Bogleheads

[grok87]

[grok87]

[grok87]

[yobria]

[speedbump101]

[magellan]

[saurabhec]

[Lbill]

Grok said:

[logger]

[saurabhec]

[yobria]

[magellan]

[grok87]

[Rodc]

[grok87]

[Valuethinker]

[Rodc]

[Lbill]

[magellan]

[Rodc]

[saurabhec]

[ken250]

It's pretty easy to wave one's hand and tell someone to be prepared to lose every red cent they've put into the market or tell them they'll earn 10% per year. The cynical customers probably have already faced the possibility of losing everything and the optimists are already expecting 10% per year.

Maybe an MCS is a waste of time for those customer types.

But what about people who can't decide on going 80/20 vs 50/50, or how that choice will impact meeting their needs? Seems to me an MCS could be useful in this case, assuming the customer understands the results are based on combinations of historical returns.

[Rodc]

[nisiprius]

[ken250]

I didn't mean to imply precision, this caution would have to be impressed on the customer...historical returns, future may not be like the past, unpredictable markets, etc. In fact the whole investment process has to be undertaken with those warnings.

Again, not to imply precision...For example, a couple has a long time horizon and they're saving for retirement. The MCS might show that there's a 20% chance of the 50/50 of not meeting the target, while the 80/20 only has a 5% chance of not hitting it. This might encourage them to be more aggressive, not necessarily adopting 80/20 but going more aggressive than 50/50.

I do think the use of tools has to be undertaken on a case by case basis though. If you get a grognard as a customer computer-generated output could really turn him off to the process, if you and he have a good relationship then a good honest face-to-face might be preferable.

[speedbump101]

Interesting thread...

All... How would you handle this real life senario?

1) Age xxx with xxx years until planned retirement.
2) xxx dollars currently saved.
3) He / she wants to know the annual contributions and AA required to reach the critical portfolio size to allow for xxx yrs of inflation protected retirement with a high probability of success.

I can think of two methods of approaching this problem:

1) the 4% SWR (Trinity study)
2) a MC simulation

Is there a better, or at least a different way of approaching this problem? Or maybe a better question is, how are you approaching this universal problem?

SB...
_________________
“Knowledge is often mistaken for intelligence. This is like mistaking a cup of milk for a cow."- Bradford K Hull

[grok87]

[nisiprius]

[Lbill]

Nisiprius - I agree with your view. Two rules I try to keep in mind:

1. Don't take any more risk than necessary to reach your investment goals; i.e, don't load up on too much equity in your portfolio.

2. If your investment goals entail taking on a lot of risk, then you should consider changing your goals; i.e., if you have to take on more than a modest helping of stocks in your portfolio then either (a) get a higher paying job and/or retire later or (b) rethink your desired standard of living in retirement. A lot of boomers playing catchup with a stock-heavy portfolio got slaughtered in 2007-2008. They learned the hard way.

The idea that anybody should have more than one-half their investment assets (I prefer no more than one-quarter) in stocks is a bad idea IMO - unless they can afford to lose that money. Milevsky recommends taking a look at how much your portfolio can lose over the next 3 years in the worst-case scenario (odds less than 1 out of 100). If taking that loss would be psychologically damaging to you, or would seriously impair your ability to reach your income goals in current, or future, retirement then you probably have a portfolio that is too risky for you. And remember that the riskiness of stocks does not decline over time. As Samuelson recommends, you should not invest any more in stocks over the long run than you would be willing to lose over the short run.
_________________
I can tolerate risk -- it's losing money that bothers me.

[ken250]

speedbump,

There's 2 steps, greatly simplified.

1) First you need to determine how much you'll need at retirement, call that R. R will support your yearly inflation adjusted income needs, and also take into consideration how much you want to leave behind.

2) In order to meet R at retirement you need to figure out much you have now and how much more you can add.

It all boils down to gathering inputs, estimating retirement needs, making reasonable/conservative assumptions for returns and inflation, knowing your time horizon, etc. Once you collect all this you're faced with a few Time Value of Money problems.

Once you know the returns you'll need in accumulation and in retirement you can go about designing those 2 portfolios.

[saurabhec]

[Lbill]

[larryswedroe]

few thoughts
First history is one set of returns. They could have played out differently--even the same long term returns but different order matters when in withdrawal. Thus those relying on history may be understating risks

Second, as I noted earlier there is often more than one goal--besides not being broke while alive--many investors have other goals like a bequeth of a certain size and the AA impacts the two different in terms of odds of success. If not an MC how else?

Third, as also noted by others the fat tails tend to show up more in the very short term data. As you look at longer periods they tend to be less of an issue. One reason is corrective actions are taken to offset them (countercyclical policy). Not a guarantee of course.



MC not perfect but the enemy of the good is often the perfect

[grok87]

[saurabhec]

[Rodc]

[saurabhec]

[grok87]

[will23]

[Rodc]

[will23]

[speedbump101]

Do you think? .... naw, no way!

CBS Money Watch

SB...
_________________
“Knowledge is often mistaken for intelligence. This is like mistaking a cup of milk for a cow."- Bradford K Hull

[tadamsmar]

What does gauge extreme events in markets?

[Rodc]

[tadamsmar]

I think focusing on the fact that MC cannot gauge extreme events represents the failure to gauge the problems of the OP citation and, indeed, a common confusion among investors, including many Bogleheads.

The citation in the OP claims that MC did not identify events as extreme as last year's market downturn. That might be true of MC based on normal distributions, but it is clearly false for some that use historical distributions like the Trinity Study (and financialengines probably but I have not check that one.)

I reviewed my Diehard plan after last year's downturn. The most interesting insight I had from this was that the MC sources for the plan did take events as extreme as last year's downturn into account and was robust against events of that magnitude.

However, one problem I saw was failure to gauge the difference between an MC study and a retirement plan. The Trinity Study studies the ratio of nest egg at retirement vs burn rate during retirement. You can use the study to estimate the size of the nest egg you need at retirement. However, it does not solve the problem of having that nest egg available on a fixed retirement date given the tendancy of markets to fluxuate.

[grayfox]

I don't see what the problem is. Just run the MC simulation, print out a nice report with fancy color charts, leather-bound with gold-leaf lettering, and give the client a multi-media powerpoint presentation with surround sound in a fancy conference room with a huge mahogany conference table. Then just print the signature below in fine print on page 147. That should cover it.
_________________
Information is provided "as is" without expressed or implied warranty of any kind including warranties of merchantability or fitness of use. For entertainment or educational purposes only.

[drjdpowell]