Streetwise Professor

January 23, 2013

Derivatives Regulation: Risk Transformation, Not Risk Reduction

One of the themes of my writing on clearing mandates and other aspects of post-Crisis derivatives regulation has been that although these initiatives have been sold as ways of reducing risk in the financial markets, in reality, they shift many risks, transform others, and create new ones.  It is by no means clear that these new laws and regulations have reduced the overall vulnerability of the financial system.

For instance, clearing and collateral mandates are intended to reduce the amount of credit embedded in derivative trades.  But market participants can respond to this by substituting other forms of leverage.  The effects on credit risk and leverage overall are far more equivocal than the advocates of the mandates claim.  As another example, making derivatives more expensive means that some will eschew using them to manage risk.  Risk that was passed to others able to bear it at lower cost will remain with those who bear it at a high cost.  As yet another example-and the one that worries me most-is that clearing and collateral mandates have transformed credit risk to liquidity risk.  Since true crises are usually liquidity crises, this is highly disturbing.

The CFO of Rolls Royce made similar arguments forcefully at a conference a few days back:

European regulatory measures designed to make the financial system safer don’t actually eliminate risk, they simply move it to nonfinancial companies, warns Mark Morris, the finance chief for Rolls-Royce, the British aero-engine maker. And there could be a negative impact on the economy, he says.

Speaking at a conference organized by The Economist Group (a minority owner of CFO), Morris criticized European regulators for seeking to deal with over-the-counter derivatives risk in the financial sector by requiring some contracts to be centrally cleared or to have collateral posted.

The European market infrastructure regulation, known as EMIR, took effect last August but requires the European Commission to enact a series of technical standards before the rules can be fully implemented. Most are expected to be finalized by mid-2013.

“In essence, you can’t destroy risk,” Morris said. “It morphs into something else. If you take market risk and go into a foreign exchange [hedging] transaction with a bank, you’re replacing it with counterparty risk. If you try to replace the counterparty risk by using central clearing or some form of posting of collateral, what you’re doing is you’re replacing counterparty risk with liquidity risk.”

Morris makes some other good points in the article, so I recommend you read the whole thing.

It seems that the Europeans are more sensitive to these concerns than was the case in the US.  The European Parliament may demand changes in EMIR. Apparently they are aware of the legislate/regulate in haste, repent at leisure problem.  Unfortunately, with Frankendodd dismissed such concerns.  Methinks that the repentance stage will come soon.

Print Friendly

10 Comments »

  1. Professor, finally this theme is getting some hard look outside your blog. :) As the article suggests on behalf of the Rolls-Royce CFO, the credit risk is being transformed to liquidity risk. But, then, the liquidity risk is going to be transformed into price risk, which in turn is going to be transformed back to credit risk this time to be borne by a handful of clearing corporations. Unlike in OTC markets, the information on the impending crisis of the clearing houses is going to be known in a very short period of time. So, the market is going to react quickly. Quicker it reacts more chaotic that reaction is going to be. There are going to be a lot of those wanting to transfer the respective risks from their books to others. But there are going to be a very few who are going to be willing to take on such risks. For a short period of time the HFTs are going to perform the function of conduit for the transfer but this game is quickly going to reach the limit of their capacity while they can benefit from it significantly. But, then, they are going to pull out as well. And then, as you put it once, in the court of public opinion these “usual suspects” are going to be held responsible for the crisis. But it is not going to matter at that point. One way or another, we are headed for series of “French Revolutions” I think. :)
    Liquidity is a funny thing. On quantitative level it manifests volumetric risk resulting from the gap between the submitted and executed orders – whether on organized exchanges or OTC. How does one hedge itself against the volumetric risk?
    We have seen this question popup in power industry in late 1990s and early 2000s. At the time weather derivatives were introduced as one remedy. They didn’t solve the problem as the nominal size of the weather derivatives market was much smaller than what the power markets needed and, besides, the correlation between the weather and volumetric demand of energy is far less than perfect.
    So, what is the remedy, really? To structure some volumetric derivatives for financial markets? Any such effort may take years to succeed, if at all, as demonstrated by the failure of VIX.
    I cannot help but to think that one selfish (non)remedy might be setting up an HFT operation targeting moderate success in ordinary times aiming at overwhelming success for one or two days in extraordinary times :)

    Comment by MJ — January 23, 2013 @ 11:38 pm

  2. There is a sort of volumetric derivative in existence, if you consider volumetric risk as a function of overall risk, that is volatility, try the VIX. The problem with HFT at a time of stress is you can’t get the orders done when and where you think they are. Peoples usual response in the past has been not to answer the phone. I suspect after the little Knight problem the HFT guys are going to be more sensitive, and just shut down.

    A lot of financial institutions are basically short straddles: the best ones run has some butterfly aspects and are short very rich straddles, but short they are: they still have exposure to volatility. This means that as v increases, their value drops. (sorry to bring Fischer Black back, Perfesser).

    The point of the Perfesser’s article is that there is a law of conservation as to risk – it cannot be destroyed just transformed. The problem is that in the mechanism that causes the transformation, a lot of efficiency can be lost, and a lot of idiot poli.. Well, idiots will think it has gone away, when it has just changed, and maybe gotten more costly to deal with on a daily basis and worse and less containable in a crisis.

    Comment by Sotos — January 24, 2013 @ 2:02 pm

  3. This analysis seems right on to me, but I wonder: What in the world are publicly held companies like Rolls-Royce doing engaging in hedging strategies? Unless they are close to bankruptcy, their shareholders do not directly benefit from reducing the volatility of cash flows to the firm. You could posit a more complex model where the managers’ personal interest in lower volatility (due to their undiversified reputation and human capital position with the firm, say) means that you can get better managers for lower compensation if you reduce firm volatility, but my gut instinct is that this is a second-order effect that would be outweighed by the additional monitoring costs of financial speculation.

    Comment by srp — January 24, 2013 @ 8:27 pm

  4. @Sotos I understand. But as I was mentioning in my note, VIX has failed to serve the purpose it was structured to. And I think it was structurally doomed to failure at inception since, as usual, it measures risk and risk-premium with the same nomenclature (furthermore, despite the claim of its authors on its treatment being model-independent, it depends of risk-neutral modeling framework at that) and second, it takes no explicit notice of trading volumes.

    I doubt HFT guys will shut down. Surely the Knight Capital incident (along with its subsequent acquisition price) was a bizarre one. But in my view other than raising the awareness of what can go wrong in high-pressure fast-pace “I want it yesterday” environment, to me it highlighted a fundamental architectural shortcoming of the current trading systems – they are disintegrated in essence and integrated only as separate gadgets somehow brought together. I think the current market environment requires the algorithmic, execution, order and portfolio [under-the-algorithm] modules to be one integrated module in the sense of being one-piece. This is not only an issue of productivity and controls, but also of an ability to treat the volumetric and price risks in an integrated manner with the understanding of the interconnected dynamics of both.

    Comment by MJ — January 25, 2013 @ 12:17 am

  5. @Sotos P.S. Besides, volatility is just part of the story – risk means more than volatility. And I am not referring to the higher moments – to the opposite, viewing them as secondary or marginal, I am referring to the lower moment.

    Comment by MJ — January 25, 2013 @ 3:03 am

  6. @srp – the reason companies hedge is that they choose the businesses they are in: for example if RR is making engines at a fixed delivery cost and schedule, they do not want to be speculating on the future raw materials market so will buy forward rather than stockpile the materials, some rare and volatile. If they do not have fixed contracts for delivery but have to plan for optional deliveries, other strategies are used. In other words hedging the price swings in their on hand inventories, etc. The point is that inventories and future demand gives them two choices: assume price and availability risks (the latter of which can shut down the entire operation), or avoid or mitigate them. Since a lot of manufacturing enterprises view themselves as makers not speculators, they hedge. This is doubly so for dealers such as lumber yards, scrap merchants, etc. where the leverage can be huge.

    @MJ Regarding my earlier comment, I realize it sounds snarky and for that I apologize. That was not my intention.

    Yes the VIX is flawed – there have been a few of those, most notably the GNMA futures that were designed by someone who so screwed up the cheapest to deliver rules that they become a proxy for super premiums and were useless as a hedge once rates started dropping from 13% or so. The issue of dealing with volume type hedges is that 1. Volume can be very hard to measure when there are multiple markets that transactions can be made on – I am not referring to multiple exchanges but types of sales – e.g. cash vs. future vs. forward, versus retained and hedged with or without basis, etc. all are kinds of sales.
    Secondly as regards to HFT, it assumes that metaphorically someone is there to answer the phone. My experience, which I freely admit is probably out of date, has been that markets can and do stop when things get truly hairy. An alternative is that quoted prices vary so widely and the sides keep disappearing that while a number will appear, it is meaningless.

    I think the point of al this FrankenDoddishness is to somehow build a structure that the ultimate risk – market failure or collapse is avoided. The point is that risk cannot be eliminated, just transferred and transformed into other venues and forms. The Perfesser’s point is that this is a futile effort in absolute terms without a Bagehot type lender of last resort, and that the form it is beginning to take will either increase the cost of hedging to such a point that risk premia in goods and services will increase due to inefficiency, or a false sense of safety will promote riskier behavior, or both.

    All of this coming from a crisis largely created by the Government itself: through the GSE’s, CRE and the legal privileges given the RA’s by statute (Basel 2, prudent (!) man regs., etc.. with a nearly 2 decade period to build up the Government induced distortions.

    The term Ironic doesn’t begin to cover it.

    Comment by Sotos — January 25, 2013 @ 9:42 am

  7. @Sotos No offense taken and no apologies needed.

    But I think the issue of volumetric measurement problems due to the “cash vs. future vs. forward,” etc. can be resolved since corresponding roll-over rates, vols, riks-premia, etc., at least in vanilla cases can be evaluated.

    Comment by MJ — January 25, 2013 @ 11:15 am

  8. @Sotos: You’ve given the standard manager’s analysis, but it is defective. The economic cost of using raw materials to make a jet engine is the spot price (opportunity cost) of those materials at the time of manufacture, regardless of whether they were purchased in advance at some other price or if there is some hedged side-bet. All that those hedging behaviors do is reduce cash-flow volatility; they do not at all affect the ability of the firm to avoid raw material price shocks to their true production cost.

    If RR finds out at the time of manufacture that the spot price of some metal is so high that they would actually lose money (including long-term effects on their reputation of a delivery interruption) by putting it into an engine, then they shouldn’t make that engine. If it still looks like a positive-NPV move to produce, but not as profitable as it would have been under the spot prices prevailing at the time RR locked in their metal price, their investors should still properly mark to market the lower profitability of the engine sale. So the hedge has no material effect on the expected profit of the firm, as long as one can neglect the possibility of bankruptcy due to the metal price swing.

    Comment by srp — January 25, 2013 @ 4:46 pm

  9. @srp

    You have given the standard economic analysis, but it is defective in practice for at least five reasons:

    1. Time lags – production takes time, time equals uncertainty, uncertainty has cost. Hedging reduces uncertainty. Therefore the issue is comparing the cost of hedging versus the cost of added volatility. Everything that follows comes from this.

    2. The profitability of The firm is a function of Revenues less costs. Revenues in any large lead time product are usually set in contracts over time. This is particularly the case for a large component supplier to an end user – if revenues are defined and fixed, costs have to be, particularly in relatively thin margin businesses. Don’t underestimate the effects of input price volatility combined with lead time considerations in running a business.

    Yes, if the firm can price at the margin, there is no need to hedge. However orders often have fixed prices or prices determined by schedules (formulas) for deliveries running over YEARS, particularly for aircraft. Yes, the true marginal cost is the current cost of all inputs. The value received for that is not necessarily correlated with the short term swings in the MC.

    3. If they cannot make money they shouldn’t make the engine with a brief nod to the “Issue” of their reputation. Easy to say, not so easy to do.

    Suppliers do not have this really have this choice. Often they are obliged by contract to make delivery. Even if they are not formally contracted, cutting off a customer whose business depends on you basically will cut you out of their business. I have seen this happen many times, particularly in heavy industry where relationships are developed over years, and can be destroyed in a month. Twice I was given a pricing spiel from a supplier to the firm I worked at based on changes of cost, quaintly phrased as “force majeure”, and had to eat it. Within 3 months we had replaced both, and when word got around, one of the firm’s business was cut in half and the other went broke. In other words, taking a cavalier attitude to delivering what you said you would at the price you said you would and on time is very dangerous.

    4. The costs associated with supply disruption. Certain hedges involving physical delivery insure that a firm has inputs needed in a timely fashion. Until you have seen a world class integrated paper mill (current replacement value of say $2mmm++, shut down fixed cost absorption needed of 12K+ an hour, 24/7, 355 (not 365) a year) come to a screeching halt because it didn’t have its deliveries of kymene and china clays delivered on time (1-2% of final cost) can one truly appreciate the risks of not nailing down the cost and AVAILABILITY of inputs where and when they are needed.

    5. There is a false dichotomy posed in your response – either exist or go bankrupt. Doesn’t work that way. Firms are judged on a gradient. If they weren’t the Rating agencies would have 2 grades: good or dead (they would love to get away with it!). This judgement is directly affected by the volatility of cash flows, relative to comparable firms and the universe of firms as a whole. Let’s look at the possible effects –

    - A firm with a more volatile cash flow has by definition more volatile margins. In periods of stress due to CF volatility this can lead to minor inconveniences like violating Loan or Indenture covenants, putting the firm at risk. Business is a stochastic process; you only need one screw up once at one given point of time to go broke, whatever the longer term viability of a non hedged strategy is vs. a costly hedge. In $2 talk, there is an asymmetry between rewards and costs – $1 does not equal one util here.

    – Such a firm is less likely to get easy trade credit terms, will have nasty things said about them in the D&B reports, be viewed as a less desirable employer, etc.

    – It may also have a harder time with suppliers when negotiating prices – a riskier customer is a less desirable one. This can lead to a significant cost disadvantage relative to more stable if slightly less profitable competitors.

    – Finally the goal of most managers is supposed to be to maximize the value of the firm, not any one set of transactions (or lining their pockets at the expense of the shareholders). A more volatile firm is often given a very steep discount to more stable firms – it all depends on the markets’ appetite for risk at any give point. This means their cost of capital is often higher, again putting them at a competitive disadvantage. Anyone remember the capital asset pricing model?

    An interesting aside demonstrating the cost of operational volatility was a paper I heard about that compared low beta investments with high beta ones: according to this the more stable firms, when their beta was increased by leverage, outperformed the equivalent un-levered high beta firms significantly. I think this was advanced as an explanation as to Warren buffet’s returns, particularly as he ran insurance firms that usually avoided inconveniences such as marks to market, and has been known to fight ferociously against any permanent impairment charges on investments like Wells Fargo. If anyone can cite the paper, I would love to hear from them an read it myself.

    In a frictionless world, you are right. Indeed, the model you describe is a very powerful one for analyzing pricing decision making. Unfortunately we exist in time, and information and transaction costs can put paid to this argument when it is applied to an imperfect world.

    Comment by Sotos — January 25, 2013 @ 10:27 pm

  10. @sotos: Thanks for your detailed response.

    1. The timing question is irrelevant to the analysis. Even when decisions are made with long lead times, the actual expected cost of the metal in the engines is going to be the expected market price for that metal at the time of production. Another way to see this is that the ranking of any two production plans’ expected profitability can never be affected by hedging the metal price, because the value of the hedge cancels out when you subtract the expected profitability of one plan from the other.

    Since the hedge does not affect the optimal production plan, it could only be of benefit if it lowered expected fixed costs. But it doesn’t do that either; rather it adds a transaction fixed cost (constant across states of the world) while adding or subtracting payoffs depending on the future realization of the metal price.

    2. The institutional details of the contracts would only matter to an equity holder in the firm–who is risk-neutral to the idiosyncratic input price swings–if a bad realization of those prices threatened firm liquidity such that his equity might be lost to creditors. So hedges that are put in to cover extremely bad realizations that threaten firm liquidity would be good for equity holders. But hedges that just smooth earnings in states of the world far from bankruptcy are bad for equity holders. Remember that the equity holders are equally interested in benefitting from favorable price swings that generate “windfall” profits. Hedging gets rid of these. (I’m assuming that the best business model is for the firm rather than the customer to bear all the risk of input price fluctuations–i.e. no escalator clauses–so that this issue arises in the first place.)

    3. Reliable delivery is very important in a large number of industrial markets. When you are planning whether to commit to those deliveries (i.e. before you have made any promises), however, the expected market input price at the time of manufacture is the correct price to use, regardless of hedging (see point 1 above–the optimal production plan under uncertainty is unaffected by the hedge).

    4. Physical access to supplies is indeed important in capital-intensive high-throughput processes. If the spot market were to become illiquid at a crucial moment, then the spot price would no longer reflect the economic cost of the input. (Interestingly, it might be that there is some other operator even more desperate than you are who would pay you such a king’s ransom for the input that he would more than compensate you for the disruption to your operation, but in such chaotic market conditions you probably would never be able to identify him and make a deal fast enough.) So physical stockpiling (or paying for contingent access to stockpiles held by others) may make sense. I’m not sure how derivatives wouldn’t be subject to enough counterparty risk under such market-collapse conditions that you wouldn’t just rather buy and hold the stuff physically, but a small amount of hedging for these cases might make sense.

    5. Let me start with your last point, about equity holders and volatility. Beta is systematic (economy-wide) risk, not the risk of a narrow-sector input price fluctuation, which is a quintessential unsystematic risk. The latter is easily avoided by diversified equity investors, so they are not rewarded for holding it and they are naturally not excited by managers who hedge against it. A Rolls-Royce investor doesn’t really care about the volatility of vanadium prices because they also own companies that are unaffected by or even benefit from higher vanadium prices. Their wealth is almost independent of relative input prices.

    Your earlier points here are about the gradient between “risk-free” and “bankruptcy risk.” Point 2 above deals with the issue in general terms, but does not address this continuum explicitly. You are 100% correct that there is a continuum here; the empirical question is how steep the knee of the cost-of-capital curve is (also proxying supplier willingness to extend trade credit) as the firm moves away from zero risk. I’m sure there are rafts of papers that link firm cost of capital to rating agency scores and so on, but I am not expert in that literature. My intuition is that there are regime effects, in the sense that in the years before the recent financial crisis, until you get pretty close to not being able to pay people back your cost of capital didn’t go up very much. Today, the risk-sensitivity of creditors is higher (and there are regulatory effects, some pushing in the opposite direction).

    But that is by no means a carte blanche for hedging to reduce cash-flow volatility. You have to be in the region of the risk curve where it matters, you have to show that hedging is superior to holding more cash or engaging in less debt, etc. Basically, this just reverts to the vast Modigliani-Miller literature on when the capital structure should be relevant to the valuation of the firm. By and large the empirical data I’ve seen say “not much,” but in a specific case your mileage may vary.

    Thanks for a thought-provoking exchange.

    Comment by srp — January 26, 2013 @ 6:56 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress