JPM - JP Morgan

[valueinvesting101]

Buffet also mentioned in his interview on CNBC around Feb end that BRK is buying more WFC and held 400+ million shares. He also said BRK is buying more shares of two companies which are already part of the portfolio. I think, one of them could easily be WFC.

[Mephistopheles]

Buffet also mentioned in his interview on CNBC around Feb end that BRK is buying more WFC and held 400+ million shares. He also said BRK is buying more shares of two companies which are already part of the portfolio. I think, one of them could easily be WFC.

 

Yea it might be WFC. But I was just pointing out that he owns WFC in his personal portfolio as well, even though he says he can't own it because of BRK's position.

[HJ]

Zerohedge just put out a seemingly quite informed guess as to what happened. 

 

http://www.zerohedge.com/news/irony-101-or-how-fed-blew-jpmorgans-hedge-22-tweets

 

Basically they shorted mezzanine tranches of the CDX 9 to hedge systematic risks, but then "delta hedged" that short to minimize mark to market on that hedge.  Then in "delta hedging", through the 4Q-1Q credit rally, they are forced to buy more and more of the underlying index, cause the "hedge ratio" need to be dynamically adjusted.  The amount was so large that it distorted the relationship between the index and underlying individual CDS's.  The CDX market is just not that deep to absorb this amount of hedge.  Based on this guess, the publicised $100 billion notional, when adjusted for the short tranches, really end up being long a ridiculous amount of super senior risk, aka AIG in subprime, but this is referencing investment corporates, true systemic risk.  You expect to take MTM on the position, but hard to imagine huge economic losses on that position, unless the underlying unravels ala subprime.  But in winding down that position, they may realize some of the MTM's.  JPM book value will now swing with the credit prices a lot more than before.

 

Part of the problem is caused by having a very large amount of security on a bank balance sheet as opposed to loans.  You have to take mark to market swings.  If you hedge, you give up all the economics.  And if you try to be cute on the hedge, you have accidents like this.  The mortgage focused banks will face a version of this in the interest rate sphere if rates spike, the historical precedent being the S&L crisis.

 

[twacowfca]

Zerohedge just put out a seemingly quite informed guess as to what happened. 

 

http://www.zerohedge.com/news/irony-101-or-how-fed-blew-jpmorgans-hedge-22-tweets

 

Basically they shorted mezzanine tranches of the CDX 9 to hedge systematic risks, but then "delta hedged" that short to minimize mark to market on that hedge.  Then in "delta hedging", through the 4Q-1Q credit rally, they are forced to buy more and more of the underlying index, cause the "hedge ratio" need to be dynamically adjusted.  The amount was so large that it distorted the relationship between the index and underlying individual CDS's.  The CDX market is just not that deep to absorb this amount of hedge.  Based on this guess, the publicised $100 billion notional, when adjusted for the short tranches, really end up being long a ridiculous amount of super senior risk, aka AIG in subprime, but this is referencing investment corporates, true systemic risk.  You expect to take MTM on the position, but hard to imagine huge economic losses on that position, unless the underlying unravels ala subprime.  But in winding down that position, they may realize some of the MTM's.  JPM book value will now swing with the credit prices a lot more than before.

 

Part of the problem is caused by having a very large amount of security on a bank balance sheet as opposed to loans.  You have to take mark to market swings.  If you hedge, you give up all the economics.  And if you try to be cute on the hedge, you have accidents like this.  The mortgage focused banks will face a version of this in the interest rate sphere if rates spike, the historical precedent being the S&L crisis.

 

 

This situation highlights the Achilles Heel of delta hedging.  There is always a transaction cost of some amount in dynamic hedging.  The larger the position in relation to the value of the securities, the greater the proportional cost. This cost continues over time.  The cost may be small usually if it is a major market maker in the underlying security that does the dynamic balancing, but some transactional drag is everpresent.

 

This looks like a derivatives trader's worst nightmare where the trader is moving a market by the size of the trade and then moves the market in a related security even more in an effort to hedge the first trade.  Then, when other market participants see what's going on, they force the company in distress to pay up to rebalance.  Could this result in JPM's being jerked in the other direction when they balance the current position?

[Mephistopheles]

Zerohedge just put out a seemingly quite informed guess as to what happened. 

 

http://www.zerohedge.com/news/irony-101-or-how-fed-blew-jpmorgans-hedge-22-tweets

 

Basically they shorted mezzanine tranches of the CDX 9 to hedge systematic risks, but then "delta hedged" that short to minimize mark to market on that hedge.  Then in "delta hedging", through the 4Q-1Q credit rally, they are forced to buy more and more of the underlying index, cause the "hedge ratio" need to be dynamically adjusted.  The amount was so large that it distorted the relationship between the index and underlying individual CDS's.  The CDX market is just not that deep to absorb this amount of hedge.  Based on this guess, the publicised $100 billion notional, when adjusted for the short tranches, really end up being long a ridiculous amount of super senior risk, aka AIG in subprime, but this is referencing investment corporates, true systemic risk.  You expect to take MTM on the position, but hard to imagine huge economic losses on that position, unless the underlying unravels ala subprime.  But in winding down that position, they may realize some of the MTM's.  JPM book value will now swing with the credit prices a lot more than before.

 

Part of the problem is caused by having a very large amount of security on a bank balance sheet as opposed to loans.  You have to take mark to market swings.  If you hedge, you give up all the economics.  And if you try to be cute on the hedge, you have accidents like this.  The mortgage focused banks will face a version of this in the interest rate sphere if rates spike, the historical precedent being the S&L crisis.

 

Sounds interesting, but I'm still having trouble following all the technicals. Can anyone be kind enough to explain in layman's terms?

[twacowfca]

Zerohedge just put out a seemingly quite informed guess as to what happened. 

 

http://www.zerohedge.com/news/irony-101-or-how-fed-blew-jpmorgans-hedge-22-tweets

 

Basically they shorted mezzanine tranches of the CDX 9 to hedge systematic risks, but then "delta hedged" that short to minimize mark to market on that hedge.  Then in "delta hedging", through the 4Q-1Q credit rally, they are forced to buy more and more of the underlying index, cause the "hedge ratio" need to be dynamically adjusted.  The amount was so large that it distorted the relationship between the index and underlying individual CDS's.  The CDX market is just not that deep to absorb this amount of hedge.  Based on this guess, the publicised $100 billion notional, when adjusted for the short tranches, really end up being long a ridiculous amount of super senior risk, aka AIG in subprime, but this is referencing investment corporates, true systemic risk.  You expect to take MTM on the position, but hard to imagine huge economic losses on that position, unless the underlying unravels ala subprime.  But in winding down that position, they may realize some of the MTM's.  JPM book value will now swing with the credit prices a lot more than before.

 

Part of the problem is caused by having a very large amount of security on a bank balance sheet as opposed to loans.  You have to take mark to market swings.  If you hedge, you give up all the economics.  And if you try to be cute on the hedge, you have accidents like this.  The mortgage focused banks will face a version of this in the interest rate sphere if rates spike, the historical precedent being the S&L crisis.

 

 

This situation highlights the Achilles Heel of delta hedging.  There is always a transaction cost of some amount in dynamic hedging.  The larger the position in relation to the value of the securities, the greater the proportional cost. This cost continues over time.  The cost may be small usually if it is a major market maker in the underlying security that does the dynamic balancing, but some transactional drag is everpresent.

 

This looks like a derivatives trader's worst nightmare where the trader is moving a market by the size of the trade and then moves the market in a related security even more in an effort to hedge the first trade.  Then, when other market participants see what's going on, they force the company in distress to pay up to rebalance.  Could this result in JPM's being jerked in the other direction when they balance the current position?

 

 

You're right about what's going to happen when the banks start to increasingly hedge interest rate risk when the rates eventually start to move up in a big way in late 2012, 2013 or whenever.  The Achilles heel of that increased hedging is that some bank or other entitity has to be the counterparty and take the other side of those trades.  This means that the system is going to be net unhedged.  Those banks  that are on the wrong side of those trades are  going to bleed and possibly need another transfusion from the government or else go under.

[HJ]

Sounds interesting, but I'm still having trouble following all the technicals. Can anyone be kind enough to explain in layman's terms?

Depend on how much you know about tranches and indices already.  Basically say you carve up a credit index into 3 tranches, the bottom 10% take the first loss, the next say 20% take the next bunch of losses, and the top 70% being the super senior, all losses would be tallied at maturity.  JP Morgan says the market expect me to take the first 10% loss as a lender.  No problem there.  But losses beyond that impact my solvency.  So I'm going to short exactly that.  But then because all of this exist in derivative world, there will be a lot of mark to market P&L noises surrounding the hedge.  So together with shorting that tranche, I'm going to go long the index as a hedge to shorting that tranche to mimize the mark to market (but also to cheapen the hedge, since I don't want to just pay out the insurance premiums for that tranche, now I collect some premiums as well).  If I'm smart, I'll also buy some protection on the worst index names on top of that, basically end up being in a long super senior position.

 

The question now arises as to how much index to buy to offset that tranche price movement, what should be the delta, aka hedge ratio.  So the question becomes given a certain amount of MTM loss on the underlying index, how much of that should be attributed to the equity tranche, how much to the mezz piece and how much to the super senior.  Notice that it's not economic loss that I'm hedging, since for that, the pieces will just take losses in reverse order.  As it turns out, the PhD's discover that theoretically the way to distribute those losses should be based on this parameter called correlation.  Intuitively speaking, if the market movements is not that correlated, then all the MTM loss should all be absorbed by equity, since they likely reflect all future economic losses at maturity, so take them in reverse order.  But if the market is moving in a highly correlated fashion, then the market movement indicates a high likelihood of systematic impact, and the top pieces would likely suffer losses as well in the future.  So now for a given loss on the index, a substantial portion of it should be attributed to the more senior pieces.  Correlation can be measured by the covariance of the stock price volatility of the underlying names.  Now correlation changes over time.  During 3Q/4Q 2011, correlation across all industries and underlying names are quite high.  Therefore a higher than normal portion of index loss would be attributable to the more senior pieces.  i.e. delta was quite high.  Based on that, they probably didn't buy a lot of the underlying index to hedge the short, which was intended to hedge the systematic risk of the bank.  But as the systematic risk concern went away in 1Q, correlation went down.  Their model tells them now all movements reflect underlying economic concerns.  They realize that their short mezz position is losing value much faster than the appreciation of the index long.  So they "dynamically" adjust the hedge to buy more indexes all along the way.  Notice that they enter into contract on indexes, not on individual names.  The assumption is that there'll be arbs out there to arb out that index differences.  But there wasn't enough, and the index traded way tight compared to the underlying because of their buying.  Now all of a sudden, correlation spiked again in April.  They now realize they are too long.  Not only that, if they break out the index into underlying components, they are miss priced. 

 

If I were the risk manager now, I'd first pay up to buy protection on the worst names in the index.  And then take my chances on owning the super senior.  Take a look at that index components.  Imagine a scenario under which you suffer 10-20% loss, which probably correspond to something like 40% default rate (since bonds recover say 30%). 

[HJ]

You're right about what's going to happen when the banks start to increasingly hedge interest rate risk when the rates eventually start to move up in a big way in late 2012, 2013 or whenever.  The Achilles heel of that increased hedging is that some bank or other entitity has to be the counterparty and take the other side of those trades.  This means that the system is going to be net unhedged.  Those banks  that are on the wrong side of those trades are  going to bleed and possibly need another transfusion from the government or else go under.

 

The mortgage equivalent to correlation would be convexity.  Really the only way to avoid risk at the system level is not to take out the debt in the first place.  If you are going to need to hedge a risk after you take it out, why don't you just not originate the risk at all!

[returnonmycapital]

Having glanced at the balance sheet, I would hesitate to include JPM in the same company as USB/WFC.

 

Loans = 30% of assets (USB/WFC have on the order of 60%).

Trading & other securities = 36% of assets (USB/WFC closer to 20%).

Common equity + deposits = 56% of assets (USB/WFC around 80%).

 

JPM looks more like a leveraged investment/trading fund with a high cost of capital. Its wholesale funding requirement looks European. Its earnings will be less certain than either USB or WFC.

 

[Mephistopheles]

Sounds interesting, but I'm still having trouble following all the technicals. Can anyone be kind enough to explain in layman's terms?

Depend on how much you know about tranches and indices already.  Basically say you carve up a credit index into 3 tranches, the bottom 10% take the first loss, the next say 20% take the next bunch of losses, and the top 70% being the super senior, all losses would be tallied at maturity.  JP Morgan says the market expect me to take the first 10% loss as a lender.  No problem there.  But losses beyond that impact my solvency.  So I'm going to short exactly that.  But then because all of this exist in derivative world, there will be a lot of mark to market P&L noises surrounding the hedge.  So together with shorting that tranche, I'm going to go long the index as a hedge to shorting that tranche to mimize the mark to market (but also to cheapen the hedge, since I don't want to just pay out the insurance premiums for that tranche, now I collect some premiums as well).  If I'm smart, I'll also buy some protection on the worst index names on top of that, basically end up being in a long super senior position.

 

The question now arises as to how much index to buy to offset that tranche price movement, what should be the delta, aka hedge ratio.  So the question becomes given a certain amount of MTM loss on the underlying index, how much of that should be attributed to the equity tranche, how much to the mezz piece and how much to the super senior.  Notice that it's not economic loss that I'm hedging, since for that, the pieces will just take losses in reverse order.  As it turns out, the PhD's discover that theoretically the way to distribute those losses should be based on this parameter called correlation.  Intuitively speaking, if the market movements is not that correlated, then all the MTM loss should all be absorbed by equity, since they likely reflect all future economic losses at maturity, so take them in reverse order.  But if the market is moving in a highly correlated fashion, then the market movement indicates a high likelihood of systematic impact, and the top pieces would likely suffer losses as well in the future.  So now for a given loss on the index, a substantial portion of it should be attributed to the more senior pieces.  Correlation can be measured by the covariance of the stock price volatility of the underlying names.  Now correlation changes over time.  During 3Q/4Q 2011, correlation across all industries and underlying names are quite high.  Therefore a higher than normal portion of index loss would be attributable to the more senior pieces.  i.e. delta was quite high.  Based on that, they probably didn't buy a lot of the underlying index to hedge the short, which was intended to hedge the systematic risk of the bank.  But as the systematic risk concern went away in 1Q, correlation went down.  Their model tells them now all movements reflect underlying economic concerns.  They realize that their short mezz position is losing value much faster than the appreciation of the index long.  So they "dynamically" adjust the hedge to buy more indexes all along the way.  Notice that they enter into contract on indexes, not on individual names.  The assumption is that there'll be arbs out there to arb out that index differences.  But there wasn't enough, and the index traded way tight compared to the underlying because of their buying.  Now all of a sudden, correlation spiked again in April.  They now realize they are too long.  Not only that, if they break out the index into underlying components, they are miss priced. 

 

If I were the risk manager now, I'd first pay up to buy protection on the worst names in the index.  And then take my chances on owning the super senior.  Take a look at that index components.  Imagine a scenario under which you suffer 10-20% loss, which probably correspond to something like 40% default rate (since bonds recover say 30%).

 

Wow, thanks a lot! I understand it much better now. But I still have a few questions. When they say "correlation", is that the correlation between the different tranches? Or the correlation between different names in the index (which causes MTM losses across all tranches)?

 

Also, why do they care so much about MTM losses, and not economic losses? Is it because it is a highly levered institution and so volatility matters a lot? It would be nice if they can just not care about quarterly billion dollar MTM swings, like Berkshire with its equity index puts. In the end when all of this plays out, will this be mainly a MTM loss or will JPM suffer a heavy economic loss?

[nkp007]

I just bought a bunch of the TARP warrants. The coverage seems overblown.

 

In five years, either

 

1) I'll feel real good because all my TARP warrants (AIG, BAC, JPM), and Citigroup equity are reflecting intrinsic value because the U.S. financial system didn't blow up and hyper-analyzed balance sheets are on the whole, quite reliable.

 

or

 

2) The world financial system has blown up and I'm eating canned beans. All banks have been replaced by local credit unions and government controlled processes.

[dcollon]

For whatever it's worth, below are some brief comments by Bill Ackman

 

http://www.forbes.com/sites/monteburke/2012/05/16/bill-ackman-the-jpmorgan-debacle-is-not-a-big-deal/

 

[HJ]

Wow, thanks a lot! I understand it much better now. But I still have a few questions. When they say "correlation", is that the correlation between the different tranches? Or the correlation between different names in the index (which causes MTM losses across all tranches)?

 

Also, why do they care so much about MTM losses, and not economic losses? Is it because it is a highly levered institution and so volatility matters a lot? It would be nice if they can just not care about quarterly billion dollar MTM swings, like Berkshire with its equity index puts. In the end when all of this plays out, will this be mainly a MTM loss or will JPM suffer a heavy economic loss?

 

Correlation references the correlation of price movements of the underlying components to the index.  They observe it by calculating the covariance of stock price movements of those components.  Technically, the correlation across tranches of the same index is 1. 

 

As to why the concern of MTM, you should ask the genius lawyers at Basel.  It affects capital ratios, which affect all the capital management activities.  The difference between a well capitalized bank and a poorly capitalized bank is just a couple percentage points of capital buffer.

 

Bear in mind that this is just a guess of what happened.  Still quite shocking as to the size of this trade. 

 

This said, it's not a foregone conclusion having just mortgage loans is always going to be the favored asset mix on bank balance sheet.  Interest rate risks on mortgages are notoriously difficult to hedge.  Among the many causes of the S&L crisis, one of them was that for an extended period of time, they had mortgages on their books which was yielding less than what they had to pay out to attract deposits. 

But that's a discussion for a different environment.

 

[PlanMaestro]

This is the only JP Morgan thread I found in the idea section. I think it is time to analyze it as one. This is the infamous trade:

 

How JPMorgan’s storm in a teapot grew

http://www.ft.com/cms/s/2/6197eb2a-9f64-11e1-8b84-00144feabdc0.html#ixzz1v8vKyvYL

 

Two problems faced the bank. First, simply shorting credit was an expensive affair: JPMorgan was not the only institution with concerns about the future. And second, market volatility promised to make any directional position – short or long – punishing in the short term.

 

The trades the CIO embarked on, it hoped, would overcome both problems. Consistent with the view that credit was due a correction, CIO traders looked to “short” credit indices. Rather than short the indices as a whole, however, they bought credit protection on specially constructed baskets of subordinated credits known as tranches, say people familiar with the supposed trade.

 

Buying such protection was expensive and prone to volatility. So for the second part of the trade, JPMorgan hedged by selling protection on the IG.9 as a whole.

 

Because of the mechanics of the trade, in order to achieve a “market neutral” position, whereby JPMorgan hedged the bet against volatility as best it could and offset the cost of its short positions, the bank had to sell far more units of cheap protection on the IG.9 as a whole than it bought on short, more expensive tranches.

 

Taken as a whole, the trade would pay off if a credit market correction damaged some investment-grade companies but was not so severe to dramatically impact them all.

 

[Liberty]

This is the only JP Morgan thread I found in the idea section.

 

There's one here:

 

http://www.cornerofberkshireandfairfax.ca/forum/investment-ideas/jp-morgan/

[PlanMaestro]

http://dealbook.nytimes.com/2012/05/16/jpmorgans-trading-loss-is-said-to-rise-at-least-50/?hp

 

At the bank’s annual meeting in Tampa, Fla., on Tuesday, Mr. Dimon did not definitively rule out cutting the dividend, although he said that he “hoped” it would not be cut.

 

John Lackey, a shareholder from Richmond, Va., who attended the meeting precisely to ask about the dividend, was not reassured. “That wasn’t a very clear answer,” he said of Mr. Dimon’s response. “I expect that shareholders are going to suffer because of this.”

 

Analysts expect the bank to earn $4 billion in the second quarter, factoring in the original estimated loss of $2 billion. Even if the additional trading losses were to double, the bank could still earn a profit of $2 billion.

[PlanMaestro]

The infamous trade:

 

How JPMorgan’s storm in a teapot grew

http://www.ft.com/cms/s/2/6197eb2a-9f64-11e1-8b84-00144feabdc0.html#ixzz1v8vKyvYL

 

Two problems faced the bank. First, simply shorting credit was an expensive affair: JPMorgan was not the only institution with concerns about the future. And second, market volatility promised to make any directional position – short or long – punishing in the short term.

 

The trades the CIO embarked on, it hoped, would overcome both problems. Consistent with the view that credit was due a correction, CIO traders looked to “short” credit indices. Rather than short the indices as a whole, however, they bought credit protection on specially constructed baskets of subordinated credits known as tranches, say people familiar with the supposed trade.

 

Buying such protection was expensive and prone to volatility. So for the second part of the trade, JPMorgan hedged by selling protection on the IG.9 as a whole.

 

Because of the mechanics of the trade, in order to achieve a “market neutral” position, whereby JPMorgan hedged the bet against volatility as best it could and offset the cost of its short positions, the bank had to sell far more units of cheap protection on the IG.9 as a whole than it bought on short, more expensive tranches.

 

Taken as a whole, the trade would pay off if a credit market correction damaged some investment-grade companies but was not so severe to dramatically impact them all.

[Guest hellsten]

For what it's worth, here's a free report on JPM from ValueLine (Updated May 18th):

http://www3.valueline.com/dow30/f1899.pdf

[Rabbitisrich]

http://www.americanbanker.com/bankthink/JPM-Risk-Management-CIO-VaR-hedging-1049268-1.html

 

 

If there had been offsetting profits somewhere, JPM would not have been compelled to make this announcement.  Similarly, if other activities were generating losses they would have been added to the preannouncement. These positions lost some $150 million to $200 million a day and as volatility increased the models should have been showing higher VaR levels than the $178 million and brought the position to management's attention.  This is what VaR is meant to do. The problem was the position could not be liquidated in a day. When that is the case, VaR is no longer the appropriate management tool. Stress tests are.  JPM has been silent about what the stress tests showed and the bank should be clear about that.

[Uccmal]

So JPM suspends the buyback to save cash.... And the stock... Well, goes up. 

[WarrenWatsa]

I agree with this comment I read below the CNBC article on the buybacks being suspended:

Just so I'm clear, JPM was buying back stock at $43.00, but now that it is $33.00, it will halt buybacks.  This is why stock buybacks are the biggest waste of shareholder money!

 

Good job Diamond.  Let this blow over, and next year you can resume the buybacks when the stock is in the $40s.

 

http://www.cnbc.com/id/47503768

[alertmeipp]

Well, which part you don't understand? Things have changed, you can't deal with different things with the same plan.

[petec]

I doubt this is by choice.  I thought Dimon's explanation in his letter of how he thought about the buyback was excellent and in stark contrast to this decision, which makes me think there is regulator pressure here.

[ERICOPOLY]

I agree with this comment I read below the CNBC article on the buybacks being suspended:

Just so I'm clear, JPM was buying back stock at $43.00, but now that it is $33.00, it will halt buybacks.  This is why stock buybacks are the biggest waste of shareholder money!

 

Good job Diamond.  Let this blow over, and next year you can resume the buybacks when the stock is in the $40s.

 

http://www.cnbc.com/id/47503768

 

The banks argued that the Fed stress tests and capital plans taken together don't make sense because it doesn't allow for the banks to cut their repurchase program during times of losses.  The banks argued they would eliminate/curtail share buybacks if they were taking on water..

 

The Fed argued that even as the losses were mounting in 2008, the banks kept on buying back shares.  So they effectively told the banks that past behavior says otherwise.

 

So, is JPM going to buy back shares now when they said they would act otherwise?

 

Might be related to the spat with the Fed over the capital planning, might not be.

[Viking]

This looks to me to be a classic case of a one time issue creating a great buying opportunity in a world class company. Will JMP lose a chunk of money? Yes. Is it material? No, especially when looking at it over what the company will earn over the next 5 or 10 year period. Is the moat at JPM breached? Not at all.

 

The company has lost more than $50 billion in market cap since March 30 and almost $30 billion in market cap since disclosing the $2 to $3 billion issue (give or take a couple of billion).

 

Given the halt on stock repurchases, my guess is Mr. Market will continue to drive the stock lower. Great buying opportunity for those with a time horizon greater than 3-6 months. Should the stock sell off much more from current levels, buyers may want to use one of those once in a lifetime puch cards that Buffett talks about (i.e. back up the truck).

 

Perhaps we could benefit from Ericopoly educating us all a little and providing a suggested strategy of how to benefit from the current situation with JPM (just thinking back to FFH a few years ago and BAC last year)?

 

Here is the link to the presentation Dimon made today (45 minutes) for exactly where things stand today with JPM. http://investor.shareholder.com/jpmorganchase/presentations.cfm