The Disciplines of M&A
The three families of valuation method and where each breaks, what diligence is actually looking for, and what the evidence says about whether acquisitions create value.
Mergers and acquisitions rest on three analytical disciplines that run in sequence. Valuation establishes what a business is worth and what a buyer can afford to pay. Due diligence tests the assumptions inside that valuation and converts what it finds into price, structure or withdrawal. Post-merger integration determines whether the value assumed at signing is realised. Valuation technique is highly developed and diligence well institutionalised; integration is where the largest share of expected value is lost.
Intrinsic valuation derives value from the cash a business is expected to generate, discounted at a rate reflecting the risk of those cash flows. The discounted cash flow (DCF) model is the canonical form. Its assumptions about growth, margin, capital intensity and risk are explicit and contestable, but the output is only as good as a forecast extending beyond any credible horizon.
Relative valuation infers value from prices at which similar assets trade or have changed hands. Trading comparables use market prices of listed peers and produce a minority, marketable value; precedent transactions use prices paid in completed deals and so embed control premiums and whatever synergies previous buyers paid for. The two answer different questions.
Transaction-specific methods derive value from the constraints of a particular buyer or exit. The leveraged buyout or ability-to-pay analysis solves backwards: given assumed debt capacity, holding period and required equity return, it computes the maximum entry price consistent with that return — what a sponsor can pay, not what an asset is worth. Sum-of-the-parts values each division on its own terms and aggregates, netting central costs; it suits conglomerates and tests whether a break-up creates value. Liquidation value estimates proceeds from an orderly or forced asset sale and functions as a floor in distress.
The standard enterprise DCF discounts unlevered free cash flow — earnings before interest and after tax, plus depreciation and amortisation, less capital expenditure and the increase in working capital — at the weighted average cost of capital (WACC), keeping financing effects out of the numerator so they are handled once, in the discount rate.
WACC weights the after-tax cost of debt and the cost of equity by target capital structure at market values. The cost of equity is conventionally estimated using the capital asset pricing model: a risk-free rate plus beta multiplied by an equity risk premium, sometimes with size or country premia. Beta is estimated from historical returns over a chosen window and is unstable, the equity risk premium is not observable and varies materially by method, and the target capital structure is a judgement rather than an observation.
Terminal value handles the period beyond the explicit forecast. The perpetuity growth (Gordon) method capitalises a normalised final-year cash flow at the discount rate less a constant growth rate; the exit multiple method applies a multiple, usually EV/EBITDA, to terminal-year earnings. The perpetuity growth rate cannot exceed long-run nominal growth in the economy where the business operates without implying the firm eventually outgrows that economy; the risk-free rate is often used as a practical ceiling, since it embeds expected long-run nominal growth.
The share of value sitting in terminal value is the most important diagnostic in a DCF. For a five-to-ten-year forecast with moderate growth it routinely accounts for most of computed enterprise value, and for high-growth assets the large bulk of it, so most of the answer rests on a period nobody has modelled in detail.
The mid-year convention discounts cash flows as though received evenly through the year rather than at year end, raising value by roughly the square root of one plus the discount rate. Sensitivity tables flex one or two variables, typically WACC against terminal growth or exit multiple; scenario analysis instead changes a coherent set of operating assumptions together, which is more informative because real downside cases involve growth, margin and capital intensity deteriorating at once.
A multiple is a compressed DCF: EV/EBITDA can be derived algebraically from a perpetuity model, rising with expected growth and cash conversion and falling with risk and capital intensity. Two companies trade at different multiples because the market takes different views of those variables, not because one is arbitrarily "cheap".
The general-purpose multiples are EV/EBITDA (capital-structure neutral and insensitive to depreciation policy, but blind to capital intensity), EV/EBIT (captures depreciation, so penalises capital-hungry businesses, but sensitive to accounting policy), EV/Revenue (usable when earnings are negative, but assumes comparable margins) and P/E (an equity multiple, contaminated by leverage and non-operating items).
Sector conventions exist where the general multiples fail:
What makes comparables not comparable is a longer list than most screens acknowledge: accounting policy (revenue recognition, capitalised development costs, lease treatment), business mix within a nominal sector, growth and reinvestment needs, geographic and regulatory exposure, and position in the cycle. Precedent transactions add further problems — they reflect the financing conditions of their moment, and disclosed multiples often rest on synergised rather than standalone earnings.
Most multiples and the DCF produce enterprise value: the value of the operating business, attributable to all capital providers. Converting to equity value requires a bridge, and that bridge is where much of the negotiation occurs. The core adjustment is net debt — gross financial debt less cash and equivalents. Beyond it the bridge captures minority interests, deducted at fair value because consolidated EBITDA includes all of a partly owned subsidiary; associates and joint ventures, added back at fair value because equity-accounted profits are excluded from EBITDA; and pension deficits, treated as debt-like net of deferred tax, with judgement required over whether to use the accounting, funding or buyout measure, which can differ by large multiples.
Enterprise value is not what the seller receives
Leases now sit in the bridge explicitly. IFRS 16, effective for annual periods beginning on or after 1 January 2019, requires lessees to recognise a right-of-use asset and a lease liability for substantially all leases over twelve months, so obligations previously in the notes appear on the balance sheet and rent splits between depreciation and interest, raising reported EBITDA. ASC 842 also brings operating leases onto the US balance sheet but retains a single straight-line operating expense in the income statement, so US-reported EBITDA is not lifted the same way. Multiples must therefore capitalise leases in both enterprise value and earnings, or in neither, and cross-border comparables need normalisation.
Other debt-like items include unfunded restructuring and environmental provisions, deferred and contingent consideration from prior acquisitions, factored receivables, disputed tax liabilities and abnormal payables; surplus assets and non-operating property are cash-like. Whether an item is debt-like or a working capital item has no fixed technical answer. It is negotiated, which is why the definitions of net debt and normalised working capital are among the sale agreement's most contested clauses.
Control confers the ability to change how a business is run — capital structure, cost base, asset portfolio, management. Aswath Damodaran's treatment makes its value situation-specific: control is worth the difference between the firm's value optimally run and as currently run, multiplied by the probability of change. Control is therefore worth much in a badly managed company and little in a well-run one, and a uniform percentage premium is a category error.
Empirically, target announcement returns are large and have grown. A survey of the takeover literature reports them rising from roughly 6 per cent in the 1960s and 1970s to about 16 per cent in the 1990s, around 24 per cent in the 2000s and about 29 per cent in the 2010s. These are abnormal returns, not offer premiums to an unaffected price, which are quoted higher because they measure against a pre-speculation reference. The minority discount is the arithmetic converse of the control premium, not an independent observation.
Illiquidity discounts for private companies are more contested. The traditional bases — restricted stock studies comparing registered and unregistered shares of the same issuer, and pre-IPO studies comparing private transaction prices to subsequent offering prices — produce very wide ranges and are criticised for small samples, selection bias and weak controls. A discount is defensible in principle; a specific percentage drawn from them is not well supported.
Evidence on synergy realisation is asymmetric. A McKinsey analysis of around 160 mergers found roughly 60 per cent delivered planned cost synergies close to in full, while about a quarter overestimated them by at least 25 per cent. Almost 70 per cent of mergers in the same database failed to achieve expected revenue synergies, and across 124 mergers with usable data the combined entity typically lost 2 to 5 per cent of the combined customer base. Cost synergies — headcount, procurement, facilities, duplicated overhead — are estimable and largely within management control; revenue synergies depend on customer behaviour and are systematically over-forecast.
Who captures synergies is settled by the price. If a buyer pays a premium equal to the present value of expected synergies, the target's shareholders capture all of it and the buyer bears all the execution risk — the mechanism behind the finding that targets do well and acquirers do not.
Accretion/dilution analysis computes pro forma earnings per share for a public acquirer after financing costs, foregone interest on cash used, new share issuance and purchase accounting effects, and compares it to standalone EPS. It is universally prepared and a poor proxy for value creation, depending mechanically on the relationship between the target's earnings yield and the after-tax cost of consideration: a high-P/E company can issue stock to buy almost any lower-P/E business and report accretion whether or not the deal creates value, while a value-creating acquisition of a fast-growing, loss-making business will be dilutive. It measures the arithmetic of one year's exchange, not the economics of the combination over its life.
A fairness opinion is a letter from a financial adviser to a board stating that, as of its date and subject to stated assumptions and qualifications, the consideration is fair from a financial point of view to a specified group of holders. It does not say the price is the best obtainable, does not opine on the merits of the transaction against alternatives, does not address allocation among classes or the fairness of management compensation unless expressly stated, and is not a voting recommendation. FINRA Rule 5150 requires member firms whose fairness opinions reach public shareholders to disclose whether the firm advised any party, whether compensation is contingent on completion, material relationships with the parties over the prior two years, whether company-supplied information was verified, and whether a fairness committee approved the opinion.
The literature is consistent on the distribution of gains. Andrade, Mitchell and Stafford's survey in the Journal of Economic Perspectives (2001) reports a positive announcement response for the combined parties — roughly 1.5 per cent in the 1970s and 2.6 per cent in the 1980s — and improved post-merger operating performance relative to industry peers. Targets capture most of that gain.
What targets capture has risen for sixty years
Acquirers on average do not gain. Robert Bruner's survey of 114 studies published between 1971 and 2001 concludes that bidders, with interesting exceptions, earn approximately zero adjusted returns. Moeller, Schlingemann and Stulz (2005) find acquiring-firm shareholders lost 12 cents for every dollar spent on acquisitions announced between 1998 and 2001, $240 billion in total, against $7 billion (1.6 cents per dollar) across the whole of the 1980s. The loss was concentrated: 87 transactions, 2.1 per cent of deals, accounted for $397 billion, and excluding them acquiring shareholders would have gained $157 billion.
The size effect is among the field's most robust findings. Moeller, Schlingemann and Stulz (2004), studying 12,023 completed acquisitions by US public acquirers from 1980 to 2001, find small acquirers earn announcement returns roughly 2.24 percentage points higher than large ones (about 2.32 per cent versus 0.08 per cent), holding across payment methods and across public, private and subsidiary targets.
Method of payment matters. Cash-financed acquisitions outperform stock-financed ones for acquirers at announcement and over long horizons, usually explained by signalling — managers issue stock when they believe it overvalued — and reinforced by the long-run underperformance of stock acquirers found by Loughran and Vijh (1997).
The listing effect interacts with both. Fuller, Netter and Stegemoller (2002), examining 3,135 bids by 539 frequent acquirers between 1990 and 2000, find acquirer five-day abnormal returns of −1.00 per cent for public targets, +2.08 per cent for private and +2.75 per cent for subsidiaries; within public targets, stock offers produced −1.86 per cent and cash offers were indistinguishable from zero.
Serial acquirers show declining returns with experience: the same study reports about 2.74 per cent on first bids falling to about 0.52 per cent by the fifth and subsequent. This sits in tension with consultancy findings — McKinsey's analysis of the largest 2,000 global companies over more than two decades finds "programmatic" acquirers, making frequent small deals, delivered around 2 percentage points more in annual excess total shareholder returns than peers, while organic, selective and large-deal approaches produced none on average. The academic work measures announcement reactions to single deals and the consultancy work long-run returns to a strategy, so the divergence is better treated as open than resolved.
Alexandridis, Antypas and Travlos (2017) find post-2009 deals create more value for acquiring shareholders than in any prior period studied — acquirers gaining roughly $62 million around announcement of mega-deals of $500 million or more, an improvement of about $325 million on 1990–2009, with combined synergistic gains above $542 million — and that stock-financed deals no longer destroyed value, which they attribute to improved acquirer governance.
Financial diligence tests the quality and sustainability of reported earnings, the adequacy of working capital, the split between maintenance and growth capital expenditure, and the run-rate position entering the forecast; it is the workstream most directly connected to price. Commercial diligence tests the market outside the company — size and growth, competitive position, pricing power, customer concentration, contract terms and renewal rights, churn and net revenue retention, and pipeline credibility.
Legal diligence establishes title to assets and shares, reviews material contracts for change-of-control and assignment provisions, and identifies litigation and contingent liabilities. Tax quantifies historical exposures, tests the survival of losses after a change of control, reviews transfer pricing and permanent establishment risk, and shapes the structure. Technology and IT assesses the application estate, technical debt, scalability, key-person dependency and the cost of separation or integration. Cyber security tests the control environment, breach history, third-party exposure and notification obligations. HR and benefits covers employment terms, pension obligations, incentive plans and their change-of-control treatment, consultation requirements and retention risk. Environmental covers contamination, remediation and permits; insurance, cover, claims history and run-off; ESG, reporting obligations, supply chain and human rights exposure, and transition risk; regulatory, licences, sector approvals, merger control and foreign investment screening. IP establishes ownership and validity of registered rights, chain of title from contractors and employees, freedom to operate and, increasingly, open-source licence compliance and the provenance of training data.
The quality of earnings (QoE) exercise produces an adjusted EBITDA bridge: a reconciliation from statutory reported EBITDA to a normalised, run-rate figure a buyer would capitalise. The adjustment categories are conventionally non-recurring items (restructuring, transaction fees, one-off settlements, insurance recoveries); normalisation (owner compensation above or below market, non-arm's-length related-party transactions, rent on vendor-owned property); accounting corrections (cut-off errors, capitalisation policies inconsistent with the buyer's, inadequate provisioning, changes in estimate); pro forma and run-rate items (annualising price increases, contract wins and losses, and part-year acquisitions or disposals); and standalone costs a divisional target will incur but has not historically borne.
The QoE number rather than the reported number becomes the price, because the transaction multiple is applied to it. The bridge is therefore negotiated line by line: at a 10x multiple, a disputed £1 million adjustment is a £10 million argument. Buyers also scrutinise the direction of adjustments — a bridge composed almost entirely of upward add-backs invites the question of whether "non-recurring" items recur.
Diligence runs through a virtual data room controlled by the seller, with staged access, question-and-answer logs and management presentations. In auctions, sellers often commission vendor diligence reports to standardise what bidders see.
Buy-side reports take two forms. A red-flag report is a rapid, exception-based review covering only issues above a materiality threshold, common at an early bid stage. A full-scope report covers the agreed scope completely, including matters found satisfactory. Reports are issued to defined addressees: a reliance letter extends the adviser's duty of care to a third party, typically a lender or co-investor, usually subject to a liability cap; a non-reliance letter confirms a recipient takes the report for information only, with no right of action.
Findings translate into outcomes along four routes. They can move price, most often through the QoE bridge or a revised net debt or working capital calculation. They can change structure — a share purchase converted to an asset purchase, deferred consideration, an earn-out bridging disagreement on forecasts, or an escrow. They can be met with contractual protection — a specific indemnity for a quantified exposure, a warranty for the unknown, or warranty and indemnity insurance. Or, where a finding undermines the investment case rather than merely pricing it, they lead to a walk. Findings that end deals typically invalidate a core assumption: undisclosed customer losses, revenue recognition that misstates growth, unresolvable title or IP defects, and regulatory exposures of uncertain magnitude.
Day one readiness is what must be true at completion for the combined business to trade legally and safely: legal entity and governance in place, employees paid, customers able to order and be invoiced, suppliers paid, regulatory permissions transferred, reporting able to run, and communications delivered to staff, customers, suppliers and regulators. Day one is a threshold, not an achievement.
The integration management office runs the programme: a central team with a defined mandate, functional workstream leads, an escalation route to a steering committee, control of synergy tracking, and authority to settle disputes quickly. The 100-day plan sequences the decisions everything else depends on: operating model and organisation design at least two levels down, brand and go-to-market approach, product roadmap rationalisation, systems strategy and location footprint. Delay here is a commonly cited cause of value loss, because uncertainty drives attrition among exactly the people the acquirer needs.
Synergy tracking must reconcile to the deal model. Each synergy needs an owner, a baseline, a phasing, milestones and a line in the management accounts where it becomes visible; without that, reporting drifts into initiative counting that cannot be tied to earnings. Costs to achieve — severance, systems, advisers, property exits — belong in the same tracking, since they routinely consume the first year or two of savings.
Retention of key people turns on retention arrangements, role clarity and rapid appointment decisions. Systems and data consolidation is generally the longest workstream: ERP, CRM, HR and finance consolidation frequently takes years, and data migration and master data reconciliation are consistently underestimated. Customer and channel overlap requires deliberate handling of accounts served by both parties, channel conflict and exclusivity. Culture is often treated as soft but shows up in decision rights, approval norms, risk appetite and pay systems — all observable and open to explicit reconciliation.
The claim that 70 to 90 per cent of acquisitions fail is among the most repeated statistics in business, and its provenance is weak. The most-cited modern instance is the 2011 Harvard Business Review article "The Big Idea: The New M&A Playbook", which states that "study after study puts the failure rate of mergers and acquisitions somewhere between 70% and 90%" without identifying a single study. McKinsey material has likewise asserted that roughly 70 per cent of mergers fail, with no citation in the text. The figure circulates by repetition, and the studies occasionally invoked differ so radically in what they measure — announcement returns, long-run stock returns, accounting performance, executive self-assessment, subsequent divestment — that no single rate can be derived from them.
What the better evidence supports is narrower. Targets capture most of the gain and acquirers on average close to none — a statement about distribution rather than failure. A small number of very large deals account for a disproportionate share of aggregate value destruction. Synergy categories fail at measurable rates: roughly 70 per cent of mergers in one consultancy database missed expected revenue synergies, while cost synergies were largely delivered in about 60 per cent of cases. And work using explicit operating criteria does reach high failure rates — Baruch Lev and Feng Gu report, from an analysis of some 40,000 acquisitions over 40 years using a model with 43 variables, that 70 to 75 per cent fail to achieve stated objectives of post-acquisition sales growth, cost savings or share price maintenance, and that the rate has risen over time. That is a defensible finding attached to a stated definition, which the free-floating claim lacks.
Causes of shortfall are consistent across sources: overpayment on optimistic synergy forecasts, particularly revenue; loss of key employees and customers during prolonged uncertainty; delayed decisions on operating model and organisation; underestimation of systems integration cost and duration; and diligence that priced the business without testing whether the integration assumed in the price was achievable.
Carve-outs are harder than whole-company acquisitions because the target does not exist as a standalone entity until it is made to. Stranded costs are the parent's costs remaining after a division departs — shared services capacity, IT licences, property and central overhead previously allocated to the divested unit. They fall on the seller and are a real transaction cost often omitted from proceeds analysis.
Transitional service agreements (TSAs) let the seller continue providing services — payroll, IT hosting, procurement, finance processing — for a defined period at defined prices. Scope must be specified in enough detail to be enforceable, service levels and exit assistance defined, pricing agreed in advance including for extensions, and both parties given incentives to exit on time. A business still dependent on a TSA when it expires has no supplier.
Standing up a standalone finance function is the characteristic difficulty. A divested unit typically lacks its own general ledger, statutory accounts, treasury and banking arrangements, tax registrations, insurance, audit relationship and reporting capability. Carve-out financial statements prepared for the sale rest on allocations and are not the accounts the standalone business will produce, which is why standalone cost estimates are a central negotiating issue and a frequent source of disappointment after completion.