When a business signs a lease, borrows money, or gets sued, a fair question is: who, exactly, is on the hook? The answer is almost never "the founder" or "the shareholders." It is a legal entity, an artificial person the law treats as separate from the humans who own and run it. That separation is the quiet machinery underneath every board agenda, and it is worth understanding before you approve the next acquisition, guarantee, or new subsidiary.

The quick version

  • A legal entity (a company) is treated by law as a separate "person": it can own assets, owe debts, sign contracts, and sue or be sued in its own name, distinct from its owners.
  • Limited liability is the payoff of that separation: shareholders can lose what they put in, but their personal assets are generally shielded from the company's debts.
  • A corporate structure is how you arrange multiple entities, a parent or holding company owning subsidiaries, to ring-fence risk, organise operations, or hold assets separately.
  • The shield has limits. Courts can pierce the corporate veil in cases of fraud or sham, and structure used to hide rather than organise (Enron) is a governance failure, not a clever trick.

The idea in depth: a company is a person you made up

The whole edifice rests on one strange, powerful fiction: that a company is a legal person in its own right. The case that nailed this down is Salomon v A Salomon & Co Ltd (House of Lords, 1897). Aron Salomon turned his boot-making business into a limited company, took almost all the shares himself, and lent the company money secured by debentures. When the business failed, unsecured creditors argued the company was really just Salomon under another name and he should pay. The Lords disagreed, unanimously. Lord Macnaghten put it plainly: "The company is at law a different person altogether from the subscribers to the memorandum… the company is not in law the agent of the subscribers." Once incorporated, the company stands on its own feet, even when one person owns it entirely.

So the move is: stop thinking of "the company" as shorthand for its owners. When your board approves a contract or a borrowing, be clear which entity's name is on it and which entity's balance sheet carries the liability. A guarantee given by the parent over a subsidiary's loan deliberately punctures the separation you spent money to create, so treat every parent guarantee as a board-level decision, not a paperwork formality.

This separate-personality fiction did not appear fully formed. It was assembled in stages: Britain's Joint Stock Companies Act 1844 let ordinary groups incorporate by registration, and the Limited Liability Act 1855 added the shield that made shareholding safe for people who were not running the business (the two were consolidated in the Companies Act 1862). Salomon then confirmed how far the principle reached. The point of limited liability was never to help people dodge debts, it was to let strangers pool capital in a venture without betting their houses on a business they did not control.

flowchart TD
  S(["Shareholders
(owners)"]) -->|"own shares,
elect the board"| C(["The company
(a separate legal person)"])
  C -->|"owns"| A(["Assets, contracts,
employees"])
  C -->|"owes"| D(["Debts & liabilities"])
  D -.->|"limited liability:
stops here, not at the owners"| S
The company is a legal person in its own right; liability generally stops at the entity, not the owner. Leaders Loop

Limited liability, and why it is a deliberate trade

Limited liability is the feature most people mean when they say "limited company." It caps what an owner can lose at what they invested, the company can go bankrupt, but the receiver cannot come for a shareholder's savings. That is what makes public markets possible: you can buy shares in a company run by people you will never meet, knowing your downside is bounded.

The trade is real, though, and worth naming honestly. The same shield that encourages investment also lets a company's failure land on its creditors, suppliers and sometimes employees rather than its owners. This is why the choice of entity type matters and why it differs by jurisdiction. In broad strokes: a sole proprietorship is not a separate entity at all, the owner is the business, with unlimited personal liability. A corporation (a C-corp in the US, a "limited" company in the UK and Australia) is fully separate with the strongest liability shield, but typically carries more formality and, for a US C-corp, tax at both the company and dividend level. The limited liability company (LLC), the most popular US structure, is the hybrid: corporate-style liability protection with partnership-style "pass-through" taxation, where profits are taxed once, on the owners' returns (per the US Small Business Administration and IRS).

So the move is: when the structure of a new venture, joint venture or acquisition comes to the board, ask two questions before the lawyers' detail, what risk are we ring-fencing, and what is the tax and reporting cost of doing it this way? The right entity is the one that isolates the risk you are worried about at a formality you are willing to maintain. The wrong one is whichever a previous deal happened to use.

Limited liability is not a loophole. It is a bargain: bounded downside for owners, in exchange for risk borne by creditors, which is exactly why structuring it is a board decision.

An honest limitation. Entity choice is genuinely jurisdiction-specific and tax-sensitive, and the labels do not travel, a US "LLC," a UK "Ltd," and an Australian "Pty Ltd" are not interchangeable, and the tax treatment of each turns on local rules that change. Use this article to ask the right questions; use a qualified lawyer and tax adviser in the relevant country to answer them. Nothing here is legal or tax advice.

Structure: parents, subsidiaries and the holding company

Most organisations of any size are not one entity but a group of them. A parent company owns enough of another company, a subsidiary, to control it. A holding company is a parent whose main job is to hold the shares of its subsidiaries rather than trade itself; the operating businesses sit beneath it. Push this further and you get a special purpose vehicle (SPV): a separate subsidiary created for one project or asset, so that its risks live and die on their own and do not infect the parent's balance sheet.

Because each entity has its own separate personality, the group structure is a way of drawing fences. A property developer puts each building in its own SPV so a problem with one does not sink the others. A bank ring-fences a risky venture so its failure is contained. Done well, this is prudent risk isolation: the SPV is a separate legal entity with its own assets and liabilities, which is precisely the point.

flowchart TD
  H(["Holding / parent company
(owns the shares)"]) --> O1(["Operating subsidiary A
(the trading business)"])
  H --> O2(["Operating subsidiary B
(another market)"])
  H --> SPV(["Special purpose vehicle
(one project / asset, ring-fenced)"])
  O1 --> P1(["Local entity
(e.g. country subsidiary)"])
A simple group: a holding company over operating subsidiaries and an SPV, each a separate entity, each its own fence. Leaders Loop

So the move is: when a structure chart lands in your board pack, do not just admire its tidiness, ask what each box is for. Which entity earns the revenue, which holds the assets, which owes the debt, and where are the guarantees and inter-company loans that quietly stitch the "separate" entities back together? Complexity that no one in the room can explain is a warning sign, not sophistication.

The honest limitation here is the dark side of the same tool. Structure can organise risk or hide it. Enron used a web of special purpose entities not to ring-fence risk but to keep billions of dollars of debt off its own balance sheet, making a failing company look healthy until it collapsed in December 2001. The vehicles were legal in form; the use was deception. The post-Enron Sarbanes-Oxley Act exists partly because separate entities are as good at concealment as at containment. The governance test is not "is this structure permitted?" but "does this structure make our true position clearer or murkier to the people relying on our accounts?"

A worked example

Take a mid-sized renewables firm, call it Bright Field Energy, with a board weighing its first overseas solar project. (Illustrative figures throughout; this is a teaching example, not a real company.) The opportunity is a £40m solar farm in a new country. The temptation is to build it inside the main operating company, where the balance sheet is strongest.

The board, asking the questions above, structures it differently. It creates a special purpose vehicle, a wholly owned subsidiary, "Bright Field Solar One Ltd", to own and finance that single project. Lenders provide, say, an illustrative £30m of project finance secured against the solar farm and its future cash flows, not against the parent. If the project underperforms, the loss is bounded by the SPV's separate personality (the Salomon principle doing exactly the job it was built for): the lenders' recourse is to the project, and the rest of Bright Field's business is insulated.

flowchart TD
  PB(["Bright Field Energy
(parent, existing business)"]) --> SPV(["Bright Field Solar One Ltd
(SPV for one £40m project)"])
  L(["Project lenders"]) -->|"£30m secured on the
project, not the parent"| SPV
  SPV --> ASSET(["The solar farm
+ its cash flows"])
  PB -.->|"NO blanket parent guarantee:
risk stays ring-fenced"| SPV
An SPV puts one project's risk in its own entity, the parent is exposed to its equity, not the whole loan. Leaders Loop

The decisive board moment is the one that almost gets waved through: the lenders ask the parent to guarantee the SPV's loan. A blanket parent guarantee would quietly undo the entire ring-fence, collapsing the careful separation back into a single risk. So the board grants a narrow, time-limited completion guarantee instead of an open-ended one, and records why. That is corporate structure used as governance: not a diagram for diagram's sake, but a deliberate decision about exactly how much of the company is allowed to be exposed to one bet.

Frequently asked questions

What's the difference between a company and its shareholders?

The company is a separate legal person; the shareholders are its owners. The company owns the assets, signs the contracts and owes the debts in its own name. Shareholders own shares in the company, a claim on its value, but they do not personally own its assets or, in the normal case, owe its debts. Salomon v Salomon (1897) confirmed this holds even when one person owns nearly all the shares.

Does limited liability mean directors can never be personally liable?

No. Limited liability protects shareholders from the company's debts; it does not give directors blanket immunity. Directors owe duties to the company and can be personally liable for breaching them, for example, for fraud, for trading while insolvent, or under specific statutes. The shield is for owning the company, not for misrunning it. (See the Toolkit on director duties & fiduciary liability.)

What does "piercing the corporate veil" mean?

It is when a court sets aside the company's separate personality and holds the owners or controllers personally responsible. Courts do this sparingly and reluctantly, usually only where the company is a mere façade for fraud or to evade an existing obligation. The veil is strong by design; Salomon is still the rule, and piercing is the rare exception, not a routine remedy.

Why would a company create lots of subsidiaries and SPVs?

Legitimately, to isolate risk (one project's failure should not sink the group), to organise operations (a separate entity per country or business line), to ring-fence assets, or to partner with co-investors in a defined vehicle. The danger is when the same tool is used to hide debt or obscure the true financial position, which is what turned Enron's special purpose entities from prudent structure into fraud.

How do I choose the right entity type for a new venture?

Start from two questions: what risk are you ring-fencing, and what tax and reporting burden are you willing to carry to do it? Sole proprietorships are simplest but offer no liability shield; corporations offer the strongest separation with more formality; LLCs (in the US) blend the two. Because the rules and labels differ sharply by country, decide the intent at board level, then take jurisdiction-specific advice on the form.

Related in the Toolkit

Structure is only half the picture, the other half is the people who steer each entity. The same separation that limits owners' liability is exactly why director duties bite at the entity level, and why a clear view of board roles and committees matters across a group of companies.

Where to go next