Was it correct for the judge and the Court of Appeal to reduce the arithmetical multiplier, and therefore, in effect, override the expectation of life agreed by the doctors? Mr. Owen submitted that there could be no rational basis for applying a further discount for "contingencies," since the doctors had already taken account of all the contingencies that might affect the plaintiff, such as the increased risk of accident, chest infection, and so on. The only reason given by the judge was that the courts had "tended to reduce multipliers by about 20 per cent." The Court of Appeal took the same line.

      I can see no answer to Mr. Owen's argument. The inevitable result of reducing the multiplier to 17, as Mr. Havers Q.C. pointed out, will be that the plaintiff's damages will run out when he is 39. He will have nothing to cover his needs for the remaining 21 years of his life.

      Mr. Havers conceded that there is room for a judicial discount when calculating the loss of future earnings, when contingencies may indeed affect the result. But there is no room for any discount in the case of a whole life multiplier with an agreed expectation of life. In the case of loss of earnings the contingencies can work in only one direction--in favour of the defendant. But in the case of life expectancy, the contingency can work in either direction. The plaintiff may exceed his normal expectation of life, or he may fall short of it.

      There is no purpose in the courts making as accurate a prediction as they can of the plaintiff's future needs if the resulting sum is arbitrarily reduced for no better reason than that the prediction might be wrong. A prediction remains a prediction. Contingencies should be taken into account where they work in one direction, but not where they cancel out. There is no more logic or justice in reducing the whole life multiplier by 15 per cent. or 20 per cent. on an agreed expectation of life than there would be in increasing it by the same amount.

      It follows from what I have said that I do not agree with the discount which McCullough J. allowed in Janardan v. East Berkshire Health Authority [1990] 2 Med.L.R. 1. In that case the plaintiff, aged five at the time of trial, had a life expectancy to 55. This indicated a multiplier of just under 20 by reference to a 4.5 per cent. discount rate. McCullough J. held that a discount was required to allow for the possibility that the plaintiff might not survive to 55. I do not accept this; and it may be that McCullough J. would not have accepted it either, if he had not felt constrained by previous authority. Left to himself, he said, he would have taken a whole life multiplier of 17.5 to 18. But the multiplier chosen by the Court of Appeal in Croke v. Wiseman [1982] 1 W.L.R. 71 and by the House in Lim Poh Choo v. Camden and Islington Area Health Authority [1980] A.C. 174 required him to choose 17 instead. But this meant, as Mr. Havers pointed out, that the plaintiff in Janardan's case would have run out of damages at the age of 33, although he was expected to live to 55.

      In Hunt v. Severs [1994] 2 A.C. 350 the plaintiff had a life expectancy of 25 years. The appropriate multiplier by reference to a 4.5 per cent. discount rate was 14.821. But the judge reduced this figure to 14 because 14 seemed more in line with the multiplier applied in other comparable cases. The Court of Appeal, correctly in my view, substituted a multiplier of 15, as being the nearest round figure to 14.821. Sir Thomas Bingham M.R. observed that an allowance for contingencies may sometimes be appropriate. But he continued:

 "Such an allowance is not appropriate in the present case, where the agreed life expectancy of the plaintiff is 25 years. That is a fact, or rather an agreed assumption, upon which the damages payable for future care must be based."

But the House disagreed. Lord Bridge of Harwich said, at p. 365:

 "The passage I have cited from the judgment of the Court of Appeal appears to show the court as treating the circumstance that both doctors in evidence estimated the plaintiff's expectation of life at 25 years as establishing the 'fact' or 'assumption' that she would live for 25 years and thus converting the process of assessing future loss into 'a simple arithmetical calculation.' I cannot think that this was the correct approach to the evidence. A man or woman in normal health, at a given age, no doubt has an ascertainable statistical life expectancy. But in using such a figure as the basis for assessment of damages with respect to future losses, some discount in respect of life's manifold contingencies is invariably made."

      I have some difficulty with this passage. The plaintiff's life expectancy was not derived from any tables. It was the agreed life expectancy of this particular plaintiff, taking her individual characteristics into account. I cannot for my part see what further room there was for "life's manifold contingencies." The whole point of agreeing a life expectancy, if it can be done, is to exclude any further speculation. With respect therefore I prefer the approach of Sir Thomas Bingham M.R. and the Court of Appeal.

      The explanation for the different approach of the House in Hunt v. Severs may be a continuing hesitation to embrace the actuarial tables. I do not suggest that the judge should be a slave to the tables. There may well be special factors in particular cases. But the tables should now be regarded as the starting-point, rather than a check. A judge should be slow to depart from the relevant actuarial multiplier on impressionistic grounds, or by reference to "a spread of multipliers in comparable cases" especially when the multipliers were fixed before actuarial tables were widely used. This may be the explanation for the relatively low multiplier chosen by the House in Lim Poh Choo's case.

      For the reasons I have given, I consider that the Court of Appeal in the present case were wrong to substitute a multiplier of 17 for the judge's 23. But the judge himself was also too low. The appropriate multiplier derived from the tables on the agreed life expectancy was 26.58.

      I can deal with the second of the two miscellaneous points in Thomas v. Brighton Health Authority quite shortly. In October 1990, 15 months after the plaintiff's birth, and five years before the trial, the plaintiff's parents moved into a larger house. They needed more space, because of his disability. The additional cost was some £60,000 which they raised by way of a mortgage. The question is how the additional cost should be reflected in the award of damages.

      Obviously the plaintiff is not entitled to the additional capital cost, since the larger house is a permanent addition to the family's assets. It will be there, and could be realised, at the end of the period covered by the award. How then should this head of damages be calculated? Should it be the interest on the mortgage? or interest calculated in some other way?

      The answer to this question, described in Kemp and Kemp as "a satisfactory and elegant solution," was provided by the Court of Appeal in Roberts v. Johnstone [1989] Q.B. 878. It is to be assumed that the plaintiff will pay for the additional accommodation out of his own capital. It is further to be assumed that the capital input will be risk-free over the period of the award, and protected against inflation, by a corresponding increase in the value of the house. What the plaintiff has therefore lost is the income which the capital would have earned over the period of the award after deduction of tax.

      But the lost income is not to be calculated by reference to a normal commercial rate of interest. For interest, as Lord Diplock explained in Wright v. British Railways Board at p. 781, normally includes two elements, "a reward for taking a risk of loss or reduction of capital," and "a reward for foregoing the use of the capital sum for the time being." Since the capital input in the new accommodation is free of risk, or virtually free of risk, it is only the second of the two elements of interest that the plaintiff has lost, namely, the "going rate" for forgoing the use of money. The Court of Appeal in Roberts v. Johnstone took 2 per cent. as the "going rate." This was the figure originally chosen by Lord Denning M.R. in Birkett v. Hayes, and accepted by Lord Diplock in Wright v. British Railways Board. Birkett v. Hayes and Wright v. British Railways Board were both cases of non-pecuniary loss, but the point is the same.

      Both sides accept that the correct approach is that adopted by the Court of Appeal in Roberts v. Johnstone. The only question is how that approach should be applied. Collins J. arrived at the "going rate" by taking the average return on I.L.G.S. as the best possible indicator of the real return on a risk-free investment over the period of the award. In other words, he took the same discount of 3 per cent. net of tax as he had taken for the calculation of future loss. The Court of Appeal disagreed. They took the "conventional rate" of 2 per cent., pointing out that Stocker L.J. had not tied his 2 per cent. to the return on any particular form of investment.

      It is true that there is no reference to I.L.G.S. in Roberts v. Johnstone. But in Wright v. British Railways Board Lord Diplock chose the return on I.L.G.S. as the first (and in my view simpler) of the two routes by which courts can arrive at the appropriate or "conventional" rate of interest for foregoing the use of capital. At that time the net return on 15-year and 25-year index-linked stocks was 2 per cent. I can see no reason for regarding 2 per cent. as sacrosanct now that the average net return on I.L.G.S. has changed. The current rate is 3 per cent. This therefore is the rate which should now be taken for calculating the cost of additional accommodation. It has two advantages. In the first place it is the same as the rate for calculating future loss. Secondly it will be kept up to date by the Lord Chancellor when exercising his powers under section 1 of the Damages Act. On this point I would restore the order of Collins J.

Page v. Sheerness Steel Co. Plc.

      The judge took a loss of earnings multiplier of 20.28, which he reduced by 8.65 per cent. to 19 in order to take account of the substantial risk that the plaintiff would not have continued working until the age of 62. The Court of Appeal took a multiplier of 14 including a discount of 17.9 per cent., on the ground that there was merit in Mr. Leighton Williams' submission that the discount should be at least 12.5 per cent. But the judge had already said that he was giving a larger discount than average, because of the nature of the plaintiff's work. There was no need for the Court of Appeal to intervene.

      The judge took a whole life multiplier of 24 based on a rate of return of 3 per cent. net on a normal life expectancy. The Court of Appeal started with a multiplier of 19 and reduced it to 17. For the reasons already mentioned there was no justification for a discount on the whole life multiplier. I would therefore restore the judge's multiplier of 24 as the correct starting-point.

      But as the plaintiff is currently being looked after by his wife, the judge sensibly and correctly split the multiplier in two. He applied a multiplier of 12 to the additional cost currently incurred while Mrs. Page is the prime carer, and the same multiplier for the further costs when she is no longer able to act as the prime carer. There is no dispute on that score.

      Part of the current cost of care is to provide the services of an "enabler," that is to say, "someone who will take the plaintiff out of the house, stimulate his interest in sporting and other activities and generally seek to motivate him and improve the quality of his life." Mrs. Gipson said that four sessions a week would be appropriate, but the judge reduced this to three sessions or 12 hours a week for the reasons he gave, and calculated the damages on the basis of a 52-week year. The Court of Appeal accepted that the cost of an enabler was justified, but reduced the period from 52 weeks a year to 40 weeks, on the ground that there would be breaks in the period of employment, as when the plaintiff is on holiday. It may be that some account should have been taken of holiday periods. But the judge's estimate was not so far wrong as to justify the Court of Appeal's interference.

      Nor was there any sound reason for increasing the discount on damages for loss of pension rights from 10 per cent. allowed (reluctantly) by the judge in the light of Auty v. National Coal Board [1985] 1 W.L.R. 784 to 15 per cent. allowed by the Court of Appeal.

      Finally Mr. Purchas argued that there ought not to be a reduction for the proceeds of permanent health insurance, on the ground that the plaintiff has contributed 4.5 per cent. of his earnings to the scheme. Mr. Purchas sought to distinguish Hussain v. New Taplow Paper Mills Ltd. [1988] A.C. 514 on that ground. The judge did not accept Mr. Purchas's argument. He dealt with this point at some length and his views were upheld in the Court of Appeal. In my view they were right.

For the reasons I have given I would allow all three appeals.

LORD STEYN

My Lords,

      I confine my observations to the principal issue about the discount rate to be applied.

The conventional rate

      It has for many years been settled practice, endorsed by decisions of the House of Lords, that the lump sum to be awarded in a personal injury action for the present value of a plaintiff's future losses of earnings, and the present cost of his future expenses, ought to be determined by using in the calculations a discount rate of 4 to 5 per cent. The rationale was that there was no other practicable basis of calculation that is capable of dealing with so conjectural a factor as inflation with greater precision: Cookson v. Knowles [1979] A.C. 556, 571G, per Lord Diplock; see also the earlier and subsequent decisions of the House of Lords in Mallett v. McMonagle [1970] A.C. 166 and Lim Poh Choo v. Camden and Islington Area Health Authority [1980] A.C. 74. The question before the House is whether this practice should now be modified in changed economic circumstances by adopting a discount figure assessed by reference to index-linked government securities, the suggested figure being 3 per cent.

The importance of the issue

      The importance of the issue is shown by a comparison of the awards of the three trial judges, who relied on index-linked government securities to fix the discount rate, and the figures substituted by the Court of Appeal, who used the conventional rate. Ignoring for present purposes that in Wells v. Wells the trial judge adopted a discount rate of 2.5 per cent. the use of a 3 per cent. discount rate instead of 4.5 per cent. would increase the awards by very roughly the following sums:

      These figures show the impact of the reduction of the discount rate in cases where damages are calculated over many years. But, since a judicial decision ought to take into account general as well as particular consequences, it is important to realise that the proposed modification of the discount rate would lead to an enormous general increase in the size of awards for future losses. Nobody has ventured a prediction of the likely cost. The sums involved would undoubtedly be huge. The implications of a modification of the conventional rate for the insurance industry would be considerable. Inevitably, it would be reflected in increased premiums. The one certain thing is that if the right decision is to make the suggested modification of the discount rate the public would by and large have to pay for the increase in awards.

The premise of the debate

      The premise of the debate was that as a matter of law a victim of a tort is entitled to be compensated as nearly as possible in full for all pecuniary losses. For present purposes this mainly means compensation for loss of earnings and medical care, both past and future. Subject to the obvious qualification that perfection in the assessment of future compensation is unattainable, the 100 per cent. principle is well established and based on high authority: Livingstone v. Rawyards Coal Co. (1880) 5 App.Cas. 25, 39; Lim Poh Choo v. Camden and Islington Area Health Authority [1980] A.C. 175, at 187E, per Lord Scarman. The technique employed to achieve this result is to provide an annuity of an annual amount equivalent to the streams of future losses of earnings and cost of future expenses: Hodgson v. Trapp [1989] A.C. 807, per Lord Oliver of Aylmerton, at 826D-E.