MaxedOutMama
Monday, March 09, 2009
Poker And The Need For Glass-Steagall
Update: A community banker sent me an alternate plan for a federal mortgage bailout program. I am thinking it over today, so I probably will not post my reply to Carl's demurrals until tomorrow.
See the "Reality Bites" issue of John Mauldin's newsletter, which includes a discussion of risk:
In order to try to compress this a little, I am going to refer back to several prior posts.
The first is How Did We Get Here, written in September of 08. This covers some of the basics about the difference between banks and SEC regulated financial institutions, the role of REITs, and the repeal of Glass-Steagall (which essentially occurred piecemeal through the 90s, but was finalized in 1999). Please also see the Wikipedia entry, which is very informative.
(Btw, for those who persist in believing either party is the sole culprit, the Clinton administration was a major driver of GLB reform, and the final vote in the Senate was 38 Aye, 7 Nay Dems and 52 Aye, 1 Nay GOP. Our current situation is as strongly bipartisan as it gets. McCain did not vote. Fitzgerald (R) voted present, as did seven Dem senators. In the house, the GOP voted it in 207-5, and the Dems voted it in 155-51. Thus each party now has a very strong incentive to obfuscate the issues here. Pelosi voted for it. Reid voted for it. House vote. Senate vote.)
The next part of the puzzle is basically covered by this 2007 post Long-Short Financial Gearing - as long as there was a crying need for the ABCP securities used as the fodder for the gearing system (money markets), nobody in investment banking gave a flip about the underlying risks. The reason why shoddy securities looked so appetizing was that the return on the securities was leveraged up by the spread between long and short rates.
Carl over at No Oil For Pacifists does not see my argument about why conglomerating financial businesses produces far more systemic risk. That is because Carl is a very logical thinker, and he operates by trying sort things into discrete sets (i.e. does this always have to mean this?). But that is not the type of thinking that is involved in banking and investment decisions, which always contain a degree of risk, i.e., the sets are always intermingled, and not only that, the degree of intersection is unknowable in any working system. Thus banking compensates for unknowability by requiring more profits to set up a reserve.
I tried to figure out many ways to get it across to him, and I finally concluded that I could do no better than by using Ben England's (he's a doctor) poker analogy:
In poker, the decent player stays or folds based on a relatively strict calculation involving the money at stake and the chance of winning the pot. In other words, if a relatively small bet gives you a small chance at winning a very big pot, you should stay in, or If:
Chance of Winning * Amount in Pot > $ amount to stay in, stay, but if
Chance of Winning * Amount in Pot < $ amount to stay in, fold. So a 30% chance of winning a $10 pot = $3.00 winnings. If I have to come up with $5 for the chance to win $3.00, I fold, because this is obviously a losing strategy over time. But if I have a 70% chance of winning a $10 pot ($7.00), that $5 bet may well be the best course. A 30% chance of winning a $30 dollar pot = $9 (not $12!) winnings, and it is generally rational to spend $5 for that chance. Beyond that one simply goes on knowledge of the other players' likely strategies to modify your calculations. So what I am trying to explain is that the recent breed of the best-and-brightest investment bankers are decent poker players, although terrible bankers. The difference between traditional bankers (commercial bankers) and investment bankers is that commercial bankers must pay attention to systemic risk. Investment bankers don't.
Allowing financial businesses to conglomerate and verticalize lines of business which have traditionally been handled by counterparties creates a situation in which the pot (potential return) gets much larger for each decision point. In other words, whereas in commercial banking the decision to say yes or no to a car loan or a home loan provides profit X, in a verticalized, conglomerated transaction the total profit from saying yes to a car loan or home loan may be more like 2X or 3X. Therefore, a straight line analysis will tell the poker-playing financial conglomerate to take on more risk.
Furthermore, allowing financial companies to be in the insurance business, the securitization and investment selling business, plus own the banking side (taking deposits and writing loans), plus verticalize (the holding company owns investment houses, banks, mortgage origination, servicing, and even insurance) destroys the independence of decision making.
Reprinting from the first post linked above, I reiterate that parties under the non-conglomerated financial system have a much lower incentive to take on risk in any single transaction, whereas when you allow the conglomeration of business lines, the profit from writing any particular appears much higher. This is because when you have a bunch of counterparties, each of whom is liable for all of the risk in their part of the transaction, but each of whom receives only a portion of the overall profit, the pot for each party is smaller. A lot more loans will be refused under this system:
The primary reason for the enhanced risks to a financial system that involves vertical lines of production and conglomerated functions is that it destroys counterparty checks and balances. Here is a somewhat simple example using mortgages.
Old way:
Bank writes mortgages. It lays off portions of its larger commercial loans to other banks (participations), keeps some shorter-term and variable mortgages, plus home equities, and sells the other consumer mortgage loans either to GSEs or to larger institutions which then batch them and securitize them. When selling mortgages the bank often retains servicing rights on the mortgages.
In writing its mortgages, the bank employs:
- An independent appraisal staff. These appraisers are reviewed annually by the board to see if their loans have produced losses on default, and the approved appraiser list is checked against the lists provided by the mortgage buying entities. The appraisers know that, and know that their profitable relationship with the bank will be jeopardized if they overestimate valuations. Thus their incentive is to be conservative and protect the bank from loss. By doing so they are maintaining their business lines.
- Independent title companies. The title companies used are bonded, licensed and and insured. If there is a problem, they will end up covering it.
- Property insurance companies. The approved ones are state-licensed and examined. The incentive to the bank is just to make sure that they are covered and their risk of loss is controlled. Property insurance companies are also rated and approved by the entities buying the mortgages.
- Mortgage insurance, if required by the bank's loan risk assessment and/or the mortgage buyer. Again, these companies are reviewed by the major players, such as GSEs.
When the bank sells mortgages, it does so under a system of covenants and representations, which basically require the bank to take back loans that are not properly documented and evaluated. So, for example, if a loan defaults almost immediately after initiation, the bank is going to be stuck with it or get it back. Likewise, if the loan later defaults and loss review at the buyer turns up an erroneous original verifications of income, or bad DTIs, or misrepresentations of income, or a funky appraisal, the bank gets its bouncing baby back. If the buyer doesn't review on default but MI is involved, the mortgage insurance company does review.
Further, the fundamental structure of Fannie-type mortgage securities was the pass-through, in which Fannie bought back defaulted loans from the investment pool. The way this system evolved tended to maximize longevity of players and accountability of players. To preserve access to Fannie, one generally anted up when asked to do so. And the same was true for the title insurance company, etc. In general, it was a trust-based system with a lot of accountability for the end results. The aggregator (the GSE) took nearly full responsibility for the outcome and passed that responsibility back down the line.
In this system, the profit of each counterparty largely depends on meeting mutual commitments (sustaining line of business) plus controlling one's own risks. Management of risks is independent. If one of the companies becomes financially unstable, it will lose business rapidly and pose potential losses to the other parties, so company "certification", whether formal or informal, tends to take place at several levels. Adequate capitalization is one of the prerequisites for maintaining your business relationships. The system as a whole is much bigger than any one bank or company, so participants can fail or be shuffled off to pasture without rocking the entire system.
As it ended up in the GLB era of bright and shining financial innovation:
Investment bank buys holding company. Holding company owns a bank and a bunch of mortgage companies that are non-banks, an insurance company, a title firm, and an outfit that contracts for appraisals. In addition, IB wants to maximize profits, so it sets up a subprime servicing company of its own. IB will generate mortgages through this origination arm, plus securitize them, plus write interest rate swaps on the resulting securities. In order to get large enough pools of mortgages, the IB also buys from other such non-bank companies and banks. In order for the banks to get enough mortgages to sell, many of the banks do wholesale mortgages, buying loans from mortgage brokers to sell through to the aggregators.
Note that the IB gets profit from virtually every step in this process. When one leg of it fails, it can plausibly argue that it is merely moving money from one arm to another. Further, the eventual investment product does not truly protect the investors. Unlike the old-fashioned, non-innovative, Fannie pass-through mortgage bond, buyers of the resulting tranched bonds do accept the risk of default in most circumstances.
See the "Reality Bites" issue of John Mauldin's newsletter, which includes a discussion of risk:
The biggest challenge facing policymakers is not short-term recovery, however. Eventually, stimulus is likely to arrest the forces of economic collapse and stabilize matters – at least temporarily. But the real problem is sowing the seeds of long-term, sustainable, organic economic growth. This is really the crux of the policy challenge. The United States in the midst of the worst economic downturn in 80 years as the result of a panoply of extremely poor economic policy choices. Economist Roger W. Garrison draws an important distinction between "healthy economic growth, which is saving-induced (and hence sustainable), and artificial booms, which are policy-induced (and hence unsustainable)."2 In other words, monetary policy that kept interest rates low for an extended period of time, tax policy that favored debt over equity, regulatory policy that allowed financial institutions to operate opaquely, and social policy that pushed home ownership regardless of affordability, all combined to create artificial economic demand that could only be financed with debt because the savings (i.e. equity) to purchase them did not exist.This issue (written by the author of the HCM Market Letter) goes on to discuss the problem of creating sustainable economic growth conjoined with policies that are jacking up government debt:
Another bout of policy-induced growth will not only repeat the mistakes of the past, but leave the economy even weaker, teetering on an unstable foundation of government support that cannot be sustained indefinitely without impairing America's balance sheet, credit rating, and ultimately its geopolitical might. Whether America's short-term political orientation can ever address this conundrum is the greatest question facing policymakers today. HCM has no hesitation in saying that much of what the government has proposed thus far to deal with the crisis won't come close to dealing with the long-term issue of creating savings-induced or organic growth. This means that any near-term relief (i.e. relief that occurs within the next five years) is most likely to give way to years of below trend growth because the economy will be lacking the organic foundation of growth it needs.End update.
In order to try to compress this a little, I am going to refer back to several prior posts.
The first is How Did We Get Here, written in September of 08. This covers some of the basics about the difference between banks and SEC regulated financial institutions, the role of REITs, and the repeal of Glass-Steagall (which essentially occurred piecemeal through the 90s, but was finalized in 1999). Please also see the Wikipedia entry, which is very informative.
(Btw, for those who persist in believing either party is the sole culprit, the Clinton administration was a major driver of GLB reform, and the final vote in the Senate was 38 Aye, 7 Nay Dems and 52 Aye, 1 Nay GOP. Our current situation is as strongly bipartisan as it gets. McCain did not vote. Fitzgerald (R) voted present, as did seven Dem senators. In the house, the GOP voted it in 207-5, and the Dems voted it in 155-51. Thus each party now has a very strong incentive to obfuscate the issues here. Pelosi voted for it. Reid voted for it. House vote. Senate vote.)
The next part of the puzzle is basically covered by this 2007 post Long-Short Financial Gearing - as long as there was a crying need for the ABCP securities used as the fodder for the gearing system (money markets), nobody in investment banking gave a flip about the underlying risks. The reason why shoddy securities looked so appetizing was that the return on the securities was leveraged up by the spread between long and short rates.
Carl over at No Oil For Pacifists does not see my argument about why conglomerating financial businesses produces far more systemic risk. That is because Carl is a very logical thinker, and he operates by trying sort things into discrete sets (i.e. does this always have to mean this?). But that is not the type of thinking that is involved in banking and investment decisions, which always contain a degree of risk, i.e., the sets are always intermingled, and not only that, the degree of intersection is unknowable in any working system. Thus banking compensates for unknowability by requiring more profits to set up a reserve.
I tried to figure out many ways to get it across to him, and I finally concluded that I could do no better than by using Ben England's (he's a doctor) poker analogy:
I used to play a fair amount of Texas Hold'em a while back. Particularly bad players like to bet a lot on marginal hands, and refuse to fold despite worsening odds on the upcoming cards. I've seen plenty of them get lucky and amass a pile of chips early in the game, but usually their sloppy play will catch up with them and they'll start losing hands.This is probably a better analogy than he even knows, although the moment I read it I was struck by the hilarious appropriateness of his thinking.
Once their stack is down, their play will get even more desperate, making larger bets on poor hands in hopes of hitting a big pot and restoring their bankroll. Soon their stack is completely gone, and they're rebuying back in the game, still playing aggressively in hopes of getting back.
Usually they'll leave the table down 80 bucks in a 20 dollar buy in game after being up 60-70 dollars.
In poker, the decent player stays or folds based on a relatively strict calculation involving the money at stake and the chance of winning the pot. In other words, if a relatively small bet gives you a small chance at winning a very big pot, you should stay in, or If:
Chance of Winning * Amount in Pot > $ amount to stay in, stay, but if
Chance of Winning * Amount in Pot < $ amount to stay in, fold. So a 30% chance of winning a $10 pot = $3.00 winnings. If I have to come up with $5 for the chance to win $3.00, I fold, because this is obviously a losing strategy over time. But if I have a 70% chance of winning a $10 pot ($7.00), that $5 bet may well be the best course. A 30% chance of winning a $30 dollar pot = $9 (not $12!) winnings, and it is generally rational to spend $5 for that chance. Beyond that one simply goes on knowledge of the other players' likely strategies to modify your calculations. So what I am trying to explain is that the recent breed of the best-and-brightest investment bankers are decent poker players, although terrible bankers. The difference between traditional bankers (commercial bankers) and investment bankers is that commercial bankers must pay attention to systemic risk. Investment bankers don't.
Allowing financial businesses to conglomerate and verticalize lines of business which have traditionally been handled by counterparties creates a situation in which the pot (potential return) gets much larger for each decision point. In other words, whereas in commercial banking the decision to say yes or no to a car loan or a home loan provides profit X, in a verticalized, conglomerated transaction the total profit from saying yes to a car loan or home loan may be more like 2X or 3X. Therefore, a straight line analysis will tell the poker-playing financial conglomerate to take on more risk.
Furthermore, allowing financial companies to be in the insurance business, the securitization and investment selling business, plus own the banking side (taking deposits and writing loans), plus verticalize (the holding company owns investment houses, banks, mortgage origination, servicing, and even insurance) destroys the independence of decision making.
Reprinting from the first post linked above, I reiterate that parties under the non-conglomerated financial system have a much lower incentive to take on risk in any single transaction, whereas when you allow the conglomeration of business lines, the profit from writing any particular appears much higher. This is because when you have a bunch of counterparties, each of whom is liable for all of the risk in their part of the transaction, but each of whom receives only a portion of the overall profit, the pot for each party is smaller. A lot more loans will be refused under this system:
The primary reason for the enhanced risks to a financial system that involves vertical lines of production and conglomerated functions is that it destroys counterparty checks and balances. Here is a somewhat simple example using mortgages.
Old way:
Bank writes mortgages. It lays off portions of its larger commercial loans to other banks (participations), keeps some shorter-term and variable mortgages, plus home equities, and sells the other consumer mortgage loans either to GSEs or to larger institutions which then batch them and securitize them. When selling mortgages the bank often retains servicing rights on the mortgages.
In writing its mortgages, the bank employs:
- An independent appraisal staff. These appraisers are reviewed annually by the board to see if their loans have produced losses on default, and the approved appraiser list is checked against the lists provided by the mortgage buying entities. The appraisers know that, and know that their profitable relationship with the bank will be jeopardized if they overestimate valuations. Thus their incentive is to be conservative and protect the bank from loss. By doing so they are maintaining their business lines.
- Independent title companies. The title companies used are bonded, licensed and and insured. If there is a problem, they will end up covering it.
- Property insurance companies. The approved ones are state-licensed and examined. The incentive to the bank is just to make sure that they are covered and their risk of loss is controlled. Property insurance companies are also rated and approved by the entities buying the mortgages.
- Mortgage insurance, if required by the bank's loan risk assessment and/or the mortgage buyer. Again, these companies are reviewed by the major players, such as GSEs.
When the bank sells mortgages, it does so under a system of covenants and representations, which basically require the bank to take back loans that are not properly documented and evaluated. So, for example, if a loan defaults almost immediately after initiation, the bank is going to be stuck with it or get it back. Likewise, if the loan later defaults and loss review at the buyer turns up an erroneous original verifications of income, or bad DTIs, or misrepresentations of income, or a funky appraisal, the bank gets its bouncing baby back. If the buyer doesn't review on default but MI is involved, the mortgage insurance company does review.
Further, the fundamental structure of Fannie-type mortgage securities was the pass-through, in which Fannie bought back defaulted loans from the investment pool. The way this system evolved tended to maximize longevity of players and accountability of players. To preserve access to Fannie, one generally anted up when asked to do so. And the same was true for the title insurance company, etc. In general, it was a trust-based system with a lot of accountability for the end results. The aggregator (the GSE) took nearly full responsibility for the outcome and passed that responsibility back down the line.
In this system, the profit of each counterparty largely depends on meeting mutual commitments (sustaining line of business) plus controlling one's own risks. Management of risks is independent. If one of the companies becomes financially unstable, it will lose business rapidly and pose potential losses to the other parties, so company "certification", whether formal or informal, tends to take place at several levels. Adequate capitalization is one of the prerequisites for maintaining your business relationships. The system as a whole is much bigger than any one bank or company, so participants can fail or be shuffled off to pasture without rocking the entire system.
As it ended up in the GLB era of bright and shining financial innovation:
Investment bank buys holding company. Holding company owns a bank and a bunch of mortgage companies that are non-banks, an insurance company, a title firm, and an outfit that contracts for appraisals. In addition, IB wants to maximize profits, so it sets up a subprime servicing company of its own. IB will generate mortgages through this origination arm, plus securitize them, plus write interest rate swaps on the resulting securities. In order to get large enough pools of mortgages, the IB also buys from other such non-bank companies and banks. In order for the banks to get enough mortgages to sell, many of the banks do wholesale mortgages, buying loans from mortgage brokers to sell through to the aggregators.
Note that the IB gets profit from virtually every step in this process. When one leg of it fails, it can plausibly argue that it is merely moving money from one arm to another. Further, the eventual investment product does not truly protect the investors. Unlike the old-fashioned, non-innovative, Fannie pass-through mortgage bond, buyers of the resulting tranched bonds do accept the risk of default in most circumstances.
