It seems you have it right. There are three portfolios--the joint non-retirement assets (regular assets) and the one for husband and one for wife. So the program knows the current balance on each of those accounts as you say because you entered it in the Assets and Saving folder and in the Retirement Accounts folder. The retirement account balance that you call "joint portfolio" needs to be assigned either to the husband or the wife since it's a retirement portfolio (qualified account, i.e., not regular assets).

Using Economics Mode, you would enter an assumed nominal rate of return for each of these three pools of money in the Assumptions area. But since you are using Monte Carlo (MC) mode, the rate of return on these pools of money is determine by the historical rate of return and variance of those asset classes.

Obviously the variance and rate of return you actually get going five, ten, fifteen years into the future may be higher or lower than the historical return/variance of any particular asset class. Will we get the epic risk-adjusted return in bonds going forward that we've had over the last ten years? Who knows. Keep in mind that some of your asset classes may have just a ten-year history making their future predictive ability for the next 40 years based on a relatively brief snapshot in a particular economic time period. The long history on equities for the past 70 years may or may not be representative of the next five or ten (or the next 70 for that matter). That's the nature of MC analysis and using historical returns to predict future returns. But it is a common way to do probability analysis and since nobody has a crystal ball, historical returns and variability is all we have to do distribution analysis like MC provides.

I like to use Economics mode and run my model using very modest assumptions on all three pools of money, note the discretionary spending, and then walk up the assumptions in increments making more optimistic assumptions.

What my approach ignores, however, is that having a relatively large pool of regular and retirement assets exposes one's personal economy to much more risk than does having a relatively small amount--that is, relative to fixed assets like a pension and Social Security. As Kotlikoff says in Spend Til the End, the poor can afford to take MORE risk than the rich (because a portfolio loss would matter less).

I believe one value of using non-MC analysis is that you can stipulate the average nominal return on each of the three pools of money and experiment to find out how low the return can be while still having happy or acceptable results in discretionary spending--then go adjust the asset allocation to match. The question becomes: how little risk can I take? instead of asking, how high of a return can I chase?

Just some thoughts there . .. for what it's worth.

I created a set of portfolios representing all of my tax-deferred holdings at 40/60, 50/50 and 60/40 asset allocations, did the same for my wife's tax-deferred holdings, and a set for our joint holdings. My holdings represent the investments available to me through my employer's plan and my retail IRAs. My wife's represent the investments available through her employer and her retail IRAs.

Until retirement I assign the respective "my" and "her" portfolios to each of our accounts in ESP. I change the portfolios over time to reflect the asset allocation in each set as it changes to conform to our asset allocation plans.

When I retire I switch my accounts from "my" to the "joint" portfolios. That's what will result of my rolling over my employer accounts into my IRA. My wife will do the same when she retires. All of our retail accounts are held at the same broker. Our post-retirement asset allocation will be uniform at that point.

Chris
(a user)