Joint venture, partnership, or just a contract — how to pick the right one

A useful working definition, from the Legal Information Institute: a joint venture is "a combination of two or more parties that seek the development of a single enterprise or project for profit, sharing the risks associated with its development". The same source adds the caveat that matters here — a joint venture "is not a partnership or a corporation, although some legal aspects of a joint venture (such as income tax treatment) may be ruled by partnership laws". A joint venture is a commercial description. It is not, by itself, a legal form.

The three structures, and what each one costs

The honest default is the left-hand column. A great many arrangements described as joint ventures are one company buying something from another with an unusual pricing model, and a well-drafted commercial agreement handles that better than a venture structure does. Reach for a JV when the collaboration creates something neither party would own alone — and where the answer is a JV company, the document that actually governs it is the shareholder agreement between the parents.

The unintended partnership is the real risk

Partnership statutes do not wait to be invoked. Delaware's, a clean enactment of the revised uniform act, provides that "the association of 2 or more persons to carry on as co-owners a business for profit forms a partnership, whether or not the persons intend to form a partnership". It adds a presumption that a person receiving a share of the profits of a business is a partner in it, subject to a list of exceptions — repayment of debt, wages, rent, interest on a loan, the price of goodwill — none of which covers "we agreed to split what was left".

If that test is met, two further rules arrive with it. Each partner is an agent of the partnership, so an act by either party apparently carrying on the venture's ordinary business binds it. And partners "are liable jointly and severally for all obligations of the partnership". Your collaborator's commitment becomes your personal exposure, for the whole amount rather than for your share.

If the venture is genuinely a business the two of you will run together rather than a project you will finish, stop resisting the label and write a proper partnership agreement — or put a company between yourselves and the liability.

Choosing between the three

Five questions get most decisions to the right answer.

• Will the venture own anything? Premises, equipment, a jointly developed product, a brand. Assets need an owner, and co-ownership between two companies is a dispute waiting for a trigger. Assets point to a JV company.
• Will it employ anyone? If people are to be hired by the venture rather than seconded from the parents, it needs to be an employer, which means an entity.
• Does a third party need one counterparty? Customers, lenders and regulators frequently need a single legal person to contract with, licence, or lend to. That is the most common reason a contractual JV is upgraded.
• How long is it? A defined project with an end date suits a contract. Something indefinite accumulates shared assets, shared staff and shared goodwill until an entity becomes the tidy answer anyway.
• Is external investment coming? Investors buy shares. They do not buy a position in a contract.

Two "no"s and a short duration means a contract. Two or more "yes"es means a company. The middle — a contractual JV — is for collaborations that are more than a supply relationship and less than a business: joint bids, co-marketing, shared development where each side keeps its own IP.

What a contractual JV has to do that a company does automatically

A JV company arrives with governance, a bank account, a balance sheet and a share register built in. A contractual JV has none of those, so the agreement has to supply them in writing.

• Governance. A named steering group, how often it meets, what a majority decides, and the reserved matters that need both parties.
• Money. Who invoices the customer, where the money lands, how costs are evidenced, when distributions are made, and who audits the numbers.
• Authority. An express statement that neither party may bind the other, incur liabilities in its name, or hold itself out as its agent, with an indemnity if it does.
• Liability to the outside world. Either each party contracts with the customer for its own scope, or one is prime and subcontracts the other back to back. Both work. What does not work is leaving it unstated and discovering the answer during a claim.
• Intellectual property. Three buckets: what each side brings, what the venture creates, and what each side may do with all of it afterwards. Background IP stays put and is licensed for the venture's purpose only; foreground IP needs an owner named now.
• Confidentiality. Mutual, and drafted for a situation where both sides are simultaneously partners and potential competitors — see mutual versus one-way NDAs.

Tax does not read your disclaimer

For US federal tax purposes, a domestic eligible entity with two or more members is classified as a partnership by default if it makes no election otherwise. So a JV formed as an LLC is taxed as a partnership unless the members affirmatively choose corporate treatment — filing a partnership return and issuing each member a Schedule K-1 showing their share of income or loss.

Whether a purely contractual arrangement is itself a separate entity for tax purposes is a factual question, and the clause in your agreement saying "nothing in this agreement creates a partnership" governs the relationship between the parties rather than the classification. It is a cheap question to put to an accountant at the outset and an expensive one to answer after two years of unfiled returns.

Deadlock, and the exit you have to write while everyone is still happy

Fifty-fifty ventures are chosen because they feel fair and they are the structure most likely to seize. Every JV agreement needs a route out of a tied vote, and the choice of route says a lot about the balance of power.

• Escalation. The first step and the one that resolves most of it: the matter goes to the chief executives of both parents, who have to meet within a stated period before anything else happens.
• Expert determination. For technical or valuation disagreements, an independent expert decides and both sides are bound. Fast and cheap compared with the alternatives, and inappropriate for strategy.
• Casting vote. One party, or an independent chair, breaks ties — sometimes only on defined categories. Clean, and only acceptable where the parties are genuinely unequal.
• Buy-sell, sometimes called a shotgun. One party names a price; the other must either buy at that price or sell at it. Elegant, and it favours the party with more cash, so it belongs between parties of similar size.
• Termination for deadlock. The venture unwinds on an agreed basis. The point of writing it down is that the basis is agreed rather than fought over.

Alongside deadlock, list the events that let either side leave: change of control of the other parent, insolvency, failure to fund an agreed contribution, material breach uncured after notice, and failure to hit defined milestones by defined dates. Then say what happens to the shared assets, the customer relationships, the jointly developed IP and any brand — because those are the four things a departing party will want and the four things that will be argued about.

Before you sign

The pattern across all of this is that joint ventures fail commercially far more often than legally. What the agreement can do is make the failure orderly: an agreed way to disagree, an agreed way to value what has been built, and an agreed way to leave. Negotiating those three things at the start costs a fortnight. Negotiating them at the end costs the venture.