277 to 6810. This corresponds to capital growth in that 30 year period of 5.74% annually. Over the same period the broader based FTSE-250 has grown by almost 8.8%, while a portfolio based on the US Dow Jones Industrial index (after allowing for the variance in the exchange rate) over the same period would have grown by over 6.9%. Thus it can be seen that the rate of capital return assumed in the Duxbury algorithm of just 3.75% is somewhat cautious, recognising that the recipient of an award is unlikely to be advised to invest the whole of her fund in equities, and that other kinds of investment are likely to achieve lower capital returns. Dividends (i.e. income yield) would also have been paid, perhaps at 3% a year. An alternative perspective is to look at P/E ratios over a long period. From 1990 to 2014 they have averaged for the UK 12.6, which translates to annual growth of 7.9%. So on the basis of history 6.75% per annum is a very reasonable guess.

Over the same period the RPI index has moved from 91.2 to 255.4. This corresponds to inflation of 3.49% annually. Again, 3% is a reasonable guess, even if there is actually no inflation right now.’ (original emphasis)

Mostyn J went on later in the judgment to say:

‘Of course there is no “standard” rate in the sense that the economic assumptions underpinning the formula are written in marble from which there can be no deviation. But the Duxbury tables are used in countless cases. Their underlying methodology and assumptions are widely accepted as the usual starting point, and where there is no countervailing evidence, the usual finishing point. In that sense they do represent an “industry standard”.’ [2]

‘As I have explained, the central datum underpinning the Duxbury algorithm is a gross actual rate of return of 6.75%. And I have further explained that, irrespective of income yield, in terms of capital growth alone the FTSE 100 has grown annually over the last 30 years by 5.74% and better than that by reference to the FTSE-250 and the Dow Jones Industrial Index.’ [3]

In 2017 in HC v FW [2017] EWHC 3162 (Fam), Cobb J was invited to consider whether the future income requirement should be computed by reference to the Ogden tables (the actuarial tables used for assessing the sum to be awarded as general damages for future pecuniary loss) or by a Duxbury calculation. The choice of tables was of particular relevance given the quasi-personal injury character of the wife’s claim, but at trial the wife’s case was presented on the basis she was content to proceed on Duxbury assumptions (producing a lower capital sum) which the court subsequently adopted.

The court was also asked to consider the Ogden tables the following year in Tattersall v Tattersall [2018] EWCA Civ 1978. The court at first instance had decided to adopt the Ogden tables in favour of Duxbury. The question for the court on appeal was whether the use of the Ogden tables was an error of law or an error which meant the judge’s award was wrong. Although the appeal court expressed a view that it would expect judges typically to use Duxbury, it held that a judge could decide to use a different method of calculation and to do so was not an error of law.

In the same year in WG v HG [2018] EWFC 84 the court was asked to select a term during which maintenance should be payable (not referable to the wife’s life expectancy) and then capitalise that term into a lump sum. The court held that this would be unfair on the wife who had income needs which needed to be met and no earning capacity with which to meet them. The court therefore concluded that if there was to be a clean break it could only be on the basis of a Duxbury calculation, commenting as follows:

‘Duxbury is no more than a tool; it is not in any way a rule that has to be followed. It has been subjected to considerable criticism not least in the return that it assumes will be made. However, it is still the tool used by judges and family lawyers alike in these cases and nobody has sought to argue in this case that there is a better way of assessing the way to capitalise lifetime maintenance. Accordingly, I will use the Duxbury tables on a normal full-life basis.’ [4]

2018 also saw the well-known decision of Waggott v Waggott [2018] EWCA Civ 727. At first instance the wife received assets worth £9.76m and a periodical payments order of £115,000 per annum (based on an income need of £175,000 per annum less an income she could generate from her free capital of £60,000 per annum on an assumed net return of 1.75%). On appeal, when discussing the scope for different rates of return which could be attributed to ‘free capital’, Moylan LJ held:

‘There are, however, clearly advantages – both in terms of providing clarity and of consistency – if the Duxbury model and the assumptions within it were to be used at least as a starting point.’ [5]

Shortly after Waggott came O’Dwyer v O’Dwyer [2019] EWHC 1838 (Fam). Mr and Mrs O’Dwyer married in 1988 and separated in 2016. For many years the husband had run a McDonalds’ franchise which had an agreed net value of just over £2.4m. The total assets were just over £5.8m which on a sharing basis meant each should leave the marriage with just over £2.9m. This part of the decision was not challenged.

The appeal related to the approach the court at first instance took to maintenance (£150,000 per annum for a term of 5 years) insofar as it amounted to the sharing of the future income stream the husband would receive through the McDonalds’ franchise which he retained in the settlement and which exceeded the wife’s needs.

When considering the appeal Francis J deducted the sum of £1.2m (being the figure the trial judge had allowed the wife to meet housing and other capital costs) which left the wife was left with £1,732,739 of free capital. Francis J held that it was then necessary to attribute an income to this sum which he did as follows:

‘One can argue endlessly about the appropriate rate of return but it seems to me that time and again the [Duxbury] assumptions in At A Glance are used and I use them for the purposes of this case. At paragraph 136 in Waggott, Moylan LJ saw clear advantages in adopting the Duxbury assumptions, and whilst of course I have a discretion to depart from them it seems to me that I should follow the lead so clearly given by him.’ [6]

On that basis Francis J found the wife could secure a return of 3.75% on her free capital which equated to an income of £64,978 per annum. This figure would be taxable in the wife’s hands in the United States (where she had returned to live) which reduced the net receipt to the wife to £52,000. On the basis that a fairer figure for the wife’s income needs was £120,000 (rather than £150,000) per annum, Francis J held that the periodical payments should be reduced to a rate of £68,000 per annum for the same term as ordered by the trial judge.

Lastly in this review, in 2021 Peel J considered a number of bespoke calculations that had been carried out by a financial adviser in ND v GD [2021] EWFC 53. He commented that the range of outcomes was enormous because he had been invited to consider life expectancy from 5 to 30 years as well as numerous options regarding the cost of care and other income and living costs. The financial adviser had also assumed combined income and capital growth of 3.62% (whereas Duxbury assumes 6.75%) and inflation of 2% (whereas Duxbury assumes 3%). After considering the calculations Peel J held (at [22]):

‘I have to say, with due respect to all who requested and sanctioned this exercise, that it has been of negligible value to me in resolving this case. In my view the parties could very easily have used the Capitalise programme to generate bespoke calculations. What matters is the figures which are put into the programme by each party to calculate the outcome contended for. Often during a hearing, as issues crystallise, the judge will ask for specific calculations to be carried out; indeed, I did just that in this case. The underlying assumptions can be adjusted on the Capitalise programme if required. I do not see that the SJE was asked to do any more than create his own Duxbury style calculations, but, perhaps inevitably, he adopted different underlying assumptions. The result is a quasi-Duxbury calculation, inconsistent with the specific Duxbury model which has stood the test of time for decades in financial remedy cases. This is not to criticise Mr Hutton-Attenborough [the expert]; he did exactly what he was asked to do, conscientiously and fairly. In my view, it was never “necessary” (to apply the Part 25 test) for him to have been instructed. Indeed, as things have transpired, and perhaps unsurprisingly, neither party really sought to rely on his figures which are so wide ranging as to be of minimal value.

Although I acknowledge that there may be the odd case where an expert is required to carry out a very clearly defined and tailored Duxbury calculation, in the vast run of cases it is inappropriate to reach beyond the Duxbury tables in At A Glance, or the Capitalise programme for a more advanced formula. For my own part, I strongly caution against the sort of exercise which was carried out here which has been a largely futile and costly exercise. There should rarely, if ever, be a need for an IFA to carry out a Duxbury style exercise which adds cost, delay, and confusion.’ [7]

‘During the hearing, as I have observed, I was presented with a number of capitalisation calculations by an IFA who used underlying assumptions which differ from the Duxbury model. In JL v SL (No 3) Mostyn J reviewed the Duxbury assumptions and concluded that they remain sound. The Duxbury model has stood the test of time since the eponymous case of Duxbury v Duxbury some 30 years ago. As has been often stated, it is a tool and not a rule. The court has the flexibility to depart from it to the extent necessary in any given case, as, for example, in A v A. There, an elderly applicant was, by reason of her age and length of the marriage, entitled to less under the Duxbury model than would have been the case had she been younger and/or the marriage shorter; the so-called Duxbury paradox. Singer J departed from the strict Duxbury application to meet this unfairness. True, the court is not barred from considering capitalisation calculations other than by the Duxbury methodology (e.g. Tattersall v Tattersall) but I am firmly of the view that there would have to be a very good reason to go down a different route.’ [8]

Drawing these cases together it can be seen that although the family court retains its notoriously wide discretion to do justice on a case-by-case basis and, despite some of the potential shortcomings of the Duxbury formula and the regular reminder that Duxbury is ‘a tool rather than a rule’, in practice the Duxbury assumptions are likely to be adopted by the court in the vast majority of cases as the net value of a right to receive periodical payments at a target annual rate for (often) the remainder of a payee’s life. Any attempt to challenge them would require the court’s permission to adduce evidence in the form of a different set of instructions which absent very good reasons would be unlikely to succeed.

Against that backdrop this article now considers (with the glorious benefit of hindsight and financial analysis) the potential net impact on recipients of Duxbury awards in previous reported decisions. In some instances assumptions have been made where the exact information was not ascertainable from the judgment. In some cases it was not possible from the information contained within the judgment to undertake this exercise but we have endeavoured to provide at least a short narrative summary of those decisions in the hope of collating a decent sized catalogue of previous reported cases involving a Duxbury award.

Although we have endeavoured to capture as many cases as possible, this article does not, for the reasons above, claim to have recorded every single Duxbury case. For a comprehensive categorisation of recent cases, see the excellent academic work of Emma Hitchings, ‘Reconsidering the Duxbury default’ [9]. The article does, however, seek to pick up on the theme mooted as early 1998 by Jeremy Posnansky when he asked whether it was time to find out what happened to the Duxbury wives and whether they had lived their lives ‘like the computer models on which their awards were based’ [10] – based on a set of calculations which may be more akin to those adopted by some of those recipients.

Assumptions

In contrasting how the model should perform compared to how often it has in practice, it is important to bear in mind the profile of a likely recipient of a Duxbury award. Duxbury recipients are often not commercially experienced individuals with a strong understanding of or appetite for risky investments [11]. They are more often likely to favour the safety of less risky financial dealings with which they may be familiar [12]. The assumptions we have used in the Appendix to this article are therefore based on how a typical cautious risk portfolio might have performed based on historic data because, in our experience, that tends to be the approach to risk favoured by the majority of recipients of Duxbury awards.

The assumptions we have used are also based on how the portfolios under the management of Saltus Partners LLP have performed historically over the last 12 years (as that is the period for which this data is available). For those cases which were decided more than 12 years ago the same data has been used. Moreover, for those cases, which are the majority, where the receiving party has not yet reached their actuarial life expectancy age, the same 12-year data has been extrapolated to project whether there will be a shortfall or surplus when they reach their actuarial life expectancy age. We have assumed inflation is the consumer prices index (CPI) measure at 2.5% per annum as the Bank of England’s target rate of inflation is 2% and the average inflation rate in the United Kingdom between 1991 and 2021 is 2.31%. This has been used for income/withdrawals and expenditure.

The Duxbury formula does not take account of how different recipients approach investing, particularly their tolerance and appetite for taking risk. We have assumed the capital sum awarded will be invested in line with a Saltus Cautious portfolio using their full Discretionary Fund Management proposition. We have also included the usual ongoing management fees that would be charged to manage an investment portfolio on a recipient’s behalf, which average around 2% per annum and include ongoing investment management, financial planning and underlying investment costs.

Ordinarily, Saltus Partners LLP would build in a variety of tax wrappers depending on the client’s individual circumstances. For the assumptions used in this article we have, however, opted for the use of an ISA which would be fully subscribed to each tax year using the General Investment Account as, without knowing an individual’s circumstances, it would be remiss to factor in other tax wrappers, such as pensions, whereas every individual (as long as they are UK tax resident) has an annual ISA allowance. We have assumed the cashflow and life expectancy of the client is in line with the ONS statistics, which in most cases is 87. The State Pension is projected at the full amount and payable from the appropriate age (usually between 66 and 68 based upon the client’s age/date of birth).

Discussion

The calculations based on how a cautious risk portfolio may have performed during the economic climate over the last 12 years indicate that every capital fund apart from two (Vaughan and S v AG [2011] EWHC 2637 (Fam), both of which involved low levels of annual income need) would have been exhausted before the end of the recipient’s actuarial life expectancy. This might not come as a surprise to the vast majority of financial remedy practitioners doing so-called bigger money work who will have probably tried (and failed) at least once to persuade the court to adopt a bespoke model rather than the Duxbury formula when capitalising a maintenance payment. That lack of surprise may, however, be scant consolation to the recipients of Duxbury awards who run out of money to meet their needs before they pass away.

It is important, however, to remember that the purpose of the Duxbury calculations is not to provide a guaranteed income for life. It is also worth bearing in mind that if the recipient had received an open-ended periodical payments order instead of a capitalised lump sum, they would have been subject to various risks including early variation or termination of the order by virtue of the death of the paying party, remarriage of the receiving party or court order following a significant change of circumstances. These were some of the very well-made points by Lewis Marks KC in ‘An Alternative View of Duxbury: A Reply’ in 2010 [13].

In addition to the (perhaps unsurprising) illustration as to the likely performance of a Duxbury fund, the analysis also suggests that there may be a difference in trends between cases involving modest Duxbury awards and larger ones. The former seem to perform less badly than the latter and one reason might be the ability of those with lower target rates of income to make greater use of tax-free allowances. Another reason may be that in cases involving lower levels of income needs a larger portion is covered by state pension, which means the recipient is not required to draw as heavily on their capital fund.

Another trend seems to be that Duxbury sums given to recipients with a longer period of life expectancy appear to perform less well, which may (despite the Duxbury adjustment to income yield in the first year) be attributable to the impact of compounding based on artificially high rates of return and/or the increased risk that a maintenance award would come to an end the longer the term. Another important factor to consider (again despite the first-year adjustment to income yield) is ‘sequencing risk’, which is the risk that investment will be subject to the worst returns at the worst time, for example the returns during the COVID-19 pandemic in March 2020. In the early years of drawing from capital when it is at its ‘peak’, weak returns could increase the risk of the recipient’s capital running out at a faster rate. Investing and immediately starting to draw from capital in periods of high volatility may not be wise. Therefore, it is often advisable to build in an emergency fund and potentially hold back the first year of income need in cash which will act as a buffer in volatile times. Whilst for the purpose of these scenarios it has been assumed that the entire sum is invested, in practice, creating this buffer will be a key consideration.

Whilst the significant benefits associated with having a universal formula which can be used to project what capital fund a party may require in lieu of a periodical payments order (including increased certainty, improved prospects of settlement and reduced cost – financial and otherwise – of prolonged litigation) should not be underestimated, it is submitted that the process by which the calculations are determined could be improved. At the moment it is a small group and little is known externally as to how often the assumptions are reviewed. It is suggested that group might, instead, be expanded into a statutory multi-disciplinary committee akin to those used to determine the band of awards in personal injury cases to recommend and review the assumptions in a transparent manner. This would retain all the benefits of a Duxbury yardstick whilst ensuring the assumptions are both appropriate and kept under review by a more independent body that is democratically mandated.

Note as a standard warning: Investments do not guarantee a return, the value and the income from them can fall as well as rise. You may not get back the amount originally invested. Past performance is no guarantee of future returns.

• 04 Appendix.docx40kb

References

[1] Tim Lawrence of what was then Coopers and Lybrand.

[2] ⁠[2015] EWHC 555 (Fam) at [17]

[3] [2015] EWHC 555 (Fam) at [18].⁠

[4] ⁠[2018] EWFC 84 at [85].⁠

[5] ⁠[2018] EWCA Civ 727 at [136].⁠

[6] ⁠[2019] EWHC 1838 (Fam) at [39].⁠

[7] ⁠[2021] EWFC 53 at [28] and [29].⁠

[8] ⁠[2021] EWFC 53 at [53].⁠

[9] ⁠[2021] CFLQ 275.⁠

[10] ⁠J Posnansky, ‘Whatever Happened to the Duxbury Wives’ [1998] Fam Law 447.⁠

[11] ⁠Sometimes in contrast to the paying party.⁠

[12] Moreover, often unlike the paying party who may have continuing income from other sources, the Duxbury payment may be the only significant funds for the recipient for the rest of their life so a cautious approach as distinct from the assumptions of the model is perhaps understandable.⁠

[13] ⁠[2010] Fam Law 614.⁠