Friday, April 26, 2013

Tweets from the Pension Research Council Conference

For the past two days I've attended the Pension Research Council annual conference at the Wharton School. Though the conference focuses more on pension management for institutional investors, there was still lots of interesting food for thought for individual investors. What follows is a collection of tweets I made during the conference.  Since this is a Twitter feed, it starts from most recent to oldest. For this to make sense, you will need to scroll down to the bottom and then work your way back up.



Wade Pfau@WadePfau 39m

Evan Inglis of Vanguard: Is the future just going to happen and is it misleading to attach parameters and expected returns to it?

Expand


Wade Pfau@WadePfau 45m

Anna Rappaport: for middle-market, asset management is not so important. Housing is a key component of wealth and need to incorporate that

Expand

Wade Pfau@WadePfau 49m


Me: Retirees may still have a long-term investment horizon, but it is key to realize sequence of returns risk effectively shortens that

Expand

Wade Pfau@WadePfau 51m

Marlena Lee of DFA: in low yield environment, it is important to stay focused on total returns portfolio, not income portfolio

Expand

Wade Pfau@WadePfau 1h


Guyle Wilson of Mercer: normal distribution assumptions works fairly well for longer return periods such as a year

Expand

Wade Pfau@WadePfau 1h

Marlena Lee of DFA: Key assumption for Monte Carlo which affects results: the expected returns assumption

Expand

Wade Pfau@WadePfau 1h

Marlena Lee of DFA: including mean reversion (bootstrapping 10 year sequences) reduces range of wealth accumulations slightly

Expand

Wade Pfau@WadePfau 1h

Marlena Lee of DFA confirms what I thought as well: normal distributions (no fat tails) give similar results as bootstrapping (fat tails)

Expand

Wade Pfau@WadePfau 19h

...  from the last several decades."

Expand

Wade Pfau@WadePfau 19h

Hodgson: "Retirement as currently configured was never affordable, but this fact was hidden by demographic and debt trends ...

Expand

Wade Pfau@WadePfau 20h

Stefan Lundbergh of Cardano: For inspiration on pension reform, study: (1) Sweden (2) The Netherlands (3) UK

Expand

Wade Pfau@WadePfau 20h

Hodgson's retirement anomaly: Global pension assets add to 78% of GDP, but they need to be 235% of GDP... can amt grow w/o decreased return?

Expand

Wade Pfau@WadePfau 20h

Tim Hodgson of Towers Watson: For risk, can we say it won't happen  in the future vs. it just hasn't happened in the past

Expand

Wade Pfau@WadePfau 23h

Kessler of SOA: mortality data is bad and out of date, worst for extreme ages, complexity may not help, but some known factors such as educ.

Expand

Wade Pfau@WadePfau 23h

Kessler of SOA: mortality varies by period, age, cohort and country, and so not easy to model

Expand

Wade Pfau@WadePfau 23h

mortality cohort effects: follows the same individuals over time and see how individual cohorts enjoy improved mortality over time

Expand

Wade Pfau@WadePfau 23h

mortality age effects: not clear patterns for one age over time

Expand

Wade Pfau@WadePfau 23h

Kessler of SOA: Mortality improvements combine into period effects (mortality rates decline across ages at same time - such as less smoking)

Expand

Wade Pfau@WadePfau 23h

Developing mortality improvement rates combines science, art, and educated guesses

Expand

Wade Pfau@WadePfau 23h

Kessler of SOA: now research is focusing on developing specific mortality improvement rates for age and calendar year

Expand

Wade Pfau@WadePfau 23h

Kessler of SOA: History shows steady longevity improvements, which must be built into future mortality assumptions

Expand

Wade Pfau@WadePfau 23h

Emily Kessler of Society of Actuaries: Longevity risk is systematic, not idiosyncratic, and so it can't be hedged

Expand

Wade Pfau@WadePfau 25 Apr

Jim Moore of PIMCO: missing fat tails in multivariate normal can give misleading answers in other parts of the model

Expand

Wade Pfau@WadePfau 25 Apr

Long-term returns modeling: assuming multivariate normal may miss fat tails, but nonetheless provides more variation in outcomes

Expand

Wade Pfau@WadePfau 25 Apr

Jim Moore of PIMCO: since 1871, geometric real equity returns over 30 year periods ranged from 3.5 to 10%... what to assume for modeling?

Expand
Reply Delete Favorite More

Wade Pfau@WadePfau 25 Apr

Jim Moore of PIMCO: modeling long-term returns, long-term convergence (less to more): bootstrapping, multivariate normal, regime switching

Expand

Wade Pfau@WadePfau 25 Apr

Live Tweets from the Pension Research Council Conference at the Wharton School



2 comments:

  1. Wade, thank you for these useful excerpts. They could almost make me take Twitter seriously.

    Would you care to expand at all on the Hodgson remark? (Hodgson: "Retirement as currently configured was never affordable, but this fact was hidden by demographic and debt trends from the last several decades.") I looked online for something of his that might address this, bugt couldn't find anything. Did he go into any specifics, or was it just a hook?

    ReplyDelete
    Replies
    1. Hi, I've heard this sort of idea expressed before. Essentially, it's the idea that people can't really expect to support 20-30 years of non-work with only 30-40 years of work. But this was masked by the pay-as-you-go nature of pensions as well as excellent market returns.

      Delete