Wednesday, July 10, 2019

My Preferred Planning Software is MaxiFi

I've been working on a research paper with UNC econometrician, Neville Francis for the past year and that has given me the opportunity to look at several free online retirement planners. Overall, I have to say that most were disappointing.

I have also worked for several years with another online retirement planning tool that is not free but is quite affordable, economist Laurence Kotlikoff's MaxiFi.[1] I recently asked Dr. Kotlikoff some questions about his product.

Dr. Kotlikoff, you say that MaxiFi is based on "consumption smoothing", the "proposition that households want to have a stable standard of living through time as well as across good times and bad times." What does that mean to a retiree or to someone saving for retirement?
Consumption smoothing is at the heart of economics-based financial planning. It's firmly anchored in human physiology. None of us wants to splurge today and starve tomorrow. Nor do we seek the opposite. Whether retired or still working, rich or poor, we're after the same thing — a highly stable living standard. Leaving aside issues of investment risk, the core financial planning question is how much to save each year to achieve a smooth consumption ride. MaxiFi calculates this directly based on your lifetime resources net of future taxes and gross of future Social Security benefits. In so doing, MaxiFi eliminates the guesswork in planning your retirement finances. It also helps you find investment strategies that limit your investment risk. In contrast, conventional financial planning asks you to set a goal for annual retirement spending. My goal is $1 billion.
A retirement planner recently commented to me that retirees don't all want "smooth consumption"; some want to spend more early in retirement. But spending more at some ages than others isn't inconsistent with "smooth" consumption, is it?  
MaxiFi has a Standard of Living Index that lets you tell the program you'd like to have a higher living standard earlier in life and a lower one later on. The tool will recommend discretionary spending that follows your desired living standard path as closely as possible subject to not putting you in debt. You can also specify special expenditures, like a major trip when you reach 70. MaxiFi will budget for this and have you pay for it by spending less ever year before and after the trip.
Most retirement planning tools measure success with "probability of ruin", or the percentage of simulated future scenarios in which a retiree can expect to not outlive their savings. Please explain why you prefer consumption smoothing.
Conventional planning is built on three mistakes. First, it asks people their retirement spending targets. Mine is $1 billion a week. So right away I've made a mistake. But even if I guess a "reasonable" number, I'm going to be miles off the level that MaxiFi will calculate. Second, conventional planning assumes you'll keep saving what you are now saving. That's mistake number 2. What you are now saving is surely wrong. The third mistake is assuming you'll spend your targeted amount year after year in retirement whether your assets go through the roof or fall through the floor.

Conventional planning's "probability of ruin" Monte Carlo simulations calculate the chance you'll run out of money if you make all three mistakes, i.e., if you a) save the wrong amount each year before retirement, b) spend the wrong amount year after year after retirement, and c) never adjust your annual spending once you retire. I can't fathom why anyone would wish to know the probability of financial survival in the context of making three major financial mistakes. Financial planning is supposed to help us make the right financial decisions, not tell us something we don't want to know about something we shouldn't be doing.
I can find lots of free "single-purpose" planning tools on the internet, tax planners, sustainable withdrawal rate calculators, life expectancy calculators, Social Security optimizers, RMD calculators, asset allocators, etc. Is there an advantage to incorporating them into a single program like MaxiFi?
All our financial decisions are interconnected. Take life insurance. You can't decide how much to buy until you know the living standard you need to insure. But your sustainable living standard (if no one dies) depends on the amount of insurance premiums you'll be paying. So, your living standard and life insurance needs must be jointly calculated. MaxiFi does this. It jointly handles all the factors you mention and more. The advantage of MaxiFi's integrated financial planning is that all its suggestions and calculations, including federal and state taxes, are absolutely internally consistent. If you use piecemeal calculators you'll get a set of suggestions that don't add up.
MaxiFi asks for only a few of my expenses as input. Why is that?
MaxiFi asks you to specify your "off the top" expenses on housing and other must-spend items, like alimony payments, out-of-pocket medical expenses, or college tuition. These expenditures are like negative income. Your other resources less a) these off-the-top expenses and b) your lifetime taxes determine your lifetime budget — what you can spend on a discretionary basis over the rest of your life. MaxiFi then smooths this spending. If we were to ask you to specify everything you were going to spend each year, year in and year out, you'd give us amounts that were either a) unaffordable or b) left some of your lifetime budget on the table.
Is MaxiFi a "Monte Carlo" simulator?
MaxiFi does Monte Carlo simulations on your living standard. It calculates 500 living standard trajectories you might experience based on how you are investing. It then compares these 500 trajectories with 500 based on investing more safely and 500 based on investing at greater risk. These trajectories take into account that you'll adjust your spending annually in light of how well your investments fare, always with the goal of having a stable living standard. Best yet, MaxiFi combines all of the 500 trajectories in a single index of your average lifetime happiness — what economists call your Expected Lifetime Utility. This index, which takes into account your tolerance for risk, lets you compare in terms of three numbers (one for each of the three sets of 500 trajectories) how your current investment strategy stacks up against investing at less or more risk. Lifetime expected utility maximization is the gold standard of economics-based portfolio guidance.
Can MaxiFi tell me if I should purchase life insurance or an annuity?
Absolutely. It calculates how much term life insurance you need to hold each year to ensure survivors have the same living standard to the dollar had you not died. It also shows you how much higher or lower your living standard will be if you purchase an annuity.
Can I perform what-if analyses with MaxiFi? What kinds of things can I test?
You can set up as many alternative profiles as you'd like and compare them against your base case in terms of their lifetime discretionary spending. For example, you can easily learn how much more you'll get to spend if you downsize or if you go back to work or if you switch jobs or if you annuitize your retirement accounts or if you wait to take your Social Security benefits.

But MaxiFi also does its own what-ifs for you. Once you run your base plan, MaxiFi asks you to MaxiFi It. When you run this report, MaxiFi looks for safe ways to raise your living standard by maximizing your lifetime Social Security benefits and finding the retirement account withdrawal strategy that will reduce your lifetime taxes.
Where can I learn more about how MaxiFi works?
Go to www.maxifi.com. Check out the videos, the case studies, and other descriptions posted there. And then try it! I promise, you'll get hooked on its ability to safely raise your living standard and finally take the guess work out of financial planning.
(Note: If you prefer video instruction, I have added two links below to recent MaxiFi Webinars.)[2,3]

Those are some of the reasons Dr. Kotlikoff believes MaxiFi's economics-based approach is best. Now, here's why I like it.

At $99 per year with $70 renewals, it's quite affordable for the do-it-yourselfer.

Dr. Kotlikoff and his team have steadily improved and refined the product, beginning with E$Planner, for over 25 years. That leaves the others with a lot of catching up to do with both the economics and the technology.

As a computer scientist, I know from experience that Dr. Kotlikoff has a top-notch technical staff and their help desk has always been available when I needed it with real people who know their product.

MaxiFi completely avoids the limitations of probability-of-ruin estimation. Instead, it incorporates consumption smoothing and maximizies the utility of achievable spending.

Many retirement planning tools address only the decumulation phase, when we retire and begin spending down our wealth. MaxiFi is a life-cycle planner and is useful at any stage.

Lastly, as Dr. Kotlikoff mentions, MaxFi integrates many calculations into a single model. Most free online simulators handle only a part of the problem, like maximizing Social Security benefits or modeling investment returns. Retirement planning isn't a problem that can be solved by solving many individual sub-problems independently.

If you're interested in financial planning software, give MaxiFi a try. You can use it to build a retirement plan or to create a "second opinion" of one you already have. It's also a good tool for your annual retirement plan checkup.

I rarely promote products at my blog but I know that many of my readers are do-it-yourselfers and many have expressed interest in software tools. I have a lot of confidence in MaxiFi. A multi-client version called MaxiFi Pro is available for advisors.

There are a number of new entrants into the online retirement planning field and I'll keep looking for free or affordable, unbiased, comprehensive planning tools. If you are especially fond of another tool that shares these attributes, please add a comment below.

To be clear, I don't believe that software can effectively replace a good human retirement planner given the current state of the technology, though the latter will no doubt cost more. I think you'd be way better off using a good human planner who uses good planning software. But for now, at least, I prefer MaxiFi for the do-it-yourselfer.



OTHER RESOURCES

Economist, Zvi Bodie now links to his "trusted sources" at https://zvibodie.com/trusted-resources/. I find the entire website very useful and particularly the videos.

NewRetirement.com provides a wealth of retirement planning software. I encourage you to take a look. Full disclosure, I act as an advisor to NewRetirement.


REFERENCES

[1] MaxiFi web-based planner, website.



[2] MaxiFi Webinar, June 26, 2019, VIDEO.



[3] MaxiFi Webinar, June 13, 2019,VIDEO.





9 comments:

  1. The link to the Maxifi web-based planner site in the first footnote is broken.

    ReplyDelete
  2. Hi Dirk. I reviewed the video and web site. It looks like there is a MaxiFi Planner for $99 and a Premium edition for $139 that includes Monte Carlo analysis and perhaps other features I'm not yet aware of. Do you have a recommendation for your readers on which edition to purchase?

    ReplyDelete
  3. Having played with Maxifi for a couple of years, it is a very powerful tool for looking comprehensively at consumption smoothing.
    As with all of these programs, there are some monumental assumptions that are applied that will dramatically impact the output. With Maxifi, I find the input variable that remains the most problematic is that of future growth rate of ones portfolio. Guess high (overestimate rate of return), and the risk of outliving one's assets is front and center. Guess low (underestimate rate of return), and one leaves assets behind. That being said, I have found the program both robust, and challenging. Each time I come back to it, I learn something new.
    An attached user's manual would be of great help.
    Thanks for all of the info you have provided to us lurkers over these many years. You have helped me immensely.

    ReplyDelete
    Replies
    1. Good comments. I suspect Dr. Kotlikoff would say that you should guess on the low side of market returns and create a plan that works even if the market performs poorly. I guess the bigger point is that market returns are unpredictable, as are many other retirement planning assumptions, and running simulations on a supercomputer until the cows come home isn't going to fix that. A physicist friend who wrote his dissertation on Monte Carlo simulation likes to tell me that you cannot remove uncertainty with simulation.

      When we make assumptions about our life spans, we assume that it could be a long and expensive retirement and that's probably the best way to look at future returns — they could be low.

      Regarding a user's manual, I would express the same concern with most software products. I will say that there is an active MaxiFi forum to answer questions and a lot of answers are there among the threads. But I get your point. MaxiFi is complex and generates a lot of questions. Still, the support is orders of magnitude better than you will find for most online planners.

      Thanks for sharing your experience. Great information.

      Delete
  4. I think of the MC capability more as a nice-to-have than a necessity with MaxiFi. I used the former product, E$Planner without MC for years and I think you will find great value in the standard product. However, if the extra $40 isn't a big issue, then go with MC capability for now and renew without it next year if you don't find it valuable. (It's always fun to play with.)

    It's worth noting that MaxiFi's MC does not estimate probability of ruin, as most retirement models do. It creates future scenarios of living standards, which is a far better use of MC than estimating p(ruin), IMHO. One of my pet peeves is columns that say "Monte Carlo models do this or assume that." MC simulation is a tool that can be used in many ways and what they really mean is that a particular simulation model uses MC in a particular way. When the model performs poorly, they blame it on MC simulation rather than a poorly built model. MC does whatever the model builder wants it to do.

    Thanks for the question!

    ReplyDelete
  5. I used ESPlanner for years after reading Larry's excellent book The Coming Generational Storm. A question: what's the difference between traditional planning MC where the discretionary spending variable is reduced until I get 95% - 100% chance of success or letting maxfi calculate that level of spending? Are they not the same?

    ReplyDelete
    Replies
    1. There is no similarity in those two approaches.

      Dr. Kotlikoff explains above the three errors made by probability-of-ruin estimators. I would add that p(ruin) measures the probability of a shortfall but not its magnitude or utility. Consequently, p(ruin) counts a scenario that funds 29 of 30 years a failure and one that funds 50 years no more a success than one that funds only 30. To make matters worse, MC models that estimate p(ruin) are not stable in the sense that the results can vary dramatically by changing nothing but the random number draw. P(ruin) is, then, a very flawed measure of success.

      If I had a magic wand and could eliminate one retirement strategy, I might well get rid of p(ruin).

      This is easily seen if you run MaxiFi and your MC model that estimates p(ruin). You are likely to see far different spending recommendations.

      Thanks for the question!

      Delete