1. 1 Design features of an IA System: What Can We Learn from Other Countries? By Estelle James

2. 2 Reasons for pre-funding Pre-funding makes system more sustainable, less sensitive to demographic change Earns investment return, requires lower payroll tax to maintain benefits (after transition stage) Avoids passing large debt to our children Pre-funding can help to increase national saving, therefore productivity and growth

3. 3 Problems with public management of funds Under current arrangement for trust fund, treasury gets exclusive access to funds, iou’s don’t count in public debt so government may borrow more today—this becomes taxpayers problem tomorrow If trust fund is invested in stock market leads to: –conflict of interest between govt as regulator & investor –political lobbying for inclusion, exclusion In other countries public funds get lower rate of return, political manipulation, misallocated capital Piecemeal reforms are necessary but inevitably involve build-up of trust funds—how can we avoid their use to finance larger government debt? IA’s are one way out.

4. 4 Private investment also entails problems High administrative costs, financial market risk, accumulation may be used up too quickly, some workers may fall below poverty, transition cost problem I will outline how other countries solved these problems in their IA systems Devil is in the details, in answers to key design issues.

5. 5 1. What is best mix of funded and PAYG benefits? Big range: Chile and most Latin America—almost 100% of benefits from IA’s, except for safety net— minimum pension guarantee (MPG) Sweden—2.5% tax goes into IA’s, 16% goes into PAYG, safety net financed by general revenues. IA’s only 15% of total payroll tax but expected to be 30% of total benefit Australia, Switzerland, Netherlands, UK, Poland, Hungary—private funded pillar 40-50% of total Mixture of PAYG & funding diversifies risks

6. 6 2. If private, who chooses investment managers? Workers—Latin America, Eastern & Central Europe (retail market-may lead to high costs) Employers—Switzerland, Netherlands, Denmark, Australia, Hong Kong –Long tradition of employer-sponsored plans, government recently made them mandatory –Group plans cheaper to administer (institutional market ) –Initially DB, but shift toward defined contribution (DC) –I f DC worker bears risk and will demand choice of investment manager & strategy Thrift Saving Plan in US, Bolivia, Kosovo— –competitive bidding process for fund managers— –workers get lower cost with limited choice

7. 7 3. How to keep administrative costs low Important to keep costs low: If expenses are 1% of assets annually, pension falls 20% Chile—administrative costs high at start-up but about 1% of assets now (< US mutual funds) Costs higher in new plans in rest of Latin America, E. Europe; also US 401(k)’s and Australian retail plans But costs only.1 of assets in TSP,.3-.5% in large employer plans in Holland, Australia, Switzerland Difficult to keep costs low when DC accounts are small, especially at start-up. How can we solve this problem?

8. 8 For low investment & marketing costs use institutional market—limit choice Most Latin American and Eastern European countries use retail market—many asset managers sell directly to individuals. High marketing cost. Bolivia, Kosovo, TSP, large employer plans, use institutional (wholesale) market— –small number of asset managers through competitive bidding –Aggregate small accounts to capture scale economies, increase bargaining power, reduce marketing costs. Use passive investing (indexing to benchmark like S&P500 or global index) Collect contributions through tax system, centralize record-keeping Amortize high start-up costs over time Should keep our IA costs to.3% of assets

9. 9 4. How to reduce risk and provide safety net to low earners Stock market volatility is definitely a problem Every country with IA has minimum pension or flat benefit that sets floor (15-30% average wage) –MPG or floor in most Latin America, Eastern Europe –Flat or almost flat benefit in Western Europe, Australia –We should consider floor tied to years of work Require broad diversification investments Portfolios restricted in Latin America &E. Europe –But not enough international diversification –Chile limits % in stocks after age 55 (less return, less volatility, less subject to date of retirement risk) –UK has too much choice and workers made mistakes –Carefully structure limited choice (like TSP) is best

10. 10 5. Guarantees Relative rate of return guarantee common, herding, index to benchmarks better Switzerland also sets floor on interest rate paid by pension funds & insurance companies—problems Private market can provide “collar”--floor + upside potential--but must be carefully regulated (costly, credibility of guarantor, hard to evaluate price) Public guarantees often cost more than expected, moral hazard problems

11. 11 6. How to handle payouts How can we be sure that worker won’t spend all his money before he dies? Annuities guarantee workers life-long income. Can private sector handle this? Will annuities be offered on good terms? Will workers buy? Chile has had IA system since 1982 –many workers have retired, government requires annuities or gradual withdrawals up to threshold of 70% replacement rate--2/3 have annuitized –Chile requires that pensions are price-indexed, joint for married men--inflation insurance, protection for widows, at no public cost (indexed annuities difficult) –Pensioners who work don’t have to contribute to their account—encourages work Private sector can handle payout stage –But only if government sets careful rules of the game –Competitive bidding process for annuities (TSP)?

12. 12 7. How to protect women Women work less, earn less, live longer, and if married their husband are older, so they become widows with low incomes. Very old women are pockets of poverty in many countries. Women are the biggest gainers in lifetime income from the pension reform in Latin America. Why? –Minimum pension guarantee or flat benefit helps low earners –Survivors benefits purchased through group insurance contract during working stage –At retirement, husbands required to purchase joint pension--financed by husbands (in US by taxpayers) –Widows keep their own pension plus survivor’s pension (in US widows must choose between the two; many women who work & contribute get no additional benefit)

13. 13 8. How to handle the transition Chile, Latin America, Central & Eastern Europe— used carve-out from existing payroll tax that was high at start. Therefore transition financing gap –Most downsized current benefits that were too generous –Chile accumulated fiscal surplus before reform but still has large financing deficit (also deficit in old system) –Debt finance played large role in most cases Australia, Netherlands, Switzerland, Denmark, Hong Kong—add-on, no transition cost problem US- We have relatively low pension debt and trust fund surplus so carve-out possible; but low contribution rate so add-on possible. Mixture might be ideal. –to increase national saving we shouldn’t finance transition mainly by increasing public debt (otherwise higher personal saving offset by public dissaving)

14. 14 Conclusion Mixture of PAYG DB and funded IA would make social security sustainable, protect workers, and improve economy—at least cost. The devil is in the details. IA’s can make us better or worse off. What is best design? –Competitive bidding, index funds, limited choice, diversification, gradual move out of stocks –Minimum pension tied to years of work –Annuities up to reasonable threshold –Strategy for covering transition financing gap –Avoid too much choice, retail market, no safety net, inadequate annuitization, pure debt-financed transition Ultimately we will have to make value judgment between reducing benefits vs. raising payroll taxes —but IA can give us better terms of trade