The Banker Who Said No | Hacker News Reader

The Banker Who Said No

forbes.com

135 pointsdangoldin63 comments

He's also behind the Beal Conjecture which offers a $100,000 prize for either the proof or counterexample:

http://www.math.unt.edu/~mauldin/beal.html

BEAL'S CONJECTURE: If A^x +B^y = C^z , where A, B, C, x, y and z are positive integers and x, y and z are all greater than 2, then A, B and C must have a common prime factor.

FTA:

> "This is the opportunity of my lifetime," says Beal. "We are going to be a $30 billion bank without any help from the government." (A slight overstatement: He is quick to say he relies on federal deposit insurance.)

And he pays for it, too (premiums, y'know). That's not "help" as we usually mean the word; that's a purchased service. Now, as for the bailouts -- which Beal is not getting -- that is "help".

It is a good read and he obviously is a "good" banker for not overextending his bank. But to say he doesn't rely on government help makes it sound like he runs a normal corporation. Banks are not normal corporations. Mine can't loan out money at 26 times my deposits. Can yours? Of course not, your not a bank. This is the rub with banks. They want to behave like private entities, but they leverage government backed currency which normal corps cannot do.

I think we need a new class of corporation for banks and possibly insurance companies. A class that recognized they are hybrid entities, and not the same as a typical private company. It turns out we sort of do have this class (there are lots of hoops to jump through to become a bank). But this class is built on top of the normal rights and privileges of the standard corporate entity.

> Mine can't loan out money at 26 times my deposits. Can yours? Of course not, your not a bank.

Does your corporation have to abide by all of the banking rules? Of course not, you're a regular corporation. If you want to be a bank, start one. It's a pain in the ass but can be lucrative.

Even very large companies, such as PayPal, find that its far easier to buy an existing licensed bank than to start a new one.

Lot of upmods for an extremely confused post.

Mine can't loan out money at 26 times my deposits. Can yours? Of course not, your not a bank.

...oh kay.

A class that recognized they are hybrid entities, and not the same as a typical private company

Umm. You seem to be getting at GSEs, a standard corporate form.

http://en.wikipedia.org/wiki/Government-owned_corporation

Generally, this doesn't seem to work out too well with financial companies, see Fannie Mae/Freddie Mac (which weren't GSEs, but close enough...) The truth is, the vast bulk of banks were perfectly healthy and responsible. Almost all the trouble came from a few gigantic ones, and the reasons for their behavior were many.

Is it really responsible for a bank to back 30-year loans with demand deposits, at 25-to-1 leverage? I really can't imagine that happening without government intervention, and if it only happens when the government makes it happen, there's a good reason to expect that it's irrational.

Banks are hybrid companies in the sense that FDIC insurance means they are basically conduits for lending to the government. If I deposit $100 with my local bank, I don't care whether the banker keeps it in cash, invests it in penny stocks, or spends it on whores -- no matter what, the government backs my deposit.

This is for other promises from companies. E.g. if I pay someone $100 now for $100 worth of lawn care over the next five years, I make damn sure I can trust that person. If I pay the bank $100 for a CD maturing in five years, I don't.

This gives all banks an incentive to take undue risk, balanced by government regulation, which gives them a secondary incentive to find clever ways to take big risks (which they do by hiring clever people who would otherwise do more socially useful stuff).

>E.g. if I pay someone $100 now for $100 worth of lawn care over the next five years, I make damn sure I can trust that person. If I pay the bank $100 for a CD maturing in five years, I don't.

> This gives all banks an incentive to take undue risk

Actually, you've just demonstrated that FDIC insurance gives you an incentive to take undue risk. You'll deposit your money at any bank without regard for whether said bank is trustworthy.

BTW - It's curious that you seem to believe that better rules will help since you also seem to believe that "clever people" can always get around the rules.

As to whether a govt-run bank would do any better, feel free to point to the fraction of govt enterprises which are run as well as you'd require of govt run banks. Which states' DMV is "good enough"? How about the post office?

The trouble is that the trust a bank or any such institution receives is not, and probably cannot be, well correlated to its trustworthiness. Cf. Bernard Madoff, and the vulnerability of any bank (especially when no insurance exists) to a run.

The threat of a run is only available if you lend out demand deposits, which is something that could be eliminated.

I have no clue what you are talking about. I'm an anarcho-capitalist.

I was talking about how the FDIC creates bad incentives. I don't think I said anything about how new regulations would help, since the bit you quote is about how the difference between good regulations and bad regulations is in what kind of talent is misallocated.

Edit: Perhaps the part about being 'balanced by regulations' threw you off. What I meant is that the government writes rules to keep people from doing what the FDIC gives them an incentive to do. Those rules, of course, do not work.

> I was talking about how the FDIC creates bad incentives.

And I'm pointing out that the bad incentives affect depositor behavior. Thanks to FDIC, you have little incentive to find a trustworthy bank. Instead, you judge entirely on other criteria.

Bank of North Dakota. A government run one..

There haven't been enough defaults in North Dakota for that to be a serious issue. Some places are able to have lots of state involvement without getting into lots of trouble, e.g. the Scandinavian socialist states, Singapore, etc. Most of the defaults were concentrated in the 'sand states' -- California, Florida, Arizona, and Nevada.

> Bank of North Dakota. A government run one..

Majority-Scandanavian states seem capable of lots of things that other states can't manage. Since the majority of US states are not majority-Scandanavian, we can't use the majority-Scandanavian ones as a model.

Well, ND is a German majority state. I understand that it is tempting to say that it has something to deal with scandinavian personality, but look up a bit up north you'll see Canada, with very non-scandinavian in it's population makeup, however I'd think it is a lot like Northern Europe in its political and economical structure

Not only is it not responsible to maintain a 4% reserve ratio, in the US it's illegal, and has been for decades. (The legal requirement is 10%.) That also means that it pretty much doesn't happen; banks are pretty tightly regulated. Until the repeal of Glass-Steagall, if I understand correctly, it was also illegal for your banker to invest your money in stocks, penny or otherwise.

The extent to which banks are conduits for lending to the government is independent of FDIC insurance. FDIC insurance is pretty much normal insurance, in that it is funded by premiums charged to policyholders, rather than, say, tax money. Banks are conduits for lending to the government in the sense that they maintain much of their reserves in US Treasury bonds.

Your comment gives the appearance of someone commenting vehemently on a topic about which they know very little. I mean, it's hard for me to remember that there are people in the world who know even less about banking than I do, but it sounds like you're a member of that elite group.

The truth is, the vast bulk of banks were perfectly healthy and responsible. Almost all the trouble came from a few gigantic ones, and the reasons for their behavior were many.

You're way too optimistic, IMO. 21 US banks have failed so far this year, and many more are going to fail. Many commercial real estate and development loans are going to default in the next couple of years, mostly provided by many small and medium local and regional banks.

sorry, my language may be confusing as I certainly didn't want to write extensively on how money expansion works (I assume most here are read up), but I don't think I'm too confused in my head...maybe not, never know ;) For the record, I have written several large banking systems. I actually do know how transactions and liabilities get handled. And normal corps don't get to behave like this.

In my suggestion to recognize banks as something different than a normal private corp, I'm not thinking of GSEs as you link to. I'm thinking more about transparency and liability issues being different for banks than for a normal private corp.

If the gigantic banks that are currently the focus of such various problems would have had to have more open financial liabilities, the risk _should_ have spread less.

> But to say he doesn't rely on government help makes it sound like he runs a normal corporation. Banks are not normal corporations. Mine can't loan out money at 26 times my deposits. Can yours?

You might have a point there. I guess the question (regarding whether my point was valid) is whether federal deposit insurance is underpriced. I suppose there is a good chance that it is.

It's the kind of thing that can't be priced (as AIG found out.) The distribution of the default rate is unknown and unknowable (fat tails and black swans and so on.)

Also, if you increased the price of FDIC insurance, banks would need to make greater interest rates, and therefore would need to make riskier loans. You could make FDIC insurance contingent on taking less risk, but that sort of regulation always seems to backfire. (Of course we already have such regulation, the article even alludes to it. But increasing the strictness of risk-taking regulation would just solidify the major players.)

You ask:

> Mine can't loan out money at 26 times my deposits. Can yours? Of course not; you're not a bank.

Actually, I think that if someone loans a hypothetical non-bank US corporation money, it can legally loan out 100% of that money, without retaining any of it as a reserve. This is probably not a good idea, since it means it won't be able to pay any of its bills next week, but it's not illegal.

What you wrote makes it sound as if, when depositors loan a bank $100 000, the bank can then loan out $2 600 000. That is not the way fractional-reserve banking works. The actual amount the bank can legally loan out in that case is more like $90 000. A non-bank corporation would be able to loan out a larger amount (up to $100 000) in that case, not a smaller amount, as your post suggests.

sorry, I was vastly oversimplifying the mechanism. I was speaking off-hand assuming most on this forum are well read on how these mechanisms work, it certainly sounds like you are.

To be more precise: The effect of banks, in aggregate, on the money supply results in a multiplier, which today runs at, in average, 26 times deposits.

Here are a few overviews: http://en.wikipedia.org/wiki/Fractional-reserve_banking http://en.wikipedia.org/wiki/Reserve_requirements http://economics.about.com/cs/money/a/reserve_ratio.htm

To prepare bids he locked himself in his office to write a computer program with 50 variables (now 250), ranging from home price changes by neighborhood to interest rates to origination dates.

Yep, he's a hacker.

This sounds like a multi-variable linear regression he did in Microsoft Excel?

That's very possible. Reporters tend to exaggerate certain elements, but he has a pretty strong math background. Either way you look at it, he's hacking the numbers to increase his own odds of investing in the right stuff.

Actually, from the description of his high stakes poker games, he ran computer simulations in order to gain a competitive advantage so I wouldn't be surprised if he actually wrote a program.

Beal, for his part, took a mathematical approach, at one point running millions of computer simulations of various poker problems, in search of an edge against the pros, who rely on an uncanny intuition honed by thousands of hands

http://www.amazon.com/Professor-Banker-Suicide-King-Richest/...

Programming is more than software engineering. Some things don't need to be engineered.

Other than excel having some floating-point arithmetic problems, what difference does it make what tool he uses?

Most things are linear to a first approximation anyway.

Only because most first approximations are linear.

I can't prove it, but I will assert without much fear of a counterexample, that, for most if not all systems, linear models describe most of the variance with the fewest number of free parameters. This may be as much as 90% in some cases, but almost certainly more than 50%.

In other words you get the most bang for your buck. This is not to say that a linear approximation might not be very inaccurate on important parts of the problem domain, but the variance that a higher order model would describe would be less, perhaps much less, than the portion the linear part describes.

And even the most fancy statistical methods are generally about proving that some mild generalization of linear regression is sufficient to solve a particular problem. Non-linear models often result in a great increases in model complexity which leads to issues of overfitting, computational intractability, and instability.

First you say, "linear models describe most of the variance with the fewest number of free parameters," and then later you say, "In other words ... the variance that a higher order model would describe would be less, perhaps much less, than the portion the linear part describes."

In my limited statistical experience, the problem with higher-order models is not that you get less bang, but that you need more buck: that is, there are too many free parameters. But in your second quote, you seem to be talking about higher-order models that have as few parameters as a linear model, i.e. one parameter, plus one per independent variable. What kind of higher-order models are you thinking of?

Sorry, I mistyped.

I meant that the additional amount of variance described by a more complex model beyond that described by a linear model is much less than that described by the linear model in the first place. Obviously the total amount will be more, or else your model is both complex and wrong. :-)

Consider:

  Model A - 1 degree of freedom - 60% of variance
  Model B - 2 degrees of freedom - 75% of variance

That extra DOF has gotten you 15% better description of the variance, but at the cost of complexity. Perhaps that is worth it, perhaps not. As has been noted above, that complexity has a real cost that can manifest itself as overfitting, instability, and lack of generalization. The curse of dimensionality is very real.

All that I meant was the linear model will probably capture the most variance per unit complexity. Which gets back to my original point that most (all?) problems are linear to a first approximation. It's not just that people are lazy.

One has to consider that he is simply investing after almost everyone else screwed up. This could be a great opportunity but results are not yet in. It important to consider this.

I think it's pretty telling that this guy was investigated for not consuming the fraudulent crap coming out of Wall Street.

Wouldn't that be like a 1999 startup being investigated because they didn't have exponential growth projections in the IPO prospectus?

Well, if someone does something completely out of the ordinary for [what the investigator thinks are] no apparent reason, then that should raise red flags to investigate.

I'm not surprised he was investigated; Massively different investments, raising money without buying anything apparent, and not trying to cash in whenever possible do point to an anomaly -- possibly even a fraud -- happening.

So long as he's not wrongfully convicted, I don't have a problem with investigations.

Well, if someone does something completely out of the ordinary for [what the investigator thinks are] no apparent reason, then that should raise red flags to investigate.

Yeah, that's the point...that it was extraordinary to behave sanely (as opposed to "consuming the fraudulent crap coming out of Wall Street").

> Beal shrank his bank's assets because he thought the loans were going to blow up. He cut his staff in half and killed time playing backgammon or racing cars.

I have always respected a man who drives his cars hard and plays a mean game of backgammon -- and actually makes money while doing it. Upvoted.

This is a great article.

If anyone is interested in learning more about Andy Beal's million-dollar poker games with the best players in the world in 2001, you should check out this book: "The Professor, the Banker, and the Suicide King: Inside the Richest Poker Game of All Time", by Michael Craig. It's got a pretty complete profile of Beal, and tells all about the games.

Does anyone know if it's possible for an individual to buy a CDO?

Supposedly everyone wants to get rid of these things and is desperate for cash, which makes it seem like a decent opportunity for buyers. If I had $10,000 I could afford to lose (and was willing to take a long shot with) is there some market where I can buy those sort of assets?

To start with, you have to be what is commonly called an "accredited investor." This is a requirement because almost none of the "investments" sold as CDOs are registered with the SEC. The accredited investor loop-hole is the one you'll need to invest in hedge funds and the like: it means something like you're a big boy now, and you don't need the feds protecting you from crooks. Alternatively, your organization will need to be qualified as an institutional investor.

http://en.wikipedia.org/wiki/Accredited_investor

http://en.wikipedia.org/wiki/Institutional_investor

Ask your broker if they've got any they want to unload.

I'm a bit short of the wealth requirements to be called an accredited investor ... (1M net worth + 200K/year income).

That just seems blatantly unfair to me. It's the government setting up a different set of rules for the rich.

> I'm a bit short of the wealth requirements to be called an accredited investor ... (1M net worth + 200K/year income).

>That just seems blatantly unfair to me. It's the government setting up a different set of rules for the rich.

Those rules are set up to protect you from unsafe investments, scams, and the like. After all, you need to have enough money to keep paying taxes....

> That just seems blatantly unfair to me. It's the government setting up a different set of rules for the rich.

There was one VC here that mentioned how someone emailed him about also investing in 3 of the companies that he had invested in, and how he couldn't let the person do it because the person didn't fit the requirements to be an "accredited investor" then he mentioned what three companies they were. They were all money pits with no path to profitability. I believe the "accredited investor" bit helps people from having there life's savings thrown away on a unregulated bad investment. The ultra rich either inherited money or built a real business from the ground up and sold it. I would suggest one of those paths instead :)

But doesn't the current crash highlight the superfluousness of the 'accredited investor' limits? Even investments poor people are allowed to make -- like common stock of giant financial institutions, even those with government sponsorship (Fannie/Freddie) -- can essentially go to zero.

The general markets are down 45% from their peaks. Why should submillionaires be denied the chance to put 55% of their portfolio in T-Bills, and 45% in highly-risky unregistered private securities? That wouldn't have done any worse than the public stock market... and might do a lot better, if you understand the private companies involved.

And if submillionaires are such easy marks, why not any limits on how much they can gamble in casinos or even state lotteries?

The 'accredited investor' limits are silly; a phony security blanket at best, an unfair impediment to broad-based entrepreneurship and investing at worst.

Just a slight quibble. It's 1M net worth or 200k/year. You make it sound like it's "and."

I'd stay away from CDOs, but there are great opportunities in cash subprime and Alt-A mortgage bonds. If you know how to analyze the deal structures, it can be very lucrative. If you run the bonds to extremely draconian default and loss scenarios, you can still end up with 15-20% yields.

$10,000 is too little -- for agency MBS, $15,000 is the minimum purchase size, and I think the rules are similar for non-agency bonds. Most of the high-end brokerage firms can help you, though. I know Merrill Lynch and Bear Stearns Private Client Services (now a division of JPM) deal in mortgage bonds for individual investors.

I really doubt you could buy a CDO for a fair price. The banking crisis basically happened because banks had their assets overvalued and over leveraged their securities.

And I believe that through some accounting tricks, most bulge bracket banks are still overvaluing their securities. And in some cases once the securities are correctly valued, the banks will be insolvent. That is one problem the bailout money is for, to enable banks to correctly value securities.

So I believe that if a bank actually sells any CDO's at a fair price, there will be major problems with keeping the rest of the securities overvalued.

And from the way I understand it, most of these over valued securities are based on multiple assets, so one CDO is going to track other CDO's, you can't pick a CDO that is solely based on correctly valued assets.

Compare it to a local bank that is insolvent.And this bank has overvalued their assets. So I am looking to buy a house, and see that the bank has called the loans on two houses that would be in my price range. Except one of the houses has been used as a meth house, and the owner would have to pay an extra 60-70k to make the house habitable. Well, it would be an easy choice for me, I would want to buy the other house for less than the value of the loan. Only in this comparison that doesn't work. See, the bank found out that if they grouped the mortgages together, they could sell overpriced pieces of the mortages. So I would not be able to buy anything from the bank that would give me legal ownership of property. And the mortgages are grouped together, so for every dollar I invested in one property I would be investing a dollar in that meth house.

Nobody really knows what the CDO's will be worth in 10 or 15 years, but I do know that the banks don't have a strong desire to price CDO's low enough to actually sell them.

"To prepare bids he locked himself in his office to write a computer program with 50 variables (now 250), ranging from home price changes by neighborhood to interest rates to origination dates." - Forbes.com 2009

"Beware of geeks bearing formulas" - Warren Buffet 2008

On the other hand, Warren Buffett is equally contrarian, and is doing a lot of similar things with his money, though on a larger scale (writing massive loans to Goldman, Harley, GE etc for 10-15% interest rates). He also has the same complaint about the government. His AAA rated business pays much higher rates to borrow money than Citi, BoA or any of the now gov't sponsored entities.

In any case, everyone loves a good contrarian.

Didn't Berkshire Hathaway loose a lot of money because of the current mess?

They went down as the market went down and did have some losses, but they also had an enormous pile of cash and not very much debt. Buffett has been waiting for the market to go down like it has so he could make his move with that cash. As a result, I think he's going to make out extraordinarily well in the coming years.

Sorry for the Buffett fanboyism, but he's the man.

Yes. He was one of the early warning voices (2003?) about the dangers of complex derivatives.

As I understand it he made some serious missteps in the recession. WP:

> Berkshire Hathaway acquired 10% perpetual preferred stock of Goldman Sachs at $123[40] only for it to fall to below $60. Furthermore some of Buffett's Index put options (European exercise at expiry only) that he wrote (sold) are currently running around $6.73 billion mark-to-market losses.

Warren Buffett tarnished as Main Street oracle http://www.thestar.com/Business/article/604619

Buffett suffers big losses at Berkshire Hathaway http://www.bloggingstocks.com/2009/02/28/buffet-suffers-big-...

(d Berkshire's net worth dropped a whopping $10.9 billion in the final three months of 2008.

Berkshire's shares have fallen 44% since the end of February 2008. )

etc...

Actually, the preferred stock in Goldman is an option. He has the option to buy the stock at $123 indefinitely, meaning if it ever goes above that amount he can then choose to buy it and sell it for a profit.

And like I said, the stock went down with the market. Yes, at one point his stock was down 44%, but so was everything else. And if you look at the fundamentals, they have a small leverage ratio, close to zero actually if I remember correctly, lots of cash, and are one of a handful of companies still rated AAA.

The guy in the article keeps saying if only he had access to more cash he would be making piles more money. Well Buffett has the cash, and he's going to rake it in when the market rebounds.

You understand incorrectly, for the most part. A sinking tide lowers all boats. Buffett doesn't care what the stock market prices his company at in the short term, only how the intrinsic value of his company improves over decades.

Also, the financial press loves to take pot shots, even though Buffett repeatedly and thoroughly explains that Berkshire's stock price will go through severe beatings from time to time.

Beal Bank? I am more interested in the story of how he started the bank. How can an individual with the government's blessing can start a bank like that. Don't they need huge funds and deposits and a million financial and security hurdles? Where did he get it all?

Furhtermore, I wouldn't trust my money to a bank where the owner spends millions playing poker and brags about it. WTF?

Poker inculcates many of the skills needed for finance, and if you're good it can simply be a profitable investment.

Exactly! When interviewing for hedge funds for a quantitative analyst / trader position one should expect to be asked about poker / blackjack / backgammon.

A classical interview question / challenge would be to invent a new card game and ask the interviewees to devise an optimal strategy for that game.

Buy low, sell high. This is not rocket science.

However, he does know that too.

Reply on news.ycombinator.com