The Crossroads of the Family Business: Cognitive Decline, Law 8/2021, and Corporate Control

Series: support measures · 1 · 2 · 3 · 4 · 5 · 6

Introduction

In almost every family business, a moment arrives when someone begins to notice that the founder is no longer entirely herself. At first it’s small things: a repeated question, a figure that slips away, a meeting that becomes hard to follow. And the telling feature isn’t that no one notices — it’s that everyone notices, and no one says so.

That silence isn’t negligence or a lack of affection. It has a fairly precise explanation. In a family business, every person occupies two seats at once, and what would be a management observation in the boardroom sounds like an accusation at home. On top of that, the cost of saying it falls entirely on whoever says it: the one who speaks up gets branded as the one who wanted to push their father or mother aside, while the relief is shared by everyone. When the price falls on one person and the benefit on the group, the predictable outcome is that no one speaks. And meanwhile, the person no one is talking about keeps signing.

When that person is also the majority shareholder and chairs the board of directors, what opens up isn’t one problem but two, and they resemble each other closely enough to be confused. One is who votes her shares once she can no longer do so in person, and that one is solved by the preventive power of attorney and the autocuratela, albeit with a handful of conditions almost no one checks until it’s too late. The other is who takes her seat on the board, and no power of attorney solves that one, because a corporate office isn’t delegated: it is granted by the general meeting through an appointment, and nothing happens on its own.

What follows traces that distinction and what flows from it: how far a preventive power of attorney really reaches and what needs to be written into it so it works the day it’s needed; what happens when nothing has been planned for — which is either a company paralyzed for months or a meeting called by surprise that removes the very person who founded the company; and the measures that avoid ever getting there, almost all of them inexpensive and almost all of them contained in a single deed and a bylaw amendment.

They all share something, and it’s what this article ultimately argues for: they let the decision be made by the person who can still make it, about herself, at a moment when the matter is still abstract and doesn’t hurt. Seen this way, this isn’t a legal problem disguised as a family conflict. It’s a family problem that can only be fully closed with legal tools, and there’s no reason a family should have to go through it alone or improvising.

The Facts of the Case

The scenario I describe is fictional for reasons of confidentiality. I mean the details, not the situation.

The board of directors of the family business has spent twenty years meeting on the same day of every month, in the same room, with the founder presiding. It’s a five-member board: the founder, three of her children, and the representative of a minority shareholder. The agenda hasn’t changed in two decades: the founder speaks first, then everyone else, if asked. The CFO is occasionally invited to attend. The HR director is occasionally invited too. Everything unfolds as always, though lately the children have noticed their mother repeating a question that’s already been answered several times over, and they’ve discussed it more than once in the hallways and at home. In recent meetings she has lost the thread of a basic process in the business she herself created and has run her whole life. The children glance at each other, but no one says anything, and they move to the next item.

The moment to resolve this wasn’t that day. It was ten years earlier, in an equally routine meeting where nothing in particular was happening.

In the previous piece in this series we talked about “the conversation” a family needs to have about the loss of capacity, and we closed with this: when it happens inside a family business, the problem multiplies. It is no longer just a matter of planning for the loss of capacity of the person who makes decisions at home; you also have to plan for the loss of capacity of the person who makes decisions at the company. This article deals with that second layer of ambiguity. And it does so from a premise worth stating up front: everything proposed here has to be done beforehand. Once the decline has already set in, almost none of these tools can be put in place, because all of them require the signature of someone who can no longer sign.

A point of vocabulary before we start. Since Law 8/2021, judicial incapacitation no longer exists in Spain. There are no “incapacitated persons.” There are people with impaired capacity who need support, to varying degrees and in varying forms, and whose will, wishes and preferences remain the guiding criterion. This isn’t merely a matter of vocabulary. Almost every planning mistake we’ll see below comes from continuing to think in terms of a switch that flips off, when what actually exists is a slope.

The Part That’s Already Solved, and Under What Conditions

Let’s start with the good news, because there is some. If the founder and majority shareholder has executed a preventive power of attorney or an autocuratela (see my earlier posts), the estate side is better covered than it looks, and so is the most visible corporate side: the vote.

A very widespread misunderstanding is worth clearing up here, because there’s a common belief that a power of attorney to represent someone at a shareholders’ meeting has to be a general power, and that isn’t always true. What determines whether the power works for the meeting isn’t its scope, but who the attorney-in-fact is.

In the limited liability company (SL), art. 183 of the Spanish Capital Companies Act (LSC) doesn’t always require a general power: it lists who may represent the shareholder. The list is the spouse, an ascendant, a descendant, another shareholder, or a person holding a general power in a public deed with authority to manage the whole of the represented person’s assets within Spanish territory. The general power is only the entry point for someone who is neither a relative nor a shareholder. The same article adds that the representation must be conferred in writing, and only has to be special for each meeting if it isn’t recorded in a public deed — which a notarial power of attorney satisfies by definition.

In the public limited company (SA) the mechanism differs even though the result is the same. Art. 184 LSC allows representation by any person but requires the representation to be conferred specially for each meeting; and art. 187 LSC lifts that requirement precisely when the representative is the spouse, an ascendant or a descendant, or when they hold that general power.

Translated into the typical family business, where the attorney-in-fact is a child or the spouse: the preventive power of attorney works even if it’s limited to corporate matters. It doesn’t need to be general. The general-power requirement only applies when someone from outside the immediate family circle is appointed — which is exactly what’s done when the goal is to avoid friction among siblings. The irony is notable: the solution to the conflict of interest, appointing a third party so there’s no friction between the children, takes you out of the cleanest route under art. 183 LSC. Though there is a fix.

For that case, the fix isn’t in the power of attorney but in the bylaws. Art. 183 LSC itself ends by saying the bylaws may authorize representation by other persons. With that clause, a neutral third party — the family’s longtime advisor or an independent professional — can represent the founder at the meeting with a power limited to corporate matters, without needing any general power at all. It’s an easy solution, yes, but it’s important that this clause actually be in the bylaws if you want to rely on it.

It’s worth weighing how wide to draft the clause, because the expansion applies to every shareholder, not just the founder, and because the spouse, ascendants and descendants were already covered by law. What the clause adds is third parties: one branch’s lawyer, another branch’s advisor, a child’s unmarried partner, and in the worst case someone outside the family with interests of their own. Opening it all the way changes the temperature in the room. Requiring that the representation be recorded in a public deed lets the trusted attorney-in-fact in and keeps out improvised authorizations, which is roughly the point.

This shortcut, mind you, exists only in the SL. In the SA there’s no subjective restriction to lift, because art. 184 LSC already admits any representative; what exists is a formal requirement — that the representation be special for each meeting — from which art. 187 LSC only exempts the spouse, an ascendant or descendant, and whoever holds a general power. For a third party in an SA, the general power remains the only safe route. One more asymmetry worth knowing before deciding on the corporate form.

And only if you go down that road does the final caution appear. Consider what has just happened: for a third party to be able to vote at a meeting, they’ve had to be given a power over the founder’s entire estate in Spain. The goal was to hand over the vote, and what ended up being handed over was the signature.

That drags in something almost no one anticipates. Art. 259 of the Civil Code (CC) provides that when a power of attorney carries a survival clause, or is granted solely for the case of a need for support, and covers all of the grantor’s affairs, the attorney-in-fact becomes subject to the rules of the curatela in everything the power did not provide for, unless the grantor stated otherwise. Among those rules is the judicial authorization that art. 287 CC requires to sell real estate or business establishments, to dispose of assets gratuitously, or to borrow money. Precisely the judicial intervention the preventive power was meant to avoid.

None of this touches the vote, which is cast without anyone’s permission. It touches everything else that broad power allows, starting with the shares themselves, which couldn’t be sold or pledged without going through the courts.

It isn’t automatic, because a power of administration over Spanish assets isn’t necessarily a power over all of the grantor’s affairs — but that discussion isn’t settled by a court: it’s settled by the notary authorizing the deed or the registrar reviewing it. And if either of them reads it as all-encompassing, the transaction stalls. Being right doesn’t matter: by the time the dispute is sorted out, the buyer will have moved on.

So the fix isn’t to fine-tune the interpretation, but to expressly exclude the curatela rules — which art. 259 CC itself authorizes in its final clause — and replace judicial oversight with controls of your own: joint action for disposals above a threshold, annual accountability to a family council. Control is preserved without having to go to court every time.

One detail that surprises whoever hears it: preventive powers of attorney executed before September 3, 2021 carry a good part of that exclusion built in by law. The third transitional provision of Law 8/2021 exempts them from arts. 284 to 290 CC, precisely the ones covering the inventory, judicial authorization and the voidability of acts done without it. It doesn’t free them from the whole curatela regime, but it does free them from the block that gets in the way. Later powers, by contrast, carry nothing: it has to be written in. So a power signed in 2015 can actually be better positioned on this point than one signed last year off a careless template. Anyone holding an older one shouldn’t rush to redo it without first checking what they’d gain and what they’d lose.

What never changes, whichever route is chosen, is that the representation has to be in writing and, to be valid for any future meeting rather than just one, has to be recorded in a public deed. Being a relative qualifies you to serve as a representative; it isn’t itself a title of representation. A son doesn’t vote his mother’s shares simply because he’s her son. A notary is still required, and someone still has to have gone to see one while the founder was still able to.

With those precautions, yes: the shareholder side is solved. The problem isn’t there. It’s one level up, on the board.

What the Power of Attorney Doesn’t Cover

Being a director isn’t the same as being a shareholder. A shareholder exercises a right that belongs to them and can delegate it to a representative. A director holds an office, and an office isn’t delegated by power of attorney: it is granted, through an appointment, by resolution of the general meeting. No one can become a company’s director because another director signed them a power of attorney. The general meeting has to appoint them.

There’s a partial palliative worth knowing even though it doesn’t solve the underlying problem. A director may delegate their representation to another director for a specific session, when the bylaws or the board’s internal regulations allow it. It’s useful for travel, for a passing illness, for the day you simply can’t attend. It doesn’t help here, and for an obvious reason: someone who can no longer form a judgment about board matters also cannot validly grant that representation, so the remedy stops existing at the exact moment it would be needed.

And this has a consequence families rarely anticipate: a director’s loss of capacity does not automatically remove them from office. No mechanism acts on its own. Art. 223 LSC allows the general meeting to remove a director at any time, without cause and even if the removal isn’t on the agenda. It’s an extremely broad power. But someone has to call that meeting, and someone has to put that item on it.

Art. 224 LSC is sometimes cited in this context, which requires the removal of a director who falls under a statutory prohibition, at the request of any shareholder. Here it’s worth being honest with the reader: that article probably doesn’t help here. The list of prohibitions is in art. 213 LSC, and it still literally reads today, “persons judicially declared incapacitated.” It’s a fossil. Law 8/2021 abolished judicial incapacitation five years ago, and the legislature never updated the corporate statute, which has since been amended for other reasons without touching this point. The provision refers to a category that no longer exists.

The practical consequence is twofold. Cognitive decline is not, by itself, a statutory prohibition on serving as director, and anyone relying on art. 224 LSC to force a removal will run into a skeptical registrar. And for planning purposes: the court order establishing support measures can expressly rule on the exercise of the office, and it’s worth asking it to. What there isn’t is any automatic mechanism.

The preventive power of attorney settles who votes. It does not settle who serves as director.

There Are Two Endings, and Both Are Bad

At this point, the usual story continues with paralysis: no one does anything, the signature carries less and less weight, the meeting is never called. That’s one possible ending. But it isn’t the only one, and the other is almost never told, because it’s more uncomfortable.

Let’s put numbers to the board from the opening paragraph. The founder presides and holds sixty percent of the capital. Her three children are directors, one of them vice-chair and another secretary, and together they hold thirty percent. The minority shareholder holds the remaining ten percent, and their representative occupies the fifth board seat.

One day, instead of exchanging glances and moving to the next item, the children call a general meeting. Two items on the agenda: removing their mother as director, and changing the way the board is organized, moving from a board of directors to two directors acting jointly.

The founder doesn’t attend. She doesn’t attend because she’s no longer aware of what’s happening.

The question is whether that holds up. And the uncomfortable answer is that, in essence, it does.

The Notice of Meeting

Calling the meeting is a matter for the governing body, and with a board that means a board resolution, which is also a non-delegable power. The children have this settled by arithmetic. The board is called by the chair “or whoever acts in her stead,” and the vice-chair is one of them; and in any case, directors who together make up at least a third of the board can call it if they ask the chair and she fails to act within a month. The board is validly constituted with a majority of its members: three out of five. And the resolution is adopted by a majority of those present: three out of three.

The one real point of failure is procedural, and it matters: the founder has to be notified of that board session, and so does the minority shareholder’s director. If they aren’t notified, the resolution is void, and that voidness carries over to the meeting called on the strength of it. It’s the flaw that most often brings these operations down. It’s also fixed with a certified letter.

The Meeting With Thirty Percent

Here the answer forks depending on the type of company, and the difference is enormous.

In the SL there’s no quorum requirement to constitute the meeting, but there is a voting floor. Resolutions are adopted by a majority of the votes validly cast, provided they represent at least a third of the votes corresponding to the entire share capital, not just the capital present. The three children together hold thirty percent. They don’t reach a third. On their own, absolutely nothing passes.

So they need the minority shareholder. That ten percent that meant nothing at any meeting for twenty years turns out, on the day it matters, to be the piece that decides everything. With their vote added, the children reach forty percent, clear the one-third floor, and the founder’s removal goes through.

The second item, on the other hand, doesn’t pass. Changing how the board is organized, if it amounts to a bylaw amendment, requires more than half the votes of the entire share capital, and forty percent doesn’t reach it. Unless the bylaws already provide for alternative modes of administration, in which case the meeting can choose one by ordinary majority without amending anything — and then it does pass. It’s a fairly common clause: worth checking before sleeping soundly.

In the SA the outcome is much worse for the founder — the exact opposite of what its elaborate quorum machinery would suggest. There’s no floor over total capital; there’s an attendance quorum, measured against whoever shows up. On first call, twenty-five percent is needed for ordinary business, which the three children already have with their thirty percent, and the removal passes by simple majority of those present, without needing the minority shareholder at all. Bylaw amendments require fifty percent on first call, but only twenty-five on second call, with the added majority of two-thirds of the capital present. Two-thirds of thirty is twenty. They have thirty. Everything passes, and it passes on their own.

And the quorum is neutralized without any maneuvering at all: the notice can announce both calls at once, twenty-four hours apart. A single announcement, the first call lapses, the second is held the next day, and it approves the entire agenda.

The Shortcut, in Case Something Doesn’t Work

Even if the change of structure falls through on the majorities, the same result is reached anyway, and it doesn’t even take forcing it.

Once the founder is removed, the board is left with four members: the three children and the minority shareholder’s representative. It remains a valid board, because the legal minimum is three. Except now the children control three of the four seats. And the permanent delegation of powers to a managing director requires the favorable vote of two-thirds of the board’s members: two-thirds of four is two point sixty-seven, meaning three votes. Exactly what they have.

They don’t need to amend the bylaws. It’s enough to appoint one of themselves managing director and delegate the entire management to him or her. The company has changed hands without a single bylaw amendment.

On a smaller board, the path would differ but would arrive at the same place. If the board had three members, removing the founder would drop it below the legal minimum, and it would have to be reconstituted at the same meeting, appointing a third director by ordinary majority — which is exactly the majority they already have.

The barrier of reinforced majorities protects the letter of the bylaws, but not control of the company, and control is reached anyway through the door next door.

Sixty Percent Defends No One

This is the point that changes how everything before it should be read.

With her sixty percent, the founder could reverse all of it in a single afternoon: call a meeting, remove her children, restore the board, appoint whoever she wanted. All she has to do is vote. And if she can’t vote, and there’s no attorney-in-fact to vote on her behalf, that sixty percent doesn’t exist operationally. It’s capital with no voice.

What unplanned cognitive decline produces, then, isn’t a symmetrical deadlock in which no one can do anything. It’s an inversion of power. The minority keeps every lever: it’s on the board, it can call meetings, ordinary majorities are enough for it. The majority is left holding a stake no one can exercise. The controlling shareholder becomes the most vulnerable person in the company, and precisely because of that control, since her entire defense depended on a personal act she can no longer perform.

The tools of defense exist, and they’re good ones. Resolutions can be challenged for breach of law or of the bylaws and for harm to the corporate interest, and the law expressly addresses resolutions abusively imposed by the majority. A precautionary suspension of the challenged resolutions can be sought, for which five percent of the capital in an unlisted company suffices, together with a preventive notation of the claim in the Commercial Registry. The children can be held liable for breach of their duty of loyalty, for using their position as directors to seize control. And there’s a built-in early warning: before recording the appointment, the Commercial Registry Regulations (RRM) require formal notice to the previous holder of the certifying power, with fifteen days to object. That certified letter is usually the moment the branch of the family that knew nothing finds out.

There’s also a route that isn’t corporate at all and is worth knowing, because it’s the only one that works when there’s no power of attorney and no curatela. Art. 757 of the Civil Procedure Act (LEC) entitles the person themselves, their spouse, a descendant, an ascendant or a sibling to initiate proceedings for the provision of support, and it obliges the Public Prosecutor’s Office to initiate them if no one else does. And art. 762 LEC allows the court to adopt, on its own motion, whatever measures it deems necessary to protect the person or their assets — which in practice usually means appointing a judicial guardian (defensor judicial), provisional administration, or a prohibition on disposing of assets. A judicial guardian, once appointed, can indeed attend the meeting and vote the founder’s shares.

It’s worth understanding its scope correctly, because more is expected of it than it delivers. These measures protect the person and their assets; they don’t govern the company. No judge is going to suspend a company’s operations through this route. To stop corporate resolutions you have to go through a challenge and a precautionary suspension, and there you act as a shareholder, not as a child.

And every one of these tools, corporate and civil alike, presupposes someone who activates them within the deadline. A person no longer aware of what’s happening on the board isn’t going to file a claim either. The children who might want to defend her face weeks of proceedings; the meeting is called with fifteen days’ notice. It’s a race almost always lost. Without an attorney-in-fact, the toolbox is full and locked.

And Yet, This Isn’t Always a Coup

It’s worth saying, because the story above lends itself to a moral reading that doesn’t always fit.

Sometimes those children aren’t heartless at all. They’re three people watching the company slide month after month, who’ve spent two years trying to raise the subject at family lunches and running into the same polite silence every time, and who eventually discover they have no clean path to do anything, because their mother never signed anything. What the law offers them is exactly this — call a meeting, remove her, reorganize — and doing it looks every bit like a betrayal even when it’s a rescue. There’s no graceful version of this operation, simply because the law never anticipated it.

And there’s a detail that makes it worse: to pull it off in an SL, they’ve had to ask for the vote of a shareholder outside the family. Whatever the outcome, the day the founder is no longer around, that family will carry the memory that the children allied with the outsider to push their mother aside.

Which is the argument running through this whole series: a lack of planning doesn’t prevent the decision — it only guarantees it gets made in the worst possible way. By surprise, without the person concerned present, with a certified letter involved, and leaving the family split into two camps that never wanted to be camps at all.

What Comes If None of This Happens

If paralysis comes instead of a takeover, the outcome usually cited is dissolution due to paralysis of the corporate bodies. It does exist, it’s true, but it requires permanent, insurmountable paralysis, and we’ve already seen that the meeting can almost always convene and decide something. It’s the extreme scenario.

What arrives much sooner, and is far more likely, is more mundane: annual accounts that never get drawn up, the filing that never happens, and the closure of the registry file that follows. From that point on, nothing gets recorded. Not a change of registered address, not a power of attorney, not the appointment of a successor. The company keeps operating, but it stops being able to change.

And there’s a symmetrical risk families take even longer to see, because they only watch for what stops getting signed: what keeps getting signed. Contracts, guarantees, deeds. The day a notary refuses to authorize a transaction because they doubt the capacity of the director appearing before them, the conversation no one wanted to have opens up on its own, at the worst possible moment and in front of third parties. That scenario — what happens once nothing at all has been planned for — is the subject of the next article in this series.

Why It Weighs More Here Than in Any Other Company

In a non-family company, replacing a director who can no longer serve is an uncomfortable but neutral formality: a meeting is called, someone else is appointed, life goes on. In a family business, that same formality is also a public declaration. A director’s removal is recorded in the Commercial Registry — it’s a formal, visible act, dated and signed. Raising it on the board isn’t like broaching an awkward conversation over a family lunch. It turns that conversation into a corporate resolution, drafted by a notary or certified by a secretary, that anyone can look up afterward.

It’s worth defusing part of that fear, though, because it’s a poorly calibrated one. The Registry publishes the removal, not its cause. The entry reflects that the meeting resolved the removal or that the director resigned. It says nothing about why. No one looking up the registry file ten years from now will read anything about the founder’s health. Being able to tell a family this, with the law in hand, unlocks more conversations than you’d expect.

And there’s another reason, less noble and far more common, why no one proposes anything, and one that’s rarely said out loud: what the founder lives on. In many family businesses, stepping down as director means giving up the compensation that comes with the office, or the managing director’s contract. When that income supports the first generation, raising the question of succession in governance is, without saying so, raising the question of a lower standard of living for her. Any plan that doesn’t resolve the financial arrangement first will run into a resistance the family will justify with arguments about respect, because those are more presentable.

That’s why, in the family business, ambiguity doesn’t just multiply — it freezes solid. The children who are also directors hesitate to propose something that, outside the boardroom, feels like a lack of respect toward their mother. The director representing the minority shareholder, the one person who could see it clearly and without emotional cost, would rather not be the one to raise it, because they know how it will read if the outsider is the one pointing at the founder. And meanwhile the company keeps running, with a signature that carries less and less weight and a meeting that never gets called, until the day someone decides enough is enough and calls one.

How to Avoid It — Not When It’s Needed, but Before

The solution isn’t legally complex. It is, again, a matter of getting ahead of it. These are ordered from most to least important, and the first is the one that’s usually missing.

Let the founder herself organize her own succession. This is the piece that closes the loop on this entire series, and the one almost never included. Nothing prevents her from leaving express instructions, in the preventive power of attorney or the autocuratela, on how her own removal as director should be voted, and on who should succeed her on the board. She can set the criterion, the procedure, and the timing. She can say, before a notary and years in advance: when this happens, my attorney-in-fact will vote my shares in favor of my removal and of the appointment of whoever the family council designates.

This can be done, and it isn’t a strained reading of the law: art. 258 CC allows the grantor to set “conditions and instructions for the exercise of the powers,” and art. 271 CC, for the autocuratela, allows provisions on the rules for managing and disposing of one’s assets. It’s worth knowing, though, that the instruction operates in the internal relationship. If the attorney-in-fact votes against it, the vote remains valid as against the company; what exists is a breach that carries liability and can lead to the judicial termination of the power. It’s a binding instruction, not a button that gets pressed automatically.

Read slowly what that means. There’s no longer a child who has to propose pushing their mother aside. There’s an attorney-in-fact carrying out what she decided when she was able to decide it. The corporate act is identical, but the emotional weight has shifted entirely, and it has shifted onto the one person who could bear it without harm: the founder herself, at a time when talking about it was still talking in the abstract. Law 8/2021 was built on the idea of respecting a person’s will as expressed before support became necessary. This is exactly that, applied to the board.

A preventive power of attorney drafted to actually work on the day it’s needed. We’ve already seen the mechanics, and they’re not the same in every case. If the attorney-in-fact will be a child or the spouse, a power limited to corporate matters is enough, always in a public deed, with express mention of representation at meetings and of voting instructions. If it will be a neutral third party and the company is an SL, the bylaw clause expanding the circle of representatives is also needed. And only if it will be a third party in an SA is the general power required, in which case you must remember to expressly exclude the curatela rules. Under any of the three routes, it’s worth adding the express power to request that a meeting be called, as a shareholder, because when the moment comes that may be the first thing that needs doing.

This document isn’t just a continuity tool. After everything we’ve seen, it’s above all a defense tool: the only thing that turns sixty percent of the estate into sixty percent of the votes on the day someone calls a meeting the founder can no longer understand.

Avoid the conflict of interest before it exists. The attorney-in-fact is usually the eldest child, who is usually also a director. And what he’ll be asked to do is vote his mother’s shares to remove his mother and, quite often, to appoint himself. Art. 251 CC prohibits providing support where there’s a conflict of interest, and the fallback is a judicial guardian: proceedings, a court, and months of delay — precisely what you were trying to avoid. On the corporate side, moreover, a resolution adopted with that vote is challengeable, and the burden of proving it served the corporate interest falls on whoever adopted it. This is prevented by appointing separate attorneys-in-fact for estate matters and for corporate matters, or a neutral third party for this act alone. It’s one more paragraph in the deed, and it avoids having to argue it out in court later.

Shield the bylaws, knowing you can’t fully do so in an SA. There’s an asymmetry here worth knowing before choosing the corporate form.

In the SL, shielding is simple and effective. The law allows the bylaws to require a reinforced majority of up to two-thirds of the votes corresponding to the entire share capital for removing a director. With that clause, neither the children’s thirty percent nor the forty percent they’d reach with the minority shareholder removes anyone. Ordinary and reinforced majorities can also be raised, with the sole limit of not reaching unanimity, and the door to alternative modes of administration can be closed.

In the SA, that shielding isn’t available. The general meeting’s power to freely remove directors is treated as a non-waivable principle, and the possibility of requiring a reinforced majority for removal is recognized by law only for the SL. What can be done is raising the constitution quorums for both calls, with the limit that the second call’s quorum must be lower than the first’s, and raising the majorities — which makes changing the structure more expensive but doesn’t prevent removal. For anything more, you have to step outside the bylaws: syndicating the vote in an agreement signed by all shareholders, or interposing a holding company between the family and the operating business, so that the meeting that decides is the holding’s, and the founder’s stake isn’t directly exposed.

And it’s worth noticing something. Expanding the circle of representatives at meetings, a reinforced majority for removing directors, alternate directors, a fixed term of office, closing off alternative modes of administration. These are five clauses that fit into a single bylaw amendment, and none of them names anyone. That’s the moment when this gets done with no emotional cost at all: a general review presented as modernizing the company’s governance, voted on while the founder still presides and still votes. No one is deciding what to do about Mother. They’re updating a set of bylaws from 1994. Ten years later, those five clauses are the difference between a family that manages a succession and one that finds out by certified letter.

It’s also worth not putting this off, for a very specific reason: amending the bylaws requires a meeting and, in the SL, more than half the votes of the entire share capital. If the founder holds sixty percent and can no longer vote it, the clause that would have allowed appointing an attorney-in-fact can never be approved. The bylaw clause is a purely preventive tool. Once the day it’s needed arrives, it can no longer be put in place.

Design a management structure that holds up. Four tools, all inexpensive, none of which requires singling anyone out.

Alternate directors: the bylaws or the meeting can appoint in advance whoever will automatically step in upon a removal, with no need for a new meeting. It costs one line in the bylaws and, on its own, covers the most common scenario.

A fixed-term appointment. In the SA, the office lasts a maximum of six years and then expires; in the SL, it’s indefinite unless the bylaws say otherwise. Setting short, renewable terms turns a traumatic removal into something far more manageable: a non-renewal. No one has to propose anything. The date arrives, and the board decides on renewal using management criteria, the way it decides everything else. Moving the problem from the terrain of decision to the terrain of the calendar is, for a family that dreads the moment of saying it, an enormous relief.

The corporate-entity director: having the family holding company itself occupy the board seat, with a natural person designated as its representative. The seat survives the person, and changing the representative is a decision of the holding company, not a removal of the founder.

And, in smaller companies, two directors acting severally instead of a sole director. With directors acting severally, the company retains valid representation even if one of the two can no longer act. It carries control trade-offs that need weighing, but for the risk we’re discussing here, it’s almost free insurance.

Delegate powers to a managing director, knowing what the delegation doesn’t cover. The law allows broad management powers to be permanently delegated to one or more managing directors or to an executive committee, with the favorable vote of two-thirds of the board. Done in advance, this separates day-to-day management from the specific person chairing the board.

Here’s the essential catch, though: certain powers can’t be delegated, including drawing up the annual accounts and calling the general meeting. And the accounts must be signed by all the directors; if a signature is missing, the reason has to be stated. So you can have all ordinary management flawlessly delegated to a managing director and still be unable to close out the financial year, because the founder’s signature is missing. Delegation solves three hundred sixty of the days. It leaves the one that actually matters untouched.

Register broad commercial powers of attorney, knowing where their limits lie. Beyond the office of director, it’s possible to record general powers of attorney in the Commercial Registry in favor of trusted individuals for the ordinary management of the business. These belong to the company, not to the director who granted them, and so they aren’t affected by whatever happens to that director. They provide operational continuity while the question of who sits on the board is worked out calmly.

Its limit runs the other way: an attorney-in-fact cannot do what the law reserves to the governing body. They cannot call the meeting, cannot draw up the accounts, cannot file for insolvency. And there’s a sequencing trap: if the person in charge is the sole director, that same person is precisely the one who would have to grant or renew those powers. These powers of attorney have to be put in place beforehand, not once they’re needed.

Put succession in governance into the family protocol, then take it back out. The protocol usually regulates in detail who inherits what. Less often does it regulate who replaces whom on the board, by what procedure and under what trigger. Agreeing on it in advance, before anyone is singled out, is far less emotionally costly than deciding it once a specific name is already on the table.

But it has to be said clearly what a protocol is and isn’t. The family protocol is a side agreement among shareholders, and art. 29 LSC establishes that agreements kept private among shareholders are not enforceable against the company. A clause saying “the founder steps down at seventy-five” does not, by itself, produce any removal. At most it creates an obligation among those who signed it, enforceable through contractual liability — meaning, with a lawsuit involved, and years later.

There is a form of registry disclosure for the protocol, regulated under Royal Decree 171/2007, with four routes the governing body can choose: publishing it on the company’s website, recording in the Commercial Registry that it exists without disclosing its content, filing it together with the annual accounts, or recording the corporate resolutions adopted under it with express mention of that fact. It’s worth not confusing what that achieves: publishing the protocol doesn’t make it enforceable against the company. It prevents anyone from claiming ignorance of it and leaves a trail of which resolutions carry it out, which is no small thing, but it doesn’t turn the pact into a rule.

The real effectiveness lies elsewhere: transposing into the bylaws whatever can be transposed — terms of office, requirements, objective grounds for removal, alternates, and reinforced majorities. A protocol left in a drawer is worth only as much as the good faith of whoever signed it.

Design the trigger carefully, and appoint the messenger. It’s common to propose “a medical report” as the activating criterion. It sounds objective and is a minefield. It’s health data subject to the General Data Protection Regulation (GDPR), it requires a legal basis to process, and it immediately raises three questions the clause almost never answers: who chooses the physician, who can request the examination, and what happens if the person refuses to undergo it.

The clean way to solve it is the same one we’ve already seen: let the founder herself design the trigger in her preventive power of attorney or her autocuratela, designating the verification mechanism she accepts in advance. It stops being an exam the family imposes and becomes a condition the person concerned set for herself.

And there’s a second piece, more about mediation than about law, that in practice carries as much weight as everything above: agreeing in advance on who raises the conversation. The independent director, the non-director secretary, the outside advisor, the family mediator. Someone who is none of the children, and preferably not the minority shareholder’s director either, so the conversation doesn’t arrive tinged with suspicion. Designating the messenger today, while the message is still hypothetical, keeps the cost of saying it, eight years from now, from falling on whoever in that family can least afford it. Combined with a letter the founder writes now, to be read on that day, it’s probably the single most effective tool that fits inside a protocol.

Keep the family council separate from the board of directors. The first is the place to talk about people: how each person feels, what belongs to whom, the things we called role ambiguity in the previous article in this series. The second is the place to decide using management criteria. Mixing both conversations in the same room is the surest way to make sure neither one lands anywhere good.

The Tax Bill No One Runs

There’s one dimension left that is neither corporate nor emotional, but fiscal. I’ll flag it here because it deserves an article of its own, but it can’t be passed over.

The family-business exemption under the Wealth Tax, and with it the reduction under the Inheritance and Gift Tax, depend on someone within the family group actually performing management functions at the company and receiving meaningful compensation for doing so. For decades, that someone has been the founder. The day she steps down, if no family member takes on those functions with that same compensation, the benefit lapses. It isn’t a penalty — it’s that the company stops meeting the requirements that gave it family-business tax treatment. When succession arrives, the bill runs into the hundreds of thousands of euros, or into the millions. And the fix is simple if it’s thought through in advance — having the handover of management functions happen at the same time as the handover of the office — and impossible to improvise on the day of the meeting itself.

There’s something worse, and it’s the least reversible consequence of all: once capacity is lost, the window for lifetime succession planning closes. Gifts, succession agreements where regional civil law allows them, reorganizations ahead of the handover. All of that requires a founder capable of granting it, and the curator cannot stand in for her in a gratuitous act except with judicial authorization, which is exceptional.

Put differently: cognitive decline doesn’t just complicate the governance of the company. It shuts down the most efficient way of passing it on. And it does so silently, with no one signing anything, on some ordinary day when a line gets crossed that’s only visible in the rearview mirror.

The specific requirements are set out in Law 19/1991, on the Wealth Tax, and in Law 29/1987, on the Inheritance and Gift Tax, and each autonomous region improves on them with its own rules. That’s a matter for a tax specialist, and for another article.

A Note on Regional Civil Law

Everything above is written from the standpoint of the national Civil Code. Catalonia, Aragón, Navarre, Galicia, the Balearic Islands and the Basque Country have their own regulation of support measures and succession agreements, in some cases more flexible. Since a large share of Spain’s family-business fabric sits in those territories, the first question before designing anything is what the founder’s regional civil-law status (vecindad civil) is, because depending on the answer, the tools available aren’t the same.

In Summary

The preventive power of attorney and the autocuratela solve the shareholder side, and they solve it better than people think: in a family business, where the attorney-in-fact is usually a child or the spouse, the power doesn’t even need to be general. What it needs is to be executed in a public deed and drafted with the day it will actually be used in mind.

They don’t solve the director side, because a corporate office isn’t delegated by power of attorney: it’s granted by appointment, and no one loses it by default. The LSC, which still talks about “incapacitated persons,” offers no automatic mechanism at all.

And there’s something worth understanding before deciding it can wait: holding a majority of the capital protects no one. It protects whoever can vote it. A founder with sixty percent and no attorney-in-fact is, within her own company, the most exposed person of all, because the minority keeps every lever and she keeps none.

That second layer of ambiguity, the one on the board, doesn’t get closed with one more notarial document. It closes with four things at once: a preventive power of attorney drafted to actually work, not to sit in a drawer; bylaws shielded as far as the corporate form allows, which is a great deal in the SL and less in the SA; corporate governance planned ahead of time, with alternates, fixed terms, delegation of powers that accounts for what can’t be delegated, and powers of attorney independent of the office; and, above all, a founder who organizes her own succession while she still can, so that when the day comes, no one in her family has to propose what she had already decided.

None of this is expensive or complicated. One afternoon at the notary’s office and a meeting to amend a set of bylaws. The hard part is doing it in the year it still seems unnecessary — which is the only year you can.

One question remains that this series hasn’t answered yet, and it’s the most uncomfortable one of all: what happens when none of this has been done. When there’s no power of attorney, no protocol, no alternates, and what there is instead is a director who keeps on signing. That’s the subject of the next article.

---

Notice

This text is informational and does not constitute legal or tax advice, nor does it create any professional relationship. Everything explained here depends on the specific case: the type of company, what its bylaws say, the person’s regional civil-law status, and the autonomous region where they reside. The rules cited are in force as of the publication date. Before making any decision on the matters discussed here, you should consult a professional who can review the actual documents involved.

Has your family business settled succession in ownership but not in governance?

At Alta Mediación we combine corporate law and family mediation to review both sides of the problem at once: the family protocol and how it fits the bylaws, the existing powers of attorney and what they actually cover, the majorities and quorums that do, or don’t, protect the majority shareholder, the structure of the board, and the tax impact of the handover. In the first consultation, we identify what’s already covered and what’s still exposed.