Hacker News Reader

Comment by logicalmind | Hacker News Reader

I work in finance and there has been a lot of interest recently in creating financial products that are eerily like mortgage backed securities, but for other types of loans. A lot of these are (yikes) unsecured loans. A person loses their job and needs a personal loan from a bank or credit union to cover their expenses until they find a job. A person buys a car, an RV, etc.

In order to offload the risk of these loans, the original lender is packaging them up together. Then they can sell ownership of the consolidated loans to other financial institutions. The original lender is only required to retain around 10% of the consolidated loans.

Needless to say, the institution originating the loan ends up disconnected from the risk once it is sold to other financial institutions. At least they're insured, right?

View on HN

"The original lender is only required to retain around 10% of the consolidated loans."

That part may be the only redeemable part of this process.

Because, hey...let's package these into collateral debt obligations, and then hey...let's engineer these CDOs with tranches of various credit quality as rated by rating organizations not understanding what they are rating and afraid of asking for more information because their profit is based on volume of these ratings and not quality, and hey...let's create a second CDO based on the lowest rated tranche of the first CDO, and boom, that low credit score is magically now a AAA, and why not create tranches of this second CDO too. Then, let's create a market out of this with a way to short the CDOs, and let's call this new instrument something ridiculous as Credit Default Swap. No one is ever going to need those, because everybody in the financial industry believes the underlying assumptions that 1) the real-estate prices will forever go up, 2) and there is no chance that more than 4% of sub-prime mortgages are going to default. Seeing that this makes a lot of sense, let's make a trillion dollars worth of these and we don't have to put them on the balance sheet because they are AAA rated, safe as the US Treasury bonds, and so why not get leveraged 40 to 1 on these. Oh wait, that's already been tried.

Yep, it is the same thing over again with a different underlying loan source. And why shouldn't it be. The people responsible for the 2008 crisis are mostly still in place. In fact, many of the people who survived were promoted to more prominent positions. In fact, some say the loans of 2008 are said to be "recovering" and it's "not so bad". This translates from the expectation that these loans would lose 90% of their value and they ended up only losing 75%.

Exactly. And 10 years after the crash - the fact that the first paragraph of this article is factually wrong (thereby illustrating a lot of what is still wrong with macro-economic analysis) - goes unremarked.

Banks do not lend their deposits. (That's what a fund does.) Banks create deposit money when they make a loan, and statistically multiplex asset cash against deposits to manage transfers of money within the banking system. Repayment of loans then removes the created deposit money from the system.

I might have an inkling of what this comment is saying, but I'm not sure. As written it almost sounds like the description of a ponzi scheme or a more pedantic means of saying "they lend out the money given to them by depositors" ... anyone around who could clarify this a bit?

It's called fractional reserve lending... and it works until it doesn't. As long as your "divedsified" losses are less than (~10% of loans) your reserve collateral (actual Treasuries and Rentable real estate) which generate cash or can be used to pay taxes... then you're solvent. If not then assume the Fed will bail you out... they did last time.

Meanwhile, you can lever your "low risk" loan income 10x so that 1-3% marginal yield looks like 10-30% profit. Hard to give up that crack pipe!

Sorry to be the bearer of bad news.. The fault tolerance of banks using FRB, is approximately 1% of loan capital per year. Any more than that, and they are driven into regulatory incompliance - which then has monetary significance.

Reserves don't actually play a roll in loss management, bad loans have to be written off against loss provisions and/or profits - liability/equity accounts. Basel 3 has put mandatory limits on how much loss provisions/capital has to be held, but it hasn't solved the fundamental problem with the interaction with the money supply.

Yes it's a multiplier scheme and if you squint just right, it's pretty much a ponzi too. But it isn't a pedantic way of restating "lend out".

Discussion of the issue is often fraught with accusations of conspiracy theory. Austrian economists (widely regarded as kooky by the establishment) have a bunch of books related to these issues perhaps the best of which is Rothbard.

_0 https://mises.org/library/mystery-banking _1 http://www.bundesbank.de/download/bildung/geld_sec2/geld2_ge... _2 http://www.bankofengland.co.uk/publications/Documents/quarte... _3 https://www.youtube.com/watch?v=CvRAqR2pAgw

Austrian economists make a few valid criticisms of mainstream economic theory. The thing is, that's very easy to do. What's not so easy is to offer a superior economic model. Their ideas, last I checked, don't involve any math; that is fine, but it makes Austrian theory unfalsifiable, and therefore it can't really be shown to be a valid alternative.

Remember, fractional reserve banking started when the Rothschildren decided it was OK to tell their depositors their deposits were in the safe, when the bank had actually given it to other people.

Our banking system is founded on a lie.

I suggest a valid alternative is the truth. If someone gives you their money and it's your job to store it safely you should to that. For 100% of that money. Not 10% of it.

There's a banking product for that, it's called a safe deposit box.

No. Those are for jewels and other goods that can't be represented with a number typed into a computer.

That is both historically untrue and stinks of anti-Semitic slander passed as fact - following the age old tradition of blaming the Jews for mismanagement by nobles and royals because that wouldn't get you executed. The Rothschildren were certainly /not/ the first being established in the 18th century. Given that the more reputable central Swedish bank used it in the 15th century - let alone the shadier practices of other banknote institutions that predated them.

As for the 100% retention idea it is like complaining that a car isn't a faster horse. There may be some fringe cases when the horse is better but most of the time the primary use was going fast across paved roads.

To be a good steward involves investment - just any asset sitting stored is a waste of value and will not even preserve it against inflation. It is the red queen's race.

OK i restate. Replace, in my original post, "Rothschildren" with "goldsmithers." I didn't even know they were Jews and that whole bit is ad hominem at best. But my point remains.

My money is not the bank's asset. Neither is the stuff I put in a storage unit at public storage.

Imagine if they were loaning out your stuff while you were away.

If I choose to invest my money that's my right. It's not your right to choose how I invest my money.

Going faster isn't a reason why it should be OK to steal people's money when you're saying you're keeping it safe.

Lies like this brought down the entire world economy. Wars were started. Lives were lost. But you suggest it's better except when it's not?

I don't agree.

I'm not an economist, so I could be wrong.

Austrians say they use praxeology. They approach economic problems through deduction. This should make some of their theories falsifiable. After all, if you can't find evidence for the notion of opportunity cost, then it can't be a valid concept.

I think that the problem with Austrian economics is that relying on praxeology and deduction is like doing theoretical physics - you make assumptions, do logic based on those assumptions and see if it fits the real world. It can work, but it's slow.

> don't involve any math [and so is] ... unfalsifiable

A theory involving math is only better than the alternative if it is also correct. Theories can be falsifiable if they don't have math either.

For example, I might have a theory that water always flows downhill. Then I am confronted with a pump, and my theory is now falsified. Fitting a mathematical model involving gravitational constants and laminar flows will be a better theory because it fits the observations very well.

A theory that I have porridge for breakfast in a way that can be modeled with an iid normal distribution is a terrible theory, because if I find out about it I'll start playing with my breakfast just to spite the modeler. The inclusion of a precise mathematical model just paints a bigger target for me to hit. Economics is much more in that vein of things - the people being modeled can respond to the models themselves.

The mainstream economics theories are all going to be based on non-mathematical assumptions that link reality to a math-y model. Those assumptions are the weak point of economics. We can trust the academics to get the math right once they've finished making assumptions.

Your interpretation of my statement isn't accurate. Austrian theory's lack of mathematical formalization is what makes it unfalsifiable. It's not your water example; if it were, then you would be correct.

Again, it is easy and often valid to criticize the methodologies associated with mainstream econ, but the question is whether one can offer a demonstrably superior alternative. You have not done that here.

> but the question is whether one can offer a demonstrably superior alternative

That isn't the only question. There is also a question "is there something here that we can model?". There is no requirement to provide a better model if the existing model is inaccurate enough.

I've never actually seen an economic model that uses maths beyond accounting balances or basic calculus/timeseries so I'm certainly not qualified to critique them - whatever they are.

But the Austrain critiques seem to be something like that if you blow a credit bubble then manipulating accounting identities isn't enough to avoid having to pay back the credit at some point and taking an economic fall where the credit gave you a boost. That is a falsifiable position. The Americans and others are running a big experiment that has so far provided potentially falsifying evidence. They can still be proven wrong even though they havn't published an equation, so their theories are obviously falsifiable.

>There is no requirement to provide a better model if the existing model is inaccurate enough.

We disagree here. I think this notion is what allows people to arbitrarily dismiss theories that they personally don't like.

>a more pedantic means of saying "they lend out the money given to them by depositors"

As the other commenter mentioned: they lend out more money than the depositors gave them. It works roughly along these lines: depositors give the bank $10 million. The bank can now make loans worth $100 million (or some such number depending on the type of debt). They're creating money "out of thin air".

Neffy probably knows more about how it actually works, but what I described is the basic principle of fractional reserve banking.

It’s important to understand that they don’t typically lend out 10x. My business banker (WF) said they have trouble lending out 0.5x. There’s just not enough high quality lending opportunities.

There's no limit to the multiple.

The 'fractional reserve banking' model is just wrong. (That there are no real reserve requirements in places like Canada or the UK, should really tell people this).

In reality it works like this. Loans create deposits. So you lend $10 million and end up with $10 million of loans and $10 million of deposits. The regulator then says you have insufficient capital, so you issue equity or bonds to the tune of $1 million and convince 10% of your depositors to swap their deposits (that you created) for that equity/bond by setting an appropriate interest rate on it.

Rinse and repeat until you run out of people to lend to at a price they are prepared to pay.

Neither deposits, nor equity have a quantity control function. It's all about the price, not the quantity.

In other words the amount of money in the system floats at the current price of money.

> There's no limit to the multiple.

That is factually wrong. You have significant capital requirement both in term of RWA or Leverage Exposure (the latter is more likely to be binding for mortgages), particularly in the UK which along with Switzerland gold pleated every international (BIS, EU) regulations.

Banks cannot extend their balance sheet indefinitely, and if you look at UK banks, they significantly deleveraged since the financial crisis, as they adapted to new regulations.

In fact these capital requirements, along with increased liquidity requirement are probably why the multiplier effect considerably reduced after the crisis.

It's factually correct. Deposits are created and then converted into the required capital at a price. To buy equity you need to use a deposit, and that comes from a prior loan. That's just how it works.

Same with liquidity which is just increased amounts of loans to the central bank (government) in the form of government bonds. Collateral uplift can give you that.

All these just add to the cost of the bank, which changes the price and that reduces the number of people taking the loans. But the fact remains that the quantity floats at the current price of money. Liquidity and capital requirements just change the current price of money.

> It works roughly along these lines: depositors give the bank $10 million. The bank can now make loans worth $100 million (or some such number depending on the type of debt). They're creating money "out of thin air".

I think this is an oversimplification to the point where it is misleading.

Banks can't literally create money like this. What happens is that if a bank has $10m in deposits, they have $10m to loan out, but not more. In practice, they are limited by how much they can loan out by the fractional reserve. If that limit is 10%, then this bank can loan out $9m, so there is now $19m in existence.

If you follow this out to the limit, you get $100m with $10m of initial deposits and a fractional reserve of 10%, but a bank can't just multiply its money by 100%/10% and loan out that much money.

Sadly it does: in practice there are 5 big banks in an economy. Maybe less.

So let's say a bank has $10m deposits. You're right: they will lend out $10 million, most of that in mortgages, a bit in business loans, a tiny bit in personal loans. Mortgages come with the condition that you don't get the money, you get to write a checque (or do a bank transfer), verified by an official like a notary public. Now you might notice ... so wait ... they lend you money that you can only deposit into another bank account (and then very likely stays in that bank account) ?

So that means they can pull this: $10 million in deposits. They lend out $10 million, $9.5 ends up into accounts in that same bank and become deposits. So now they have $9.5 million in new deposits and $10 million in assets (money people owe them). Plus an additional $1 million in interests. Because there's only 5 banks where money can be kept, they're pretty much guaranteed, once above a certain size, to keep or exchange 1:1 deposits created from loans in their bank.

So now they lend out the new deposits ... and of course also the assets. So now they lend out a further $21.5 million. Next step is a further ~$40 million. In most of the world laws specify that they can only multiply ~50 times this way (Australia being a very notable exception).

So with $10 million in source deposits a bank would lend out a little less than half a billion $.

Obviously in practice it's 100 times more complex than described above. For example, banks will usually sell a good part of those assets, like loans, and then take that as a profit, and pay it out to shareholders. This changes their risk profile in positive ways. There's business loans which mostly (the larger amounts at least) work like mortgages but not entirely. There's international transactions, ...

The thing to remember is: large amounts of money simply don't "really exist". Starting at maybe $100 million it is simply not possible to "have" that amount of money. Something else is happening, some financial construction is involved.

This is actually one of the most succinct (ever so slightly hyperbolic) write-ups of the 2008 financial crisis I've seen.

Check out "Diary of a Very Bad Year" for a longer but equally, ah, education-packed treatment.

That’s because OP ripped it off word for word from Wolf of Wall Street

Reads like that, but, if you've seen it (or haven't) how else would you describe the concept?

Not sure what you mean.

What OP copypastaed is word-for-word from a monologue in the movie.

That is impossible. I've never seen the movie.

I've never read Wolf of Wall Street.

Pedantic I know but I wanted to check if it was in Wolf of Wall Street and having searched the script it would appear it wasn't...

Maybe another movie?

"That part may be the only redeemable part of this process."

Until they start hedging their exposure in some more or less creative way.

Well the 10% is I assume an equity tranche. Ie any loss on any loan in the portfolio goes to that tranche first up until the total losses exceed 10%. So the bank still has significant skin in the game, unless their short term profit exceeds 10%, which seems kind of high to me (but I don’t know the terms of the transaction you refer to).

The whole issue is so grey, that it'll be hard to find a robust solution to it, without some sort of wholesale change.

There is a lot of overall sense in having retailers focus on retailing, and leave the high finance stuff to underwriters, like insurance.

The problems is that the laissez faire market dynamics degenerate at this point, Retailers get too good at retailing, selling people bad deals that could land them in trouble later. Underwriters focus on securitization, which is ultimately a sister to the money-creation process (like "fractional reserve" banking).

Pretty much any solution is clunky. You can try to force "originators" to hold risk, but this limits scale. You can regulate underwriting rules (affordability & doc checking standards, for instance). You can regulate underwriting, but the high finance sector is (a) good at influencing regulators and (b) good at finding ways around regulations.

|regulating for risk is a tricky business. You end up trying to regulate away risk, which is impossible.

Asset backed securities aren't really new: funnily enough, the same models that are used to price and work with mortgage backed securities can be rapidly hacked to work for credit card debt, mobile homes and well, pretty much anything else out there.

Someone's buying it, so the banks are making the sausage again. Such is life.

Isn't it the job of Credit Rating Agencies to give poor ratings to these packages if the loans in them are bad?

What makes you think they don’t? Rating agencies have been burned by the last crisis and are poised not to make the same mistakes.

Is there a /s I'm missing? What evidence is there that they are doing anything differently, and what incentive would they have to do so?

What evidence do you have? I work with the teams who structure these transactions and get them rated on the bank side. Their methodologies have been revised, every communication with the originating bank is recorded and actively monitored (they have a compliance officer on every call), they do not allow feedback or revising their initial rating to optimise the structure. They have become paranoid. What makes you think they are not?

And their incentive is the same as an audit company, it is an existential threat.

I don't think you understand how trust works. Rating agencies have failed so many times, with bad consequences mainly for common folks instead of higher management (I mean long prison terms for whole C-suite/senior mgmt + all personal assets lost) that they need to bring some seriously good performance, over long time (say at least 2 economy cycles) to gain back at least some portion of trust.

Sharp suites and clever talk only won't impress anyone anymore.

> What evidence do you have?

The business model is the same; big failures are still common (e.g. the recent case of Oi SA); the same Big Three still control the market; "no meaningful reform of the credit-rating agencies has been undertaken" (Paul Volcker).

> And their incentive is the same as an audit company, it is an existential threat.

What existential threat? Which relevant (and this mostly means the Big Three, which "issued 97%–98% of all credit ratings in the United States and roughly 95% worldwide") rating agency has went bankrupt or even faced a major threat to its dominance?

What's the incentive to buy such a toxic bundle?

High returns at low risk (as long as you believe the rating agencies).

The risk is always pretty low if you can leverage taxes for the worst case.

Why would the agency rate it as low-risk though? Because they are assuming that individual defaults are uncorrelated? The variance of the distribution of returns of a portfolio like that must be huge, not to mention the risk of an economic downturn putting the entire portfolio into default.

I'm just an armchair analyst, but I'd love to know what the people buying these loan packages know. Because it doesn't seem low risk to me.

>Needless to say, the institution originating the loan ends up disconnected from the risk once it is sold to other financial institutions.

In many cases, isn't this already the case with the largest consumer loans (mortgages)? If the buyer understands the risks of the consolidated loans, what's the beef? If the bank is hiding the risks, that's another issue - I agree.

The house is the collateral in a mortgage though.

The very-very upside-down house...

Which you have the privilege of holding onto for years after a 12-24 month foreclosure process. “Collateral”.

>If the bank is hiding the risks, that's another issue - I agree.

I think the point is that this creates a lot of incentive for the bank to hide the risks that they wouldn't otherwise have. Perhaps more importantly, it doesn't incentive them to prioritize things that might help them measure the risk better, so those sorts of improvements never get made and it just deteriorates over time.

The loan originator has a different risk calculation when their ultimate intent is to offload 90% of the loan(s). Some might say it would be in their own financial interest to minimize the amount of due diligence involved in determining whether to give such loans. Plausible deniability.

From a buyer's perspective, they receive the due diligence that the originator provided. There are various risk models put in place with that information, but they are also "insured".

These are the same fundamentals that led to the 2008 crisis. Giving loans to people who weren't actually qualified. Miscalculating the risk, intentionally or not. Buyers ok with the risk, assuming the insurance would come through in the worst case.

If there is an event, or a series of events, that cause these loan takes to become insolvent, say losing jobs with no hope of getting a new one (truck drivers being replaced with self-driving trucks?), that would cause large amount of loan defaults...

Except the 90% is not a vertical slice, it is horizontal. I.e, the first 10% of notional in losses on the whole portfolio goes back to the originator. The profit on the transaction would have to be huge for this not to realign the incentives.

How long until these unsecured loans go bust and the big banks beg for a bailout again?

When the next recession comes.

no they are not inf act disconnected as 2008 disaster showed with those saqme banks buying insurance on such new finance dreviatives

Reply on news.ycombinator.com