Is Wall Street Meeting the SEC's Rule (Reg BI) to Act in the Retail Investor's Best Interest?

This is the Question the SEC and FINRA Should Be Asking.

Photo by Francisco De Legarreta C. on Unsplash

Tens of millions of American households rely on the financial industry to provide investment and related financial advice that is in their best interest. The requirement that the industry recommend financial products that serve retail investors’ interests ahead of their own is stipulated in the Security and Exchange Commission’s (SEC) 2019 Best Interest Regulation (Reg BI).

… regardless of whether a retail investor chooses a broker-dealer or an investment adviser (or both), the retail investor will be entitled to a recommendation (from a broker-dealer) or advice (from an investment adviser) that is in the best interest of the retail investor and that does not place the interests of the firm or the financial professional ahead of the interests of the retail investor.

Reg BI includes four “Care Obligations”.

1. Understanding the potential risks, rewards, and costs associated with a recommendation

2. Having a “reasonable understanding” of a retail investor’s investment profile

3. Based on those understandings and a consideration of reasonably available alternatives, having a reasonable basis to believe the recommendation provided is in the retail investor’s best interest

4. Exercise (of) reasonable diligence, care, and skill in making the recommendation

Is the Financial Industry Systematically Violating Reg BI?

In my opinion the answer is yes. Upholding Reg BI requires using the best available lifecycle investment-guidance software — software that’s not decades old, that’s not based on arbitrary rules of thumb, and that accords with a century of economic financial science — whether or not doing so leads to the highest profits.

I think the industry is using such substandard software to form its investment advice. In saying this, I must immediately disclose my conflict of interest. My software company, Economic Security Planning, Inc. (ESP), develops and markets MaxiFi Planner (MaxiFi). MaxiFi delivers economics-based financial planning guidance, including investment guidance that can be at great odds with that provided by the industry. Given my conflict, please read everything that follows knowing that it is a) self serving and b) may be perfectly true.

A Word About MaxiFi

ESP has spent three decades carefully constructing MaxiFi to fully align with economic science, specifically the subfield of lifecycle personal finance as developed by a Who’sWho of economists and one exceptional polymath — John von Neuman — over a century. Building MaxiFi has taken a very long time for a very good reason. Personal financial decisions are remarkably complicated and their joint solution required developing advanced algorithms, including a new, patent-winning, method of dynamic programming.

A good metaphor for MaxiFi’s inner workings is the 2024 Vacheron Constantin Berkley Grand Complication — the most complicated watch ever built. The watch has thousands of interconnected, moving parts, including a Chinese perpetual calendar with three mechanical brains. It too wasn’t built in a day.

MaxiFi’s parts are a multitude of equations, many highly non-linear, and many encoding vast numbers of fiscal provisions, including Social Security’s thousands of benefit rules. All the equations need to work together — be simultaneously solved to an extreme degree of accuracy — even to answer seemingly simply questions such as “Can I afford to retire early?”

Like The Grand Complication, MaxiFi’s innards are crazy complex. But grasping its results is no harder than telling time. The tool’s user friendliness plus its economics-based planning is, presumably, why Bankrate co-named MaxiFi “Best Financial Planning Software of 2025.”

Reg BI

Reg BI focuses on investment advice. Let’s compare, then, conventional and economics-based investment guidance and ask which comprises best practice. Certainly, the “exercise (of) reasonable diligence, care, and skill in making the (investment) recommendation” requires use of economically appropriate, state-of-the-art technology.

Wall Street’s Investment Advice

The financial industry’s investment advice overwhelming relies on either Modern Portfolio Theory (MPT), including standard Mean Variance Optimization (MVO), or Monte Carlo Targeted-Spending Simulations (MCTSS). This, at least, was the assessment of each LLM that I queried.

MPT is now APT – Ancient Portfolio Theory. It was developed 73 years ago and, cannot, in its standard form, deliver portfolio advice appropriate to households’ complex, time-varying financial lives. As for MCTSS, its rule-of-thumb, portfolio-advice methodology contradicts basic tenants of economic science, not to mention common sense. In short, MPT violates the industry’s Best Interest obligation by omission. MCTSS violates it by commission.

Re my critique of MPT, don’t just take it from me. Read this article, “The Importance of Joining Lifecycle Models with Mean Variance Optimization,” by Thomas Idzorek, CFA and current Chief Investment Officer-Retirement at Morningstar and Paul Kaplan, CFA and former Director of Research at Morningstar. The article’s abstract begins,

For nearly three-quarters of a century, there has been a large separation between lifecycle finance models stemming from numerous Nobel laureates and the single-period mean-variance optimization-oriented models starting with Markowitz.

Idzorek and Kaplan also just published a tour-de-force book called Lifetime Financial Advice. The book devotes 250 pages to detailing, primarily in mathematical terms, precisely how conventional financial planning, in general, and conventional investment advice, in particular, needs to change to accord with lifecycle economics.

The Noble Laureates — 11 in total — that Idzorek and Kaplan reference are all economists. They, together with Irving Fisher, John von Neuman, Oscar Morgenstern, and Menachem Yarri, developed the equations of economics-based financial planning — the ones MaxiFi jointly solves. One of the 11 Nobel Laureates is MIT’s Robert Merton. Merton teaches MaxiFi in his graduate finance course. Here is his reason.

I assign MaxiFi Planner in my Asset Management Course at MIT’s Sloan School of Management as an Outstanding Science-Based Lifecycle and Retirement Management Platform.

Lifetime Expected Utility Maximization (LEUM)

LEUM, which underlies MaxiFi’s investment guidance, represents economics’ framework for providing investment advice. LEUM is taught in graduate courses in economics and finance departments throughout the world. LEUM considers the full array of possible living standard paths a household or client may experience in each future year as a result of following a particular investment strategy, including time-varying asset allocations. Each living standard path is evaluated in terms of its lifetime utility or happiness with bad paths receiving heavier weight in impacted the expected (average) level of utility. This downside weighting captures risk aversion.

The investment strategy — which assets to hold in each future year — that produces the highest level of lifetime expected utility is the one economics recommends. As Idzorek and Kaplan make analytically clear and as I demonstrate here using MaxiFi, optimal, i.e., LEUM-maximizing, investment strategies are heavily dependent on a host of investor-profile factors, such as the level of their Social Security benefits, the taxable nature of their retirement accounts, and their need for short-run liquidity, which MPT ignores and MCTSS either ignores, mis-measures, or mishandles.

Modern Portfolio Theory

MPT was developed by Harry Markowitz in 1952, for which he received the Nobel Prize. At the time, the model represented a true breakthrough -- the first use of expected utility maximization to guide portfolio choice. But its simplifications – compressing the lifecycle into a single period, ignoring the endogeneity of our highly non-linear fiscal system, ignoring changing household demographics, ignoring cash-flow constraints, and ignoring non-financial assets, including labor earnings and Social Security benefits, make it wholly inappropriate for providing retail investors portfolio advice.

If you consume only at the end of your life, your consumption is just your terminal wealth. But your terminal wealth equals your initial wealth multiplied by 1 plus your portfolio return. Thus, Markowitz’s single-period assumption plus several other mathematical short cuts let him express expected utility as depending on just two things — the mean (average or expected) value of your portfolio’s return and its variance (variability). Next, Markowitz showed how to find efficient portfolios — portfolios that had the highest mean return for a given variance. Which efficient portfolio to choose then boiled down to the one that maximized expected utility. And this boiled down to a household’s risk aversion. More risk-averse households would hold a larger share of their wealth in bonds and other safe assets and a smaller share of their wealth in stock and other risky assets, while less risk-averse households would do the opposite. Thus, was born mean-variance portfolio optimization, which Markowitz called Modern Portfolio Theory.

MPT’s investment recommendations connected to economic theory. The framework formally controlled for the correlation of asset returns, and MPT’s mathematics could be expressed in simple terms. This made MPT highly modern relative to what passed for portfolio advice in 1952. But people don’t live for just one period. They live for many years and face a host of factors that Markowitz’s framework entirely ignores. The list includes marital status, the presence and ages of children, economies of shared living, the relative cost of children, career, housing, and retirement decisions, federal and state income taxes, payroll taxes, Social Security benefits, cash-flow constraints, college support for kids, Roth and non-Roth tax-favored retirement accounts, life insurance needs and premium payments, Medicare IRMAA premiums, Roth conversion strategies, RMDs, QCDs, annuitization, annual taxes, and annual Social Security benefits. Each of these factors enters into expected lifetime utility maximization. And each will materially impact a client’s optimal, potentially time-varying investment strategy.

In ignoring these factors, MPT fails to satisfy the first three of Reg BI’s requirements. As for the fourth requirement, using a 73 year-old asset-allocation method that ignores these factors when software is available that incorporates them all and comports with modern finance surely fails to “exercise (of) reasonable diligence, care, and skill in making the recommendation.”

Monte Carlo Target-Spending Simulations

MCTSS starts by setting a rigid annual retirement spending target that is generally unaffordable absent high-yield, high-risk investing. I.e., the target generally fails to satisfy lifetime budget balance based on a fiduciarily-prudent TIPS-investment assumption. MCTSS recommends portfolios that provide an 80 percent or higher probability of “success.” Success references avoiding complete insolvency. MCTSS appears to take no account whatsoever of a retail investor’s risk tolerance or risk-absorption capacity. Nor does it appear to disclose either the extent to which “successful” investment trajectories entail losing most of one’s wealth or the earliest and average ages when such insolvency arises. LEUM would not view even a 1 percent probability of financial insolvency as “safe” if losing all one’s wealth meant a far lower living standard for even a single year, let alone for the rest of one’s life.

MCTSS’ retirement-income replacement-rate formula bears no necessary connection to what a client can safely afford. As Perplexity:AI states, “The financial industry’s commonly used retirement-income replacement ratio—usually 70% to 85% of preretirement earnings—does not necessarily reflect what a client could afford using only safe, inflation-indexed bonds such as TIPS.”

Having set a retirement-spending target that is, fiduciarily speaking, likely too high, if not far too high, the financial industry runs Monte-Carlo simulations that accumulate up the client’s initial assets (typically less than should have been saved) as well as the client’s assumed-constant-through-retirement current saving (typically less than would be associated with safe consumption smoothing) and accumulate down the client’s targeted spending. This accumulation and decumulation is based on returns drawn at random from the client’s likely conservative portfolio allocation.

Because initial wealth is generally too low, because assumed ongoing saving is generally too low, because the client’s current portfolio is generally conservative, and because the post-retirement spending target is generally far above what is remotely affordable given prudent investment, the Monte Carlo simulations will generally show a low probability of meeting “the client’s” target without running out of wealth.

Next, MCTSS practitioners simulate the probability of making the “client’s” target while investing in the higher yield assets marketed by the investment advisor. Because the securities used in this second-round Monte Carlo have higher average yields, the success probability rises and the failure probability falls. But the client is typically still left with an up to 20 percent chance of destitution that could occur as early as the first day of retirement. Taking on more risk is never free. Yes, the probability of success rises and the probability of failure falls. But the probability that losses will be very large when they do arise also rises. So does the probability of early insolvency.

MCTSS bears no relationship to LEUM. Indeed, LEUM adjusts spending on an annual basis -- before as well as after retirement -- to the client’s current financial situation, not to an arbitrary target. MCTSS’ reassuring “stay-the-course” allure reflects Wall Street’s routine assertion that long-run equity investing is safe, i.e., that declines in the client’s portfolio will self-correct in the future – an assumption for which there is no statistically valid empirical, let alone theoretical support. Finally, MCTSS fails to value living-standard trajectories based on the client’s degree of risk aversion – something that lies at the heart of LEUM. For these and other reasons discussed here, including the industry’s cherry-picking of returns series in its Monte Carlo runs discussed here, MCTSS fails to meet SEC Reg BI’s requirements.

Conclusion

No physician would be permitted to use decades-old technology or one-size-fits-all rules of thumb to diagnose illness or prescribe treatment. Doing so would lead to a swift ouster from the profession. The SEC and FINRA have permitted such practices in the realm of investment advice because they have ignored the prevailing scientific standard of care, namely that set by the field of economics, specifically its subfield of finance. At this point, such ignorance is willful — serving the industry while disserving the public.

Is this over the top? Is it self serving? Is it sour grapes?

I can’t claim to be an impartial judge. But I can claim that the use of outmoded, primitive financial-planning methods is producing investment and other financial “advice” that, far too often, is either bad, terrible, or extremely dangerous, not to mention expensive.

---

Economics Matters — Blog/Podcast/Financial Riddler/MaxiFi Puzzler

Please sign up for a free subscription to all four products at larrykotlikoff.substack.com. I’ll immediately make it a lifetime subscription. But please support Economics Matters. Just unsubscribe and resubscribe here on a paid basis. Paid subscriptions come with a one-hour, free financial consultation. Email me to find a time at kotlikoff@gmail.com. You can unsubscribe and resubscribe for free. Also, please sign up kids, friends, colleagues, students, … !