Summary of Life Cycle Investing PDF

The Economics & Fin.of Pensions Summary Roel Mehlkopf: Life cycle investing (11/10 & 23/10) Thought experiment - If human capital (future income) is riskfree  optimal for young workers to invest all their financial assets into risky assets. Amount that young want to invest in risky a...

The Economics & Fin.of Pensions Summary Roel Mehlkopf: Life cycle investing (11/10 & 23/10) Thought experiment - If human capital (future income) is riskfree  optimal for young workers to invest all their financial assets into risky assets. Amount that young want to invest in risky assets exceeds their financial assets (borrow money and invest this in risky assets) - If labor income is risky  less willing to take risk - Other financial assets (f.e. a house) can have important impact on optimal lifecycle investment strategy Lifecycles in practice: United States - When people become older  less human capital and smaller investment horizon so will invest less in stocks - Duration of bonds align with ages - Target Date Funds: type of mutual fund/ETF designed to simplify retirement investing by automatically adjusting asset allocation based on a target retirement date. Observations in US: o High percentage of wealth invested in stocks for young people, this will be reduced later in life o Most show no lifecycle pattern for the duration of bonds - Possible explanations observation on duration of bonds: o US pension market doesn’t have a mandatory conversion of pension wealth into lifelong income stream o Us pension market is business-to-consumer (hard to explain long-term bonds to consumers: benefit levels more stable but pension pots more volatile (interest rate and inflation risk)) Lifecycles in practice: the Netherlands - Less percentage invested in risky assets when people become older - Pension funds GP (huisartsen): large amount of certainty from human capital (certain to keep job as GP,stable income) and diversified investment portfolio  reducing investment risk just 10 year before pension age - 1-year higher average age  reduction of the strategic equity exposure by around 0.5 percentage point - Average age active participants influences investment behaviour more strongly than average age of all participants (retirees no longer possess human capital) - When retirement age coming closer  higher interest rate hedging Life cycle investment theory - Brownian motion to model return risky assets (returns are logN distributed) - Financial market: o Continuous time and continuous trading o One aggregate risk factor and 2 assets:  Risk-free asset: (doesn’t exist in real word, government bond closest)  Risky asset: Zt follows BM; stock return is driven by iid random shocks o , drift (higher  higher avg return) higher then interest rate reflecting positive risk premium - Benchmark model: o Labor market: future income is deterministic (fully predictable) & Labor supply is exogenous o Exogeneous retirement age T and exogenous lifespan D o CRRA preferences: no habits, no behavioural features (loss-aversion), no bequest motive (you only care about your own consumption level), subjective discount factor (rate of time preference), constant relative risk aversion (between 1 and 5)  o  Default parameters - Share of financial wealth invested in risk asset with riskfree human capital: - Optimal portfolio mix in terms of total wealth (human capital + financial capital) = everyone 20% in risky asset, rest in risk-free - Optimal portfolio mix in terms of financial wealth: age 20-55 invest more in stocks Effects decline interest rates on optimal saving for an individual with a utility function that features CRRA - Substitution effect: lower interest rate decrease optimal slope of consumption path: people will save less as the reward for delaying consumption is lower. Substitution effect reduces savings in a low interest rate environment: reward for saving is low and that you better consume more today, at the cost of a lower pension in the future - Income effect: a lower interest rate makes future consumption more expensive, and hence a larger amount of savings is required to realize a certain level of consumption at old age; this can be regarded as a decrease in the budget available for consumption over the life-cycle. The income effect increase the savings rate in a low interest rate environment: if you want to keep up consumption after retirement you have to put extra money in your pension and consume less today - Wealth effect: the income effect may be mitigated if the lower interest rates lead to revaluation of the households’ assets (e.g. bonds). This provided a ‘hedge’ against the fall in interest rate. Also, future labour income can be seen as a hedge as it becomes more worth in present value terms. The wealth effect (partially) mitigates the income effect. Roel Mehlkopf: Life cycle investing continued (2/12) Previous lecture: risk aversion, risk premium risky asset, volatility risky asset, future income and social security constant. Standard model is quite simple Model extensions Extension: Interest rate risk - Add interest rate risk and choice between short- and long-term bonds (instead of risk-free asset in standard model) o People in 40 and 50 years invest in long-term bond (besides risk asset) o From age 60 invest also in short-term bonds - If risk aversion is doubled and risk premium on stocks twice as small (risk aversion = 10, risk premium= 2%)  invest already early in long-term bonds - The introduction of interest rate risk introduces a (hedging) demand for long term bonds, as interest rate risk introduces reinvestment risk for a long-term investor - The demand for long-bonds may be small for a young investor, as this demand ‘competes’ (horse race) against stocks as young investors also try to get a large exposure to stock market risk - The demand for long-term bonds also reduces at higher ages, where the remaining investment horizon becomes shorter so that long term bonds are gradually replaced with short-term bonds as the investor becomes older Extension: Inflation risk - Real interest rate = nominal interest rate – inflation rate - If inflation goes up  investors demand higher return  interest rate goes up - If you buy nominal annuity (or long-term nominal bond that matches the duration of your future consumption) you fix your consumption level in euros so with inflation rate increase and interest rate increase your consumption level will go down - Inflation linked annuity if hard to sell to consumers because people only look at the short term but in the long term the value will go up increasingly - If the inflation risk is added to the model and only nominal bonds are available in the market  less interest in long-term bond because they are vulnerable to inflation so people will invest in risky assets or short-term bonds - If a rise in inflation expectations leads to an increase in nominal interest rates (with real interest rates unchanged) then the price of long-term nominal bonds falls, while the price of future (real) consumption increases - Hence, investments in long-term bonds can make an investor vulnerable to inflation risk. Long-term nominal bonds exposure an investor to changes in expected inflation while liabilities (i.e. consumption needs) are likely to be linked to inflation - Hence, holding a large portfolio of nominal bonds can be seen as unattractive by investors with inflation- linked objectives. - Inflation-linked bonds may be more attractive for an investor, but these may not be available in the market (due to market incompleteness) or only in limited amounts. Extension: Labour income risk / employment risk - Presence of labor income can lead to a reduced capacity to take risk - Labor income risk linked to macro-economic developments (e.g. employment risk) introduces a correlation between human capital and stock market risk, and reduces the optimal allocation towards stocks - The type of occupation may be relevant o Public administration (for the government) has stable income (relatively safe) o Construction and agriculture can fluctuate more - A long-term interrelation (cointegration) between stock markets and labor markets further reduces demand for stocks of young Extension: Endogenous labour supply - Endogenous labor supply can take the form of: flexibility in the number of hours worked per week or flexibility in the choice of the retirement date (required flexible labor market for the elderly, allowing them to work longer if they want to, also in an economic recession) - Flexible labor supply leads to a higher allocation to stocks when working, because labor supply can be used as a buffer against disappointing stock returns: labor supply can be increased in response to negative financial stocks Extension: Non-standard preferences - Extension: preferences - o Leads to lower allocation toward stocks o Investor need to hedge the risk-free liability in preferences o At the same time, the future income from the state pension may already partially or fully meet minimum consumption needs o Social security and minimum consumption cancel out against each other - Extension: loss aversion and reference point o Relative risk aversion may not be constant but depend on the actual consumption level and the particular losses and gains in the past. o Myopic loss aversion: relative risk aversion in large close reference point o Risk aversion depends on if you are close to the reference point o In NL: reference point is current pension level - Extension: habit formation in preferences o Internal habits: less willingness to fall below the consumption level that was anticipated or experienced earlier in the lifecycle o External habits:  Consumption requirements depend on aggregate consumption of the economy as a whole or another group or other workers  Elderly compare their consumption level relative to the developments in the consumption level of workers, so that developments in wage levels are more relevant than price levels  Elderly may invest more in equity if equity is correlated with wages Extension: mean-reversion in stock returns - The young have a longer investment horizon than the old - If belief is that stock returns are mean-reverting: extra leverage that people will invest in risky asset Conclusion - There is no such thing as the optimal lifecycle: depends on preferences, investment beliefs, etc which are subjective or difficult to estimate - Investing in stock becomes more attractive for young investors o If human capital is safe o If one believes in mean-reversion in stock returns o If labor supply is flexible (retirement date, working hours) o If you own a house that you consider to be bond-like - Investing in stocks becomes less attractive for young investors o If there is a risk that you lose your job o If human capital moves together with stock markets in the long run o If you own a house that you consider to be stock-like o If you have a minimum subsistent level for consumption Casper van Ewijk : Ageing, retirement age and budgetary consequences (28/10) Old age dependency (number of individuals ages 65 or older per 100 working age(65+/(24-64)) rising until 2040-2050 Population ageing and why it’s a problem - Explanations rise in old-age dependency: o Baby boom (they aged 75 in 2022), noticeable but small effect (dependency should decrease if baby boom was the cause) o Increasing life expectancy (also convergence between men and women in the future)  fix by increasing retirement age in line with life expectancy o Main cause: less children per family (fall in fertility; anticonception and emancipation in 1960/1970s)  In NL: fertility rate (children per woman) is lower than 2  Not easily repaired, especially for countries that rely on payg more than on funding - Ageing problem in PAYG pensions: o Burder for the young may more than double (because of longer life expectancy and fall in fertility rate). PAYG pensions is an intergenerational contract - Health care and long-term care have become the main threat to financial sustainability! Generation accounts (GA) - Generation accounts measures net benefit (net contribution) of a generation from (to) a PAYG arrangement, taken over the entire remaining lifecycle starting from current age. Important measure of redistributive impact of government finances - Net benefit = PV of pensions received minus contributions paid - Net (PV) contribution of generation born in k to the government over their remaining life cycle o T=net payment=taxes paid minus transfer received, D=max length of life, P=number of surviving members, r=real discount rate - o Contribution current generations + contributions future generations = real government (G) expenditures in PV terms + initial debt  burden of government finance is divided over current and future generations o In pure PAYG Gs and Wt zero  Generations accounts are zero sum - Unstustainable if future generations pay more on average than current generations - Results from Auerbach in US (makes projection of total G with taking into account aging  then projection of the contribution of current generations and what is left to be paid by the future generations) shows that the government finance is unsustainable (difference of 111%) - Kotlikoff redid the analysis in US and found that current generations actually receive more from the government than they contribute as tax to the government (negative GA for current generations)  US public finances unsustainable from an intergenerational POV - Explanation negative GA in US: increasing social security and care expenditures while not compensating by higher taxes (around 80% public debt as percentage of the GDP) - CPB has expanded the GA framework to include benefits of government expenditures o F.e. education (for young) and health care (for old) is easy to determine which groups are benefitting  o Costs of roads and public administration is difficult to assign to groups so its evenly distributed over all age group o Age profile of net benefits (benefits-taxes) from the public sector Sustainability of government finances in the Netherlands (and the EU) - Sustainability: offering future generations same arrangements as current generations taking into account of population ageing - Sustainability gap: immediate and permanent adjustment in the government budget that is needed to make arrangements sustainable from a long term perspective (by how much should we improve the primary surplus (T-G % GDP) to make government finances sustainable). Determine by: o Make projections for government revenues and outlays using the age profiles of net benefits, and projections for demographic change (ageing) assuming constant arrangements (constant average tax rates and constant social security arrangements and constant government provisions as % of income) o Find what adjustments in government consumption is needed to make government finances sustainable (satisfying intertemporal budget constraint) (find policy reform necessary to avoid exploding public debt in the distant future) - Impact of immediate and permanent increase in (T-G)/Y on the evolution of the fiscal balance (T-G-iD) and public debt (% GDP) - High sustainability gap in NL in 2022 - Pension reform 2019: second pillar pension from DB to DC, more personal pension - Sustainability gap measures the adjustment needed to make government finances stable from a long term perspective o Fiscal sustainability requires the government to build up reserves to prepare for the increasing cost of ageing in the future - Public debt should be considered from a long-term perspective (only change in debt matters) - Sustainability has been the leading parameter for fiscal policy in NL since 2000 and one of the indicators in the EU - Despite reforms still a fiscal sustainability gap in almost all EU countries - Public pensions and health care cost have been the most important drivers in the rising cost of ageing in the past - For the future, the rising cost of public pensions are generally contained by raising retirement age and cuts in pension benefits, but expenditures on health care and long term care keep rising and have become the main threat to fiscal sustainability. Application of generation accounts: pension reform: intergenerational aspects - Foreign students who stay contribute 1.5 euro billion to the government - Options for pension reform of public pensions: o Cutting indexation of pensions (indexation to prices only, thus nu indexation to real wage growth)  lower growth of pensions   Affects all generations, the loss is small for the very old cohort and becomes larger for later cohorts. Impact is largest for cohorts just to retire by 2011. For later generations the impact is again smaller as the pension period is only part of their future lifecycle o Raising retirement age (constant thereafter or raising it further) (no effect on cohorts already retired in 2011, strongest impact for generations just to retire when retirement age is raised) o (2011)(impact is measured as the average loss per year over the remaining lifecycle) - Generation accounts = net benefit over the remaining lifecycle (useful to compare generations, important tool for policy making and assessing intergenerational impact of policy reforms) - Transition to the new pension contract in NL will lead to negative net benefit for people born between 1976- - - Intergenerational effects: o Double transition: no implicit subsidy (from younger workers to older workers) + lower buffer requirements o Abolishing implicit subsidy: current young worker pay too much and older worker too little due to uniform pricing o Smaller reserves in new contract: current DB requires reserved to warrant the nominal guarantee in pensions, new features DC pensions without guarantee Casper van Ewijk: Intergenerational risk sharing and payg vs funded systems (30/10) Introduction PAYG - Non-contributory: pension unrelated to the individual’s contribution to the pension system; aimed to poverty alleviation (AOW) - Contributory: pension related to past earning and contributions; aimed at consumption smoothing; can be defined benefit (DB) or defined contribution (DC). DB is costly for firms and pension funds - Continental (Bismarckian) o Government provides pensions to all citizens, means-tested to individual financial situation (point system; pension points linked to their individual contributions or Notional DC system) o Large EU countries are Bismarckian (PAYG) - Anglo-Saxan (Beveridgean) o Government provides basic pension (aiming at poverty alleviation) o Supplementary pensions aiming at consumption smoothing (keep up their consumption after retirement) are left to private institutions - PAYG: less sensitive to financial market risks, redistribution intra- and intergenerational, democratic decision making, solves market failure; pensions linked to wage growth, intergenerational risk sharing (improves welfare by smoothing shocks over more generations) - Funding: less sensitive to political risks (even better when spreading investment internationally), contributes to savings, more transparency, investments of pension funds contribute to deeper capital markets and thus enhance economic growth, allows pensioners to benefit from higher return on financial markets - Pension funds deepen capital markets: o Some Pension funds investing in illiquid domestic markets or aiming at climate and other ESG goals o National savings through pension funds  financial stability (private investment less dependent on domestic banks) o Institutional investors (pension funds and insurance companies) promoting economic growth - Funded system may include payg elements: under-funding of pensions implies transfer from younger future generations to current generations - Payg system may include a funding elements: include capital reserve to save for rising pension burden in the future and to accommodate temporary shock PAYG vs funding: intergenerational distribution (no risk) - Recap: intertemporal budget constraint - PAYG implies redistribution of income by taxing the young and paying pensions to the old of the previous generation - Simple 2x2 OLG (overlapping generations) model: start payg pension in t, all consumption after retirement w=wage, P=pension, T=tax rate, c=consumption, r=interest rate, n=population growth  - GA for PAYG pension - Intergenerational effects of introducing PAYG: Windfall gain for first generation: pension without having paid pension to previous generations. Burden of the transfer to the first generation is shared among all successive generation - Consider with constant contribution rate (Tt=T) and taking wage growth and population growth together 1+g = (1+n)(1+λ) we get g this negative GS can be regarded as implicit tax necessary to finance the windfall gain (normally we assume r>g) - In a world without risk the choice between funding and PAYG affects the intergenerational distribution (zero sum) PAYG vs funded in the presence of risk - Insurance: pooling of idiosyncratic risks (small relative to aggregate wealth or consumption and uncorrelated to systematic (macro) risk): distributing small shocks over many people  cost tend to 0 - Risk sharing: distribution of aggregate risks over individuals. In optimum: minimize aggregate welfare loss - If number of individuals tend to infinity: perfect pooling of independent shocks = insurance - Risk should be shared because of concave utility (diminishing marginal utility)  smoothing consumption over time and smoothing over different states of the world (risk) - Risk aversion: marginal welfare of cost of a downward shock increases with the size of the shock - Risk sharing via private markets o Smoothing of macro risks over individuals takes place through financial markets by trading assets (distribute shock over many individuals) o Financial risks can be spread over many individuals if individual portfolios are sufficiently diversified o Can provide international risk sharing o Can be subject market failures - Role for policy intervention when markets fail o Missing markets: no well-developed markets for f.e. wage or inflation liked assets and no market for trading risks between current and future generations o Constrained individuals: not all individuals have access to financial markets o Non-optimizing behaviour: not all individuals make the right decisions due to lack of knowledge Welfare gain of intergenerational risk sharing - No market solution for risk sharing with future generations (shocks revealed before young are born): o Return on saving old generation revealed before young generation able to trade  cannot share in financial risk of elderly o Young generation cant invest in financial assets before birth  misses opportunity for diversification -  market failures that allows for welfare improving interventions  government intervention: o F.e. tax on financial returns for elderly and distribute revenues to next generation - Sharing financial risk o Government taxes half of shock in each period t  pareto improvement (beneficial for all) o Introducing PAYG: first generation had windfall gain, next generation is better off if welfare gain of risk sharing outweighs the implicit tax  Pareto improvement in welfare by risk sharing in wages - PAYG pensions may produce Pareto improvement in welfare in presence of financial risk and wage risk  - We assumed zero growth and interest rates but also with r>g pareto improvement is possible as long as welfare gain of risk sharing outweighs the implicit tax due to first generation effect. More likely in presence of large uninsured risks and high risk aversion. Case: Intergenerational risk sharing (IGR) in 2nd pillar pensions in NL - Before: shocks in 2nd pillar pensions were smoothed over a longer time horizon  pension funds contributed to intergenerational risk sharing  this solidarity between generations was seen as one of the big assets  but due to ageing population this risk sharing has become increasingly more difficult - Fall in stock prices in 2010 was distributed over generations: largest for birthyear 1950 and only limited risk sharing with future generations  optimal to smooth over all current and future generations  but this implies that the fund may run deficits and surpluses over a very long period - Discontinuity (deficits and surpluses) of intergenerational contract may result in less inflow of new participants  solve by making participation in the pension fund mandatory (in NL: mandatory per sector or firm) - Estimates of the welfare gains vary a lot - Maturing of pension funds and population ageing make it more difficult to share shocks with future generations. Smoothing periods were shortened to contain the risk of discontinuity (risk that younger generations don’t longer want to participate in the fund if the burden in the system is too large) - IGR has become difficult/ is limited in the Dutch 2nd pillar: growing pension wealth relative to contribution base - Government is better equipped for organising IGT as the discontinuity risk is virtually absent thanks to the solid contribution base of taxpayers  better to organize risk sharing with future generation via the government for smoothing shocks. PAYGO pension in 1st pillar may contribute to IGR. With tax payers provide a solid contribution base for the government. - Arguments for funding: o Better prepared for rising cost of pensions due to ageing population o Better diversification of risks by investing in worldwide portfolios o Allows participants to benefit from the risk premium of assets o Contributes to deeper capital markets (by providing risky bearing capital to the private sector)  economic growth o Pension investors play useful role in governance of companies (long-term focus) - Arguments for PAYG o Provides wage linked pensions and thus contributes to solidarity o Can smooth risks that are not traded in the market (aggregate longevity risk) - PAYG can be welfare-improving in the presence of uncertainty and capital market failures. In the absences of uncertainty the choice between PAYG and funding is purely a matter of intergenerational distribution (zero sum) Casper van Ewijk: Pensions in low interest rate environments and pension savings (4/11) - (nominal and real) Interest rates on the decline since 1980 - Recent rise in interest rates due to surge (piek) in inflation - Replacement rate at retirement higher when real interest rate higher - DB pension more expensive at low interest rate (fall in interest rate raised cost of pension liabilities  lower funding ratio (asset/liabilities) of pension funds, effect more heavily with longer duration) Underlying fundamentals: CAPM model: (formulas on slides) - Interest can be interpreted as the rate than ensures that the growth rate is equal to the aggregate growth rate of GDP (if growth rate is dependent on other variables) - Drop In growth by 1% can explain a drop in interest rate by 2 to 5% - Both (nominal) interest rates and growth rates are falling since 1980; interest close to zero bound - Economic growth has declined with about 1%-point between 1990 and 2021. This can explain a fall in the interest rate by 2 to 5 %-points! Actual decline in the interest rate was about 2%-points in this same period. - Also increased uncertainty and higher risk aversion may have contributed to lower interest rates - Population growth slowing down due to lower fertility - In EU declining work force - Technological progress has slowed down - Equity returns o We assume returns on equity co-move with state of economy o However: if decline in risk free interest rate is due to increased uncertainty of risk aversion, the fall in equity returns may be mitigated by an increase in the equity premium. This is important for pensions as it make them less sensitive to the interest rate o Estimates for the equity premium, show increases by some 4 %-points since 2000 - Return on stocks was not reduced in a one for one manner with interest rate: it was assumed that the fall in interest rate was mitigated (beperkt) by a rise in equity premium Low interest rates: Are they to stay? (based on Rachel and Summers 2019) - Studies on natural interest rate (equilibrium real interest rate in absence of cyclical factors) also point to persistent low interest rates - Over a longer time horizon (1870-2015): more periods with low (and negative) real rates - Government: increased public debt (more bonds), higher government spending, better pensions, and improved social insurance all contribute to raising the equilibrium real interest rate (r*) by either increasing demand for funds or decreasing savings. - Since 1970: government policy pushed up r* by larger debt, higher health care expenditures and improved social security causing hh savings to fell. Large standard errors - Private sector: Demographic factors (lower population growth, longevity), lower productivity growth, and rising income inequality all lead to: Lower investment (I ↓) or higher savings (S ↑). These effects cause the equilibrium real interest rate (r*) to decline. - Conclusion: fall in interest rate mainly driven by private sector: low investment (due to low growth) and high savings private sector (due to uncertainty) - Policy implications: monetary policies less effective because of 0 lower bound, mature economies may be prone to secular stagnation (slow recovery, no bouncing back) (or even stagflation: high inflation and low growth (Kaas Knot)) - Uncertainty about future: new events may cause higher real interest rates (trump) - Conclusion: o Low interest rates likely to persist in the future o Also equity returns expected to be lower, but maybe less so than interest rates (because of larger equity premium) o Low interest rate is mainly a demand side problem: too high savings and depressed demand in the private sector  o Uncertain future: some point to new upward pressures on interest rates. Inflation may persist together with low growth (stagflation) Low interest rates and funded pensions - Low interest rates make pensions more costly (fall in interest rates 3% to 0% would require the contribution rate to more than double for a young person, alternatives are lower replacement rate or higher retirement age) - Effects lower interest rate o Substitution effect: Decrease the optimal slope of the consumption path: people will save less o Income effect: a lower interest rate makes future consumption more expensive; this can be regarded as a decrease in the budget available for consumption over the life-cycle. o Wealth effect: the income effect may be mitigated if the lower interest rates lead to revaluation of the households’ assets, e.g. bonds. This provides a ‘hedge’ against the fall in interest rate. Also, future labor income can be seen as a hedge as it becomes more worth in present value terms. - Taking the effects together, lower interest rate leads to: o Lower optimal pension: The income effect and substitution effect together always imply consumption should fall after retirement. o Ambiguous effect on contribution rate: Here the substitution effect and income effect point in different directions. It depends on the size of these two effects whether pensions savings ( = contribution rates) should go up or down.  Idea: If you want to keep up consumption after retirement you have put extra money in your pension and consume less today.  But you may also decide that the reward for saving is too low and that you better consume more today, at the cost of an extra low pension in the future - Impact of 1% lower r on optimal consumption path: slope of C decrease (substitution effect), level of C decreases on average (income effect) - Only at very young ages the substitution effect dominates the income effect. So ideally, the contribution rates should be increased only in later stages of your career Implications for the balance between funding and PAYG - Popular story: o private saving gives a return of r o a PAYG system with constant a contribution rate (T) gives a ‘return’ of g, where g is the growth of the tax base (wage growth + population growth) - So the popular (but wrong) story is that r versus g determines which system funding/PAYG is to be preferred - As long as r > g the younger generation pays an implicit tax, needed to finance the windfall gain for the first generation that receives pensions without having contributed themselves - Choosing between PAYG and funding is about redistribution between generations, not about efficiency (‘better’). - As long as r > g intergenerational distribution is zero sum: that is, what is given to one generation is taken from other generations - Note that r>g is required to have a finite budget constraint - One should measure r and g properly, that is taking account of the risk premium on investment on capital markets - In world with risk one should not compare the risk free interest rate with a (uncertain) rate of growth. - In stead, one should compare the expected growth rate, with the return on capital including a risk premium to account for the risk in the economy Non-financial DC (NDC) - 5 pillars: 0th pillar is basic provision for poverty alleviation, 1st pillar (government) is mandatory contributory pension for all working people (PAYG), so including self employed, for consumption smoothing -  - DC (Defined Contribution) like in the 2nd pillar but now on PAYG basis organised by the government in the 1st pillar. - Pension linked to contributions in the past, each individual has an individual account recording contributions paid - Every year accounts are increased by a virtual return (wage growth or inflation). - After retirement this account is decumulated paying a pension until death (annuity). Retirement age is flexible within a certain window (e.g. between age 60 and 70). The pension age is regularly adjusted to changing life expectancy. - There is a balancing mechanism to ensure sustainability over long time horizon. - NDC can be introduced without initial transfers to the current generations, if current working generations start building up accounts in the NDC system. - But, it requires that the government should not spend the contributions to NDC for current expenditures; this may require considerable fiscal discipline - But funded pensions are important for diversification of risks, deepening capital markets and thus fostering economic growth! Key results - Low interest rates require to rethink the balance between pension ambition, contribution rates, and retirement age: invest in human capital rather than financial capital - the Aaron condition (r-g) is NOT a sound criterion for choosing between PAYG and funding; as long as r > g then shifting between PAYG and funding is just a redistribution between generations (zero sum, under certainty) - PAYG should be motivated from the solution to market failures: o PAYG may enable wage linked pensions not available in the market o PAYG may contribute to risk sharing between generations, for example allowing younger generations to share their wage risk with elderly of previous generations - Other aspects to be taken into account: o trade-off between political vs financial market risk o funding provides better commitment to saving (needed for ageing) o funding makes capital markets deeper and stimulate growth Bas Werker: Longevity risk (6/11) Longevity (without risk) and pension premiums - Work 40 year and live 20 years after retirement, pension premium that lead to a 100% replacement rate is 20/60=33% - Less pension premium may suffice through; the replacement rate is lower than 100% or there is positive real interest rates or risk premiums or lower taxes for retirees (less gross income needed to achieve same net level) or first pillar pension or other wealth (housing or other savings) - Actual pension premiums: Dutch DB: 20-30%, Dutch DC: 10-20%, worldwide almost 0% (robust state pension or less developed private pension system) Longevity (without risk) and pension age - Reduce pension deficits through increase of the retirement age - V = 2/3*(L-20.64)-(P-67) (V is verhoging pensioenleeftijd, L is resterende levensverwachting op 65jarige leeftijd in jaar v verhoging, P pensioenleeftijd voorgaande jaar, V Interest rate risky: Payments start higher to match the AIR assumption but decrease over time as the shortfall depletes the principal faster than expected. Smoothing (“spreiding”) - Lose 10% wealth o No smoothing: reduce each bucket by 10% (optimal when Merton model with CRRA preferences) (proportional utility derived from consumption the same, even after the wealth loss)(people don’t like this) o Smoothing over 5 year: reduce each subsequent bucket by more than 10/5=2% (optimal with habit formation preferences) Reductions increase over time. Risk can average out over time - Smoothing: o your income decreases more slowly, but ultimately more o doesn’t lead to stability (reduce short-horizon volatility but increase longer-horizon volatility) Investing (100-age)% of wealth in equities vs Merton model - Both imply life-cycle investing: more risk when young - Merton model links the optimal asset allocation to the risk aversion of the individual. - Merton model can yield asset allocation above 100% when young. - In the Merton model the optimal asset allocation no longer depends on age once retired. - In the Merton model the optimal asset allocation depends on financial wealth (on past returns, not just on age) Longevity risk in variable annuities - pt(h) probability that an individual doesn’t survive till time h - Use AIR for sharing micro longevity risk  mortality credit Bas Werker: Guarantees and loss aversion (13/11) Solvency Based pension schemes - Defined pension schemes are typically regulated based on a notion of solvency (having a buffer) - Insurance and reinsurance companies have the solvency capital requirement - Pensions funds have the funding ratio = value of assets/value of liabilities (for solvency) - Value of asset: market value available assets (objectively observable with possible exception of illiquid assets as private equity) - Value of liabilities: calculated by discounting the future payments (entitlements multiplies by survival probabilities (then the entitlements will lead to actual pension payment)) - Discount rate is debated (lower discount rate leads to higher present value of liabilities so weaker solvency) o NL: yield on default-free bonds, other countries: expected return on assets o DB pension payments are risk  include risk premium o Bonds are sensitive to discount rate. The more sensitive (or duration-heavy) the portfolio is to discount rate changes, the larger the fluctuation in total asset value. o Higher allocation to stocks  higher expected return  lower present value of liabilities (if discount rate is based on expected returns assets) Risk-sharing-based pension schemes - Pension fund: collective where individuals have entitlements. Shocks (financial and longevity) are allocated to those entitlements (entitlements can be promised benefit (DB) or individual wealth (DC)) - All risks are allocated to the participants with pension funds but with insurance companies, the share holders are the external risk bearers - In mutual the financial market risk is allocated one-for-one to individual wealth, relation between wealth and income depends on interest rate in a pension fund (Interest rates determine how much income can be derived from wealth (pension fund assets)) - Dutch pension contract had first a buffer but no in the new one Nominal guarantee in the decumulation phase - Not possible to give active fund participants a nominally guaranteed pension (no default-free bonds exist with the appropriate maturity and its sensitive to inflation risk) - Retirees can get a nominally guaranteed pension (guaranteed relative to standard financial and longevity risk, active participants needed to bear the residual risks that cant be hedged on financial market) - Nominally guaranteed pension: starting income uncertain but afterwards remains fixed through: o Protection return to offset changes in AIR (nominal income remain stable) (achieved by hedging pension fund payments by nominal default-free bonds) o Zero exposure to excess return fund (allocated to active participants if its appropriate for them (works only with foreseeable risks)) o Protection return to offset longevity risk - Dutch pension sector provide indexation: take risks such that pension can increase by 2% a year (or it can lower pensions) - Options to pay for indexation: o Give retirees risk on a pension decrease o Make retirees pays for future indexation by giving them a lower starting pension  Use f.e. 90% of pension wealth to crease guaranteed pension invest remainder in risky (when remainder goes up -> increase guarantee otherwise decrease risk exposure) o Pay-as-you-go where active participants pay for pension increases of retirees  With use of solidarity reserve in NL, but only 5% of wealth is allocated in pension funds by rules of solidarity reserve (rest is allocated using protection and excess returns) - In current NL: pay indexation by giving retirees the risk of extreme pension reductions (tail risks) and PAYG Theo Nijman: Basic features of pension systems (18/11 & 20/11) Role of government - Objectives public pensions systems: alleviating old age poverty and intertemporal consumption smoothing - Intertemporal consumption smoothing through government in Bismarck (close link value of contributions and pension rights public pension on individual level, high contribution rates, Germany) but not in Beveridge (weak link, AOW, low contribution rates, more private provision of pensions, NL) - Pillar 1 in NL: AOW. PAYG, redistribution between and within generations, prevent old-age poverty, income- specific taxes and transfers. AOW age linked to estimated life expectancy. Mandatory participations of workers and paid out as annuity (micro longevity risk insured) - Pillar 2 in NL: Occupational pension in NL. Maintain standard of living middle class. Funding (protection bankruptcy employer), actuarially fair (closer link benefit and contributions on individual level) - Pillar 3 in NL: Private personal pensions. To tailor pensions to specific individual circumstances of if not covered by 2nd pillar (self-employed). Actuarially fair. Voluntary but often favoured, high transaction costs PAYG versus funded - Funded: participant has paid for his own pension during working years. Invested in risky assets to get risk premium but have also risk factor - PAYG: current working generations pay for current retirees, via government, AOW - Main difference: in PAYG one pays the pension of the previous generation while in funded pensions on saves one own wealth. In ageing society (increasing dependency ratio) and in particular if the cause is low fertility PAYG gets quite expensive for new generations. PAYG can cover wage risks for individuals even if there are not traded in financial markets. Note that funded pensions can be hit by other risk factors such as equity risk, interest rate risk and inflation. Funded pensions have better property rights. Defined Benefit versus Defined Contribution (benefits or contributions depending on risk factors) - Pure DB: employer guarantees benefits and bears risk, sponsor take all decisions, benefits guaranteed - pure DC: individual participants bear risk (f.e. micro-longevity risk) and has many choice options, contribution is fixed. No redistribution - in DC benefits are adjusted to maintain balance and in DB recovery contribution of injected funds - issues with design of balancing mechanism: which adjustments preferable, who shares in the risk, built into the contract or into legislation in advance or periodic legislative changes or decisions by a board of trustees - old Dutch 2nd pillar: after poor investment returns both benefits can be cut and recovery contributions charged from participants  intermediate between DB/DC - DC in US private sector: participation voluntary (contribution partly paid by employer: nudge towards participation), participation can choose their own asset allocation, possible to withdraw pension capital in case of hardship, possible to withdraw money now and contribute additionally in later years, access to capital at retirement date (no annuities), pension capital is part of bequest (erfenis) - DB in US state and local funds (f.e. teachers, police etc) : different than in private sector. Mandatory participations, guarantees nominal benefits (inflation adjustments partly return dependent), investment decisions by trustees, mandatory annuities, high income close to retirement age, usually underfunded and even more if adequate actuarial valuation is used. Pension liabilities affect public spending. - Key findings giesecke and rauh: introducing DC (ipv DB) could improve the sustainability of the pension plans without negatively impacting participant satisfaction. Side step: Transition from DB to DC in NL - Current: DB but benefits can be cut in case of low funded rates - Drivers of major reform: lack of trust if their will be capital for current young in future and conflicts of interest between generations - Step 1 valuation: current pension wealth allocated to individual accounts - Step 2 adequate exposure: Select adequate investment strategies and plan for both accumulation (saving) and decumulation (spending) phases.  - Social partners (employers and unions) can choose between two plan set-ups under the new law. - Main difference: choice options offered to trustees or to individuals on asset allocation - Solidarity contract is usually selected: still many decisions with trustees (strategies chosen for homogenous groups) and absence of value transfers between individuals in new law (actuarial fairness) - DB requirement for trustees to take adequate investment decisions for all is not well defined if individuals have different characteristics and if in investment decisions moreover affect outsiders (current tax payers) - Conflict of interest also because adequate investment risk differs between participants - Switching to adequate exposures can be attractive for all  Pareto improvement - Dutch approach to break the conflict of interest: o Allocate accrued pension capital to individual accounts o Tailor the investment and decumulation strategies etc. to the characteristics of the individual - Tailored strategies can be collectively or based on individual choice - Avoid allocation to individual accounts and debt recognition by soft (no DB accrual for new employees) or hard freeze (no new DB accrual for everyone) (only step 2) - Arguments for DC: flexibility employees to tailor contributions and benefits to their personal circumstances and preferences, discretion about investment decisions and portability (take money along with you) - Arguments against DC: lack of financial literacy, behavioural biases - DB/DC can be sees an continuum with different options. New dutch law chose: o No sponsor support (no guarantee nominal benefit level but individual guarantee) o No redistribution of wealth to future participants or outsiders (tax payers) o Life long (variable) monthly payments o Benefit level fluctuating up and down with the economy o Mandatory participation and contributions o Asset allocation: imposed individual choice - Other relevant elements Dutch design: o NL focus on level and stability annual benefits o Managing/hedging interest rate risk is key in NL to stabilize annual benefits Mandatory versus voluntary elements - Argument in favor of choice options: o Individuals best know their own preferences and characteristics - Arguments against choice options: o If choices are not actuarily fair introducing them could be costly for the pension provider or other participants o Behavioural biases, such as overconfidence and procrastination (they make behavioural mistakes) o Financial Literacy - A choice option is actuarially fair if the economic value (cost of option) for both options is the same - Example: partner pension (participant pays cost effective contribution) o Alternative: insure all participants and spread the cost across everyone (redistribution) o Sub-issue: contributions vary based on age, but is this based on fully fair choice or solidarity? - Claiming annual benefits early usually implies that the annual benefits will be lower. The choice is actuarially fair if the total value of both income streams is the same Welfare loss of lack of differentiation - Differentiation can be by mandating specific choice for individual with specific characteristics or by offering characteristic dependent defaults (individuals can still opt out) - Examples: more or less offensive investments strategies (based on age, risk tolerance f.e.), implement partner pension, level of required contributions dependent on existing pension wealth, free wealth and housing wealth - The magnitude of welfare changes from both gains (e.g., prepaying a mortgage or reducing contributions) and losses (e.g., poor diversification) are of a similar order of magnitude, roughly 5–7%. This reinforces the importance of tailoring pension contributions and investment strategies to the needs and circumstances of individuals to achieve optimal welfare outcomes. - Standard model: no trigger for lumpsum withdrawals (withdrawing a portion of pension savings upfront rather than receiving it as a lifetime annuity or regular payments) - As the percentage of lump-sum withdrawal increases, welfare losses become significant, especially for those who rely on their pension for long-term income stability - When people get old, welfare loss is quite substantial - Welfare loss due to inadequate product / choice o Using incomplete or generalized assumptions leads to suboptimal strategies o Uniform strategies optimized for an average participant might work for some but can lead to significant welfare losses for individuals with characteristics far from the average. Examples of potential choice options: - Choice to participate in a fund - Choice of contribution rate - Choice of lump sum at retirement - Choice of asset allocation (whether or not to choose ESG investments) - Choice to have or not have partner pension provisions - Choice of date as of which accumulated pension wealth is decumulated - Choice of pension provider Theo Nijman: Financial literacy and (individual) pension choices (20/11) - Differentiation via specific choice for individuals with specific characteristics or by offering characteristics- dependent defaults (individual can still opt out) (f.e. age, risk tolerance, risk capacity, partner pension, housing wealth) - Behavioural arguments: people prefer choices, they don’t use the existing choices, they don’t understand choice options (financial literacy is quite low), procrastination (uitstellen), defaults have a quite important impact - Studies are in favor or libertarian (respect for individual freedom) paternalism (guiding individuals toward better decisions): aims to influence (nudge) peoples choices in a way that improves their welfare while preserving their freedom to choose. Nudged choice = default Preference for choice - In many countries the freedom to choose is valued, in NL/Europa this is less obvious - Henkens and van Dalen (2016): o Considered 5 domains for choice options: Pension contribution, lumpsum withdrawal, fund choice, contract specifics, asset allocations o 2 question: do you value choice in this domain, do you value that others take decision on your behalf o Classification agents: 0 (indifferent), 1 (paternalists; high score 2nd question and low on first), 2 (libertarian paternalists; high score both), 3 (libertarians; high first, low second) o Conclusion: with respect to many potential choice options the Dutch want to have the option to choose, but don’t intent to use the opportunities to choice (libertarian paternalists highest) Behavioural considerations: financial literacy - Financial literacy scores around the world are dramatic; people don’t even understand % - Policy makers embraced financial education (improve education, more emphasis on finance)  to increase financial literacy and financial decision making (But people will always trust f.e. doctor more with healthcare than themselves ) Behaviour and choice architecture - Humans don’t take decisions the way econs (rational decision makers, in economic models) do - Humans use two cognitive systems: o Automatic: uncontrolled, effortless, associative, fast, unconscious, skilled (first) (airplane shakes  I will die) o Reflective: controlled, effortful, deductive, slow, self-aware, rule-following (then) (its turbulence) - Individual judgments are based on: o Anchoring: compare with somewhat similar question o Availability: if certain risk appeared in an example that comes directly to your mind o Representativeness: how similar is what I see to what I expect; people see patterns in independent random draws (they think it’s not possible to throw the 5th time in a row a 4 on the dice) - Drivers individual decision making o Overconfidence: can trigger a lot of risk taking in the domains of investments, life risks (I will not die at a young age) and health risks o Gains and losses: people hate losses, loss aversion presses not to change even if changes would rationally be in your interest (people are not willing to switch: they don’t want to lose what they initially had) o Status quo bias/procrastination: avoids that people adjust to new circumstances, partially due to lack of attention (not cancelling magazine subscription that was free in first month) o Framing: works because individuals are often mindless passive decision makers - Nudges o Nudgeable: can be stimulated to make a certain decision. Pure nudges are liberty preserving. Choice set is not restricted o Libertarian paternalists favor free choice with adequate nudging o Libertarians want to avoid nudging (not always possible; sometimes you have to make a choice) o Others would want to avoid too many choice options or avoid specific choice Theo Nijman: Survivor pensions (25/11) Partner pensions - Partner pensions provide an income for the partner if the insured (e.g. the employee) passes away (similar to orphan pensions, but that is till a certain age. Partner and orphan pensions are survivor pensions) - Relevant questions e.g. on desirable exposure/insurance/risk sharing o Level of income required (NL 70% of 70% so 49% of final/average income) o Should partner pension be lifelong or for a fixed number of years o Who do you want to insure? Only current employees or also former employees? o What happens if one finds new relation? If remarriage do you lose the pension of the first married one? o How can partner pension be tailer to income? What if someone didn’t have an income (only AOW) or the widow earned more? - How are the costs allocated? o Should singles pay for partner pension? Solidarity, is the case in NL. o Should couples with same age pay for partner pension also where partner is much younger? Costfull if widow is 20 and death one is 80 - Goals of partner pensions o Basic income if partner passes away before retirement date (no poverty, similar to AOW) o Provide adequate standard of living for a partner in case of death of the insured (piggy bank) o Deliver product such that its supported in society o Fair cost allocation of the insurance - When is the coverage adequate? o Required income level when living as a couple compared to alone: 70% (scale efficiencies in hh) o PP income lifelong (also if some die very young) in NL o Avoid PP yields incentive no to take up life again after mourning period o PP has no relation on income/ wealth of widow in NL - Poverty alleviation for widows and widowers o In NL: bijstandsuitkering for every inhabitant, bijstandsuitkering levels are lower than minimum wage  Test if eligible by assessing financial situation (available wealth and current income) o Insurance in first pillar after the loss of a partner if one has not yet reached the statutory retirement age  NL: ANW: algemene nabestaande wet (provide similar coverage as AOW) (only if >40% disabled or children below 18 years) (this rule is since 1998, before had everyone ANW)  Stimulate to get back to work (and also cheaper for government)  In NL: only 15% of the relevant population eligible for ANW - First pillar arrangements in a number of European countries o Often not related to number of dependent children or disability o In many countries requires marriage o Available for a limited time only (France, Germany 2 years) o In Switzerland clear differences widow and widowers (currently under discussion) - First pillar coverage is always PAYG so not actuarially fair - Goal second pillar: o Arrange affordable safety net for partners of most employees in case their death with fair cost allocation (retain their standard of living - Goals third pillar: o Offer option to buy supplementary PP to those without second pillar coverage (self employed) or if income level that is insured in PP is too low Current Dutch Partner Pension Provisions - 2nd pillar PP provide income assuming basic ANW level = f * (Y – 15000) (Y is current income and f is fraction), often f = 50% (70*70) of the threshold-adjusted income is provided as partner pension benefits  o Franchise: portion of income excluded from pension accrual because it is covered by state pensions (close to minimum: 15000) - If ANW lacking  gap is not insured  ANW gap - 2nd pillar PP can insurance against lack of ANW coverage or risk van be insured in third pillar - If someone don’t get ANW but only PP: huge drop for low income groups if passes away before 67 (even when 2nd pillar PP if fully covered)  income level will increase with AOW of the living partner - If passes away at age 77: OP before death: 70%*(Y-X), PP after death 50%*(Y-X), X is AOW. So drop of income at retirement age and further drop if partner dies - Problem in current Dutch PP: they assume that everyone get the ANW Second pillar: what is the adequate coverage? - Two main contract specification: o Insurance basis: Insurance for current employees. 70% * 70% * (Y-15000) (partner income is fraction of maximum attainable old age pension) (lower cost) o Capital basis: Insurance for current employees as well as partial insurance for employees that left before passing away: partial coverage partners of past employees (optimal risk coverage) - In both cases similar coverage if risk of passing away during retirement - If someone is consistently employed and fully covered by second pillar pensions, the income streams (retirement or partner pensions) are straightforward and aligned. - Insurance basic becomes problematic if individual becomes self-employed, works abroad or works part-time - For new workers, the costs of setting up insurance basis are lower compared to capital basis systems. - Single employees contribute under both systems: o Under the insurance basis, their contributions don’t generate benefits o Under the capital basis, contributions accrue as savings for potential future partners or can convert into additional old-age pensions for themselves. - The lower costs probably motivated the switch to insurance basis designs in 1998-2006 (rather than view on optimal risk coverage) Concerns about Dutch PP contracts - Lack of 1st pillar PP coverage - 2nd pillar PP o Lack of integration 1st and 2nd pillar (often no 2nd pillar coverage for gaps not covered by 1st pillar) o Lack of transparency :  Many people do not know whether their second pillar contract is based on an insurance or capital basis (big risk if no longer employee and no action taken)  Heterogeneity in percentage of income above franchise that is covered o Insurance-based coverage often depends on the number of years worked with the same employer. o Inadequate Definitions of “Partner” Proposed adjustment Dutch partner pension provisions Reform proposal PP second pillar - Dutch PP remains lifelong, irrespective of current age (first pillar (ANW) starts later (at retirement date) to stimulate widows to go to work) - PP will be percentage of full current salary Y if death employee before retirement date : , fiscal maximum: (substantially higher PP for low income workers), attainable old age pension replaced by current salary (major improvement older workers who didn’t work many years as employee) o Attainable OP: a projection of what the pension would be if the individual worked full-time until retirement - Cost concerns: If budget neutrality (no extra cost for employers) is maintained, better coverage for some may imply worse coverage for others - Lower transparency: differ per pension fund - In proposed law: much smaller income drop low income groups in case of death before retirement  - PP in case death before retirement will be purely insurance basis - PP in case of death before retirement date will be on insurance basis (risk basis), implying that one is not insured any more if one is no longer an employee (unless other insurance is arranged) Employees who leave their job face significant risks, but: o There will be a mandatory extension of insurance coverage for three months after employment ends o Individuals will have the option to voluntarily prolong insurance beyond this period, with the cost deducted from their pension wealth (old-age pension wealth) (max 15 years) - Policy decisions: o What the default should be of voluntarily prolong insurance o Instead of sending an invoice to highlight the cost of continuing insurance (increase cost awareness and reduces double insurance (if found new job), but not sure if people pay and than they are not insured), the proposal suggests deducting costs directly from pension wealth (could deplete OP capital, especially dramatic for those with limited accrued OP capital of for those paying for PP contributions for many years) - PP in case of death after retirement date purely capital basis: benefit depend on individual’s labor market history and investment returns of other accumulated assets - The reform addresses potential issues when death occurs close to the retirement date, where the switch from insurance to capital basis could create gaps or confusion o When death before retirement date PP don’t depend on number of year worked but after it does - Definition of partner for pension purposes will be updated - Pension income for a partner in the event of death before retirement will no longer depend on the number of years the deceased worked with the same employer - Number of years one has worked with an employer will not be a determinant of the level of partner pension in case of death before the retirement date anymore The costs of partner pension insurance Cost of capital basis insurance in case of death after retirement - income for the still alive partner - If there is a guaranteed level of partner income, the cost of the insurance equal the present value of the expected cash flows - Net present value: o Pt: probability that insured is alive in t period from now o P*_{s+t} probability the partner is alive in s+t periods from now o Assume interest rate curve is flat at R - PP substantially cheaper than OP Cost of insurance based coverage in case of death before retirement - Individuals pays contributions for old-age pension and for capital-based partner pension and also for insurance-based coverage for death before retirement - income for the still alive partner - Net present value of insurance at age t o A(t) NP annuity factor, captures the discounted expected value of payments to the partner - As a fraction of the pension base Y-X (income above the franchise), the actuarially fair personal contribution rate is: - Cost must be high for incomes just above X - Better coverage for low income people, PP more expensive for older people (not actuarially fair). New system is more expensive than old system, Income and age effect in personal contribution rate - In new PP (if death before retirement age): income level is 50% of full salary - Required fraction of Y-X to cover costs based on realistic assumption on survival rates and interest rate  - Contribution rates increase with age under the new system: Younger participant face lower rates, rates peak at 57 years. (heterogeneity in age) - Uniform contribution rate across all incomes would be actuarially unfair; higher income groups would subsidize protection for lower-income households - If the contribution rate is tied to personal income, the burden for low-income households becomes disproportionately high under the new system. - Setting the contribution as a percentage of full income shifts the cost burden to high-income earners, potentially addressing inequities. - The new proposal significantly increases the contribution rate for low-income earners, especially as they age, which raises concerns about affordability and equity. - Policymakers face a trade-off between uniform rates (equity concerns) and actuarial fairness (higher personal rates for low-income households). Determinant of cost of insurance - Which fraction to be covered - Full income Y or pension-base Y-X - Age, partner age - Discount rate - Survival probabilities (gender composition, education level) - If accrued PP is used to lower insurance costs: accrued NP capital - Partner frequency: do bachelors pay in for this insurance - Payment length (currently lifelong) - Economics costs also depends on factors such as gender, education level and health status but not allow to use these in pricing Extension and impact on OP pension - If insurer still employed, all mandatory contribution rates subtracted from gross wage to find net wage - Impact on OP in case of uniform contribution rates: o Exhaustion depends on age and number of years accrued so far o In case of full accrual short period of coverage small impact on OP o If less accrual and long coverage impact can ger substantial - When people choose to voluntarily extend their coverage, they need to make personal contributions (This is because competition and the option to get insurance elsewhere make it necessary to fund the extension personally.) For people with low incomes, the amount required for personal contributions may be higher than the uniform contributions typically applied. - When personal contributions are used to fund the extension, the occupational pension capital is affected more significantly for low-income individuals than what was shown in a previous table. Miscellaneous (diversen) - More design issues: like duration, tax policy, divorce, remarriage, and multiple widowhoods. These points aim to address fairness, sustainability, and incentives for partners to remain engaged in the workforce or receive adequate support. Conclusion - Reasons to reconsider current PP legislation: lack of adequate coverage for low incomes, lack of transparency who is covered, lack of transparency on who is a partner - New law provides lot of scope for discussion on adequate implementation Roel Mehlkopf: Pension institutions around the world (4/12) - Why look at other countries’ pension systems? Population ageing, less guarantees, more choice, move from DB to DC how to ensure adequate saving, how to structure the pay-out phase,…… (similar trends and challenges, different responses) - Mercer CFA institute Global Pension Index 2024 o Looks at different aspects of the entire pension system: across all pillars o Scores 3 subcategories- adequacy, sustainability, integrity – which result in an overall reating o Index uses world bank’s 4 pillar approach o 40% adequacy, 35% sustainability and 25% integrity - NL, Iceland, Denmark and Israel are top-rated (grade A), but have very different pension systems - Japan’s index score is in the lowest category – along with countries with much lower GDP/capita o They score low in sustainability because their population is really ageing. 50/50 retirees and workers  so score can also say something about the underlying and not the pension system itself S1: pension plan coverage - Proportion of the working age population that are members of private pension plans - Highest coverage when pension saving is mandatory, either through law or labour agreement - Alternative automatic enrolment with opt-out - Case: Automatic enrolment in the UK o Participation of occupational pension plans was voluntary but from 2012 there was automatic enrolment (participation in workplace pensions increase by 30%) o But the required contributions rates are low - Voluntary participation rates in eg Italy and France are very low (10-20%), where is there is no mandatory participation of private pension scheme. This is not a problem because they have a large public first pillar.  high adequacy of pensions but sustainability is a challenge in view of population ageing - S1 is relevant but coverage alone is not enough if the contributions rates are very low o Relevant for sustainability but not for adequacy if the first pillar already proves enough benefits S2: level of pension assets - Amount of pension assets as a % of GDP - Large variation between funded and PAYG countries; developed and developing funded pension pillars - High level of pension assets good indicator because the availability of funded system provided diversification for different risks and having a pool of assets to back up a pension system increases the likelihood that the pension system will be able to make future payments  so level of assets set aside for pension is positive feature of pension system (the higher, the better) - Use different funding methods (PAYG, capital-funding) to diversifies risks - Risks of high level of pension assets: o Used as other objectives than pension saving: Examples from Hungary and Poland (nationalised private pension systems) o Large pool of assets with a long term objective may attract unwelcome attention: there may be political pressure to use these assets for other purposes if the need arises, for instance in times of economic distress o Large funded pension pillar make pensions relatively more vulnerable to financial market risks o Large funded pension pillar makes pensions relatively more vulnerable to inflation risk and/or interest rate risk (what matters for pensions is the combi of both: fluctuations in real interest rate) o In closed economy, large funded pension pillar can drive down the interest rates and thereby lower the returns from pension saving o When investment returns are lower than population and productivity growth, the rate of return of a funded pension system is lower than a PAYG system S3: Population ageing - Pensions become more expensive when retirement lasts longer  - Most public pensions are PAYG financed: old-age dependency ration and fertility rate - Solutions to reduce population ageing o Increase pension age and/or link it to life expectancy o Stimulate immigration o Reduce dependency on PAYG pensions, which are more sensitive to population ageing o Stimulate birth rate, e.g. by providing free child care - Population ageing is a good indicator in relation to public and PAYG pensions and in relation to pension systems with guaranteed and lifelong pension payments. Both need different indicators (old age dependency ratio or fertility rate vs life expectancy at retirement) - Increasing the statutory pension (legal) age is only part of the solution: also the effective pension age (when people actually stop) needs to increase. A6: Proportion of benefits takes as income stream - Pensions could be taken as a income stream (annuity) or/and as (tax-free) lump sum - Potential pitfall of greater pension freedom is that people make flawed decisions and pension assets leaking from the system - Potential benefit of giving people freedom to take out (some of) their pension savings: create confidence among the members, especially in an environment where confidence in financial markets or service providers is low (in Ghana the saving go up when members had the possibility to take out 50% of their savings) - Discussion on pay-out phase more relevant since many pension systems are moving from DB to DC - It is advisable that there is a requirement to convert (large) part of the pension pot into an income stream, eg an annuity, a pooled arrangement or a programmed withdrawal product. - Allowing retirees to take out a limited amount of their pension as a ‘lump-sum’ provides greater flexibility. - It makes sense to develop products like reverse mortgages that enable pensioners to make more effective use of their assets (including non-financial assets). A10: Growth assets - The % of growth assets (equities, property,..) of total assets - A reasonable % of growth assets gives an indication of pension plan’s ability to generate long-term returns, which translates into higher benefits  adequacy - The assets mix should allow for diversification of asset classes to optimise risk/returns - The optimal % of growth assets depends on, among other factors, age and risk tolerance of the population - Free choice of investments has pitfalls: design adequate choice architecture - It’s not problematic that the 2nd pillar in Sweden allows for a lot of choice in investments, because the 2nd pillar is small and mainly relevant of higher incomes so taking more risk in the 2nd pillar may therefore be less problematic and match the participants’ risk profiles S6: Government debt: - Level of adjust ed government debt of GDP - Higher government debt restricts the (future) government in supporting the elderly  sustainability - Public pensions are often the largest single item of social expenditure, accounting for 18.4% of total government spending on average in 2017 - Large variation among OECD countries: especially with a large earnings-related PAYG first pillar score poorly on this indicator, mainly due to high public pension expenditure Conclusion - No single, perfect solution to organize a pension system - However, some overall recommendations: - Increase coverage of pension systems –including self-employed. Many individuals will not save without compulsion or auto-enrollment!  - Increase pension age, including effective retirement age, while life expectancy increases, to reduce costs of pensions. This includes promoting higher labor force participation at older ages. - Ensure that a sizeable part of the pension is taken as an income stream, and sensible defaults are in place for investing pension capital. - Diversify risks within a pension system, by including both a PAYG and a funded element. Make sure that the funding is used for what it is meant: paying pensions! Sample Exam Doorsneesystematiek (uniform accrual and contribution rates) - Doorsneesystematiek implies that young workers and older worker pay the same contribution rate and receive the same accrual (transactie) of their pension income (both as a % of pensionable wage) (case in the current Dutch pension contract) - Actuarially unfair year by year (assuming positive interest rates) since the young have to wait more years before they receive the pension income. This is less of an issue if all employees stay employee during their full working years and all have the same wage profile. If that is not the case those with little income increase over the years and those that are not an employee withing the system in later working years (e.f. worked abroad, self-employed, part-time, or didn’t work) pay in more than what they receive which is often seen as undesirable - Transition to a new system without the doorsnee-systematiek is costly because many current employees have paid for previous generations but would not receive subsidy in later years. Depending on the interest rate the costs can be say 5% of total Dutch pension wealth. 2 35-year old individuals with same characteristics but 1 is a homeowner and 1 is a renter - There can be a rationale to invest pension assets less risky under the assumption that a house is a ‘stock’-like asset that is strongly correlated with the returns of risk assets such as stocks in the investment portfolio. Under this assumption, a homeowner already has a large exposure to a risky stock-like asset and this reduced the appetite for risky assets in the investment mix of the pension wealth - There can be a rationale to invest pension assets more risky under the assumption that a house is a bond-like asset that provides an individual with a lifelong income stream in the form of reduced expenditures of the cost-of-living (rent on a house). In particular, a house can be regarded as a bond that pays of cashflows that are linked to the cost-of-living, and thence have a similarity with an inflation-linked bond. Under this assumption, a homeowner already has a large exposure to a risky bond-like assets and this can increase the appetite for risky assts in the investment mix of the pension wealth Economic rationale to share macro-longevity risk between young and old - Assumptions needed for risk-sharing of macro-longevity risk between young and old can be welfare enhancing: o if young and old differ from each other in terms of their level of risk aversion o if young and old differ from each other in terms of their risk exposure to macro-longevity shocks - Risk sharing can enhance welfare if the impact of the longevity shock on the consumption is smaller for the young than for the old - Risk sharing solution can be exante welfare improving for young and old: o For old: have reduced exposure to longevity socks o For young: if young are less affected by longevity shock than the old (e.g. by postponed retirement date due to increased life expectancy  human capital increases: natural hedge against longevity risk via human capital), as can make the young willing to absorb a part of the longevity risk of the old o In an ideal case for risk sharing: the young may not need to reduce their consumption level at all in response to longevity shocks if the ratio between the number of remaining working years versus the number of retirement years remains unaffected by a longevity shock. o For young: if risk sharing contract includes a risk compensation (risk premium) to be paid by the old. In absence of risk compensation, the consumption of the young becomes more volatile without  being higher in expectation. If the young are risk averse, then this would reduced their ex ante welfare instead of improving it Risk sharing in a pre-funded (contributions are invested in financial markets to provide for future retirement benefits) pension system between current and future generations - How risk sharing between current and future generations can be Pareto-improving (beneficial for everyone) from an ex-ante perspective. o Future generations get the opportunity to invest early and benefit from potential risk premium. (don’t have a choice for this) o Current generations can reduce their exposure to financial risk by spreading it across more people (including future generations). o This approach removes the biological trading constraint; the fact that people typically cany invest in financial markets before they are born or start working - From an ex-ante perspective: intergenerational risk sharing can be welfare improving for all generation, but from an ex-post perspective some generations can be worse off. - The first-best risk sharing solution can be vulnerable to discontinuity risk when the expost realization of investment returns is much lower than its ex ante expectations o Young generation may not be willing to join the risk sharing arrangement (and not take on a deficit that is shifted upon them from old generations) because it has become clear that the risk sharing contract makes them worse off - Why it can also be vulnerable if ex-post investment returns are much higher than expected. o Old generations may have an incentive to terminate the risk sharing arrangement (and not pass a surplus onto young/future generation) because it has become clear that the risk sharing contract makes them worse off. Guest Lecture Marcus Haveman: A suitable individual pension (NN) (27/11) Suitable individual pension options DC-contracts; collective and individual - Solidary pension agreement: collective lifecycle, no individual choices, with compulsory obligate solidarity elements like a protection yield and a solidarity reserve - Individual Flexible Premium agreement: Investment choices available, right to shop on retirement date, with the option to purchase a fixed pension on retirement date, but not before the retirement date, with risk- sharing reserve (solidarity) (optional). Lot of options to make it tailored to your own preferences. Participants are employers and employees. - Default lifecycle: which type of LC participants standard starts, decreases variable LC, on pension date, participants can choose themselves, social partners opt for variable LC (otherwise the default LC is fixed, 20% expected difference between those two options) Different types of Lifecycles and investments - How much higher is the expected pension compared to the invested DC-premiums: more than 3.5 time (3.5 times the paid premium will be the expected pension) for fixed and more than 4 for variable - You need to choose fixed or variable annuity and the pension provider long before retirement - What percentage of participants filled in the risk profile? 30%-50% but they need to be activated - ALM study: o in younger years, equity risk gives best risk-return ratio on pension date because of a long horizon (hedging interest rate risk reduces interest rate risk but investing in equity’s gives better risk return ratio no pension date. In this way, we make trade-offs between different risk sources) o 9-18 years phase out period to pension date o During pension phase we see different risk profiles - Lifecycle based on academic models? o No consensus on the best method to measure risk preferences o Gives extreme lifecycles for young participants where participants can lose al their financial capital o Offer no ALM optimization of investments (not enough asset categories, academic lifecycle can be optimized with ALM study) - NN: offers 5 risk profiles (differences between risk preferences and risk capacity between participants) o Fixed guaranteed pension o Other 4 has steps of 15% more return (so more variable) o Spreading investment results over time to avoid extreme fluctuations - What percentage of pensioners make active choice on pension date for a specific profile in an individual DC- contract? 99.9% - 1/3 that chose for the variable annuity chose the most risky profile - There is different between participants (scientific methods and their own research showed that) - Lifecycles: which choices can you make? o Choice of risk profile o 15 years before retirement date: choice to reduce to fixed or variable pension o Choose which retirement date you want to reduce the LifeCycle to o Right to shop on retirement date or roll over to fixed or different variable pension profiles to the pension provider with the best offer o Types of investment: responsible (passively optimized, general exclusions of companies that have damaged their reputation, extra investments in companies that support transition, 1% tracking error), sustainable (More companies that support certain SDG’s, partly in impact investing such as green/social bonds, tracking error=3%) or impact investing (best in class in combination with substantial exclusions, other themes besides green and social bonds, tracking error=6-8%) o Tracking error: deviation from the benchmark - Pension funds don’t ask about inflation to their participants, they think for the participants and think indexation is the best (than the pension is more stable in real terms) Determining a suitable risk profile and activation How do you ask appropriate questions? - Make sure you have a communication expert on your team who can help you write understandable questions - Avoid Framing: explain for example the pros and cons of a variable pension. Add a behavioural scientists to your team who can help you conduct a food qualitative research - Qualitative testing of questions through in-dept interviews with several participants by a qualified interviewer. To ensure that participants understand the question and fill it in accordance with the objective - Quantitative test of question under a group of participants. Never be surprised by unexpected guidance towards certain profiles. Necessary research to develop a suitable questionnaire? - With sufficient capacity to test qualitatively and quantitatively - Qualitative questions o + Explanation about return and risk investing o - lack of concrete numbers makes participants cautious/defensive - Show concrete numbers on a simple basis o Simple example about the amount expected and pessimistic pension of various lifecycles  + simple way to provide insight: concrete and insightful  - simplification of reality - Including simple figure, not too difficult but simplification of reality - Pension amounts over time (URM numbers in table with weather scenario) o + provides insight into changes in benefits o – complex due to large number of numbers - Qualitative questions + CS method + Amounts over time Choice guidance and activation - Experimental tests: an pension adjusted by inflation increases the risk preferences - Activation o Persoonlijke welkomstvideo o Young people need different activation than older people - Heatmap when most people log in  when is the best time to send an activation mail Guest Lecture DNB: Pension adequacy (toereikenheid) in the Netherlands (21/10) - Low poverty rates among elderly in OECD, NL highest score in pension adequacy - Goal recent research: examine adequacy of pension savings in NL: who has too little or too much pension savings, provide policy options to improve retirement adequacy - Level of retirement savings should depend on income, inflation, fixed expenditure Dutch pension system - 1st: State pension: AOW, PAYG, depend on min wage, retirement age linked to life expectancy - 2nd: Occupational pensions: capital funded, income-related contributions by employers and employees, >85% coverage. Mandatory for some self-employed occupations but in general not o New 2nd pillar: DC ipv DB. Individual pension capital with collective reserves. - 3 : Private pension products: supplementary, tax-facilitated for those who don’t make full use of 2nd pillar rd o New pension law: increase in tax-facilitation th - “4 ”: Savings, private wealth Data (Microdata from statistics NL)pension adequacy - 1st pillar: previous pension rights observed until 2022, future: full pension right accrual (assume no emigration) and assume no changes in hh composition - 2nd: based on 2022 pension system, assumption: job and income doesn’t change - 3rd: annual contribution observed since 2011 (from taxed), future: 5y-average contribution until retirement, assumption: 2% real annual return - 4th: liquid (immediately available) is lower bound and total private wealth (incl housing wealth) is upper bound. Assume: sell everything immediately when retired. Assume: 2% real annual return Results - Expected gross retirement income in median euros and Expected gross replacement rates in median % - Median doesn’t tell full story because expected gross annual pension income from 2nd and 4th pillar is heavily right skewed (very high for pensions with highest current income) - First pillar replaced >90% of previous income for lowest decile - Households with expected retirement income below poverty line are mainly hh born in other countries (refugees and people that worked not long enough in NL to get 100% pension) - Self-employed rely more on illiquid wealth (4th pillar) - More variation replacement rate at lowest income deciles - Median retirement income strongly related to worksector Conclusion IBO - Dutch pension system is in general effective: 1st pillar prevents poverty for majority of HH, most workers are able to maintain their living standard to a certain degree in retirement, for specific groups the expected income is too low or too high - Ongoing reforms in the 2nd pillar pension are expected to improve the balance in pension adequacy - At this moment, a substantial reform of the pension system is not recommended, however pension adequacy should be monitored the coming years Policy options for poverty prevention - Improve take-up of allowance for hh with incomplete first pillar - Reduce first pillar accrual period from 50 to 40 years (get 100% full pension after 40 work years)  not recommended because reduce total funding available - Enhance pension accrual in other pillars  reduce reliance on first pillar alone Policy options to increase replacement rates - Increase pension contributions of employees with insufficient pensions - Mandatory 2nd pillar pension with or without opt-out