Carve-outs are value creation drivers on both sides of the deal. Counter-intuitive, but the data supports it.
The seller unlocks capital tied up in a non-core asset, redirects management bandwidth to the businesses that fit the strategy, and usually sees a multiple expansion on the remaining portfolio. EY-Parthenon's analysis of enterprise software divestments (Jan. 2025) shows the parent companies outperformed the S&P Global 1200 IT Index by 9% TSR over the two years following the divestiture. The stock goes up because of the carve-out, not despite it.
The buyer, typically PE, applies the focus, capital, and operating discipline that the corporate parent could not. Bain's data through 2012 showed PE carve-outs delivering 3.0x MOIC against 1.8x for the average buyout -- carve-outs were the highest-returning category in the asset class. Top-quartile carve-outs still deliver 2.5x MOIC today, but average carve-out MOIC has fallen from 3.0x pre-2012 to 1.5x since 2012.
Both sides of the deal can win. The data says they should. But many don't. Why?
Alvarez and Marsal name four causes:
- A lack of clear information at the point of signing a deal.
- The difficulty for buyers of assessing separation costs and phasing.
- The pressure to get the deal done without coming across as 'difficult'.
- The architecture of a typical auction sales process.
The first three are Buyer problems, the fourth is a Seller dynamic that leads to the other three.
Carve-outs have become an exercise in allocating positive and negative value-enhancements among Buyer and Seller. It's an imbalanced trade. The Seller is incentivized to allocate people, tools, tech debt and liabilities to the carved out asset, while maximizing retention of IP and key people that enhance value. The Buyer needs to accept this and build around that thesis. Or not, and walk away.
The Buyer mis-pricing largely hangs on the false assumption that the Seller is transacting a clean asset.
Consequently, sell-side bankers drive carve-out auction processes on extremely aggressive timelines. They are excellent at the art of choreographing the release of information to compress the cycle time of factual understanding and technological depth. Bidder pools are carefully structured. Headline price thresholds are established. Risk transmission is spelled out in pre-determined interim operating covenant terms and TSA architectures. The standalone operating model is designed by the Seller without regard for how a Buyer wants it. The table is set, but there are Buyer leverage points too.
By the time the carve-out auction is running, the deal architecture has already decided where the value goes.
Buyers have an advantage because much of the diligence has been done for them -- operators within the Sell-side company have conducted internal diligence to land at the optimum model for them. You just need to uncover it. The Seller likely won't offer it, but there's no disincentive for them to provide it either because the table has been set.
What the architecture actually does
Sell-side carve-out and Buy-side carve-out integration processes are similar, just a different landing. The work breaks into four pieces.
Standalone operating-model design is the work of building the carved-out business as a standalone before the auction and preparing materials to articulate it to the Seller. Org structure, leadership, finance function, IT estate, vendor contracts, real-estate footprint, shared services. The covenant package reflects that. The price reflects that. The Buyers job is to force n-levels down discovery, so they can map it to their integration plan and operating model.
Separation & Integration planning is the function-by-function review of what comes with the carved-out business and what stays. Legal. Tax. HR. IT. Finance. GTM. Product. The default Seller answer -- "we'll figure it out in the TSA" -- is how Buyer covenant exposure compounds and incremental Seller value accrues. The operator answer -- on both sides -- is a written model that names every shared service, every shared system, every shared contract, and every shared person, with a cutover plan for each. For the Buyer forced into a time box, integration planning is diligence.
TSA architecture is the contract that supports the buyer's operations for six to eighteen months after close. Most TSAs are written as back-stops. They should be written as operating contracts interwoven with the fabric of the Buyers integration plan. The architecture work is granular: which services, at what cost basis, with what SLAs, with what exit criteria, with what step-down schedule. Done right, the TSA reads like a service agreement. Done wrong, it reads like a liability schedule. A properly priced TSA provides a mutually incentivized framework where the Seller wants to stop servicing it and the Buyer wants to stop paying for it.
Removing financing friction by providing good-faith financials -- like a high-quality QOE, carve-out adjustments documented, separation costs scoped -- enable Buyer underwriting that does not disrupt the aggressive carve-out auction timetable. If Bidders can't fully model the business quickly, especially those who have to assemble debt, will price the financing risk into the discount. If enough bidders run into this challenge, the price threshold may not be breached. Similarly, Buyers should pre-socialize a broader M&A thesis with their banks prior to going to market so that they are able to model the asset within their bank's underwriting window. The architecture on both sides removes the friction and allows focus on other transaction details.
How the architecture work shows up in the returns
The MOIC collapse is the simplest way to see it. Pre-2012, PE carve-outs delivered 3.0x MOIC against 1.8x for the average buyout. Then: the carve-out discount was wide enough that even mediocre architecture work produced strong returns. Now: average carve-out MOIC has fallen to 1.5x. The discount got bid away by competition, and sponsors stopped delivering the operational improvements they once did -- revenue growth under PE ownership fell from 31% pre-2012 to 17%, and margin expansion fell from 29% to 2%. The architecture work actually had to deliver.
Top-quartile carve-outs still hit 2.5x. The average has stopped working.
The Alvarez and Marsal four causes map back to the architecture cleanly. Causes one through three -- unclear information at signing, difficulty assessing separation costs, pressure to close without pushing back -- are buyer-side symptoms that the architecture work fixes from the buyer side.
Operator-led integration planning surfaces what the diligence book hides. A standalone operating model built n-levels down kills the assumption that the asset is clean. Pre-socialized financing replaces deal-clock pressure with an underwriting window. Cause four -- the architecture of the auction process -- is the seller-side dynamic that creates the tension of the first three, and it goes away when the seller does the architecture work too. The two sides don't trade off. They compound. A seller who hands the buyer a properly architected asset gets a tighter bid range. A buyer who shows up with integration planning treated as diligence captures more of the standalone value post-close.
Both sides need an operator-architect. Most don't, because PE diligence still gets staffed financial and legal, with operations bolted on at the end. The transaction is treated as a spreadsheet exercise. The architecture is treated as someone else's job. The 2.5x top quartile reflects what happens when at least one side does the work, and both sides win when both do.
The work is unglamorous. The architecture is what separates the top-quartile carve-out from the average. On the sell side, the price holds. On the buy side, the MOIC delivers. On both sides, the architecture is the difference. The data has been telling us this for over a decade. Most of the asset class has not been listening.