Recovery plans for Dutch pension plans Theo Nijman, Tilburg University January 2009

Dutch second pillar pensions • Standard characteristics:
Mandatory collective products
Nominal guarantee on pension income
Indexation ambition (indexation typically provided if (nominal) funded rate is sufficient) • Prudential supervision on promise to provide nominal guarantee • Continuity analysis checks contract consistency in terms of contribution rules, investments strategy and indexation ambition

2

Prudential supervision Dutch pension funds • Funded ratio (frt): • Market value assets / Market value liabilities • Risk based solvency criteria (FTK): • frt > 105% • Pr (frt+1 < 100%) < 2.5% • Motivation for solvency criteria: • Protect current participants for risk of discontinuity of collective • Protect new entrants for costs of underfunding 3

Impact of credit crisis on funded ratios • During the years 2006-2007 the (nominal) funded ratios of most pension funds were in the range of 130-150%. • Two main effects of credit crisis on funded rates • Drop in equity markets • Reduction in interest rates • Average funded ratio January 2009: 101% • Weighted average might be substantially less • Many funds have funded ratios that are substantially lower 4

Recovery plans • In “normal market conditions” recovery plans are to be submitted within three months once the level of 105% is hit. The maximum recovery period is 3 years. Within that period the expected funded rate should meet the target again. • Likewise the recovery period for “reserve shortage” is 15 years. • The prudential supervisor, DNB, has extended this period: recovery plans are to be submitted before April 1, 2009. • By mid February decisions will be taken whether “extra leeway” will be given (Donner; 22/1; IPE) • Some argue that recovery period for funding shortage is to be extended to 5 years 5

Recovery options • Recovery options • Indexation cuts, • Recovery contributions (“have to contribute”) • Return on mismatch • If no recovery plan can satisfy the criteria: • Reduction of guarantees (negative indexation) • Recovery of reduction and missed indexation in good scenario’s • Reduction can be effective directly or can be through announcement that rights will be reduced in poor subsequent scenarios • Framing of reductions can be in terms of current entitlements, in cash flows or in terms of age at which benefits can be claimed 6

Issues addressed in this presentation • Options for recovery packages • Criteria for recovery plans • Additional criterion on probability of negative buffer: • Natural extension of risk based solvency rules • Values announced cuts in bad scenarios • Encourages more complete contracts to recapitalize the pension funds • Motivates extended recovery period if additional criteria are met • Numerical illustration • Parameter settings • Conclusions

7

Recovery options

Recovery options • Do nothing: hope that financial markets will recover and the interest rates will rise • Use control instruments:
No indexation of entitlements
Obtain additional sponsor contributions
Charge recovery contributions, i.e. pay more than what you get,
Real cost effective but no indexation,
Constant premium but reduced new entitlements, e.g. increased age at which new entitlements can be claimed
The asset allocation could be adjusted NOTE: Current formal criteria for recovery plans favor risk taking … 9

Recovery options II • Note that different options are attractive for different subgroups in the collective • Risky asset mixes are usually attractive for the young • Absence of indexation is costly for the very old, who might never get recovery indexation • Recovery contributions are paid by the active members • Benefit cuts might undermine trust • Young funds are likely to choose other recovery options than older funds. • If the contract can be age dependent some choices might be simpler 10

Recovery package • Standard recovery package: • Retain asset allocation • Skip indexation of entitlements • Small contribution adjustments • For a number of pension funds this package will not satisfy the criterion of an expected funded rate above 105% in three years. • Large contribution adjustments are not attractive given recession and will not really help if the recovery period is short.

11

Recovery package II • Reducing entitlements now is unattractive • Participants are hardly aware that this might once be required • The funding shortage might be solved within a few years • However the message must be communicated now that cutting benefits might be required if the markets do not recover adequately • Not communicating and just allowing a longer recovery period implies shifting the burden to young generations and new participants in not transparent ways who might try to avoid the contract. It is not consistent with fair risk sharing, 12 pension system itself would be at stake

Criteria for recovery plans

DNB requirements recovery plans • DNB Document 2007: No requirements, but suggestions for continuity analysis • DNB Document Dec 2008: Instructions for recovery plans “to enable efficient processing” • (Too much) emphasis on recovery on expected path • Recovery should be “smooth” • Only formal quantitative criterion for DNB is that expected funded rate in 3 or 15 years meets the minimum threshold.

14

Additional criterion recover plans • Proposal: Add criterion that probability that buffer would not be adequate at the horizon is sufficiently small, e.g. Pr (ft+5 < 100%) < 33%. • Compare well known reserve shortage criterion: Pr (ft+1 < 100%) < 2.5%. • The additional criterion emphasizes risks in the recovery plan to balance the current focus on the expected funded rate • The plea for a longer (5 year) recovery period could be better motivated if both this criterion and the one on the expected funded rate are to be met. 15

Impact of the additional criterion • One way to satisfy the additional criterion is to announce which instruments will be used in poor future scenarios e.g. the sponsor will step in or guarantees will be reduced • Plans with recovery packages with a too large probability that the buffer will be too small are unattractive for the active participants. • By assigning both the up-side and the down-side of scenario’s which deviate from the expected path behind the “veil of ignorance”, i.e. ex ante, conflicts of interest can be avoided and the fund remains more attractive for all 16

A numerical example

Assumptions numerical results • Initial funding rate 90% nominal • Model of stock returns, nominal and real term structure and inflation as in Koijen, Nijman and Werker (2007), i.e. US data • Parameters adjusted manually to reflect current Dutch settings.

• Note: Software (Tilburg Finance Tool) readily available

18

Proposed initial recovery plan • Asset allocation: • 50% stocks; • 50% in 5-yr zero coupon bonds • Required funding rate: 125%. • Indexation policy: • Standard conditional indexation contract • Thresholds 110 / 140 • Contribution policy: • Recovery contributions towards level of 125% in 15 yrs • Total contribution level 20-25% (approx. cost effective real). Current contribution levels in sector hardly above nominal cost effective level, as interest rates are low.19

Funded Rate

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024

145,00 135,00 125,00 115,00 105,00 95,00 85,00 75,00 65,00

Expectation

33% Quantile

10% Quantile

Numerical results funded rate in example • Expected funded rate after 3 years: 102% • Expected funded rate after 5 years: 110% • Only looking at expected returns implies that strategies with more exposure to priced risks would always be preferred • Probability funded rate after 5 years less than 100% exceeds 33% • Proposed criterion would require that an additional (gradual) recovery strategy is announced now in case of bad scenarios in next 5 years. • Expected return after 15 years: 135% (OK) • Probability funded rate after 15 years less than 120% is 33% (less than 3 yrs behind schedule) 21

Purchasing Power L oss

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024

50 40 30 20 10 0

E xpectation

33% Quantile

10% Quantile 22

Other criteria for recovery plans • Supervisors and social partners alike will not just look at the funded ratio but also at the other key variables • contribution level • purchasing power loss and to what extent these are consistent with previous communication • Decision makers might also want to see the market value of the contribution of subgroups to the recovery plan • Heterogeneity in funds: Adequate length of recovery period dependent on discontinuity risk for collective and strength of sponsor 23

Parameter settings

Parameters • In valuation od assets and liabilities much of the required information is observable in the market: asset prices, term structure Issues: illiquid trading, forced trading (swap curve) • In risk based solvency regulation, recovery plans and continuity analysis expected returns and second moments are required: more scope for differences of opinion • For now discussion is on unconditional parameters • Some input requirements are observable in the market (if liquid and not forced..) • Nominal term structure • Term structure of inflation expectations 25

Nominal term structure • Current curve contains the rates at which future term structure can be fixed (subject to liquidity etc.) R(t,T) = R(0,t+T) + t/T [R(0,t+T) – R(0,t)] • Requirements for recovery plans for reserve shortage allow use t+4 curve. Motivation: illiquidity and volatility long end (substantial impact !) • In case of funding shortage DNB rules impose that increases in forward curve can not play role in expected recovery (M4 = 0). Adequate if liabilities “fade out”. Consistency with definition of a cost effective premiums to be considered further. 26

Implied nominal term s truc tures (t=0 is 2009)

5,0% 4,5%

4,0% 3,5%

3,0% 2,5%

R (0,T) R (3,T)

2,0%

R (4,T) R (5,T)

1,5%

R (15,T)

1,0%

0,5% 0,0% 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Ma turity in ye a rs

27

Inflation expectations • Quite important if no indexation would be given or several years • Differences between real and nominal term structure caused by differences in expected risk premiums and by expected inflation • Inflation risk premiums likely to be small (Rudebusch et al (2008)). • Under this assumption “break even inflation” indicator of inflation expectation R(0,T) – r(0,T)= E0 Σt [R(t,1) – r(t,1)] = E0 Σt π(t,1) • Data for US as for UK, Data for Euro inflation not directly available. Expected 3,5% wage inflation might be high for short run and not even enough for very 28 long run.

Inflation expectations Implied forward inflation curve UK; december 22, 2008 6,00

5,00

%

4,00

3,00

2,00

1,00

0,00 0

5

10

15

20

25

years Series1

29

Conclusions

Conclusions • Use more market data to set parameters in continuity analysis and recovery plans • In analogy with risk based solvency regulation, the continuity analyses and recovery plans should not just focus on the expected path: additional criterion for recovery plans is natural next step in FTK • Where appropriate, communication that in further bad scenario’s some reduction of guarantees can not be avoided should start to avoid discontinuity of system • The additional criterion on the probability of sufficient recovery supports the continuity of the fund and enables participants to prepare for the future • Such an additional criterion could motivate an extended recovery period 31