EUROSTAT. September 2, Ernst & Young Actuaires-Conseils Contract No Eurostat

Transcription

1

2 EUROSTAT Independent examination of the 2009 update of the 2004 actuarial assessment of the balance of the Pension Scheme of European Officials (PSEO) in accordance with Annex XII Rules for implementing Article 83a of the Staff Regulations September 2, 2009 Ernst & Young Actuaires-Conseils Contract No Eurostat

3 Table of contents 1. Executive summary 3 2. Introduction Subject and objectives Scope of work Distribution & Use 6 3. Current actuarial framework Content of Annex XII Concepts and definition of Annex XII 8 4. Eurostat valuation Census data Eurostat results 9 5. Examination of the actuarial modelling Examination of assumptions Preliminary considerations Main findings 12

4 1. Executive summary In accordance with its assignment entrusted by Eurostat, Ernst & Young Actuaires- Conseils, as an independent expert, has conducted an actuarial examination of the 2009 update of the 2004 actuarial assessment of the pension scheme referred to in Article 83a (3) of the new Staff Regulations for which Eurostat endorsed the technical responsibility towards the Commission according to the Annex XII Rules for implementing Article 83a of the Staff Regulations. This examination consisted in confirming the relevance and the reliability of the actuarial processes and assumptions used in accordance with the methodology described in Annex XII to evaluate the contribution rate of officials. For the aspects explicitly mentioned in Annex XII, we did not detect any elements likely to cast doubt that the calculations determined by Eurostat are not compliant with the rules of Annex XII. For the aspects not described explicitly in Annex XII, we have checked their compliance with the generally accepted actuarial practices and we did not detect any significant difference with the methodology and the assumptions we would have chosen ourselves to perform the calculations. Ernst & Young Actuaires-Conseils has no reasons to doubt that the official s contribution rate (11.3%) calculated by Eurostat is a sufficiently accurate estimate of the reality. 3

5 2. Introduction 2.1 Subject and objectives The assignment of Ernst & Young Actuaires-Conseils (hereafter the Independent Expert ) under the Contract Eurostat consisted in the performance of several tasks relating to the cost of the PSEO. The main task was to provide assistance to Eurostat as a qualified independent expert by performing an independent review of the 2009 actuarial assessment of the balance of the pension scheme referred to in Article 83a (3) of the new Staff Regulations for which Eurostat endorsed the technical responsibility towards the Commission according to the Annex XII Rules for implementing Article 83a of the Staff Regulations (hereafter Annex XII ). The assessment is used to determine the contribution of the officials to the PSEO. This report concerns only the mentioned main tasks 5a and 5b i.e. Validate the Eurostat actuarial assessment of the balance pension scheme referred to the Article 83a and Annex XII of the Staff Regulations. Annex XII entered into force in May 2004 and defines the technical rules for the assessment of the balance of the PSEO. According to Annex XII article 13, Eurostat was to be assisted by the Independent Expert on the implementation of the methodology described in this document, and on the definition and calculation of the corresponding actuarial assumptions. We have then conducted an actuarial examination of the contribution rate calculated by Eurostat. This examination consisted in confirming the relevance and reliability of the actuarial processes and assumptions used in accordance with the methodology described in Annex XII. Nevertheless, the decisions of the Working Group Art.83 are predominant. When no rules or less precise guidance were provided in Annex XII or by the Working Group Art.83 to Eurostat regarding the methodology to be applied and/or the determination of certain assumptions, the assistance has allowed the Independent Expert to make recommendations on acceptable rules in accordance with the generally accepted international practices of actuarial assessment of pension schemes. Concerning actuarial assumptions, the Independent Expert performed diligences to ensure that the underlying data provided by Eurostat were used correctly. 4

6 We raise your attention that this review should not be considered as a critical review of the Annex XII rules that reflect negotiations on the split of the cost of the PSEO between parties, but as independent analyses of the Eurostat implementation of this Annex XII. 2.2 Scope of work Work performed Assistance work and examination diligences have been performed in tight cooperation with the Unit C6 Remuneration and Pensions; Purchasing Power Parities of Eurostat - leaded by Mr. Jean-Claude Roman and particularly the Pension team. The team included Mr. Jose Maria Olivares Ramos, Mr. Christopher Garland and Mr. Alan Zakar. We within this report refer to this team as the Eurostat Team. The methodology described in Annex XII specifies some aspects of the actuarial assessment in great details; however, in some respects, only general guidance is given so that it was necessary for Eurostat to choose between various different options for the purposes of modelling. We discussed the relevance of these options with the Eurostat Team during several technical meetings. Within the scope of our assignment, we performed the following work on the review of the actuarial assessment of the PSEO: 1. We have examined the global methodology used by Eurostat for the modelling of the different components of the contribution rate; this part of our work consisted in two different phases: a. Check that the different points set out in Annex XII were taken into account as described and were also duly taken into consideration in the Excel worksheets prepared by Eurostat Team; b. Check that the proposals made by Eurostat Team for the different subjects that were not addressed precisely in Annex XII were in conformity with the generally accepted international practices of actuarial assessment of pension schemes. 2. We have examined the assumptions used for the computation of the contribution rate. We have particularly focused our review on the compliance with the statements made in Annex XII. As for some methodological aspects, Annex XII only gives some general recommendations for the assumptions and it was necessary for the Eurostat Team to work on them to some extent. We gave our opinion on the way they were constructed. We raise your attention to the fact that: 5

7 We have considered that the formulas proposed in Annex XII had to be used as they were, We did not undertake any procedure to check the accuracy or the validity of provided data Reliances & Limitations The report presents an actuarial analysis of data, conditions, and practices communicated to Ernst & Young Actuaires-Conseils and described in the section entitled Work Performed. Ernst & Young Actuaires-Conseils believes these communications to be reliable. An actuarial assessment is an estimate of future events, the outcomes of which are unknown at this time. A certain level of uncertainty and variability are inherent to any actuarial assessment. As a result, it is possible that actual experience may be different than the estimates prepared by Eurostat, and such difference may be material. As such, we cannot guarantee that future experience will be as expected in this report. In performing our examination, we relied on data and qualitative information provided by the Eurostat team regarding the quality, accuracy, and completeness of the data and information supplied to us. Any inaccuracies or inconsistencies in the data could have a significant effect on our analysis. In the course of our analysis we tested the data provided to us for reasonableness and consistency according to our experience. We found no material discrepancies or inconsistencies in the data. Should any inaccuracies be found in the data, we should be notified so that our analysis can be adjusted accordingly. 2.3 Distribution & Use This report has been prepared for internal use of the EU Institutions. Any other use or distribution of this report without our express written consent is not authorized. Ernst &Young Actuaires-Conseils assumes no responsibility for any loss or damage that might arise from the use of, or reliance on, this report for any purpose other than for a description of the analyses underlying the opinion relating to the actuarial assessment of the balance of the PSEO in accordance with Annex XII Rules for implementing Article 83a of the Staff Regulations. 6

8 3. Current actuarial framework Annex XII defines the current actuarial framework to perform the actuarial assessment of the balance of the PSEO. It is divided into five chapters, the terms of which are commented briefly below. 3.1 Content of Annex XII The 1 st chapter states that the Commission will carry out, every five years starting in 2004, the actuarial assessment of the balance of the pension scheme. Moreover the Commission shall every year update the actuarial assessment, having regard to changes in some defined variables. Consequently the theoretical level of the contribution rate of officials is to be checked each year. The 2 nd chapter describes how the actuarial balance of the scheme has to be calculated, that is to say it defines the service cost as the expected increase in rights during the coming year and explains how to compute this service cost, that is the sum of three amounts: the service cost for retirement is the expected value of rights to be accrued during the coming year under certain assumptions; the service cost for invalidity is the expected value of rights due to newly invalid officials during the coming year; the service cost for survival is the expected value of rights due to survivors of a working official. Chapter 2 gives in great detail how the different annuities will be calculated, in respect of both direct annuities and reversionary annuities. The 3 rd chapter of Annex XII prescribes the different assumptions to be used in the computation as following: the demographic assumptions that are divided into two parts: - the mortality table will be updated every five years and has to be constructed by reference to a population similar to the one of the members of the scheme; - the invalidity table, the average age of retirement and the variables linked to the population structure may change annually and are partly based on the demographic data collected by Eurostat concerning the members of the scheme. the economic assumptions that are also divided into two parts: - the discount rate called the interest rate has to be defined by reference to the 12-year average for governments bonds of Member States; 7

9 - the rate of increase of salaries. The Staff Regulations precise that the general salary increase (to be distinguished from the individual salary increase) is defined by reference to the 12-year average of the specific indicator specified in Annex XI. The 4 th chapter states that Eurostat is responsible for all the calculations but may be assisted by independent experts. The 5 th chapter states that Annex XII may be modified after a period of five years in light of budgetary implications. 3.2 Concepts and definition of Annex XII Annex XII states that the contribution rate will be based on the service cost, defined as the expected increase in the liability of the PSEO during the coming year with no regards to the interest cost on the actuarial liability as of the calculation date. This approach is based on the assumption that the past liability is completely financed by the past contributions. It should also be noted that Annex XII prevents any dramatic changes in the contribution rate from one year to the next, stating that the variations in the contribution rate should not exceed a certain level set at 1.00% per year. Furthermore, the maximum contribution rates applicable from July 2004 and 2005 are set at 9.75% and 10.25% respectively. 8

10 4. Eurostat valuation 4.1 Census data Participants taken into account by Eurostat for the valuation at December 31, 2008 break down as shown in the below table. Workforce Avg. Age (year) Avg. seniority (year) Avg. salary and allocation (in ) Actives - contributing members ,7 11, Retired officials - contributing members , Invalids - contributing members , Total ,9 11, Eurostat results This table presents the Eurostat results for the 2009 valuation of the contribution rate. Eurostat in m results Total service cost Sum of basic salaries and invalidity allowances Official contribution rate 11,3% The increase of 0,4% between the 2008 and 2009 contribution rate is mainly explained by : the decrease of the real discount rate (3,1% last year compared to 2,8% this year), the change in methodology regarding the revalorization of the minimum subsistence in the pension amount, according to our audit of final valuation of the liability of the PSEO : + 0,09%. 9

11 5. Examination of the actuarial modelling As mentioned in the first section of this report, Annex XII does not clearly specify every point of the modelling and some points are not stated at all, so that it is possible and even necessary for Eurostat to complete those parts of the modelling that are not dealt with by Annex XII. It should be noted in particular that Annex XII gives explicit formulas only for the computation of the annuities and for the other terms used in the calculation of the service cost of the participants during the year. These service costs correspond to three different scenarios of accrual of rights that are equal to the product of three different terms: the probability that the scenario will occur; this probability is equal to: the probability of becoming invalid for the service cost for invalidity; the probability of dying for the service cost for survival; the probability of remaining in activity for the service cost for pensions; the rights accrued under the scenario, i.e.: a percentage of the present salary for the service cost for invalidity and the service cost for survival; a percentage of the expected final salary for the service cost for pensions based on the assumptions of future salary increases; the annuity coefficient, i.e.: the coefficient for an immediate lifetime annuity for the service cost for survival; the coefficient for a temporary annuity for the service cost for invalidity; the coefficient for a deferred lifetime annuity for the service cost for pensions. The calculation of the contribution coefficient has to be made in accordance with the statements of Annex XII, with no consideration of what would have been preferable. However, Eurostat may decide on some parts of the modelling, since recommendations are not directly available. Concerning the parts that are stated, it may be noted, for example, that Annex XII makes the assumption that the annuities are paid to the beneficiaries yearly in the middle of the year, since they are actually paid monthly; this corresponds to a slight difference with what would have been modelled on the basis of the generally accepted actuarial international practices. However, it should be noted that such a difference is not major and the formulas presented can be considered as conforming on an overall basis to acceptable practices. 10

12 6. Examination of assumptions 6.1 Preliminary considerations As it was the case for modelling, Annex XII does not describe exhaustively the methodology to be used in order to choose the actuarial assumptions but only gives general guidance, so that the Eurostat Team had to make a certain number of methodological choices in order to build a set of convenient actuarial assumptions. Annex XII allows some flexibility in the methodology as well as in the implementation of the construction of the assumptions. However, this flexibility is not the same for every assumption so that it is necessary to define each assumption very carefully at the outset. The following table shows the frequency of changes in the set of assumptions used: Frequency of Assumptions changes (in years) Discount rate 1 Mortality table 5 Invalidity table Not defined Salary increase : GSG ISP 1 Not defined Average retirement age 5 Other assumptions Not defined However, the Eurostat Team prefers to determine an overall methodology for choosing the assumption that will be used over the 5-year period. The application of such a methodology may lead to different values for the various parameters from one year to the next. 11

13 6.2 Main findings Discount rate Statements in Annex XII With respect to the discount rate, that is also called interest rate in Annex XII, it is stated in Article 5 6 that: "A smoothing process shall be applied to determine the real discount rate ( ). The smoothing shall be obtained through a 12-year moving average for the interest rate." Article 10 gives further details concerning this rate: "The interest rates to be taken into consideration for the actuarial calculations shall be based on the observed average annual interest rates on the long-term public debt of Member States as published by the Commission. An appropriate consumer price index shall be used to calculate the corresponding interest rate net of inflation as needed for the actuarial calculations." Choice made by the Eurostat Team No other information is provided by the Staff Regulations concerning the discount rate, particularly concerning the duration of the government bonds to be used, and concerning the reference currency. It is then necessary to define the duration of the government bonds to be used. Eurostat has considered a duration that is equal to the duration of the scheme. This approach complies with usual practice for actuarial valuations. The use of a discount rate based on market yields of bonds with the same duration as that of the scheme is justified by two underlying assumptions used in the actuarial valuation: 1. the first assumption is that the different flows could be hedged by similar flows from government bonds; the basket of such bonds is considered as a notional plan asset ; 2. the second assumption is that the sensitivity to changes in the market rate should be identical for the obligation and for the notional plan asset. The sensitivity of the notional plan asset to slight changes in the market rate is measured with its duration. The only way to hedge the obligation against changes of the first order in the market yield is for the duration of the notional plan asset and the duration of the obligation to be exactly the same. This condition is met by choosing a discount rate based on government bonds with the same duration as the scheme. 12

14 In order to determine the duration of the scheme, it was necessary to put in place an iterative process based on the initial choice of the discount rate and measurement of the sensitivity to slight changes in this discount rate in the actuarial valuation in order to confirm that the initial discount rate was correct or if it should be changed. Duration is approximately 21 years for the scheme. In compliance with Annex XII rules, it was applied an average of the corresponding rates determined by Eurostat over a period of 12 years and not to determine, for the 12 previous years, a duration equal to that of the scheme at that date. We think this approach is consistent with Annex XII. Furthermore, the rate to be chosen must be a real discount rate, and it is therefore necessary to perform the same type of work with regard to inflation. Eurostat has chosen to use the 12-year average of the 1-year inflation rates in the different countries. We consider that it might have been more appropriate to use an average of the expected future inflation rates over a duration equal to that of the scheme instead of the 1-year inflation rates actually observed; however, we note that such data is not easy to reconstruct for various reasons, the most important being the lack of available historical data, since inflation-indexed bonds do not represent a big market, and such bonds have not existed for long with lengthy maturities. The approximation made by Eurostat could thus be considered as correct in general. The 2008 yield government bonds with a maturity of 21 years as calculated by the ECB is 4.8%. The nominal discount rate (NDR) used by Eurostat in the valuation is an average over 12 years: 5.1%. Regarding the 1-year inflation rate (IR), the value retained this year is 2.2%. The real discount rate (RDR) is calculated as 100*(100+NDR)/(100+IR)-100 and is equals to 2.8% for the 2009 assessment of the contribution rate The mortality table Statements in Annex XII Annex XII states the following in Article 9 2: The mortality table shall relate to a population which has characteristics as close as possible to those of the population of members of the scheme. It shall be updated only on the occasion of the five-yearly actuarial assessment referred to in Article 1 of this Annex Issues raised In 2009, the same table used last year was retained. It is the static 2008 ICSLT or 2008 ICSLT Pro Life table. This table is used for healthy people (active and Degagement population). For the invalids, the mortality rate corresponds to a person 3 years older than an invalid person. Therefore, the mortality table used is 2008 ICSLT + 3 years for this type of person. In addition, we would like to point out that progressive life tables are increasingly used in the private sector and in international 13

15 organizations, and that a progressive table should eventually be used in the pension assessment to better reflect the life improvement. Use of such progressive table would lead to an increase of the contribution rate Invalidity table The invalidity table is identical as last year. It is the 2009 EU invalidity table. It is based on historical statistics of news invalids occurred during the 2004 to 2008 period. We draw your attention to the fact that the base of population used to build the table is quite limited (historic of 5 years). We believe it is relevant to apply invalidity statistics that are specific to the EU official s population, even if we recognized that the underlying data for the updated invalidity table are based on limited historic. We recommend a follow up of the invalidity statistics to eventually amend the invalidity table, next year Salary increases Statements in Annex XII Annex XII states in Article 6.2 that: The Projected basic Salary (PS) at retirement shall be calculated starting from the basic salary at 31 December of the previous year and taking into account (i) the rate of annual increase in the salary scales and (ii) the estimated annual rate of increase due to seniority and promotions. Furthermore, in Article 5.6, it states that: A smoothing process shall be applied to determine ( ) the rate of annual change in the salary scales of officials of the Communities. The smoothing shall be obtained through a 12-year moving average for ( ) the increase in the salary scales. In addition, Article 10 states that: 1. The annual change in the salary scales of officials to be taken into consideration for the actuarial calculations shall be based on the specific indicators referred to in Article 1(4) of Annex XI. 2. The effective annual rate to be taken into consideration for the actuarial calculations shall be the average of the net specific indicators for the European Union for the 12 years preceding the current year. 14

16 Issues raised Thus, the statements made in Annex XII concerning the rate of salary increases are the following: 1. the rate is composed of two different components that are respectively the General Salary Growth and the Individual Salary Progression; 2. the first component of the rate must be based on data from the 12 years preceding the calculation date. There was a high degree of freedom for Eurostat in determining the rate of individual salary increase. Therefore Eurostat decided to develop two different methodologies to be used in determining the two components of the salary increases: - the General Salary Growth (GSG) is based on the average of the various overall salary progressions (measured by the Specific Indicator) over the 12 past years, which is equals to 0.3% this year; - the new Individual Salary Progression (ISP) is based on a study (described below) carried out by the Eurostat Team to take into account the new Staff Regulations stipulations about promotions, transitional measures included. Several rates from a table are used in computation so a weighted average of rates is estimated at 1.3% for the 2009 valuation of contribution rate. The study carried out by the Eurostat Team consisted in determining a global annualized rate of salary increase, depending on the grade of the officials concerned and the number of years of activity remaining. It consisted in anticipating future promotions by reference to the theoretical rules mentioned in the new Staff Regulations (including the transitional measures described in Annex XIII). The ISP table used in this 2009 valuation takes into account individual promotions and seniority step progressions observed in years 2005 to Promotion rates are the mean between the theoretical promotion rates and those observed in years 2005 to The average retirement age Statements in Annex XII Annex XII states the following in Article 4 3: It is assumed that all retirements will occur at a fixed average age (r). The average retirement age shall be updated only on the occasion of the fiveyearly actuarial assessment referred to in Article 1 of this Annex and may be different for different groups of staff. Article 9 1 states in addition that: 15

17 From the observation of this population [The population of active staff and pensioners] shall be deduced in particular ( ) the average age of retirement Issues raised The Eurostat Team decided to develop a methodology consisting of several steps: 1. active officials are individually assumed to work until an expected retirement age at which they benefit from the maximum pension, to the extent that this age does not exceed 65 and that it is not below 55; i.e., the age at which they could benefit from the percentage of 70% of their final salary is computed according to new rules regarding the accrual of rights; 2. the officials are then classified according to their actual age, and for the different age classes, the expected retirement age calculated on the basis of the methodology described in point one above is averaged, giving an average retirement age for each class of active official. This methodology respects the general statements of Annex XII that assumes that all retirements will occur at a common average age, which could be different for different groups of staff. The methodology also takes into account the individual characteristics of the participants and is based on rational behaviour, for instance, if an official is older than the expected retirement age at the evaluation date, he retires during the year Other assumptions Probability of being married The probability of being married has been considered 0 or 1 according to the present marital status of the official in the calculation of the service cost for invalidity and in the service cost for death. However, for the service cost for pension, it is defined as the expected probability of being married at retirement age (84% for the men and 38% for the women). We still believe that this approach is acceptable. It has to be noted that some non-marital partnerships were included in the marital status of active officials ( Union Libre and Fin de cohabitation, indexed as UL and CX in the database) Weight of survivors The methodology used remained unchanged compared to the 2008 assessment. The "weight" of survivors is more than 1 due to the fact that the husband or wife is not the only potential survivor but that other survivors such as surviving divorcees, children, or assimilated persons may exist. In order to take into account this factor, the Eurostat Team decided to determine the actuarial value of the rights for survivors other than widows or widowers, and compare this value with the actuarial value of the rights for widows and widowers. The percentage obtained in this way increases the part of the service cost related to survivors. 16