Close Close

Financial Planning > Behavioral Finance

My eMoney Review: Good, but What’s With Monte Carlo?

Your article was successfully shared with the contacts you provided.

In my last few posts (My Wish List for Redtail and eMoneyViewing Financial Planning Software From a Client’s Perspective) I’ve been discussing some new software I’ve adopted this year. In this writing, I’ll continue with some of the functionality of eMoney.

As you know, I’ve been using eMoney for a number of months now and have had nothing but good things to say. After getting into the details of the program, I have found a few shortfalls. For instance, with residential rental property, the net rental income is calculated on IRS Form 1040, Schedule E. This schedule begins with the gross rental income which is reduced by the expenses associated with the property. Then there is a deduction for depreciation. If there is a mortgage on it, the interest is part of the deductible expenses.

In eMoney, the interest is also included in the client’s rental property expenses, but so is the entire mortgage payment. Therefore, the total expenses are overstated by the amount of the interest. The program should differentiate between rental real estate and a personal residence. 

I have also found another curious issue. At the end of one plan I created (age 100) the client has over $30 million in financial assets using a linear forecast. At age 100, the probability of success (defined as the probability of not running out of money) using Monte Carlo simulation (MCS) is only around 64%. This is highly unlikely. The longer you project into the future the more uncertain the outcome. This is measured by the cone of uncertainty. However, when the slope of the financial assets continues to rise on a linear basis, especially when it reaches such a high level ($30 million in this case), the success rate should be much higher. Why would it forecast a 64% success rate? Here’s what I suspect may be occurring.

Let’s assume from today through the end of the forecast period (age 100) there are 60 years. When applying MCS (ex: 1,000 trials), some years will result in positive returns and some will be negative. MCS uses random sampling within the stated parameters (i.e. the distribution surrounding the MCS variables) and varies the results accordingly. With one trial, it might model negative returns in years 1, 2, 5, 10, 12, 13, 15-18, etc. With another trial, the positive and negative returns will vary. Each time a trial is run, the program remembers the terminal value of the financial assets. 

Since there is no serial correlation in stock market returns, if one year is good, the following year may be good or bad. Unlike inflation or money market returns, stock market returns from one year to the next are uncorrelated. If the program models mostly negative returns in one trial and mostly positive returns in another trial, the cone of uncertainty would continue to expand as the time horizon increases. However, this would occur in a horizontal pattern rather than an upward or downward sloping cone. In this case, the longer the period modeled, the more probable it would be that the financial assets would fall below zero, reducing the success rate. I suspect this may be the case, but I am only speculating. Having run thousands of simulations and researching MCS over the past 15 years, the success rate in a situation such as described above would have to be higher.

Even with these issues, I find eMoney to be one of the best options on the market today. There may be a financial planning program which is slightly better, but when you add the ability to link outside accounts, the client’s personal website, and online document storage functionality, eMoney is a top-shelf program.

The specific curiosities I’ve described can be remedied. And, once they’ve been addressed, eMoney will be that much better. I think the future is bright with this program and I’m still a big fan!