When variation margin attacks: Difference between revisions

When variation margin attacks (view source)

Revision as of 13:33, 15 December 2023

14,192 bytes added , 15 December 2023

no edit summary

Amwelladmin

Bureaucrats, Interface administrators, Administrators

82,853

edits

@@ Line 154: / Line 154: @@
 Under a swap at any time, the [[out-of-the-money]] counterparty has [[market risk]], because it is losing on the trade, and the [[in-the-money]] one has [[credit exposure]], because it stands to lose if the other guy can’t pay out its profit. [[Variation margin]] addresses that [[Credit risk|''credit'' exposure]] by requiring the [[out-of-the-money]] counterparty pays out variation margin equal to its moneyness, in full, every day even though the game is still going. If things get worse, you have to give some of the variation margin back. If you start losing, you have to pay variation margin to the house.
+=== On the case for one-way margin ===
+In recent years — ironically, just as the “dealer” vs “customer” dynamic has become more pronounced<ref>After the [[GFC]], bank proprietary trading fell away to almost nothing.</ref> — the global regulatory-industrial complex,<ref>This label is not just sardonic: there really is a cottage industry of “regulatory change management professionals”, in-house and out, who owe their last decade’s livelihood to ''accommodating'' quixotic regulatory initiatives like this. They are a powerful lobby with a direct interest in maintaining the rate of regulatory churn.</ref> still fighting last decade’s war, forged rules which overlook this plain asymmetry. Notably, the coordinated worldwide approach to bilateral [[regulatory margin]]. As swap positions move in and out of the market, counterparties must post each other the cash value of the net market movements each day. This is a little like closing positions out at the end of each day and settling up, with a key difference: you ''don’t'' close out your positions. The valuations at which the parties exchange margin are guesstimates. The parties stay on risk.
+Well — ''one'' of them does — as per the above, the customer has risk; the dealer does not. The customer was the one who initiated the trade, to put itself into a market position of some sort. The dealer didn’t initiate the trade, but accommodated it in the expectation only of commission and on the explicit grounds that its market position would not change and the customer’s credit position would be satisfactory.
+Requiring margin — even guesstimated margin —from a ''customer'' who is net [[out-of-the-money]] makes sense: if the customer fails, the dealer’s hedges are defeated and it will be have open market exposures to the customer’s positions. From the point of view of systemic risk, the last thing anyone wants is a dealer whose hedges fail. That is when it can go bust. So, daily [[variation margin]] ''to the dealer'' mitigates that risk to date; [[initial margin]] covers it for the future, should the dealer have to close out hedges against a defaulting customer.
+As long as the dealer is covered, there will be minimal market disruption and the dealer’s own solvency is not threatened.
+But requiring a ''dealer'' to post margin to its customer to cover the customer’s net in-the-money positions makes no sense whatsoever.
+First, customers — and here I mean [[buy-side]] market participants ''who do not themselves post systemic risk''<ref>There are different considerations for those who do pose systemic risk, but these should be dealt with by equivalent capital regulation and limitations on leverage and so on: in a perfect world, buy-side entities would never get so big as to pose systemic risk.</ref> — are trading on their capital, dealers are not.<ref>Dealers hold capital primarily against counterparty failure, remember, not market risk itself: absent counterparty failure they should have none.</ref> They ''willingly'' put themselves in “harm’s way” in the hopeful expectation of a return on their equity. Dealers do not. ''Customers take risk'': that is what they are there for. Except through customer misadventure, ''dealers do not''.
+Of course, dealers ''do'' present some risk of insolvency, and customers should only tolerate so much exposure to that risk, but the customer has other levers to manage it. They can close out their positions, take profits and re-establish their position at the current level, or with another dealer, for one thing. If they do that, the dealer can close out its hedge, pass on gains whilst being off risk, and then restrike its hedges and initial margin at the higher level if need be.<ref>A grave factor in [[Credit Suisse]]’s losses on [[Archegos]] was “margin erosion” caused by massive appreciation on its swap positions. While Credit Suisse was unusual in not using “dynamic margining” (which solves the “margin erosion” problem) to its [[Synthetic prime brokerage|synthetic equity derivatives]] book, “static” [[initial margin]] is the rule for other asset classes, and for [[regulatory IM]].</ref> ''This is not the same as paying out the [[mark-to-market]] of a unrealised swap''.
+To be sure, customers might not ''like'' doing this — realising a taxable gain and having to stump up more [[initial margin]] when re-establishing positions blows the kumara, for sure — but none of these are good reasons for anyone but the customer. Withholding [[variation margin]] on profitable positions gives customers the choice: you can ''either'' keep your position open, but your money with the dealer, avoid tax and live with the “dealer risk”, ''or'' book your gain and get your money back and start again. This encourages prudent behaviour. If nothing else, it incentivises customers to diversify their risk across dealers.
+And it does not automatically lever up the customer’s portfolio. For what do we think a customer will ''do'' with all that free cash [[VM]] its dealer keeps sending it? If it was planning to just sit on it, wouldn’t just — ''leave it at the bank''?
+Secondly, [[dealer]]s and [[bank]]s are already capitalised and regulated for systemic risk.<ref>[[Broker/dealer]]s that are not deposit-taking banks are more lightly capitalised. But nor — for that very reason — can they hold customer assets and cash on their balance sheet, but must hold it on trust for customers with a client money bank that ''is'' capital regulated.</ref> There are already constraints on how they must operate, and how much capital they must hold against the contingency of portfolio losses. Dealers hold this capital, in large part, to protect against the risks presented to them ''by customers''. Customers like thinly capitalised, highly-levered, investment funds. If no customer ever fails, nor will a delta-hedging dealer.
+That risk is amplified if dealers must pay away their own cash to reflecting their customers’ unrealised gains on a derivative portfolio ''already 70% financed by the dealer''. It’s just mad: “Hi. You already owe me 70% of the value of the stock you bought largely with my money, and you want ''me'' to pay you margin if the stock goes up?”
+This is all the more mad if the dealer is hedging with a physical asset. ''No-one pays variation margin on gains on a physical asset''.<ref>Dealers can, and do, manage this by financing their physical portfolios. They would do this anyway, but variation margin requirements more or less oblige then to.</ref>
+Now you might make the case, and some have,<ref>Notably Gerd Gigerenzer, who has tracked the expansion in length of the Basel accords against the persistent rate of bank failure.</ref> that capital regulation has been a bit of disaster, but one lot of crappy regulations is not a prescription for ''more'' crappy regulations. Even if, as in this case, the new regulations were also proposed by the Basel committee too.
+For this is ''exactly'' what bilateral variation margin does. Capital is the measure of “unallocated cash” available to meet the claims of general creditors. Cash being fungible, ''any'' cash on the balance sheet counts towards the capital ratio. A counterparty with an uncollateralised paper gain of $100m against a dealer still has a claim to that $100m: it can close out at any time, and even if the dealer fails first ''it still has a claim on that amount from the dealer’s capital reserves''. It is just lining up with other creditors who also have claims.
+==== Other dumb use cases ====
+And leaving aside the risk consequences of bilateral variation margin, don’t forget the ''operational hassle'' it presents. The financial system is complicated enough without gobs of cash flying to and fro just to offset changing risk positions on principal contracts. All other things being equal, it would be better not to post margin than to post it.
+If we accept for a minute dealers shouldn’t have to margin customers without good reason, we notice a class of transactions in which customers presents ''no'' risk to the system at all but, because of the bilaterality of margin regulations, they are obliged to receive, and then return, daily margin anyway. These are fully-paid option contracts. Here, the customer pays its premium up front. Anything it can be liable for, it pays at inception. Thereafter, worst case, the option expires out-of-the-money and the customer gets nothing.
+But if the option trades into the money, the dealer is obliged to pay variation margin to the customer.
+Not only is this counterproductive to the interests of systemic stability, as per the above, but it also means the customer has to get involved with margining and cash management, because it will have to post this excess margin back to the dealer if the position moves back the other way. This is operational faff for no good reason.
+===Voluntary margin===
+Now, none of this stops a dealer recognising the “equity” in a customer’s unrealised [[mark-to-market]] gains and lending against it. This is what [[Margin lending|margin lenders]] and physical prime brokers do, every day of the week. But this kind of lending is, as lending should be, ''discretionary''. Dealers don’t have to accept the new trade, and if they do, they can impose whatever [[haircut]]s, credit terms, diversification criteria and other conditions they like. A customer who diesn’t like the terms can take its business elsewhere. This, as the [[Archegos]] situation illustrated, is plenty compulsion enough.
+For, as a customer’s unrealised “profit” increases, you would expect the dealer’s lending appetite to diminish. Because unrealised profit ''is not profit''. It is the supposed attitude of the market without the single, vital signal that actual profit implies: ''sell''.
+In times of orderly function and normal liquidity, of course, that signal will be swamped by the hullabaloo of impulses flying around the market. It will be drowned out. But no-one risk manages for times of orderly function and normal liquidity. You risk manage for dislocation, sudden expected panic and market stampedes. Here the difference between your dealer’s last mark and your actual realised profit when you managed to get out can be vast. How vast? More than ten billion dollars, in [[Archegos]]’s case.
+But regulatory VM rules ''force'' dealers to lend in full, and in cash against untested marks, ''while investors’ capital is still at risk''. This is like expecting the house to pay out before the roulette wheel has landed, on an optimistic assessment of where it might. Imagine if retail banks were forced to cash collateralise mortgage customers the value of unrealised gains on their properties!
+“But, but, but, JC: there is a difference. A mortgage customer ''owns'' her house. She has no credit exposure to the bank for the house. If the bank fails, the customer keeps the house. With a swap, the dealer owns the hedge. If it fails, the customer would lose everything.”
+All this is true. But, equally, the customer’s personal capital outlay for that house — her real investment — is small. A house is a levered play. A customer might technically “own” her house, but only in the very contingent sense that she keeps up her mortgage payments. ''She only owns the house as long as the bank gets its money''.
+The same is true of an equity swap. It is, as above, an implied loan. The customer puts down initial margin — the economic equivalent of a deposit — to get a levered return on the whole asset. The dealer earns only its commission when it opens and closes the trade and, because of the implied loan, takes a funding rate from the customer. Economically, the dealer has lent 70% of the initial value of the asset. If the customer wanted to isolate its exposure to the dealer for its equity in that investment, it could take out a margin loan against a physical asset, just like a mortgage. Or it could frequently close out its trade, take its profit and restrike par and margin with the dealer.
+[[Buy-side]] counterparties are, [[Q.E.D.]], sophisticated professionals.<ref>They don’t get through onboarding if they are not. Sophistication is a condition to entry to the game.</ref> They have the tools, resources and skills to monitor their dealers’ credit standing. It is a ''much'' better discipline for them to prudently manage their own dealer credit exposure than to have dealers send them hard cash so they don’t have to. It is much more efficient. It simplifies the operational system.
+A customer’s failure shouldn’t be a systemic risk unless its unusual size, interconnectedness or other unintended system effects ''make'' it systematically important, in which case it should be regulated as if it is systemically important, and made to hold capital reserves.
+Every dollar of margin a dealer pays to one customer reduces the capital it has available for everyone else. Dealer margin is a preference to that creditor over others. Unlike the preference afforded to retail depositors — on whose confidence a bank relies for its ongoing viability —there is no good grounds for preferring leveraged buy-side professionals in this way.
+Certainly not for daily [[mark-to-market]] moves, which are are ''mainly'' “noise”. ''Signal'' emerges only over time. The back-and-forth of [[variation margin]] therefore, simply ''accommodates noise''. Downward spikes in noise can certainly lay out levered investment funds, as we have repeatedly seenOver the short run posted collateral can, as we know a system effect: if I double down on an illiquid position, it will tend to rise, and I will get more margin, and — this is the story of [[Archegos]].
+The systemic risk caused by interconnected financial institutions failing — which is what the margin regs were designed to address — is not caused by the banks themselves, but by their customer exposures. Dealer risk is a function of customer failures, which in turn are a function of leverage. Paying variation margin to customers invites more leverage.
 {{Sa}}