MEMORANDUM TO: Members of the CAA FROM: SUBJECT: Neil Dingwall, Chairman, CAA Standards Steering Committee Principles and Considerations for the selection of the Discount Rate Curve – Discussion Draft DATE: 22 April, 2013 At the 2010 Caribbean Actuarial Association Annual Meeting, Council approved the development of a comprehensive set of Actuarial Standards that will apply to work performed in the Caribbean region. The objectives of the project were outlined as follows: Caribbean Actuarial Standards will become the Caribbean Standard. That the Caribbean Actuarial Standards will be consistent with International Actuarial Standards. The overriding purpose of CAA Practice Standards is to serve the public interest by ensuring that the users of actuarial services in the Caribbean benefit from a high quality of actuarial work. An additional motivation has come from the regulators in the region who have expressed a desire for the C.A.A. to take an active role. As a profession, selfgovernance and self-regulation should be our preference. In accordance with due process, Council appointed a CAA Standards Steering Committee to produce the Standards. The CAA Standards Steering Committee has divided the project into various segments, some of which are able to run concurrently and some which will need to be sequential. The drafting work has been performed by the Practice Committees and a number of new and revised standards were exposed and adopted by the membership. Attached is a discussion document on the Principles and Considerations for developing the discount rate curve in preparation for the development of a separate practice standard on this issue Also included is the Actuarial Standards Architecture background paper. In commenting on and reviewing the attached Discussion Draft, we would ask that, as far as is practicable, you follow the instructions in Appendix 1 of this memo. Also, it Page | 1 CAA DD APS2.1 22 April 2013 would be of benefit to the committee, if you could answer the questions contained in Appendix 2. After receiving and discussing comments it is planned to draft a practice standard and expose for comment. Sessions will be held at the Annual meeting to allow for further discussion and allow the Practice Committees to report both the comments received and what further amendments they would recommend in making their recommendation for adoption by Council at their meeting. The CAA due process for the approval of new standards requires an initial six month exposure period. As this is a discussion paper the steering committee has set a deadline of 30 June 2013 for receiving all comments and feedback. Comments, suggestions, recommendations for amendments and any other feedback on Discussion Draft APS2.1 should be submitted by email to the chairs of the Life subcommittee, The Steering Committee looks forward to communication with the members and with other interested parties, who wish to raise issues of concern throughout this process. Neil Dingwall. Page | 2 CAA DD APS2.1 22 April 2013 Appendix 1: Instructions for Commenting 1. The deadline for official submission of comments is 30 June 2013. 2. The draft has been circulated in Microsoft Word format in a two column layout. Please insert your comments in line with the relevant paragraph in the second column. 3. Unless necessary, please do not use Microsoft Word’s “Track Changes” features to indicate your changes. 4. As part of the review process, your comments may be shared with the wider membership but individual views will not be identified with any specific member. 5. Submissions specific to the Principles and Considerations should be emailed to the relevant Sub-committee Chair as outlined in the memorandum. 6. Any comments on the overall approach or related to the CAA’s due process can be addressed to the Co-Chairs of the CAA Life Committee, Simone Braithwaite at simone.brathwaite@oliverwyman.com and Marcia Tam-Marks at mtammarks@morneaushepell.com 7. Please include a completed version of Appendix 2 with your submission. 22 April 2013 Page | 3 CAA DD APS2.1 22 April 2013 Appendix 2: Questions for Members, CAA Advisory Council, Regulators and other interested persons. This is a discussion draft on Principles and Considerations in determining the discount rate curve. This is a working document and therefore your comments and feedback are required. Q1: Does the draft Principles and Considerations document adequately explore the issue of discount rate setting? Q2: Is there any additional guidance that you would like to see included? Q3: Are there other issues relevant to the discussion that are not addressed in the document and that you would like to see addressed? Q4: Is there any aspect of the Principles and Considerations document that you consider inappropriate or unnecessary? Q5: Are there other sources of information that you would recommend we look at to facilitate the discussion? Q6. The focus of the draft Principles and Considerations document is primarily with respect to the Life Insurance practice. Do you think that the application of this discussion document should be extended to include Pensions and Social Insurance and if so, what are the additional issues that should be addressed? Q7: Do you have any comments or suggestions on the next steps in the process? 22 April 2013 Page | 4 CAA DD APS2.1 22 April 2013 CAA LIFE COMMITTEE CARIBBEAN ACTUARIAL ASSOCIATION PRELIMINARY DISCUSSION DRAFT Modification of Existing Standards Caribbean Actuarial Association Standard of Practice APS2.1: Principles and Considerations for the selection of the Discount Rate Curve A Preliminary Discussion Draft of the Committee on Actuarial Standards of the CAA Issued: 22 April 2013 Comments Due: 30 June 2013 This Exposure Draft of a discussion paper for Actuarial Practice Standards APS2.1 is issued by CAA Life Committee for comment. Written comments should be addressed to: simone.brathwaite@oliverwyman.con or mtammarks@morneaushepell.com by 30 June 2013 CAA DD APS2.1 22 April 2013 6.2CAA Life Committee Initial Views on UFR Contents 7 Question 5: What is the best extrapolation method? 1 Executive Summary 2 1.1Summary Action List on Discount Rate 2 1.2Summary of Principles and Industry standard-setters 4 2 Introduction 2.1Five key questions 3 Question 1: What are the available sources of market data? 8 9 11 3.1Principle – Use of all relevant observed market data 11 3.2CAA – LC initial view 14 4 Question 2: How should credit risk, sovereign risk and illiquidity premium be accounted for? 17 4.1International views on Credit risk and Sovereign risk 17 4.2CAA-LC Initial Views on Credit Risk and Sovereign Risk 22 4.3Illiquidity risk 23 4.4CAA-LC initial views on illiquidity risk 24 5 Question 3: At which point is the last liquidity point? 39 42 7.1How to grade from the last observed liquid forward rate to the ultimate rate 42 7.2Principles and key criteria for the choice of the extrapolation method 44 7.3Convergence 46 7.4Initial CAA-LC View 48 7.5The extrapolation should follow a smooth path from the entry point to the unconditional ultimate long-term level 48 Appendix A A.1. Survey on asset portfolio of insurance companies Appendix B 49 49 56 B.1. Principle 1: Extrapolated market data should be arbitrage-free56 B.2. Principle 3: The extrapolation method should be theoretically sound 56 26 5.1Summary of methods and considerations used in more developed markets 27 5.2CAA-LC initial views on the last liquidity point 28 5.3Survey of 7 Expert practitioners in the region (TBC) 33 6 Question 4: Choice for the unconditional ultimate long-term level 35 6.1Solvency II considerations 37 Page | 1 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 1 Executive Summary This paper summarises considerations and principles promulgated by the various International bodies and forums on the development of a discount curve which could be adopted for the Caribbean insurance industry. It also provides a discussion and outlines the CAA Life Committee’s (CAA-LC) initial views on how these international principles and guidance can be applied in the Caribbean context. Summary Action List on Discount Rate 1.1 1. Agree on Principles (Life Committee). 2. Develop a document with the CAA-LC’s initial views on Principles and Guidance that should underlie the development of the discount curve in the Caribbean market. 3. Conduct research on what could be considered all relevant observable market data in the Caribbean market to base the discount curve on. Many insurers have the available longer tenor local government bonds in their asset portfolio. These bonds while typically not very liquid can give a good reflection of the cost of the Page | 2 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments replicating portfolio and should therefore be considered in building the risk free curve. Research: ¯ Nature of the most common and/or highest quality assets commonly used to support life insurance liabilities ¯ Request companies to provide more detailed information on feasible reinvestment strategies – target mix, duration etc. 4. Liaise with Central Banks and other industry experts in the region to solicit their input on what could be considered as the last liquidity point for each currency. 5. Test and create an appendix comparing potential extrapolation approaches/methods for the development of a discount rate curve. 6. Draft a Practice Standard for the Development of the Discount Rate Curve based on the information gathered, discussions held, and comments received. Page | 3 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 1.2 Summary of Principles and Industry standard-setters 1.2.1 Draft Principles and Educational Notes which are to be considered Body Key Reference papers Issue Date Link IASB -IFRS Phase 2 Update on changes to 2010 ED April 2012 http://www.ifrs.org/CurrentProjects/IASBProjects/InsuranceContracts/Documents/compar isonEDApri2012.pdf IASB IFRS Phase 2 Feedback on Discount rate decisions December 2011 http://www.ifrs.org/CurrentProjects/IASBProjects/InsuranceContracts/Documents/7Bdisc ountrate.pdf IASB-IFRS Phase 2 Exposure draft July 2010 http://www.ifrs.org/CurrentProjects/IASBProjects/InsuranceContracts/Exposure-draft2010/Documents/ED_Insuran ce_Contracts_Standard_WEB. pdf IAA IAA Educational Monograph_2012 0810 FINAL September 2012 http://www.actuaries.org/CTT EES_ACTSTD/Documents/Disc ount_Rate_Monograph_Cover ing_Note.pdf EIOPASolvency II QIS5 Technical Specification – Risk-free interest rates & Omnibus 2 June 2012 Page | 4 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments EIOPASolvency II CEIOPS-L2-FinalAdvice-on-TPRisk-freerate[1].pdf – CEIOPS-DOC-3409 CRO forum October 2009 https://eiopa.europa.eu/publi cations/sii-final-l2advice/index.html http://www.ceiops.eu/conten t/view/14/18/ MCEV (CFO forum) Page | 5 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 1.3. CAA Life Committee initial views on the most Frequently Referenced International Principles Principles Source Details CAA-LC View Reflects the characteristic of the underlying liability cash flows IFRS 4 An insurer shall adjust the future cash flows for the time value of money, using discount rates that reflect the characteristics of the insurance contract liabilities. Such rates: a. Shall be consistent with observable current market prices for instruments with cash flows whose characteristics reflect those of the insurance contract liability, in terms of, for example, timing, currency and liquidity. b. Exclude any factors that influence the observed rates but are not relevant to the insurance contract liability (e.g. risk not present in the liability but present in the instrument for which the market prices are observed).” Agree with the Principle. It is high-level but all encompassing Exclude credit risk/Sovereign risk IFRS/Solv ency II The proposed IFRS 4 Phase II rules [Ref 12] suggest the use of discount rates derived from market values where the impact of credit risk is minimal, i.e. “risk free interest rates”. Agree to exclude credit risk where possible/ measurable. The exclusion of Page | 6 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Hence, if the underlying asset used has credit risk associated with it, the impact of the credit risk on the yield must be removed to produce the discount rates used Include a matching adjustment or illiquidity premium Proportionality :Keep method simple and pragmatic Sovereign risk is subject to debate Agree where characteristic of the underlying liability warrants and where quantifiable EIOPA – L2 Advice on the risk free interest rate term structure The choice of the risk-free rate should not depend on the nature, scale and complexity of the risks of the undertaking. However, as all undertakings should be able to discount their liabilities, consideration needs to be given to how the risk-free interest rate term structure should be made available, particularly for small and medium sized undertakings which might not be in a position to derive the term structure themselves. EIOPA believes that it is necessary to provide both the term structure and the methodology used to derive the term structure for all major currencies. Page | 7 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 2 Introduction A properly chosen discount rate would provide the value of insurance liabilities being adequate and market consistent. In small economies this is difficult to achieve due to the unavailability of deep and liquid market for bonds. There exist industry papers, other academic investigations and modern practices dealing with the question of discounting in insurance accounting and for solvency purposes. However, a very small part of the studies really touch such a sensitive subject as the choice of the relevant discount rate in the small economy environment, where the securities denominated in the local currency are not traded in a liquid and deep market. We appreciate that most of the key questions to be answered in the development of an appropriate discount rate vector are fundamentally the same in all markets as such the CAA will aim to be consistent with international principles and educational notes were possible. However, the challenge is that to date, despite comprehensive analysis of the related issues by various bodies, there is no clear Page | 8 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments guidance for setting up the interest rate term structure for non-Euro zone smaller economies. As such, the CAA-LC is aiming to develop a proposal that will be reflective of the economic characteristics of the Caribbean and that needs to be pragmatic given our specific limitations. We will also need to be explicit as to the reasons for any deviation from international standards and principles. 2.1 Five key questions The five key questions which need to be addressed in the development of a discount rate vector for insurance liabilities in all markets are: 1. Which assets categories are appropriate for use as a starting point to develop market based risk free rates? 2. How should credit risk, sovereign risk and liquidity premium be accounted for? What methods and what source of information is available and what process would be appropriate? 3. At which point is the last liquidity point-: At what duration does the data become unreliable and where extrapolation becomes Page | 9 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments necessary? What is the appropriate method and process for determining this point? • For each currency, the “last liquid point” needs to be determined – i.e. the point at which we ignore market data and start to extrapolate. 4. What is an appropriate approach for determining the ultimate forward rate? 5. What is the best extrapolation method? Each of the following sections aims to address the five questions in the context of the Caribbean market giving consideration to proposed international standards, educational papers and any relevant useful methods already in use in the regional insurance industry. Page | 10 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 3 Question 1: What are the available sources of market data? 3.1 Principle – Use of all relevant observed market data A fundamental principle for the development of an appropriate discount curve is the use of all relevant market data. The principle is fundamental as if ignored then objectivity of the valuation would be compromised and the cost of the replicating portfolio would differ from the calculated liability. 3.1.1 Relevant market data as per International guidance Solvency II – criteria and 3 stage approach Ideal characteristics for risk free rates: Solvency II Criteria1 1. No credit risk: the rate should be free of credit risk 2. Realism: It should be possible to earn the rates in practice 3. Reliability: the determination of the rates should be reliable and robust 4. High liquidity: the rates should be based on financial instruments from deep, liquid and transparent market 5. No technical bias CEIOPS' Advice for L2 Implementing Measures on SII: Technical Provisions Risk free interest rate 1 Page | 11 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Government bonds rates of AAA rated governments should be considered as the benchmark for credit risk-free rates. Swap rates are not credit risk-free and for this reason unadjusted swap rates should not be used to discount technical provisions. For each currency, CEIOPS proposes to follow a three stage approach to determine the relevant risk-free interest rate term structure: First stage If government bonds are available that meet the risk-free criteria as defined (minimal credit risk, reliable, realistic etc.)then government bonds should be used to determine the relevant risk-free rates. Second stage If government bonds are available, but they do not meet the risk-free rate criteria, then they should be adjusted for their deficiencies relating to these criteria. The adjusted rates should approximate government bond rates which meet the risk-free criteria. The adjusted rates should be used to determine the relevant risk-free rates. Third stage If government bonds are not available or if government bond rates cannot be adjusted to meet the risk-free rate criteria for practical or theoretical reasons, other financial instruments can be used to derive Page | 12 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments the risk-free interest rates. These instruments should be as similar to government bonds as possible. Their rates should be adjusted for credit risk and any other deviations from the criteria with the objective to approximate government bond rates which meet the risk-free criteria. Where government bonds do meet the risk-free rate criteria (or can be adjusted to meet them) for some maturities but not for all maturities, they should be used to derive the relevant risk-free rate for these maturities only. CRO forum Definition of relevant market data [ref CRO forum] Applicable to Caribbean Market instrument has characteristics that are similar to a component of the liability Yes? Data is from a liquid market – that is, participants can rapidly execute large-volume transactions with little impact on the price OR No Based on expert judgment allow to include market data that does not meet the strict liquidity requirements, but is still considered to be a good reflection of the cost of the replicating portfolio Yes Market data is reliable, trade and quote information of these prices is supplied by third parties and accessible and verifiable (e.g. multiple providers similar price) to market participants No Page | 13 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments IFRS Phase 2 The proposed IFRS 4 method allows for the use of either the top-down or bottom up approach to developing the risk free discount curve. The IAA monograph does not indicate a preference for one approach over the other. The bottom up approach – is a “build up” approach – to the appropriate discount curve, which is possible if there is a reliable liquid market source for “risk free rates” e.g. such as high quality Government bonds and swaps. The top down approach is a “peel back” approach whereby, market rates on a greater set of assets can be used as a starting point however those elements not characteristic of the liability cash flows must be identified quantified and eliminated. In theory, the top-down approach should provide more flexibility but that is only to the extent there are other appropriate sources of liquid assets available to be used as a starting point. 3.2 CAA – LC initial view In most Caribbean currencies there are no available assets that would meet the criteria of “risk free” without needing adjustment. In particular, the AAA rating benchmark criteria is difficult to meet in the Page | 14 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments current environment anywhere. However based on the applicable guidance from the EIOPA Solvency II/CRO forum and the IFRS proposal, there are possible other sources of market data in the Caribbean market to consider as an appropriate starting point and are listed in the table below: Asset type Appropriateness as a starting point Available Liquidity/ Reliability Comment Government bonds/T bills Yes Yes but Limited Limited Potential need to be adjusted for sovereign risk and possibly technical bias at some maturities to proxy risk free Swaps other derivatives Yes No N/A Corporate bonds Yes Limited Limited Publicly rated corporate bonds can be adjusted to proxy risk free Mortgages No Yes No Subject to prepayment risk Policy loans No Yes Yes Equity No Yes Limited Not reflective of NonPar liability characteristics Page | 15 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Based on our survey of the supporting asset portfolios of selected insurance companies in the Caribbean market – we have observed that Government bonds are still the primary asset type followed by mortgages and corporate bonds. CAA life committee believes that this is a reflection of what is available and probably most appropriate in the local capital markets. See Appendix A for survey of asset portfolios of insurance companies in the region. 3.2.1. Initial Recommendation: CAA Life Committee recommends that Government bonds (both on-therun and off-the-run securities) and Corporate bonds be considered as possible starting points to develop appropriate discount rate. Page | 16 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 4 Question 2: How should credit risk, sovereign risk and illiquidity premium be accounted for? 4.1 International views on Credit risk and Sovereign risk Body Position on credit risk Position on Sovereign risk IFRS Exclude – Exclude Solvency II Exclude Depends on materiality IAA Exclude Exclude The IFRS principles suggest that all elements of the market rate used which are not characteristic of the insurance liability be removed. As credit risk is not relevant to nonparticipating insurance contract liabilities, a risk-free rate is usually used as the base discount rate. A risk-free interest/discount rate represents a limiting point so in general it is accepted that it would contain minimal risk and exclude credit risk and sovereign risk [Ref IAA monograph]. Page | 17 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments There are more consensuses among the international standard-setters that spreads related to credit risk are not characteristic of the underlying liability and should be removed. There is more debate on whether sovereign risk, the illiquidity risk premium and other non-credit related spreads should also be eliminated. Solvency II The credit standing of an AAA rated government should serve therefore as a benchmark… An adjustment should be made only in cases where government bonds are inappropriate, for example because of technical bias or liquidity considerations. This adjustment should be made following a clear, reliable and well-established methodology IAA In the past, actuaries measured and analyzed credit risk by looking at historical experience. The historical experience was presumed to be reasonably reflective of expected future conditions and was therefore used to project credit risk and defaults. However, the historical approach is of limited use when analyzing sovereign risk. Among other factors, historical experience of sovereign default is extremely limited Page | 18 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments and the facts and circumstances of each country are unique. The following approaches are sometimes used to measure sovereign risk: credit default swaps, Z-spreads, the Solvency II approach, historical data and expert opinion. Since 2009, mounting fears over a Eurozone sovereign debt crisis have lengthened risks of sovereign debt restructuring and outright default. More globally, the financial turmoil in these years has also changed the classical view to the ‘risk-free’ notion of default on government bonds. For this reason, government bonds are no longer immediately accepted as “risk-free”. Measurement of Credit risk and Sovereign risk The components of the market spread over the risk-free rate consist of risk premiums for credit risk/sovereign risk, illiquidity risk and other residual items. However, trying to separate credit-risk and the non-credit related risk premiums in the market spread is not an easy task. There is substantial academic discussion on the measurement of credit risk in the return on financial instruments. Several approaches for the quantification of credit spreads separately from the other spread components could be considered. It appears that the more sophisticated methods require a Page | 19 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments liquid market as well. Best practice which would ensure comparable quantification outcomes has not yet emerged. Current quantification methods: credit risk It would be ideal to develop a method which would derive a basis point adjustment to the observed rate corporate bonds based on their rating. Methods which rely on a CDS market are not feasible; as such the traditional historical factor approaches, despite the cons (outlined in the IAA monograph) are still the most realistic option to be applied in this region. Given the lack of historical data specific to the region, we may still need to rely on the assumptions used for other bonds with similar ratings. Currently some Caribbean insurers rely on North American benchmark factors for the credit spread based on the public rating of the bond. These benchmark factors for credit risk are generally taken from analysis and actuarial guidance developed in North America based on North American credit markets over 10 years ago. Even though the market data is outdated and is not specific to the Caribbean market, the rational is that the qualitative analysis to develop a public rating on a security is generally the same and as such the use of updated North American benchmarks for credit spread by rating (and Page | 20 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments possibly maturity) may still be the most pragmatic option until adequate credible risk data specific to the Caribbean is available. Current quantification methods: sovereign risk There is a growing literature on sovereign default risk premia for developed countries, which show that the government’s ability and willingness to service its debt is country specific. The ability can be assesses by looking at the macroeconomic indicators, but not the “willingness” cannot be assessed solely by looking at these indicators the political factor is to be considered which varies by country. Given the lack of readily available literature on the sovereign default risk premium in smaller economies and the country-specific nature of its assessment, we recommend that the CAA Life Committee to liaise with and review the methods of, regional agencies, such as, CDB CariCRIS, which also rate sovereign debt and regional corporate debt before adopting the ratings of International rating agencies. The CDB recently conducted a study on sovereign risk in the region from which we may also be able to leverage. As mentioned above, before this is quantified we need to assess at what level of provisioning against sovereign risk in the reserves is necessary and appropriate. Page | 21 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 4.2 CAA-LC Initial Views on Credit Risk and Sovereign Risk CAA-LC also believes that the main objective for the derivation of the risk-free term structure is to ensure that it includes as little credit spread from the private sector instruments as possible. With respect to sovereign risk, Government bonds in many regions are generally accepted as risk-free because the likelihood of government failing to honor their commitments is extremely low in most cases. The credit standing of an AAA rated government and AAA rated corporate bonds should serve therefore as a benchmark. In the Caribbean, Government bonds are currently rated well below this, see table below. Given the current lack of higher quality fixed instruments in the region, these government bonds are still the best starting point. We need to assess whether insurers should be required to insure against sovereign risk in this context and to what extent. That is, CAA-LC believes that we first need to agree in principle as to what level of provision for sovereign risk is appropriate if any and whether it should be held in reserves and/or in capital. To the extent that insurance companies are required to insure against sovereign risk and/or the risk is “insurable”, then the use of government bonds to derive the underlying discount rate may not eliminate the need for a potential allowance for sovereign risk. Page | 22 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments S&P Current Country Rating2 Outlook Date of rating Jamaica B- Negative 2011-11-29 Trinidad A Stable 2011-11-29 Barbados BBB- Negative 2012-02-20 Bahamas BBB- Stable 2012-02-20 4.3 Illiquidity risk The concept of the illiquidity premium is associated with the “predictability” of the liability cash flows. There is less consensus in the global insurance industry on the principles related to its inclusion or exclusion and its measurement. 2 Body Position on inclusion/exclusion Position on measurement IFRS Include Not clear Solvency II Include Referred to as Matching adjustment IAA Not clear None CRO Forum Include For S&P, a bond is considered investment grade if its credit rating is BBB- or higher. Bonds rated BB+ and below are considered to be speculative grade, sometimes also referred to as "junk" bonds. Page | 23 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 4.4 CAA-LC initial views on illiquidity risk At a theoretical level, the CAA Life Committee understands the rationale behind the inclusion of an illiquidity premium and agrees with the Solvency II CEIOPS task force principles on illiquidity which state that: (1) The products to which the adjustment is applied have no future premiums, no possibility of lapse (or if so, there is a surrender value such that no loss is made) and the only underwriting risks that exist are longevity and expense. Therefore, we are basically looking at immediate annuity contracts. (2) principle #3 states that “The liquidity premium applicable to a liability should not exceed the extra return which can be earned by the insurer by holding illiquid assets free of credit risk, available in the financial markets and matching the cash flows of the liability.” However, there are also other allowances within that spread which do not relate to default, downgrade or liquidity. To avoid debate about whether the “premium” actually relates to liquidity or other elements as well, the term “liquidity premium” was replaced by “matching premium”. More recently, it has been viewed that the term premium Page | 24 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments is potentially confusing and therefore, the latest Solvency II text refers to a “matching adjustment”. 4.4.1. CAA-LC initial recommendation The Caribbean capital market is very illiquid and transaction costs are high. As such most insurance companies utilize a buy and hold strategy, especially for Government bonds and other high-quality/investment grade securities. As such the “matching adjustment” is usually earned unless there is a default. Given this practice, for the sake of pragmatism, until there is more direction from international on the separate quantification of the illiquidity premium to be added under the “Bottom-up approach”, the Life committee will not be proposing a separate method to identify and quantify an illiquidity premium which is to be added to the risk-free rate. Initial Recommendation CAA life committee advocates the inclusion of illiquidity premium/matching adjustment to the extent it can be reasonably estimated separately from default/credit spreads Page | 25 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 5 Question 3: At which point is the last liquidity point? That is, at what duration does the data become unreliable? The most available instruments in the region are typically short duration instruments. However, insurance liabilities are typically of a significantly longer duration. Therefore, it is necessary to determine the longest duration on the for which we have reliable data and extrapolate until we have a curve that is of long enough duration to be usable for discounting insurance liability cash flows (A Whole Life product to a woman aged 25, the cash flows could easily extend for 70 years). For each currency, the “last liquid point needs to be determined. – i.e. the point at which we ignore market data and start to extrapolate. Currently, under the proposed Solvency II the last liquid point for the major currencies has been assumed as follows: Currency Last Liquid point EUR 20 years GBP 50 years USD 30 years Page | 26 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Summary of methods and considerations used in more developed markets 5.1 Solvency II regulators and international accountants (IASB/FASB) have drafted rules and guidance with the aim of defining which prices should be used and in what circumstances unreliable prices may be replaced with other valuation measures. However, the current rules and guidance are not well aligned (between SII and IASB/FASB) and there is a dearth of detailed thinking on how to turn aspirational regulation into practical reality. IASB/FASB (FAS 157/IFRST 7 Fair value Hierarchy • Active Markets: ─ Level 1 – Mark-to Market using quoted prices • No Active Market ─ Level 2 - Mark-to model using observable inputs ─ Level 3 - Mark-to-model using significant unobservable inputs Where an active market provides for “quoted prices which are readily and regularly available from an exchange, dealer, broker, industry Page | 27 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments group, pricing service or regulatory agency, and those prices represent actual and regularly occurring market transactions on an arm's length basis.” EIOPA- Solvency II (DLT test) As stated in the CEIOPS—DOC-35/09 advice on the calculation of technical provisions as a whole (Article 86c of the Level 1 text), 'deep, liquid and transparent markets', are required to meet all the following requirements: 1. DEEP & LIQUID - market participants can rapidly execute largevolume transactions with little impact on the prices of the financial instruments used in the replication 2. TRANSPARENT - current trade and quote information of those prices is readily available to the public 5.2 CAA-LC initial views on the last liquidity point Most of us would struggle to say where the cut-off lies for Level 1Active market vs. Level 2/3-No Active market under the accounting guidance. Solvency II DLT criteria is even more problematic in practice which define that reliable prices must trade in deep, liquid and transparent markets. It seeks to measure properties of the market that are difficult to assess in even developed capital markets: Page | 28 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments • Depth: requires an understanding of the normal market size of trades – is this an absolute or relative amount, and if relative, relative to what? • Liquidity – this can be interpreted as a measure of the impact of abnormal trade sizes – again this requires clarification of what would considered abnormal relative to the definition of normal above • Transparency – some further qualification on the information required which would make a market transparent is needed. Even with further future clarification on the DLT criteria from EIOPA, our apriori assumption is that, in the Caribbean market, very few instruments would pass this strong DLT test. However, where market prices are observable but fail the DLT requirement we believe that some weight should be given to these points in the observed part of the curve. There is not a very clear line between market data points that are liquid/active and should be fully incorporated in the observed part of the curve and market data points that are not liquid/not active and should not be taken into account. As such, the CAA life committee believes that it is difficult to develop easy to apply “criteria” and “techniques.” We should rather refer to “considerations” that need to Page | 29 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments be taken into account for determining the last observed liquid data point. Some level of expert judgment will ultimately be required to determine the last liquid market data point based on such considerations. Key is that the market data used should be “a good reflection of the cost of the replicating portfolio”. Below are a summary of some considerations and a survey of opinions of various practitioners in the regional industry. 5.2.1. Key Considerations Below a few considerations on how to deal with data points before and after the last liquid market data point. 1. Use of multiple data sources • CAA-LC recommends the use of Government bonds market in assessing the last liquidity point. However, if other relevant data (Corporate bonds, swaps) has availability at tenors beyond Government bonds then this should also be used. • The question is how these extra market data points from different sources can be included in the risk free yield curve while maintaining consistency with the principles of yield Page | 30 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments curve extension; In particular avoiding discontinuities in the curve that could result in arbitrage opportunities. One should also take into account that both sources of market data (e.g. corporates and government bonds) show a difference (spread) for the observed part of the curve. ─ Given the importance of the slope of the forward curve as the starting point for the extrapolation one has to be careful in fitting the slope between two market data points based on different sources 2. The actual investment of insurance companies • Especially in less developed market (E.g. Caribbean) there is often no deep and liquid market. However, many insurance companies have the available longer tenor local government bonds in their asset portfolio. These bonds while typically not very liquid can give a good reflection of the cost of the replicating portfolio and should therefore be considered in building the risk free curve. 3. Consider quantitative measures if available • Quantitative measures include bid-ask spread, trading volume or frequency. However, not straight forward to Page | 31 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments arrive at a consistent rule-based approach that works for all economies/currencies. 4. Consider the circumstance under which the market price was arrived • Market data that is biased because of a forced liquidation or distress sale should not be directly used. 5. The slope of the last observed liquid forward rate should be relatively stable • CRO Forum – Extrapolation Paper recommends that market data points that are subject to significant unbalanced market conditions will promote unnecessary instability in the liability and should be ignored. Extrapolation starts from the last observed liquid market data. So any instability in such point is extended to longer tenors. • The mechanism that provides stability is that in a crisis situation the transition point of where market data is used and where extrapolation starts is moved to an earlier point. • How can such an unbalance be observed? Page | 32 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 6. Considerations for using interest rate data points that are not fully incorporated. Market data points not included in the observed part of the curve. Such points should give a sensibility check on the slope of the extrapolated part of the curve. 5.3 Survey of 7 Expert practitioners in the region (TBC) Currency Apriori view on liquidity point Trinidad TT$ 20 yrs (3 responses) 15 yrs ( 2 responses) Barbados 15 yrs (1 response) Bahamas 1 yr (1response) 5.3.1. Initial recommendations and next steps Possible Considerations CAA-LC View Action/Timeline The actual investments of insurance companies Yes – survey completed 2012 Quantitative measures – such as bid- ask spread N/A None The circumstance under which the market price was arrived at – cannot use forced liquidation or distress sale data Agree None Page | 33 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Market best practice on last liquid data point Yes – market opinion to be surveyed annually 2012 –market opinion survey by currency September 2012 Market data beyond the last observed liquid market data point Agree None Extrapolation using multiple data sources Agree – Subject to peer review by Life Committee Stressed market conditions Agree None OTC markets N/A Page | 34 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 6 Question 4: Choice for the unconditional ultimate long-term level In the previous section the CAA Life Committee is in agreement that the extrapolation should follow a smooth path from the entry point to the unconditional ultimate long term level. Both the entry point and the unconditional ultimate long term level are key assumptions. The entry point assumption is addressed in Section 4. The unconditional ultimate long term level is addressed in this section. There is still much global industry debate regarding the ultimate forward rate (UFR) Theoretically the view on the ultimate long-term level should be forward looking, reflecting: • market analysis, • future expectations expected real interest rates and inflation, • economic theory and • historic experience. This view should be documented appropriately and regularly reviewed. In practice, The UFR is driven by the long term expected interest rates and long term expected inflation. The other theoretical Page | 35 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments drivers of the long term forward rate are ignored and are not often explicitly considered. One can distinguish three types of extrapolation: Method Application Use Applicable to Caribbean 1. A simple direct extrapolation of the market price; Keeping the last observed liquid market forward data point constant for the entire extrapolation Short extrapolations Investment banks Not appropriate where extrapolations longer than 5 years from last liquid point are required. for long-term insurance liabilities beyond 5-15 years depending on the currency 2a. Solvency II approach -components of the market price are extrapolated based on a macro-economic view of the long-term market price (i.e. not necessarily the same as in the last observed liquid market data). Insurance liabilities Solvency II Yes Considered best practice by Solvency II Irrelevant of the amount of market data available Page | 36 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 2b. extrapolation to a best-estimate long-term level; i.e. expected “real world” realization of interest rates or volatilities. Current Canadian approach where UFR is based on rolling average of historic rates Life insurance Practical but no econometric link to future rates Conservative approach when rates are rising Solvency II considerations 6.1 Under SII principles the unconditional long-term forward rate should be relatively stable over time and only change due to fundamental changes in long-term expectations and should not be affected materially by short-term economic changes. The unconditional long-term forward rate is the sum of forward looking expectations of the: i. Long-term Inflation Level ii. Long-term Real Rate iii. Long-term Term Premium Page | 37 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments The current Solvency II regulations set a UFR of 4.2% for every economy. This is split as 2.2% long term growth and 2% inflation. The 2% long term inflation assumption is consistent with the inflation target of most central banks, although there are various economists who argue that this inflation target (particularly for the European Central Bank) is unrealistically low and should be revised. The current Solvency II regulations assume that the UFR is the same for all economies. Again, this assumption is open to debate. Work performed for QIS5 demonstrated that there should be at least 3 different “buckets” of economies to reflect different long term inflation environments – for example, Japan should arguably have a lower long term inflation assumption given the long period of deflation seen there, whereas Turkey has had consistently high inflation and has had a target of between 5% and 7.5% over the last 4 years. The long term real growth rate assumption is also seen to be quite different across economies over the long term. Even when considering only the second half of the 20th century (the first half being heavily impacted by high inflation due to the wars), real returns ranged from 1% in the Netherlands to 4.7% in France. However, the Page | 38 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments average return over all economies over the second half of the century is 2.3%, which gives some justification to the 2.2% assumption. CAA Life Committee Initial Views on UFR 6.2 In the Caribbean context, we have situations where there are only observable interest rate forwards for the first 5-10 years. This creates a number of complications: • Yield curves tend to be steep on the shorter end of the curve. This could result in extrapolation to extraordinary levels. • Yield curves tend to be more volatile in the shorter end of the curve. This could result in extraordinary volatility in the extrapolated curve. • The fact that the liquid curve ends at 5 years indicates that there is no natural supply of long dated fixed income. • Any rational player who would offer a 45 year guarantee would be exposed to a very large amount of risk and would want to be compensated for this in excess of the best estimate level. Page | 39 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments For this type of situation it is desirable to use SII method, as it allows taking into account the long term nature of the product and can be based on external economic forecast. 6.2.1. Initial CAA-LC Recommendation on UFR Extrapolation should be based on forward rates converging from one or a set of last observed liquid market data points to an unconditional ultimate long-term forward rate to be determined for each currency by macro-economic methods (e.g. a long term growth assumption plus long term inflation assumption). The table below provides the 2016 (4th year) forecast rates of real GDP and inflation (from Oxford Economics) which would be an indication of their long term view. Other regional sources should be used and compared before setting the UFR. Country Expected long term Real GDP forecast Expected inflation UFR Bahamas 2.6% 2.5% 5.5% Barbados 3.0% 2.2% 5.2% Jamaica 1.7% 5.2% 6.9% Trinidad 3.9% 3.3% 7.2% United States 2.9% 2.3% 5.1% Page | 40 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments We are also of the view that should the UFR should be relatively stable over time and only change due to fundamental changes in long-term expectations and should not be affected materially by short term economic changes. Page | 41 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 7 Question 5: What is the best extrapolation method? 7.1 How to grade from the last observed liquid forward rate to the ultimate rate Defining an ultimate rate translates the extrapolation problem of yield curves into an interpolation (“grading”) problem so that it can be treated with known methods as described in the following. The grading methodology chosen has a large effect on the variation in yields beyond the last observed liquid market data point. Spurious variation in these yields can feed through to give excessive balance sheet and valuation variation that should be avoided. With this in mind it becomes important to develop a robust extrapolation method that reflects both current market conditions and empirical views of long rate volatility, while simultaneously displaying adequate stability. Page | 42 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments There are many different extrapolation methods in existence which have been considered by international standard setters: 1. The Smith-Wilson method 2. The Nelson-Siegel method 3. The Svensson method (in a macroeconomic formulation) 4. The method used and developed by Barrie & Hibbert 5. A linear approach (as adopted for QIS5)?? Body Position on Extrapolation method Solvency II The Smith-Wilson method is currently the one being proposed for use by Solvency II CRO forum The CRO Forum supports the grading method based on Nelson-Siegel developed by B&H as being appropriate, but alternatives can be evaluated in order to achieve the goal of a curve with the desired properties with less technical efforts. IASB No preferred method identified IAA No preferred method identified As with any extrapolation method, certain assumptions are required – the key assumptions being: Page | 43 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 1. Entry point of extrapolation (i.e. last liquid point as discussed in Section X) 2. The ultimate forward rate (UFR) (i.e. the rate that the curve should converge, addressed in Section X) 3. The maximum period of convergence (i.e. at what point the curve will be considered close enough to the UFR) 4. The speed of convergence (i.e. how quickly the curve will converge) Each of these assumptions is open to judgment – for example, we have already discussed the considerations on determining the entry point of extrapolation. 7.2 Principles and key criteria for the choice of the extrapolation method Once we know where we start (last liquidity point) and where we should converge to (UFR), we then need to determine how that should happen and how quickly. The extrapolation method determines the grading process, the shape and speed of convergence to the UFR. Page | 44 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Principles on extrapolation promulgated by international standard setters and other bodies/educational notes are listed below and should be considered in the choice of method to grade from the entry point (last observed liquid forward rate) to the ultimate forward rate. Principle Source/Body CAA Life Committee view 1 Use all relevant observed market data where available CRO forum Agree 2 Extrapolated market data should be arbitrage free CRO forum Agree. Where possible 3 Extrapolation method – theoretically sound CRO forum Agree. Where possible 4 Extrapolation should follow a smooth path from the entry point to the unconditional ultimate long term level From a financial stability perspective extrapolation methodology is important as it can exuberate or dampen volatility in the financial markets into the entire industry CRO forum Agree 5 Keep method simple and pragmatic 9 The extrapolated part of the basis risk free interest rate Agree Solvency IICEIOPs Agree Page | 45 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments 10 7.3 curve should be calculated and published by a central institution, based on transparent procedures and methodologies, Taskforce report 2010 The ultimate forward rate should be compatible with the criteria of realism and the principles used to determine the macroeconomic long-term forward rate should be explicitly communicated. Solvency IICEIOPs Taskforce report 2010 Agree Convergence As the last observed liquid forward rate is the starting point of the extrapolation, such forward rate has a high impact on the start and direction of the forward curve. Therefore, it is key that in fitting the curve to the observed market data that there is some smoothing mechanism that avoids that small inconsistencies in the last observed liquid market data points can result in big swings in the level and the shape of the last observed liquid forward rates. The Economic Monetary Affairs Committee (ECON) has suggested that the risk free curve should converge to the UFR (or, at least, not be materially different) 10 years after the last liquid point. However, EIOPA have suggested that the period of time should be 40 years. Page | 46 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Clearly these are significant differences in approach that could have material impacts on the technical provisions. For the Euro, this would mean that we would assume that after 30 years we reach the ultimate forward rate (i.e. the last liquid point is 20 years and the period to convergence is 10 years). Some observers have noted that this is too short a period of time and it is worth noting that the shorter the period of time until the UFR is reached, the more sensitive the extrapolated portion of the yield curve is to the choice of UFR. Note that the Wilson-Smith model does not specify a point when the UFR is reached – instead it incorporates a speed of convergence (∝). The Solvency II default value for ∝ is 0.1. Furthermore, the extrapolation method should not result in abrupt changes in the shape of the forward curve. This would result in spurious volatility in the extrapolated part of the curve. Any method used should be back-tested for a range of currencies over a significant time horizon (5 years) to show that the volatility in the extrapolated part of the curve shows the same declining pattern as observed in the observed part of the curve. Lastly, there are often market instruments available beyond the last liquid market data point. Therefore, such instruments lie in the Page | 47 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments extrapolated part of the curve. Although these instruments are not extremely liquid, they can still be used in effectively managing the risks. 7.4 Initial CAA-LC View 7.4.1. Initial recommendation and next steps on extrapolation method: The two extrapolated methods recommended by CRO forum and Solvency II – the Smith Wilson and Nelson-Siegel method developed by B&H should be back-tested. If both methods are equivalent in their fit – perhaps the most simple method should be used – or alternatively a linear method can be developed as done for QIS5 in Solvency II (TBD). The speed of convergence should consider the strength and reliability of the available market data, specifically the last liquidity points. The impact of variance convergence speeds should also be tested and understood. See Appendix B for further background on extrapolation. 7.5 The extrapolation should follow a smooth path from the entry point to the unconditional ultimate longterm level Page | 48 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Appendix A A.1. Survey on asset portfolio of insurance companies Asset distribution data in the life insurance industry was gathered for the following countries: Asset distribution data in the life insurance industry was gathered for the following countries: • Jamaica –2011 data for 3 companies • Trinidad and Tobago – 2011 data from the Central Bank of Trinidad and Tobago mid-year report • Barbados – 2010 data for 1 company, published financials • Bahamas – 2011 data for 3 companies, 2 published financials, 1 regulator • Belize - 2010 data for 6 companies, from regulator • ECCU – 2011 data for 1 company, published financials Asset Category Bahamas Cash 2% Short term Deposits 6% Barbados 9% Belize ECCU Jamaica 3% 3% 1% 56% 56% 2% T&T 8% Page | 49 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Government Securities 27% 50% 16% 16% 67% Regional Bonds 4% Extra Regional Bonds 0% 48% Corporate Bonds 5% 0% 4% 4% 2% 16% Equities 4% 7% 1% 1% 1% 13% Preferred Shares 2% Mortgages 20% 25% 5% 5% Policy Loans 10% 8% 2% 2% 1% Fixed Assets 5% - 4% 4% 1% Subsidiaries 2% - 3% 3% 1% Receivables 9% - 3% 3% 13% Goodwill 1% - Investment Properties 8% - Mutual Funds Longest Bond Term 4% 2037 6% 2046 The objective was to determine the category of assets backing the liabilities and the distribution of these assets among the various categories. The information would also serve as an indicator of the availability of types of assets. Page | 50 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Another point of observation was to determine the longest term of government bond being held by the companies. Based on the above, we also gathered information from the various Central Bank websites to determine the duration of the Government bonds being issued and those that are outstanding. Outstanding Bond Issues As at September 30 2011 Country Type Yrs to Maturity Local Currency Barbados Fixed Rate (Gov't) 22.8 Jamaica Fixed Rate (Gov't) 12.4 Floating Rate (Gov't) 20.4 Fixed Rate (Gov't) 13.4 Floating Rate (Gov't) 17.0 Fixed Rate (Gov't Agency) 18.7 Floating Rate (Gov't Agency) 11.3 Fixed Rate (Corp) 11.8 Floating Rate (Corp) 18.8 St. Lucia Fixed Rate (Gov't) 6.8 St. Vincent Fixed Rate (Gov't) 4.9 Bahamas Floating Rate (Gov't) 26.0 Belize Fixed Rate (Gov't) 18.0 Trinidad and Tobago US$ 4.4 15.6 6.0 Sources: First Citizen's Investment Services: Fixed Income IQ 30 September 2011 Central Bank of Bahamas Central Bank of Belize Page | 51 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Country Maturity of existing issues Term of new Issues 1 Bahamas 2029 Barbados 20342 Belize n/a ECCU 20183 90 days 180 days 2 Jamaica 2032 Trinidad and Tobago 20302 19 yrs in 20114 1 Central Bank of Bahamas 2 Fixed income IQ September 2011, First Citizens Investment Services 3 Easter Caribbean Central Bank 4 Central Bank of Trinidad and Tobago Our research indicated that there are at least 2 agencies that develop yield curves for Barbados, Jamaica and Trinidad and Tobago on a regular basis. • First Citizens Investment Services Research Department (CMMB) publishes quarterly its own development of the yield curve and a list of bonds in the marketplace and their corresponding yields. Here is a list of the bond information available at September 2011, for Trinidad and Tobago, Barbados, Jamaica and other countries in the region. Page | 52 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments ─ Government of Trinidad & Tobago Fixed Rate Bonds (TT$) ─ Government of Trinidad & Tobago Fixed Rate Eurobonds (US$) ─ Government of Trinidad & Tobago Floating Rate Bonds (TT$) ─ Government of Trinidad & Tobago Agency Fixed Rate Bonds (TT$) ─ Government of Trinidad & Tobago Agency Floating Rate Bonds (TT$) ─ Trinidad and Tobago Agency Amortizing Fixed Rate Bonds (TT$) ─ Trinidad & Tobago Corporate Fixed Rate Bonds (TT$) ─ Trinidad & Tobago Corporate Floating Rate Bonds (TT$) ─ Trinidad and Tobago Zero-Coupon Bonds (TT$) ─ Government of Barbados Fixed Rate Bonds (BB$) ─ Regional Fixed Rate Bonds (US$, TT$ & EC$) ─ Government of Jamaica Fixed Rate Bonds (J$) ─ Government of Jamaica Fixed Rate Bonds (US$) ─ Government of Jamaica Floating Rate Bonds (J$) The following method is used to determine the Spot (Zero) Curve: Page | 53 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments ‘The spot curve used in the calculation outlined above is generated by our proprietary valuation system. The Research Department uses recent market issues and market reads to construct a par yield curve each month. These par rates are uploaded into our valuation model upon approval of the Group’s Market Risk Committee. Our valuation model generates the key points on the zero curves using a bootstrapping methodology and uses cubic spline interpolation to generate all the points between the key points. For example if the 30 day, 90 day , 1 year, 2year…..20 year par points are uploaded, the system will generate the matching zero points and use cubic spline interpolation to determine the points between 30 day, 90 day , 1 year, 2year…..20 year spot rates. The spot rates generated are entered in the respective sheets in the bond pricing model. Further cubic spline interpolation is done between the points generated using an excel Spline function to get the exact yield for the respective time to the cash-flow.’ • Bourse Securities out of Trinidad and Tobago produces similar information. It is not clear with what frequency the yield curve is developed. The Central Bank of Trinidad and Tobago has recently issued a guideline for determining residential real estate mortgage reference Page | 54 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments rates. These rates will be calculated quarterly by the Central Bank and would be based on applicable Treasury bond yields and the cost of funds in the banking system. Banks are to use this as a reference rate as they negotiate the actual mortgage rate with the individual which would include a risk premium for individual credit rating, term and amount of the mortgage, etc. As at the end of June 2012 the rate is 3.25%. Page | 55 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments Appendix B B.1. Principle 1: Extrapolated market data should be arbitrage-free B.1.1. Rational Arbitrage opportunities should be avoided to ensure that the bestestimate is appropriate. For example, when the interest rate curve is extrapolated based on the last observed liquid spot rate this implies in most cases that you observe an abrupt changing in slope of the curve, which implies a jump in the instantaneous forward interest rate. Any arbitrage free extrapolation therefore should have a continuous curve plus a continuous forward curve. Extrapolating based on forward rate achieves this by construction and is therefore a good starting point for any extrapolation method. B.2. Principle 3: The extrapolation method should be theoretically sound B.2.1. Rational It is important that the chosen method is theoretically sound. Extrapolation methods that are in violation of generally accepted Page | 56 CAA DD APS2.1 22 April 2013 Preliminary Discussion Draft 22 April 2013 APS2.1: Principles and Consideration of the selection of the Discount Rate Curve Comments economic theory or clearly observed historical patterns should be avoided. From a financial stability perspective extrapolation methodology is important as it can exuberate or dampen volatility in the financial markets into the entire industry. The extrapolation method should not result in spurious volatility. Volatility arises from genuine changes in expected levels of long-term best-estimates and changes in risk premia. Thus the results of the extrapolation method should be consistent with the observed patterns of market data. For example, realized interest rate volatility is typically declining with the tenor of the rate and the result of the extrapolation method should be consistent with that. Theoretical models calibrated using available market data typically try to capture well described properties of market data such as mean reverting behavior. Therefore such theoretical models provide a solid basis for any extrapolation method. Moreover, by construction such models are arbitrage-free. Page | 57 CAA DD APS2.1 22 April 2013