Joint ventures are corporate marriages of convenience. They promise the best of both parents—capital from one side, technology from the other, local reach from a third, and a shared appetite for risk. On the slide deck, the math is irresistible: split the investment, pool the know-how, halve the downside. In practice, the structure often dulls incentives, complicates accountability, and slows decisions until the market changes faster than the partners can. Some joint ventures work, usually because one party truly leads or the mission is narrow and clockwork precise. Most of the rest end in buyouts, quiet wind-downs or litigation, and the reason is almost always the same: no one is paid to maximize the joint venture’s value as an independent business.

The incentive problem starts the day the ink dries. Managers seconded to the venture still report—formally or informally—to their home company. Their bonuses depend on metrics that rarely align perfectly with the joint venture’s P&L. A procurement chief from Parent A is nudged to favor Parent A’s suppliers; a sales lead from Parent B is careful not to cannibalize Parent B’s flagship product. Transfer pricing becomes a trench war disguised as a spreadsheet. Even when good people try to do the right thing, the human instinct is to avoid harming the hand that will write the next performance review. That skews everyday choices—what to build, whose IP to prioritize, which customers to call—away from the joint venture’s optimal path.

Governance is the next trap. The romantic ideal is the 50/50 venture: equal capital, equal votes, equal say. In the real world, 50/50 often means a permanent threat of deadlock and a permanent incentive to avoid hard calls. Capital allocation turns into an annual negotiation; product roadmaps bend toward the least objectionable option; sales territories are drawn to avoid offense. Cost control, especially in cyclical businesses, is chronically late because no one wants to be the partner who cut too deeply if the upturn arrives. When a downturn does hit, one parent sees the venture as strategic and the other sees it as expendable. The boardroom fills with the wrong kind of urgency.

Culture, time horizons and attention spans finish the job. Joint ventures thrive on senior attention, and senior attention is scarce. The executives who championed the deal may rotate out a year later, replaced by successors with fresh priorities. Cost and compliance creep in; top engineers drift back to the parents where career paths are clearer and stock awards richer. As market conditions change, the original logic frays. What began as a clever way to enter China or share a factory becomes a defense of yesterday’s model against new competitors or technologies. If there isn’t a clear, contractually reinforced “lead parent” empowered to make final calls, the venture spends its best years debating itself.

Still, some joint ventures do work—and spectacularly so. CFM International, the engine maker jointly owned by GE Aerospace and Safran, is the canonical success. It has endured for decades because the mission is narrow, the stakes are enormous, and the partners’ roles are clean. Jet engines are long-cycle programs with sky-high technical and financial barriers. The incentives align around reliability, product support and long-term service revenue, not around quarterly share grabs between the parents. Equally important, CFM has always had a de facto captain on program decisions, and its economics are designed so both sides win as the installed base grows. It feels more like a single franchise than a forced marriage.

Dow Corning was another long runner. For decades, Dow and Corning combined to dominate silicones with a structure that pooled chemistry, plants and distribution more efficiently than either could alone. The venture survived cycles and antitrust scrutiny in part because its boundaries were clear and each parent had plenty of other businesses to focus on, leaving the JV team to operate with autonomy. In time, the logic of full ownership prevailed, but that is a sign of success: a venture that created so much value that one partner eventually took it in-house.

Consumer goods offer quieter examples. Cereal Partners Worldwide—Nestlé and General Mills—divided geographies and capabilities in a way that limited overlap and internal rivalry. The point was not to invent a new cereal so much as to exploit distribution, brands and shelf-space economics in markets where one partner was stronger. Because the business is relatively stable and brand-driven, the costs of coordination are lower and the benefits of shared reach are tangible.

In market-access deals, automotive has produced winners and warnings. Western carmakers have long used Chinese joint ventures to build locally, share costs and navigate regulation. Many of those ventures delivered volume and profits for years because they combined global product platforms with state-backed distribution. Yet the same structures transferred know-how and supplier ecosystems that later empowered local competitors, and as the industry pivots to electric vehicles the old incentives misalign again. The strategic question becomes whether the venture is a bridge to a new platform or an anchor to an old one.

When joint ventures fail, they tend to do so in familiar patterns. Sony Ericsson, created to merge handset hardware and consumer electronics chops, looked sensible until the smartphone platform war rendered its middle-of-the-road strategy untenable. The parents had different views on software ecosystems, brand priorities and capital commitment; by the time a decisive bet was needed, the market had chosen for them. S-LCD, the Samsung-Sony display venture, suffered a different version of the same problem: as panels commoditized and scale economics tilted, it made more sense for one parent to own the assets outright than to argue over incremental capital and margins through a shared vehicle. Nokia Siemens Networks endured turbulent years of culture clash and cost cutting before ending up under one roof. Caradigm, the Microsoft–GE Healthcare venture, showed how hard it is to stitch together overlapping software and sales channels without a single product owner with the authority to say no.

None of these outcomes are mysterious. Joint ventures are inherently incomplete contracts. You cannot write down, in advance, how to react to every technological shift, pricing cycle or regulatory change over a decade. That is why control and residual claimancy matter. When one party has clear decision rights and the strongest claim on the upside from hard choices, the “incomplete” parts of the contract are filled by incentives. When control is split and the economics are muddled, gaps are filled by delay and politeness.

There are ways to improve the odds. The first is to avoid symmetrical control. The most successful ventures have a single captain—contractually empowered, not just informally respected. The second is to keep the mission narrow and the scope disciplined. JVs built around a single platform, plant or service tend to fare better than ones asked to be a mini-conglomerate. The third is to hard-code the critical economics: transfer-pricing bands, access to IP, service-level agreements and, above all, clean exit provisions. A pre-agreed path for one parent to buy out the other at a transparent formula turns many potential fights into negotiations over timing rather than existential battles. The fourth is talent. Ventures need their own bench and compensation plans that don’t make their best people feel like visitors waiting to go home.

Even with those guardrails, investors should treat joint ventures as transitional structures. They are bridges into new markets, laboratories to test technologies and training wheels for supply chains. As bridges, labs and training wheels, they can be excellent. As permanent homes, they are less convincing. The market moves; strategy shifts; one parent decides to double down while the other wants to de-risk. The venture can either adapt at the speed of the faster parent or slow to the speed of compromise. That choice usually determines the endgame.

So when a company touts a new JV, skip the slide that lists the synergies and ask five questions in plain language—who is actually in charge; how exactly do both parents get paid; where are the conflict lines with the parents’ existing businesses; what, specifically, will the venture stop doing that the parents can do better; and how does this end if it works. If the answers sound like shared steering wheels, shared wallets and an exit left to fate, history suggests you’re buying optionality with a side of friction. If the answers point to a single accountable leader, a narrow mission and an exit that’s already sketched out, you might have found the rare joint venture that behaves like a real company.

The paradox of joint ventures is that they are often launched to avoid hard decisions—build alone or partner, buy now or wait, commit fully or hedge. The ventures that succeed are the ones that are allowed to make hard decisions anyway. Everyone else eventually learns what the term sheet didn’t say: when incentives are split, value creation usually is too.