FINANCIAL PLANNING ENGINES REVISITED
Financial Planning Engines Revisited
In the Q3 issue of the Regional Review, Boston Fed Senior Economist Peter Fortune reviewed two financial planning tools that have been developed by leading economists and that are currently offered for sale. In his review of ESPlanner, Fortune reported that his simulations produced some puzzling results, and he critiqued the tool for a lack of transparency that would make the basis for the results apparent. Professor Laurence Kotlikoff, one of the creators of ESPlanner, responds below. A brief note from Professor William F. Sharpe, who developed financialengines.com, also appears.
Peter Fortune's article, Financial Planning Engines: Motoring Toward a Better Future, misrepresented my company's financial planning product, ESPlanner in seven ways.
First, Dr. Fortune suggested that the program recommends an unreasonably large amount of life insurance. He reached this conclusion because of the data he gave the program. Specifically, Dr. Fortune told the program that each of his hypothetical spouses would stop working if the other died. Under those circumstances, the spouses would, indeed, need a tremendous amount of insurance. Dr. Fortune apparently intended to have survivors keep working. He could have told the program that with just two clicks of the mouse. Instructions for doing so are directly available on the earnings folder, in the manual, and in the software's tutorial. Apparently, Dr. Fortune read none of these guides and ignored the check box and copy button on the earnings screen. With the corrected inputs, the program recommends less than $600,000 in life insurance for the two spouses combined, not the $2.3 million reported by Dr. Fortune.
Second, the large dissaving recommendation reported by Dr. Fortune reflects addition errors he made in inputting data. While Dr. Fortune entered the couple's current tax-deferred contributions to 401(k) and IRA plans, he failed to enter any of their future contributions. ESPlanner, having been told that the couple had this large off-the-top contribution to make this year and this year only, properly told the couple to cover this expense from its previous savings, rather than experience a large decline in its living standard. The program didn't tell the couple to go to the mall as Dr. Fortune alleges. Indeed, the whole point of the program is to maintain a household's living standard over time and to keep it from splurging or starving.
Dr. Fortune also told ESPlanner that the couple was willing to go into debt and borrow up to $30,000 against its future earnings. This isn't something most couples would want to do. In addition, most couples would want to set aside a modest emergency fund. Such set-asides are inputted in ESPlanner's special expenditures folder. Finally, Dr. Fortune told ESPlanner that each spouse would experience real wage growth of 2 percent throughout their late 50s and early 60s. Most couples are likely to see their real earnings peak well before retirement or at least wouldn't count on such earnings growth.
With these modifications, ESPlanner recommends that Dr. Fortune's couple save a positive, albeit modest amount in 2000. The reason ESPlanner doesn't recommend large amounts of non-tax-deferred saving in 2000 is that the couple is already saving a large amount in tax-deferred form. Furthermore, when retired, the couple will receive Social Security benefits and will have paid off their mortgage. Indeed, if ESPlanner is run without letting the couple go into debt, the software, while smoothing the couple's living standard to the maximum extent possible, can no longer make it perfectly smooth. For the case being discussed, the couple doesn't save much in 2000 because it will otherwise be enjoying an 83 percent higher living standard in old age.
Third, Dr. Fortune states that ESPlanner is a black box whose calculations cannot be checked. This is unfair. Had Dr. Fortune read our manual's tutorial (which can be downloaded for free at www.esplanner.com), he would have learned how easy it is to check that the program is making the right calculations. One does so by observing from the generated balance sheet that households die broke, by observing from our consumption recommendations that households have the same living standard, adjusted for household composition each year unless they are borrowing-constrained, and by observing that survivors are able to afford the same living standard if the couple purchases the recommended level of life insurance.
Fourth, Dr. Fortune states: The mortality assumptions and insurance premiums embedded in the analysis are not clear. The manual and our research papers, to which Dr. Fortune refers, indicate that we are using actual insurance premiums charged in the marketplace, in this case those charged by TIAA-CREF. The program doesn't use mortality probabilities because it plans for all possible dates of death and considers the case that survivors will live to their maximum ages of life.
Fifth, Dr. Fortune says, One takes on faith the modeling of federal and state taxes, Social Security calculations... Not so. Our manual spends thirty pages detailing the federal income tax calculator, each state's income tax calculator, our FICA tax calculator, and our Social Security benefit calculator. No other financial planning program on the market provides as much detail about its tax and Social Security benefit calculations. For example, in discussing the program's federal income tax calculations, we point out that these calculations a) are made separately for each future year and for each potential survival configuration of the household; b) determine whether the household should itemize its deductions; c) determine whether the household is eligible for the earned income tax credit; d) determine whether the household is eligible for the child-tax credit; e) determine whether the household's Social Security benefits are taxable; f) adjust for the indexation of deductions, exemptions, and tax brackets; g) adjust for capital gains taxation; and h) properly tax 401(k) and other tax-favored saving vehicles. The manual not only indicates that these elements are included in our federal income tax calculator, it specifies prevailing tax rates, exemption and deduction levels, tax credit provisions, and so on.
Sixth, Dr. Fortune states: The emphasis on taxation and Social Security benefits necessitates a change in the program whenever new laws are passed. One wonders how ESPlanner will keep itself up to date, and I already suspect that it has not. According to information in one of the scholarly papers, ESPlanner assumes that Massachusetts residents pay a 12 percent tax rate on interest and dividend income, and on short-term capital gains. This was the law before 1999, but for 1999 and after, the tax rate on interest and dividends is only 5.95 percent. Again, had Dr. Fortune bothered to glance at our manual, he would have learned that the recent change in the Massachusetts tax rate on capital income is incorporated in ESPlanner. Indeed, the first line of our description of Massachusetts income tax calculations, which appears on page 99 of our manual, is Massachusetts taxes labor and interest and dividend income at a 5.95 percent rate. This change in our program was made available in a free update on our website the same day we learned about the tax change. As another example, the recent change in the earnings test was incorporated in an update to ESPlanner the same day that President Clinton signed the law. The research for the scholarly paper to which Dr. Fortune refers was done prior to the change in the law. This paper, by the way, is posted on our website, www.esplanner.com, and shows huge discrepancies in saving and insurance recommendations between ESPlanner and Quicken Financial Planner.
Seventh, Dr. Fortune suggests that ESPlanner requires inputting an immense amount of data. In fact, once one is familiar with the interface, inputting the data takes about 30 minutes, roughly the same amount of time needed to input data in Quicken Financial Planner. Moreover, all the data requested by ESPlanner are critical to generate an appropriate plan.
Finally, Dr. Fortune fairly criticizes ESPlanner for providing no dynamic analysis of portfolio positions. But had he asked us about that or anything else, he would have learned that ESPlanner2001 will include Monte Carlo simulations that show how a household's portfolio holdings affect the variability of its future living standard.
ESPlanner is really very easy to use if you are willing to read the brief instructions available on each input page and each report. Reviewers have a special obligation to understand a product before they start criticizing it.
Professor of Economics
William F. Sharpe
Professor of Finance, Emeritus