The Shift from DB to DC: Economic Impact on Plan Members Presentation to: Workplace Pensions: Next Generation or Final Frontier Centre for Law in the Contemporary Workplace, Queen’s Law School Toronto, May 2014 Bob Baldwin

Cautions Terms DB and DC lack precise meanings Heterogeneity within each of DB and DC DB and DC data may reflect characteristics other than inherent differences EG: size, differences in governance and management, level of earnings (RPP needs) Data not perfectly fit for purpose Relevance will trump reliability

Extent of the Shift from DB to DC: PPIC Data All DB94%91%85%73% DC5%8%13%16% Private Sector DB90%85%77%51% DC8%13%20%29% Note too: In 2012, 488,000 members in plans transitioning from DB to DC (compared to 1 m in DC plans) PPIC source picks up registered RPPs but not group RRSPs In 2011, 3m participate in group RRSPs (3 x the number in registered DC) Migration from small registered DC to group RRSP Growth in small DB in the private sector

Retirement Wealth Participation in an employment pension matters (SFS, 2012) Among households aged 55 to 64, 77% have private pension assets with a median value of $304K 69% have RRSPs with a median value of $50K Will aim for SFS data on DB vs DC wealth MVA per member in DB and DC plans (PPIC) In 2012, the average MVA/member in DB plans was $252K vs $50K for DC plans DB plan members more likely to identify their workplace pension as their primary source of retirement income 64% vs 47% (Survey of Financial Preparedness, 2009)

An Accounting Identity Contributions + investment returns = benefits + expenses Underlines a conceptual point and helps organize information Where do investment and other uncertainties show up in DB and DC Pure DB: all on the contribution rate Pure DC: all on the benefits A reformulation: In DB the rate of saving is adjusted regularly to meet the benefit obligations; in DC the benefit is adjusted to reflect the contributions and investment returns

DB vs DC: The General Picture Lower level of contributions in DC Dollar amounts of contributions per member higher in DB than in registered DC ($9K vs $3.8K, 2010) – normalized rate data would be preferred Actuarial firm identifies 8% as typical DC contribution rate Life industry spokesperson: employer GRRSP contributions 4% Marginally lower investment returns in DC Evidence is from US, sample bias (DC all members) Slight advantage to DB but not consistent from year to year (TW study: 0.76 bps over 17 yrs through 2011; 0.39 in last 5 yrs – net of estimated costs) Well documented problems in US 401K investments But vs a DC counterfactual, in DB group and individual returns can diverge Higher fund management and admin expenses in DC Jog to F/P/T ministers on fund management fees: Private DB bps; Public DB bps; corp DC 60 bps; corp RRSP 92 bps; retail 40 for ETF/index and 200 on average for active with advice. Other sources provide a mixed picture Large size (to a point) and capacity for in-house fund management are clearest determinants of cost

The Predictability Problem with DC DC outcomes depend heavily on conditions in financial markets around retirement age Purchases of deferred annuities and life cycle investing can limit the impact – at a price MS modelling DC replacement rates vary from a high of 48% (2,000) to 14% (1977): investment volatility and salary growth Blake modelling: 2 scenarios Scenario 1: DC portfolio and DB are identical M result, DC = 100% of DB, 23% chance DC>DB and 44% chance of DC < 25% DB Scenario 2 (life cycle): DC portfolio shifts to low risk assets 5 years before retirement M result, DC = 72 % of DB result, only 7% chance of DC DB NB: mirror image in DB Variable contributions (consumption opportunities foregone) Amplitude of variation may be smaller in DB (amortization periods, smoothing of values, etc)

Comfort: An Additional Consideration No need to worry about Investment decisions (low priced access to expertise) Longevity risk Converting assets from accumulation to decumulation But DC is entrenched Can be organized to reduce discomfort

Outcomes with No Change to DC Impacts greatest for people earning above half AWS Greatest need for RPP income to meet earnings replacement goals Impacts on retirement income and retirement age Less certainty about amounts of income and retirement age Likely lower income and later retirement Some attenuation thanks to interactions with other components of the RIS, income tested benefits and progressive PIT Impacts more clear on gross versus net replacement rates DB entails variable foregone consumption opportunities prior to retirement

Final Reflections 1 Future context may be difficult for DB and DC Ratio of the retirement period to the period of working life Gap between returns on financial assets and salary growth But, tighter labour market context may make DB pensions more interesting for employers retention, orderly exit, and mid career recruitment will become more important Pension design is not a binary choice between DB and DC Need to explore designs that involve elements of DB and DC – especially in the private sector Predictability of contributions and benefits are both desired

Final Reflections 2 Single firms are not appropriate organizational platforms for pensions Too small, lack expertise, and conflicts Even with CPP expansion or ORPP, it is important to get workplace pensions right Next generation or final frontier Next generation: pensions as we know them are young; we are learning how they function in different environments