Shareholders' agreement: what it should include and how to draft it
Shareholders' agreement: what it should include and how to draft it
A shareholders' agreement is a contract among a company's owners that sets how they decide, how they share results and how someone leaves when the day comes. It is not paperwork for show: it is what keeps a company from breaking over a disagreement that could have been settled in one afternoon. Here you will find which clauses it needs, how it differs from the bylaws, what can be agreed depending on the corporate form, how to resolve a deadlock, and a worked example with three shareholders holding 50%, 30% and 20%.
The warning comes first: this is a content guide, not a template to sign blindly. A lawyer should review the final wording with the bylaws, the certificate of existence and prior agreements on the table.
What a shareholders' agreement is and what it is not
The shareholders' agreement is a contract. It governs three things: how shareholders deal with each other, how they make decisions and how one of them exits. Its force comes from the meeting of minds, not from registration. It binds those who sign it.
It is not a deed of incorporation and not an ornamental internal rulebook. Nor is it a tool for distrusting your partner: quite the opposite. It is signed while the relationship is healthy, precisely because that is when uncomfortable topics can be discussed with a cool head.
There is a simple test to know whether your company needs one. Ask yourself what would happen tomorrow if the majority shareholder sold his stake to an outsider, if one shareholder stopped showing up for work, or if two blocks tied on a critical decision. If the answer is that you have never discussed it, that gap is exactly what the agreement covers.
The difference from the bylaws and from the incorporation deed
The bylaws are the company's own rulebook. They are filed with the chamber of commerce, they are public and they are enforceable against third parties: anyone negotiating with the company can read them. The shareholders' agreement, by contrast, is a deal among the parties and binds those who sign. That difference is not a technicality; it changes who you can hold accountable.
The incorporation deed creates the company. The bylaws set its name, corporate purpose, domicile, capital, management and decision rules. The agreement steps in where the bylaws do not reach or where it is better that they do not: personal commitments among shareholders, valuation formulas, non-compete obligations, exit mechanisms.
Hence the practical rule: anything you want the company itself to respect, and any outsider to be able to verify, goes into the bylaws. Anything that is a commitment between people goes into the agreement. If a clause needs to be enforceable against the company or a future buyer, it must go into the bylaws and be registered. If it stays only in the agreement, you can pursue the shareholder who signed it, not the company.
Why a SAS allows more freedom to agree than an Ltda
In Colombia, a sociedad por acciones simplificada (SAS, a simplified joint-stock company) is incorporated by private document filed with the chamber of commerce. Its bylaws are more flexible than those of a sociedad limitada (Ltda, a limited liability company), and that flexibility lets many rules move into the bylaws that in other corporate forms would stay only in the agreement.
Why does it matter? Because a rule written in the bylaws becomes enforceable against the company, not only against the shareholder who accepted it. If you want share transfers to require prior approval, certain decisions to need a qualified majority, or restrictions on the circulation of stakes, the SAS gives you more room to write it into the document everyone reads.
In an Ltda the margin is narrower and the weight of statute is heavier: several rules that SAS shareholders can adopt freely are treated differently there. The practical conclusion is not that a SAS is better, but that the corporate form changes what can be agreed and where it belongs. That decision is made before incorporating, not after the problem appears.
Contributions, stakes and what happens if a shareholder does not pay in
The agreement must state who contributes what and how it is valued. Contributions are not only cash: they can be goods, equipment, premises, a customer portfolio or work. Each one needs a valuation rule and a deadline. Without that, each stake floats and any discussion about percentages turns into a discussion about perceptions.
In a SAS, contributions in services or industry can be agreed under their own rules. Other corporate forms treat them differently, so confirm the specific case before copying a formula from one company to another. What works in a SAS does not automatically transfer to an Ltda.
The point almost nobody writes down is default. If a shareholder promises to contribute and does not, the agreement must set consequences: a cure period, a reduction of the stake, loss of voting rights while the default continues, an obligation to sell to the other shareholders or compensation. Leaving that scenario silent is the fastest route to a lawsuit between people who once got along.
Governance: majorities, vetoes and who decides what
The central question of any agreement is who decides. It is not enough to say decisions are taken by majority: you must define majority for what. A simple majority for day-to-day operations is healthy, since it lets the team run the business without convening everyone for every minor purchase. But that same majority cannot be enough to sell the main asset or to take on debt above a defined amount.
| Decision | Majority | Who reaches it with 50 / 30 / 20 | What it prevents |
|---|---|---|---|
| Ordinary operations | Simple | The 50% alone, or 30% plus 20% | Day-to-day paralysis |
| Selling main assets | 75% | Requires 50% and 30% | One shareholder selling the company piece by piece |
| Borrowing above the agreed amount | 75% | Requires 50% and 30% | Debt that binds everyone without consensus |
| Changing the corporate purpose | 75% | Requires 50% and 30% | Pulling the company into another business |
| Admitting a new shareholder | 75% | Requires 50% and 30% | An unwanted outsider coming in |
| Matters on the veto list | Individual veto | Any of the three can stop it | Changes that hit only the minority |
Look at the third column: with a 75% threshold, 50% alone is not enough for anything important. The majority holder keeps control of operations and loses unilateral control over decisions that change the nature of the company.
The veto deserves its own paragraph. Used well, it protects the minority from having the rules changed behind their back; used badly, it turns any shareholder into a hostage. The answer is a short, closed list of veto matters and everything else by majority: requiring unanimity for everything sounds prudent and paralyses.
Right of first refusal, tag along and drag along
All three tackle the same problem from different angles: what happens when someone wants to sell. Without them, a shareholder can show up one day with a buyer and change the neighbourhood for everyone else.
The right of first refusal obliges the selling shareholder to notify the others before closing, so they can buy proportionally on the same terms. In the example, the seller gives 30 days' notice and the other two can take their share. It prevents an outsider from coming in without existing shareholders having had the option.
Tag along protects the minority: if the 50% sells, the minority may sell on the same terms. Without this clause, the majority holder walks out with the buyer and leaves the 20% sitting next to a new partner they did not choose.
Drag along does the opposite: if 75% accepts an offer, the rest must sell at the same price. It sounds aggressive, and it is, but it is what lets a buyer acquire 100% of the company instead of negotiating with three owners separately. A serious buyer rarely agrees to buy alongside a minority holder who can block the deal.
| Clause | What it sets with 50 / 30 / 20 | What it prevents | What it costs to agree |
|---|---|---|---|
| Right of first refusal | 30 days' notice; the others buy proportionally | An outsider coming in with no option for existing holders | Lengthens every sale and demands clear terms |
| Tag along | If 50% sells, 30% and 20% may sell too | The minority tied to a new partner | The buyer must buy more than it wanted |
| Drag along | If 75% accepts, everyone sells at the same price | A minority blocking a good sale | The minority loses a say on when to exit |
| Deadlock | 15 days of mediation, then forced purchase | A blockage with no exit date | Price and buyer must be defined in advance |
None of the four is free. All of them limit someone's freedom and all of them buy something in return: predictability. The useful debate is not whether they restrict, but whether the limit is worth it for this company's size and risk.
The exit clause and how the stake is valued
The exit clause says how a shareholder leaves. It must answer three questions: who buys, at what price and on what timeline. The most neglected is price, and it is precisely the one that decides whether the exit ends in agreement or litigation.
The valuation formula is chosen before you need it: a multiple of earnings, book value, an independent appraisal or a combination. Each produces different numbers and favours a different party. Book value favours whoever stays in a company with assets; the earnings multiple favours whoever exits when the business is profitable.
See the difference in round figures. Take a company with annual earnings of 300 million Colombian pesos and equity of 500 million. If the agreement sets a multiple of 3 on earnings, the value is 900 million and the 20% stake is worth 180 million. If it sets book value, the 20% is worth 100 million. The 80 million difference is not produced by business performance: it is produced by the formula that was signed.
That is why the clause must also cover what happens if the parties reject the appraisal, whether payment is in cash or instalments, what happens to retained earnings and when the clock starts. An exit clause without a price is an invitation to negotiate at the worst possible moment.
Deadlock: what to do when shareholders cannot agree
A deadlock is a blockage. Two shareholders of equal weight, or two blocks that will not move, and the company stands still. In a 50/50 company it is almost inevitable without a prior rule. In a 50/30/20 it appears when the 50% needs the 30% to reach 75% and the 30% decides not to go along.
The rule is written before the deadlock, not during it. Common exits: a mediator on a short deadline, a tie-breaking vote assigned to someone, an offer to buy the other stake, a cooling-off clause that forces a pause and a second attempt, or an agreed dissolution if nothing works.
In the example: 15 days of mediation and, if the disagreement persists, a forced purchase mechanism. That mechanism needs a price and a deadline, just like the exit clause. A deadlock without a purchase formula is a lawsuit with a scheduled date.
Non-compete, confidentiality and intellectual property
Three clauses signed without discussion and missed when they matter most. Non-compete stops a shareholder from setting up the same business or taking customers while inside the company and, if agreed, for a period after leaving. It must define territory, activity and duration: a vague prohibition is hard to enforce.
Confidentiality covers what is not public: customer lists, prices, formulas, plans. Without it, the departing shareholder takes the knowledge with no consequences. Intellectual property is the third point: if someone develops a product, a brand or a piece of software inside the company, the agreement must say who owns it. If it does not, the argument comes later, when the asset is already worth something.
All three have a limit: they protect what is legitimate, not abusive restriction. An agreement that stops a shareholder from working forever, in any activity, probably will not hold. Draft them with a lawyer and a sense of proportion.
How to draft it, step by step
- Define each shareholder's stake and contribution, with valuation and deadline.
- Separate what goes into the bylaws from what goes into the agreement.
- Set majorities for ordinary and extraordinary decisions, plus a short veto list.
- Write the sale rules: first refusal, tag along and drag along.
- Resolve the exit: who buys, with which formula and on what timeline.
- Anticipate deadlock with a mechanism and a concrete deadline.
- Add non-compete, confidentiality and intellectual property.
- Choose how conflicts are resolved: conciliation first, arbitration if that fails.
- Sign before a notary or with an electronic signature, and register what must be enforceable.
- Set an annual review date, with an owner and the document at hand.
On step 8: conciliation is cheaper and faster than a lawsuit, and arbitration avoids a court with long timelines. What matters is that the agreement states which forum comes first, in which city and under which rules. The phrase "we'll work it out" is not a clause.
Mistakes that cost dearly and when to update it
The mistakes repeat, and they are almost never about wording. They are about substance: signing something that does not answer the questions that will come up.
| Mistake | How it shows up in practice | What it causes |
|---|---|---|
| Signing a generic template | Clauses that never mention the real business | Useless when conflict arrives |
| Leaving valuation for later | "We'll calculate it when it happens" | The exit becomes a negotiation with no rules |
| Confusing agreement with bylaws | Rules only in the private contract | Not enforceable against the company or third parties |
| Majorities without detail | "By majority", without saying which | Every decision is argued from scratch |
| Ignoring contribution default | Total silence on the subject | The shareholder who never paid keeps the stake |
| Ignoring deadlock | No mechanism in place | Months of paralysis or a long lawsuit |
| Never updating it | An agreement from eight years ago | It governs a company that no longer exists |
When do you update it? When a shareholder joins, the corporate purpose changes, investment comes in, majorities shift or a succession approaches. Those are the moments when the company stops being the one it was when the document was signed. An agreement from eight years ago usually governs another company: other partners, another size, another business.
You do not have to rewrite everything: often a signed amendment, with the same formality as the original, and the corresponding filing are enough.
An honest closing before you sign
This article does not replace legal advice. It is a content guide: it tells you which questions the document must answer, not how each clause should be worded. A lawyer should review the final text with the bylaws, the certificate of existence and legal representation, prior agreements and each shareholder's real situation in view.
If your company is single-owner, if the shareholders are family members with working rules of their own, or if the conflict has already erupted, the order of the steps changes. An agreement is best negotiated before there is something to lose; once there is a lawsuit, the document serves another purpose.
One last warning about expectations: signing an agreement does not eliminate disagreements. It makes them manageable. The difference between a company that argues and keeps going and one that breaks apart is rarely the quality of the friendship, but whether someone sat down one afternoon to write what happens when friendship is not enough.
