All podcasts / Invest Like the Best / Summary

Vlad Barbalat - Investing $120 Billion in Permanent Capital

2026-06-23 - 71 min - source - Read full transcript

Patrick O'Shaughnessy (host)Vlad Barbalat

Key insights

Owning permanent, single-balance-sheet capital removes the business-strategy distortions that come from managing third-party fund capital.

Barbalat argues that fund managers are ultimately running a business - they have to think about fundraising cycles, investor updates, and the multiple the market will pay for their manager - and that inevitably dilutes the pure craft of investing. Because Liberty Mutual Investments answers only to its own balance sheet and policyholders, it can sustain what he calls better 'investment hygiene': no reacting to investor circumstances, no fund-cycle pressure.

permanent-capital

The insurance float model works because protecting risk on one side of the balance sheet funds growing the economy on the other.

Liberty takes in premiums to protect individuals and businesses, then invests that float (the Buffett playbook) into infrastructure, entrepreneurs, and job creation. Barbalat frames this as a genuinely unique position in the economy: doing risk protection and capital formation simultaneously rather than as separate businesses.

insurance-as-investment-platform

Liberty's $120 billion splits into roughly $70-75 billion of tightly managed reserves and the remainder in growth credit and growth equity that scale with surplus capital.

The reserve portion must always be able to fulfill policyholder promises, but Barbalat says even that piece is managed actively (acting as a liquidity provider) rather than passively holding investment-grade bonds to maturity. The growth pools (credit and equity, each with several sub-businesses like direct lending, private equity, real estate, energy/infrastructure, alternative credit) are the more flexible, opportunistic capital.

insurance-as-investment-platform

The choice of access format (direct, co-invest, club deal, or LP) matters as much as the choice of exposure itself.

Once LMI decides what economic exposure it wants, it treats the access route as a second, equally deliberate decision: direct origination when it has genuine expertise, LP allocation when a risk is too specialized to replicate in-house, club/co-invest structures in between. Barbalat says this optionality is a structural advantage most institutions lack because they're typically locked into one access model.

permanent-capital

'Branded capital' is built by being fast and creative, not by writing the biggest check.

Barbalat distinguishes LMI's brand from mega-LPs like large state pensions that win allocation purely on check size. Liberty instead hires people with GP/operator backgrounds, moves quickly to a yes/no so it doesn't waste originators' time, and tries to be the LP a GP wants on its capital roster the way Yale's endorsement signals quality to other investors - while also being willing to take risks that brand-name institutions typically won't.

permanent-capital

A mutual (vs. public) insurance structure is a deliberate choice to forgo the shareholder discipline that forces most public insurers to stay conservative investors.

Barbalat argues public insurer shareholders generally want a narrow, consistent underwriting margin and their capital back via dividends/buybacks, not a shadow asset-management build-out - a structure he calls a 'conglomerate' problem shareholders resist. Because Liberty is mutual, it lacks that external forcing function, but has chosen internally to hold itself to the same high bar on both underwriting and investing.

insurance-as-investment-platform

Diversified, long-tail insurance books require a fundamentally different balance sheet than narrow, short-tail insurers.

Comparing Liberty to Progressive, Barbalat notes Progressive's motor-focused book is not particularly long-tailed, while Liberty's commercial and specialty lines can generate claims decades later, requiring 'fatter tails' of balance sheet strength. This is why he frames underwriting and investing as a mutually reinforcing flywheel rather than separate functions.

insurance-as-investment-platform

AI is producing a genuinely new kind of valuation uncertainty: not about macro variables, but about which businesses will even exist in ten years.

Barbalat says he's never before questioned equity and credit multiples based on not knowing which companies will survive - previously multiples moved with inflation or rates, not existential doubt about business survival. He extends this to companies that seem 'outside the AI crossfire' (Home Depot, John Deere) and to credit duration (comfortable with 4-year Salesforce/Oracle paper, uneasy holding 30-year paper on the same names).

ai-and-valuation-uncertainty

Private markets grew not from prestige-seeking but because they solved the capital-availability problem while public markets got structurally more costly to inhabit.

Barbalat traces the historic reason for going public - needing capital unavailable privately, plus the prestige of a public listing - and argues both have eroded: private markets can now raise enormous sums, and public markets impose compliance costs and quarterly pressure that many businesses can't tolerate if they need a genuine three-to-five-year operating horizon. He expects this balance to persist even if some regulatory burden reverts.

public-vs-private-markets

Equity exposure should be decided first; public vs. private is a secondary access decision, not a separate asset-allocation lever.

He explicitly rejects moving allocation between public and private equity as a tactical dial - fundamentally, equity risk is equity risk. Liberty's own preference for private equity exposure is described as balance-sheet-specific (public equity doesn't suit LMI's structure) rather than a market-timing call.

public-vs-private-markets

Growing up as a persecuted minority in the Soviet Union instilled a permanent sense of non-entitlement that now shapes his investing and leadership approach.

Barbalat describes explicit, socially normalized antisemitic persecution in Soviet Moldova (quotas on professions, being called out in school) and contrasts it with the individualism and agency available in the U.S. He connects this directly to an investing culture at LMI built around never assuming you're entitled to a deal, a career, or a result, and always pushing to make things better.

immigrant-perspective-on-risk

Heavy reliance on AI risks displacing the 'messy' human relationships that generate investing insight, and Barbalat has no resolved answer for that tradeoff.

He describes AI as a genuine superpower for editing and sharpening thinking (warning that taking the first AI output uncritically is 'where slop tends to live'), but flags a real worry: the more time he spends interacting with AI, the less he spends with colleagues, and he doesn't know how that isolation dynamic resolves for investing, which he sees as fundamentally a human, relationship-driven craft.

ai-and-valuation-uncertainty

Media referenced

Colossus - other - Positive Sum's quarterly print, digital, and private audio publication that Patrick plugs at the open and close of the episode.

Companies

Liberty Mutual Investments - The $120 billion investment platform Barbalat runs as CIO, sitting inside Liberty Mutual Group's insurance balance sheet.
Liberty Mutual Group - The parent mutual insurer; personal lines (Liberty 'jingle', home/auto) plus a global commercial and specialty insurance business feed the investment platform.
Berkshire Hathaway - Reference point for the insurance-float-funds-investing model; Barbalat contrasts LMI's more programmatic diversification with Berkshire's outsized, one-off underwriting bets.
Geico - Cited as Berkshire's most visible insurance brand and example of the insurer-of-last-resort posture a fortress balance sheet enables.
Goldman Sachs - Barbalat and much of his senior team came from Goldman; he credits it with instilling a 'drive for excellence' culture he brought to Liberty.
Progressive - Used as a contrast case: a highly successful but narrowly focused (U.S. motor) insurer whose balance sheet needs differ sharply from Liberty's diversified, longer-tail book.
Salesforce - Central example in the AI-and-valuation discussion: even if every enterprise keeps using Salesforce, the real risk is whether the next generation of trillion-dollar companies will ever adopt it.
Home Depot - Named as an example of a business that seems outside the obvious 'AI crossfire' but may still face the same multiple-compression question.
John Deere - Same point as Home Depot - a seemingly AI-insulated business whose long-term multiple is now less certain.
Oracle - Used alongside Salesforce as an example of long-duration software credit that Barbalat says he'd be warier of holding at 30-year tenor than at 4-year tenor.

Techniques and frameworks

Permanent capital / no third-party LPs - Because Liberty Mutual Investments deploys its own insurance balance sheet rather than fund capital, Barbalat says it avoids the 'investment hygiene' problems that come from managing other people's money on a fund cycle.
Branded capital - Barbalat's term for cultivating a reputation as a fast, creative, GP-like LP that GPs actively want on their capital roster, rather than competing purely on check size.
Fortress balance sheet - The strategic goal of building enough capital strength (via disciplined underwriting plus strong investment returns) to absorb risks - like data-center-scale assets - that smaller insurance balance sheets can't.
House view built on exposures, not forecasts - LMI explicitly avoids trying to predict macro outcomes (overweight Europe, etc.); instead it defines the long-term business exposures it wants and picks the best access route for each.

Summary

Vlad Barbalat, chief investment officer of Liberty Mutual Investments, walks Patrick O'Shaughnessy through what it actually means to run a $120 billion balance sheet that has no third-party investors. The core thesis of the conversation is that permanent capital - money that answers only to Liberty's own policyholders rather than to fund LPs on a raise-invest-return cycle - changes the investing craft at a structural level. It removes the business-strategy pressures (fundraising, investor updates, multiple management) that Barbalat argues inevitably dilute even excellent fund managers' investment process, letting LMI maintain what he repeatedly calls "investment hygiene."

Barbalat breaks down the $120 billion into roughly $70-75 billion of reserves (actively but conservatively managed to guarantee policyholder promises) and a growth pool split between credit and equity businesses - private equity, real estate, energy/infrastructure, alternative credit, direct lending, and public/private corporate credit - organized around exposure rather than product. A recurring theme is that LMI treats "what exposure do we want" and "how do we best access that exposure" (direct, co-invest, club deal, or LP) as two separate, equally deliberate decisions, an optionality he says most institutions don't have. He also details LMI's "branded capital" strategy: rather than competing on check size like a mega state pension, Liberty aims to be the fast, creative, GP-like partner that originators want on their capital roster - hiring operators rather than traditional LP staff and moving quickly to a yes or no.

The conversation moves into a substantial discussion of AI's effect on valuation. Barbalat says he's never before faced a question like the current one: not "should multiples move with macro variables" but "do we even know which businesses will exist in ten to fifteen years." He extends the uncertainty beyond obvious software names to businesses like Home Depot and John Deere, and into credit duration - comfortable holding four-year Salesforce or Oracle paper, much less comfortable at thirty years. He separately voices an unresolved personal worry: heavier AI use is displacing time he'd otherwise spend with colleagues, and he doesn't know how that tradeoff plays out for a craft he still sees as fundamentally human and relationship-driven.

A significant portion of the episode is biographical. Barbalat recounts growing up Jewish in Soviet Moldova, experiencing explicit, normalized persecution (professional quotas, being singled out in school) before his family emigrated to the U.S. in 1990. He ties that experience directly to an investing and leadership philosophy built on non-entitlement - the idea that nothing, from a deal to a career outcome, is owed - and to Liberty's internal culture of continuous improvement, which he compares to Americans' habit of endlessly reinventing something as simple as a croissant rather than accepting "bread is bread."

The episode closes on public-versus-private markets and on the meaning of permanence itself. Barbalat argues private markets grew not from prestige but because they solved the capital-availability problem while public markets became structurally more costly and short-horizon; he treats equity exposure as the primary decision and public/private access as secondary. On permanence, he pushes back on the idea that permanent capital is simply "nice to have," noting that individuals and teams within a permanent platform are not themselves permanent, and that the real discipline is holding both the three-to-five-year strategic horizon and the one-year stakeholder reality at once, sustained through transparency: "no transparency, no autonomy."

Notable Quotes

"We are not in the business for predicting the future. We're in the business being prepared for all its eventualities." - Vlad Barbalat

"If you just ask for something and you get it back, it will give you generalities, and it will drive everything to kind of an average, right? That's what these models are. So in order to get the best out of them, you need to engage with your knowledge, your experience, your ideas and creativity." - Vlad Barbalat

"You're literally saying the future is so unpredictable that how could I possibly place some higher multiple on something?" - Vlad Barbalat

"No transparency, no autonomy." - Vlad Barbalat

"The craft of investing is inherently diluted one way or the other. It just is." - Vlad Barbalat