1000 Alitalia in one shot: so' forti 'sti amerikani ... (I)

22 settembre 2008 michele boldrin
Trying to make sense of what is happening in the US financial markets, and of what the Fed and the US Government are doing, is not that simple. I start today, in the hope to finish the second and third part by the end of the week. The Italian versions should also be appearing, with a one-day delay.

''Equities and derivatives''

Because we will be talking about "financial assets" and "financial derivatives" let me start by making clear to those who are not familiar with this terminology what these things are. A financial asset is, in general, a title to a future (and uncertain) stream of payments coming from somewhere under certain circumstances. The "somewhere" in the last statement matters as, for our purposes, it is useful to distinguish between two types of financial assets depending on the nature of their "somewhere". We call the first type "positive net supply securities" and the second "zero net supply securities" or, for short, "equities" and "derivatives". Obviously both actual equities (e.g. stocks and secured bonds) and derivatives (e.g. options) are aspecial case of what I call here "equities" and "derivatives".

To have an equity there must be some real asset out there,to the (uncertain) fruits of which the owner of the equity has a title: a tree, a horse, a house, an oil field, a company, and so on. These are real material assets, the value of which the equities are supposed to reflect: hence the positive net supply feature. Notice that, even if there is no slavery, the existence of intellectual property rights and other kinds of contractual arrangements imply that equities may entitle someone to the (uncertain) fruits of someone else's labor, through ownership of the firm where the someone is employed, and so on. I am stressing this to make clear that lots of "things" that are out there can be owned via one form of equity or another. In fact, most of them are and, from the theoretical point of view at least, EVERYTHING that exists out there and has some productive potential is a material asset the ownership of which can be structured via equities. Because the annual flow of goods and services we call GNP (in fact, more than that, but forget details) is the product of the material assets existing out there, and the value of equities is nothing but the present discounted value of the (uncertain) fruits produced by something that exists out there, we have the following simple implication, which is actually relevant to understand the mess we are in.

The total market value of all "equities" is equal at most (i.e. when the ownership of every productive asset is represented by an equity) to the present discounted value of all the (expected) future GNPs that existing assets can produce. In other words, the market value of all equities of a country (or of the world, if you like big things) is bounded above by a multiple of current GNP (of the country, of the world).

Just to give an idea of the numbers involved, follow me into some algebra. Currently we estimate the 2008 US GNP at almost 15 trillion USD and my friend Ellen McGrattan (who knows these numbers very well) estimates that the US capital stock (including houses) is about 4 times that (Ellen, I am rounding it up).That makes about $60 trillion. Now, this calculation is a bit of a cheat, because GNP contains also the fruits of labor, and labor is accounted for in the capital stock, in the form of equities, only very very partially. So, if you want to follow the theoretical argument strictly, you may add to those $60 trillion the (implicit) capitalized value of labor and human capital, reaching a number around $150 trillion - to do this use the fact that capital income is a tad more than 1/3 of GNP and assume the return to human capital and labor interms of GNP is about the same as the return on what economists call "physical capital"). If you add to this the (implicit) value of all those other things that are produced but not marketed (e.g. your dinner at home, for the part that has to do with cooking, or your clean and ironed shirt if you do it yourself, or the market value of you resting or doing other, nicer, things in bed, etcetera) you may get an even bigger number, which I have no idea what it may be. Say $300 trillion: still a finite number, which is what matters.

Now, let me consider derivatives: financial securities that are in zero net supply. Contrary to an equity, which requires only an asset and a person, a derivative security is a bit like tango: it takes two (people) to make it but the asset "out there" is not strictly needed. It works like this. Person A tells person B: if X takes place on day D I pay you $100, if not I pay you nothing; how much are you willing to give me today in exchange for my signature on such a promise? If B says: I give you $P and A says"ok", a derivative is born. Obviously what A tells B may be very complicated, and it may involve lots of different circumstances (i.e. if in D we get X I pay 100, if we get Y we wait for D' in which, if X' occurs then I pay 40, if Y' occurs then I pay 3 and if Z' occurs then we wait for D" in which etcetera). The things that can happen are anything conceivable and observable - better: anything that A and B agree that, when they tango (oops, create) the derivative, they will be able to observe when D comes ... once D comes, who knows ... which is also relevant for our argument, but let me not jump ahead - and the payments can go either way. In any case, when A and B create the derivative, they agree on a price P>0, which is paid (say) by B to A. At that point we are in the hands of Fortuna: B hopes that in D the good (for him) thing happens and also hopes that A keeps its (that's a substitute for the annoying his/her) word or, better, is capable to keep it and pays up.

Now, notice this: a derivative has, in a sense that I hope is clear, nothing to do with outstanding assets and with their products, at least in principle. Certainly, many derivatives are defined with respect to the behavior of some underlying real assets or, at least, equity (see here for slightly more technical details) but this is not necessary. When you bet $10 with your friend that tomorrow it will rain in San Francisco you have created a derivative, and so you do when you purchase some form of insurance on your car, or when you buy a lottery ticket: everyone trades in derivatives and humans have been trading in derivatives since the very far past; no need to have a PhD in physics or to go to the Chicago Board of Trade to do it! Because of this fact, the number and, more important, the value of POTENTIALLY outstanding derivatives at any given point in time is ... infinite! Well, not really, because IF they were properly priced and IF things were done properly, their value would in fact always add up to zero: they are zero net supply securities or, as I like to call them, "zero sum games, if no one cheats".

Summarizing:(I) an unbounded number of derivatives can be created, irrespective of the existence of actual productive assets "out there"; (II) if "properly priced" and if everyone involved in the game plays fair, the net total value of outstanding derivative is nevertheless zero because for every winner there is a loser of equal magnitude (in net present value); (III) derivatives are instruments for either insuring or gambling, the two things being in fact one and the same thing with the sign inverted; (IV) derivatives are re-distributive securities: they are not associated to the creation of a new productive asset (that is the role of equities, in the extended sense), instead they redistribute wealth from A to B or from B to A depending on the outcomes of random events people have agreed upon before hand, events people believe they can observe and the probability of which they can assess.

''How derivative markets should work''

Among other things, the facts listed so far imply that, when markets function "properly" (i.e. according to theory!) a change in the value of a derivative may, per se, imply nothing about the value of outstanding real assets and viceversa. A change in the value of a real asset will be reflected by a change in the value of the equity representing it. The change in value of the derivative that may have been written "on" that equity (real asset) only determines WHO, between our fictional A and B characters, will gain from the change in the value of the asset and who will lose. Moreover, the facts listed above also imply that when the value of the underlying asset (equity) changes by △ the associated derivatives may imply gains (losses) for A (B) equal to manytimes △.


Because this aspect is very important, let me elaborate on it using the simplest example given before. Say B offers $50 to A, in exchange for the promise that "if the event X takes place on day D I pay you $100, if not I pay you nothing", and A accepts, so they can tango. Good. Say A believes - for whatever reason, I have no intention in this article to get into the story of how people form their expectations, talk to De Finetti (or to my friend Paolo Siconolfi who, contrary to De Finetti, is still alive and very well) for that topic - that there is a 10% chance that X will happen on day D, and 90% that it will not. If A does nothing it (A) is taking a risk: true, there is only one out of ten chances that X will happen in D but, if it does, A has to give $100 to B and B just gave A only $50. Since, to make things simple, D is tomorrow and it is now 11:53 p.m. of today, there is no interest to be earned in the meantime on those $50. Hence, either (i) A has its own funds to shovel up the extra $50 if shit (pardon,X) happens or (ii) A insures itself in the way I describe in a few lines, or, (iii) it is planning to cheat, which is the same thing as to say A is taking a risk: the risk of defaulting on its promise (derivative) to B. Before following the path into the dark woods of contemporary Wall Street finance that (iii) opens to us, let me follow (ii) for a bit, i.e. the path that finance textbooks teach normal and well regulated financial or insurance markets should follow.

In "normal and well regulated financial markets" adhering to (ii) one of the following (essentially equivalent) things should happen. A regulator - it need not be the government, it could be a reliable third party overseer that both A and B trust and who is REALLY impartial (i.e. not S&P or Moodys ...) and has the power to enforce its ruling - will somehow force A to set aside either another $50 (over and above the $50 A just received from B) or whatever a reasonable pater familias would expect to be worth $50 in case X (i.e. shit) happens. This can be done in a number of ways: for example, A could be asked to look around for lots (N) of people like B, sign similar derivative contracts with them until the law of large number can be really expected to hold. Say N=100 is enough for the law to hold. Then A has received $5000 and may have to pay up to $10000 if X happens. Because the probability that X happens 100 times in D is infinitely small (and so are the probabilities that it happens 99, 98 ..., 51 times) A (and the many Bs) are facing no risk of disaster. As long as no more than 50 Xs happen in D, A has funds to pay to the designated Bs what it owes them. There are other ways in which a well regulated market can achieve the same outcome. For example it may ask A to "reinsure" its risk with C, who has the required funds in case X happens more than 50 times; alternatively it may ask A to sell some of the derivatives it is holding to other kinds of Cs who also can cover those payments, and so on. Notice that, in so doing, many more derivatives are created from the original one and, because the process of selling the same bet (or bets on the original bet, or bets on betson bets ...) can be repeated an endless number of times, the notional value of outstanding derivatives originated by that initial simple contract between the first A and the first B can become very very very large.

In all these cases, though, if the theoretical mechanism of properly functioning derivative markets were working properly the (many) As that had taken out the bets would be able to honor them when shit (oops, X) happened. To understand this, imagine that the event X in the original contract was "the market price of your (B's) house drops by $10". The house is a nice physical asset and B owns it entirely (the argument needs no mortgages, for now ...) but has decided to "super-insure" itself. That is fine, and derivatives allow just that. The point is that the drop in the value of real assets is just $10 for the whole economy, and B is the one suffering it, which is why it is (super) insured. Now A owes $100 to B and, even if there is a very very long chain of derivatives going from A to B from B to C ... all the way to Z'''', as long as $100 are collected (as all those derivative contracts are settled) and they end up in the hands of A, A can pay B its dues. To achieve this result three things are needed: (1) a total of at least $100 in either bills or "credit" must be available in the system; (2) none of the individuals involved cheats or finds itself able to grab the bills it must pass on to the next person in the chain; and, last but not least, (3) $100 of actual real nice and edible output(GNP) is available for delivery to B in exchange for those $100. These three conditions are all crucial, but (3) is especially subtle: if there is no real nice and edible output available worth $100, but just $100 dollar bills made of unedible paper, prices will go up proportionally (maybe with the long and variable lags of Milton) and B will not really receive $100. It will receive only worthless inflated pieces of green paper.

''Curious things may happen''

Interestingly enough, failure of (3) is, in our actual historical circumstances, strictly linked to the occurrence of (iii) (not (ii), (iii)) and it seems to be the way in which we (well, not me: the US Fed, the US Treasury and all those other powerful guys) have decided to pretend that (ii) happened when, in fact, it was (iii) that really tookplace. Before getting to this point, though, we will need to go through other two steps. First we will need to have an idea of what the value of outstanding derivatives is today in the US, so we can talk real numbers. Second, we will have to figure out why neither (i) nor (ii) happened, but instead (iii) did and none (none?) of us figured it out until the other day. The first part I do now, the second, which is longer, will have to wait for tomorrow or the day after. The third part will then discuss why the "solution" we are apparently heading toward may be anything but a solution. Better said, it is a solution for the few lucky As, but not for the many and unlucky Bs. Derivatives, remember, are a distributive tool and it is (gigantic) income and wealth redistribution we are talking about these days.

''Some numbers, and a puzzle''

Anyhow, for tonight let me just end with the estimates I promised. How big can the notional value of outstanding derivatives become? Very large indeed. Recall that, using my friend Ellen's careful estimates, I un-carefully valued the total amount of outstanding US real assets at around $300 trillion. In the Wikipedia page on derivatives linked above we read:

Over-the-counter (OTC) derivatives are contracts that are traded (and privately negotiated) directly between two parties, without going through an exchange or other intermediary. Products such as swaps, forward rate agreements, and exotic options are almost always traded in this way. The OTC derivative market is the largest market for derivatives, and is unregulated. According to the Bank for International Settlements, the total outstanding notional amount is $596 trillion (as of December 2007)[1]. Of this total notional amount, 66% are interest rate contracts, 10% are credit default swaps (CDS), 9% are foreign exchange contracts, 2% are commodity contracts, 1% are equity contracts, and 12% are other. OTC derivatives are largely subject to counterparty risk, as the validity of a contract depends on the counterparty's solvency and ability to honor its obligations.

That is: just the OTC derivatives (which, as you will see, are our main concern in this saga) have a notional value that is twice my bold estimate of total US equities of any kind, which included those that do not exist because, for example and for good reasons, slavery is forbidden. If we stick to what I claim Ellen's estimate of the outstanding capital stock of the US is, the ratio is 10 (ten). To these OTC derivatives one should add the exchange traded ones, that the IBS estimated to be at around $400 trillion in 2006. Now, these numbers are for the whole world, but even if we assume that those "related to the US economy" are only 20% of the total (and they are much more) we are talking $200 trillion in notional value of derivatives, versus an actual capital stock which is about 1/4 of that and a GNP that is ...$5 trillion short to make it 1/10 of it.

Hence the puzzle: if only 1/10 of them derivatives "had to be paid" (what this means and who should pay whom, we will see in the second part of this trilogy) where the hell do we find the money? Even by loading the entire GNP of the US into the backyards of the lucky winners of the "tango game", call them Bs, we would not only be going hungry ...we would be short of $5 trillion. Peanuts, no?

Well, don't the Fed andthe Treasury own two printing presses, one for $$ and one for debt? Yes indeed, they do. Stay tuned.

The second part is here

28 commenti (espandi tutti)

Thank you for the explanation. Long‚ but could not have been shorter.

I have a question. While we can justify the utility of an equity market (to allocate resources to the most productive...) can we justify the need to play these zero-sum-games outside of Vegas? It might not be convincing‚ but I am sure there is some sort of justification. Is there a textbook definition of why this "insurance" helps the market?

You can justify them through consumption smoothing. If I have an asset that pays in each period 1 in state 1 and 0 in state 2 and you have an asset that pays 0 in state 1 and 1 in state 2 we can be better off by exchanging some claims on each other's asset instead of staying in an autarky position.

But then if we can "both" be better off, it's not a zero-sum-game. I am thinking out loud...

No, the fact that we are both better off depends on the assumption of concavity of our utility function. It is the base of the life-cycle theory of consumption. That is, I am better off if I consume 0.5 in each period instead of consuming 1 in one period and zero in the other. To give an example, I prefer to eat a sandwich everyday instead of eating once per week at the restaurant.

I get it: U(1/2) > [U(0) + U(1)]/2, or something like that. The point is there is some value from this service... Actually, the point is I just wanted to be the first making a post with "when X hits the fan" ;-)

I have a question. While we can justify the utility of an equity market
(to allocate resources to the most productive...) can we justify the
need to play these zero-sum-games outside of Vegas? It might not be
convincing‚ but I am sure there is some sort of justification. Is there
a textbook definition of why this "insurance" helps the market?

The "textbook definitions" usually given are hedging, speculation and arbitrage (see Wikipedia).
Hedging works like this. Say you are an airline and you wish to hedge your exposure to the price
of oil. You can enter into a forward/futures contract to buy X amount
of oil for Y price at time Z, or you can buy an option. Say you are
based in the EU, do business in the US, are due to receive/make a USD
1,000,000 payment in six months, and want to hedge your exposure to the
EURUSD exchange rate. Again, you can enter into a forward contract or
an FX option. Or: a bank is willing to lend you money, but it's in the
"wrong" currency and you want to hedge away any risk that the FX rate
or the interest rate in that currency might move against you in the
future, so you enter into a cross currency swap. Etc.

Non vi dispiace, vero, parlare come mangiate ? When you are in US do like America does, ma siamo tutti italiani...anche se parliamo e scriviamo un buon inglese.

Non vorrei dirlo, ma penso che Boldrin ha scritto in inglese perchè, forse, ha preparato qualche lezione per i suoi studenti, o cose simili, se poi si vuole far partecipare al dibattito qualcuno più importante , tipo BS o GT,stiamo freschi con l'inglese.

Marco, come sai cerchiamo sempre di parlare come mangiamo. L'articolo di Michele appartiene a una categoria (anglonoise) in cui mettiamo riflessioni che possono essere di interesse anche a colleghi e lettori che non parlano italiano. L'attuale crisi finanziaria USA è un ovvio candidato, per cui Michele ha scelto di scrivere il pezzo in inglese.

Siccome questo è un sito in italiano letto prevalentemente da italiani, mettiamo anche la traduzione, come peraltro tendiamo a fare quando riportiamo pezzi lunghi in inglese nei nostri pezzi. Devi solo avere un attimo di pazienza; non abbiamo segretarie e traduttori, quindi le traduzioni le facciamo  nel tempo libero, che non è molto. Comunque adesso la traduzione è disponibile schiacciando la bandierina italianasotto il titolo.

Grazie Sandro, ma il mio scritto era diretto ai commentatori, nulla da eccepire sullo scritto in inglese di Michele (anche se la traduzione è molto divertente). Poi se uno è inglese, o americano, e commenta in lingua, no problem, doesn't matter.

Per gli altri è un problema commentare in italiano ? Guardate che se ci legge BS non capisce un'acca e se ne va !

Pensate che ai nostri seminari cafoscarini si parla in inglese anche se tutti italiani :-)

Eh beh... presumo anche tu sia stato presente al Liquidity and Credit Risk oggi. Attendiamo il proseguio delle ormai epiche "Boldrin Trilogy", che tanto appassionano i lettori di NfA, me compreso che ho un background accademico aziendalistico, ma ora proiettato verso economia e finanza.

no, quella conferenza non è roba per me!

allora a tor vergata non so l'unici! :O

 

che è 'n gioco de pupi che imitano i grandi?

Io non ho idea della cinghia di trasmissione dei problemi nei derivati nell'economia reale, nelle cose che produco e vendo, ma ho l'impressione che l'intervento della Fed e del Tesoro USA serva più a salvare degli immensi patrimoni personali, fatti in un Casinò stile Las Vegas,così che quando si è scoperto che il Banco era sotto e non poteva pagare c'è stato il panico, ed adesso interviene il fesso che paga.

Questo mi ricorda "L'aereo più pazzo del mondo" con la lucetta "panic" o "don't panic" che si accende a comando.

Forse andavano salvate Freddie Mae e Fannie Mac, ma anche per un keynesiano come me questo intervento puzza molto (anche perchè per Keynes i mercati dovevano essere regolati, non aiutati)  non parliamo del tentativo (?) di reintrodurre Bretton Woods, come argomentato dal "Sole24ore", assolutamente improponibile. 

Seguirò la "saga" boldriniana, ma avendo investito dei soldi (non molti, calmi) nelle azioni delle società energetiche USA, mi dite come mi vedete ? -)

Bloomberg informa che le banche d'investimento USA spariscono. Niente male ...

Direte voi, cosa c'entra l'inconsistenza temporale? C'entra.

Nel caso classico il governo, dopo aver promesso di non tassare, tassa colui che ha investito dopo che il capitale è stato accumulato.

Nel caso Bernanke-Paulson (così passerà alla storia) il governo assicura e protegge dopo che si è mangiato capitale e rendimenti colui che, non essendo regolato ed investendo a modo suo, è finito nei guai.

As I said: 1000 Alitalia. Il più grande "reverse Robin Hood" della storia contemporanea. Not bad, professor Bernanke, not bad.

1. La prima inconsistenza temporale mi sembra essere la tua notte, niente male come riposo, praticamente 0.

2. Poichè sono molto d'accordo con te, ci spieghi perchè, come mi sembra di ricordare tu abbia scritto, Fannie Mae e Feddie mac andavano salvate, mentre i giocatori di hold'em no ? Io la vedo così: Fannie Mae e Freddie Mac erano coinvolte nell'economia reale, i mitici mutui subprime, quelli a clientela non proprio  di prim'ordine, ovvero avevano a che fare con equities, roba che prima o poi indietro ci torna, i giocatori di hold'em stavano per cavoli loro ad un tavolo verde, e nessuno gli ha chiesto di giocarsi le mutande. Sbaglio ?

Gli Stati Uniti sono (o erano?) famosi per avere introdotto da tempo una legislazione antitrust che non avrebbe dovuto consentire posizioni monopolistiche o comunque un peso eccessivo di una singola società nel mercato di appartenenza. Ma le dimensioni di Fannie e Freddy o della AIG erano talmente enormi che non si è potuto non intervenire. Forse era il caso di intervenire prima che diventassero cosi grandi.

Pero' non lasciamoci prendere la mano col populismo di dagli ai ricchi. O meglio, facciamolo pure ma senza dimenticare l'analisi. Il prior che Bernanke-Paulson adesso stiano aiutando gli investment bankers ce l'ho anch'io. Ma questo e' un prior, non analisi. L'analisi dice che per poter dire se l'aiuto c'e' o meno bisogna vedere a che prezzo comprano quello che comprano. Il fatto che Bernanke-Paulson vogliano carta bianca e definiscano in modo vago cosa compreranno e a che prezzo; e soprattutto il fatto che qualcuno queste cose che comprano le deve gestire e poi vendere e che di conseguenza oggi a Wall Street ci siano migliaia di persone con la mano alzata a dire "io, io, io, le gestisco; io" mi fa pensar male. Pero' una giustificazione razionale per un intervento che non sia un bail-out, cioe' un acquisto a prezzo inferiore al "valore reale se ben calcolato" ma superiore al valore di mercato ci sarebbe.  Naturalmente, questo solo se i) il "valore reale se ben calcolato" e' superiore al valore di mercato (a causa di panico, informazione asimmetrica,..), ii) la Fed puo' calcolarlo bene il "valore reale se ben calcolato". Discutiamo se questo sia il caso o meno, piuttosto che non se ci sono simpatici gli investment bankers. 

Alberto, che la Fed e il tesoro rischino di fare un affare è un dato di fatto, anche perchè possono agire su più fronti, ad esempio nazionalizzando freddie mae e fannie mac hanno acquisito i mutui subprime molto al di sotto del nominale, e se fra dieci anni il valore delle case torna quasi ai livelli pre-bolla, et voilà il gioco è fatto, e per loro 10 anni non significano niente, allo stesso tempo risaliranno le quotazioni dei derivati che si poggiavano su quei crediti, e, bingo !

Quello che penso, e che così come Alitalia è un cadavere, e solo la cocciutaggine non ne fa constatare il decesso, così non mi piace (e a molti altri, a quanto pare) che i contribuenti debbano pagare quello che altri hanno danneggiato, anche se nel futuro potrebbe essere un affare (allora ditelo: lo facciamo perchè sembra un buon affare, e per noi 10 anni non sono un lasso di tempo irragionevole).

Domandiamoci, anche, Keynes e Friedman che avrebbero detto ? (il secondo nulla, avrebbe detto lasciamo in pace i mercati..)

 

domande mica da ridere, se fossimo in grado di dimostrare che i) real value > market value è vera, si potrebbe pensare di mettere su un vulture fund, qualcosa tipo nFA asset management

Dear Michele,

Just to get your point straight...it sounds like you are blaming the cheaters.

I think is useful to split the picture into two slices: 1) mortgages and the first layers of them and;  2) any  derivatives related to 1).

I really don't think we can characterize 1) as cheating. A lot of players bought the very risky assets  either because a bad precification or in the expectation that they could sell them before the crash.

In 2) perhaps we can say something about cheating. For example, AIG and Lehman have tons of contracts in the swap default kind of contract. I think is reasonable to say that they knew the aggregate risk and so, given the amount of contratcts, that can be characterized as cheating.

Is that more or less what you have in mind?

What about optimal policy in the current situation? I guess this is what we really care about  by the end of the day.

To the Minnesotan (e firmatevi, che vi costa?): I am trying to think the problem through by writing it down. I am not sure I have it clear, yet ... I do have some ideas in mind, yes, and having gone over them at lunch with, among other, Narayana who's here today has helped. As other colleagues have, in previous days. I am not blaming it on the cheaters, per se. I argue that there are four causes, interacting: (1) bad Fed policies since the late 1990s until now, these policies have fueled the localized inflation and created a gigantic moral hazard problem in the banking sector; (2) the nature of OTC derivatives allows for improper "profit taking" when profits are not necessarily there in an economic sense; (3) very poor market design and even poorer regulation at all levels, from the mortgage originators to the investment bankers trading CDSs; (4) an essentially oligopolistic situation in the "Wall Street industry", with a captured regulator (but this may be just an implication of (1) and (3)). It is clear that (2)-(4) also create incentives for cheating, and people cheat when allowed to.

I have no idea what an "optimal" policy is, for a simple reason: there is a conflict of interest and the issue is one of redistribution. So optimal for whom? Forget Pareto optimality with side payments, please. It is informationally unfeasible.

My opinion on the current measure is reflected, in part, by what David Levine says here.

Conflitto d'interessi. D'ora in poi, purtroppo, non potremmo più usare l'esempio USA quando si parla di conflitti d'interesse. Paulson è l'ex CEO di GS, e vale (mi dicono) nell'ordine di $800 million i quali sono probabilmente investiti in un trust fund che sarà anche blind ma che contiene di certo grossi pezzi di Wall Street firms. Ora, a questo signore vengono dati $ 700 billion per comprare quello che vuole, come vuole ed ai prezzi che vuole dai suoi amichetti e previous business associates di Wall Street. Questo, ovviamente, batte qualsiasi conflitto d'interessi italiano hands down.

Credo sia meglio mi dedichi alla seconda ed alla terza puntata, in ogni caso.

Non penso che il problema principale sia "cheating" quanto piuttosto "leveraging". Le Investment Bank non hanno ingannato il mercato con premeditazione, almeno inizialmente. Semplicemente si imbottite di titoli che dapprima sembravano redditizi, anche se non moltissimo in termini unitari, ma che moltiplicati per la leva finanziaria che ha caratterizzato i bilanci delle IB, hanno prodotto lauti guadagni (insieme alle commissioni di collocamento di tutta questa robaccia). Oggi la leva continua a produrre il suo effetto moltiplicativo amplificando le perdite e si mangia il capitale. Devo altresì aggiungere che "cheating" c'è stato. Ma non parlo dei derivati OTC quanto piuttosto di tutti i SIV/SPVs (off balance sheet vehicles) che sono stati usati 1) dapprima per accrescere ulteriormente la leva finanziaria e 2) poi per nascondere le perdite. Una questione interessante da chiarire è: perchè le IB si sono riempite di titoli la cui liquidità è peggiorata progressivamente ? In qualche modo si è trattato di una sorta di illusione finanziaria collettiva in cui anche le agenzie di rating hanno avuto un ruolo da protagoniste continuando ad attribuire giudizi lusinghieri su titoli sempre meno presentabili per non prosciugare il ricco flusso di commissioni che le stava beneficiando come mai prima (lo stesso W. Buffett si è comprato una fetta di Moody's). Le banche, in cui probabilmente la mano dx ignorava quanto faceva la sx (in certi casi le "muraglie cinesi" funzionano benissimo...), hanno continuato ad accumulare i titoli perchè gli stessi avevano dei rating elevati. Poi i pagamenti delle rate dei mutui hanno cominciato a perdere qualche colpo e il castello ha cominciato a crollare.

Visto che in questo giorni sentiremo ad nauseam argomenti di policy basati sull'irrazionalità dei mercati, segnalo a fini preventivi dal blog inglese stumbling and mumbling la seguente osservazione, a cui non avevo inizialmente pensato: chi ritiene che la gente sia irrazionale e possa essere guidata meglio nelle proprie scelte dagli ''esperti'' dovrebbe fermarsi un attimo e riflettere che gli ''esperti'' sono proprio quelli che hanno creato tutto questo casino. Veramente dobbiamo avere più fiducia del loro giudizio su ciò che è bene per noi, piuttosto che del nostro? Essendo il blog inglese e di orientamento left-libertarian l'osservazione è costruita come polemica con il leader conservatore David Cameron, uno degli ultimi entusiasti convertiti a behavioral economics, ma il punto è valido più in generale.

Un paio di questioni sulle definizioni. Sembri suggerire che derivatives sia una forma di inside money. Mentre per assegni e cambiali non ho problemi ho difficoltà a  pensare ad un contratto di assicurazione come una forma di inside money. E' questo quello che mi stai dicendo?   

Prendo il tuo riassunto:

 (I) an unbounded number of derivatives can be created, irrespective of the existence of actual productive assets "out there"; (II) if "properly priced" and if everyone involved in the game plays fair, the net total value of outstanding derivative is nevertheless zero because for every winner there is a loser of equal magnitude (in net present value); (III) derivatives are instruments for either insuring or gambling, the two things being in fact one and the same thing with the sign inverted; (IV) derivatives are re-distributive securities: they are not associated to the creation of a new productive assets (that is the role of equities, in the extended sense) but redistribute wealth from A to B or from B to A depending on the outcomes of random events people have agreed upon beforehand, events people believe they can observe and the probability of which they can assess.

Sul punto 1 inoltre: 

everyone trades in derivatives and humans have been trading in derivatives since the very far past; no need to have a PhD in physics or to go to the Chicago Board of Trade to do it! Because of this fact, the number and, more important, the value of POTENTIALLY outstanding derivatives in existence at any given point in time is ... infinite!  

 La domanda è questa: dal punto 1 sembra che stai dando la definizione di moneta. Mi stai dicendo che derivatives sono (inside  money) e l'origine del casino è che sono state usate come tale in modo disinvolto?  

Ma se invece sono contratti di assicurazione (dalla definizione riportata avanti sembra proprio di si) perchè mai il valore dei potenziali contratti di assicurazione dovrebbe essere infinito, l'importante è che siano in zero net supply e assimilabile a inside money? Che differenza c'è tra il contratto che dici tu:

Person A tells person B: if X takes place on day D I pay you $100, if not I pay you nothing; how much are you willing to give me today in exchange for my signature on such a promise? If B says: I give you $P and A says "ok", a derivative is born.

e il mio contratto RCA (è l'unico contratto di assicurazione che mi posso permettere  e quindi l'unico di cui ho un'idea)? Il valore dei contratti RCA è potenzialmente infinito? Non capisco.

Dopo aver descritto derivative come impegni reciproci di agenti privati dici anche che  "derivative has, in a sense that I hope is clear, nothing to do with outstanding assets and with their products, at least in principle".

Non capisco. Siamo in endowment economies? Inoltre  scambiare state contingent claims ad un certo livello- certo non quelli che ci scambiamo io e mia moglie- lo possono fare solo in pochi con le tecnologie giuste, come fare le automobili o i missili intercontinentali o le torte. Queste torte, per quel che ne so, si fanno con  precisi ingredienti (riserve tecniche, tavole di probabilità, copayments etc.)  perchè abbiamo permesso di farle fare ai banchieri? Mica facciamo operare al cuore i dentisti o cavare i denti ai cardiologi! Questo mi sembra il costo vero del bail out. Non solo ti ho fatto fare un mestiere che non eri in grado di fare (fesso chi ci è capitato ma questo è un altro problema), ma ti salvo pure dopo che hai combinato casini. Prima ancora che una questione redistributiva (which matters only ex-post e che sarebbe sorta in ogni caso come sempre quando lo stato del mondo si rivela, qui il punto è che la fregatura sembra essere ex-ante) qui il punto è di efficienza. E' qui il punto di regolamentazione.

 

 

dubbio sul puzzle finale e i numeri in gioco: ma perché il nozionale dovrebbe essere una proxy più o meno buona dell'effettivo ammontare che le controparti devono liquidarsi? O la definizione di nozionale è un'altra rispetto a quella canonica ?

dubbio legittimo, risolto alla puntata che, spero, apparirà domani mattina italiana. È un blog, non Econometrica quindi ... mi permetto dei trucchi retorici che altrove sarebbero vietati :-)

bubbles

giulio zanella 23/9/2008 - 21:07

The total market value of all "equities" is equal at
most (i.e. when the ownership of every productive asset is represented
by an equity)
to the
present discounted value of all the (expected) future GNPs that existing
assets can produce

What about bubbles? Equities can be overpriced (this is how all this began, inflated housing prices and subprimes, no?), GNP cannot (either there's something to eat now and in the future or not).

 

Inizia una nuova discussione

Login o registrati per inviare commenti