In Part 1 of this series, I examined how companies can be controlled using the voting rights attached to shares.
For example, this corporate network graph of Pearson New Zealand Limited is underpinned by a series of several underlying statements, most of which are about shares. You’ll see that in addition to shareholdings, we list another type of statement, called a subsidiary.
Part 2 is about this second type of relationship, the subsidiary. I’ll explore what this means with reference to the UK and USA, and I’ll also show why it’s important to know the provenance of your information: unless you understand the where the fact came from, you can’t know what the word “subsidiary” actually means. Our commitment to showing you all the sources for all our information is what makes OpenCorporates unique.
A word of warning: Part 1 was somewhat technical, but when you look at ways control is exercised without voting rights, it gets worse. Welcome to the world of consolidation, off-balance sheet reporting, GAAP, IFRS, VIEs, and SPVs; a world with such mind-numbing terminology, it makes it hard for the lay person to concentrate long enough to understand.
If you’re not an accounting specialist, I hope this article is interesting enough for you to make it to the end. This is important stuff: the subject of corporate control is increasingly defining the world we live in. From Enron to Lehman Brothers, from Apple to Starbucks, the structure of corporate networks underlies many recent big news stories.
So what does OpenCorporates mean by “subsidiary”? We use the following definition:
A Subsidiary is anything defined as a subsidiary according to the laws and regulations of the reporting jurisdiction.
Let’s start with the UK, and look at a relatively simple legal definition.
In the UK a subsidiary is defined in primary legislation:
A company is a “subsidiary” of another company, its “holding company”, if that other company—
(a) holds a majority of the voting rights in it, or
(b) is a member of it and has the right to appoint or remove a majority of its board of directors, or
(c) is a member of it and controls alone, pursuant to an agreement with other members, a majority of the voting rights in it
The definition starts with the traditional view of control through voting interest, and goes on to extend our notion of control beyond equity in two important new ways: the right to appoint the board and voting rights conferred by contract.
Two of the three clauses above state that a subsidiary must be a member of the company (member means nothing more than “an entity listed in the register of members”). This opens a loophole whereby a holding company can control another company via voting rights, but it is not a subsidiary under law — a situation confirmed in the UK Supreme Court with the comment, “the legislation does lead to a result which is certainly odd and possibly absurd.”
It’s interesting to see how loopholes in formal definitions can be exploited, but a statutory definition of a “subsidiary” only takes us so far. We need publicly available lists of subsidiaries, and these tend to be made available as part of financial and accounting regulations.
Many corporate registers around the world require formal accounts to be deposited annually. The standards which govern how accounts are prepared are usually termed Generally Accepted Accounting Principles (GAAP), and are different around the world. However, they share many essential concepts. When talking about subsidiaries, an important accounting concept is consolidation.
Consolidation versus off balance sheet reporting
The primary rationale behind publishing a company’s accounts is to give consumers, investors and regulators a good idea of the risks associated with a company: is it about to go bust, is it stable, is it undergoing growth? For this reason, accounts are expected to consolidate the accounts of their subsidiaries, as the debts of a subsidiary should also count as debts of the parent company. The concept of a subsidiary is, therefore, closely tied up with the question of when a company should consolidate its accounts.
In the common standards, there are exceptions to when a company should consolidate accounts. For example, a bank that invests money on behalf of a client is not required to register this money as a debt; this money can be taken off-balance sheet.
But while there are legitimate reasons for off-balance sheet accounting, it can be used to make it look like a company has far less debt than it actually does. Under older accounting standards, which used a voting interest definition of control, a company could take a subsidiary off their balance sheet if they owned less than 51% of the voting rights. In the US, the accounting standards of 1959 defined control in this way, and these rules were not replaced until 2003.
So if you must give up voting rights in a company order to get it off your balance sheet, how can you continue to benefit from controlling it?
One approach might be to appoint your children as shareholders, but US legislation, as in many places, has had this covered for a long time, by specifying, for example, that “shares owned or controlled by a member of an individual’s immediate family are considered to be held by the individual.”
It’s relatively easy to follow voting rights to an ultimate owner, so it shouldn’t surprise you to learn that methods exist for taking companies off your balance sheet which bypass voting altogether.
Orphan companies and special purpose entities
This sales brochure explains how one law firm, for example, can offer an off-the-peg control structure in the Cayman Islands that avoids using voting rights entirely. A common approach is to create a trust, put the equity you want to move off your balance sheet into the trust, and then sit on the board of the trust. There’s no longer formal connection between the parent company and the subsidiary, which is now an orphan company.
Enron is perhaps the most famous case of fraudulent off-balance sheet accounting. They wanted to maintain a high credit rating, so with the help of their accountants, they created numerous legal entities known as Special Purpose Vehicles (SPVs) in the Cayman Islands (and other off-shore havens), designed to hide their debt. The SPVs were controlled via contractual mechanisms (such as officers from Enron having the controlling stakes) rather than directly via Enron itself.
SPVs first began to be used extensively in the 1980s, following the invention of new ways to repackage and sell risk in financial markets (such as securitisation). They’re not necessarily fraudulent, but they do provide a convenient location where the boundary between legitimate isolation of risk and fraudulent hiding of risk is blurred.
In the US, the fact that voting interests were an inadequate way of tracing power was recognised in 1987, when it was “tentatively concluded that the concept should be based primarily on control rather than on ownership of a majority voting interest”. However, regulators did not actually take action until the Enron scandal unfolded.
… the power to govern the financial and operating policies of an entity so as to obtain benefits from its activities.
The definition goes on to illustrate ways such power can be exercised, going from specific to generic. It starts with things like voting interest and the ability to appoint the board. It covers “having an agent with the ability to direct the activities for the benefit of the controlling entity”, and ends in the most general sense: “the parent has the power to exercise, or actually exercises, dominant influence or control.”
A similar definition, relating specifically to banks, is contained in the US Bank Holding Company Act, and at least one court has stated that this can mean “the mere potential for manipulation of a bank“.
Broad definitions: useful and problematic
While it has the advantage of being very broad, this latter definition is subjective, and the evolving regulatory story in the US illustrates this. In 2003, in the wake of the Enron scandal, the financial regulators tried to make the definition of control more precise, by introducing the concept of a Variable Interest Entity (VIEs) to US GAAP – essentially, any subsidiary not controlled via voting.
It attempts to define VIEs in terms of the flow of economic benefits, and goes into excruciating detail (summarised here) about how a company must consider how the entity was first set up, how money flows to and from it, and so on.
The original version of the regulations only focused on economic benefits, a definition which brought huge swathes of previously off-balance subsidiaries into scope, and was subject to a lot of criticism. So an updated version was introduced which also required the holding company also to wield “power“.
“Power” is, again, a vague concept. Its assessment requires significant degree of judgement – a judgement which is the responsibility of the reporting company. Accordingly, the introduction of this concept into the regulations delighted some in the finance industry, including Deloitte, who wrote a paper advising clients that
“power” … probably looks and smells a lot more like the concept of “control” that most are familiar with under AR51 [i.e. 51% voting control] … [this] might just be the remedy the doctor ordered.
A spectrum of subsidiaries
As we’ve seen, in the US and the UK alone there is a spectrum of definitions, from the narrow and precise, to the broad and vague. Starting with the most narrow definition, a company can be said to control a subsidiary when:
- it holds majority voting shares (with a threshold of 50%)
- it holds minority voting shares with a threshold of 20%
- it holds minority voting shares which are in a larger block than any other shares
- its directors and their family members together hold a majority of voting shares
- it has had voting rights contractually signed over to it from other shareholders
- it can control the membership of the board
- it can direct activities in the subsidiary to its own benefit
- it derives rights and obligations with respect of revenue and debt from the subsidiary
- it has the potential to manipulate the subsidiary
Because OpenCorporates provides a source for every statement, you can decide for yourself if our statements about control meet your requirements. For example, our corporate network visualisations combine both share-based and subsidiary-based data:
The links we show between companies have been computed based on our own decision of what makes an interesting subsidiary.
Because everything is Open Data, you are free to reconstruct networks using your own decisions – but remember to check the sources for what they mean.
For example, much of the data in our bank visualisations comes from the US Federal Reserve, whose basic rule is a 20% threshold.
In New Zealand, on the other hand, much of our data is from Share Parcels we gathered from the New Zealand official company register; but this may be incomplete, because in New Zealand companies can opt out of showing all the shareholdings in them if they are listed companies, or if they have more than 20 such shareholdings.
In Part 3, I will look at the regulations that dictate when a subsidiary should be named. Without these, they would only exist as consolidated figures in accounts (if that), and we might never be able to tell that they exist in the first place.