Discussion of the paper by Daniele Fano:

1
Discussion of the paper by Daniele
FanoAssessing risk in household portfolios
progress and open issues at the macro and
micro/survey level

• Monica Paiella
• Bank of Italy, Research Department

2
Defining portfolio risk (1)

• ... as volatility of returns
• straightforward implications as to the
  classification of assets
• implication that there is a unique optimal
  risky portfolio
• Is volatility a good measure of portfolio
  risk?
• It depends on the question one wants to address
• and on the data one uses!

3
Defining portfolio risk (2)

• Generally speaking, the assessment of hh.
  portfolio risk cannot ignore
• saving motives/LC issues (which determine the
  length of the investors horizon)
• background risk (labor income risk,
  health/longevity risk, risk from ownership of
  illiquid assets, liquidity constraints)
• transaction costs
• implication that different risky portfolios
  might be appropriate for different investors
• these are issues that micro data could help
  address

4
NA/FA data (1)

• They are useful for
• across sector and across country comparison of
  portfolio risk, as measured by the volatility of
  returns
• an assessment of the time trends in portfolio
  risk
• but ...

5
NA/FA data (2)

• they are not informative as to who bears the
  risk among hhs
• extensive vs. intensive margin
• issues of information on financial investment
  opportunities
• young vs. elderly (horizon of the investor)
• issues of savings for retirement
• rich vs. poor
• transaction costs/minimum investment requirements
• diversification issues/imperfection in capital
  markets
• DRRA

6
Survey data

• They allow to
• evaluate individual portfolio differences in
  terms of composition
• relate the differences to individual
  socio-demographic characteristics
• Hence, they allow to appraise individual
  portfolio risk in terms of
• return volatility
• time horizon of investors
• covariance with income/consumption risk

7
Time-horizon of investors (1)

• It changes over the LC
• It is crucial to appraise the riskiness of hh
  portfolios and of specific financial instruments
• cash is risky in the LR, even though it is safe
  in the SR, because it must be reinvested in the
  future at unknown real interest rates
• inflation-indexed bonds provide a known stream of
  LR real payments even though their capital value
  is unknown
• evidence that stocks can support a stable
  standard of living more successfully than their
  SR price variability would indicate (Modigliani
  and Stuch, 1966 Stiglitz, 1970, Rubinstein,
  1976)
• These issues are crucial from a policy
  perspective when appraising hh savings for
  retirement

8
Time-horizon of investors (2)

• Related empirical issue
• conditional portfolio shares appear not to vary
  across age groups.
• Appraising portfolio shares based on (single)
  surveys is a tricky issue due to cohort effects!

9
Covariance between financial and consumption risks

• Labor income can be thought of as the dividend of
  human wealth
• Investors cannot trade in their human wealth, but
  they can adjust their financial asset holdings to
  take into account the riskiness of their implicit
  holdings of human wealth
• Liquidity constraints

10
Consistency of information based on macro and
micro data (1)

• Different types of data give somewhat different
  pictures of household wealth and portfolios
• per capita aggregate wealth gt per (survey) hh
  wealth
• the difference is larger, the more concentrated
  the asset
• These differences are due to measurement issues
• reporting errors, evaluation of (non-traded)
  stock
• sampling/coverage errors (under-representation
  of the rich)

11
Consistency of information (2)

• Combining the information helps to highlight
• the differences between the portfolio of the
  representative/average household and that of the
  median household
• hence, the implications deriving from the
  distribution of wealth on risk

12
Concluding

• That of the assessment of household portfolio
  riskiness is very interesting issue and the
  author raises several crucial points, but
  overall the paper is still at a preliminary
  stage
• what is the question the paper wants to address?
• Methodological paper?
• Survey of issues?
• Policy question?