John and I were talking about the Geithner plan at dinner last night (at Bistro Petit Salut in Holland Village, which does a nice steak frites. It and the other restaurants along that stretch were noticeably empty last night.) I realized that I just don't understand it. Not just, I don't think it's a good idea, but actually I am confused about what's being proposed.
Right now nobody wants to buy these crappy mortgage-backed securities with their own money. Or, as I understand it, there are people out there willing to pay 20 cents on the dollar or something, but the banks can't accept that because it would devalue all their holdings; if they had to admit that these things were only worth 20c they would be manifestly insolvent. So the government is saying, look investors, we'll give you $93 to buy $100 worth of this paper. Obviously this is a great deal, so the people doing the buying are going to be willing to pay much more than 20c on the dollar--let's say they bid it up to 60c on the dollar. I take at that the plan is: we imagine that this is the new "market price"; the banks get a lot of cash from the government; and somehow private investors going forward are inclined to treat this "market price" as a reality, i.e. they will be willing to pay 60c on the dollar for this stuff, even without getting money from the government. Is this right?
Here are my questions. If I as an investor get the government's money only if I win the bidding war and wind up buying this paper, why won't I bid it up just ridiculously high? Isn't it the case that I will make some money as long as the paper is worth more than 7c on the dollar?
The government is guaranteeing the losses. Does this mean that they are just saying, you put up all the money ($100), but if the thing you buy turns out to be worth less we will reimburse you to the tune of $93? This would be cheaper for the government because they don't have to come up with cash now. Or is there not even enough investor money out there for that, so that instead the government is literally handing the investors cash up front so long as they use it to buy mortgage-backed securities?
I also wonder what price level corresponds to what view about how much more average housing prices are likely to fall (and whether people with good credit will walk away from deeply underwater loans). If I offer 50c on the dollar, what does that say about my belief in future housing prices? That they will decline 25% more? I guess this will be a bet on the combination of both factors, how much prices will decline and whether historical average rates of default will occur, or possibly there will be much higher levels of default.
I ask that because I was considering, if private investors were allowed to take part in this buying things with the government's money scheme, would I want to put down some of my money? It seems like a sweet deal but does the government really eat all the losses up to $97 on the $100 or does my $7 investment also decline proportionately? Wait, it can't, though, because how else would I be protected from the downside risk?
It's clear why this is a great deal for the banks, because they get rid of some of this crappy paper and get real money instead to try and fill the hole in their balance sheets. So in that sense it's just a roundabout way of handing money to the banks via some third party. But is it a great deal for the investors? It seems like it must be, because they are getting free money also; even if the paper is worth only 20c on the dollar they will make money because they mostly used the government's money to buy the thing. But how could this serve any purpose of "price discovery" or setting a market rate? Why won't the situation just be the same afterward, with private investors being willing to buy only at 20c on the dollar? Does this all rest on the assumption that the assets are mispriced due to irrational fear in the market, and once people see government-backed investors paying 60c on the dollar they will jump in and do the same with their own money? That seems crazy.
I don't get it either, but Mark CC takes a shot at explaining it to us here.
Posted by: Dave M | March 27, 2009 at 10:30 AM
Does this all rest on the assumption that the assets are mispriced due to irrational fear in the market
I think there's also an assumption that economic conditions are holding down the value, and the bank problems are worsening economic conditions in a vicious circle. They think they can break the log jam.
Or so the story goes. More likely it's just a backhanded bailout necessitated by political constraints.
Posted by: gnarlytrombone | March 27, 2009 at 11:25 AM
It seems like a sweet deal but does the government really eat all the losses up to $97 on the $100 or does my $7 investment also decline proportionately?
I'm pretty sure you lose all your money first, before the government loses anything -- the meaning of "the government takes the downside risk" is just that you don't have much riding on it, so you don't care if you lose your $7.
A non-recourse loan is like a mortgage where you always have the option of walking away and giving the bank your house. If you put down $20K, and borrow $80K, the bank has four times as much downside risk as you do. But the bank doesn't lose a dime until you've lost all of your $20K. For example, the value of the house drops to $75K. You walk away having lost your down payment, and the bank forecloses and sells for $75K, losing only $5K. But if the value of the house went to zero, you'd still only lose $20K, while the bank would lose $80K.
So, not as sweet a deal as you were thinking. But the problem is that we've got no idea at all what this stuff is worth, so we've got no idea how much it's reasonable to lend on it.
Posted by: LizardBreath | March 27, 2009 at 11:32 PM
What puzzles me is why we're buying up all the financial products that were based on underlying assets and loans, instead of just dealing with those assets and loans. For example, from what I understand, one of the big burdens on AIG is that it allowed lots of people to take out insurance policies on mortgages that those people weren't even holding; those people basically were just betting on whether the mortgage actually would get paid, without having the underlying interest. If a few million unpaid mortgages are enough to make all those policies come due and AIG have to pay up, in addition to the MBSes that were created, why not just prop up the mortgages themselves instead of the financial product superstructure built on them?
Posted by: PG | March 28, 2009 at 01:06 AM
Think of it as market prices on stilts.
Posted by: Jon H | March 28, 2009 at 09:27 AM
What LizardBreath Said. If you pay 50, the loss from 50 to 46.5 is yours (at 93% participation). The loss from 46.5 to 0 is on the U.S. taxpayer.
So there are three possible scenarios:
(1) You deliberately bid no higher than the amount that makes the government responsible for the current price. As a ballpark, this takes the bid price from 30 to 32. Transactions don't happen in much greater volume, because 44% of Citi's "assets" are priced at 80, and they won't sell unless you buy all of them for something close to that level.
(2) You agree to bid some inflated price (say, 55, midway between 30 and 80). You get some more flow, and are only putting up about $5 of that yourself. (You can pick up other assets around 30, with that price slowly going up--and if you're PIMCO, you've been doing that for the past several months--so if there ends up being a market-cleaning price after the Fed goes away around 45-50, you're in good shape. Citi, btw, is still bankrupt.)
(3) You bid $80 or thereabouts, taking a bullet to save Citi, taking a noticeable loss when the assets are ultimately revealed to be worth no more than 50-ish, given default and housing-price trends. This scenario is the "Santa Claus/Easter Bunny/Excellent Violist" in those old "you're driving a two-seater in the desert and see...broken down at the side of the road" jokes.
Note that the plan does nothing for the underlying assets--all it does is support the securities, which are now unbundled. This is why I'm generally Not Nice about Brad DeLong's "save the economy or punish the bankers" claim--the plan may save the bankers (security prices rise) but it does bupkus for the actual economy (the assets continue depreciating at speed).
Posted by: Ken Houghton | March 29, 2009 at 05:12 AM
The way it works is similar to "downpayment assistance" arranged for a mortgage borrower by a homebuilder's captive finance company.
First the bank calls up a friend (or creates an "off balance sheet entity") and gives the friend enough capital to pay 7% of the book value of a really bad asset package.
The friend bids book value and all the money is given to the bank, 7% coming from the friend and 93% from the government.
The bank walks away with cash and the friend loses the 7% but doesn't care.
The taxpayers get the assets, which in the cases where real liquidation occurred (mostly in relation to the Lehman collapse) have turned out to be worth somewhere between 9 and 12 cents on the dollar on average.
See, the important thing to understand is that the really bad assets aren't mortgages, they are magical securities that have an extremely tenuous legal relationship to mortgages. The "collateral" on these securities isn't a house, it is just another security that is slightly less distantly related to the original mortgage.
Proponents say the government is going to prevent banks from gaming the system. But the plan can't work unless it is gamed, because the banks are never going to accept a sale based on the real value of these assets.
Posted by: albrt | April 07, 2009 at 03:36 AM