MaxiFi Planner's new Roth Conversion Optimizer is uniquely able to precisely calculate lifetime-tax minimizing conversions. Its frequent answer: Convert most/all of your IRAs!
Jack, I appreciate your comments. There are two ways economics handles uncertainty -- full bore expected utility maximization and certainty equivalent risk adjustment. The price of every financial asset is the discounted stream of its payouts, where. the discount rate is the TIPS return (term structure) with an adjustment for risk. MaxiFi does both types of planning. In its risk-adjusted/deterministic/conservative/"fiduciarily appropriate" planning, MaxiFi takes the TIPS return and max age of life as default values. It asks users to enter conservative earnings, healthcare outlays, and other assumptions. This is not hidden. We say this in many different places in our website and in my columns. Then we ask people to do stress tests. The Roth conversion case I ran was of a single person. The issue of his max age changing due to changes in health seems of second order wrt Roth conversions for such a person. If he were married it would seem of third order. Were he married with kids, it would seem of fourth order. Plus, we have some, albeit rather crude ways, of stress testing this issue. I'm sorry I played tough cop in our conversation. But I could see we were going down a rabbit hole with you claiming to be as much an expert on economics as me and then claiming economists are off base. Anyway, we've exhausted this topic and ourselves. I'll reference your concern in my future writings about Roth conversions. My best, Larry
I appreciate your comments as well Larry. I agree that we have exhausted the topic. I need to point out, however, that my last comment has been deleted. I will post it again here. Otherwise your last response has no counterpart or reference. I would only modify your response above, to point out that I never claimed to be an expert on economics, in fact the contrary. I have said, and it deserves repeating, economists are not the best guides to investment decisions, IMO of course. Here is my last response, please ensure that it doesn't get deleted. Best wishes, Jack
MY LAST RESPONSE, TO WHICH LARRY IS RESPONDING ABOVE
Larry, I am not speaking for Peter Coy, but rather citing his endorsement of the Roth conversion strategy based on your modeling software in his NYT column recently. He did criticize the policy implications of using Roth IRA rules in this respect, and I agree with him in that criticism. On the other hand, I agree with your point about not assuming what financial consultants using your software are advising their clients, and I withdraw that comment and apologize for being presumptuous in that regard.
I reiterate the point I made before about your model being impressively capable of taking into account a wide range of possibilities and preferences. I am not criticizing MaxFi, but rather the
absence of qualifiers and cautions accompanying your substack post above of the 65 year old man with $2 million in savings who does a conversion and can expect a return of $175K at age 100. I assumed that reference point throughout my comments.
Again, I never claimed to be an economist, and my approach to this discussion is as a non-academician, who read two articles that you recommended to me as being central to your approach. I think is is understandable that I read them in light of the discussion we have been engaged in, not as an expert that was likely to pick up the nuances and details of the debates on Yarri over the last 60 years. If I was overly sharp in my responses about these articles, it was because I felt that I was being led on a wild-goose chase in that these articles were not really relevant to what I see as the key issues. I think this conclusion is buttressed by the fact that only now are you addressing my key points about changing time horizons, and preferences, in a strategy that requires large up front costs, and decades to achieve compensating benefits. I think most investors (as opposed to economists) reading through this discussion will conclude that you have been dilatory in coming to grips with my key points.
Your "Fourth, Sixth and Seventh" points, are good examples of what I am referring to above. To take one example, I have never come across a financial analyst who would consider that "spending more when young and less when old" and "going to Settings and Assumptions and altering their Standard of Living Index, telling the program to spend more, if doing so doesn't violate the household's cash flow, when young and less when old" would be helpful, or even comprehensible, advice. To accuse me of unjustified criticism because I haven't addressed these opaque features in your model, and these arcane points of theory, is strange, to say the least. I think I can be excused for assuming that you are trying to intimidate and impress me with economic jargon and waving of hands.
In your "Eighth" paragraph, you finally acknowledge what I have been trying to get you to address through all of our exchanges. Indeed, considering worst case scenarios (and more realistic scenarios like I only live to 90) should be addressed. "This is the fiducially responsible thing to encourage, not that we are fiduciaries." I was a pension investment consultant, and I was very aware of my fiduciary responsibilities under ERISA. You do not have those responsibilities. Customers should note this fact.
"Ninth" and "Tenth" paragraphs: It was not necessary for me to get into a discussion of the economic sacred texts and internecine jousting among the economic cognoscenti. None of it is relevant to my arguments or to non-academicians. I defer to your expert opinions in these matters.
Your last paragraph: I agree, I should have simply said that people should run MaxiFi with alternative max ages of life because their most desirable annuitization conversion plan will depend on this. People have varying risk tolerances, preferences for how much to spend and when to spend it. Some will not agree with you that the greatest risk to hedge is the risk of living so long that they run out of money. These judgments are subjective as well as objective, and models are good ways, at least for sophisticated and informed investors, to explore these alternatives. You may disagree, but I think our debate is useful for people who are using, or thinking about using, MaxFi to explore financial decisions.
Jack, Let me respond point by point. First, Yarri's and all subsequent longevity risk papers, including Kent's with Reichling, assume a maximum age of life. Yes, Riechling and Smeeters consider one's learning that their max is no longer their assumed age 120 because they, say, learn they have pancreatic cancer. MaxiFi's max Max Age is 110. But we will, on reflection, change that to 120. Our default is 100.
Second, since every economics model incorporates a maximum age of life, MaxiFi, which is an economics-based financial planning educational tool, does as well.
Third, I had detailed conversations with Peter Coy. I don't think you should speak for him about what he did and didn't get. Nor should you speak for the financial planners using our program.
Fourth, the example I provide of a household who has a max age of 100 does incorporate the probability that he may get inoperable cancer in 6 years. As Yarri shows, economics handles the high probability of dying before your max age, either through inoperable cancer or something else, by tilting your age-consumption profile -- spending more when young and less when old. MaxiFi lets users take this calculated gamble by going to Settings and Assumptions and altering their Standard of Living Index, telling the program to spend more, if doing so doesn't violate the household's cash flow, when young and less when old. Have you looked at this setting? If not, you have engaged in lots of unjustified criticism.
Fifth, Yarri considers uncertainty over the max age of life and does so at the beginning of his article. He handles this by including utility of the household's estate. The estate proxies for having money in reserve if the max age of life turns out to be lower than the true realized max. MaxiFi lets you specify bequests in four ways a) specify that some regular assets are reserve funds making them unavailable for spending, b) (This is brand new. specify a ceiling on the amount of discretionary spending you wish to do, c) specify that a share of non-annuitized retirement account assets are to be bequeathed to the extent any are available after RMDs, and d) specify gifts as special expenses at the end of life.
Sixth, Kent is my close friend and former co-author. But I think his paper with Reichling treats Yarri as a straw man. Yarri explicitly contemplates annuities whose loads are so high that they aren't purchased. He's also looking at being able to tailor one's annuities to match one's desired age-consumption profile. His analysis can easily be extended to show that households wouldn't necessarily fully annuitize if they were forced to buy fixed nominal annuities as opposed to specially tailored real annuities. This is doubly true if future inflation is uncertain. MaxiFi's baseline assumption treats annuities as unavailable and then lets users examine the value of purchasing alternative annuities. You can enter them in our annuities screen and compare their value under different assumed future inflation rates (set up as alternative profiles).
Seventh, Reichling and Smetters are focused on changes in the future in the probability distributions of survival rates. As is apparent from Yarri's paper, if there were annuities available at time zero whose payoffs were dependent on the future resolution of those distributions, the single agent with no bequest motive would fully annuitize. Reichling and Smetters are assuming away the existence of such financial instruments. I.e., they are ruling out Arrow state-contingent longevity insurance markets. Yarri does the same when he considers the case that non-state-contingent annuities are unavailable in the context of probability distributions not being part of the future state space. This is the straw man point. Yarri is not claiming that everyone should annuitize given extant markets. He's showing that with perfect insurance markets everyone will annuitize. That's just as true whether future survival probability distributions change or not. If they do, the insurance markets will handle this situation making full insurance the right answer.
Eight, MaxiFi let's users explore future changes in beliefs. For example, suppose a 50 year old expects to work till 65 but realizes they may get fired at 60. Well, you can run MaxiFi under both assumptions. We encourage users to assume worst-case scenarios. This is the fiducially responsible thing to encourage, not that we are fiduciaries. Could we formally incorporate the risk of being fired each year and then do maximum expected utility maximization? Yes, but considering this and all manner of uncertainties makes the problem intractable. This is why we are taking the most conservative assumptions as our base case and letting users explore other outcomes manually. This includes having a maximum age of life that, in six years, you learn is not 40 years, but two years. If one runs Roth conversions assuming an 8 year horizon, the calculated optimal conversion plan will surely be not to convert.
Ninth, You wrote, This is directly addressed by Reichling and Smetters. They show that unexpected illness and increased uninsured medical and other costs (not to mention the change in investment horizon) appropriately change anyone's time horizon and risk tolerance. If someone is already committed to a strategy like a Roth Conversion, they do not have the ability to reflect that change in their time horizon because the costs are sunk and unrecoverable at that point. That is why they avoid annuities and hedge their bets on life expectancy!
This is wrong on two counts. Learning this information doesn't, in R&S, change their risk tolerance. Nor does their analysis connect to the term "life expectancy." It does connect to contingencies that can be explored, if not perfectly, by setting up alternative profiles with different maximum ages of life. But that's true of all kinds of contingencies. If I plan to work to 65, but can get fired at 60, I will spend through 60 as if I was able to work through 65. This will be a sunk cost. MaxiFi would encourage someone who could be fired at 60 to run their plan under that conservative assumption as well as assuming retirement at 65. In the context of longevity risk, the conservative assumption is living to your max age of life. Yes, if you learn at 60 that you have pancreatic cancer, you won't be able to keep working and won't be able to benefit to the same degree as would otherwise have been the case from Roth conversions. But the bigger longevity risk is not leaving money on the table, but running out because you live too long.
Tenth, Economists, like you and Yarri, have been telling them that they were irrational! Yarri and a whole generation of economists after him saw this lack of interest in annuities as irrational, or a puzzle needing further research. Well, people turned out to be right, and economists turned out to be wrong.
Sorry, but show me where Yarri or I or any decent economist has been saying this. (It's certainly not in Yarri's paper, which you claim to have read.) This is a figment of your imagination. Kotlikoff and Spivak (1981) provides one of many solid explanations for the observed lack of annuitization. You are just making up a straw man to say economists are wrong. Not the case. Economists have spent a lot of time explaining why people don't purchase annuities. R&S is a good example. They add to a long list of reasons -- reasons they list, all of which economists pointed out.
I've given you as much of my time on this, Jack, as I can. So, I'm not going to respond to any more of your comments. You could simply have written that people should run MaxiFi with alternative max ages of life because their optimal annuitization conversion plan will depend on this. But, in the end, I'd still focus attention on the max max age of life, which is the truly catastrophic financial scenario, not having paid taxes early, which will likely mean your kids will pay less.
Jack, I suggest you build your own software program and market it. I'm following what economics says. Extending Yarri to consider annuities with loads leads people not to purchase annuities or to purchase fewer annuities. This wasn't ignored by Yarri. His central case of interest is that of people buying no annuities. As for people's maximum ages changing, you can run MaxiFi for different assumed maximum ages of life, weigh the results based on the probability of those maximums, and now we're back forming expected utility. You can also run cases with future tax hikes and future tax cuts. This is why we have alternative profiles that you can run. Larry PS, the rational failure to buy annuities does not, as I say, matter to the issue of maximizing lifetime utility through the max age of life. So, I stand by that point wrt to all the papers that don't consider max age of life uncertainty. That concern can be dealt with in our tool as just described. But one wants to plan cautiously. If it turns out you get pancreatic cancer at 90 and can't enjoy your tax savings thereafter, you'll be in heaven. But if you run out of money at 90 because you said, no way, I can't make it to 100, you'll be in trouble. Larry
Larry, No need for another program, just more cautionary notes on risk tolerance, variable discounting on FV, and time horizon variability. Best wishes, Jack
Jack, Good you read Yarri. Yarri's surprising point is not about pooling the risk of longevity. This goes back to Tontine in the 1600s. It's the simple point that the max age of life is the appropriate planning horizon and that focusing on earlier dates of death never arises. None of the papers you mention matter to the point you keep making and keep getting wrong. In all papers with uncertain lifetimes, the agent maximizes lifetime utility through the maximum age of life. MaxiFi's focus on maximum age of life, not expected age of life is consonant with this key point. And yes, some households may be altruistic and enjoy higher expected utility from conversions if they provide higher spending power for their kids. But knowing if that's true, requires running the kids through MaxiFi in parallel. Whether conversions are a good or bad thing wrt leaving or gifting money to kids requires knowing the kids' tax situations. So, if that's your concern, purchase a license for your kids and run them with alternative assumptions about your date of death. Yarri wasn't disproved by the subsequent papers. Those papers just introduced more realistic assumptions than Yarri made, including the possibility that annuities are available, but aren't actuarially fair. Larry
Larry, you say that "None of the papers you mention matter to the point you keep making and keep getting wrong. In all papers with uncertain lifetimes, the agent maximizes lifetime utility through the maximum age of life." That is misleading in the way that you are interpreting it. Allow me to illustrate why. Putting this into the real world context of your Roth IRA conversion model and your advice to people to use a maximum life assumption of 100 years, you are either implicitly or explicitly advising people to use their maximum life expectancy as a time horizon for decision making despite your claim that you don't give advice. (Peter Coy didn't get that impression, and neither did the financial planners that use your software.) Your example in this substack, in which your retired person spreads out conversions over 5-6 years, gives a completely different result if he finds out in year 6 that he has inoperable cancer and has a life expectancy of 3-5 years. This is directly addressed by Reichling and Smetters. They show that unexpected illness and increased uninsured medical and other costs (not to mention the change in investment horizon) appropriately change anyone's time horizon and risk tolerance. If someone is already committed to a strategy like a Roth Conversion, they do not have the ability to reflect that change in their time horizon because the costs are sunk and unrecoverable at that point. That is why they avoid annuities and hedge their bets on life expectancy! Economists, like you and Yarri, have been telling them that they were irrational! Yarri and a whole generation of economists after him saw this lack of interest in annuities as irrational, or a puzzle needing further research. Well, people turned out to be right, and economists turned out to be wrong. Maybe people just got lucky when they ignored economic advice, or maybe their intuition and common sense proved to be better than that of academicians. I don't know the answer to that, but it seems that you have not recognized that mistake.
I urge any reader of our debate to read these papers themselves. It involves a slog through a lot of economic jargon but I think many people reading these papers will realize that Larry is not addressing important aspects of financial decision making. Larry says that he is not giving advice, but he has been, and is, touting his modeling software to NYT columnists, individuals, and financial advisors because it reveals several strategies that provide large financial benefits, most saliently delaying social security payments and front loading Trad.IRA's withdrawals into Roth IRA's. These strategies are based on facts and they are well analyzed by Larry's software, but they involve assumptions about the future. Those assumptions could be wrong. I am as competent to give an informed opinion on these strategies as Larry, Peter Coy, and all of the financial advisors using his software are. My informed opinion is to diversify, hedge your bets when you reasonably can, and heavily discount assumed returns 30 years in the future, whether they derive from dividends, interest, stock prices or lower taxes.
Jack, You are apparently unable to follow the math in Yarri's paper or you would understand what I wrote. You could just admit that. Let me explain the "weird" point this way. If I'm buying homeowners insure, I buy full coverage to protect myself against the worst case scenario - the house burying down. I don't buy coverage for the expected loss, but for the max loss. And, in doing so, I'm also buying coverage for all smaller losses. Same with lifespan. In setting a spending plan for the worst case scenario -- living to my max age of life, I am also setting a spending plan for all preceding years. This will be my last reply to you. You don't have the background to criticize economics-based financial planning. That's what MaxiFi delivers. If we modified MaxiFi along the lines you suggest, it would no longer comprise economics-based planning. best, Larry
OK, I have now read Yarri. I am puzzled by your emphasis on the math because it is actually straight forward, as is the principal argument that pooled annuities create a surplus value from those who die before the annuity values they receive match the cost of the annuity provider, which earns a profit and presumably passes some of that profit to annuitants with longer lifetimes and lower income in old age. The math, simple or complex, does not settle the question, however. It is generally recognized now that Yarri’s assumptions are unrealistic and fail to capture important real world impacts on these decisions. See Uncertain Lifetime, The Theory of the Consumer, and the Life Cycle Hypothesis, Siu Fai Leung, 1994
I believe it is noteworthy that in two posts now you have drawn our attention to early studies ((Merton 1969 and Yarri 1965) that have been largely modified, even supplanted, by more recent studies that significantly revised, even contradicted, that earlier work. I’m sure that you are familiar with Optimal Annuitization with Stochastic Mortality and Correlated Medical Costs, Felix Reichling and Kent Smetters, 2015 (Spoiler alert: I know that you were consulted on this work and it was supported by BU). This paper throughly revises Yarri’s conclusions, and provides very different conclusions about time horizons and evolving life cycle utility. They found “that including asset management fees and bequest motives….at least, nine out of ten households (should) not hold any annuities.” Also, “depending on our calibration assumptions, between 64% and 76% of households should not annuitize any wealth, even with no transaction costs, bequest motives, ad-hoc liquidity constraints, or asymmetric information problems. In contrast, the Yaari model predicts 0%.” By “should not annuitizing wealth”, Reichling and Smetters explain that the incidence of sickness and shortened life expectancies are, at this point at least, largely stochastic and greatly influence time horizons and optimal investing strategies of investors in dynamic and rational ways. In short, one time horizon (maximum life) and therefore optimal strategy does not exist for all people, even if they have the same risk tolerance. If these conclusions somehow agree with your take on investing for maximum possible age, I think you have substantial explaining to do, having pointed to an outdated and simplistic analysis that has been shown to be misleading, and then having ridiculed our inability to understand these mathematic “truths.” Forgive me if my perception of your end of the dialogue up to this point is one of an intellectual bully trying to awe the bothersome rubes. Best, Jack
Jack, Pls read Yarri, Uncertain Lifetimes, RESTUD, 1965. You claim to be an economist. Let's see if you can follow this fundamental paper. It shows that the parents should value their net resources through their max age of life. As for children, they can be run in their own MaxiFi program based on different assumptions about when their parents will die. But, just like their parents, they need to plan to live to their maximum age of life. If the parents care about the expected utility of their children, the gains to the parents from conversions may or may not be larger. If the kids are in a lower tax bracket than the parents, conversion followed by early death may lead to lower lifetime spending capacity of the kids than absent conversion.
Larry, I'm disappointed to see you return to the ad hominem questioning of my fitness as a critic. I am a Chartered Financial Analyst, who was able to retire at the age of 44, after following a strategy of super saving and standard index investing. I am generally well-read in economics, as I am in a wide variety of disciplines, as you have conceded, and most importantly you have not been able to counter my arguments convincingly. I have never claimed to be an economist, and I actually regard that as a strength in these debates because you appear to be a person who is so dazzled by your (rather eccentric) interpretation of economic theory that you ignore common sense and obvious facts. Your appeal to "it's weird but that's economics" has been almost completely lacking in specifics. You contemptuously dismiss any criticisms of your advice here as uninformed and incompetent, but you fail to link hopelessly vague claims like "parents should value their net resources through their max age of life" to the real choices facing parents with respect to tax and investmtent decisions. Bottom line: it is rational for a parent to look at your model's output that shows (in one example) that someone can covert his Trad. IRA to a Roth at age 65, pay $350K in taxes up front, and eventually make $175K PV in profit by doing so if he lives to be 100. If he lives to 95, he makes 25% less, but still $125K or whatever, if he lives to his average life expectancy, the ROR is rather meager, but economic theory tells us "don't worry!" this is all just life insurance and every other consideration is irrelevant because maximum age of life is supposedly a time-horizon decreed by the theorists and their simulations. In light of your recommendations, my conclusion is that you are lost in an academic fog. You have a stronger case to make with Social Security strategies and your modeling of those. Stick with that and backoff this absolutist pitch for Roth conversions for everyone who shows a half-decent increase in FV from your model. For very wealthy people looking for tax savings, it can be attractive if one has faith in the stability of public policy. For others, it can be reasonable to avoid it.
Let me explain it this way. Suppose you will either live for 10 years or live for 20 years. You have to plan for living for 20 years and the consumption you do under that plan for the first 10 years is what the live-for-20-years plan dictates. And the tax moves you make are based on living to 20 years. If you die in 10 years and don't get to reap the benefit of being able to spend more in the next 10 years, well, you're in heaven, so it doesn't matter. Yes, this is surprising and slightly weird, but it's what economics says. best, Larry
Hi Benjamin, Yes, lowering your max age will lower your conversion gain. But your max age is what it is. You can't count on dying before it. You can't play the averages. This exclusive focus on the worst-case scenario -- living to your max age -- is surprising, but it's what economic theory says to do. best, Larry
Jack, MaxiFi has a standard of living index that let's you effectively include time preference. Frankly, I don't think it will matter much to the gains from Roth conversions.
For those readers who are not familiar with the concepts, here are explanations:
To calculate expected lifetime utility, you would sum the discounted utility of consumption in each period of life, taking into account the probability of reaching each period.
Similarly, when evaluating investment options, the concept of FV discounting is used to compare the present value of future cash flows, allowing for a more accurate comparison between different investment choices with different payout timelines.
To calculate expected lifetime utility, economists use a "discount factor" which represents the value placed on future utility compared to present utility, effectively discounting the value of future consumption, and reflecting the uncertainties of future consumption.
The discount factor reflects an individual's "time preference," which is their tendency to prefer immediate consumption over future consumption, even if the future consumption is larger. The discount factor also should reflect the uncertainty if achieving a FV.
Hi Jack, I appreciate your reply, but still don't get your discounting point. Let's take this off line. Pls email me at kotlikoff@gmail.com and we'll find a time to chat on the phone. best, Larry
My point about TIPS rate discounting of FV is straightforward. I recognize that it makes sense from your point of view to use the TIP rate because you regard the strategy as very low risk, but I think it entails substantial risk, and therefore requires a higher discount rate during the 30-40 year wait for FV. In other words, the benefits of the conversion strategy come very far in the future, and should be discounted not only for lost income, but also for uncertainty in the outcome. As for exploring this further with you, I think you would agree with me that we have pretty thoroughly covered (exhausted) the topic. It's been fun, honestly. Best wishes, Jack
Jack, As you know, one can't characterize all the risks that we face and do full expected lifetime utility maximization over all such risks. But one can do scenario analysis -- stress testing if you like. For example, you can consider your apparent fear that your Roth account will be taxed upon withdraw at some point down the road. You can just run MaxiFi treating the extra future taxes as special non-tax-related special expenses. MaxiFi is an educational tool. It's not giving advice. It's here for you and others to explore different scenarios. An important scenario, arguably the most likely, is that current tax law (TCJA) will prevail. But you can run MaxiFi's Roth Conversion Optimizer under different assumptions about future tax increases or tax cuts and, as indicated, adding additional future taxes as special expenses. If you can find a better program for exploring Roth conversions, use it. I know none exists. None of the other programs produces a single result that is internally consistent. Given your fears, you'd presumably advise everyone to take Social Security as early as possible because anyone who waits will be viewed as rich enough to be fleeced. Again, you can run our Social Security optimization specifying future benefit cuts to explore this. In short, I don't know what you want. You are clearly using MaxiFi to help guide your planning. If you'd like some other features, let me know. But I've tried to give you what I think you and others, including myself, need -- namely a tool that solves all the key equations of economics-based personal finance. Best, Larry
You're a real mench Larry, and crazier than I am to be up this late--I'm two hours earlier in Mtn. time zone. I agree that your model is a useful planning tool. You have created a marvelous, sophisticated model, and I believe it has no match in other PC-based models. . However, you mischaracterize my legitimate criticisms of your discount rate for backloaded tax benefits--these are not fear-based, and should not be characterized as a risk-aversion anomaly on my part. Moreover, as the next four years unfold, I reasonably believe that prior norms of risk assessment will be significantly reassessed . Take care, Jack
⚖️Hi Larry, My Underlining Didn't Copy To Substack. I'll Capitalize And Skip A Line And Add A Scale Of Justice Emoji To My Responses.. Hopefully That Will Make Easier To Follow, But I Hope Reading Capitalized Text Is Easier For Others Than It Is For Me!
HI JACK, I REPLY IN BOLD.
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen").
OUR MAXIFI SOFTWARE HANDLES CASH FLOW CONCERNS. YOU HAVE RUN THE ROTH CONVERTER OF LATE, YES? SO I DON'T UNDERSTAND YOUR SUGGESTIONS THAT WE ARE IGNORING THAT MAJOR ISSUE. WE AREN'T. I AM NOT
⚖️ I Understand That Maxfi Deals With Cash Flow And Monte Carlo Simulations. That Is All Fine And Good, And The Complexity And Detail Of Your Model Is Truly Impressive. My Point Is That The Uninitiated Consumer Is Likely To Misinterpret The Narrative That The Model Can Maximize Lifetime Spending/Utility. It Can Do That If The Inputs And Simulations Capture Future Market And Regulatory Events. You And I Know That But I’m Hoping To Give Some Cautionary Information To People Who Might Not Understand That, Or Understand It Fully.
All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design.
MAXIFI'S FULL RISK MONTE CARLO SIMULATION FULLY INCORPORATES CASH FLOW CONSTRAINTS ALONG EACH LIVING STANDARD TRAJECTORY IN DETERMINING INVESTMENT STRATEGIES THAT MAXIMIZE LIFETIME UTILITY. SO, AGAIN, I THINK YOU ARE CRITICIZING OUR TOOL WITHOUT HAVING RUN IT OR UNDERSTOOD IT.
⚖️ As I stated above, I am not criticizing your tool, but the marketing of it.
I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does.
LOVE TO HEAR FROM GENE. NOTHING I'M SAYING OR THAT MAXIFI IS DOING IS ANYTHING BUT FULLY IN TUNE WITH MODERN FINANCE.
⚖️ Ditto. The Question IS Not About Modern Finance, But The Risk/Reward Profile Of The Roth Conversion Assessment.
Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions.
THIS IS, FRANKLY, CONTENT FREE.
⚖️ OK, My Point Is That The Merton/Samuelson 1969 Paper, Which You Cited As Ruling Authority, Considered Time Horizon To Retirement To Be Irrelevant To Portfolio Construction. They Changed That Appraisal In The 90's, And The Current Consensus Is That Time Horizon Is A Critical Variable In Portfolio Construction, Which You Failed to Acknowledge In The Comment That I Was Responding To.
Although Merton And Samuelson In 1969 Suggested That Time Horizon Is Irrelevant For Asset Allocation, Merton And Samuelson In 1992 Argued That Workers Who Are Sure Of Future Employment Should Increase Their Investment In Stocks, While For Retired People It Would Not Be Optimal To Do The Same, Indicating That They Believed Time Horizon Is Relevant To Portfolio Allocation Decisions.
⚖️ In Important Respects They Are Inconsistent. The 1969 Paper Did Not Recognize That With Time Dependent Investment Opportunities, Retirement And Death Are Key To Determining Investment Horizon And Portfolio Construction. Remember That I Am Responding To Your Previous Claim That Time Horizon Is Irrelevant To Investment Decisions. You Partly Acknowledge That Below, But Still Fail To Address That Much Research Has Shown That Risk Tolerance Decreases With An Increase In Undiversifiable Risk, Which Is Likely To Happen, And Therefore Time Horizons Shorten. In Your View This Is Largely Irrelevant To Portfolio Decisions, But I Think Most (Non Academicians) Disagree.
Other distinguished financial economists, like burt malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
YES, THE TIME HORIZON IS IMPORTANT TO THE PORTFOLIO CHOICE, BUT PRIMARILY BECAUSE OF THE COMPOSITION OF ONE'S INCOME FLOWS -- FIRST LABOR EARNINGS AND THEN SOCIAL SECURITY (ANOTHER BOND).
⚖️ Note That The Composition/Amount Of Income Flows Is Relevant To My Point When I Focus Below On Changing Exogenous Factors.
Your advise to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer.
OF COURSE. EXPECTED UTILITY MAXIMIZATION ENTAILS ALWAYS REOPTIMIZING BASED ON THE LATEST INFORMATION.
⚖️ And My Point Is That The Efficacy Of Reoptimizing Is Significantly Constrained By The Upfront Costs Of A Roth Conversion.
OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established.
THERE IS NO FIRST-ORDER EXTERNALITY ISSUE INVOLVED IN ROTH OPTIMIZATION OR PORTFOLIO CHOICE FOR AN INDIVIDUAL HOUSEHOLD. SO, I HAVE NO IDEA TO WHAT YOU REFER.
⚖️ I Am Using “Externality” Metaphorically Here. There Are Many First And Second Order Economic Externalities Impacting All Of Us These Days, But I Am Referring To Rapidly Rising Political, Economic, Etc. Risks That Would Affect Your Model’s Optimization.
Many Of These Risks Will Be Difficult Or Impossible To Hedge. Your Model Assumes General Stability While Changing Conditions Are Supposedly Accommodated By Yearly Reevaluations.
YES, WE COULD HAVE NUCLEAR WAR AND MAXIFI DOESN'T CONSIDER ITS LIKELIHOOD IN DECIDING WHETHER TO DO ROTH CONVERSIONS. IF THIS IS YOUR OBJECTION, I THINK YOU SHOULD ASK YOURSELF WHY YOU WOULD WANT TO DO ANYTHING EXCEPT PUT A BLANKET OVER YOUR HEAD AND SCREAM.
⚖️ We Agree That Nuclear War Is Not A Diversifiable Risk, Nor Can It Be Hedged By Investment Decisions. Your Argument Doesn’t Apply, However, To Increased Insecurity Of Income And Expenses In All Time-Horizons, Coming From Political Instability, Trade Wars, Climate Change, Etc. These Risks Could Plausibly Reduce One’s Willingness To Bet On A Strategy That Involves A 35-Year Time Horizon.
I think that is an assumption people should not make. Well, you may reasonably reply, these (rising risks) “externalities” impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do.
LEAVING TAX BREAKS ON THE TABLE IS NOT SUPERCHARGING ANYONE'S RISK. PLANNING FOR PEOPLE TO LIVE TO THER MAX AGE OF LIFE IS THE RIGHT APPROACH TO LONGEVITY RISK. READ YARRI 1969.
⚖️ Again, These Risks Could Reasonably Reduce One’s Willingness To Bet On A Strategy That Involves A 35-Year Time Horizon And Stable Govt. Policy.
You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not.
YES, ECONOMICS SAYS THAT PEOPLE SHOULD CAUTIOUSLY SPEND OUT OF THEIR FUTURE RESOURCES, CAUTIOUSLY ESTIMATED. THAT'S WHAT BOTH OUR DETERMINISTIC AS WELL AS STOCHASTIC PLANNING REPORTS DO.
⚖️ I Agree, The Disagreement Is In What Constitutes A Cautious Approach To Increasing Risk.
My Argument, In Part, Is That Those Expectations Of Future Benefits Are Less Likely To Be Realized Than They Were A Few Years Ago, And Will Be Less Likely Still Five Years Hence. If I Were To Execute Such A Strategy Today, 5-6 Years From Now I Would Have No Way To Change My Strategy If My Assumptions Have Changed.
WHEN IT COMES TO ROTH CONVERSIONS, THE COUNTRY'S FISCAL INSOLVENCY POINTS TO TAX HIKES DOWN THE ROAD. THAT WOULD MAKE CONVERSIONS EVEN MORE ADVANTAGEOUS. USERS CAN ADD THAT TO THEIR PLANS UNDER SETTINGS AND ASSUMPTIONS. BUT OUR DEFAULT DOESN'T INCLUDE THIS HAPPENING TO MAKE THE CONSERVATIVE ASSUMPTION. THE WORRY YOU ARE REFERENCING IS, FRANKLY, RIDICULOUS GIVEN OUR FISCAL GAP, OF WHICH I'VE WRITTEN AS MUCH AS ANYONE. YOU ARE AFRAID TO DO A ROTH CONVERSION BECAUSE YOU ARE WORRIED THAT TAXES WILL FALL IN THE FUTURE. WELL, MAYBE YOU SHOULD RETHINK THAT POSITION.
⚖️ You’re Cherry Picking Your Scenarios. I Never Said That I Think Tax Rates Will Fall—That Is A Straw Man. The Threat Of Insolvency Is Just As Likely To Lead To Taxation Of Roth Tax Havens For The Wealthy (See Peter Coy’s Comments On That Subject), Or Some Other Rollback Of Those Conversion Benefits. I Don’t Think That Would Be Ridiculous Or Even Particularly Unlikely. I Doubt That Most Of Your Customers Understand The Number Of Contingencies Involved In This Strategy, And The Limitation On Their Choices If Prior Assumptions Turn Out To Be Wrong.
SEE ABOVE.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
JACK, GOOD TO DISCUSS. BUT YOU ARE BEING NIHILISTIC. I APPRECIATE THAT YOU KNOW MORE REAL FINANCE THAN I THOUGHT, BUT YOU ARE SUGGESTING THINGS THAT REAL FINANCE DOESN'T COUNTENANCE, NAMELY PUTTING YOUR HEAD IN THE SAND BECAUSE THE FUTURE IS UNCERTAIN. THE FUTURE HAS ALWAYS BEEN UNCERTAIN AND IT SOMEONE FINDS A WAY TO LOWER THEIR FUTURE TAXES WITH VERY HIGH PROBABILITY, THEY SHOULD RATIONALLY SPEND MORE TODAY UNLESS THEY ARE EXTREMELY RISK AVERSE. I DON'T BELIEVE YOU ARE EITHER IRRATIONAL OR EXTREMELY RISK AVERSE. HENCE, I SUGGEST YOU TAKE MAXIFI'S ROTH SUGGESTIONS SERIOUSLY. BEST, LARRY
⚖️ I Am Guilty Of Using “Externality” Metaphorically Instead Of Maybe Exogenous, You Are Guilty Of Using Nihilistic When You Should Say Pessimistic. I Cop To That, And Polls Show That Is A Majority Position These Days. I Likewise Appreciate That You Know More Real Finance Than I Thought, But You Are Egregiously Misrepresenting My Points And Baselessly Characterizing Them As Inconsistent With "Real Finance". Likewise, I Think It Is You Who Are Putting Your Head In The Sand To Ignore That The Future Is Becoming Much More Uncertain In Many Ways; To Expect That A 35-Year Payoff Of $175K Pv After A $300K Up Front Bet Is Questionable, And To Propose That A Discount Rate Set By The Tips Rate Is Sufficient For Future Benefits. You Said That The Roth Conversion Model Produces A “Very High Probability” Of Lowering Taxes. My Main Point Is That Those Probabilities Are Subject To Many Assumptions And Risks. Buyer Beware. Best Wishes, Jack
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen"). OUR MAXIFI SOFTWARE HANDLES CASH FLOW CONCERNS. YOU HAVE RUN THE ROTH CONVERTER OF LATE, YES? SO I DON'T UNDERSTAND YOUR SUGGESTIONS THAT WE ARE IGNORING THAT MAJOR ISSUE. WE AREN'T.
All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design.
MAXIFI'S FULL RISK MONTE CARLO SIMULATION FULLY INCORPORATES CASH FLOW CONSTRAINTS ALONG EACH LIVING STANDARD TRAJECTORY IN DETERMINING INVESTMENT STRATEGIES THAT MAXIMIZE LIFETIME UTILITY. SO, AGAIN, I THINK YOU ARE CRITICIZING OUR TOOL WITHOUT HAVING RUN IT OR UNDERSTOOD IT.
I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does.
LOVE TO HEAR FROM GENE. NOTHING I'M SAYING OR THAT MAXIFI IS DOING IS ANYTHING BUT FULLY IN TUNE WITH MODERN FINANCE.
Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions.
THIS IS, FRANKLY, CONTENT FREE.
Although Merton and Samuelson in 1969 suggested that time horizon is irrelevant for asset allocation, Merton and Samuelson in 1992 argued that workers who are sure of future employment should increase their investment in stocks, while for retired people it would not be optimal to do the same, indicating that they believed time horizon is relevant to portfolio allocation decisions.
Other distinguished financial economists, like Burt Malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
YES, THE TIME HORIZON IS IMPORTANT TO THE PORTFOLIO CHOICE, BUT PRIMARILY BECAUSE OF THE COMPOSITION OF ONE'S INCOME FLOWS -- FIRST LABOR EARNINGS AND THEN SOCIAL SECURITY (ANOTHER BOND).
Your advise to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer.
OF COURSE. EXPECTED UTILITY MAXIMIZATION ENTAILS ALWAYS REOPTIMIZING BASED ON THE LATEST INFORMATION.
OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established.
THERE IS NO FIRST-ORDER EXTERNALITY ISSUE INVOLVED IN ROTH OPTIMIZATION OR PORTFOLIO CHOICE FOR AN INDIVIDUAL HOUSEHOLD. SO, I HAVE NO IDEA TO WHAT YOU REFER.
Many of these risks will be difficult or impossible to hedge. Your model assumes general stability while changing conditions are supposedly accommodated by yearly reevaluations.
YES, WE COULD HAVE NUCLEAR WAR AND MAXIFI DOESN'T CONSIDER ITS LIKELIHOOD IN DECIDING WHETHER TO DO ROTH CONVERSIONS. IF THIS IS YOUR OBJECTION, I THINK YOU SHOULD ASK YOURSELF WHY YOU WOULD WANT TO DO ANYTHING EXCEPT PUT A BLANKET OVER YOUR HEAD AND SCREAM.
I think that is an assumption people should not make. Well, you may reasonably reply, these externalities impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do.
LEAVING TAX BREAKS ON THE TABLE IS NOT SUPERCHARGING ANYONE'S RISK. PLANNING FOR PEOPLE TO LIVE TO THER MAX AGE OF LIFE IS THE RIGHT APPROACH TO LONGEVITY RISK. READ YARRI 1969.
You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not.
YES, ECONOMICS SAYS THAT PEOPLE SHOULD CAUTIOUSLY SPEND OUT OF THEIR FUTURE RESOURCES, CAUTIOUSLY ESTIMATED. THAT'S WHAT BOTH OUR DETERMINISTIC AS WELL AS STOCHASTIC PLANNING REPORTS DO.
My argument, in part, is that those expectations of future benefits are less likely to be realized than they were a few years ago, and will be less likely still five years hence. If I were to execute such a strategy today, 5-6 years from now I would have no way to change my strategy if my assumptions have changed.
WHEN IT COMES TO ROTH CONVERSIONS, THE COUNTRY'S FISCAL INSOLVENCY POINTS TO TAX HIKES DOWN THE ROAD. THAT WOULD MAKE CONVERSIONS EVEN MORE ADVANTAGEOUS. USERS CAN ADD THAT TO THEIR PLANS UNDER SETTINGS AND ASSUMPTIONS. BUT OUR DEFAULT DOESN'T INCLUDE THIS HAPPENING TO MAKE THE CONSERVATIVE ASSUMPTION. THE WORRY YOU ARE REFERENCING IS, FRANKLY, RIDICULOUS GIVEN OUR FISCAL GAP, OF WHICH I'VE WRITTEN AS MUCH AS ANYONE. YOU ARE AFRAID TO DO A ROTH CONVERSION BECAUSE YOU ARE WORRIED THAT TAXES WILL FALL IN THE FUTURE. WELL, MAYBE YOU SHOULD RETHINK THAT POSITION.
I doubt that most of your customers understand the number of contingencies involved in this strategy, and the limitation on their choices if prior assumptions turn out to be wrong.
SEE ABOVE.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
JACK, GOOD TO DISCUSS. BUT YOU ARE BEING NIHILISTIC. I APPRECIATE THAT YOU KNOW MORE REAL FINANCE THAN I THOUGHT, BUT YOU ARE SUGGESTING THINGS THAT REAL FINANCE DOESN'T COUNTENANCE, NAMELY PUTTING YOUR HEAD IN THE SAND BECAUSE THE FUTURE IS UNCERTAIN. THE FUTURE HAS ALWAYS BEEN UNCERTAIN AND IT SOMEONE FINDS A WAY TO LOWER THEIR FUTURE TAXES WITH VERY HIGH PROBABILITY, THEY SHOULD RATIONALLY SPEND MORE TODAY UNLESS THEY ARE EXTREMELY RISK AVERSE. I DON'T BELIEVE YOU ARE EITHER IRRATIONAL OR EXTREMELY RISK AVERSE. HENCE, I SUGGEST YOU TAKE MAXIFI'S ROTH SUGGESTIONS SERIOUSLY. BEST, LARRY
Larry, I tried to keep our conversation in one post, but it was too long, and Substack made me break it up. Somehow it got moved down below your response to Gerald. Look for it there.
Hi Jack, Sorry for the tough love. But you have to study economic theory before you can understand it. Have you done so? Do you understand the economics of uncertainty as laid out by John von Neumann (co-developer of the A bomb) and Oscar Morgenstern? This is taught in every PhD program in economics in year 1. It's the theory that underlies the classic 1969 results of Merton and Samuelson that household's optimal portfolios are independent of their age (ignoring cash flow considerations). So, no, when we invest in stocks, we don't fundamentally play the averages. We concern ourself with the entire distribution of returns, not just the mean. Gene teaches and respects expected lifetime utility maximization as much as any economist. Just call him and ask. So, I don't know what you mean when you say he recommends investing for long-term horizons. That sounds like he believes that stocks are safe in the long run. Call and ask. He'll refute that in a minute. Or check how much it costs to insure that the S&P will beat TIPS over horizon X. It costs most the larger is X. Hence, the markets are directly refuting this proposition. And, no, life expectancy is not critical, indeed has no impact whatsoever, on one's investment strategy. Read Merton and Samuelson's independent 1969 articles on his. Re MaxiFi's Roth Conversion Optimizer, it takes explicit account of the degree to which households are cash-flow constrained. So, I have to object to your statement on this as well. Finally, re fire insurance, everyone buys catastrophic insurance even if they are willing, due to loads, to self insure small losses. In the lifespan context, this means planning for the catastrophic event, namely that you live to your max age of life. As for your statement -- paying $300 in current taxes for a chance at more spending late in life -- is missing the point of MaxiFi. When it shows your lifetime discretionary spending goes up, it's saying that your lifetime tax savings are larger, potentially far larger than the current $300K tax cost. It's also saying that you can have a higher living standard EVERY YEAR -- starting this year --- unless you are cash-flow constrained. And the program shows you whether such constraints will arise and also, as I mentioned, lets you optimize subject to not producing cash-flow problems. This may mean not doing any conversions. But in your case, you don't sound cash-flow constraints. If I've got you right, not doing conversions that lower your lifetime taxes is paying extra taxes for no reason. It's throwing away alpha based on a misunderstanding of economics and even your own financial situation.
All this said, some people have particular habit-preferences, not considered by Merton and Samuelson. They are extremely concerned about downside risk, but still want to experience upside to their living standard floor. That's why we developed MaxiFi's Upside Investing -- to help those types of households. It entails long-term stock investing that is not exposed to sequence of return risk, delivering only increases in one's sustainable living standard. Gene might have said whatever you are thinking he said in the context of such households or, indeed, foundations, universities, endowments, etc.
Glad we had this exchange. I'm a tough bird. Don't take this badly. I wish I had your investing acumen.
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen"). All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design. I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does. Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions. Although Merton and Samuelson in 1969 suggested that time horizon is irrelevant for asset allocation, Merton and Samuelson in 1992 argued that workers who are sure of future employment should increase their investment in stocks, while for retired people it would not be optimal to do the same, indicating that they believed time horizon is relevant to portfolio allocation decisions. Other distinguished financial economists, like Burt Malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
Your advice to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer. OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established. Many of these risks will be difficult or impossible to hedge. Your model assumes general stability while changing conditions are supposedly accommodated by yearly reevaluations. I think that is an assumption people should not make. Well, you may reasonably reply, these externalities impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do. You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not. My argument, in part, is that those expectations of future benefits are less likely to be realized than they were a few years ago, and will be less likely still five years hence. If I were to execute such a strategy today, 5-6 years from now I would have no way to change my strategy if my assumptions have changed. I doubt that most of your customers understand the number of contingencies involved in this strategy, and the limitation on their choices if prior assumptions turn out to be wrong.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
Robb, It most certainly does. Happy to discuss. Cell is 617 834-2148. best, Larry
Jack, I appreciate your comments. There are two ways economics handles uncertainty -- full bore expected utility maximization and certainty equivalent risk adjustment. The price of every financial asset is the discounted stream of its payouts, where. the discount rate is the TIPS return (term structure) with an adjustment for risk. MaxiFi does both types of planning. In its risk-adjusted/deterministic/conservative/"fiduciarily appropriate" planning, MaxiFi takes the TIPS return and max age of life as default values. It asks users to enter conservative earnings, healthcare outlays, and other assumptions. This is not hidden. We say this in many different places in our website and in my columns. Then we ask people to do stress tests. The Roth conversion case I ran was of a single person. The issue of his max age changing due to changes in health seems of second order wrt Roth conversions for such a person. If he were married it would seem of third order. Were he married with kids, it would seem of fourth order. Plus, we have some, albeit rather crude ways, of stress testing this issue. I'm sorry I played tough cop in our conversation. But I could see we were going down a rabbit hole with you claiming to be as much an expert on economics as me and then claiming economists are off base. Anyway, we've exhausted this topic and ourselves. I'll reference your concern in my future writings about Roth conversions. My best, Larry
I appreciate your comments as well Larry. I agree that we have exhausted the topic. I need to point out, however, that my last comment has been deleted. I will post it again here. Otherwise your last response has no counterpart or reference. I would only modify your response above, to point out that I never claimed to be an expert on economics, in fact the contrary. I have said, and it deserves repeating, economists are not the best guides to investment decisions, IMO of course. Here is my last response, please ensure that it doesn't get deleted. Best wishes, Jack
MY LAST RESPONSE, TO WHICH LARRY IS RESPONDING ABOVE
Larry, I am not speaking for Peter Coy, but rather citing his endorsement of the Roth conversion strategy based on your modeling software in his NYT column recently. He did criticize the policy implications of using Roth IRA rules in this respect, and I agree with him in that criticism. On the other hand, I agree with your point about not assuming what financial consultants using your software are advising their clients, and I withdraw that comment and apologize for being presumptuous in that regard.
I reiterate the point I made before about your model being impressively capable of taking into account a wide range of possibilities and preferences. I am not criticizing MaxFi, but rather the
absence of qualifiers and cautions accompanying your substack post above of the 65 year old man with $2 million in savings who does a conversion and can expect a return of $175K at age 100. I assumed that reference point throughout my comments.
Again, I never claimed to be an economist, and my approach to this discussion is as a non-academician, who read two articles that you recommended to me as being central to your approach. I think is is understandable that I read them in light of the discussion we have been engaged in, not as an expert that was likely to pick up the nuances and details of the debates on Yarri over the last 60 years. If I was overly sharp in my responses about these articles, it was because I felt that I was being led on a wild-goose chase in that these articles were not really relevant to what I see as the key issues. I think this conclusion is buttressed by the fact that only now are you addressing my key points about changing time horizons, and preferences, in a strategy that requires large up front costs, and decades to achieve compensating benefits. I think most investors (as opposed to economists) reading through this discussion will conclude that you have been dilatory in coming to grips with my key points.
Your "Fourth, Sixth and Seventh" points, are good examples of what I am referring to above. To take one example, I have never come across a financial analyst who would consider that "spending more when young and less when old" and "going to Settings and Assumptions and altering their Standard of Living Index, telling the program to spend more, if doing so doesn't violate the household's cash flow, when young and less when old" would be helpful, or even comprehensible, advice. To accuse me of unjustified criticism because I haven't addressed these opaque features in your model, and these arcane points of theory, is strange, to say the least. I think I can be excused for assuming that you are trying to intimidate and impress me with economic jargon and waving of hands.
In your "Eighth" paragraph, you finally acknowledge what I have been trying to get you to address through all of our exchanges. Indeed, considering worst case scenarios (and more realistic scenarios like I only live to 90) should be addressed. "This is the fiducially responsible thing to encourage, not that we are fiduciaries." I was a pension investment consultant, and I was very aware of my fiduciary responsibilities under ERISA. You do not have those responsibilities. Customers should note this fact.
"Ninth" and "Tenth" paragraphs: It was not necessary for me to get into a discussion of the economic sacred texts and internecine jousting among the economic cognoscenti. None of it is relevant to my arguments or to non-academicians. I defer to your expert opinions in these matters.
Your last paragraph: I agree, I should have simply said that people should run MaxiFi with alternative max ages of life because their most desirable annuitization conversion plan will depend on this. People have varying risk tolerances, preferences for how much to spend and when to spend it. Some will not agree with you that the greatest risk to hedge is the risk of living so long that they run out of money. These judgments are subjective as well as objective, and models are good ways, at least for sophisticated and informed investors, to explore these alternatives. You may disagree, but I think our debate is useful for people who are using, or thinking about using, MaxFi to explore financial decisions.
Jack, Let me respond point by point. First, Yarri's and all subsequent longevity risk papers, including Kent's with Reichling, assume a maximum age of life. Yes, Riechling and Smeeters consider one's learning that their max is no longer their assumed age 120 because they, say, learn they have pancreatic cancer. MaxiFi's max Max Age is 110. But we will, on reflection, change that to 120. Our default is 100.
Second, since every economics model incorporates a maximum age of life, MaxiFi, which is an economics-based financial planning educational tool, does as well.
Third, I had detailed conversations with Peter Coy. I don't think you should speak for him about what he did and didn't get. Nor should you speak for the financial planners using our program.
Fourth, the example I provide of a household who has a max age of 100 does incorporate the probability that he may get inoperable cancer in 6 years. As Yarri shows, economics handles the high probability of dying before your max age, either through inoperable cancer or something else, by tilting your age-consumption profile -- spending more when young and less when old. MaxiFi lets users take this calculated gamble by going to Settings and Assumptions and altering their Standard of Living Index, telling the program to spend more, if doing so doesn't violate the household's cash flow, when young and less when old. Have you looked at this setting? If not, you have engaged in lots of unjustified criticism.
Fifth, Yarri considers uncertainty over the max age of life and does so at the beginning of his article. He handles this by including utility of the household's estate. The estate proxies for having money in reserve if the max age of life turns out to be lower than the true realized max. MaxiFi lets you specify bequests in four ways a) specify that some regular assets are reserve funds making them unavailable for spending, b) (This is brand new. specify a ceiling on the amount of discretionary spending you wish to do, c) specify that a share of non-annuitized retirement account assets are to be bequeathed to the extent any are available after RMDs, and d) specify gifts as special expenses at the end of life.
Sixth, Kent is my close friend and former co-author. But I think his paper with Reichling treats Yarri as a straw man. Yarri explicitly contemplates annuities whose loads are so high that they aren't purchased. He's also looking at being able to tailor one's annuities to match one's desired age-consumption profile. His analysis can easily be extended to show that households wouldn't necessarily fully annuitize if they were forced to buy fixed nominal annuities as opposed to specially tailored real annuities. This is doubly true if future inflation is uncertain. MaxiFi's baseline assumption treats annuities as unavailable and then lets users examine the value of purchasing alternative annuities. You can enter them in our annuities screen and compare their value under different assumed future inflation rates (set up as alternative profiles).
Seventh, Reichling and Smetters are focused on changes in the future in the probability distributions of survival rates. As is apparent from Yarri's paper, if there were annuities available at time zero whose payoffs were dependent on the future resolution of those distributions, the single agent with no bequest motive would fully annuitize. Reichling and Smetters are assuming away the existence of such financial instruments. I.e., they are ruling out Arrow state-contingent longevity insurance markets. Yarri does the same when he considers the case that non-state-contingent annuities are unavailable in the context of probability distributions not being part of the future state space. This is the straw man point. Yarri is not claiming that everyone should annuitize given extant markets. He's showing that with perfect insurance markets everyone will annuitize. That's just as true whether future survival probability distributions change or not. If they do, the insurance markets will handle this situation making full insurance the right answer.
Eight, MaxiFi let's users explore future changes in beliefs. For example, suppose a 50 year old expects to work till 65 but realizes they may get fired at 60. Well, you can run MaxiFi under both assumptions. We encourage users to assume worst-case scenarios. This is the fiducially responsible thing to encourage, not that we are fiduciaries. Could we formally incorporate the risk of being fired each year and then do maximum expected utility maximization? Yes, but considering this and all manner of uncertainties makes the problem intractable. This is why we are taking the most conservative assumptions as our base case and letting users explore other outcomes manually. This includes having a maximum age of life that, in six years, you learn is not 40 years, but two years. If one runs Roth conversions assuming an 8 year horizon, the calculated optimal conversion plan will surely be not to convert.
Ninth, You wrote, This is directly addressed by Reichling and Smetters. They show that unexpected illness and increased uninsured medical and other costs (not to mention the change in investment horizon) appropriately change anyone's time horizon and risk tolerance. If someone is already committed to a strategy like a Roth Conversion, they do not have the ability to reflect that change in their time horizon because the costs are sunk and unrecoverable at that point. That is why they avoid annuities and hedge their bets on life expectancy!
This is wrong on two counts. Learning this information doesn't, in R&S, change their risk tolerance. Nor does their analysis connect to the term "life expectancy." It does connect to contingencies that can be explored, if not perfectly, by setting up alternative profiles with different maximum ages of life. But that's true of all kinds of contingencies. If I plan to work to 65, but can get fired at 60, I will spend through 60 as if I was able to work through 65. This will be a sunk cost. MaxiFi would encourage someone who could be fired at 60 to run their plan under that conservative assumption as well as assuming retirement at 65. In the context of longevity risk, the conservative assumption is living to your max age of life. Yes, if you learn at 60 that you have pancreatic cancer, you won't be able to keep working and won't be able to benefit to the same degree as would otherwise have been the case from Roth conversions. But the bigger longevity risk is not leaving money on the table, but running out because you live too long.
Tenth, Economists, like you and Yarri, have been telling them that they were irrational! Yarri and a whole generation of economists after him saw this lack of interest in annuities as irrational, or a puzzle needing further research. Well, people turned out to be right, and economists turned out to be wrong.
Sorry, but show me where Yarri or I or any decent economist has been saying this. (It's certainly not in Yarri's paper, which you claim to have read.) This is a figment of your imagination. Kotlikoff and Spivak (1981) provides one of many solid explanations for the observed lack of annuitization. You are just making up a straw man to say economists are wrong. Not the case. Economists have spent a lot of time explaining why people don't purchase annuities. R&S is a good example. They add to a long list of reasons -- reasons they list, all of which economists pointed out.
I've given you as much of my time on this, Jack, as I can. So, I'm not going to respond to any more of your comments. You could simply have written that people should run MaxiFi with alternative max ages of life because their optimal annuitization conversion plan will depend on this. But, in the end, I'd still focus attention on the max max age of life, which is the truly catastrophic financial scenario, not having paid taxes early, which will likely mean your kids will pay less.
best, Larry
Jack, I suggest you build your own software program and market it. I'm following what economics says. Extending Yarri to consider annuities with loads leads people not to purchase annuities or to purchase fewer annuities. This wasn't ignored by Yarri. His central case of interest is that of people buying no annuities. As for people's maximum ages changing, you can run MaxiFi for different assumed maximum ages of life, weigh the results based on the probability of those maximums, and now we're back forming expected utility. You can also run cases with future tax hikes and future tax cuts. This is why we have alternative profiles that you can run. Larry PS, the rational failure to buy annuities does not, as I say, matter to the issue of maximizing lifetime utility through the max age of life. So, I stand by that point wrt to all the papers that don't consider max age of life uncertainty. That concern can be dealt with in our tool as just described. But one wants to plan cautiously. If it turns out you get pancreatic cancer at 90 and can't enjoy your tax savings thereafter, you'll be in heaven. But if you run out of money at 90 because you said, no way, I can't make it to 100, you'll be in trouble. Larry
Larry, No need for another program, just more cautionary notes on risk tolerance, variable discounting on FV, and time horizon variability. Best wishes, Jack
Jack, Good you read Yarri. Yarri's surprising point is not about pooling the risk of longevity. This goes back to Tontine in the 1600s. It's the simple point that the max age of life is the appropriate planning horizon and that focusing on earlier dates of death never arises. None of the papers you mention matter to the point you keep making and keep getting wrong. In all papers with uncertain lifetimes, the agent maximizes lifetime utility through the maximum age of life. MaxiFi's focus on maximum age of life, not expected age of life is consonant with this key point. And yes, some households may be altruistic and enjoy higher expected utility from conversions if they provide higher spending power for their kids. But knowing if that's true, requires running the kids through MaxiFi in parallel. Whether conversions are a good or bad thing wrt leaving or gifting money to kids requires knowing the kids' tax situations. So, if that's your concern, purchase a license for your kids and run them with alternative assumptions about your date of death. Yarri wasn't disproved by the subsequent papers. Those papers just introduced more realistic assumptions than Yarri made, including the possibility that annuities are available, but aren't actuarially fair. Larry
Larry, you say that "None of the papers you mention matter to the point you keep making and keep getting wrong. In all papers with uncertain lifetimes, the agent maximizes lifetime utility through the maximum age of life." That is misleading in the way that you are interpreting it. Allow me to illustrate why. Putting this into the real world context of your Roth IRA conversion model and your advice to people to use a maximum life assumption of 100 years, you are either implicitly or explicitly advising people to use their maximum life expectancy as a time horizon for decision making despite your claim that you don't give advice. (Peter Coy didn't get that impression, and neither did the financial planners that use your software.) Your example in this substack, in which your retired person spreads out conversions over 5-6 years, gives a completely different result if he finds out in year 6 that he has inoperable cancer and has a life expectancy of 3-5 years. This is directly addressed by Reichling and Smetters. They show that unexpected illness and increased uninsured medical and other costs (not to mention the change in investment horizon) appropriately change anyone's time horizon and risk tolerance. If someone is already committed to a strategy like a Roth Conversion, they do not have the ability to reflect that change in their time horizon because the costs are sunk and unrecoverable at that point. That is why they avoid annuities and hedge their bets on life expectancy! Economists, like you and Yarri, have been telling them that they were irrational! Yarri and a whole generation of economists after him saw this lack of interest in annuities as irrational, or a puzzle needing further research. Well, people turned out to be right, and economists turned out to be wrong. Maybe people just got lucky when they ignored economic advice, or maybe their intuition and common sense proved to be better than that of academicians. I don't know the answer to that, but it seems that you have not recognized that mistake.
I urge any reader of our debate to read these papers themselves. It involves a slog through a lot of economic jargon but I think many people reading these papers will realize that Larry is not addressing important aspects of financial decision making. Larry says that he is not giving advice, but he has been, and is, touting his modeling software to NYT columnists, individuals, and financial advisors because it reveals several strategies that provide large financial benefits, most saliently delaying social security payments and front loading Trad.IRA's withdrawals into Roth IRA's. These strategies are based on facts and they are well analyzed by Larry's software, but they involve assumptions about the future. Those assumptions could be wrong. I am as competent to give an informed opinion on these strategies as Larry, Peter Coy, and all of the financial advisors using his software are. My informed opinion is to diversify, hedge your bets when you reasonably can, and heavily discount assumed returns 30 years in the future, whether they derive from dividends, interest, stock prices or lower taxes.
Jack, You are apparently unable to follow the math in Yarri's paper or you would understand what I wrote. You could just admit that. Let me explain the "weird" point this way. If I'm buying homeowners insure, I buy full coverage to protect myself against the worst case scenario - the house burying down. I don't buy coverage for the expected loss, but for the max loss. And, in doing so, I'm also buying coverage for all smaller losses. Same with lifespan. In setting a spending plan for the worst case scenario -- living to my max age of life, I am also setting a spending plan for all preceding years. This will be my last reply to you. You don't have the background to criticize economics-based financial planning. That's what MaxiFi delivers. If we modified MaxiFi along the lines you suggest, it would no longer comprise economics-based planning. best, Larry
OK, I have now read Yarri. I am puzzled by your emphasis on the math because it is actually straight forward, as is the principal argument that pooled annuities create a surplus value from those who die before the annuity values they receive match the cost of the annuity provider, which earns a profit and presumably passes some of that profit to annuitants with longer lifetimes and lower income in old age. The math, simple or complex, does not settle the question, however. It is generally recognized now that Yarri’s assumptions are unrealistic and fail to capture important real world impacts on these decisions. See Uncertain Lifetime, The Theory of the Consumer, and the Life Cycle Hypothesis, Siu Fai Leung, 1994
I believe it is noteworthy that in two posts now you have drawn our attention to early studies ((Merton 1969 and Yarri 1965) that have been largely modified, even supplanted, by more recent studies that significantly revised, even contradicted, that earlier work. I’m sure that you are familiar with Optimal Annuitization with Stochastic Mortality and Correlated Medical Costs, Felix Reichling and Kent Smetters, 2015 (Spoiler alert: I know that you were consulted on this work and it was supported by BU). This paper throughly revises Yarri’s conclusions, and provides very different conclusions about time horizons and evolving life cycle utility. They found “that including asset management fees and bequest motives….at least, nine out of ten households (should) not hold any annuities.” Also, “depending on our calibration assumptions, between 64% and 76% of households should not annuitize any wealth, even with no transaction costs, bequest motives, ad-hoc liquidity constraints, or asymmetric information problems. In contrast, the Yaari model predicts 0%.” By “should not annuitizing wealth”, Reichling and Smetters explain that the incidence of sickness and shortened life expectancies are, at this point at least, largely stochastic and greatly influence time horizons and optimal investing strategies of investors in dynamic and rational ways. In short, one time horizon (maximum life) and therefore optimal strategy does not exist for all people, even if they have the same risk tolerance. If these conclusions somehow agree with your take on investing for maximum possible age, I think you have substantial explaining to do, having pointed to an outdated and simplistic analysis that has been shown to be misleading, and then having ridiculed our inability to understand these mathematic “truths.” Forgive me if my perception of your end of the dialogue up to this point is one of an intellectual bully trying to awe the bothersome rubes. Best, Jack
Jack, Pls read Yarri, Uncertain Lifetimes, RESTUD, 1965. You claim to be an economist. Let's see if you can follow this fundamental paper. It shows that the parents should value their net resources through their max age of life. As for children, they can be run in their own MaxiFi program based on different assumptions about when their parents will die. But, just like their parents, they need to plan to live to their maximum age of life. If the parents care about the expected utility of their children, the gains to the parents from conversions may or may not be larger. If the kids are in a lower tax bracket than the parents, conversion followed by early death may lead to lower lifetime spending capacity of the kids than absent conversion.
Larry, I'm disappointed to see you return to the ad hominem questioning of my fitness as a critic. I am a Chartered Financial Analyst, who was able to retire at the age of 44, after following a strategy of super saving and standard index investing. I am generally well-read in economics, as I am in a wide variety of disciplines, as you have conceded, and most importantly you have not been able to counter my arguments convincingly. I have never claimed to be an economist, and I actually regard that as a strength in these debates because you appear to be a person who is so dazzled by your (rather eccentric) interpretation of economic theory that you ignore common sense and obvious facts. Your appeal to "it's weird but that's economics" has been almost completely lacking in specifics. You contemptuously dismiss any criticisms of your advice here as uninformed and incompetent, but you fail to link hopelessly vague claims like "parents should value their net resources through their max age of life" to the real choices facing parents with respect to tax and investmtent decisions. Bottom line: it is rational for a parent to look at your model's output that shows (in one example) that someone can covert his Trad. IRA to a Roth at age 65, pay $350K in taxes up front, and eventually make $175K PV in profit by doing so if he lives to be 100. If he lives to 95, he makes 25% less, but still $125K or whatever, if he lives to his average life expectancy, the ROR is rather meager, but economic theory tells us "don't worry!" this is all just life insurance and every other consideration is irrelevant because maximum age of life is supposedly a time-horizon decreed by the theorists and their simulations. In light of your recommendations, my conclusion is that you are lost in an academic fog. You have a stronger case to make with Social Security strategies and your modeling of those. Stick with that and backoff this absolutist pitch for Roth conversions for everyone who shows a half-decent increase in FV from your model. For very wealthy people looking for tax savings, it can be attractive if one has faith in the stability of public policy. For others, it can be reasonable to avoid it.
Let me explain it this way. Suppose you will either live for 10 years or live for 20 years. You have to plan for living for 20 years and the consumption you do under that plan for the first 10 years is what the live-for-20-years plan dictates. And the tax moves you make are based on living to 20 years. If you die in 10 years and don't get to reap the benefit of being able to spend more in the next 10 years, well, you're in heaven, so it doesn't matter. Yes, this is surprising and slightly weird, but it's what economics says. best, Larry
Hi Benjamin, Yes, lowering your max age will lower your conversion gain. But your max age is what it is. You can't count on dying before it. You can't play the averages. This exclusive focus on the worst-case scenario -- living to your max age -- is surprising, but it's what economic theory says to do. best, Larry
Jack, MaxiFi has a standard of living index that let's you effectively include time preference. Frankly, I don't think it will matter much to the gains from Roth conversions.
best, Larry
Jack, Think about the problem in terms of expected lifetime utility. best, Larry
Yes, that's how I am evaluating this decision.
For those readers who are not familiar with the concepts, here are explanations:
To calculate expected lifetime utility, you would sum the discounted utility of consumption in each period of life, taking into account the probability of reaching each period.
Similarly, when evaluating investment options, the concept of FV discounting is used to compare the present value of future cash flows, allowing for a more accurate comparison between different investment choices with different payout timelines.
To calculate expected lifetime utility, economists use a "discount factor" which represents the value placed on future utility compared to present utility, effectively discounting the value of future consumption, and reflecting the uncertainties of future consumption.
The discount factor reflects an individual's "time preference," which is their tendency to prefer immediate consumption over future consumption, even if the future consumption is larger. The discount factor also should reflect the uncertainty if achieving a FV.
Hi Jack, I appreciate your reply, but still don't get your discounting point. Let's take this off line. Pls email me at kotlikoff@gmail.com and we'll find a time to chat on the phone. best, Larry
My point about TIPS rate discounting of FV is straightforward. I recognize that it makes sense from your point of view to use the TIP rate because you regard the strategy as very low risk, but I think it entails substantial risk, and therefore requires a higher discount rate during the 30-40 year wait for FV. In other words, the benefits of the conversion strategy come very far in the future, and should be discounted not only for lost income, but also for uncertainty in the outcome. As for exploring this further with you, I think you would agree with me that we have pretty thoroughly covered (exhausted) the topic. It's been fun, honestly. Best wishes, Jack
Jack, As you know, one can't characterize all the risks that we face and do full expected lifetime utility maximization over all such risks. But one can do scenario analysis -- stress testing if you like. For example, you can consider your apparent fear that your Roth account will be taxed upon withdraw at some point down the road. You can just run MaxiFi treating the extra future taxes as special non-tax-related special expenses. MaxiFi is an educational tool. It's not giving advice. It's here for you and others to explore different scenarios. An important scenario, arguably the most likely, is that current tax law (TCJA) will prevail. But you can run MaxiFi's Roth Conversion Optimizer under different assumptions about future tax increases or tax cuts and, as indicated, adding additional future taxes as special expenses. If you can find a better program for exploring Roth conversions, use it. I know none exists. None of the other programs produces a single result that is internally consistent. Given your fears, you'd presumably advise everyone to take Social Security as early as possible because anyone who waits will be viewed as rich enough to be fleeced. Again, you can run our Social Security optimization specifying future benefit cuts to explore this. In short, I don't know what you want. You are clearly using MaxiFi to help guide your planning. If you'd like some other features, let me know. But I've tried to give you what I think you and others, including myself, need -- namely a tool that solves all the key equations of economics-based personal finance. Best, Larry
You're a real mench Larry, and crazier than I am to be up this late--I'm two hours earlier in Mtn. time zone. I agree that your model is a useful planning tool. You have created a marvelous, sophisticated model, and I believe it has no match in other PC-based models. . However, you mischaracterize my legitimate criticisms of your discount rate for backloaded tax benefits--these are not fear-based, and should not be characterized as a risk-aversion anomaly on my part. Moreover, as the next four years unfold, I reasonably believe that prior norms of risk assessment will be significantly reassessed . Take care, Jack
⚖️Hi Larry, My Underlining Didn't Copy To Substack. I'll Capitalize And Skip A Line And Add A Scale Of Justice Emoji To My Responses.. Hopefully That Will Make Easier To Follow, But I Hope Reading Capitalized Text Is Easier For Others Than It Is For Me!
HI JACK, I REPLY IN BOLD.
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen").
OUR MAXIFI SOFTWARE HANDLES CASH FLOW CONCERNS. YOU HAVE RUN THE ROTH CONVERTER OF LATE, YES? SO I DON'T UNDERSTAND YOUR SUGGESTIONS THAT WE ARE IGNORING THAT MAJOR ISSUE. WE AREN'T. I AM NOT
⚖️ I Understand That Maxfi Deals With Cash Flow And Monte Carlo Simulations. That Is All Fine And Good, And The Complexity And Detail Of Your Model Is Truly Impressive. My Point Is That The Uninitiated Consumer Is Likely To Misinterpret The Narrative That The Model Can Maximize Lifetime Spending/Utility. It Can Do That If The Inputs And Simulations Capture Future Market And Regulatory Events. You And I Know That But I’m Hoping To Give Some Cautionary Information To People Who Might Not Understand That, Or Understand It Fully.
All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design.
MAXIFI'S FULL RISK MONTE CARLO SIMULATION FULLY INCORPORATES CASH FLOW CONSTRAINTS ALONG EACH LIVING STANDARD TRAJECTORY IN DETERMINING INVESTMENT STRATEGIES THAT MAXIMIZE LIFETIME UTILITY. SO, AGAIN, I THINK YOU ARE CRITICIZING OUR TOOL WITHOUT HAVING RUN IT OR UNDERSTOOD IT.
⚖️ As I stated above, I am not criticizing your tool, but the marketing of it.
I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does.
LOVE TO HEAR FROM GENE. NOTHING I'M SAYING OR THAT MAXIFI IS DOING IS ANYTHING BUT FULLY IN TUNE WITH MODERN FINANCE.
⚖️ Ditto. The Question IS Not About Modern Finance, But The Risk/Reward Profile Of The Roth Conversion Assessment.
Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions.
THIS IS, FRANKLY, CONTENT FREE.
⚖️ OK, My Point Is That The Merton/Samuelson 1969 Paper, Which You Cited As Ruling Authority, Considered Time Horizon To Retirement To Be Irrelevant To Portfolio Construction. They Changed That Appraisal In The 90's, And The Current Consensus Is That Time Horizon Is A Critical Variable In Portfolio Construction, Which You Failed to Acknowledge In The Comment That I Was Responding To.
Although Merton And Samuelson In 1969 Suggested That Time Horizon Is Irrelevant For Asset Allocation, Merton And Samuelson In 1992 Argued That Workers Who Are Sure Of Future Employment Should Increase Their Investment In Stocks, While For Retired People It Would Not Be Optimal To Do The Same, Indicating That They Believed Time Horizon Is Relevant To Portfolio Allocation Decisions.
THESE TWO PAPERS AREN'T INCONSISTENT. THEY ARE COMPLEMENTARY. THOSE WITH SAFE FUTURE EARNINGS ARE EFFECTIVELY HOLDING BONDS. THAT'S WHY THEY SHOULD HOLD MORE STOCK TO KEEP THE RATIO OF STOCKS TO BONDS (INCLUDING SAFE EARNINGS) FIXED AS THEY AGE. SEE THIS ARTICLE: https://kotlikoff.net/wp-content/uploads/2019/04/Optimal-Life-Cycle-Investing-AER-2008.pdf
⚖️ In Important Respects They Are Inconsistent. The 1969 Paper Did Not Recognize That With Time Dependent Investment Opportunities, Retirement And Death Are Key To Determining Investment Horizon And Portfolio Construction. Remember That I Am Responding To Your Previous Claim That Time Horizon Is Irrelevant To Investment Decisions. You Partly Acknowledge That Below, But Still Fail To Address That Much Research Has Shown That Risk Tolerance Decreases With An Increase In Undiversifiable Risk, Which Is Likely To Happen, And Therefore Time Horizons Shorten. In Your View This Is Largely Irrelevant To Portfolio Decisions, But I Think Most (Non Academicians) Disagree.
Other distinguished financial economists, like burt malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
YES, THE TIME HORIZON IS IMPORTANT TO THE PORTFOLIO CHOICE, BUT PRIMARILY BECAUSE OF THE COMPOSITION OF ONE'S INCOME FLOWS -- FIRST LABOR EARNINGS AND THEN SOCIAL SECURITY (ANOTHER BOND).
⚖️ Note That The Composition/Amount Of Income Flows Is Relevant To My Point When I Focus Below On Changing Exogenous Factors.
Your advise to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer.
OF COURSE. EXPECTED UTILITY MAXIMIZATION ENTAILS ALWAYS REOPTIMIZING BASED ON THE LATEST INFORMATION.
⚖️ And My Point Is That The Efficacy Of Reoptimizing Is Significantly Constrained By The Upfront Costs Of A Roth Conversion.
OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established.
THERE IS NO FIRST-ORDER EXTERNALITY ISSUE INVOLVED IN ROTH OPTIMIZATION OR PORTFOLIO CHOICE FOR AN INDIVIDUAL HOUSEHOLD. SO, I HAVE NO IDEA TO WHAT YOU REFER.
⚖️ I Am Using “Externality” Metaphorically Here. There Are Many First And Second Order Economic Externalities Impacting All Of Us These Days, But I Am Referring To Rapidly Rising Political, Economic, Etc. Risks That Would Affect Your Model’s Optimization.
Many Of These Risks Will Be Difficult Or Impossible To Hedge. Your Model Assumes General Stability While Changing Conditions Are Supposedly Accommodated By Yearly Reevaluations.
YES, WE COULD HAVE NUCLEAR WAR AND MAXIFI DOESN'T CONSIDER ITS LIKELIHOOD IN DECIDING WHETHER TO DO ROTH CONVERSIONS. IF THIS IS YOUR OBJECTION, I THINK YOU SHOULD ASK YOURSELF WHY YOU WOULD WANT TO DO ANYTHING EXCEPT PUT A BLANKET OVER YOUR HEAD AND SCREAM.
⚖️ We Agree That Nuclear War Is Not A Diversifiable Risk, Nor Can It Be Hedged By Investment Decisions. Your Argument Doesn’t Apply, However, To Increased Insecurity Of Income And Expenses In All Time-Horizons, Coming From Political Instability, Trade Wars, Climate Change, Etc. These Risks Could Plausibly Reduce One’s Willingness To Bet On A Strategy That Involves A 35-Year Time Horizon.
I think that is an assumption people should not make. Well, you may reasonably reply, these (rising risks) “externalities” impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do.
LEAVING TAX BREAKS ON THE TABLE IS NOT SUPERCHARGING ANYONE'S RISK. PLANNING FOR PEOPLE TO LIVE TO THER MAX AGE OF LIFE IS THE RIGHT APPROACH TO LONGEVITY RISK. READ YARRI 1969.
⚖️ Again, These Risks Could Reasonably Reduce One’s Willingness To Bet On A Strategy That Involves A 35-Year Time Horizon And Stable Govt. Policy.
You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not.
YES, ECONOMICS SAYS THAT PEOPLE SHOULD CAUTIOUSLY SPEND OUT OF THEIR FUTURE RESOURCES, CAUTIOUSLY ESTIMATED. THAT'S WHAT BOTH OUR DETERMINISTIC AS WELL AS STOCHASTIC PLANNING REPORTS DO.
Continued Below
LIKE
REPLY (1)
Jack Cargill
8h
⚖️ I Agree, The Disagreement Is In What Constitutes A Cautious Approach To Increasing Risk.
My Argument, In Part, Is That Those Expectations Of Future Benefits Are Less Likely To Be Realized Than They Were A Few Years Ago, And Will Be Less Likely Still Five Years Hence. If I Were To Execute Such A Strategy Today, 5-6 Years From Now I Would Have No Way To Change My Strategy If My Assumptions Have Changed.
WHEN IT COMES TO ROTH CONVERSIONS, THE COUNTRY'S FISCAL INSOLVENCY POINTS TO TAX HIKES DOWN THE ROAD. THAT WOULD MAKE CONVERSIONS EVEN MORE ADVANTAGEOUS. USERS CAN ADD THAT TO THEIR PLANS UNDER SETTINGS AND ASSUMPTIONS. BUT OUR DEFAULT DOESN'T INCLUDE THIS HAPPENING TO MAKE THE CONSERVATIVE ASSUMPTION. THE WORRY YOU ARE REFERENCING IS, FRANKLY, RIDICULOUS GIVEN OUR FISCAL GAP, OF WHICH I'VE WRITTEN AS MUCH AS ANYONE. YOU ARE AFRAID TO DO A ROTH CONVERSION BECAUSE YOU ARE WORRIED THAT TAXES WILL FALL IN THE FUTURE. WELL, MAYBE YOU SHOULD RETHINK THAT POSITION.
⚖️ You’re Cherry Picking Your Scenarios. I Never Said That I Think Tax Rates Will Fall—That Is A Straw Man. The Threat Of Insolvency Is Just As Likely To Lead To Taxation Of Roth Tax Havens For The Wealthy (See Peter Coy’s Comments On That Subject), Or Some Other Rollback Of Those Conversion Benefits. I Don’t Think That Would Be Ridiculous Or Even Particularly Unlikely. I Doubt That Most Of Your Customers Understand The Number Of Contingencies Involved In This Strategy, And The Limitation On Their Choices If Prior Assumptions Turn Out To Be Wrong.
SEE ABOVE.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
JACK, GOOD TO DISCUSS. BUT YOU ARE BEING NIHILISTIC. I APPRECIATE THAT YOU KNOW MORE REAL FINANCE THAN I THOUGHT, BUT YOU ARE SUGGESTING THINGS THAT REAL FINANCE DOESN'T COUNTENANCE, NAMELY PUTTING YOUR HEAD IN THE SAND BECAUSE THE FUTURE IS UNCERTAIN. THE FUTURE HAS ALWAYS BEEN UNCERTAIN AND IT SOMEONE FINDS A WAY TO LOWER THEIR FUTURE TAXES WITH VERY HIGH PROBABILITY, THEY SHOULD RATIONALLY SPEND MORE TODAY UNLESS THEY ARE EXTREMELY RISK AVERSE. I DON'T BELIEVE YOU ARE EITHER IRRATIONAL OR EXTREMELY RISK AVERSE. HENCE, I SUGGEST YOU TAKE MAXIFI'S ROTH SUGGESTIONS SERIOUSLY. BEST, LARRY
⚖️ I Am Guilty Of Using “Externality” Metaphorically Instead Of Maybe Exogenous, You Are Guilty Of Using Nihilistic When You Should Say Pessimistic. I Cop To That, And Polls Show That Is A Majority Position These Days. I Likewise Appreciate That You Know More Real Finance Than I Thought, But You Are Egregiously Misrepresenting My Points And Baselessly Characterizing Them As Inconsistent With "Real Finance". Likewise, I Think It Is You Who Are Putting Your Head In The Sand To Ignore That The Future Is Becoming Much More Uncertain In Many Ways; To Expect That A 35-Year Payoff Of $175K Pv After A $300K Up Front Bet Is Questionable, And To Propose That A Discount Rate Set By The Tips Rate Is Sufficient For Future Benefits. You Said That The Roth Conversion Model Produces A “Very High Probability” Of Lowering Taxes. My Main Point Is That Those Probabilities Are Subject To Many Assumptions And Risks. Buyer Beware. Best Wishes, Jack
HI JACK, I REPLY IN BOLD.
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen"). OUR MAXIFI SOFTWARE HANDLES CASH FLOW CONCERNS. YOU HAVE RUN THE ROTH CONVERTER OF LATE, YES? SO I DON'T UNDERSTAND YOUR SUGGESTIONS THAT WE ARE IGNORING THAT MAJOR ISSUE. WE AREN'T.
All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design.
MAXIFI'S FULL RISK MONTE CARLO SIMULATION FULLY INCORPORATES CASH FLOW CONSTRAINTS ALONG EACH LIVING STANDARD TRAJECTORY IN DETERMINING INVESTMENT STRATEGIES THAT MAXIMIZE LIFETIME UTILITY. SO, AGAIN, I THINK YOU ARE CRITICIZING OUR TOOL WITHOUT HAVING RUN IT OR UNDERSTOOD IT.
I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does.
LOVE TO HEAR FROM GENE. NOTHING I'M SAYING OR THAT MAXIFI IS DOING IS ANYTHING BUT FULLY IN TUNE WITH MODERN FINANCE.
Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions.
THIS IS, FRANKLY, CONTENT FREE.
Although Merton and Samuelson in 1969 suggested that time horizon is irrelevant for asset allocation, Merton and Samuelson in 1992 argued that workers who are sure of future employment should increase their investment in stocks, while for retired people it would not be optimal to do the same, indicating that they believed time horizon is relevant to portfolio allocation decisions.
THESE TWO PAPERS AREN'T INCONSISTENT. THEY ARE COMPLEMENTARY. THOSE WITH SAFE FUTURE EARNINGS ARE EFFECTIVELY HOLDING BONDS. THAT'S WHY THEY SHOULD HOLD MORE STOCK TO KEEP THE RATIO OF STOCKS TO BONDS (INCLUDING SAFE EARNINGS) FIXED AS THEY AGE. SEE THIS ARTICLE: https://kotlikoff.net/wp-content/uploads/2019/04/Optimal-Life-Cycle-Investing-AER-2008.pdf
Other distinguished financial economists, like Burt Malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
YES, THE TIME HORIZON IS IMPORTANT TO THE PORTFOLIO CHOICE, BUT PRIMARILY BECAUSE OF THE COMPOSITION OF ONE'S INCOME FLOWS -- FIRST LABOR EARNINGS AND THEN SOCIAL SECURITY (ANOTHER BOND).
Your advise to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer.
OF COURSE. EXPECTED UTILITY MAXIMIZATION ENTAILS ALWAYS REOPTIMIZING BASED ON THE LATEST INFORMATION.
OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established.
THERE IS NO FIRST-ORDER EXTERNALITY ISSUE INVOLVED IN ROTH OPTIMIZATION OR PORTFOLIO CHOICE FOR AN INDIVIDUAL HOUSEHOLD. SO, I HAVE NO IDEA TO WHAT YOU REFER.
Many of these risks will be difficult or impossible to hedge. Your model assumes general stability while changing conditions are supposedly accommodated by yearly reevaluations.
YES, WE COULD HAVE NUCLEAR WAR AND MAXIFI DOESN'T CONSIDER ITS LIKELIHOOD IN DECIDING WHETHER TO DO ROTH CONVERSIONS. IF THIS IS YOUR OBJECTION, I THINK YOU SHOULD ASK YOURSELF WHY YOU WOULD WANT TO DO ANYTHING EXCEPT PUT A BLANKET OVER YOUR HEAD AND SCREAM.
I think that is an assumption people should not make. Well, you may reasonably reply, these externalities impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do.
LEAVING TAX BREAKS ON THE TABLE IS NOT SUPERCHARGING ANYONE'S RISK. PLANNING FOR PEOPLE TO LIVE TO THER MAX AGE OF LIFE IS THE RIGHT APPROACH TO LONGEVITY RISK. READ YARRI 1969.
You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not.
YES, ECONOMICS SAYS THAT PEOPLE SHOULD CAUTIOUSLY SPEND OUT OF THEIR FUTURE RESOURCES, CAUTIOUSLY ESTIMATED. THAT'S WHAT BOTH OUR DETERMINISTIC AS WELL AS STOCHASTIC PLANNING REPORTS DO.
My argument, in part, is that those expectations of future benefits are less likely to be realized than they were a few years ago, and will be less likely still five years hence. If I were to execute such a strategy today, 5-6 years from now I would have no way to change my strategy if my assumptions have changed.
WHEN IT COMES TO ROTH CONVERSIONS, THE COUNTRY'S FISCAL INSOLVENCY POINTS TO TAX HIKES DOWN THE ROAD. THAT WOULD MAKE CONVERSIONS EVEN MORE ADVANTAGEOUS. USERS CAN ADD THAT TO THEIR PLANS UNDER SETTINGS AND ASSUMPTIONS. BUT OUR DEFAULT DOESN'T INCLUDE THIS HAPPENING TO MAKE THE CONSERVATIVE ASSUMPTION. THE WORRY YOU ARE REFERENCING IS, FRANKLY, RIDICULOUS GIVEN OUR FISCAL GAP, OF WHICH I'VE WRITTEN AS MUCH AS ANYONE. YOU ARE AFRAID TO DO A ROTH CONVERSION BECAUSE YOU ARE WORRIED THAT TAXES WILL FALL IN THE FUTURE. WELL, MAYBE YOU SHOULD RETHINK THAT POSITION.
I doubt that most of your customers understand the number of contingencies involved in this strategy, and the limitation on their choices if prior assumptions turn out to be wrong.
SEE ABOVE.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
JACK, GOOD TO DISCUSS. BUT YOU ARE BEING NIHILISTIC. I APPRECIATE THAT YOU KNOW MORE REAL FINANCE THAN I THOUGHT, BUT YOU ARE SUGGESTING THINGS THAT REAL FINANCE DOESN'T COUNTENANCE, NAMELY PUTTING YOUR HEAD IN THE SAND BECAUSE THE FUTURE IS UNCERTAIN. THE FUTURE HAS ALWAYS BEEN UNCERTAIN AND IT SOMEONE FINDS A WAY TO LOWER THEIR FUTURE TAXES WITH VERY HIGH PROBABILITY, THEY SHOULD RATIONALLY SPEND MORE TODAY UNLESS THEY ARE EXTREMELY RISK AVERSE. I DON'T BELIEVE YOU ARE EITHER IRRATIONAL OR EXTREMELY RISK AVERSE. HENCE, I SUGGEST YOU TAKE MAXIFI'S ROTH SUGGESTIONS SERIOUSLY. BEST, LARRY
Larry, I tried to keep our conversation in one post, but it was too long, and Substack made me break it up. Somehow it got moved down below your response to Gerald. Look for it there.
Hi Jack, Sorry for the tough love. But you have to study economic theory before you can understand it. Have you done so? Do you understand the economics of uncertainty as laid out by John von Neumann (co-developer of the A bomb) and Oscar Morgenstern? This is taught in every PhD program in economics in year 1. It's the theory that underlies the classic 1969 results of Merton and Samuelson that household's optimal portfolios are independent of their age (ignoring cash flow considerations). So, no, when we invest in stocks, we don't fundamentally play the averages. We concern ourself with the entire distribution of returns, not just the mean. Gene teaches and respects expected lifetime utility maximization as much as any economist. Just call him and ask. So, I don't know what you mean when you say he recommends investing for long-term horizons. That sounds like he believes that stocks are safe in the long run. Call and ask. He'll refute that in a minute. Or check how much it costs to insure that the S&P will beat TIPS over horizon X. It costs most the larger is X. Hence, the markets are directly refuting this proposition. And, no, life expectancy is not critical, indeed has no impact whatsoever, on one's investment strategy. Read Merton and Samuelson's independent 1969 articles on his. Re MaxiFi's Roth Conversion Optimizer, it takes explicit account of the degree to which households are cash-flow constrained. So, I have to object to your statement on this as well. Finally, re fire insurance, everyone buys catastrophic insurance even if they are willing, due to loads, to self insure small losses. In the lifespan context, this means planning for the catastrophic event, namely that you live to your max age of life. As for your statement -- paying $300 in current taxes for a chance at more spending late in life -- is missing the point of MaxiFi. When it shows your lifetime discretionary spending goes up, it's saying that your lifetime tax savings are larger, potentially far larger than the current $300K tax cost. It's also saying that you can have a higher living standard EVERY YEAR -- starting this year --- unless you are cash-flow constrained. And the program shows you whether such constraints will arise and also, as I mentioned, lets you optimize subject to not producing cash-flow problems. This may mean not doing any conversions. But in your case, you don't sound cash-flow constraints. If I've got you right, not doing conversions that lower your lifetime taxes is paying extra taxes for no reason. It's throwing away alpha based on a misunderstanding of economics and even your own financial situation.
All this said, some people have particular habit-preferences, not considered by Merton and Samuelson. They are extremely concerned about downside risk, but still want to experience upside to their living standard floor. That's why we developed MaxiFi's Upside Investing -- to help those types of households. It entails long-term stock investing that is not exposed to sequence of return risk, delivering only increases in one's sustainable living standard. Gene might have said whatever you are thinking he said in the context of such households or, indeed, foundations, universities, endowments, etc.
Glad we had this exchange. I'm a tough bird. Don't take this badly. I wish I had your investing acumen.
Larry
Hi Larry, Don’t worry about hurting my feelings, I enjoy exchanging opinions and don’t bruise easily. I don’t understand your comment about my lack of attention to cash flow constraints, alpha, and wasted tax payments (not to mention my "investing acumen"). All I can say at this point in response to those comments is that I don’t generally believe in predictable alpha in public markets with no insider information, and I think cash flow constraints are crucial considerations in portfolio design. I mentioned Gene Fama as a contributor to, and symbol of, modern portfolio theory, as a small counter balance to your list of notables. Nevertheless, because you think he will back your position in this argument, I am sending Gene a copy of our exchanges and asked him to weigh in, I hope he does. Your invocation of economic expertise and settled science in these matters doth strike me as protesting too much, or at least conveniently ignoring the assumptions in these theories that justify extreme caution when using them to guide real world investment decisions. Although Merton and Samuelson in 1969 suggested that time horizon is irrelevant for asset allocation, Merton and Samuelson in 1992 argued that workers who are sure of future employment should increase their investment in stocks, while for retired people it would not be optimal to do the same, indicating that they believed time horizon is relevant to portfolio allocation decisions. Other distinguished financial economists, like Burt Malkiel, also argue that time horizon is an important element in portfolio allocation decisions.
Your advice to users of your model is to frequently take into account changed conditions (one should ponder the scenario in which after 6 years of IRA conversions, the world changes, and the model tells you that what was optimal no longer is). What kind of changes might plausibly occur in a world moving toward oligarchy, fascism, international conflict, AI-enabled mass unemployment, and climate and weather disasters? These concerns seemed speculative a few years ago, but no longer. OK, I think we have a new context for evaluating the stabilities assumed in your models: Economic theorists have always had a problem dealing with externalities, and we have externalities in spades, some vague and some already well established. Many of these risks will be difficult or impossible to hedge. Your model assumes general stability while changing conditions are supposedly accommodated by yearly reevaluations. I think that is an assumption people should not make. Well, you may reasonably reply, these externalities impact other investment strategies also, including indexing and time-horizon duration modification. Yes, but for retirement age investors standard portfolio theory is telling them to reduce their risk as they get older, not supercharge it, as your model and theory do. You say that lifetime discretionary spending goes up “Every Year, starting this year, unless you are cash-flow constrained.” IMO, this is misleading, because it requires that many possible developments have to occur or not occur. You pay the taxes up front, and trust that the tax-protected earnings in your Roth IRA eventually make up those tax payments in 20-35 years. I may be missing something, but I assume your “every-year” gain in spending depends on a willingness to spend some of your expected life-time earnings in advance. That might work and it might not. My argument, in part, is that those expectations of future benefits are less likely to be realized than they were a few years ago, and will be less likely still five years hence. If I were to execute such a strategy today, 5-6 years from now I would have no way to change my strategy if my assumptions have changed. I doubt that most of your customers understand the number of contingencies involved in this strategy, and the limitation on their choices if prior assumptions turn out to be wrong.
I'm glad we are having this exchange, and I appreciate your willingness to engage. Jack
Hi Donald, We just moved off of Dupal and need to repost that page. I'll check in with our CTO. best, Larry