In late April, Bank of America’s Michael Hartnett told clients that 2026 asset price action is starting to resemble what he saw in mid-2007 and mid-2008. He used the word allocators have spent two years avoiding out loud: subprime. He flagged tremors. The note got passed around quietly.
Most CIOs read it and went back to their meetings.
The ones who didn’t are now asking a different question. It is the question that will define this cycle, and it will not be answered by the sectors people are still arguing about. It is this: when every line item in a portfolio sells off at the same time, what was the diversification ever actually for?
This article is about what most institutional alternatives sleeves now look like under the hood. The answer is uncomfortable. In a “Sell America” environment, the implications are worse.
The 60/20/20 Story Most Allocators Are Telling Themselves
After 2022, the institutional consensus quietly walked away from 60/40. The new framework was a 60/20/20, sixty percent public equities, twenty percent fixed income, twenty percent alternatives. The 20 was meant to be the part of the book that did not move with everything else. The hedge.
Every major asset manager built a thesis around this. Apollo, Ares, Blackstone, KKR, and Carlyle, the five largest listed private markets firms, now manage roughly $1.5 trillion in perpetual capital between them. Around forty percent of their combined AUM, up from thirty-five percent in 2021. That capital was sold to LPs as the answer to the simultaneous-decline problem of 2022.
The story made sense at the time. It made less sense once you traced where the capital actually went.
The Plumbing Nobody Drew on the Slide
In April, Eileen Appelbaum at the Center for Economic and Policy Research published the chain. Private equity firms had spent the prior seven years acquiring life insurance and annuity businesses. They redirected policyholder reserves into proprietary private credit funds with thin disclosure and infrequent marks. Those funds deployed capital into PE-owned portfolio companies. A material concentration sat in mid-market application software businesses bought during the period of peak enterprise software valuations.
So the policyholder, who thought they owned an annuity, was actually a senior creditor to a software roll-up.
Now layer the AI capex on top. Goldman Sachs estimates $539 billion in AI capital expenditure for 2026 alone. Cumulative AI infrastructure spend is projected to hit $3 trillion by 2028. Morgan Stanley puts the share financed by private credit at roughly half. Hyperscalers issued more than $100 billion in bond paper in the last six months. Data center developers are sourcing capital from the same private credit vehicles that institutional LPs allocated to as their uncorrelated sleeve.
This is not a theoretical concern anymore. The Bank for International Settlements published the numbers in March: software stocks fell almost thirty percent between October 2025 and February 2026. Business development company stock prices fell about ten percent on average. Discounts to net asset value deepened on private credit vehicles whose books are largely opaque illiquid loans. Software, where AI is doing most of its disruption, accounts for about seventeen percent of BDC investments by deal count.
The Morgan Stanley note was specific. Default rates in private credit direct lending could surge to eight percent, well above the historical 2-2.5% range. Pressure concentrated, predictably, in sectors most exposed to AI displacement.
The Definition of Uncorrelated, Re-checked
Allocators have stopped using the word correlation precisely. It got reduced to “low beta to the S&P 500” and circulated in factsheets.
The proper definition is not statistical. It is mechanical. An asset is genuinely uncorrelated when the underlying force that drives its returns does not respond to the same forces that move the rest of the portfolio.
Apply that test to the typical 2026 alternatives sleeve.
Private equity returns hinge on entry-to-exit equity multiples. Equity multiples respond to rates, growth expectations, and risk appetite. Same forces driving public equities.
Private credit returns hinge on borrower cash flows, default probability, and refinancing conditions. All three are tightening simultaneously right now. Same forces.
Real estate, even before you account for the data center concentration, is rate-sensitive duration with a vacancy overlay. Same forces.
Hedge funds, in aggregate, increasingly closet-track equity factors. Same forces.
Even gold and crypto have surprised people in the “Sell America” trade. Bitcoin fell 4.8 percent in a single February session when the White House announced a tariff escalation, briefly retesting $64,000, not because of anything crypto-specific, but because it had repriced as a US-linked asset. Gold rallied. The two trades cancelled each other philosophically. The portfolio still moved.
The list of major institutional asset classes whose return driver is not on that list is short. I would argue it is one item long.
The Asset Class Biology Won’t Let Get Caught
Life settlements are the secondary market for U.S. life insurance policies. An institutional fund acquires an in-force policy from its original owner, usually a senior who no longer needs or can afford it, pays the remaining premiums, and collects the death benefit at policy maturity. The return is determined by the timing of that maturity and the discount at acquisition.
The driver is actuarial. The cash flow is contractually backstopped by a US-licensed life insurer regulated at the state level. The maturity event is biological. None of those things care what Powell does in June. None of them care what tariff lands next. None of them care whether Nvidia hits its number. None of them care that BDC NAV discounts are widening.
Look at the 2018-2024 numbers honestly:
Modeled life settlement returns averaged 11.04 percent annually. Standard deviation of 3.8 percent. Sharpe ratio of 2.7. Correlation coefficient to the S&P 500 of -0.09. To high-yield corporate bonds: -0.02. To 7-10 year Treasuries: -0.10. These are not asset class adjacencies that move together with a small lag. They are operating in a different return universe.
Behavior in stress periods is more interesting than the long-run averages.
In March 2020, when the S&P 500 lost thirty-four percent in weeks and liquidity vanished, life settlement portfolios posted negligible volatility. There was nothing to mark down. Policies don’t have a panic price. They have a maturity date.
In 2022, when the 60/40 portfolio recorded its worst year since 2008 and equities and bonds fell together, life settlements held their ground. No correlation to the selloff. No liquidity-driven markdowns. No forced selling. Just the steady accrual of yield.
In February 2026, when the “Sell America” trade hit every dollar-denominated asset across the spectrum simultaneously, life settlements continued to mature on actuarial timelines. Not on CNBC’s schedule.
The Old Market vs. The New One
I have spent enough time around skeptical institutional allocators to know the two real reasons life settlements were not on their books five years ago.
The first was the underwriting problem. Traditional life expectancy providers used manual review, actuarial tables, and eight to twelve weeks of work per file. Variance in outputs across different providers on the same policy was uncomfortably wide. If you cannot price the asset accurately, you cannot construct a diversified institutional portfolio at scale.
That barrier has collapsed, and it has collapsed in a measurable way. Machine learning models trained on longitudinal clinical datasets, combined with transformer-based embeddings on unstructured medical records, are now producing probabilistic life expectancy estimates with interquartile ranges of two to three months across repeated runs. That is not a marketing claim. It is a measurable validation metric, Continuous Ranked Probability Score, concordance index, calibration analysis, that an institutional investment committee can audit. The technology shrinks both timing risk and pricing dispersion in a way that makes diversified portfolio construction tractable.
The second was the ethics problem. The “betting on death” framing made it allocator-uncomfortable to put on a diligence list. The framing was always wrong, but it was emotionally salient and it stuck.
Here is what actually happens. Roughly eighty-eight percent of life insurance policies in the United States never pay a death benefit. They lapse, or they are surrendered to the carrier for a fraction of their value. In 2023, over $700 billion in face value was lapsed or surrendered. Only about $4.7 billion was settled. Sellers in the secondary market receive three to five times the cash surrender value the carrier offers. That difference, hundreds of millions of dollars annually, flows to seniors during their lifetime, often funding healthcare, long-term care that has risen approximately sixty-nine percent in cost since 2004, or the retirement gap that affects four out of five Americans.
The ESG question, properly stated, is not whether life settlements are ethical. It is whether continuing to leave eighty-eight percent of policies to lapse worthless is ethical. The default outcome enriches the carrier. The settlement outcome enriches the policyholder. The regulatory framework, forty-six states covering more than ninety percent of the US population, NAIC model laws, the 1911 Grigsby v. Russell Supreme Court precedent, is mature and tested.
When an institution’s investment committee asks the right Social and Governance questions of a life settlement manager, quantified policyholder uplift versus surrender value, competitive bidding evidence, STOLI exclusion process, insurable interest verification, data privacy controls, independent fiduciary architecture, the answers are typically easier to provide than for most private credit funds.
The Window
Institutional capital is moving. Eighty-three percent of institutional investors are either invested in life settlements or actively considering an allocation. Forty-four percent plan to increase allocations in the coming year. Apollo, Blackstone, KKR (via Global Atlantic), Berkshire Hathaway, and PIMCO already participate. Goldman Sachs, J.P. Morgan, and Morgan Stanley provide financing and diligence support to transactions.
Current US market penetration sits at 0.7 to 2.5 percent of the available policy universe. The structural supply runway is the seventy-three million American baby boomers aging into the 65-85 corridor where life settlements become economically relevant. Face value lapsing annually exceeds face value being settled by a factor of roughly one hundred and fifty.
This is what early-cycle institutional adoption of an asset class looks like. It is what private credit looked like in 2013-2015. It is what real estate looked like before 1980s pension reform. The narrative was identical: niche, illiquid, complex, misunderstood. A decade later, each of those became a core institutional allocation generating hundreds of billions in committed capital, and, importantly, the alpha compressed.
The investors who built positions in those windows did so because they answered one specific question correctly: when allocations are still small, when supply is structurally vast, when correlations are mathematically demonstrable rather than narratively asserted, that is the window. After that, the asset class re-prices and the early-mover yield disappears.
The Question for May 2026
Hartnett’s note will get re-read when 2026 prints its full data. The 60/20/20 will be revisited. Some of the larger private credit vehicles will mark down further. LPs will demand explicit AI-displacement-risk disclosures in writing. Redemption queues will lengthen.
The CIOs reading this who already run a careful diligence process know what comes next. The portfolio review will identify three or four positions that were sold as uncorrelated and are not. The re-underwriting will look for what is actually decoupled.
If the right question is “what in my portfolio is biologically determined rather than economically determined,” there is a very short list of candidates.
It is not gold. It is not commodities. It is not a new macro hedge fund strategy.
It is life settlements. And the window to build a real position at current penetration levels is measured in months, not years.
About the author
Yaniv Bertele is Founding Partner & CEO of EverOak Innovations, an institutional life settlements fund applying proprietary AI-powered underwriting to build portfolios with target net IRRs of 14-15% and near-zero correlation to traditional asset classes. He writes regularly on alternative investments, longevity yield, and portfolio construction.
Reach: [email protected].
Disclosure: The author is the founder of an investment firm operating in the asset class discussed. This article is for informational purposes only and does not constitute investment advice, an offer, or solicitation. Past performance and modeled outcomes are not indicative of future results.
