Ruthen, Run the Numbers:

1
Ruthen,Run the Numbers

• Longitudinal Microsimulation for Individuals
• Rick Morrison
• richard.morrison3_at_sympatico.ca

2
Goals Paper a first step toward

• Substantially increasing the pool of longitudinal
  microsimulation practitioners
• Improving, the effectiveness and sophistication
  of personal financial planning, particularly as
  regards the treatment of uncertainty
• Making some of you rich
• Providing governments with another tool in
  support of intelligent tax policies

3
Weaknesses in Financial Planning

• Inherent uncertainty is largely ignored
• Simplistic rules for savings and returns
• Frequent absence of personalization to clients
  assumptions
• Advisors potential conflicts of interest
  regarding recommendations
• Little attention to spending down savings
• Limited attention to after-tax, constant dollar
  representation of lifetime finances

4
But These Same Issues Are Longitudinal
Microsimulations Strengths

• We live by probabilities, events, and
  distributions
• Were used to implementing complex algorithms
• We routinely parameterize everything
• Were used to performing objective, neutral
  analyses
• We naturally treat full lifetimes, with
  end-of-life wrap-ups
• We typically model entire tax/transfer systems,
  and carry out our analyses in constant dollars

5
Potential A New Type of Longitudinal
Microsimulation Model

• Adapt existing longitudinal technologies
• DONT project representative populations!
• INSTEAD, we generate many independent projections
  of a single (client) individual, using his/her
  (a) assumptions and(b) strategy for financial
  decisions
• Result is a reasonable distribution of the
  consequences of using that strategy.

6
What Wed Get From It

• Distributions to reflect inherent uncertainty
• Opportunity to systematically apply sophisticated
  strategies, improve them
• Personalization to whatever extent desired
• Capacity to be scrupulously objective,
  quantitative, regarding conclusions
• Explicit attention to savings drawdowns
• Natural presentation of constant-dollar,
  after-tax results as the norm

7
Ruthen in Single Strategy Mode
8
And In Comparative Mode

• Analogous to Base/Option analyses
• Simulate consequences of TWO strategies
  simultaneously, using same random s
• Results would show not only the absolute
  consequences, but the comparative consequences
  between the two strategies
• Provides a foundation for evolving good
  strategies, optimizing, measuring tradeoffs

9
Ruthen in Comparative Mode
10
Whats Adapted From Existing Models?

• Longitudinal simulation via events for all
  relevant life events and financial transactions
• Replicable streams of high-quality pseudo-random
  deviates
• Control and output structures
• Focus on distributions and their presentations
• Base versus option perspective

11
Some New Capacities Required

• Convenient specification of assumptions, e.g.,
  mortality relative to standard tables
• Strategies for saving, dis-saving, and portfolio
  adjustments, including housing
• Assessing outcomes and comparative outcomes, and
  their distributions
• Presenting results effectively in tables and
  graphs distributions, comparative distributions,
  summary measures

12
Selected Uses For Individuals

• What is the expected real after-tax rate of
  return associated with my financial strategy?
• Whats the expected estate Ill leave?
• When can I afford to retire? (constrained choice)
• What kinds of insurance should I buy, and how
  much of it?
• Should I rent or own my home?
• Do management expense ratios affect my retirement
  consumption significantly?

13
Selected Uses for Financial Planners

• Deriving personalized financial strategies for
  clients to meet their needs and reflect their
  assumptions
• Showing value-added for the financial planning
  services one offers
• Assessing whether/when various rules of thumb
  make sense for clients

14
Illustrative Institutional Use 1

• Derive for the government sector the internal
  rates of return for individuals savings
• When individuals save, the government sector
  loses revenues reduced sales taxes and possibly
  income tax revenues
• Subsequently, as savings are drawn down (by
  choice or death) there are additional revenues
  for the government sector

15
Illustrative Institutional Use 2

• Government sector sees taxes on returns to
  investments, reduced income-tested benefit
  payouts, increased sales taxes upon consumption,
  greater probate taxes
• The series of these net flows (averaged across
  independent simulations) implies an internal rate
  of return for the government sector
• Useful to gauge size of governmental IRR
  absolutely, and relative to the individuals IRR
  for the same savings strategy

16
Challenges/Caveats

• Use of strategies is not well developed (really
  an opportunity?)
• Finances not the only thing important
• Form of objective function unclear
• Degree of analyst and client comfort in dealing
  with distributions
• Applicability clearer for individuals than for
  families (but most people live in families).

17
What We Might Well Find 1

• Saving for retirement often/typically lowers real
  discounted lifetime consumption
• Its expensive, in terms of current foregone
  consumption, to push consumption into the future,
  e.g., into retirement
• Real financial returns to later-life work are
  small, particularly for those with significant
  pension entitlements

18
What We Might Well Find 2

• The primary beneficiary of individuals decision
  to save is the government sector
• Governments enjoy a much higher internal rate of
  return to citizens savings than the citizens
  obtain on those same savings
• Individuals, world-wide, are being quite rational
  in reducing their voluntary savings rates in the
  face of systemic incentives

19
Next Steps

• Building such a model will be my focus in my next
  career, post-DYNACAN
• Preliminary steps, and a modelette, are under
  way
• I think the project is going to be challenging,
  but a lot of fun, and with the potential to do a
  fair bit of good
• Theres ample room for many players

20

Thanks!
21
Inherent Uncertainty

• Seldom know when one will die
• Uncertainty about returns to ones investments
• Statistically, ones work and earnings are highly
  variable over time
• Possibility of disability, fraud and theft are
  often ignored
• Retirement is not an on/off switch

22
Simplistic Rules

• Plan on living until age 95
• Save 10 of your gross income
• At age X, hold Y of your assets in stocks
• Spend only the returns to your capital
• Keep Z of a years expenditures in an emergency
  fund
• Assume constant inflation and constant real
  returns on investments

23
Absence of Personalization

• Mortality based on family history
• Distribution of investment returns as a function
  of personal risk tolerance
• Personal assumptions about the future of
  inflation
• Details of ones employer pension plan
• Health strengths and susceptibilities

24
Potential Advisor Conflicts of Interest

• Choices of investments may have different returns
  to advisors
• Over-savings generally leads to bigger earnings
  by advisors
• Advisors typically unable to provide objective
  evidence of the value added by their services

25
Spending Down Ones Savings

• Literature is spectacularly weak in indicating
  what one can safely spend in retirement without
  leaving an undesirably large estate
• Planned retirement spending typically doesnt
  take into account statistically likely changes in
  activities and health-related spending
• Over-savings equals under-consumption

26
Limited Attention to Real After-Tax

• Many plans deal only with gross, pre-tax income,
  rather than consumable income
• Many advisors ignore impacts of income-tested
  benefits
• Many advisors deal primarily in nominal, rather
  than real, incomes and expenses
• Effects of using up/drawing down existing
  stocks in retirement may be ignored