Buy at Eight, Trade at Thirty: The Arithmetic of Serial Acquirers

In the first half of 2026, Sweden’s Lifco spent 928 million kronor buying companies and acquired roughly 110 million kronor of annual operating profit in the process. That is a price of about nine times earnings. Lifco’s own shares change hands in Stockholm at about 28 times. An entire business model lives in the gap between those two numbers — along with the question of whether that gap is value creation or an accounting entry.

Serial acquirers are the quietest success story in public equities of the past two decades. They are holding companies that do nothing but buy small, dull, highly profitable niche businesses, leave the existing managers in charge, and recycle the cash into the next purchase. No synergies, no integration, no restructuring. Just capital allocation, repeated a thousand times. Lifco compounded operating profit at 22.5 percent a year from 2001 through 2025. Constellation Software of Canada has returned roughly 150 to 170 times the original investment since its 2006 listing — depending on the measurement date — while keeping its share count essentially flat at around 21 million.

Which is exactly why the second look matters. In August 2026, Constellation trades about 35 percent below its high while its cash flow runs at record levels. The market has not written off the machine. It has repriced what the machine is worth. That distinction is the subject of this analysis, and it determines whether an investor makes money in this asset class or merely admires a model.

The number everything depends on, and it is on no balance sheet

The mechanics are disarmingly simple. A serial acquirer buys a family-owned business earning five million dollars of operating profit at five to nine times that figure. The acquirer itself trades at 25 to 35 times earnings. The instant the purchase agreement is signed, those five million dollars of profit move from a world where they are worth 25 to 45 million into a world where they are worth 125 to 175 million. The difference is not created by work. It is created by relocation.

The unromantic name for this is multiple arbitrage, and it is the real engine — even if almost no chief executive describes it that way. In the early life of a serial acquirer it is overwhelming, because the base is small: a company with a 200 million dollar market value that buys 20 million dollars of earnings has transformed itself.

Current valuations across the best-known names show how wide the spread is today, and how differently the market ranks the individual houses.

Company	Forward price/earnings, 2026	Operating profit growth, 2001–2025 p.a.
Addtech	35.2	16.7%
Lagercrantz	31.5	13.9%
Lifco	28.8	22.5%
Indutrade	25.9	13.1%
Röko	22.4	founded 2019, no comparable record

Against those figures sit entry prices of typically five to nine times earnings for the businesses being bought. The spread is enormous. But it does not belong to the acquirer. It is on loan.

The 28 or 35 times that Lifco or Addtech commands is not profit the company earned. It is an expectation the market extends — specifically, the expectation that the next decade of acquisitions will work as well as the last two did. That expectation can be withdrawn at any moment without a single portfolio company performing worse. Which is precisely what happened to Constellation Software in 2026. An investor in this asset class buys the machine and pays for confidence in it. Only the first item appears in the accounts.

Why reported earnings systematically mismeasure these companies

There is scarcely another corporate form where the headline profit number tells you less. The reason is purchase price allocation. When a company is acquired, the price must be split across identifiable assets — customer relationships, technology, trade names — and those items are then amortized over their estimated useful lives. What is left becomes goodwill, which is touched only on impairment.

Roper Technologies, from its 2025 acquisitions alone, recorded roughly 460.9 million dollars of goodwill and 210.4 million dollars of other identifiable intangibles — customer relationships with a 16.7-year weighted average life and technology at 7.5 years. That 210.4 million will now run through the income statement for a decade and a half without any further cash ever leaving the building. Danaher expects 1.7 billion dollars of such acquisition-related amortization in 2026 and reports it as a separate reconciling item.

The effect is structural. The more active a buyer is, the wider the wedge grows between reported profit and actual cash generation. A company that simply stops acquiring will look dramatically better on an earnings basis within three years without earning a dollar more.

Constellation Software turned 2025 into the definitive case study. The numbers appear to contradict each other outright.

Constellation Software, fiscal 2025	Figure	Change vs. 2024
Revenue	$11,623m	+15%
Organic growth	4% (3% in constant currency)	–
Net income to common shareholders	$512m	−30%
Cash flow from operations	$2,732m	+24%
Free cash flow available to shareholders	$1,683m	+14%
Consideration paid for acquisitions	$1,579m	–

Diluted earnings per share fell from $34.48 to $24.15 while operating cash flow rose by a quarter. Anyone reading only the first line saw a business in decline. The actual driver was the revaluation of a membership liability tied to subsidiary Topicus’ stake in Poland’s Asseco: because the value of that investment rose, the corresponding obligation had to be marked up — roughly $155 million of expense in the fourth quarter and about $225 million for the year. No cash, no operating deterioration, just accounting. An investment that appreciated produced earnings that fell.

Where adjusting stops being legitimate

None of which means adjusted figures should simply be believed. It means the opposite: an investor has to know which adjustment to accept and which to refuse.

Excluding amortization of acquired customer relationships is defensible so long as the buyer does not have to keep replacing those relationships. For niche software with a 96 percent renewal rate, that holds — the relationship does not wear out, it is merely treated as though it does. For a distribution business whose customers must be won again every seven years, it does not hold at all. Danaher itself notes in its own disclosures that these intangibles contribute to generating sales and that amortization from past acquisitions will recur. That is not a footnote; it is the operating manual.

The practical test is easier than it sounds and rests on three questions. First, does free cash flow grow at a similar rate to adjusted earnings over a multi-year window? Second, how does capital spending compare with depreciation on physical assets — if a buyer claims to be asset-light, this is where it must show. Third, have there been goodwill impairments in the past decade? An impairment is the retrospective admission of having overpaid. A house that has never taken one either bought well or is not looking closely.

The one number you cannot buy

Acquisitions manufacture growth that says nothing about quality. Enough money buys any revenue growth you like. That makes the single most informative metric for a serial acquirer the one thing it cannot purchase: organic growth, meaning the performance of the businesses it already owns.

Constellation reported 4 percent organic growth for 2025, 3 percent in constant currency. In the second quarter of 2026, organic maintenance and recurring revenue growth excluding Altera was again 4 percent in constant currency — below the historical trend, which management attributed to accounting normalization at Altera, comparison distortions at Dark Matter, recent acquisitions at Lumine, and a large customer loss in South America. Indutrade has struggled to grow operating profit organically for several quarters. Lifco chief executive Per Waldemarson has summarized the whole industry in a sentence: completing an acquisition is one thing, developing the company afterwards is the actual trick.

The judgment that follows is blunt. An acquirer with 2 percent organic growth and 20 percent total growth is not a compounder; it is a fund with operating expenses. It earns its return exclusively from the valuation spread, and that spread disappears the moment the market declines to grant it. An acquirer running 5 to 7 percent organic growth has a return that survives a year of buying nothing. That is the real dividing line in this asset class, and it runs through every portfolio rather than between countries.

The law of large numbers, or the decay rate built into the model

Every serial acquirer carries a brake that grows with its own success. The deals that built the model are small — companies with two to twenty million dollars of operating profit, often family-owned, with no investment bank running a process, and therefore cheap. Those same deals eventually become too small to matter.

Constellation had to place roughly $1.58 billion of purchase consideration in 2025, followed shortly after year-end by commitments of a further $802 million. In the second quarter of 2026 alone, total consideration reached $893 million. A deal contributing five million dollars of profit moves nothing at an $11.6 billion revenue base. So the acquirer must either do more deals, which reduces the diligence per deal, or larger deals, for which competition exists and prices are higher. Both routes lower the return.

Constellation’s answer has been notable: spin-offs. Topicus.com and Lumine Group were separated into independently listed vehicles. What was decentralized there was not the operations — those already were — but the capital allocation problem. Each unit gets a small base again, where small acquisitions are once more meaningful, and its own market price against which the valuation spread can be harvested afresh. It is elegant. It is also an admission that scale is a disadvantage here. Anyone investing in a serial acquirer should therefore carry one figure in mind: how much capital must this company deploy every year, and does its hunting ground actually contain enough targets of that size?

The American variant went up-market, and that is the whole story

North American investors often reach for Danaher and Roper Technologies as the domestic version of the model, and the comparison is instructive precisely because both companies stopped doing what made them famous. Danaher built its reputation on a business system applied to acquired industrial companies; today it is a large-cap life sciences and diagnostics group whose deals are measured in billions. Roper spent a decade deliberately shifting from niche industrial products into vertical market software, and its 2025 purchases — Spectrum AI, the legal technology assets of HerculesAI, and Valuation Pricing Director — are platform additions rather than the steady drip of small niche businesses.

That drift is not a failure of management. It is arithmetic. Once a company needs to deploy billions a year, the small-deal pond is simply too shallow, and the only remaining pool is the one where private equity, strategic buyers, and auction processes already operate. The consequence for investors is that the American names should not be valued on the logic of the Nordic ones. They are large-cap quality industrials with an acquisition history, and their future return depends far more on the organic performance of what they already own than on any spread between entry and exit multiples.

This also explains a pattern that puzzles newcomers: why the highest multiples in the sector sit on comparatively small Swedish companies rather than on the far larger American ones. The market is not paying for size. It is paying for the runway — for the belief that a company small enough to still buy small things has years of the arbitrage left.

Lifco and Storskogen: identical strategy, opposite outcome

Advocates of the model like to point to Stockholm, where an entire species of these companies has evolved. They usually omit that the most prominent counter-example is listed in the same city.

Storskogen went public in October 2021, raising 13 billion kronor at a pre-money valuation of 56 billion, and rose 30 percent on its first day. The shares now trade roughly three-quarters below the offer price. Net debt went from about three times operating cash flow across 2019 to 2021 to nine times in 2022. In the second quarter of 2024 came an impairment of approximately 920 million kronor, of which some 600 million was goodwill.

Lifco and Storskogen pursue the same basic idea. The difference sits in three variables that are observable in advance.

Distinguishing feature	The durable pattern	The brittle pattern
How deals are funded	predominantly from operating cash flow	from debt and equity issuance
Leverage	stable within a stated range, including in booms	rises with the pace of acquisition
Pace	steady across cycles	accelerates when money is cheap
Price discipline	fixed return hurdle; deals get walked away from	entry prices rise alongside the buyer’s own multiple
Target size	small, repeatable, unnoticed	grows with the capital requirement
Organic growth	consistently disclosed and positive	hidden behind acquired growth

What follows for valuation is telling. When Röko listed on Nasdaq Stockholm in March 2025 at 2,048 kronor per B share and a total value of roughly 29.95 billion kronor — founded by Tomas Billing, former chief executive of Nordstjernan, and Fredrik Karlsson, the former head of Lifco — Swedish institutional investors were conspicuously reluctant. Not because the model was misunderstood, but because it was understood too well. Storskogen had already supplied the price of getting it wrong.

What decentralization actually buys, and where it fails

The second pillar of the model is usually discussed as a culture question. It is really a risk decision. A classic conglomerate integrates: shared administration, shared purchasing, shared brand. That produces cost advantages — and chain reactions. The Nordic-style serial acquirer does the opposite. The acquired business keeps its name, its management, its site, and its customer relationships. Head office supplies capital, a reporting framework, and a return hurdle, and nothing else. Lifco ran a 22.4 percent operating margin in 2025 while consolidating 16 new businesses.

The benefit is loss containment: mistakes stay local, and no single failed acquisition drags down the group. The price is that no synergies exist to retroactively justify an inflated purchase price. You have to buy correctly the first time, because there is no second chance.

And there is a limit that is rarely named. Decentralization works so long as the portfolio companies serve structurally stable niches. It fails when a technological shift hits every subsidiary at once. A federation of forty autonomous niche vendors holds exactly zero advantage over an integrated group in the face of an industry-wide dislocation — worse, it lacks the center capable of forcing a common answer. Anyone examining a serially acquired software portfolio today should ask less about how many companies it holds and more about how many of them face the same wind.

The six questions to ask before buying

What precedes yields a sequence that can be worked through in about an hour using nothing but public filings.

Question	Where it is answered	What counts as a warning
What is organic growth, disclosed separately?	quarterly report, growth bridge	not broken out, or persistently below 2 percent
How are acquisitions funded?	cash flow statement, financing section	recurring equity issuance or sharply rising net debt
How much capital must be deployed each year?	consideration paid versus free cash flow	purchase prices persistently exceed cash generated
Have there been goodwill impairments?	notes, impairment testing section	repeated impairments in the same segment
How wide is the gap between adjusted earnings and free cash flow?	non-GAAP reconciliation	the gap widens over several years
How much depends on one individual?	management history, succession arrangements	no visible second layer with a capital allocation mandate

The sixth is the most uncomfortable because it resists quantification. At Constellation Software, the name Mark Leonard stands for the method itself; the hurdle rate, the reporting culture, the refusal to dilute are his inventions. At Lifco it was Fredrik Karlsson, who now runs Röko — a personnel change that triggered an entire investor debate. Capital allocation can be encoded in process, but only up to a point. The remainder is judgment, and judgment has no succession plan.

Practical notes for investors outside Sweden

Access is the first obstacle and it is more than a technicality. Constellation Software trades primarily in Toronto; its United States over-the-counter line is thin, with wide spreads that can quietly cost more than a year of dividend yield on entry and exit combined. The Swedish names are frequently unavailable at American retail brokers altogether, and where available, often only through an OTC ticker with the same problem. For anything other than a small position, direct market access to Nasdaq Stockholm or the Toronto Stock Exchange is worth arranging deliberately rather than accepting whatever line a broker happens to offer.

Currency is the second. These are kronor and Canadian dollar earnings streams. Over a five-year holding period, the exchange rate can account for a meaningful share of the total return in either direction, and it is uncorrelated with anything discussed above. Hedging a single-stock position is rarely worth the cost, but the exposure should be a conscious choice rather than a surprise.

Withholding tax is the third and the most fixable. Sweden levies 30 percent on dividends; the United States treaty rate is 15 percent, obtainable by having a valid W-8BEN on file with the custodian, with the remainder generally creditable via the foreign tax credit. Because the Nordic acquirers deliberately retain most earnings for reinvestment, payout ratios are low and the absolute sums are small — which is exactly why the paperwork tends to get neglected and then matters more than expected once a position has compounded for a decade.

Bull and bear cases, cleanly separated

Rather than a recommendation, here are both cases in their strongest form, each with the condition it hangs on.

Case	The core assumption	Where it breaks
Bull	Europe faces a generational succession wave; hundreds of thousands of owner-managed businesses need a buyer who will not break them up. Serial acquirers are the only category of purchaser credibly offering permanent ownership, and are therefore able to pay less than financial sponsors.	if private equity and a new generation of acquirers discover the same pond and push entry prices from five to ten times earnings, the spread vanishes on the purchase side
Bear	The historical returns were earned in an era of falling rates and expanding multiples. At 4 percent organic growth and 30 times earnings, future returns are predominantly multiple risk rather than business risk.	if the best operators keep delivering — Lifco compounded profit at 22.5 percent across 24 years spanning several rate cycles and two severe recessions

Both cases share the same flaw in the other’s argument: they treat “serial acquirers” as a category. They are not one. Lifco and Storskogen pursued an identical strategy and are separated by several hundred percentage points of share price performance, and in both cases the reason was visible in the annual report years earlier.

What it comes down to

The model works. But it works for a different reason than the one usually given. It does not work because acquisitions create value — most acquisitions in recorded corporate history do not. It works because a particular kind of buyer reaches a particular kind of seller who does not want an auction, and because that buyer then refrains from breaking anything. Everything else is capital allocation, repeated to the point of tedium.

Three things are worth carrying away. First, reported earnings are the weakest available metric for these companies — Constellation Software demonstrated in 2025 how a 30 percent decline in profit and a 24 percent increase in cash flow can both be true at once. Second, the spread between what an acquirer pays and what it trades at is the engine, but it does not belong to the company; it is lent by the market and recalled without notice. Third, organic growth is the one number that cannot be bought, and therefore the only one that separates the durable from the merely energetic.

An investor who checks those three may end up owning the same shares as one who simply followed the story. But they will know what they are paying for — and, more valuable in a year like 2026, they will know what happened when the price falls 35 percent and nothing inside the business has changed at all.