Rehypothecation: a simple guide

by on January 6, 2012 at 3:55 am in Economics, Law | Permalink

From Keith Fitz-Gerald, via The Browser, here is a useful introductory article to what is likely this year’s coming hot topic.  Excerpt:

Assets in brokerage accounts can be used and re-used in such a way the credit multiples far outweigh the actual assets in the accounts. In effect, rehypothecated assets become part of a daisy chain, for lack of a better term, wherein one company’s liabilities become another’s assets.

If there is a hiccup anywhere in the chain, the effect is one of instant collateral collapse as everybody in the chain is forced to buy back, or recall, their assets. The effect is not unlike a colossal global “short” on world markets.

Imagine what happens if something goes wrong and everybody wants their $10 back, but find that there is only $1 in actual cash.

I believe this is what Federal Reserve Chairman Ben Bernanke and his counterparts at the ECB are so concerned with and why they are obsessed with liquidity. Everybody knows that too much debt caused this mess, but what they don’t realize is that it’s the use of rehypothecated assets that make collateralizing it nearly impossible barring massive injections and printing.

I would define it as “an asset used as collateral more than once, at the same time.”  And there is this:

Take the United Kingdom, for example, where there is no limit on the amount of client assets that can be rehypothecated. There, brokers have reportedly and routinely rehypothecated 100% of the value of client accounts, not just those assets pledged as collateral.

What kind of day would Arnold Kling be wishing you at this point?

david January 6, 2012 at 4:21 am

A remark – possibly by Rortybomb or similar – that it was the (state-backed) establishment of clear claims to property and a refusal to recognize non-state-registered claims to have acquired property via private agreement that enabled property-focused investment by feudal lords to take off. If you buy something, you need to have some way of reliably knowing that the person selling it to you really has the right to do so.

Financial regulation has not taken the step of simply invalidating any privately negotiated contracts that don’t make their claims publicly known, or refusing to enforce ownership of any financial instruments created without some prior, public state recognition, but I wonder whether we’re inevitably heading there nonetheless…

Andrew' January 6, 2012 at 5:30 am

Isn’t this the bankruptcy process? People make claims, a judge arbitrates. The sum of the assets is worth less than the whole of the ongoing business, so this should be avoided. Thus the need for the living will especially for financial institutions. Money and bankruptcy are, as an aside, some of the primary responsibilities of government.

David Manheim January 6, 2012 at 10:16 am

Having a process that resolves claims is different than having one that requires the claims to be real.

As a coder, this is the difference between debugging and testing – I promise, testing is a better plan, because debugging gets complex. If I heard of a programmer that relied on debugging instead of testing to figure out what their code does wrong, I’d never work with them. And Our financial system does just that – it debugs through bankruptcy. And when the debugging process is impossible because everything is too complex, instead of throwing out the code, we call it “too big to fail”. At the end of the day, bankruptcy doesn’t fix problems, it causes a crash and causes a process to restart – this is fine, as long as the process isn’t the operating system (banking system.)

If you go through bankruptcy, the court realizes that the assets you borrowed against are not really yours, and so you can’t both pay the person you borrowed from, and surrender the assets back. This doesn’t make it impossible to do, it just makes it resolve correctly, if the threads can be untangled.

question the question January 6, 2012 at 10:42 am

Another problem with relying on bankruptcy proceedings to unwind assets is that lawyers are involved. This immediately reduces the pool of paid recovery from some high percentage of lost assets (it’s never full recovery) to a low percentage.

This is what’s wrong with libertarian proscriptions for legal remedy. The remedy never fully recompenses the victim, and more usually is far short of the mark.

Short Theta January 6, 2012 at 11:49 am

The objective of bankruptcy court is not to recompense the “victim”, but to divide up any remaining assets to the stakeholders based on their pecking order. You invest in a company or project because with the risk of failure comes the prospect of reward. If you invest in a complex company, you can objectively factor in higher bankruptcy costs in your analysis and should demand a higher rate of return. This would be a better process than complaining about excessive legal fees when your investment goes south.

question the question January 6, 2012 at 1:52 pm

In some cases, the risks are known and easier to account for. In other cases – and the Lehman implosion is a good example – the risks are impossible to understand, even by insiders, and thus utterly impossible to quantify. Witness the amount of time and energy it has taken to unwind the gordian knot, the massive confusion faced by counterparty upon counterparty, etc.

In any event, I agree it’s not a good idea to leave it to a bankruptcy court to make victims whole, which is why I wrote what I wrote. As the saying goes, an ounce of prevention is worth a pound of cure. Adapting regulations to fit the ever-changing landscape and better enforcing those regulations on the books are absolutely necessary if we don’t want to imperil the entire financial system. And frankly, to protect the populace from people like you who like to blame the victim.

david January 6, 2012 at 2:37 pm

By the time the sum of the claims is known to be smaller than the sum of the assets, it’s too late for the state to resolve correlated leverages. It only works for an individual bankruptcy, but bankruptcy en masse is too socially costly… nobody can hedge away the risk. This method of preventing conflicting claims occurs too late.

Fundamentally, if we expect investors to take existing mass leverage into account, then it should necessarily be public information. If you buy a house, its value depends on the value of other houses which then depends on how much they are leveraged – which presently one has no right to know. Even quantifying the liabilities of assets which you do have a right to know is extraordinarily difficult. This is bad for a system that wants to enable lots of financialization.

Highgamma January 6, 2012 at 7:17 am

An important point about rehypothecation is that it acts like a “leverage maximization machine”. If I place assets into an account with a small amount or zero leverage and those assets are rehypothecated, they can eventually have the maximum amount of leverage that the market will bear if there is demand for that leverage. That’s what happened in prime brokerage before the crisis.

BTW, rehypothecation isn’t evil. A brokerage firm that offers a margin account to its customers will not likely have access to the funding for that loan. They rehypothecate your assets to get access to the money that they lend you. It’s only when rehypothecation takes on its “leverage maximization” feature that it becomes dangerous.

Rahul January 6, 2012 at 7:28 am

I still don’t get the legal basis of allowing the practice: Shouldn’t collateral offered be assets that the debtor actually owns ? In re-hypothication it seems that the intermediary is offering as collateral funds that it merely has a lien on . What happens legally , if the Client does exercise his obligations but the Broker cannot? ( in a Client –> Broker –>Lender transaction)

Tyler Cowen January 6, 2012 at 7:31 am

Rahul, I would say that you do get it.

Norman Pfyster January 6, 2012 at 9:08 am

In the UK and under UK law, the secured party does actually have legal title to the assets they are receiving as collateral. In the US, it is trickier, but the secured party to the pledge is deemed to have control over the collateral, and so can trade it, if permitted by the pledge documentation. Under both laws, the more relevant point is that the same assets appears on both the secured party’s and the pledging party’s balance sheets.

Norman Pfyster January 6, 2012 at 9:19 am

The legal basis (contract law) is less troubling than the accounting basis. My confusion as a lawyer the first time I worked on these issues (derivatives, securities lending, repos, etc.) was the divorce between what I would think as a lawyer and what the accountants were saying. One of the huge issues (particularly in Europe) following the Lehman bankruptcy was that a business practice guided almost entirely by accountants got yanked into the legal arena, and the results weren’t pretty. The macroeconomic issue is that the same asset can appear on multiple balance sheets under current accounting practices, effectively inflating balance sheets everywhere (which is less rehypothecation rather than the accounting for the initial hypothecation/pledge).

Highgamma January 6, 2012 at 10:49 am

Norman, the effect that you are describing is a great example of “gross” versus “net”. With rehypothecation, the net borrowing against the assets will not exceed the market value of the assets. (The “market value” is actually somewhat nebulous since the lending could occur at different times with differing market values.) This is how rehypothecation becomes a leverage maximization machine on a “net” basis.

On a gross basis, accounting has the assets appear on the balance sheet as well as the liabilities against those assets. That does inflate the balance sheet, as it should. Net leveraged positions should not be put on a balance sheet. Gross position should be put on a balance sheet. In terms of the systemic risk created by rehypothecation, it is the gross effect (the so-called “daisy chain”) that causes all of the problems.

Once again, rehypothecation has a legitimate use. It’s the foundation for the entire customer margin market.

JPC January 6, 2012 at 2:18 pm

Highgamma – it may be the foundation for the OTC customer margin market. How about the futures markets? That’s a different story, and that’s the MF Global story. There’s no argument for a firm being able to take cash from futures margin accounts and buy other types of assets with them. They are just acting as pass throughs to the exchange. IN the OTC market its different and the price the OTC participants for margin is the possibility of credit risk to the counterparty, although that should properly be maitianed by margin agreements with downgrade and default provisions.

Highgamma January 6, 2012 at 3:24 pm

JPC, it’s actually the foundation of all customer margin account. If you have a margin account with Charles Schwab, your assets can be rehypothecated so that Schwab can raise the money to lend to you (as well as for their own account, and there’s the rub). The prime brokerage extension of this model is just that, an extension. I can even see the need to use one customers assets, which may have a more ready rehypothecation market, to help a firm raise money to provide to another customer with assets that are harder to rehypothecate, but using those assets to play the ponies? We all paid for the systemic risk from that one.

In a futures account, the futures commission merchant needs to post different types of collateral at different futures clearinghouses in order to meet margin. These requirements may be in different currencies. Therefore, I can see how an FCM would need to rehypothecate assets to meet their customer margining needs.

All that being said, the use of rehypothecation to use accounts as a source of funding is a huge source of systemic risk that we’ve not dealt with (or even talked much about) even though it was at the heart of Bear Stearns’ and Lehman’s demise. I was at a conference recently with some economists who are on the “inside” at the regulators, including the new systemic regulator. Their answers to my questions told me that this issue wasn’t even on their radar screen.

MF Global nay have gone one step further and just taken the assets from their customers’ account. That’s not rehypothecation. That’s theft.

JPC January 7, 2012 at 12:21 am

HG – Understood in all cases but the futures market. Here, only spot transactions are necessary for the broker, even cross currency, cross exchange. Futures clients are never asking for more leverage than the futures exchanges will naturally provide (nor do they get it ). Additionally, there is a pretty large difference between cash/margin/spot FX management and levering up the balance sheet to take 10 yr sovereign risk a la MF Global. The OTC and equity margin accounts are different.

JPC January 7, 2012 at 12:26 am

Add TBILLS to the futures list, but the point is the same.

Silas Barta January 6, 2012 at 11:32 am

Norman_Pfyster: Sorry if I’m being dense on this, but I don’t think you answered Rahul’s question.

In a normal mortgage it works like this: I get money from the bank to buy a house. The bank holds the title to the house. I owe the bank a stream of payments. If I satisfy the full stream of payments, the bank must shift the title to me.

In rehypothetication, the bank can (somehow) pledge the house as collateral for its own asinine bets. So here’s the problem scenario.

Bank wants to take out a loan to buy toxic assets. Bank borrows from hedge fund, putting up its title to my house as collateral.

Bank’s new toxic assets plunge in value and it can’t service the debt to hedge fund.

Hedge fund seizes my house title due to bank’s failure.

I satisfy my full mortgage obligation and expect title to the house.

Oops! Now the hedge fund and I both think we own the house! But the bank should never have been able to transfer the house away without the attached obligation that it be transferred to me in the event I can pay the mortgage!

What am I missing here?

AndrewL January 6, 2012 at 12:30 pm

I don’t think the bank can put up your house title as collateral. I think it can only put up your mortgage. Your mortgage should have a credit rating attached to it that defines it’s resale value. the bank can use that value as collateral, but it typically dosn’t do that. Instead it will sell your mortgage to someone like fannie / freddie to obtain capital, while continuing to hold your title and servicing the mortgage.

I think this is how it works….

Norman Pfyster January 6, 2012 at 12:42 pm

1. Most states don’t operate mortgages that way. In most states, mortgages are liens, which is to say that the lender is a secured credtior of the house owner, where the borrower has title to the house. I realize that in some states title is with the bank until satisfaction of the loan, at which point the title passes to the borrower, which is the example you described and if you live in one of those states, it might seem typical to you. I was describing the more common US mortgage law.

2. Because title mortgages are not common, I’m not sure how the title can get passed, that is to say, whether the title can pass without the corresponding loan passing with it.

3. Ignoring the housing context for the moment, you have described the delivery failure situation I described earlier; that is what happened with Lehman. As for the law that permits it, in the UK, title passes with the pledge and in the US, it’s done by contract (somewhere in the small print of your brokerage account, there will be a provision permitting the broker to rehypothecate the collateral posted on your margin account).

derek January 6, 2012 at 12:52 pm

It isn’t a house. These are brokerage accounts where you put up collateral as a percentage of your positions.

In any case, if you are arguing against fractional banking, then make the case.

Silas Barta January 6, 2012 at 1:44 pm

So, general question then: if I put up asset A as collateral to make some transaction (say, short-selling), doesn’t that mean that once I’ve closed the transaction (say, covered my short), I get the asset back? Some more questions (not corresponding to the above 1-3):

1) Is the broker really allowed to use that collateral as his own collateral for his own transactions?

2) If yes to 1), and things go sour for the broker, is he then required/allowed to give my collateral to that counterparty even after I (or if it’s still possible for me to) close my transaction?

3) If yes to 2), how is that possibly legal, and why do we let it continue?

4) If no to 2), how did anyone lose money through MF Global’s rehypothecation practices?

Tummler January 6, 2012 at 11:36 am

Why wasn’t there a corresponding liability?

Ricardo January 7, 2012 at 1:07 am

On the value of rehypothecation, the U.S. used to restrict the use of rehypothecated assets to federal, state and local government obligations. This regulation was relaxed so that now customer assets can be used in repos or used to purchase a variety of much riskier assets including foreign sovereign debt.

Second, while the U.S. restricts rehypothecation to 140% of customer liabilities, the U.K. has no such limit and the U.S. does not restrict transferring U.S. assets to non-U.S. subsidiaries as long as a broker can get a customer to sign an agreement with small print authorizing such a transfer.

Whatever the value of margin trades, would it really be so bad to go back to the pre-2000 system the U.S. apparently had?

Silas Barta January 6, 2012 at 8:48 am

How is this remotely legal or accepted practice?

Try “rehypothecating” your mortgaged house, see what happens.

Norman Pfyster January 6, 2012 at 9:25 am

Your house has almost certainly been rehypothecated (albeit indirectly). The mortgage on the house is the pledged asset (technically, a mortgage is the pledge itself, depending on the state). The mortgagee (the bank) has probably sold/repackaged your mortgage into a mortgage-backed security, which is one of the primary assets that gets traded and rehypothecated.

Rahul January 6, 2012 at 9:46 am

So is the transaction necessarily asymmetric? i.e. If a mortgagor defaults the mortgagee has a secured interest in that asset (house); but conversely if a mortgagee is unable to discharge his obligation of redemption (on successful loan repayment) what legal recourse does the mortgagor have?

Not sure if I am clear, but another try: If the mortgagee somehow fails, do the interests of the mortgagor win over the rights of the rehypothicated MBS holder? What’s the legal precedent?

Andrew' January 6, 2012 at 9:56 am

“How is this remotely legal or accepted practice?”

It’s not against the law (yet) to not get paid. Individually it’s not a big deal to say “I promise to pay, if I have the money, and if not, you get the collateral, if I have the collateral.” Systemically, that’s a different story, not least of all because when everyone else is doing it you have to do it to keep up, and that keeps stretching the rubber band further than it should go. Eventually it snaps back. It’s still okay. We just don’t like the result.

Norman Pfyster January 6, 2012 at 10:31 am

I don’t understand your question. The failure at issue with rehypothecation is the failure to redeliver the pledged asset. When you discharge a loan, nothing gets returned to you; the effect is that the pledge/lien is also discharged.

Rahul January 6, 2012 at 11:07 am

If this unwinds who gets the higher priority over the mortgaged asset; the mortgagor or the security holder?

SirSpider January 6, 2012 at 3:48 pm

Rahul — What do you mean by “the mortgagee fails” ? You as borrower-mortgagor pay back your loan in full. The lender-mortgagor* should then file a release of lien with the local authorities, as per the terms of the mortgage contract. If the lender fails to do so, you request them to do it, and if not, you could sue them, or go to the newspaper for an embarrassing story.

*Thanks to securitization, the mortgagor with the right to foreclose might be a servicing agent, or MERS, or it might be the ultimate MBS holder with the right to payment.

Silas Barta January 6, 2012 at 10:26 am

@Norman_Pfyster: The description of rehypothecation above shows it happening to the same asset an arbitrary number of times, so the fact that the house has already been (re-)pledged to someone else shouldn’t distinguish it from the kinds of rehypothetication that seem to be accepted.

In a mortgage, the bank holds the house as collateral for your stream of payments. It is against the law to promise the same people the house as collateral. You have to set it up so that, “Bank A gets the house first to cover the debt, and THEN anything else can be claimed by Bank B.” If you apply to banks A and B to get two 50% LTV loans for the same house, without making sure the other knows, you will be in deep sh**. And probably even if they do know.

But when it comes to John CorRUPTzine, you can apparently pledge the same house to everyone in the world and get enough money to temporarily buy the entire world economy, and no one blinks twice?

Norman Pfyster January 6, 2012 at 10:39 am

You are confusing multiple hypothecation with rehypothecation. Yes, a person would have problems pledging the same asset to multiple parties because, as a contract matter, you normally have to pass the assets free and clear of other liens or obtain the permission of the counterparty to take a second lien on the asset. Without those contractual protections, the subsequent pledgees are SOL under the law (assuming perfection of the first lien). Rehypothecation, on the other hand, is the use of the pledged asset as collateral by the pledgee, not the initial pledgor.

Silas Barta January 6, 2012 at 11:20 am

I didn’t just lose billions of dollars in assets that customers thought they owned free and clear, so I’m probably confusing things a lot less than the supposed experts.

mofo January 6, 2012 at 9:32 am

So what happens in the case of like a MF global, where an entity, presumably, has done this extensivly and goes belly up? If this article is right we should be seeing some sort of chain reaction all along MF Global’s trading partners, right?

Dan January 6, 2012 at 10:19 am

The story seems to be that MF Global’s retail brokerage accounts were used as collateral for their trading partners. Their trading partners have been made whole at the expense of the retail account holders.

mofo January 6, 2012 at 1:54 pm

My point is, are we going to see a chain reaction byond MF Global, and, if so, how extensive will it be? Thanks for the link, btw, it was very informative.

question the question January 6, 2012 at 2:00 pm

Yes, but according to Short Theta upthread, these retail brokerage account holder idiots should have understood and factored in that risk when they entrusted their funds to the nefarious fly-by-night outfit known as MF Global.

Thank god we have free marketeers banging the drum for less (or no) regulation.

Short Theta January 6, 2012 at 3:29 pm

Actually, I was just explaining bankruptcy proceedings. But I wouldn’t mind explaning which firms and accounts the FDIC guarantees and which ones they don’t.

question the question January 6, 2012 at 5:18 pm

No, you really weren’t “explaining” anything. You were proselytizing about how the bankruptcy system ought to work, and laying blame at the feet of those who misprice risk often due to disinformation or outright fraud.

The FDIC doesn’t guarantee brokerage accounts; SIPC does.

Short Theta January 9, 2012 at 11:24 am

Among the investments that are ineligible for SIPC protection are commodity futures contracts

TallDave January 6, 2012 at 10:09 am

Yeah… I had to explain to someone at Megan’s the other day why a U.S. default would be so catastrophic. I don’t think people realize how much interconncted risk there is. It’s an issue that should have been more fully addressed in the bailouts.

Silas Barta January 6, 2012 at 10:28 am

That’s part of why I shrug my shoulders whenever someone warns me that a particular policy will mean “sharply curtailed lending”.

Sorry, folks, but that if that’s what a sound financial system and well-defined asset rights require…

NAME REDACTED January 6, 2012 at 10:34 am

Unfortunately, interest rates that are well below what they should be lead to excessive rehypothecation. This puts excessive risk into the market.

bluto January 6, 2012 at 10:46 am

It’s a little spooky that this is normally only discussed on goldbug pages.

Becky Hargrove January 6, 2012 at 10:52 am

What started up all the questions for me was when the evening news ran a story on midwest farmers who lost everything with MF Global, sometime in the last month or so. There was a flurry of discussion about rehypothecation in which we were basically assured: even though firms in other countries could engage in rehypothecation, firms in the U.S. could not. Although firms might shift some of their accounts overseas to take advantage, it was supposedly a minimal practice and there was no need to worry. Now the story has shifted to yes, the U.S. does it too, just not to the same degree.

nanute January 6, 2012 at 12:56 pm

Saw your comment at Aysmptosis, and wander over here. With regard to MF Global, it appears that the transactions could only take place through their UK operations. I don’t think it is legal in the US. Having said that, it appears to be a difference without a distinction. That money is long gone, and it’s doubtful it will be going back to the original parties.

derek January 6, 2012 at 3:30 pm

Yes. By law I think they can do it with something short of 50%. The farmers, if they read the contract documents, agreed that the money could be transferred between branches of the brokerage firm, into other countries where they could legally do it up to 100%.

Rahul January 6, 2012 at 3:48 pm

How long was the contract document? Honestly, other than lawyers how many people read the whole contract document? Maybe there ought to be legislation mandating the maximum complexity of a contract when drawn between asymmetric parties.

question the question January 6, 2012 at 11:36 pm

You’re clearly on the wrong blog ;)

Alan January 6, 2012 at 2:24 pm

I will be better off as a result of rehypothecation. I’m not sure when and the details of how elude me but I have faith in the Hidden Hand.

Jason January 6, 2012 at 2:31 pm

I wonder if all these tricks to effectively create money out of thin air would be necessary if the Fed relaxed monetary policy. There seems to be a market demand for money and the supply is being fed by fractional reserve banking, rehypothecation, commercial paper, etc.

Scott Sumner was interested in Dvisia M4 the other day as a more accurate monetary aggregate and it includes all kinds of stuff (like M3, MZM etc did in the past).

Another way to put this: are all these risky financial instruments the result of monetary policy on average being too tight since the Volcker Fed?

NAME REDACTED January 7, 2012 at 6:38 am

You have it backwards. When the fed relaxes monetary policy is when this sort of thing happens. Individuals and institutions needs interest rates to pay their nominal contracts so they must lever up more and more to do so and at the same time the artificially low interest rate makes the risk much cheaper.

In order to fix this, interest rates have to rise naturally and failures have to be cleared out.

Jason January 7, 2012 at 3:03 pm

Your statement is false.

There was more rehypothecation (~ $4 trillion) when interest rates were higher in 2007 (4.5% to 5%) than when they were lower in 2009 (2.5% to 3.0%) using the 10 year treasury benchmark. In 2009 there was only about ~ $2 trillion in rehypothecation. And there was a consistent reduction as interest rates went down.
See figure 1.

There was also more commercial paper outstanding in 2007 when interest rates were higher:

Olaf January 6, 2012 at 4:36 pm

As a finance lawyer / banker (non-U.S.), I make the following conceptual observations, albeit from a “micro” not “macro” perspective:

Concept 1: let’s call it custodial deposit
- Titel: generally, if ones deposits assets (e.g. securities) with a bank or similar institution, one expects to retain titel. The bank is a custodian.
- Insolvency of the bank: depositor has a priority claim for delivery of the asset (based on the title one still has) and there is no competition with any other creditors.
- Remuneration: Since the bank derives no benefit from this transaction, and the depositor bears no insolvency risk, the depositor actually pays the bank a custody fee.

Concept 2: ordinary bank deposit (usually cash but could also be other fungible assets such as securities). This is the other end of the spectrum of legal possibilities – you could call it lending to the bank
- Titel: The depositor transfers title to the bank and in return obtains (only) an unsecured claim against the bank for “repayment” in kind.
- Insolvency: Due to loss of title, the depositor has only an unsecured claim and in insolvency of the bank he is in full competition with all other unsecured creditors (and of course “behind” any secured creditors). This is the situation with any ordinary cash bank deposit (leaving aside public or private deposit insurance – another complication). As a compensation mainly for the insolvency risk of the bank, the depositor here gets a lender’s return (i.e. interest) from the bank, also because the bank obtains title can make use of the deposited assets (especially cash which it can lend further pursuant to the fractional reserve system).

Now concept 1 can be varied e.g. one can borrow by way of margin loan (or otherwise borrow from the bank) against the assets in question (let’s assume securities) as collateral. This is “hypothecation”. Leaving out complexities and peculiarities of local laws, this is akin to a mixture of concepts 1 and 2, i.e. before default one basically retains title and after a default the title passes (one way or the other) to the bank or a third-party designee of the bank. (Repo again works differently but is ultimately also a variant of hypothecation.)

So BOTH concepts (as well as combinations and variations) coexist, point is only that apparently, and perhaps surprisingly, these so-called brokerage / deposit accounts work more like concept 2 instead of concept 1 in the “anglo-saxon” legal world. Meaning that subject to all sorts of intricacies that just obfuscate the situation, e.g. the requirements and restrictions of “reypothecation”, the depositor in an insolvency of the bank (think MG Global) gets into a situation of concept 2 (unsecured creditor with no title to his assets) instead of concept 1. So question for me is just (i) do sophisticated depositors (e.g. hedge funds) and ordinary people realize this? and (ii) are they requiring and obtaining adequate remuneration i.e. a lender’s/depositor’s return (commensurate to the insolvency risk they are taking)? Remember these accounts are even more at risk than “ordinary bank accounts/deposits”, because I assume deposit insurance doesn’t cover them… It seems an information/transparency problem, not a conceptual impossibility or even intricacy – either you deposit something with a custodian or you lend someone something on an uncured basis. You just have to plan accordingly, and something seems to be going wrong at that end, judging from people’s (depositors’) surprise on Lehman and MG global at not getting their money or securities back in solvency!

Olaf January 6, 2012 at 4:42 pm

last two lines should read: “…either you deposit something with a custodian or you lend someone something on an un*se*cured basis. You just have to plan accordingly, and something seems to be going wrong at that end, judging from people’s (depositors’) surprise on Lehman and MG global at not getting their money or securities back in *in*solvency!”

mike sproul January 6, 2012 at 6:04 pm

“I would define it as “an asset used as collateral more than once, at the same time.” ”

You got this one wrong. If a broker has allowed a customer to invest $1 so as to control $10 worth of stocks, it is only because the broker has in place a system that will liquidate the customer’s position before his total losses reach $1. Either that, or the broker has a $9 lien on the customer’s house, in addition to the $1 cash. Try borrowing $200,000 against a $100,000 house sometime and you’ll see what I mean.

Olaf January 6, 2012 at 7:14 pm

What you write is 100% correct – but this is not in conflict with Tyler’s description in any way. The “system that will liquidate the customer’s position” legally breaks down into the following pieces: broker lends $9 against collateral of at least the loan amount. This initially consists of $10 worth of stocks (or $9 worth of stocks and $1 cash). Even this needs to be broken down further: the broker lends the $9, the customer utilizes that loan (together with his own $1) to buy the stocks, with the broker insisting that the orders are transacted through him and the stocks are held on in a securities account of the customer (= he has title), which account is pledged to the broker to secure repayment of the loan. This is the “system”. If the collateral value drops below (or nears) the loan amount, the loan agreement allows the broker to require repayment of the loan, unless further collateral is posted. If the customer does neither (voluntary, partial repayment nor posting of further collateral), the broker “forecloses” (assumes title) or “liquidates/enforces” the collateral i.e. the securities (exact mechanism depends on local law) and applies the proceeds as discharge of the amounts owed to him as per the required loan repayment. QED: PRIOR to default of customer/foreclosure, the securities remained “property of the customer”, who had only pledged them as collateral to the broker. To the customer’s surprise it seems possible that the broker, despite having no title/ownership on those securities and the lien on these securities not yet being enforceable, was legally allowed to grant a third party creditor of the broker a first ranking collateral over those securities (to secure BROKER’S obligations). This is what is meant with Tyler’s description “an asset used as collateral more than once, at the same time”. This result is surprising because it is more than just passing through the original collateral – the risk and the secured obligations are entirely different. For the customer the reypothecation carries the same risk as if the broker could just secretively sell those securities in the market, i.e. the securities are effectively, sneakily gone and the customer has only an unsecured claim – albeit the same situation as an ordinary bank depositor faces. BUT the brokerage customer thought the securities were HIS i.e, that they were insolvency remote in an insolvency of the broker (as long as the customer himself hasn’t defaulted and the collateral value has remained sufficient!!) whereas the ordinary bank depositor KNOWS that is money is immediately passed along the famous daisy chain i.e. the cash deposited is NOT his any longer and hence is NOT insolvency remote but is rather COMMINGLED with the bank’s OWN assets….

I think this is the usual intuitive understanding. From it flows the concept of assets used “more than once as collateral”.

However, conceptually, the entire transaction COULD be structured differently: instead of granting the customer a margin loan, the broker could buy securities from his own funds, in his own name, retain them for the time being on its own accounts and only promise to transfer to the customer those securities at some future time. Structered this way, in the meantime the broker could do what he likes with those securities – they are still his!! If broker goes bust before remitting the securities to customer at the end of transaction, customer has only unsecured claim. That’s the situation with any ordinary cash account. Noone expects the money to “stay with the bank”. That’s why deposit insurance is necessary – to “beef-up” the recovery position of depositor beyond that of an unsecured creditor. But the securities depositor – surely he thought the securities are his and are collateralised only to secure his own obligations (something he controls) so if broker goes bust and customer was always faithful regarding his own obligations – he will be quite surprised that his securities are effectively gone because they are being claimed (with priority) as collateral of a third party who loaned money to broker against those securities – lucky third party – but unlucky customer it seems …

The entire point is that under anglo-saxon law and practice of brokerage, the brokerage customers are getting a transaction structure which although quite well possible/structurable/explainable – is apparently (judging from people’s reactions) counterintuitive and hence surprising in a bad way! No magic, no logical issue, just a further unrecognized LEMON (from brokerage customers’ perspective) in the market for financial products!

mike sproul January 7, 2012 at 4:01 pm


“the broker, despite having no title/ownership on those securities and the lien on these securities not yet being enforceable, was legally allowed to grant a third party creditor of the broker a first ranking collateral over those securities (to secure BROKER’S obligations). ”

This is not clear. The broker has lent $9 to the customer so that the customer can control $10 in stock, which stock is held by the broker. The broker then borrows $9 against the stock he holds. At first it looks like $10 in stock backs $18 in loans, but this can’t happen, as no lender would lend without having a clear claim to adequate collateral. As soon as the $10 worth of stock falls to $9 the customer’s position is liquidated, leaving the broker with $9 of stock securing his $9 loan. There are no overlapping claims to the same collateral.

Henry Thornton (1801) made this same mistake in his explanation of fractional reserve banking.

Olaf January 7, 2012 at 9:32 pm

I agree there is no collateral overlap in your example (the reasons are complicated and would confuse.) However:
1. Unrelated but important: Between broker and customer, customer is not getting what he expects i.e. title to the stock which he believes no-one can take from him unless he defaults on his loan/margin call. See further discussion below and PWC description of Lehman situation. It’s the “lemon” problem.
2. Re “systemic” aspect and relation to fractional reserve banking (I am less interested in this yet): I think there WILL be collateral overlap in your example at the latest once the “lender to your broker” himself rehypothecates (second level) that stock to its own lender. You just have to take it one step further (I think this is legally possible and fear that is what is causing the “daisy chain” problem so often quoted). And another step…. At the end only one person can have a robust title to the asset itself, all others just have unsecured paper claims.

mike sproul January 8, 2012 at 12:17 pm

1) But the customer knows about margin calls, etc., so he is getting what he expects.
2) No lender will lend on collateral that is already pledged to someone else. If there is a lender to the broker, then either the broker offers some other collateral to the lender, or the lender takes the collateral previously held by the broker. The effect is to eliminate the broker from the chain, and the lender has, in effect, lent directly to the customer.

Bryan Willman January 6, 2012 at 9:27 pm

This whole discussion illustrates three things:

1. The far-libertarian view that bank/broker risk should be managed by market forces is both practically and politically hopeless. I open a brokerage account, do NOT draw margin, the broker pledges those assets for who knows what, goes broke, and I’m screwed. Regardless of whether there is some grand theory that would make the system as a whole more safe, this will not work in the real world. (Like, nobody will be able to pay their taxes, so there will be no police…)

2. The Stephan Waldman view that it’s bad for the well off to “hoard” assets by parking money in treasuries or the like is also pointless – nobody has any idea where assets go, it’s impractical to store money any way except by lending it to somebody or “hoarding” something (real estate, canned goods, whatever.) Even if a person invests in starting and growing operating companies, the money is all held in banks that do random things with it – right back to systemic risk and the need for insurance.

3. There are practical problems for even the wise avoiding this – can a retail investor actually open a custodial only account? At what cost?

At least if cash is parked in FDIC accounts, one has a clear claim (bank failing to provide money amounts to bank failure) and there’s strong insurance. But that insurance can’t (by normal insurance rules) cover a total systemic failure – if every bank folds, the only way to insure those accounts is to print insane amounts of money. Failure to do so creates collapse because nobody can pay anybody for anything.

Highgamma January 6, 2012 at 11:25 pm

Bryan, in case 1, if you do not open a margin account, then your assets cannot legally be rehypothecated. For question 3, you can open a cash account (sometimes called a “Type 1″ account) with Charles Schwab or most any other retail broker tomorrow. This account would not allow rehypothecation. You also get the added benefit that your shares cannot be lent out to short sellers. In fact, by law from my recollection, any IRA account is a cash account.

Ricardo January 7, 2012 at 12:56 am

If this article linked to above is any guide, though, a U.S. broker merely needs to get its clients to sign a securities lending or margin agreement with a U.K. subsidiary. Once the ink is dry on that agreement, the broker has legal authorization to transfer the client’s assets to the U.K. and can use them without limit as collateral for its own trades. In the U.S., rehypothecation is restricted to 140% of the client’s liabilities; in the U.K., there is no limit at all unless the contract explicitly states the limit.

So, yes, it appears to be possible for Americans to avoid this by insisting on plain vanilla brokerage accounts with no derivatives, no margin and no shorts.

On the other hand, what possible justification is there for a U.S. company to be transferring client assets to the U.K. or for the U.K.’s extremely lax regulations on this? Canada apparently restricts this practice and perhaps as a result, Canadian clients of MF Global have all gotten their money back, according to Calculated Risk.

JPC January 7, 2012 at 12:32 am

The point of the MF Global case is that the accounts were largely futures accounts, on regulated exchanges, with daily margining requirements. If that could be manipulated, the system is really broken. IE these are not “margin” accounts in the sense being talked about here, although the futures are margined on the exchange, the broker has no need to come up with more funds than the customer provides.

NAME REDACTED January 7, 2012 at 6:40 am

What you say is true, but they weren’t manipulated. They were stolden. I am not being hyperbolic here. This was the heist of the century.

nanute January 7, 2012 at 7:59 am

You are absolutely right. This was the heist of the century. When the dust settles I’d be willing to wager we’re going to find out it was all perfectly legal. There’s the problem.

Peter Russell January 7, 2012 at 10:44 am

The only way to be safe, it seems, is to put no more than ten percent of your money in any one investment, and spread those investments over the entire spectrum of investments, including everything from cash in an FDIC backed US bank, to, to…to…

Richard January 7, 2012 at 11:00 am

A basic legal observation form a U.S. lawyer that seems to be missing from the discussion:

Imagine A borrows $100 from B, and pledges collateral X in exchange. B now has a lien on X. Now, imagine that B borrows $100 from C, and wants to pledge something in exchange. B cannot pledge X, because B does not own X. All B can pledge is its own lien on X. So let’s assume that B does this, and subsequently defaults. C can foreclose on B, meaning that it can seize B’s lien on X. But C cannot seize X itself unless A defaults. In other words, B cannot pledge to C an interest in X that is greater than the interest that B itself had.

The discussion in the comments implies that rehypothecation makes A vulnerable to losing X if B defaults on its debt to C. But this is not how secured lending works, at least under US law.

Thus, is it possible that we are greatly exaggerating the risk to the original customer in these transactions?

Rahul January 7, 2012 at 11:55 am

Thanks! This was exactly what I wanted to know. If this is true I do not see any legal problem since under any scenario ownership rights are well defined. e.g. No mater what hanky panky goes on between B and C, A’s asset is always safe (so long as A pays his debt)

Olaf January 7, 2012 at 2:12 pm

what you describe is what always works, the regular case of secured lending so to speak. Actually it’s like that under most laws I assume. This is akin to the example above of the house-owner/mortgagor who’s mortgage loan (loan + attached mortgage) is packaged, repacked etc. Indeed he has no worries beyond fulfillment of his own obligations. But the consequences of Lehman UK insolvency and now MF Global have shown that this “regular secured lending structure” is simply NOT what is going on here. Otherwise noone would be having this discussion and noone would have lost his money. So the point is: what is going on here (if it ain’t “regular” secured lending)? And it seems that the brokerage customers are not transacting in the form of secured lending but rather in a different transaction type pursuant to which the customer actually loses title (or at least part of it – by way of the broker being somehow empowered to “rehypothecate” those assets i.e. creating security over them to secure his OWN obligations IRRESPECTIVE it seems of the original customers fulfillment of his obligations). I fear you haven’t found a flaw in the analysis but rather you are, as we all are, stumbling over a flaw in reality, namely the specific defect in the product “brokerage account” (at least certain types it seems). Once one accepts that, one has to figure out what is the model here – if it isn’t secured lending. Hence my ruminations above and comparisons above of how different transaction structures give customer different positions in an insolvency of the brokerage house. It doesn’t have to be regular secured lending and it seems it isn’t. Problem is that apparently noone was aware (although after the debacle in Lehman UK insolvency in 2008/2009 where hedge funds lost large amounts of securities which they thought were just pledged under secured lending transaction and instead were gone, one would have thought the word got round…).

Richard January 7, 2012 at 3:36 pm


In a standard brokerage account, the customer has title to the assets. See, e.g., So either Lehman UK and MF Global didn’t use standard brokerage accounts or the problem here is something other than rehypothecation. For example, isn’t the problem at MF Global that the company just “can’t find” its customers’ assets? This isn’t a problem of legal title per se, but rather of sloppy accounting and/or fraud, it seems to me.

Olaf January 7, 2012 at 4:02 pm

I urge you to read this reuters article which was already referenced by someone above. It is actually a legal one, and it is much better than the rather confused article that Tyler linked to in his original post. I think you will find your thoughts (e.g. “wasn’t it fraud instead of legalistic mechanics”?) addressed and largely refuted after which we can welcome you in the “surprised and concerned” camp. :)

Richard January 7, 2012 at 5:20 pm

Olaf thanks for that article. It seems not to address squarely the question you and I are both concerned with, which is, how does rehypothecating subordinate or eliminate the original customer’s lien on the collateral? So yes, I’m now with you in the “surprised and concerned” camp, and it would be nice if someone who really knew the law in this particular area in particular would weigh in.

Norman Pfyster January 7, 2012 at 9:05 pm

I did answer above. One, UK law is different than US law. Second, in the US, it would depend on the form of the pledge. In situations where rehypothecation is possible, control (if not title) over the asset passes to the secured party. Combine control with a contract right to rehypothecate the collateral assets, and the secured party can transfer the assets to another party. The party posting the collateral has a contractual right to the return of substantially similar collateral.

Olaf January 7, 2012 at 9:47 pm

Excellent, thanks. The last sentence is crucial obviously, it confirms that the rehypothecation agreement (in fine print I assume!) effectively expropriates the customer or put differently completely switches the transaction structure from being a borrower who collateralises his borrowings with his own assets (retaining ownership) to being the holder of an unsecured claim for return of a substantially similar asset. I hope he gets paid just as well for this as other unsecured lenders to the broker will require for such counterparty risk (i.e. CDS spread), especially with regard to the excess 40% over his margin loan amount, i.e. regarding that part of the assets that he hasn’t even borrowed against! (I am referring to the ominous 140% rule…in UK even unlimited expropriation of his assets I understand.)

Olaf January 7, 2012 at 6:43 pm

Richard – I agree that would be desireable and that the *exact* mechanics are not yet described in the sources/discussions at hand. But even without, I know enough to believe that it does have the described effects (one way or another). I am more concerned that this runs afoul of reasonable expectations of financial customers as to what position, structure and ranking they are getting from their brokerage or deposit or whatever accounts. (That even you as a US lawyer have difficulty reconciling this outcome with your legal understanding just underlines this for me.) Next highest concern for me are the systemic accounting and risk implications – if customers (especially commercial ones) are accounting for the deposited assets as e.g. securities to which they have title – whereas in reality they have only an unsecured claim against the broker in the latter’s insolvency, well then goodnight…and if the customers themselves don’t realise they are taking insolvency risk of the broker, then the counterparts of that customer are surely even more clueless that they are taking that risk indirectly…

JPC January 7, 2012 at 7:13 pm


So there are several questions here:

1. Are the rules in place if followed correctly, sufficient to guarantee the reasonable safety of customer funds versus broker dealer risk accounts, and do customers (or anyone else besides the brokerage lawyers) really understand those rules?

2. In any case, were these rules followed, or did something extra-legal occur in this case?

Olaf January 7, 2012 at 7:50 pm

I understand that MF Global may not be the ideal precedent for the pure legal workings of rehypothecation. It seems that they had massively commingled customers’ assets with their own, probably to a large extent illegally – so Richard had a point on that, i.e. MF Global is probably just trying to obfuscate by citing various quasi-legitimate “commingling” models such as rehypothecation. So on 1 (“if followed”) I would say: still not sufficient even in absence of fraud, thanks to the strange rehypothecation practice which apparently no-one really fathoms in practice. On 2., see above, I don’t think the rules were followed at MF Global, so that case represents the worst of all worlds probably.

Briefly back to the legal effects of rehypothecation, here a quote from an IMF paper, emphasis (***) from me:
“Thus, prime brokers and banks would rehypothecate their customers’ assets along with their own proprietary assets as collateral for funding from the global financial system. Lehman’s administrators, PriceWaterhouseCoopers (PWC), confirmed in October 2008 that certain assets provided to LBIE ***were rehypothecated and no longer held for the customer on a segregated basis and as a result the client may no longer have a proprietary interest in the assets. As such, LBIE investors (e.g., hedge funds) fell within the general body of unsecured creditors.***”

Richard January 7, 2012 at 11:08 pm

Olaf, that “may” in your quote form the IMF paper is pretty cute, isn’t it! Perhaps the answer is that the law is actually fuzzy here, that no court has had occasion to address squarely a customer’s title to collateral that has been rehypothecated. I suppose that, as a matter of contract law, the customer could empower the broker to give away the collateral if the broker wants to, perhaps in exchange for this rather unencouraging promise to substitute similar collateral. And maybe the contract is vague — perhaps deliberately so — about the customer’s rights to the original collateral once the rehypothecation occurs.

Rahul January 8, 2012 at 4:59 am

I’m no lawyer but isn’t the Equity of Redemption principle applicable here? It seems the legal systems have regarded the right of a mortgagor to get his property back pretty sacred and have struck down any contract clauses that obstruct this right.

Olaf January 8, 2012 at 5:22 am

I fear the effects of rehypothecation are “in rem” such that these kind of “position attributes” of a regular mortgagor are not applicable. If the law has intentionally established a way for you as owner to legitimately lose title in rem (for the law that’s just a binary technicality) I suspect equity is not going to prevent that or get it back to you.

Thankfully it only affects the leveraged brokerage customer (and only some account types and only subject to signing the fine print) and apparently not the leveraged homeowner. But the comparison confirms that it is slightly scary.

Rahul January 8, 2012 at 8:57 am

How much longer before an enterprising financial engineer tries similarly with house mortgages? I don’t see a fundamental legal difference between houses and other assets offered as collateral.

Olaf January 10, 2012 at 4:50 am

General comment (coming from a so-called civil law jurisdiction) and leaving aside the collateral idea (which is just one of many cases in which possession and title diverge):
- There is still the possession / control criterium to which the law usually attaches quite some significance (relinquishing control/possession reduces your legal protection against fraudulent etc transfers).
- Further, the law tends to cross the line of “loss of title to asset in exchange for claim for return of substantially similiar quality assets” only for fungible assets.

So the concerns would be low at the end of the spectrum represented by the “owner of an individual home who lives in it himself” and it would always be highest in the case of the “owner of a basket of uniform widgets/securities/dollar bills who hands them over to someone else (for custody, rental etc.)”. In the latter case you cross the line quickly to structure in which the law assumes an implied willingness not to retain title to the assets and instead be content with that “claim for return of similar items”. This can kill you in insolvency of the possessor of course but I may remind again that noone expects segregation of the cash funds he has deposited in his bank account… fungible securities are not soooo much different from money. But a house is! So I wouldn’t be worried on that.

It’s all shades of grey surely but there are patterns to be sure and they do make sense in principle I think. That’s why my mystification is still triggered more by how common yet apparently unappreciated this specific mechanism was / is in the relevant markets (at least up to Lehman). For me it’s an information problem i.e. why didn’t sophisticated people see they were getting a lemon although the information was in the public domain?

Olaf January 8, 2012 at 5:10 am

Could be but I think “may” probably just alludes to the curiosity that the “expropriation” is conditional, i.e. broker may not outright sell customer’s asset, but he may rehypothecate it. So whether customer still holds title or “may” have lost it depends on whether that third party beneficiary has exercised its collateral rights i.e. whether the broker has defaulted on his own obligations or not. (We now just have to find someone involved in the outcome of the Lehman UK (LBIE) insolvency to share with us …)

Richard January 8, 2012 at 11:26 am

It’s also possible that the IMF article was simply noting that, since the original broker was no longer in possession of the collateral, the customer no longer had a property-based claim against the broker, but rather only a general (unsecured) claim. If A entrusts property X to B, who then pledges X it to C, and B later defaults on its obligations to A, then A would have a general claim against B plus potentially a property-based claim against C for X. I say “potentially” because it may depend on legal questions such as whether C had notice that title in X belonged to A, whether C was a “good faith buyer in the ordinary course of business,” and so on. Again, these niceties may never have been litigated in the rehypothecation context, forcing observers like the IMF article writer to speculate. Olaf certainly is right that the law of real property and mortgages is much clearer in these regards. And, crucially, in that context the mortgagor retains possession of the collateral (i.e., the borrower possesses the house), making these types of shenanigans impossible.

Richard January 8, 2012 at 11:40 am

I should have added that yes, Rahul is right that the equity of redemption is relevant here. The only wrinkle is that it technically would provide a right against the original broker, not against other firms down the line to whom the collateral had been rehypothecated. My point regarding property claims was about the original customer’s power to go after those folks as well. But the more general point is a good one: the equity of redemption is evidence that courts traditionally have been wary of loan agreements whereby a mortgagor could make all payments when due and still somehow lose title to his collateral.

Rahul January 8, 2012 at 3:12 pm

The origins of rehypothecation seem respectable so long as intermediaries were offering the mortgages or liens as collateral. This is a pledge of the claim to the asset not the asset itself.

At some point down the line things got fuzzy and a mortgagee decided it was OK to re-offer the asset itself (which the mortgagee only conditionally holds title to, if at all) as collateral; and that is what stinks to high heaven.

Furkan January 31, 2012 at 7:57 pm

This is a smart blog. I mean it. You have so much knwgledoe about this issue, and so much passion. You also know how to make people rally behind it, obviously from the responses. Youve got a design here thats not too flashy, but makes a statement as big as what youre saying. Great job, indeed.

gcdzqeqzwak February 3, 2012 at 12:39 pm

w79kkl uoankrmooeaf

Comments on this entry are closed.

Previous post:

Next post: