The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Federal Reserve Bank of Chicago Interest Rates or Haircuts? Prices Versus Quantities in the Market for Collateralized Risky Loans Robert Barsky, Theodore Bogusz, and Matthew Easton November 2016 WP 2016-19 Interest Rates or Haircuts? Prices Versus Quantities in the Market for Collateralized Risky Loans∗ Robert Barsky, Theodore Bogusz, Matthew Easton November 29, 2016 Markets for risky loans clear on two dimensions - an interest rate (or equivalently a spread above the riskless rate) and a specification of the amount of collateral per dollar of lending. The latter is summarized by the margin or ”haircut” associated with the loan. Some key models of endogenous collateral constraints imply that the primary equilibrating force will be in the form of haircuts rather than movements in interest rate spreads. Indeed, an important benchmark model, derived in a two-state world, implies that haircuts will adjust to render all lending riskless, and that a loss of risk capital on the part of borrowers has profound effects on asset prices. Quantitative analysis of a model of collateral equilibrium with a continuum of states turns these results on their heads. The bulk of the response to lenders’ perception of increased default risk is in the form of higher default premia. Further, with high initial leverage, reductions in risk capital decrease equilibrium margins almost proportionately, while asset prices barely move. To the extent that one believes that it is a stylized fact that haircuts move more than spreads - as seen, for example, in bilateral repo data from 2007-2008 - this reversal is disturbing. Keywords: leverage cycle, margins, financial crises, repo, risk, collateral, belief disagreements JEL Classification: D53, E44, G00, G01 ∗ Barsky: Senior Economist and Research Advisor, Federal Reserve Bank of Chicago. Bogusz: PhD Student, University of Michigan. Easton: Associate Economist, Federal Reserve Bank of Chicago. The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Chicago or the Federal Reserve System. The authors thank Gadi Barlevy, Marco Bassetto, Ben Chabot, Mariacristina De Nardi, Ana Fostel, David Marshall, Anna Paulson, and Alp Simsek for valuable conversations at various of stages, and Stefan Nagel for comments on a very early draft. Markets for risky loans clear on two dimensions - an interest rate (or equivalently a spread above the riskless rate) and a specification of the amount of collateral per dollar of lending. The latter is summarized by the margin or ”haircut” associated with the loan. Geanakoplos (2012) stresses the strong association of major booms and busts in financial markets with substantial movements in haircuts, a phenomenon he calls the ”leverage cycle”. Interest spreads, on the other hand, show far more modest time-variation. Put differently, financial crises seem to be first and foremost periods in which the quantity of leverage falls, and only secondarily periods in which the price of leverage is high. The primacy of haircuts rather than spreads as the equilibrating mechanism shows up clearly in the Gorton-Metrick (2012) data from bilateral repo markets during the dramatic 20072008 episode. Figure I displays eight panels, each of which corresponds to a class of relatively risky collateral assets. For each asset class the figure shows data on average haircuts and repo spreads from three periods: i) the pre-crisis first half of 2007; ii) the second half of 2007, which might be thought of as the period of the relatively contained ”subprime crisis”; and iii) 2008 as a whole, the period of the general financial crisis centered on the shadow banking system. Across all eight asset classes, bilateral repo in the first half of 2007 appears nearly riskless, with both haircuts and spreads close to zero. In the second period, there are modest increases in the repo spread, and larger (in some cases an order of magnitude so) but still not dramatic increases in haircuts. Finally, the third period shows truly striking spikes in harcuts, as high as 60% for some private label assets (see Gorton and Metrick for details). While the interest spread also rises sharply in period 3, this increase is easily an order of magnitude less than the rise in haircuts, with the highest repo spreads on the order of 200 basis points. Table 1, which shows descriptive statistics, pooling the time series and cross section data, indicates that whether one focuses on means, variances (across time and across asset classes), or higher moments, variation in haircuts trumps variation in spreads in every respect. While Krishnamurthy et al (2014) cast doubt on the notion that the same dramatic increase in haircuts characterized the larger triparty repo market, one might just as readily interpret the complete disappearance of repo for many kinds of collateral in the triparty market as equivalent to a haircut 1 of 100 percent. At a broad brush level, the tendency for financial market scares to manifest themselves in a sharp drop in the equilibrium quantity of lending against risky assets without a commensurately sharp increase in the price of loans seems to go far beyond repo. For instance a key feature of the 2007-2008 crisis was the collapse of the large market for asset-backed commercial paper (Krishnamurthy, et al). Why did the quantity of securitized lending fall so dramatically, instead of contracting more modestly with a greater share of the adjustment to increased default risk occurring through spikes in interest rates on the risky loans? There are a number of reason to regard the ”spreads vs. haircuts” (or more generally, the ”prices vs. quantities” in financial markets) question as one of first order importance. If financial market scares manifest themselves in increased haircuts without large increases in risky interest rates, the monitoring of spreads alone would provide insufficient warning of financial stress. Likewise, the policy implications for issues such as the lender of last resort function of central banks might well depend on the relative importance of spreads vs. haircuts as equilibrating mechanisms, especially if increased margins are not merely an equilbrium an equilibrium outcome but a reflection of a significant degree of ”credit rationing” (Geanakoplos, 2010; Fostel and Geanakoplos, 2013). Second, margins and spreads are key (probably the key) statistics on which to evaluate the empirical relevance of models - indeed, entire classes of models - of collateralized risky lending. Finally, the question of how markets clear, particularly when non-price in addition to price mechanisms are at work, is at the very core of economists’ underlying intellectual agenda. Fostel and Geanakoplos (2015) provide an elegant theory of a collaterized loan market with heterogeneous beliefs in which haircuts are always sufficient to preclude equilibrium default and all lending is likely to occur at the riskless rate. Their version of the theory of collateral equilibrium is constructed in the context of ”binomonial economies” in which there are only two continuation states. Simsek (2013) studies an otherwise nearly identical model in which there is a continuum of states and finds that the equilibrium features default in some states of the world, that collateralized loans consequently trade at spreads above the riskless interest rate, and that the variation in those spreads depends in interesting ways on both the downside risk and uncertainty perceived by lenders 2 and the upside opportunities perceived by optimistic borrowers. What Simsek’s results do not tell us directly is whether the predictions of the theory of collateral equilibrium in the binomial case might nevertheless be a reasonable approximation to the truth. Should one expect dramatic gyrations in default premia in response to shocks such as increased fear on the part of lenders or a loss of risk capital on the part of borrowers? This paper offers quantitative theory aimed at answering that question. One might thus view our work as an examination of Fostel-Geanakoplos (2015) through the lens of Simsek (2013). The plan of the paper is as follows. In Section II-a, we present a streamlined derivation of the Simsek model, with special attention to the equilibrium haircuts and interest rates on which Simsek did not explicitly focus, while referring the reader to Simsek (2013) for proofs of existence and uniqueness, etc. In Section II-B, we discuss the somewhat degenerate case of the Simsek model in which there are only two discrete states. We show that in the two-state case the existence of any risky borrowing implies the asset must sell at the pessimist’s valuation. We also show that there is another equilibrium at the pessimist’s price that has lower leverage and zero default risk, with ex post consumption allocations for both agents that are the same as those that would be achieved with equilibrium risky borrowing. Thus the equilibrium with zero-value-at-risk haircut and no interest rate spread is ”essentially unique” in the sense of Fostel and Geanakoplos (2015), and the result is effectively a weak Fostel-Geanakoplos nondefault theorem for the Simsek model (the stronger version, which Fostel-Geanakoplos obtain in the binomial economy when there is a continuum of agents, rules out the equilibrium trading of any risky loan contracts.) Though hardly surprising, the FG result for the Simsek model is a critical benchmark because it verifies that we are not dealing with two fundamentally incompatible enviroments - on the contrary, Simsek (2013) is a natural generalization of Fostel-Geanakoplos (2015). Section III, which contains the main results of the paper, consists of quantitative theory, studying the behavior of the Simsek model in a laboratory-type setting with specific belief distributions and endowments carefully chosen to lay bare the fundamental mechanisms and to stress-test the model’s predictions for margins and spreads. Representing beliefs by means of the simple triangular family of distributions, we compute equilibrium loan size, interest rates, margins, and the 3 price of the risky asset, and present diagramatic representations of the underlying workings of the loan market equilibria that determine them. Not surprisingly, lenders’ fear of increased defaults raises both interest rates and haircuts, and (as Simsek’s logic already lead us to expect) the extent of the rise in the interest rate is greater, and the fall in leverage (rise in the endogenous haircut) smaller when the optimist’s belief distribution is skewed towards positive events. However, nothing prepared us for the sharp variation in interest rates, and the extremely limited variation in haircuts, that the model generates in response to belief shocks. The quantitative predictions of Simsek in response to a ”fear shock” are far from those of Fostel-Geanakoplos and far from what is seen in the data. Surely, however, there are disturbances other than shocks to beliefs. Might these not help us better match the data? Most observers of the 2007-2008 crisis, not least Geanakoplos (2010), consider the loss of much of the wealth that constituted the optimistic borrowers’ risk capital as the other most salient characteristic of the downward ”loss-haircut” spiral. The stylized analogue of the loss of risk capital in the model is an adverse endowment shock. In the model, such a shock is anything but helpful. The wealth loss increases the interest rate and reduces the haircut more or less proportionately - i.e. a halving in risk capital leads to an approximate halving of the haircut (i.e. a doubling of leverage at the height of financial crisis!) The underlying reason for this is the fundamentally inessential nature of risk capital in the Simsek model in the first place. The optimist’s endowment serves to reduce his desired borrowing, thereby reducing the likelihood of default and taking some pressure off of the interest rate. However, because the default premium does the heavy lifting, there is no need for loans to be overcollateralized in this model. Ineed, there is no analytical reason that loan to value ratios cannot exceed 100 percent (as they in fact sometimes did in nonconfirming mortgage markets). Simsek, the natural extension of FostelGeanakoplos beyond the binomial case, in a sense turns the FG results on their head. Instead of the simple but powerful expression anchoring leverage to the borrower’s risk capital and the lender’s worst conceivable realization of the continuation value of the collateral, we have an ”endogeneous haircut” in which risk capital plays primarily a mechanical role. The fundamental economics is about the optimist’s trading off his perceived wedge between the market price and the true value 4 of the collateral, on the one hand, and the pessimist’s demand for an ”excessive” default premium on the other. Section IV discusses the implications of the counterfactual predictions of the Simsek model for the leverage cycle and for financial theory in general. We suspect that many of the fascinating theoretical results about the leverage cycle that Geanakoplos and his coauthors have generated are largely independent of their foundations in collateral equilibrium, but should perhaps be constructed on alternative principles, probably incorporating counterparty risk on top of uncertainty about the collateral alone. II. Interest Rates and Haircuts in the Simsek Model a) The Simsek Model With General Continuous Belief Distributions Except where noted explicitly we follow closely the derivations in Simsek (2013). The model has two dates {0,1} and two types of risk neutral agents {o, p}, denoting optimists and pessimists, respectively. There is a continuum of each type of agent. There are two assets - a risky asset that we will call a tree, and a consumption good that we will call fruit. Fruit can be stored at a constant real return of 0, so that it functions as a riskless asset; we will sometimes call it ”cash”. Agents receive their endowments of the assets at date 0 but consume at date 1 only. Unlike Simsek, who has a third set of agents that are endowed with the risky asset and the sole sellers of it, we assign the initial endowments of trees (normalized to unity in the aggregate) to the pessimists, who are also endowed with cash. Without loss of generality, we endow the optimists with cash only. The important thing is that they use that fruit (which we intepret as their risk capital), along with fruit borrowed from pessimists, to buy trees. Pessimists are both lenders of fruit and sellers of trees. They cannot, however, bundle these activities - i.e. provide the buyers of their trees with cash financing. The two activities occur in separate competive securities markets.1 1 Simsek makes two assumptions concerning the size of the cash endowments of the two groups that guarantee that the optimists will hold all the trees in equilibrium and that this can be accomplished only with at least some risky borrowing. These together ensure that the set of possible equilibrium prices for the risky trees will lie strictly 5 Next we turn to the characterization of the beliefs of the two kinds of agents. The optimist first and foremest believes that the expected payoff from the trees is higher than does the pessimist. Sufficient conditions for the existence and uniqueness of the solution to the prinicipal agent problem described immediately below also require the relative optimism of the optimist to be increasing in the state s. This can be expressed either in terms of increasing vertical distance between the two agents’ inverse CDFs as s gets large, or equivalently in the non-crossing of the hazard conditions. The intuition can be understand either through the proofs contained in Simsek’s appendix or via our quantitative examples in which the hazard function and the inverse CDF are exhibited graphically in each case. The work of Geanakoplos and his coauthors on ”collateral equilibrium” is based on the application of competitive general equilibrium theory to commodities consisting of contracts - in the case of the ”simple debt contracts” which are the sole contracts in the current paper, representable by ordered pairs consisting of an interest rate and an amount borrowed per unit collateral. Except in the limiting case where the price of trees falls to the pessimists’ level , pessimists confine their period zero activities to storage and lending. If the contractual interest rate on a loan of size b collateralized by one tree is r, the payment received by the lender in state s is min[s, φ], where φ = (1 + r)b. In state s, the lender either receives his full repayment φ or the salvage value s, whichever is less. Thus under risk neutrality, arbitrage between storing and lending establishes the size of the loan collateralizable with one tree as Ep [min(s, φ)]. Simsek reformulates the determination of collateral equilbrium as a principal-agent problem, albeit with one key additional step. The optimization problem faced by optimists can be written as: between the optimist’s full price (the maximum he would pay for the asset, which is the integral of the possible payoffs weighted by his perceived probabilities and discounted at the riskless rate) and the price at which the pessimist’s short sale constraint ceases to bind and he is marginally willing to hold the trees. Although this will hold in almost all of our examples, we make neither assumption a priorri, as we find the limiting cases in which a) the optimist can purchases the entire supply of trees with riskless borrowing only and, more importantly b) the opposite case in which the price must fall to the level at which the pessimist is willing to go long in trees to be of some interest (see in particular Section II. b, where we discuss the version of the Fostel-Geanakoplos nondefault result that holds in our model). 6 max (a1 ,φ)∈R2+ ao Eo [s] − ao Eo [min(s, φ)] (1) s.t. ao p = no + ao Ep [min(s, φ)] (2) In words, the optimist chooses an amount ao of risky trees to purchase with collaterized loans (in addition to his endowment of fruit) by maximizing his expected payoff from the trees net of his expected debt repayment, subject to a budget constraint that takes the price of trees as given and the interest rate as increasing in borrowing per tree in order to satisfy the lender’s ”participation constraint” b = Ep [min(s, φ)]. How do we know that the participation constraint holds with equality? Perhaps the pessimist instead receives some surplus Ep [min(s, φ)] − b. The answer (which happens to represent that key step mentioned above) is that the expected return on collateralized lending must be precisely zero (or more generally, the return on storage). Importantly, because he perceives less lower tail risk, the optimist believes that his expected loan repayment is greater than the pessimist’s expectation; Eo [min(s, φ)] > Ep [min(s, φ)]. As we will soon see, the optimist’s belief that the system is rigged against him and that he must pay an excessive default premium will play a crucial role in discouraging the optimistic from borrowing as much as he otherwise might, and in depressing the price of the risky asset. Simsek proves that, under the above assumptions, the solution to the principal-agent problem and a characterization of the associated collateral equilibrium is characterized by the following equation, which implicitly determines a bankruptcy threshhold or ”loan riskiness” s¯: p = popt (¯ s) ≡ Z s¯ Z sdFp + (1 − Fp (¯ s)) smin smax s s¯ dFO 1 − FO (¯ s) Following Simsek we will call the first equation, the optimality condition. There is also an equilibrium condition: p = n1 + Ep [min(s, s¯)] (3) which we will refer to as the market-clearing condition. This says that the price of a tree is just 7 covered by the risk capital of the optimist plus the maximal loan per tree that the pessimist is willing to provide. Note that the pessimist’s participation constraint appears both in the decision of the optimist represented by the optimality curve, and in the market clearing condition, which combines the otpimist’s budget constraint and the pessimists’ participation constraint. In the next section of the paper we will plot the two curves and find the equilibrium at their intersection, solved numerically for exogenously determined values of the endowments and subjective probability distributions. A new way of looking at this price equation emerges from the following thought experiment. The first term represents the value of a security entitling its owner to the salvage value of a tree in bankruptcy. Suppose the lender cedes this security to the borrower in exchange for a higher ”zero recovery” interest rate. Then the relevant disagreement is fully characterized by the R smax dFO pessimist’s higher default probability. The second additive piece, (1 − Fp (¯ s)) s¯ s 1−F s) , repreO (¯ sents optimists’ valuation of the fruit he keeps in nonbankruptcy states, taking into account that it is on the margin purchased with money borrowed from someone with whom there is a fundamental disagreement about bankruptcy risk. If there were no default disagreement, given the risk neutrality of both agents the optimist would discount at the riskless rate of zero, and that upside piece R smax would be simply s¯ sdFO. Instead the optimist sees the appropriate breakeven interest rate as (1 + ropt ) = [1 − Fo (¯ s)]−1 , while the pessimist requires the higher rate (1 + r) = [1 − Fp (¯ s)]−1 ; thus s)] 0 (¯ the optimist discounts his winnings in the nondefault state by 1−F 1−Fp (¯ s) > 1, in Simsek’s terminology the optimist’s perceived [gross] interest rate. From the point of view of the optimist it might be thought of as an ”unfairness” measure. In all of our numerical exercises we will present the statistic 0≤ 1−Fp (¯ s)] 1−Fo (¯ s) ≤ 1,which we refer to as the ”agreement ratio”. When this ratio is unity, both parties are confident that loans will be paid in full, while a ratio of zero implies that the pessimist (but not the optimist) is certain that default will occur. Finally, Simsek rewrites the optimality condition in terms of conditional expectations and provides a nice compact ”price equation”: p = Fp (¯ s)Ep [s|s < s¯] + (1 − Fp (¯ s))Eo [s|s > s¯]. 8 This ”price equation” is however not a reduced form solution for the asset price. Although our focus is shifted towards the equilibrium price of the tree rather than the choice of optimal loan size s¯ for a parametrically given asset price, it is mathematically equivalent to the optimality equation as s¯ is still an endogenous variable, pinned down only by coupling the optimality condition with the equilibrium condition. This has an important implication. A key theme in Simsek is the distinction between upside and downside disagreement. Simsek is normally quite clear that whether or not the disagreement is ”downside” depends on whether or not it concerns default states, - i.e. states for which s < s¯. However, it can be easy to forget that upside versus downside, though it has much to do with the shapes of the belief distributions, cannot be inferred from looking at the agents’ pdfs alone. The crucial dividing line is the equilibrium loan riskiness s¯, indicated in the quantitative exercises of the next section by the dotted vertical line, which depends on endowments in addition to beliefs. b) Two Discrete States: A Fostel-Geanakoplos Result for the Simsek Model Suppose, as do Fostel and Geanakoplos, that there are just two discrete states, H and L. The optimist believes state H will occur with probability πH,o and the pessimist believes that it occurs with probability πH,p . Substituting into the Simsek pricing formula eq. 3, we have (1 − πH,p )L + πH,p (πH,o H) = (1 − πH,p )L + πH,p H πH,o (4) The r.h.s is precisely the formula for pessimistic valuation. Thus, if there is any risky borrowing, the asset must sell at the pessimist’s price. The pessimist’s price can be supported without any risky borrowing because pessimists are now indifferent between purchasing the asset from other pessimists and storing fruit. Thus in the two state model there is always an equilibrium that does not feature default, just as in Fostel-Geanakoplos. In the Simsek two agent model, there is non-uniqueness in the sense that an equilibrium at the pessimist’s price can also be achieved with risky borrowing – e.g. optimists can hold all of the asset by borrowing at a risky rate just high enough to discount their expected payoff 9 to the pessimistic valuation. However, just as in F-G, the non-uniqueness is “inessential” in the sense that not only the asset price, but the consumption allocations are invariant to whether or not the risky debt contract is actively in use. This is well illustrated by the following example. Suppose the two states are H = 1 and L = .5. The optimist attaches probability .75 to H, while the pessimist believes there is a 75% chance of state L. Let pessimists be endowed (in the aggregate) with one tree and no fruits, and optimists be endowed with no trees but with risk capital in the form of .1 fruit. We verify that there is an equilibrium in which all of the risky tree is held by the optimist at the pessimist’s fundamental valuation of .625 (.25 × 1 + .75 × .5). In this equilibrium the optimist borrows .525 at a risky interest rate of 14.29% - which gives the pessimist his required expected return of zero, since .25 × (1.1429 × .525) + .75 × .5 = .525. In this equilibrium the optimist consumes (1 − 1.143 × .525) ∼ = .4 in state H and zero in the default state L, while the pessimist consumes 1.143 × .525 ∼ = .6 in state H and .5 in state L. Alternatively, suppose the optimists borrow only .4, which they will be able to pay back with certainty, at an interest rate of zero. Adding this borrowing to their risk capital of .1, optimists will now hold only 80% (.5/.6265) of the trees, but at the rock-bottom price of .6265 the pessimist is willing to take up the slack and hold the remaining 20%. The optimist consumes .4 (proceeds of .8 from his tree minus his debt of .4) in state H and 0 in state L, while the pessimist consumes .6 (proceeds of .2 from equity in the tree plus debt repayment of .4) in state H and .5 (.1 from equity in the tree plus the debt repayment of .4) in state L. Thus we have established and illustrated a weak Fostel-Geanakoplos result for the two state case of the Simsek model. It is not impossible that risky debt is traded in equilibrium, but if it is, the equilbrium is ”esssentially equivalent” to one in which only riskless loans are traded. Either way, the price is at the pessimist’s fundamental valuation, and the consumption allocations for each of the two agents in each of the two states are the same whether or not risky borrowing is observed. Fostel and Geanakoplos derive a stronger result that flatly rules out the equilbrium trading of risky debt contracts in a two-state model in which agents lie along a continuum from most optimistic to least optimistic. But even a case like ours when the equilibrium is only ”essentially unique” there are at least two reasons to focus, in the two state case, on the equilibrium without risky 10 debt. First, as F-G note, the use of collateral is costly; collateral must be evaluated, contractual arrangements clearly specified, etc. Second, if the required risk spread drives the asset price down to the pessimist’s valuation, the reason d’etre for repo and other collaterized lending disappears. In a sense, this situation represents a sort of joint collapse of the market for the risky asset and the loans collateralized by it. III. Quantitative Theory We represent the beliefs of the two kinds of agents by means of the 3-parameter triangular family of distributions. We take the view that the agents have little precise knowledge, but have a sense of the worst and best case scenarios (a and b, respectively, with b strictly greater than a), as well as a ”most likely” scenario c. Our benchmark case is a ”right triangular distribution” with b = c = 1, so that the mostly likely outcome coincides with the best case scenario - which we take to be ”business as usual” with minimal defaults on the underlying mortgages or accounts receivable which constitute the cash flows beyond the security. With a satisfying 0 ≤ a < b, 1 − a has an interpretation as the agent’s maximum conceivable loss. We will be particualrly interested in sharp reductions in the pessimist’s a, what we call a ”scare” or a ”fear shock”. When we depart from the right triangular case and allow (for the optimist in particular) b > c, the interpretation is that the optimist may have picked up assets at fire sale prices during a previous scare and is hoping to experience a capital gain as the scare dies down. 0 2(x−a) (b−a)(c−a) 2 The density is given by: f (x) = b−a 2(b−x) (b−a)(b−c) 0 for x < a for a ≤ x < c for x = c for c < x ≤ b for b < x Figure II shows the belief distributions of the two agents, and the determination of equilibrium loan size, interest rate, and asset price. The optimist is confident that the asset will not fall in value by more than a couple of percent. The pessimist, on the other hand, believes the collateral 11 could lose as much as a third of its value. Neither agent sees any upside potential at this point. Note the key role of s¯. One this equilibrium value is determined in the diagram at the top right of Figure II (which was first constructed in Simsek’s paper to determine the equilibrium asset price), we can easily find the equilibrium loan size, interest rate, and the ”endogenous haircut”. Figure III contains our first comparative statics exercise, a ”fear shock”. The pessimist, who initially believes the collateral can lose at most a third of its value, comes to believe that the potential loss may be as large as fifty percent. Note the key result that the interest rate more than doubles, from 2.2% to 5.7% (more than twice the maximum spread seen in the Gorton-Metrick data for 2008), while the haircut barely budges. Figure IV repeats the exercise, with one crucial difference. The optimist is now depicted (throughout) as perceiving a substantial upside - he believes he may have purchased the collateral at a fire sale price and that it might rise in value by as much as 20%. The initial interest rate is higher, and the post-scare interest rate has risen more (in terms of basis points, though not in percentage terms) than in the absence of optimists’ perceived upside potential. A comparison of Figures III and IV (panel by panel, horizontally) can also be interpreted as revealing the effect of an ”excitement shock” - the optimist perceives an increase in upside potential holding constant his down side and that of the pessimist as well. Figure V illustrates the effect of a loss of the optimistic borrower’s risk capital. The optimist’s endowment falls from 0.1 to 0.05, while beliefs are unchanged. When initial risk capital is small relative to borrowing, the level of borrowing is nearly unchanged by this wealth shock, and thus a halving of risk capital also nearly halves the haircut. As stressed in our introduction, this fact strikes us as quite telling, both about the workings of the model and about its essential unrealism. IV. Concluding Remarks We found that the quantitative predictions of Simsek’s constinuous state model of collateral equilibrium, which constitutes a natural generalization of the Fostel-Geanakoplos two-state model, are wide of the mark relative to the Gorton-Metrick data as well more generally recognized stylized facts, according to which spreads move considerably less than margins. In particular, the predictions 12 for a fear shock, and even more strikingly a loss of risk capital are highly counterfactual. The present paper should be interpreted neither as an assault on the empirical facts underlying the leverage cycle notion nor on the rich theoretical implications derived in the fascinating series of papers by Geanakoplos and his coauthors, most of which we suspect are not dependent on the precise way in which the determination of leverage is modeled. Similarly, our results do not challenge the analytical value of the results in Simsek (2013), which we continue to regard as the appropriate way to further study the prices vs. quantitites properties of collateral equilibrium problems. Rather, our paper might be characterized in part as an indication for resting the leverage cycle on a somewhat different foundation than collateral equilibrium, in which uncertainty and disagreement focus exclusively o the properties of hte colateral itself. Perhaps collateral equilibrium should be replaced or modified to account for the importance ascribed to counterparty risk in institutionally motivated studies of markets for collateralized loans. References Copeland, Adam, Martin, Antoine, and Walker, Michael, 2014 ”Repo Runs: Evidence from the Tri-Party Repo Market.” New York Fed Staff Report no.506. Fostel, Ana and Geanakoplos, John, 2014, “Endogenous Collateral Constraints and The Leverage Cycle.”Annual Review of Economics (May). Fostel, Ana and Geanakoplos, John, 2015 ”Leverage and Default in Binomial Economies: A Complete Characterization.” Econometrica (December), 83(6): 2191-2229. Geanakoplos, John, 2010. ”The Leverage Cycle.” NBER Macroeconomics Annual, 24(1), p1-66. Gorton, Gary & Metrick, Andrew, 2012. ”Securitized banking and the run on repo.” Journal of Financial Economics, Elsevier, vol. 104(3), pages 425-451. Krishnamurthy, A., Nagel, S. and Orlov, D. 2014. ”Sizing Up Repo.” The Journal of Finance, 69: 2381–2417. Simsek, Alp. 2013. ”Belief Disagreements and Collateral Constraints.” Econometrica, 81(1) p1-53. 13 Table I: Descriptive Statistics for Haircut and Spread, Pooled Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis HAIRCUT REPO_SPREAD 0.128583 0.038000 0.680000 0.000000 0.196709 1.767162 4.910127 0.010502 0.008524 0.026839 0.000444 0.009027 0.373831 1.751665 Source: Gorton and Metrick (2012) Figure I: Mean Repo Spread and Haircut for Eight Asset Classes in Bilateral Repo Market: First Half of 2007, Second Half of 2007, All of 2008 1 2 3 .10 .20 .20 .08 .16 .16 .06 .12 .12 .04 .08 .08 .02 .04 .04 .00 .00 2007m1 2007m7 2008m1 .00 2007m1 4 2007m7 2008m1 2007m1 5 .7 2007m7 2008m1 6 .20 .6 .6 .5 .16 .5 .4 .4 .12 .3 .08 .3 .2 .2 .04 .1 .1 .0 .00 2007m1 2007m7 2008m1 .0 2007m1 7 2007m7 2008m1 8 .6 .28 .5 .24 .20 .4 .16 .3 .12 .2 .08 .1 .04 .0 .00 2007m1 2007m7 2008m1 2007m1 2007m7 REPO_SPREAD 2008m1 HAIRCUT Source: Gorton and Metrick (2012) 2007m1 2007m7 2008m1 Figure II: Belief Distributions and Determination of Equilibrium Determination of Equilibrium Price PDF 1 70 Optimist Mean: 0.99 Optimist Std Deviation: 0.0070711 Optimist skew: -0.56569 Pessimist Mean: 0.88867 Pessimist Std Deviation: 0.078725 Pessimist skew: -0.56569 Optimist Pessimist Eqlbm sBar 60 0.98 Optimality Market Clearin Eqlbm sBar Eqlbm Price 50 0.96 Optimist Endowment: 0.1 Borrowing: 0.82989 sBar: 0.84786 Price: 0.92989 Opt ExpValue > sBar: 0.99 Salvage: 0.23339 Price 40 0.94 30 0.92 20 Opt Prob. No Default: 1 Pess Prob. No Default: 0.70353 Leverage: 9.2989 Margin: 0.10754 Equilibrium Agreement Ratio: 0.70353 Risky Interest Rate: 2.1656% 0.9 10 0 0 Inverse CDF 1 0.5 1 Loan Riskiness sBar Hazard Rates Determination of Equilibrium Interest Rate 1 1.5 5 Optimist Pessimist Eqlbm sBar 0.8 4 3.5 0.6 3 0.5 2.5 0.4 2 0.3 1.5 0.2 1 0.1 0.5 0 Optimist Pessimist 4.5 0.7 0.5 1 s 1.5 0 0 1.5 0.2 0.15 Interest Rate 0.9 0 0.88 s 0.5 Interest Rate (Feasible) Interest Rate (Truncated) Eqlbm sBar Eqlbm Interest Rate 0.1 0.05 0 0 0.2 0.4 0.6 s 0.8 1 -0.05 0.4 0.6 0.8 Loan Riskiness sBar 1 Pre-Shock Beliefs Optimist Mean: 0.99333 Optimist Std Deviation: 0.004714 Pessimist Mean: 0.88867 Pessimist Std Deviation: 0.078725 Optimist Endowment: 0.1 Borrowing: 0.83129 sBar: 0.84986 Price: 0.93129 Opt ExpValue > sBar: 0.99333 Salvage: 0.23895 Opt Prob. No Default: 1 Pess Prob. No Default: 0.69698 Leverage: 9.3129 Margin: 0.10738 Equilibrium Agreement Ratio: 0.69698 Risky Interest Rate: 2.234% Figure III: A Fear Shock PDF 1: Initial Beliefs 100 90 80 70 60 50 Optimist Pessimist Eqlbm sBar 40 30 20 10 0 0 0.5 1 1.5 s Post-Shock Beliefs PDF 2: Post-Shock Beliefs Optimist Mean: 0.99333 Optimist Std Deviation: 0.004714 Pessimist Mean: 0.83333 Pessimist Std Deviation: 0.11785 100 90 80 Optimist Endowment: 0.1 Borrowing: 0.77764 sBar: 0.82227 Price: 0.87764 Opt ExpValue > sBar: 0.99333 Salvage: 0.29696 Opt Prob. No Default: 1 Pess Prob. No Default: 0.58458 Leverage: 8.7764 Margin: 0.11394 Equilibrium Agreement Ratio: 0.58458 Risky Interest Rate: 5.738% 70 60 50 Optimist Pessimist Eqlbm sBar 40 30 20 10 0 0 0.5 1 s 1.5 Pre-Shock Beliefs Optimist Mean: 1.06 Optimist Std Deviation: 0.049666 Pessimist Mean: 0.88867 Pessimist Std Deviation: 0.078725 Optimist Endowment: 0.1 Borrowing: 0.85351 sBar: 0.88481 Price: 0.95351 Opt ExpValue > sBar: 1.06 Salvage: 0.34844 Opt Prob. No Default: 1 Pess Prob. No Default: 0.57082 Leverage: 9.5351 Margin: 0.10488 Equilibrium Agreement Ratio: 0.57082 Risky Interest Rate: 3.668% Figure IV A Fear Shock in Presence of High Upside for Optimist PDF 1: Initial Beliefs 10 Optimist Pessimist Eqlbm sBar 9 8 7 6 5 4 3 2 1 0 0 0.5 1 1.5 s Post-Shock Beliefs PDF 2: Post-Shock Beliefs Optimist Mean: 1.0767 Optimist Std Deviation: 0.061418 Pessimist Mean: 0.83333 Pessimist Std Deviation: 0.11785 8 Optimist Endowment: 0.1 Borrowing: 0.79968 sBar: 0.86396 Price: 0.89968 Opt ExpValue > sBar: 1.0767 Salvage: 0.3935 6 Opt Prob. No Default: 1 Pess Prob. No Default: 0.47013 Leverage: 8.9968 Margin: 0.11115 Equilibrium Agreement Ratio: 0.47013 Risky Interest Rate: 8.039% Optimist Pessimist Eqlbm sBar 7 5 4 3 2 1 0 0 0.5 1 s 1.5 Figure V: Loss of Risk Capital Pre-Shock: Risk Capital = 0.1 Optimist Mean: 0.98333 Optimist Std Deviation: 0.011785 Pessimist Mean: 0.83333 Pessimist Std Deviation: 0.11785 Optimist Endowment: 0.1 Borrowing: 0.77422 sBar: 0.81649 Price: 0.87422 Opt ExpValue > sBar: 0.9897 Salvage: 0.28486 Opt Prob. No Default: 1 Pess Prob. No Default: 0.59935 Leverage: 8.7422 Margin: 0.11439 Equilibrium Agreement Ratio: 0.59935 Risky Interest Rate: 5.459% PDF 1: Initial Beliefs 40 35 30 Optimist Pessimist Eqlbm sBar 25 20 15 10 5 0 0 0.5 1 1.5 s sBar Post-Shock: Risk Capital = 0.05 Optimist Mean: 0.98333 Optimist Std Deviation: 0.011785 Pessimist Mean: 0.83333 Pessimist Std Deviation: 0.11785 40 Optimist Endowment: 0.05 Borrowing: 0.80319 sBar: 0.87162 Price: 0.85319 Opt ExpValue > sBar: 0.98333 Salvage: 0.41307 30 Opt Prob. No Default: 0.99988 Pess Prob. No Default: 0.44757 Leverage: 17.0639 Margin: 0.058603 Equilibrium Agreement Ratio: 0.44759 Risky Interest Rate: 8.520% PDF 2: Post-Shock Beliefs 35 Optimist Pessimist Eqlbm sBar 25 20 15 10 5 0 0 0.5 sBar 1 1.5 s Working Paper Series A series of research studies on regional economic issues relating to the Seventh Federal Reserve District, and on financial and economic topics. The Urban Density Premium across Establishments R. Jason Faberman and Matthew Freedman WP-13-01 Why Do Borrowers Make Mortgage Refinancing Mistakes? Sumit Agarwal, Richard J. Rosen, and Vincent Yao WP-13-02 Bank Panics, Government Guarantees, and the Long-Run Size of the Financial Sector: Evidence from Free-Banking America Benjamin Chabot and Charles C. Moul WP-13-03 Fiscal Consequences of Paying Interest on Reserves Marco Bassetto and Todd Messer WP-13-04 Properties of the Vacancy Statistic in the Discrete Circle Covering Problem Gadi Barlevy and H. N. Nagaraja WP-13-05 Credit Crunches and Credit Allocation in a Model of Entrepreneurship Marco Bassetto, Marco Cagetti, and Mariacristina De Nardi WP-13-06 Financial Incentives and Educational Investment: The Impact of Performance-Based Scholarships on Student Time Use Lisa Barrow and Cecilia Elena Rouse WP-13-07 The Global Welfare Impact of China: Trade Integration and Technological Change Julian di Giovanni, Andrei A. Levchenko, and Jing Zhang WP-13-08 Structural Change in an Open Economy Timothy Uy, Kei-Mu Yi, and Jing Zhang WP-13-09 The Global Labor Market Impact of Emerging Giants: a Quantitative Assessment Andrei A. Levchenko and Jing Zhang WP-13-10 Size-Dependent Regulations, Firm Size Distribution, and Reallocation François Gourio and Nicolas Roys WP-13-11 Modeling the Evolution of Expectations and Uncertainty in General Equilibrium Francesco Bianchi and Leonardo Melosi WP-13-12 Rushing into the American Dream? House Prices, the Timing of Homeownership, and the Adjustment of Consumer Credit Sumit Agarwal, Luojia Hu, and Xing Huang WP-13-13 1 Working Paper Series (continued) The Earned Income Tax Credit and Food Consumption Patterns Leslie McGranahan and Diane W. Schanzenbach WP-13-14 Agglomeration in the European automobile supplier industry Thomas Klier and Dan McMillen WP-13-15 Human Capital and Long-Run Labor Income Risk Luca Benzoni and Olena Chyruk WP-13-16 The Effects of the Saving and Banking Glut on the U.S. Economy Alejandro Justiniano, Giorgio E. Primiceri, and Andrea Tambalotti WP-13-17 A Portfolio-Balance Approach to the Nominal Term Structure Thomas B. King WP-13-18 Gross Migration, Housing and Urban Population Dynamics Morris A. Davis, Jonas D.M. Fisher, and Marcelo Veracierto WP-13-19 Very Simple Markov-Perfect Industry Dynamics Jaap H. Abbring, Jeffrey R. Campbell, Jan Tilly, and Nan Yang WP-13-20 Bubbles and Leverage: A Simple and Unified Approach Robert Barsky and Theodore Bogusz WP-13-21 The scarcity value of Treasury collateral: Repo market effects of security-specific supply and demand factors Stefania D'Amico, Roger Fan, and Yuriy Kitsul Gambling for Dollars: Strategic Hedge Fund Manager Investment Dan Bernhardt and Ed Nosal Cash-in-the-Market Pricing in a Model with Money and Over-the-Counter Financial Markets Fabrizio Mattesini and Ed Nosal WP-13-22 WP-13-23 WP-13-24 An Interview with Neil Wallace David Altig and Ed Nosal WP-13-25 Firm Dynamics and the Minimum Wage: A Putty-Clay Approach Daniel Aaronson, Eric French, and Isaac Sorkin WP-13-26 Policy Intervention in Debt Renegotiation: Evidence from the Home Affordable Modification Program Sumit Agarwal, Gene Amromin, Itzhak Ben-David, Souphala Chomsisengphet, Tomasz Piskorski, and Amit Seru WP-13-27 2 Working Paper Series (continued) The Effects of the Massachusetts Health Reform on Financial Distress Bhashkar Mazumder and Sarah Miller WP-14-01 Can Intangible Capital Explain Cyclical Movements in the Labor Wedge? François Gourio and Leena Rudanko WP-14-02 Early Public Banks William Roberds and François R. Velde WP-14-03 Mandatory Disclosure and Financial Contagion Fernando Alvarez and Gadi Barlevy WP-14-04 The Stock of External Sovereign Debt: Can We Take the Data at ‘Face Value’? Daniel A. Dias, Christine Richmond, and Mark L. J. Wright WP-14-05 Interpreting the Pari Passu Clause in Sovereign Bond Contracts: It’s All Hebrew (and Aramaic) to Me Mark L. J. Wright WP-14-06 AIG in Hindsight Robert McDonald and Anna Paulson WP-14-07 On the Structural Interpretation of the Smets-Wouters “Risk Premium” Shock Jonas D.M. Fisher WP-14-08 Human Capital Risk, Contract Enforcement, and the Macroeconomy Tom Krebs, Moritz Kuhn, and Mark L. J. Wright WP-14-09 Adverse Selection, Risk Sharing and Business Cycles Marcelo Veracierto WP-14-10 Core and ‘Crust’: Consumer Prices and the Term Structure of Interest Rates Andrea Ajello, Luca Benzoni, and Olena Chyruk WP-14-11 The Evolution of Comparative Advantage: Measurement and Implications Andrei A. Levchenko and Jing Zhang WP-14-12 Saving Europe?: The Unpleasant Arithmetic of Fiscal Austerity in Integrated Economies Enrique G. Mendoza, Linda L. Tesar, and Jing Zhang WP-14-13 Liquidity Traps and Monetary Policy: Managing a Credit Crunch Francisco Buera and Juan Pablo Nicolini WP-14-14 Quantitative Easing in Joseph’s Egypt with Keynesian Producers Jeffrey R. Campbell WP-14-15 3 Working Paper Series (continued) Constrained Discretion and Central Bank Transparency Francesco Bianchi and Leonardo Melosi WP-14-16 Escaping the Great Recession Francesco Bianchi and Leonardo Melosi WP-14-17 More on Middlemen: Equilibrium Entry and Efficiency in Intermediated Markets Ed Nosal, Yuet-Yee Wong, and Randall Wright WP-14-18 Preventing Bank Runs David Andolfatto, Ed Nosal, and Bruno Sultanum WP-14-19 The Impact of Chicago’s Small High School Initiative Lisa Barrow, Diane Whitmore Schanzenbach, and Amy Claessens WP-14-20 Credit Supply and the Housing Boom Alejandro Justiniano, Giorgio E. Primiceri, and Andrea Tambalotti WP-14-21 The Effect of Vehicle Fuel Economy Standards on Technology Adoption Thomas Klier and Joshua Linn WP-14-22 What Drives Bank Funding Spreads? Thomas B. King and Kurt F. Lewis WP-14-23 Inflation Uncertainty and Disagreement in Bond Risk Premia Stefania D’Amico and Athanasios Orphanides WP-14-24 Access to Refinancing and Mortgage Interest Rates: HARPing on the Importance of Competition Gene Amromin and Caitlin Kearns WP-14-25 Private Takings Alessandro Marchesiani and Ed Nosal WP-14-26 Momentum Trading, Return Chasing, and Predictable Crashes Benjamin Chabot, Eric Ghysels, and Ravi Jagannathan WP-14-27 Early Life Environment and Racial Inequality in Education and Earnings in the United States Kenneth Y. Chay, Jonathan Guryan, and Bhashkar Mazumder WP-14-28 Poor (Wo)man’s Bootstrap Bo E. Honoré and Luojia Hu WP-15-01 Revisiting the Role of Home Production in Life-Cycle Labor Supply R. Jason Faberman WP-15-02 4 Working Paper Series (continued) Risk Management for Monetary Policy Near the Zero Lower Bound Charles Evans, Jonas Fisher, François Gourio, and Spencer Krane Estimating the Intergenerational Elasticity and Rank Association in the US: Overcoming the Current Limitations of Tax Data Bhashkar Mazumder WP-15-03 WP-15-04 External and Public Debt Crises Cristina Arellano, Andrew Atkeson, and Mark Wright WP-15-05 The Value and Risk of Human Capital Luca Benzoni and Olena Chyruk WP-15-06 Simpler Bootstrap Estimation of the Asymptotic Variance of U-statistic Based Estimators Bo E. Honoré and Luojia Hu WP-15-07 Bad Investments and Missed Opportunities? Postwar Capital Flows to Asia and Latin America Lee E. Ohanian, Paulina Restrepo-Echavarria, and Mark L. J. Wright WP-15-08 Backtesting Systemic Risk Measures During Historical Bank Runs Christian Brownlees, Ben Chabot, Eric Ghysels, and Christopher Kurz WP-15-09 What Does Anticipated Monetary Policy Do? Stefania D’Amico and Thomas B. King WP-15-10 Firm Entry and Macroeconomic Dynamics: A State-level Analysis François Gourio, Todd Messer, and Michael Siemer WP-16-01 Measuring Interest Rate Risk in the Life Insurance Sector: the U.S. and the U.K. Daniel Hartley, Anna Paulson, and Richard J. Rosen WP-16-02 Allocating Effort and Talent in Professional Labor Markets Gadi Barlevy and Derek Neal WP-16-03 The Life Insurance Industry and Systemic Risk: A Bond Market Perspective Anna Paulson and Richard Rosen WP-16-04 Forecasting Economic Activity with Mixed Frequency Bayesian VARs Scott A. Brave, R. Andrew Butters, and Alejandro Justiniano WP-16-05 Optimal Monetary Policy in an Open Emerging Market Economy Tara Iyer WP-16-06 Forward Guidance and Macroeconomic Outcomes Since the Financial Crisis Jeffrey R. Campbell, Jonas D. M. Fisher, Alejandro Justiniano, and Leonardo Melosi WP-16-07 5 Working Paper Series (continued) Insurance in Human Capital Models with Limited Enforcement Tom Krebs, Moritz Kuhn, and Mark Wright WP-16-08 Accounting for Central Neighborhood Change, 1980-2010 Nathaniel Baum-Snow and Daniel Hartley WP-16-09 The Effect of the Patient Protection and Affordable Care Act Medicaid Expansions on Financial Wellbeing Luojia Hu, Robert Kaestner, Bhashkar Mazumder, Sarah Miller, and Ashley Wong WP-16-10 The Interplay Between Financial Conditions and Monetary Policy Shock Marco Bassetto, Luca Benzoni, and Trevor Serrao WP-16-11 Tax Credits and the Debt Position of US Households Leslie McGranahan WP-16-12 The Global Diffusion of Ideas Francisco J. Buera and Ezra Oberfield WP-16-13 Signaling Effects of Monetary Policy Leonardo Melosi WP-16-14 Constrained Discretion and Central Bank Transparency Francesco Bianchi and Leonardo Melosi WP-16-15 Escaping the Great Recession Francesco Bianchi and Leonardo Melosi WP-16-16 The Role of Selective High Schools in Equalizing Educational Outcomes: Heterogeneous Effects by Neighborhood Socioeconomic Status Lisa Barrow, Lauren Sartain, and Marisa de la Torre Monetary Policy and Durable Goods Robert B. Barsky, Christoph E. Boehm, Christopher L. House, and Miles S. Kimball Interest Rates or Haircuts? Prices Versus Quantities in the Market for Collateralized Risky Loans Robert Barsky, Theodore Bogusz, and Matthew Easton WP-16-17 WP-16-18 WP-16-19 6