I am not an economics academic. I gave that game away for the lure of lucre and funds management. But this job throws up more than a few ideas for publishable economics papers – whereas when I was a student I was desperately short good ideas.
Here is one – a pure throwaway – for anyone that wants it. (It’s a nice paper for a masters thesis.)*
It has also explained neatly the problem of not getting banks to bring assets to the Geithner Funds – and in a way which I suggest is surprising.
It started as I tried to pick apart Rortybomb’s analysis of the Geithner Plan. Rortybomb does – in more formal form what Krugman does – analyse the non-recourse financing of the plan as a subsidy. He suggests that the non-recourse nature of the funding is a put option to the Treasury/FDIC of the assets – and that the correct way to model it is using (standard) option pricing models. Rortybomb’s posts are here and here. Krugman is a little more simplistic – but the idea is the same. Krugman produces a two-outcome model (rather than the range implicit in the option pricing model) and demonstrates there is a subsidy. Krugman’s post is here.
Anyway if you use a standard option pricing model and assume some volatility of outcome it is not hard to quantify the subsidy implicit in the plan. I have borrowed Mike’s (ie Rortybomb’s) diagrams. I hope he doesn’t mind.
The subsidy is dependent – as Rortybomb would acknowlege – on the leverage of the fund, the diversification of the fund and the variability of outcomes (particularly stress outcomes). All of this is standard option theory.
Now this is all fairly convincing until you work out that the party selling the assets (presumably a large and stressed bank) is also subsidized. The policy of the US Government (stated many times) is that there should be “No More Lehmans”. You may argue with this policy (reasonable people myself not included) think that this is the wrong policy. But you can’t argue that it isn’t the policy. The demonstration is that any bank that gets into any kind of liquidity trouble gets a “Sunday Night Liquidity Fix”. The availability of that Sunday Night Fix is a subsidy for the bank – just as surely as the non-recourse funding is a subsidy for the Geithner Fund.
So the issue is whether the Geithner Funds reduce the tail risk for the government – not whether the funds are themselves subsidized. After all the assets being sold are from non-recourse finance banks (losses beyond capital borne by the taxpayer) to non-recourse financed funds (losses beyond capital borne by the government). It depends on the relative solvency of the banks and the Geithner funds.
Thinking carefully there should be four broad outcomes:
• Both the bank and the Geithner fund is solvent
• Both the bank and the Geithner fund is insolvent
• The Geithner fund is insolvent but the bank is solvent
• The Geithner fund is solvent and the bank is insolvent
When both the bank and the Geithner Fund is solvent ex-poste there was no cost to the government. Sure there was an ex-ante subsidy but it didn’t cost anything. This case should not worry us.
The second case – when both the banks and the Geithner funds are insolvent the government will lose money – but it will lose less money than it would without the Geithner Plan. After all there was some private money in the fund – and that reduced the end loss borne by the government. In other words subsidy be damned - the plan reduced government losses.
The third case is problematic. If the Geithner Fund is insolvent and the bank is solvent then the Geithner plan cost the taxpayer real money.
The fourth case where the fund is solvent and the bank is insolvent is also problematic – but in a different way. The fourth case is where the banks sold good assets to the fund (presumably for liquidity) and kept the bad book for itself (because it could not sell it). Now in this case the subsidy to the Geithner Funds is not a problem – rather it is the desperation of the banks to sell assets, any assets and only being able to sell good assets. The more subsidy you give the Geithner Funds and the more competition between Geithner Funds you have (bidding up the price of the asset) the lesser the end problem for the banks. Either way however we shouldn’t be that stressed about the subsidy to the Geithner Fund.
Indeed the only place that we should be really stressed about subsidy to the Geithner Funds is the third case – where the fund is insolvent but the banks are solvent.
Oops. The people that are really stressed about the subsidy to the Geithner Fund (Krugman, Felix Salmon, Yves Smith of Naked Capitalism, Mike of Rortybomb) are also worried about or even convinced that the banks are insolvent. Indeed several of these people just advocate nationalisation now.
This is illogical. It is the second time I have accused Krugman of gross illogic – but it is simply illogical to believe that
(a). The banks are largely insolvent,
(b). The right or actual government policy is guarantee big banks (ie no more Lehmans) and
(c). The subsidy to the Geithner Funds is a real problem.
If both (a) and (b) applied the Geithner Fund MUST save the government money - so the subsidy is irrelevant.
This illogic extends to several of the bloggers I admire most. That is why I think there is a good academic paper in there. Krugman actually expresses “despair” over the subsidy. His despair is misplaced.
I guess the extension is to model it with many Geithner Funds, some of which are solvent, and some of which are insolvent. The situation might wind up more nuanced. Indeed my rough modelling (Monte Carlo rather than rigorous maths) suggests that it is more nuanced – but only slightly. The nuance disappears if the diversity of the Geithner Funds matches the diversity of the banks.
Success of the Geithner Plan
One concern with the Geithner plan is that the banks won’t actually come to the party and sell assets. It’s a concern taken up by Charlie Rose when he interviewed Timothy Geithner and dismissed in the Bronte Capital submission on administration of the plan.
Now – thinking about it I am not quite so sure. There are simple explanations as to why banks won’t bring assets to the plan – see for instance this post from Accrued Interest. But the most obvious reason is that the relatively good assets logically belong with the party with the biggest subsidy. And that might be the banks. The fact that banks won’t bring assets to the Geithner funds is in fact a measure that the relative subsidy of the Geithner funds is too low.
John Hempton
*At one stage I tried to contact Brad DeLong possibly about being involved in a PhD program at Berkley (ideally with him). We managed never to connect. And I have since given up that goal.
31 comments:
Can you clarify how the Geithner fund can be insolvent?
Here is what I don't understand.
The Geithner fund starts out with private investment, treasury investment and a FDIC non-recourse loan. So from the start it can't be insolvent right?
If the fund than grossly over pays for problem bank assets, the loan is still non-recourse. So how could Geithner fund be insolvent?
Am I missing some technical aspect about option pricing?
Thank you.
Ok - insolvent just means cost the government - ie defaults on non-recourse loan.
J
You say: "It is simply illogical to believe that
(a). The banks are largely insolvent,
(b). The right or actual government policy is guarantee big banks (ie no more Lehmans) and
(c). The subsidy to the Geithner Funds is a real problem."
Please explain. Thanks.
I think I have. If the first two conditions apply then the Geithner funds MUST save the government money. So we should not be worried about their subsidy.
J
i don't think your logic is correct.
if a bank is deemed insolvent and the government steps in to save it, the government is going to hold these assets to maturity so the government gets the hold to maturity price. i'm ignoring financing costs here because it's the government.
if a bank sells assets using PPIF and then is deemed insolvent anyway, the government gets the price that the PPIF pays.
so if the bank is insolvent and the PPIF makes money for the private investor, the government ends up paying the difference.
i think you are assuming that the private investors will lose money in this case, but that's not clear.
when looking at a combination of relaxing mark to market alongside the geithner plan it looks like the banks get to unload the really bad stuff and hold the not so terrible stuff until it rises in value or becomes really bad. then they can dump it off to the fund/taxpayer. wash, rinse, repeat.
am i wrong on this?
John,
Enjoy your thinking.
I think of the Geithner funds as a complicated, non-transparent way of providing bridge finance. The banks are a little short this year, having partied up big. If G can somehow overpay for risky assets that will tide them over and the shareholders/unsecured creditors get all the upside.
So there are three cases:
1) The banks are already solvent, but somehow aren't able to disclose that to the market (like Saddam and his WMD's). In this case the G investors reap a fantastic return on funds at risk.
2) The banks are bust, in which case the government will own the assets anyway and if the private G money lured in is larger than the shareholder/bondholder claims wiped out (unlikely IMO) the government comes out ahead.
3) The banks will become solvent given a little time, in which case the G plan enables the transition from state 2 to state 1; an incredible subsidy in itself.
It is this inequity in potential outcomes vs potential costs, along with the moral hazard implications going forward, to which I think the bloggers you refer to object. After all, there is a simpler, more transparent, tested and legal way to accomplish the same goal with no ex ante or ex post cost to the taxpayer - a normal FDIC restructuring. For this Buiter is best.
I really think you miss the point entirely. The largest reasons against this plan, and the most important ones to Krugman from what I can tell, are not exactly in the money.
The first flaw in this plan is that if the banks are sucessfully recapitalized and become solvent, the impetus to regulate and break up the banks (necessary for future stability) will all but vanish.
Second, this plan is unclear where the losses will come from, but it seems the funds have the least to fear (but the lack of clarity is really the ploblem).
Behind closed doors meetings of huge banks and hedge funds needs to stop. Ultimately, transparancy and regulatory power (and breaking up "too big to fail") are inherent and the best qualities of nationalization.
It's ALL about confidence. The final fund results will only be known in several years, but the cash flow should make the funds look OK with no mark-to-market issues.
If PNC or WFC sell a part of their VERY marked-down NCC or GDW loan books, they will show a profit on the sale, and there is a better prospect of the clouds lifting.
john: this is another interesting post.
some thoughts:
1. i think many people object to the Geithner Plan not just because of the subsidy, but rather because of the asymetry of returns on profits.
2. another objection is that the Geithner Plans will delay taking more expedited action and allow the situation with bank balance sheets to fester and deteriorate.
3. a third objection is if case 4 applies, then the gains by the funds will further poison the well of public sentinment.
4. one has to wonder why the plan wasn't more directed at raising bank capital directly by private participation in a fund (with the mix of private/public equity and gov non recourse debt in such fund to be arrived at through auction) which fund would directly guarantee the "troubled" assets in return for an equity stake granted to such fund in the banks (which i would think would be more appealing to the private sector for the larger asset base for upside). such a plan would elimate the pricing problem, engender private participation, eliminate the participation (and knowledge asymetry)problem and focus the real price discovery to a negotiation between private capital and the gov on the appropriate "subsidy" for the private participation.
The implicit assumption of Krugman and the others is that collective net losses put to taxpayers via the Geithner funds may exceed collective net gains for private equity investors. The examples they choose lead to such asymmetric outcomes.
In such outcomes, the taxpayer subsidy via the Geithner funds is greater than via the bank debt guarantee. These outcomes are possible because the granular structure of the Geithner put means it is generally closer to the money in total dollars at risk than the aggregate bank put:
If the bank sells the assets, the assets will generally end up being distributed or subdivided amongst multiple Geithner funds with multiple debt funding and collateralization arrangements, and multiple equity funding allocations. Based on the kind of outcome distribution assumed by Krugman and the others, the Geithner private equity investors can make money while the taxpayer loses.
If the bank retains the assets, they will be supported by the bank’s equity funding. But if the bank is insolvent, and the debt is guaranteed, all remaining assets effectively will be put to the taxpayer. Any assets that would have produced net gains for equity investors in the Geithner model would be used instead here to reduce the net put to the taxpayer. Therefore, the net subsidy in the bank case is less than in the Geithner case.
This structural difference between the Geithner put and the bank put could be described variously as the portfolio effect, the diversification effect, the correlation effect, the collateralization effect, or the granular put effect.
It should not be described as a volatility effect per se. Volatility affects both the Geithner put and the bank put. It affects the degree of superior value experienced in the Geithner put. But it is not the essence of the structural difference between the Geithner put and the bank put.
One potential criticism of Krugman and his ilk in this case is that they choose outcomes to fit their conclusion. But it is precisely the outliers that create the problem and the risk to the taxpayer. So it’s not clear that even this criticism is warranted.
In any event, they’re being consistent. Therefore, they’re not being illogical.
The use of the terminology “non-recourse” in these discussions is confusing and unnecessary. The fact that both bank debt and Geithner debt are non-recourse at the fundamental level of the equity investor is irrelevant to a comparison of the differences in these funding alternatives. The fact that both bank debt and Geithner debt have puts to the taxpayer similarly adds nothing to the comparison. The comparison rests on the structure of the put. The put structure is aggregate in the case of the bank. It is subdivided in the case of the Geithner plan. This means the collective outcome is different in the two cases, depending on the constituent outcomes. It is the effect of the subdivision of the put that creates all of the extra risk for the taxpayer in the Geithner plan.
When Krugman said he was okay with a nationalization that fully protected bank bond holders he lost me.
The primary benefit of nationalization is to cut costs by forcing bondholders to take a hair cut. That is why I would support it.
If that is not there then I do not see much of an advantage to nationalization.
As you point out, the subsidies needed to keep the system solvent are still there no matter how they get paid out.
there is a difference here as well, is there not, in the capital structures of the bank as compared to the PPIFs.
once a PPIF is through the sliver of private equity, it's all taxpayer loss.
once a bank is through the equity, there's a capital structure designed (in a different age, apparently) to eat losses. correct me if i'm wrong, but if none of the debt of the banks were backstopped, there would be more than enough debt in each and every of the majors to swap to equity and restore solvency in virtually any conceivable stress case. even in the more likely case that the senior debt is guaranteed, there would still be significant structure to absorb losses ahead of the taxpayer.
i suppose this depends on how one interprets the dictum "no more lehmans", but given that citi is already converting preferred to common, it would seem the adminstration is open to wiping out some of the structure -- how much depends on how insolvent they are and what there is to spend bailing.
I think JKH has it on excess cost.
In addition I think Krugman et al believe that we must allow banks to be nationalized and broken up (with attendant bond/share holder losses) so I'm not sure they would agree with (b).
And finally, it really is about the moral hazard. I'd like it to be another 80 years before the treasury is raided by the financiers.
Per my recollections, Krugman's "despair" is not because he thinks that the Geithner plan is a disaster or even that it would be a net negative. He does think some of the money might be wasted on solvent banks but more importantly, he thinks it will not be enough to save insolvent banks and that once this is clear, Obama will not have the political capital (and congress the will) to take further steps (read nationalization).
As I have written elsewhere this plan will fail because we are in a market for lemons with high level of asymmetric information (it’s not only clarity or liquidity). Sellers and buyers of toxic assets will not reach an equilibrium price or discover the “right” price unless somebody cheats. Uncertainty on the cash flow of the toxic assets is still there so it is the lack of confidence in and among the banking counterparts. Moreover there is no money around for this further securitization of toxic…
That is what led to this problem of toxic err legacy securities in the first place.
People start with a problem, take stupid and complex assumptions and prove conclusions that defy gravity.
Congratulations. You should think of starting to charge some consultant fees for your anti-commonsensical conclusions.
Although the analysis is elegant, I think it misses some of the details about the situation where the bank is insolvent. Let's take a simple example. The bank has one asset with a book value of $90 and $80 in liabilities, so on paper it is solvent. But with perfect foresight we know that the asset will be worth $70 in the future. I use the distribution of profits as shown in your chart from Rortybomb (except for the outcome distribution for the asset, since I take that as fixed).
In the absence of the PPIP, assuming "no more Lehmans," the government will lose $10.
With the PPIP, there are a few possibilities. If the buyer pays over $85, then the government will lose more than $10, because the PPIP will lose more than $15, and the hedge fund only takes the first $5.
If the buyer pays less than $70, then again the government will lose more than $70, because the bank now has less than $70, so it takes more than $10 to plug the whole in the bank balance sheet.
Between $70 and $85, the government loses less than $10 (better than the default).
So you can think the banks are insolvent and still think PPIP is a subsidy.
However, I think the real problem is that in my example the banks will not accept less than $90 (maybe a little less); they have no reason to given the "no more Lehmans" guarantee, and the buyers will not pay anything close, so the net effect is that nothing will happen (in general - I imagine a few assets will manage to change hands).
If the bad assets purchased by the PPIF funds were bought from insolvent banks only, then your point is correct. However, there is no restriction for the PPIF on who they can purchase assets from. If they purchase assets from solvent banks, then the subsidy is a concern.
And there is a great danger that only solvent banks will be able to take advantage of the liquidity provided by the PPIF, because the insolvent banks would be unwilling to mark down the assets to prices where the PPIF buyers would be interested (because it would make obvious that they are insolvent.)
We changed the accounting rules so that likely insolvent institutions can mask their insolvency for a period of time in the hope that the underlying economy changes and the current prababilities are proven wrong. This should be enough of a give away to the creditors of these institutions- for both the solvent and insolvent creditors are spared for some time. However, to further subsidize these creditors is unethical and amounts to no more than theft in my opinion.
You allude to this at the end of your post, but I think it could be clarified a bit further. There is a simplifying assumption that we can only be in one of the 4 possible states, ie., the possible combinations of solvency/insolvency for the bank and the Geither fund. However, since there are many banks, some solvent and some insolvent, and there would be many Geithner funds, some solvent some insolvent, we would find ourselves with outcomes distributed across all four states, in varying amounts. So we will see some of the downside of the bad states as well as some of the upside of the good states. I believe this is what you were alluding to with your brief comment about a Monte Carlo analysis. However, it would require some further demonstration as to why the proportion of bad outcomes would not outweigh the proportion of good outcomes.
My problem with your logic starts where you say:
'The second case – when both the banks and the Geithner funds are insolvent the government will lose money – but it will lose less money than it would without the Geithner Plan. After all there was some private money in the fund – and that reduced the end loss borne by the government. In other words subsidy be damned - the plan reduced government losses.'
This is not true. The Geithner fund can loss money and the private portion of the PPIP can gain money. This is what Mike at Rortybomb, among others, pointed out. In this case the private money in the fund did not save the government money, it actual cost them more.
The point is that the private portion of the PPIP and the government portion of the PPIP will not necessarily have the same outcome. The private can gain while the public losses.
The phrase "right or actual" is some nice rhetorical slight-of-hand.
To paraphrase: The government has already decided to privatize the gains and socialize the losses; therefore, anything the government does to reduce the losses MUST save the taxpayers money.
Well, I'm convinced. So, when do we vote?
Do I understand your argument correctly in saying that proponents of recapitalization should stop worrying because that is exactly what they are getting? It just happens through a different mechanism.
So banks are either recapitalized enough to be solvent or if they are still insolvent after the fund, the asset purchase was really just a down payment on recapitalization.
Can we be confident that this mechanism can correctly set the level of recapitalization? As noted in JK's example above, perhaps not.
In the other direction, could this scheme overshoot and transfer money to banks that don't really need it? i.e. over-recapitalize the banks.
John:
More than anything else, I think you exposed the casual haphazardness of Krugman's Monday morning blog post.
A mere sentence is needed to explain that the presence of cheap and easily accessible funding for an asset supports/inflates/drives up the price of such asset, and hence subsidizes its -natural- holders at the expense of the funding/debt providers. Krugman explained this in a lot more detail, and none of what he said was wrong.
I do think Krugman carefully avoided making a statement about the overall cost to the tax payer, he merely focused on the equity investors in the Geithner plan. Whether you believe such investor gets a subsidy or not will depend on the price of the asset indeed; and on bloggers' semantics.
I think your comments on the overall cost to the tax payer are indeed very relevant in this discussion. It is this perspective that Krugman didn't have the time or insight for to develop, and you did.
Can you please have a go at extending the analysis to the underlying bad assets, i.e. underwater mortgages? I suggest ignoring for the moment the "cloud of unknowing" created by too-clever securitisation, which I suppose has to be fixed by unbundling.
The US Government has great influence over the value of mortgages, for example by making writedowns easy or difficult for borrowers. An optimum policy from the perspective of the finance sector as a whole (Geithner Funds plus banks) is presumably one with the minimum number of foreclosures and the greatest continuing flow of mortgage payments - the most goose feathers with the least hissing, in Colbert's phrase. The current attitude of the banks to writedowns looks self-defeating.
Krugman seems to be responding to something like your argument in this blog post (though, perhaps inadvertently, he links to an unrelated argument on a different blog):
http://krugman.blogs.nytimes.com/2009/04/02/the-banks-versus-some-banks/
The conclusion depends on the analysis that case 4 is not a problem.
Case 4 is where the “bank is insolvent” and the “Geithner fund is solvent”.
The relevant comparison is between two alternatives for the resolution of ultimate cash flow – FDIC receivership or the PPIP fund structure.
If the bank is insolvent, and we examine the alternative of FDIC receivership, then all net cash flow, positive or negative accrues, to the FDIC. This includes net cash flow that would have otherwise accrued to PPIC had assets been sold to PPIC, but which still reside in the bank.
Thus, any cash flow representing positive returns on good assets in the case of PPIP accrues to the government on the very same assets in the case of FDIC receivership.
Thus, private investor equity gains realized in the case of PPIC are a component cash flow to the government in the case of FDIC receivership.
Thus, the private equity investor gain in the case of PPIP is an opportunity cost to the government when comparing PPIP with FDIC receivership.
Apart from this opportunity cost, all other net cash flow benefits or costs accruing to government in the PPIP case also accrue in the FDIC receivership case.
The only thing that matters to any of these PPIP/FDIC receivership comparisons is the experience of net cash flow for the government in comparison to net cash for the private equity investor. E.g. in the case where both the banks and PPIP are solvent in the generalized way portrayed, both private equity investors and the government are assumed to realize net positive cash flow. There is a net gain for the government in this case. Because it is not a net loss, there is less reason to complain about the fact that the private equity investor has extracted value (unless the government gets into the business of greed).
Finally, the only way case 4 could resolve itself in such a way as argued is if the transfer of assets from an insolvent bank to a solvent fund somehow makes the bank solvent. Of course, this contradicts the assumed matrix state of (insolvency, solvency) at the outset. If you assume the banks are largely insolvent as part of the problem, you can’t go about making them solvent as a result of the problem. If so, one throws up ones arms in disgust - the premise has included an assumed counterfactual state that flexibly, suddenly, and conveniently shifts in order to support the predetermined conclusion.
Krugman may not be right, but such analysis doesn’t prove he’s wrong.
the way i understand it the geithner plan means that private investors will sacrifice some money so the government won't be worse off. however, couldn't the government force some of the banks creditors to take a small haircut which would be equivalent to the private contribution in geithner plan.
I'm not smart at this stuff, but from the Geithner plan's complex setup of each pool, I don't see how there can be a lot of pools. In that case, the strategies presented by Krugman and others seem wildly risky since the space is tiny. What people seem to not have appreciated is that the FDIC loans are the senior debt of the pool, not of each asset.
My prediction, the big banks can't sell even to a subsidized market and there will be no legal or political challenges to takeover
I agree with your assessment. Most of the criticism og Gethner's plan appears to be on principle, not principal. Partnering with private capital reduces the risk to the government and should provide some much-needed profits to investors and capital to the banks.
Post a Comment