Options for improving the Retail Prices Index

Transcription

1 Options for improving the Retail Prices Index The ABI s Response to the National Statistician s consultation The UK Insurance Industry The UK insurance industry is the third largest in the world and the largest in Europe. It is a vital part of the UK economy, managing investments amounting to 26% of the UK s total net worth and contributing 10.4 billion in taxes to the Government. Employing over 290,000 people in the UK alone, the insurance industry is also one of this country s major exporters, with 28% of its net premium income coming from overseas business. Insurance helps individuals and businesses protect themselves against the everyday risks they face, enabling people to own homes, travel overseas, provide for a financially secure future and run businesses. Insurance underpins a healthy and prosperous society, enabling businesses and individuals to thrive, safe in the knowledge that problems can be handled and risks carefully managed. Every day, our members pay out 147 million in benefits to pensioners and long-term savers as well as 60 million in general insurance claims. The ABI The ABI is the voice of insurance, representing the general insurance, protection, investment and long-term savings industry. It was formed in 1985 to represent the whole of the industry and today has over 300 members, accounting for some 90% of premiums in the UK. The ABI s role is to: Be the voice of the UK insurance industry, leading debate and speaking up for insurers. Represent the UK insurance industry to government, regulators and policy makers in the UK, EU and internationally, driving effective public policy and regulation. Advocate high standards of customer service within the industry and provide useful information to the public about insurance. Promote the benefits of insurance to the government, regulators, policy makers and the public.

2 Introduction 1. Our Members are users of, or otherwise have an interest from a number of perspectives in, the Retail Prices Index (RPI). Our overarching concern is to avoid loss of value for Members and/or their policy holders from inappropriate changes to methodology of RPI index. Inappropriate changes to RPI that reduce the value of savings and protection products linked to that index that have been purchased, typically, by those seeking secure incomes in retirement would also risk reputational damage to Government, the savings industry and the public interest objective of promotion of a responsible savings culture to meet long-term financial needs. 2. Financial services entities have both assets and liabilities linked to RPI. They are also investors in a range of industries that are subject to economic regulation based on target rates of real return. Companies in which our Members invest may also use levels of inflation for calibrating performance targets under incentivisation schemes for directors and other employees. RPI has historically been used for this purpose. 3. It has been known since 1996 that RPI tends to produce higher answers than CPI as a result of what is now termed the formula effect. The vast majority of RPI contracts, benefits and obligations have been constructed and built up, since 1996, in that very knowledge. To change the basis of calculation is to change the value of the property rights between the parties that made the contracts and their associated benefits and obligations. It would create potentially large windfall gains and losses. An example would be that pay-outs under executive remuneration incentivisation arrangements would be higher than intended for any given level of performance and shareholder value would be correspondingly reduced. 4. RPI has a history that goes back over many years, and to change the calculation so fundamentally will mean the loss of historic continuity. Economic historians using RPI to show how the cost of a basket of goods has changed over periods will have a fundamental discontinuity at We consider that the formula effect is a known and justified aspect of divergence between RPI and CPI. However, it would be helpful if ONS could produce more guidance for users as to the difference to help users determine whether RPI or CPI is more appropriate to them on future occasions when they are striking contracts that involve creating benefits and obligations that are index-linked. Main comments 6. As asset owners, our Members principal interest is in direct holdings of index-linked fixed interest instruments, both gilts and corporate and PFI bonds as well as investment exposures to property, and to utilities subject to economic regulation. The consultation document notes the potential implications of any changes to the treatment of gilts, but does not recognise the potentially disruptive effect on corporate bonds where contractual provisions vary significantly and where potentially opportunistic actions by issuers are a real risk. 2

3 7. As issuer of index-linked gilts, any de facto reneging on its debts would have significant adverse implications for the perceived creditworthiness and trustworthiness of HM Government. 8. It is evident that the concern that RPI is not providing appropriate readings vis-a-vis CPI is the result of changes introduced to clothing data collection in 2010 that were deemed to require recalibration of CPI. The increase in the wedge between CPI and RPI that has followed the changes in both indices is the legitimate focus of concern and the proper focus for action to remove unjustified differences. It is disappointing that the ONS study has not uncovered the underlying reason for the divergence which, if it needs to be emphasised, is a result of those 2010 changes and not of any change in the mathematical characteristics of the various possible averaging techniques employed at the elementary aggregation level. 9. The consequences of change to RPI construction could be profound and disruptive. This would be particularly so in the corporate index-linked market where changes that would already be detrimental to investors could lead to enforced early repayments of bonds at their indexed par. This would lead to significant danger of market value losses for investors and re-financing risks for issuers. 10. This all has the potential to severely damage confidence in the index-linked corporate market, which has until now been characterised by long term and stable debt investors offering competitive pricing via matching of assets and liabilities. UK corporate and project finance index-linked bonds have been used extensively to provide funding for UK Infrastructure. For example, of a total c. 40bn outstanding corporate index-linked bonds, some 98% is represented by UK regulated Utilities, Rail and infrastructure financing. Future financing appetite in this market has to be in question if existing investments are so damaged, aside from any wider UK financial market reputation consequences. Government's efforts to encourage investors to substantially increase the private funding of UK infrastructure, particularly bond funding as the banks continue to retrench, must be cognisant of the above risks. 11. There could also be a severely detrimental impact on equity investments in infrastructure companies, where investors have bought RPI linked businesses on the basis of existing RPI and would see an immediate loss on the change, resulting from any new index. It should also be noted that the documentation/wording is not consistent between individual index-linked corporate bonds, index-linked swaps and index-linked EIB loans i.e. as to when a change to RPI would trigger a change and as to how that change is dealt with e.g. in some cases a new index has to be created, to give the same effect as the old one; in other cases, the new index is adopted or the bonds are called at indexed par. 12. Our Members as investors wish to avoid a change that is fundamental and detrimental of the type that would require the Bank of England to make such a determination in respect of index-linked gilts. It is to be hoped (though it cannot be guaranteed) that trustees of corporate index-linked bonds and other relevant decision takers, in deciding on the implications of any change to the index, will be 3

4 guided by the decision taken by the Bank of England on its similar but not identical task in respect of the index-linked gilts market. This would allow the new RPI index to be adopted across the market and would be least disruptive, as it would avoid heavy bondholder losses, keep assets and liabilities matched for investors and issuers and reduce complexity and cost of funding for borrowers. 13. A change to the nature of the RPI, known to be constructed as a pure arithmetic average to incorporate geometric averaging techniques, would carry a high risk of being deemed fundamental. We think that such a change would in any case be unjustified as the geometric approach to be valid for inflation measurement purposes, relies on the assumption of a certain degree of product substitution inperiod 1, as embodied in the economic approach. This has never been an input to the RPI measure of the change in price of a basket of goods. Indeed, the consultation document conveys that the ONS, reflecting the nature of the latest expert advice, does not accept the applicability of this economic approach. 14. If a fundamental change to RPI is implemented, we think that the authorities would need to continue to publish the old RPI index so that borrowers do not have an excuse to call their bonds due to lack of an appropriate index. However, we do not see how that would be a better or less costly approach than retaining RPI in its current form, and it would be less useful than making appropriate modifications to the RPI index to revert to its previous nature. 15. Any fundamental change to an index that drives a significant portion of investors' asset and liability cash-flows has the potential to cause mismatch risk if the assets and liabilities are treated differently. Of significant concern to us is the documentation risk in many index-linked assets that could lead to severely detrimental outcomes in some scenarios: For investors (the vast majority of whom are insurers, pension funds and banks that have RPI linked liabilities - either in the form of annuity or pension commitments, or RPI swap payments) the adverse outcome would be one whereby there is a wholesale call of index-linked corporate bonds giving rise to a major mismatch between assets and liabilities. The risk to issuers is that, if RPI is changed such that "new" RPI is expected to be structurally lower than "old" RPI, a borrower would be at risk of having a wedge develop in their business whereby they are only earning "new" RPI but having to pay out "old" RPI. For the Authorities, the risk is that this market is brought into disrepute either through an unintentional call of substantially all of the market or as a result of protracted litigation as the fine points of each of the varying documentation of the bonds are picked over in the courts by issuers, investors and bond trustees. This has the makings of a major problem. 1 And, as noted in paragraphs 41 to 42 of the consultation document, is relevant for cost of living approximations and most accurately so in the special case where quantities consumed are in inverse proportion to the changes in relative prices 4

5 16. We have examined the arguments advanced and the options canvassed in the consultation document and supporting papers and we wish to make the following observations: 17. The RPI is what it is a measure of the price changes of basket of goods over a discrete period. It is a cost of goods index (which does not include all items of household expenditure or indeed all households). The CPI, which employs geometric averaging techniques, implies a degree of product substitution proportionate to the relative price changes of goods at the elementary aggregation stage and therefore is more in the nature of a cost of living index based on an economic approach. This economic approach may well be appropriate if an index is to be used for, say, macro-economic planning purposes, but may be less so for purposes such as approximating to any individual s experienced inflation rate. It is to be expected that the RPI s methodology will tend to give a slightly higher result. This has been priced into all contracts that are RPI-linked including RPI-linked annuities and index-linked gilts and corporate bonds. 18. If technical problems have arisen in the RPI index, then an appropriately targeted technical solution is required. We do not agree that the 2010 changes have, as suggested in Annex E of the consultation document, given an improved approach. We note that the rationale provided by ONS and by Professor Diewert has moved away from paying some regard to the economic approach to one based on whether particular index techniques pass certain statistical tests. We are aware of criticism that these tests, which suggest that the Jevons approach should be preferred on its technical merits, are, to a large extent, recapitulations of each other. More significantly, the important weakness of the geometric construction of the Jevons approach noted in the consultation document (Table 2) that if any one observation out of a set of observations is zero, their geometric mean is zero, whatever the value of the other observations, was not considered as part of this evaluation. 19. The Jevons method is alien to RPI s arithmetic construction. A Carli to Jevons swap in essence exchanges known upward bias for downward bias. The Dutot approach does not incorporate systematic bias upwards or downwards but will tend to give undue weight to the inflation in higher priced items. The choice as to use of Carli or Dutot, the two key forms of arithmetic average, is more obviously a technical matter on which users would wish statistics professionals to reach an appropriate determination. Summary of our views on the four options presented 20. We consider that the consultation paper has neither produced an adequate analysis of why RPI may since 2010 have been providing an inappropriate reading nor have options for change been canvassed that rectify the problem appropriately. This would more than justify a choice at this time of Option 1 No change. 21. Option 2, given its clear and appropriate focus on changes to calculation of inflation in clothing, represents the most plausible attempt to achieve the necessary rectification. However, it does not address the root of the problem and no evidence 5

6 is provided that the alternative techniques would provide a more correct answer, bearing in mind their acknowledged weaknesses downward bias in the case of Jevons and bias towards inflation in higher priced goods in the case of Dutot which is the reason that the latter is not already employed for clothing. We also emphasise that we do not agree with an objective of aligning inflation measurement of clothing between RPI and CPI. We therefore conclude that if Option 2 is taken forward it should be on the basis of amending the arithmetic averaging methodology but we think that more work is still required on properly understanding the relative merits of Carli and Dutot We do not support Option 3 which proposes changes that go well beyond the scope of what is required to identify and eliminate unjustified elements of the formula effect that have arisen as a consequence of changes to collection of clothing prices. 23. We do not support Option 4 as we do not support incorporation of geometric averaging techniques in RPI. If Option 4, or an unmodified Option 3, were to be introduced we consider that publication of old RPI should continue at least until maturity of all legacy gilts and bonds. ABI 30/11/12 [s:\inv\consultations\other domestic\onsrpioc12 2 We note, for example, that Annex E of the consultation document identifies that prices collected in the base month (January) now better reflect consumer spending patterns: clothing products that are in a sale are collected in January, in those locations and for those products where prices are being collected for the first time. The dangers of incorporating such data may be disproportionately significant for the Carli indexing approach but considerably less so for the Dutot arithmetic averaging technique which is acknowledged to have a bias to reflecting the inflation in higher priced items. We note the consultation document s indication that replacement of Carli with Dutot would give the same or a greater reduction in inflation than would the substitution of Jevons even with the latter s acknowledged tendency to understate due to its geometric construction. These facts tend to suggests that the Dutot result may itself be an understatement in so far as it is acceptable for prices to be included for the first time in the index by virtue of significant transactions in a sales environment. In this regard it is likely therefore that the most correct arithmetic average determination may lie at an intermediate level between Dutot and Carli. 6