Income Planning

It Pays Commissions, So I Don't Need to Do the Math

Jermaine Carter · September 30, 2026 · 12 min read

Whole life insurance cash value used as a retirement income buffer has been tested in peer-reviewed research and major independent studies. When the math is done — comparing the policy's internal rate of return against taxable bonds on an after-tax basis, accounting for how guaranteed cash value behaves when markets fall — the case for some households is real. The case against skipping that math entirely is stronger still.

The objection that ends most conversations is a compensation disclosure: "That pays commissions." It sounds like analysis. It functions as a way to avoid the arithmetic. How an advisor gets paid is always a fair question. Whether a contract's guarantees, tax treatment, and behavior in a bad year are worth studying is a different question — one that deserves its own answer.

What the Research Actually Tested

Two independent studies tested whole life cash value directly against investment-only and buy-term-and-invest alternatives. Both found that integrated strategies generally improved retirement outcomes, though the details matter considerably.

Wade Pfau and Michael Finke, both of The American College of Financial Services, published Integrating Whole Life Insurance into a Retirement Income Plan: Emphasis on Cash Value as a Volatility Buffer Asset in April 2019. The method compared couples who bought term and invested the difference against couples who bought a limited-pay whole life policy and put less into their 401(k). In retirement, the "pure buffer" strategy drew income from cash value in any year after the portfolio posted a negative return, and from the portfolio otherwise. Using 10,000 Monte Carlo simulations and a 90% success target through age 100, the model showed meaningful improvements at the median: for a couple starting at age 35, the sustainable 401(k) withdrawal rate rose from 2.85% to 4.31%, total income at 65 rose 30%, and legacy at 100 rose 32%. For a couple starting at 50, the withdrawal rate rose from 2.69% to 4.30%, with 20% more income and 13% more legacy.

The trade-offs are stated plainly in the same paper. The 401(k) balance at retirement was 14% lower at the median for the 35-year-old couple, because premiums displaced contributions. At the 90th percentile — when markets cooperate — buy-term-and-invest produced the larger legacy. The buffer strategy is built for adversity, not for the good-luck scenario.

Ernst & Young published Benefits of Integrating Insurance Products into a Retirement Plan in 2022. It ran 1,000 Monte Carlo scenarios and treated permanent life cash value as part of the fixed income allocation — replacing bonds rather than stocks. For a 25-year-old couple allocating 10% of savings to whole life, after-tax retirement income rose 1.0% and legacy rose 4.4%. At 30%, income rose 2.0% and legacy rose 13.4%. EY described allocations of 10% to 30% as generally supportable; larger allocations began to reduce income by redirecting too much away from equities.

One correction circulates widely enough to name directly. Some articles report that EY found approximately 20% more retirement income from whole life. The 20% figure in that paper is legacy, not income. The income effect of whole life alone ranged from −1.2% to +2.0% across EY's 25- and 35-year-old cases. The larger income gains in the study came from adding a deferred income annuity, which contributed up to 8.2%. Using the wrong number does not help anyone evaluate the strategy honestly.

Why a Guaranteed Asset Matters at the Wrong Moment

The logic behind the buffer strategy rests on what sequence-of-returns risk actually costs a retiree. A portfolio that loses significantly in the first years of retirement never fully recovers, because income needs force asset sales at depressed prices. A Pfau and Kitces paper published in the Journal of Financial Planning in January 2014 found that starting retirement conservatively and increasing equity exposure over time — rather than holding a fixed allocation — reduced both the probability and magnitude of portfolio failure. That paper studied stocks and bonds only. Its relevance here is the principle: the early years of retirement are the most dangerous, so the composition of the safe asset held during those years deserves scrutiny.

The question that scrutiny raised in 2022 was about bonds specifically. In that calendar year, the S&P 500 returned −18.1% and the Bloomberg U.S. Aggregate Bond Index returned −13.0%. The Bloomberg Municipal Bond Index fell −8.53%, its worst year since 1982. Stocks and bonds fell together because the Federal Reserve raised rates aggressively to address inflation, and bond prices move inversely to yields — with the severity of the move roughly proportional to the bond's duration. Guaranteed whole life cash value cannot decline by contract. That is the structural distinction the buffer research is built on: not a prediction about future markets, but an observation about how different assets behave when the correlation that retirees rely on breaks down.

The buffer rule is mechanical: after a year the portfolio lost money, income comes from cash value; otherwise it comes from the portfolio. The goal is to avoid forced sales at a loss and give the portfolio time to recover.

Pfau and Finke described cash value as not exposed to interest rate risk and capital losses, which is why they called it a buffer asset. EY likewise treated it as a fixed income substitute. Neither paper argues that whole life beats stocks — that is a different and unanswered question. The argument is narrower: that guaranteed cash value may serve the specific role bonds are supposed to serve in a retirement portfolio, more reliably, in the years when reliability matters most.

The Tax Math for High-Bracket California Households

Compensation disclosure aside, the tax treatment of policy cash value is where the math gets specific enough to be useful or irrelevant depending on the household.

Cash value growth accumulates without an annual tax event — there is no 1099 on the growth, unlike a taxable bond. Withdrawals up to the cost basis come out without tax. After basis is exhausted, income is accessed by policy loan, which under IRC §72(e) is not treated as a distribution for a non-modified-endowment-contract policy, meaning the proceeds are not taxed as ordinary income. Policy loan proceeds are not included in MAGI for Medicare's income-related surcharge (IRMAA) — while taxable bond interest and even municipal bond interest both count toward IRMAA income. That distinction matters for retirees near the IRMAA thresholds.

The fair comparison converts a policy's internal rate of return into a tax-equivalent yield: divide the tax-free return by one minus the combined marginal rate. For a California household in the 32% federal bracket with the 9.3% state rate and 3.8% net investment income tax (NIIT applies once MAGI exceeds $250,000 for joint filers), the combined rate is 45.1%. At that bracket, a 4.0% policy IRR requires a taxable corporate bond yielding 7.29% to keep pace after tax. At the top California bracket — 37% federal, 13.3% state, and 3.8% NIIT, totaling 54.1% — the required corporate bond yield rises to 8.71%.

Those figures use the policy's IRR, not its dividend rate. A dividend interest rate is not a policy's return because the cost of insurance and expenses come out first. The internal rate of return on cash value, taken from the policy illustration at the year income would begin, is the right input. California does not tax Treasury interest, so Treasuries carry a lower break-even than corporates at any given bracket. In-state municipal bonds are exempt from both federal and California tax, so their yield compares directly to the policy's IRR; the advantage shifts elsewhere — to IRMAA treatment and to the fact that muni prices can fall, as the 2022 index return showed.

These calculations apply to taxable accounts. Bonds held inside an IRA or 401(k) are already tax-deferred, and the tax advantage of a life policy largely disappears in that comparison.

What the Models Assume — and What They Can't

Both studies deducted investment fees before comparing. Pfau and Finke deducted a 0.84% fund cost plus a 0.75% advisory fee — 1.59% per year in total. EY charged a 1.25% combined advisory and investment management fee. The investment alternative in both models is already paying realistic costs.

What neither model accounts for is behavioral deviation. The investor in both studies follows the plan perfectly: buys, holds, rebalances, and never sells in a panic. Academic research finds that real investors earn less. Dichev, writing in the American Economic Review in 2007, compared dollar-weighted returns — based on when money actually moved in and out — with buy-and-hold returns across major markets. The gap was 1.3% per year for NYSE and AMEX stocks from 1926 to 2002, and 5.3% per year for NASDAQ from 1973 to 2002. Later academic reviews found the magnitude of the gap is debated and varies by period and methodology. EY raised the question in both directions — noting that do-it-yourself investors tend to lag the market, but listing the size of that lag as an open question rather than a finding.

The honest framing: the behavior gap is relevant because the buffer strategy exists specifically to remove the need to sell in a down year. That is a reason to expect the buffer to help real investors, not just modeled ones. It is not a measured result, and we do not present it as one.

The Early Years, and Who This Fits

Everything above describes a policy with years of funding behind it. A new policy behaves differently.

Cash value is premiums collected minus expenses and charges. Because costs are front-loaded, it builds slowly, and surrendering in the early years can return substantially less than what was paid in. State nonforfeiture laws require that built cash value not be forfeited if premiums stop, but the plan built around the policy is gone. No one in either study bought a policy on the eve of retirement — Pfau and Finke modeled couples buying at 35 and 50 and retiring at 65; EY modeled starting ages of 25, 35, and 45. A buffer has to be built before it is needed.

Policy design also shapes the early curve. How much of the premium goes to a paid-up additions rider versus the base policy changes how quickly early cash value accumulates, within the limits set by the seven-pay test under IRC §7702A. Exceed those limits and the policy becomes a modified endowment contract, and the tax treatment changes: distributions — loans included — are then taxed gain-first, with a 10% additional tax on the taxable portion before age 59½.

Who this may fit: someone with a genuine need for permanent coverage, the cash flow to fund premiums for many years, a high combined marginal tax rate, and bonds held in taxable accounts that could be replaced — with a long runway before income begins.
Who this likely does not fit: someone close to retirement with no policy already in place, anyone who may need liquidity in the first years, anyone who cannot commit to the premium schedule, and anyone whose bonds sit inside an IRA or 401(k) where the tax comparison largely disappears.

A disclosure that belongs upfront rather than at the bottom: Applied Wealth Management is a fee-based fiduciary with an affiliated insurance agency. If we recommend a policy, a commission is paid. That is exactly why the math should be on the table — not instead of the conversation, but as the conversation itself.

Common Questions

How does whole life cash value function as a retirement income buffer?

The strategy draws income from the policy's cash value in any year the investment portfolio posts a negative return, and from the portfolio in all other years. Because guaranteed cash value cannot decline by contract, the policy provides income without forcing the sale of depressed portfolio assets. The portfolio is left intact to recover. Pfau and Finke modeled this rule mechanically across 10,000 Monte Carlo simulations and found it improved sustainable withdrawal rates and median income at retirement in their case studies, though results depend on assumptions including a healthy insured, decades of premium funding, and dividend scales that hold.

Why does the tax comparison matter most for high-income California households?

Taxable bond interest is ordinary income, subject to federal rates, California state income tax, and — above the applicable MAGI threshold — the 3.8% net investment income tax. At a combined rate of 45.1% (32% federal, 9.3% California, 3.8% NIIT), a 4.0% policy IRR would require a corporate bond yielding 7.29% to produce the same after-tax income. At California's top combined rate of 54.1%, that break-even rises to 8.71%. The right input is always the policy's internal rate of return from the actual illustration — not the dividend interest rate — compared against the bonds a household would actually hold, in a taxable account.

What are the main risks and limitations of this strategy?

Dividends are not guaranteed and are declared annually by each insurer's board. Guarantees are subject to the claims-paying ability of the issuing insurer. Cash value builds slowly in the early years, and surrendering early can return less than premiums paid. Loan interest accrues and compounds; an outstanding loan reduces the death benefit and, if the policy lapses with a loan outstanding, the gain becomes taxable income. In strong markets, buy-term-and-invest produced larger legacy values in Pfau and Finke's results. And putting too much into insurance — EY found allocations above 30% to 50% of savings could reduce income — may redirect too much away from equities. The strategy requires underwriting, a long funding horizon, and a premium commitment the household can sustain.