Pledgeability, Industry Liquidity, and Financing Cycles. Abstract

Transcription

1 March 7 Pledgeability, Indutry Liquidity, and Financing Cycle Dougla W. Diamond Yunzhi Hu and Raghuram G. Rajan Chicago ooth and NER Chicago ooth Chicago ooth and NER Abtract Why are downturn following prolonged epiode of high valuation of firm o evere and long? Why do firm promie high external payment when they anticipate high valuation, and underperform ubequently? In thi paper, we propoe a theory of financing cycle where the control right to enforce claim in an aet price boom (right to ell aet) differ from the control right ued in more normal time (right over cah flow that we term pledgeability ). Firm management limited incentive to enhance pledgeability in an aet price boom can have long-drawn advere effect in a downturn, which may not be reolved by renegotiation. Thi can alo explain why involuntary aet turnover and aet miallocation to outider are high in a downturn, a well a why indutry productivity fall. The paper highlight an advere conequence of high anticipated liquidity, working through leverage, on the economy acce to finance and productivity when that liquidity fail to materialize. Diamond and Rajan thank the Center for Reearch in Security Price at Chicago ooth and the National Science Foundation for reearch upport. Rajan alo thank the Stigler Center. We are grateful for helpful comment from Florian Heider, Alan Morrion, Martin Oehmke, and Adriano A. Rampini, a well a workhop participant at the OXFIT 4 conference, Chicago ooth, the Federal Reerve ank of Chicago, the Federal Reerve ank of Richmond, the NER 5 Corporate Finance Summer Intitute, Science Po, American Finance Aociation meeting in 6 in San Francico, Princeton, MIT, the European Central ank and oton Univerity.

2 Why do downturn following prolonged epiode of high firm valuation prove to be detrimental to growth and reult in more protracted receion (ee Krihnamurthy and Muir (5) and López-Salido, Stein and Zakrajšek (5))? One traditional rationale i baed on the idea of debt overhang the debt built up during the boom erve to retrict invetment and borrowing during the but. However, if everyone, including the debt holder, know that debt i holding back invetment, they have an incentive to write down the debt in return for a take in the firm growth. For debt overhang to be a eriou concern, the firm and debt holder mut be unable to undertake value enhancing contractual bargain. Another view i that borrower cannot be truted to take only value enhancing invetment, even in a downturn. So debt overhang i needed to contrain the borrower invetment overhang i a econd bet olution to a fundamental moral hazard problem (ee Hart and Moore (995)). The immediate quetion raied by uch an analyi i why we want to contrain borrower more in bad time when the contraint impoed by debt are already high. Moreover, why would the moral hazard problem be o much more eriou in a downturn which follow high valuation? In thi paper, we provide an explanation of the caue and conequence of financial liability overhang (including debt overhang) and explain why it i more acute following period of high valuation and rational optimim about the future value of firm. In doing o, we differentiate between the control right that are due to high reale price for aet, which enable external claim to be enforced in a boom, and control right baed on pledging of cah flow, which facilitate the enforcement of external claim at other time, including downturn. The tranition between thee regime, in which different type of control right are operational, caue the external claim build up during the boom to have long-drawn advere effect in the downturn. Let u be more pecific. Conider an indutry that require pecial managerial knowledge. Within the indutry, there are firm run by incumbent. There are alo indutry inider (thoe who know the indutry well enough to be able to run firm a efficiently a the incumbent). Indutry outider (financier who don t really know how to run indutry firm but have general managerial/financial kill) are the other agent in the model. We firt examine the effect of financing firm with fully tate-contingent financial contract, and then we turn to tandard debt with a contant payment in a given period. Financier have two ort of control right; firt, control through the right to repoe and ell the underlying aet being financed if payment are mied and, econd, control over cah flow generated by the aet. The firt right only require the frictionle enforcement of property right in the economy, which we aume. It ha epecial value when there are a large number of capable

3 potential buyer willing to pay the full price for the firm aet. Greater wealth amongt indutry inider (which we term indutry liquidity) increae the availability of thi aet-ale-baed financing. ecaue we analyze a ingle indutry, high level of thi indutry liquidity can be interpreted a an economy-wide boom. The econd type of control right i more endogenou, and conferred on creditor by the firm incumbent manager a he make the firm cah flow more appropriable or pledgeable over the medium term for example, by improving accounting tandard and tranparency, by etting up ecrow account and monitoring arrangement, by including debt covenant and condition on dividend payment, or even by tandardizing managerial procedure o a to make herelf more replaceable a a manager. From the incumbent manager perpective, enhancing cah flow pledgeability i a double-edged word. It make it eaier for her to ell the firm when he i no longer fit to run it becaue new buyer can borrow againt future pledgeable cah flow to finance the acquiition. However, it alo enable exiting creditor to collect more if he tay in control, which reduce her incentive to enhance pledgeability. Thu the choice of cah flow pledgeability i ubject to additional moral hazard, over and above the intrinic reluctance of the incumbent to repay outide financier. Thi limit the external financing capacity of the firm. The advantage of high pledgeability for financial capacity have been tudied by Holmtröm and Tirole (998). We examine the tradeoff between the advantage and diadvantage of increaed pledgeability for the incumbent, a indutry liquidity alo varie. Our goal i to undertand how the external obligation built up in a boom affect a firm pledgeability choice, and it ubequent acce to financing. When market are buoyant and indutry inider have plenty of cah, repayment i enforced by the high reale value of aet. There i no additional need to rely on pledged cah flow. Indutry aet trade for fundamental value (with no underpricing), a in Shleifer and Vihny (99). The mot efficient uer hold the aet becaue they have enough cah and borrowing capacity to pay full value. The high anticipated reale value increae the amount of financing that a firm can credibly repay and thu the potential leverage of the firm. If competitive preure indeed force the firm to lever up, the incumbent incentive to enhance pledgeability further diminihe becaue the incumbent benefit le from a higher ale price when debt take away much of the ale proceed. Liquidity operate through leverage to crowd out pledgeability. With pledgeability low, a downturn, even one that i anticipated to occur with ignificant probability, can then impair firm performance everely. Indutry inider, alo hit by the downturn, no longer have peronal wealth to buy aet, nor doe the low cah flow pledgeability of the firm allow them to borrow againt future cah flow to pay for purchae. Since external claim are high in thee epiode, the firm may be old to outider. While indutry outider have little

4 ability to operate the aet themelve, thi may be a virtue they have a trong incentive to improve aet pledgeability becaue they do not want to own the aet long term, but intead want to ell the aet back to indutry inider at a high price. Outider play a critical role, therefore, not becaue they are fluh with fund but becaue they are not ubject to moral hazard over pledgeability. Importantly, financier have little incentive to renegotiate down fixed debt claim in a downturn, ince the reallocation of the firm to indutry outider may be the outcome that maximize their claim, given pat pledgeability choice. Conequently, in a downturn following a boom, a larger number of the new aet owner will be le-productive indutry outider, reducing average productivity. Eifeldt and Rampini (6, 8) provide evidence conitent with thi. Eventually, a the economy recover, outider ell the aet back to the more productive indutry inider, a the higher pledgeability increae the inider ability to raie money againt future cah flow. Recoverie following period of an aet price boom and high leverage are thu delayed, not jut becaue debt ha to be written down and undoubtedly friction in writing down debt would increae the length of the delay but alo becaue corporation have to retore the pledgeability of their cah flow to cope with a world where liquidity i more carce. It i the latter which may make the debt hangover more prolonged. High anticipated liquidity, therefore, not only lead to greater financial leverage, but alo the combination lead to low pledgeability being choen. Thi then lead to ditortion in allocation, a unproductive uer of the aet take control in a downturn from more productive uer. Higher anticipated liquidity in ome future tate can therefore induce more eventual miallocation in le liquid tate, a pillover effect between tate that operate through leverage and pledgeability! The liquidity-leverage overhang on pledgeability choice reemble traditional debt-overhang (Myer (977)), where firm deciion are ditorted whenever the deciion caue an increae in the value of outtanding debt. However, it differ in important way. The outide claim in our model could be any claim whoe value i boltered by the threat of outide ale or takeover in time of trong liquidity, a well a internally-et governance improvement in more normal time. So while the fixed nature of the outide claim help in making the point, the effect generalize to other variable (but not fully tate-contingent) outide claim like equity. Moreover, the underinvetment i in pledgeability or governance, and the inefficiency i oberved ex pot, not ex ante, a aet go into the hand of low-productivity outider. The effect of thi underinvetment are oberved primarily after boom which give way to downturn that were known to be poible but were rationally overlooked. In that ene, outide-claim overhang i an indutry or economy-wide effect, wherea traditional debt overhang occur in a ingle firm a it lever up exceively. Our paper explain why aet price boom baed on a combination of liquidity and leverage can be fragile (ee, for example, orio and Lowe (), Adrian and Shin (), and Rajan and 3

5 Ramcharan (5)). It alo ugget a reaon why credit cycle emerge, though a dynamic extenion to the model i needed to explain the propertie of uch cycle fully (ee, for example, Kiyotaki and Moore (997)). More broadly, it ugget theoretical underpinning for financing cycle (orio ()), where a imultaneou and utained rie in aet price and leverage could ignificantly augment, and increae the peritence of, buine cycle downturn. Our paper build on Shleifer and Vihny (99), where the high net worth of indutry participant allow aet to ell for their fundamental value becaue the bet uer of an aet can outbid le efficient uer, which lead to efficient reallocation. In their paper, reallocation to inefficient uer take place only when indutry inider are le liquid (traditional debt overhang effect) than outider. Eifeldt and Rampini (8) develop a theory where capital reallocation i more efficient in good time, with key ingredient being private information about managerial ability and cyclical effect of labor market competition for manager. Good time lead to high required cah compenation to manager becaue reervation managerial wage become elevated. A a reult, high ability manager can accept lower wage in return for the benefit of managing more aet. They ue the differential compenation to bribe low ability manager to give up their aet. In bad time, managerial compenation i lower and even if high ability manager accepted zero cah compenation, it would not be ufficient to bribe low ability manager to give up their aet. Thi lead to a more efficient reallocation of capital in good (high compenation and therefore high manager liquidity) time and le in bad. In both Shleifer and Vihny (99) and Eifeldt and Rampini (8), adjuting for current condition (uch a indutry net worth or compenation), pat action do not affect financial capacity or the efficiency of reallocation of capital today. Thi i unlike our model, where hitory matter, allowing u to explain prolonged downturn following boom, and ketch the poibility of financing cycle. Moreover, outider in our model are not necearily more liquid, but till play an important role becaue they do not uffer from moral hazard over pledgeability. They take over the firm temporarily o a to raie future pledgeability, even though they cannot generate cah flow. Finally, and perhap mot important, higher liquidity i not problematic in their model, unlike in our where it effect can be tranmitted through greater anticipatory leverage and lower pledgeability into wore allocation. The ret of the paper i a follow. In Section I, we decribe the baic benchmark model of pledgeability choice and the timing of deciion in a three-period model. In Section II, we analyze the implication of pledgeability choice when financial contract are fully tate contingent and See enmelech and ergman (), Coval and Stafford (7), and Shleifer and Vihny () for comprehenive review. 4

6 pledgeability can be choen flexibly in repone to the tate. The maximum amount that can be pledged to outide invetor i characterized, and the fundamental tradeoff in the model are explained, focuing on the lat two period. We then analyze the initial period. In Section III, we examine the implication of tandard debt contract and peritent pledgeability choice that are made before uncertainty i fully revealed. In Section IV, we dicu implication and conclude in Section V. A. The Indutry and State of Nature I. The Framework Conider an indutry with 4 date (-,,, ) and 3 period between thee date, with date t marking the end of period t. A period i a phae of the financing cycle (ee orio (4) for example), and extend over everal year. The tate of the indutry i realized at the beginning of every period. In the good tate G, the indutry proper. In the bad tate, indutry-wide ditre occur. In period, the indutry i in tate G, Similarly in period, the probability of tate G i q, with the probability of tate G being G G q (ee Figure ).. In period, we aume the indutry return to tate G for ure thi i meant to repreent the long run tate of the indutry (we model economic fluctuation and not apocalype). A full decription of the tate in period t include the tate that were realized in previou period, but where thi i unneceary we will kip it for convenience. Figure : State of Nature. Agent and the Aet There are two type of agent in the economy: High type (H) are indutry inider with high ability to produce with an aet, which we call the firm. There i ome mutual pecialization 5

7 etablihed over the period between the incumbent manager and the firm that create a value to incumbency. When the tate i G, only a high type manager in place at the beginning of a period t can produce cah flow C t with the aet over the period. In the tate, however, even a high ability manager cannot produce cah flow. A low type (L) manager ha no ability to produce cah flow regardle of the tate. Thee could be indutry outider uch a financier who hold the aet for the purpoe of reelling, or indutry inider who have lot their ability (ee below). Financier alo have fund, which they will lend to other managing the firm if they expect to break even. All agent are rik neutral. We ignore time dicounting, which i jut a matter of recaling the unit of cah flow. H A high ability manager retain her ability into the next period only with probability. Think of thi a the degree of tability of the indutry. Intuitively, the critical capabilitie for ucce are likely to be table in a mature indutry or in an indutry with little technological innovation. However, in an indutry which i young and unettled, or in an indutry with ignificant innovation, the critical capabilitie for ucce can vary over time. A manager who i very appropriate in a particular period may be ineffective in the next. Thi i the ene in which an incumbent can loe ability and thi occur with higher probability in a young or changing indutry. The incumbent lo of ability in the next period become known to all hortly before the end of the current period. Lo of ability i not an indutry wide occurrence and i independent acro manager. So even if a manager loe her ability, there are a large number of other indutry inider manager equally able to take her place next period. If a new high ability manager take over at the end of the current period, he will hape the firm toward her idioyncratic management tyle, o he can produce cah flow with the firm aet in future period in good tate. C. Financial Contract Any manager can raie money from financier againt the aet by writing one period financial contract. Although our ultimate goal i to undertand the effect of debt contract, we begin by analyzing an economy in which contract are allowed to be tate contingent, o promied payment at the end of period t are D. t t Having acquired control of the firm, a manager would like to keep the realized cah flow for herelf rather than hare it with financier. Two ort of control right force the manager to repay the external claim. Firt, the financier automatically get paid a portion that we call pledgeable of the cah flow produced over the period, up to the amount of the financier claim. Second, jut before the end of the period, the financier get the right to auction the firm to the highet bidder if he ha not been paid in full. elow we decribe the two control right in detail. 6

8 D. Control Right over Cah Flow: Pledgeability Let u define pledgeability a the fraction of realized cah flow that are automatically directed to an outide financier. In practice, it i determined by a variety of factor: the information poeed by the financier and hence the nature of the financier ( arm length like a bond invetor or relationhip like a banker); the nature of financing (for example, dipered or concentrated); the financial contract (covenant lite or covenant heavy); the quality of the accounting ytem in place; the tranparency of the organizational tructure and the ytem of contracting (e.g., the abence of pyramid, the rule governing related party tranaction, etc.); and the check and balance that are impoed on the manager by the organization (the quality and independence of the board, the replaceability of the CEO, the independence of the auditor and the audit committee, etc.). The incumbent chooe pledgeability thi period, but it i embedded only by next period, and will then perit for the entire period. So pledgeability t choen in period t i the fraction of period t+ cah flow that can be automatically paid to outide financier. [, ], where the range of feaible value i determined by the economy intitution upporting corporate governance (uch a regulator and regulation, invetigative agencie, law and the judiciary). Alo,. To et t, it cot ε, where ε i the cot of action uch a hiring a reputable accountant. Our reult will be preented primarily for the cae where, and a poitive will only alter the reult quantitatively. While a low-type incumbent cannot generate cah flow, he too can et next period pledgeability he doe not have indutry-pecific managerial capabilitie but ha governance capabilitie. E. Control Right over Aet: Auction and Reale If the financier ha not been paid in full from the pledged cah flow and any additional um the incumbent voluntarily pay, then the financier get the right to auction the firm to the highet bidder at date t. One can think of uch an auction a a form of bankruptcy. Therefore, the incumbent can retain control by either paying off the financier in full (poibly by borrowing once again againt future pledgeable cah flow) or by paying le than the full contracted amount and outbidding other bidder in the auction. F. Initial Condition and Wealth wealth At date, the incumbent ha initial wealth H, i, t. Let indutry inider tart out with which, recall, i termed indutry liquidity. If the tate i good in period t, we aume 7

9 that both the wealth of the incumbent and indutry inider go up by Ct (the indutry boom lift the private income of all inider whether owning a firm or working a contractor, conultant or employee). Furthermore, the wealth of the incumbent increae by an additional tct, the unpledged cah flow he generate within the firm. G. Efficiency The meaure of uncontrained economic efficiency we ue through the ret of thi paper i the extent to which the aet i in the hand of the mot productive owner at that time. We do not model invetment, intead auming that the aet exit and i owned by an incumbent. 3 H. Timing We will tart by examining incentive in period. The timing of event i decribed in Figure. We aume that the incumbent learn the tate, then et pledgeability, knowing the amount of payment that i due at date. Next, her ability in period i realized. Subequently, production take place and the pledgeable fraction of cah flow (et in the previou period) goe to financier automatically. She either pay the remaining due or enter the auction. The period end with potentially a new incumbent in place. Figure : Timing and Deciion in Period II. Pledgeability Choice with State-Contingent Contract We now focu on deciion in period. What determine pledgeability? How doe the level of promied payment D influence the incumbent incentive to et pledgeability? We will how t t that both the choice of pledgeability and the maximum tate-contingent payment are determined by 3 Alternatively, we can put a minimum cale on the value of real input to be aembled into the firm at the initial date -, and aume the firm tart at that date only if enough funding i available. A a reult, inefficient underinvetment may occur if incentive to make cah flow pledgeable or to tranfer the firm to more efficient producer are ufficiently weak, for bid may be reduced at date - below thi floor. 8

10 two form of interacting moral hazard. Firt, incumbent can withhold cah flow from financier except for what they are forced to pay by pre-et pledgeability or the financier threat to eize and auction aet. Second, the incumbent can et future pledgeability low, potentially reducing the amount that financier are able to collect. We aume that period tart with a high type manager in place. In thi etting, all outcome will be efficient, and many of our poitive implication will be clear. In Section.5, we extend the analyi to period and how an inefficient real outcome: the aet can poibly be old to a low type at date, and period can tart with a low type manager in place... Date Since the economy end at date, and there i no uncertainty over the tate in period, a high type indutry inider who bid for control at date can borrow up to D C where i preet by the incumbent in period. The incumbent can alo borrow up to C at date if he remain a high type and bid to retain control into period... Date Let D be the promied payment to the financier at date in tate, G,. If the incumbent in period i an indutry inider and =G, cah C goe directly to the financier (up to the value of her promied claim), where i pledgeability that wa et in period. The remaining payment due i D D Min[ C, D ]. If =, then D D. Indutry Inider id In any date auction for the firm, indutry outider or financier do not bid to take direct control of the firm ince the firm generate no cah flow in their hand in the lat period, and the firm ha no reidual value. Indutry inider, however, bid uing their date wealth and any amount that can be borrowed at date by pledging period output. Their wealth increae over period by C HG, H, H, H, in tate G, and remain unchanged in tate, i.e., C and. Together H, with the amount Cthey can borrow, the total amount that they can pay i C. Of coure, they will not bid more than the total value of cah flow, C. So the maximum auction bid at date i ( ) Min[ C, C ]. H, H, A meaure which will help undertand the model better i potential underpricing, which i the difference between the preent value of future cah flow accruing to an indutry inider if he buy the firm and the amount that he can bid if the incumbent ha et period- pledgeability to be low. It equal 9

11 C at date. y chooing different level of period- pledgeability, the incumbent can H, ( ) H, H, vary indutry inider bid between ( ) and ( ), thu altering the realized underpricing, which i the difference between the preent value of future cah flow and the actual bid, i.e., H, C ( ). Incumbent id The incumbent ha to repay the financier in full or outbid other in an auction if he want to H, Min [ D, ]. The cah he ha at date i the retain control into period. That i, he pay i, initial wealth level,, augmented by C in tate G, plu the non-pledgeable portion of cah ig, i, flow generated during period. At date, the incumbent ha cah C if the i, i, period tate i G, and if the tate i. In addition, if he know he i going to keep her ability in period, he can alo raie fund againt period output, C. Therefore, the incumbent i, i, can pay a much a min C, C to the financier. The incumbent will retain control H, Min [ D, ]. Since the continuation if the amount he can pay i (weakly) greater than value of the aet, C, i identical for the incumbent and indutry inider, the incumbent i alway willing to hold on to the aet if he i able to, unle he ha lot her ability, or he i a low type to begin with, when he will want to ell out ince he cannot generate cah flow next period. Regardle of who win, if the incumbent in period i a high type, the financier recoup H, Min[ C, D ] + Min [ D, ] if the tate i G and H, Min D, if the tate i. The financier threat of eizing and elling aet i therefore a powerful intrument for him to H, extract repayment. The value of that threat depend on the bid by indutry inider, which in turn depend on the wealth of indutry inider and the future pledgeability of the aet. D, Max We now how the incumbent choice of pledgeability and the maximal credible payment,, are jointly determined, depending on whether the incumbent can outbid indutry inider. It i eaily hown that becaue of linearity, the incumbent never et pledgeability at an interior level in the range. We identify three cae: (i) Pledgeability doe not matter for repayment. (ii) The incumbent can never outbid indutry inider. (iii) The incumbent can alway outbid indutry inider. We olve explicitly for the maximal credible payment, Max D in all thee cae.

12 (i) Pledgeability doe not matter for repayment (no potential underpricing) H, When ( ) C, indutry liquidity i ufficiently high that high-type inider can pay the full price of the aet, even if the incumbent ha choen low pledgeability, o, Max D C. In thi cae, there i no potential underpricing and pledgeability doe not matter for repayment. A a reult, the incumbent will et pledgeability to be low. External payment are committed to through the high reale price of the aet. High pledgeability i neither needed nor deired by anyone in thi cae. (ii) Incumbent cannot outbid indutry inider in an auction H, H, i, Let C ( ) o there i potential underpricing. When ( ) ( ), the indutry inider can alway outbid the incumbent no matter what level pledgeability i et at. 4 y etting payment at or below, Max H, D ( ), the incumbent i incentivized to et next period pledgeability at. She recoup the cot of etting pledgeability high becaue the promied payment i at leat below the auction bid. It i eay to check that the incumbent payoff would never increae if he et pledgeability lower. (iii) Incumbent alway can outbid indutry inider i, H, Now let ( ) ( ) o that the incumbent can outbid the indutry inider regardle of her choice of pledgeability. She chooe iff H ( C Min[ D, ( )]) ( )( ( ) Min[ D, ( )]) () ( [, ( ))]) ( )( ( ) [, ( ))]) H, H H, H, H H, H H, H, C Min D Min D The left hand ide i the incumbent continuation value if he chooe ide i the one if he chooe, while the right hand. The firt term on each ide of () i the reidual amount the incumbent expect if he remain a high type in period. The econd term on each ide i the expected reidual amount if he loe her ability and ha to auction the firm at date. Note that a higher (weakly) increae the amount the incumbent ha to pay the financier when he retain 4 Strictly peaking, there i one more cae becaue we break tie in favor of the incumbent. If H, i, H, i, C ( ) ( ) and ( ) ( ), the incumbent retain control if he chooe high general pledgeability and continue to be a high type, becaue he i able to pay the full value of the aet C, i, and inider will not outbid her. y contrat, if he chooe low pledgeability and debt i above ( ), he loe control becaue the high promied payment i enforceable and higher than what he can pay. The maximum level of debt i a in cae (ii).

13 capability and control, therefore (weakly) decreaing the firt term, while it (weakly) increae the amount the incumbent get in the auction if he loe capability, thu (weakly) increaing the econd term. The incumbent therefore trade off higher poible repayment againt higher poible reale value in chooing. Importantly, a higher outtanding promied payment reduce the incumbent incentive to chooe higher becaue more of the pledgeable cah flow are captured by financier if the incumbent tay in control, and more of the reale value alo goe to financier if the aet i old. Thi i the ource of moral hazard over pledgeability. The maximum level of promied payment D that till give her an incentive to chooe, Max PayIC H H, H H, i D ( ) ( ) ( ), where upercript PayIC indicate the payment i incentive compatible. Note that it i eaier to incentivize the incumbent, and thu upport higher payment, when the probability he loe kill ( H ale. Lemma. ummarize the reult when G, ) i high, for thi enhance the likelihood of. Lemma. (i) If ( ) C, H, D C Max, and.. For any promied payment D D Max,, the, incumbent expect V i ( D ) C D H, i, H, Max, H, (ii) If C ( ) and ( ) ( ), D ( ) and. For any Max,,, promied payment D D, the incumbent expect V i ( D ) H ( ) D if,, i ( ) D Max i, H H H, D, and expect V ( D) C( ) ( ) D if i, D ( ). H, i, H, Max, PayIC (iii) If C ( ) and ( ) ( ), D D and. For any promied Max, i, H H H, payment D D, incumbent expect V ( D ) C ( ) ( ) D. Proof: See Appendix. In Cae (i), there i no potential underpricing and the choice of pledgeability doe not matter for payment. In Cae (ii), the incumbent loe control whenever he enter an auction. The maximal promied payment i et by indutry liquidity and the need to compenate the incumbent for incurring the cot of etting pledgeability high. In cae (iii), however, the incumbent i able to hold onto the aet for any choice of pledgeability, provided he retain capability. Therefore, the maximal promied payment ha to be ignificantly lower to incentivize high pledgeability choice the

14 incumbent higher ability to retain control make a higher outide bid price le attractive and increae the moral hazard over pledgeability. Thi i why, a illutrated by the top left panel of Figure 3, the maximum credible payment D i, i, decreae (weakly) with incumbent wealth. When i low a in cae (ii), the, Max H, promied payment can be et a high a i,. However, a increae and the incumbent gain the ability to retain control, her incentive tart mattering, reulting in a lower credible payment D PayIC. Of coure if there i no potential underpricing to begin with, the incumbent wealth doe not matter for repayment. Although D i, decreae with, the continuation value accruing to an, Max incumbent for a given payment D i, alway increae with (Figure 3, the top right panel).,max Figure 3: D i, (left) and Other parameter: i, H, (top) and (bottom) V D (right) a function of.,.3,., C,. 5,, D. 6 H, H H A dicued earlier, an increae in tability decreae the maximum incentive compatible payment D PayIC, Max, and hence weakly decreae D. Intuitively, the higher i tability, 3

15 lower the probability that a ale will be neceary. For any debt level, thi increae the attractivene for the incumbent manager to chooe low pledgeability to reduce the enforceable payment. 5 H,, Max Finally, an increae in indutry liquidity alway raie D. There are two channel at work here. An increae in indutry liquidity puhe up the amount indutry inider can pay, H, ( ), for any level of pledgeability. It alo expand the parameter range in which either there i no potential underpricing or the incumbent cannot retain control. Conequently, again, the maximum pledgeable payment increae. The bottom left panel of Figure 3 illutrate thi by plotting againt fixed D. H, i,. It i eaily een that,max D V D H, varie with in a non-monotonic manner for a Corollary.:, Max D H, H i, increae (weakly) with, decreae (weakly) in and. i, V ( D i, H, ) increae with, and varie non-monotonically with. Proof: Follow from dicuion above..3. Involuntary Management Turnover with State-contingent Contract Define an involuntary turnover a one where an incumbent who retain ability ha to ell an underpriced firm. We now tudy how involuntary management turnover varie acro different tate. A we jut aw, Lemma. indicate equilibrium outcome are determined by three tate variable: indutry liquidity, whether the incumbent can make the payment, and whether the incumbent can outbid indutry inider once he enter an auction. If promied payment are choen H, o a to maximize payout, indutry liquidity,, fully determine pledgeability choice: low pledgeability i choen if and only if there i no potential underpricing. Indutry-wide liquidity i unambiguouly the highet in tate GG, which i meant to capture long- HG, term boom. There i no potential underpricing in tate GG if C C. Potential H, G, underpricing i guaranteed in all other tate if C and C C implie the reult in Propoition.. H. Thi 5 There i a parallel here to Jenen (986) argument that free cah flow in mature indutrie lead to greater wate. In hi view, the paucity of invetment need in mature indutrie reult in firm generating ubtantial free cah flow (and hence poorer governance becaue of a lower need to return to the market for funding). In our model, the lower probability of the need to ell the firm to manager with different capabilitie (or equivalently, the lower need to iue financial claim to raie finance for unmodeled invetment) in a mature or table indutry reduce the need to maintain better outide pledgeability. 4

16 C C C, low pledgeability H G Propoition.: If if and only if the tate i GG., H, i choen Clearly, by definition, involuntary management turnover take place only when there i potential underpricing. In thi cae, the other two tate variable alo become crucial. Clearly, if the incumbent can make the contracted payment D, no auction or involuntary turnover take place. Let u now ee what happen if the incumbent cannot pay and i tipped into the auction. H, Even in the abence of any moral hazard over pledgeability, actual date- repayment i capped by the maximal poible bid that an incumbent encounter in an auction. Thi necearily implie that repayment i low when indutry liquidity i low, uch a in tate. Second, the need to provide incentive for the incumbent to chooe high pledgeability will further reduce maximum repayment. If the incumbent either cannot retain the aet or can only retain by paying the full price (cae (ii) in Lemma.), maximal repayment i reduced by a mall amount,. If the incumbent can retain the aet (cae(iii)), the credible amount that he can promie i reduced by a ignificant amount to H, from D, PayIC. Thu maximal repayment fall with the everity of the incentive problem, which turn on whether the incumbent can outbid indutry inider if he were forced into an auction, which in turn depend on the wealth of the incumbent relative to indutry inider. Note that in both tate G and tate, the incumbent ha trictly le wealth than indutry inider for two reaon. Firt, he doe not produce any cah flow over the period. Second, he will ha ued up her previou wealth in making payment intuitively, when the incumbent can write tate-contingent contract, it never make ene for her to retain non-tate contingent cah when future payment are poitive it i better for her to pay out everything and make future payment lower and more favorably tate contingent. Thu, a long a he ha to make payment at date, the incumbent will not leave date poeing any cah. Conequently, in both tate G and, repayment i et at H, or below, leaving her enough to incentivize high pledgeability choice, even though he alway loe the firm involuntarily at date in thee tate. Of coure, all thi i contingent on the incumbent not being able to pay the contracted payment. The contracted payment may, however, be lower in tate G. Thi i becaue the incumbent i wealthier in tate G (he ha C C ) than in tate (), and ince he ue thi wealth to reduce what need to be raied at date, he may iue lower tate-contingent claim at date in tate G G. The incumbent may thu be able to reduce the required payment in tate G, D below what he 5

17 ig, can raie,, make the payment and thu avoid an auction for control. Thi mean involuntary turnover may be lower in tate G than in tate. In tate G, the incumbent may actually have more wealth than indutry inider. To ee thi, uppoe indutry inider tart with liquidity on date -. If the tate i G in period, the difference between incumbent and indutry inider i C. If C, the 6 incumbent i able to outbid inider in tate G, involuntary turnover do not occur even if the to H, G auction i triggered, and a ignificant reduction in required payment (from i needed to incentivize the incumbent to chooe high pledgeability. Propoition. ummarize the reult. Propoition.: If C and the condition in propoition. hold, ) Maximal repayment i lower in all tate other than GG: D Max, H, D in tate G and. D Max, PayIC G i, G ) Involuntary management turnover alway occur in tate G if D i, if D in tate G, and. Involuntary turnover occur le frequently (i.e., for a maller et of parameter) in tate G than in tate. Turnover i not involuntary in tate G..4. Dicuion D ) G, PayIC, and in tate We have outlined two kind of moral hazard moral hazard over repayment, and moral hazard over etting pledgeability. The two are related. When the economy i in a prolonged boom, indutry inider can pay full value for the firm even when pledgeability i et at a minimum level. There i no need to reduce the moral hazard over repayment by increaing pledgeability ince creditor can extract full repayment through the threat of aet ale. However, when indutry wide liquidity i lower, indutry inider bid for the firm are lower than the future cah flow it generate. Thi underpricing mean that their bid can be raied by etting pledgeability higher. Not only doe thi raie what the incumbent can get if he ha to ell the firm, it alo increae the financier ability to extract repayment from her if he doe not. Thu when the firm finance in the midt of more normal indutry liquidity, not only i higher pledgeability neceary to reduce moral hazard over repayment, but alo the incumbent face moral hazard over pledgeability if payment are et too high. Importantly, thi lat moral hazard i quite different than the effect of fixed payment in the tandard riky debt overhang model. In thoe model, the incumbent avoid invetment becaue the return are largely captured by exiting debt in certain tate of the world. Outide claim overhang in

18 our model tem from the incumbent trying to reduce payment he mut make in the future, by trategically reducing financier enforcement capability. She ha a particular incentive to do thi, not in the tate where everyone i liquid o he ha little capability of altering enforcement (uch a tate GG), nor in the tate where he i illiquid o he ha little propect of retaining control (uch a tate G and ), but in tate of moderate overall liquidity (uch a tate G). The fixed nature of the outide claim acro tate i not critical, only that the claim enforcement can be altered by the incumbent. So outide equity can alo be a ource of overhang. We will how later that riky debt ha intereting additional effect. The pat and the future of the indutry thu interact in intereting way. When the indutry experienced good outcome in the pat, and the future i alo expected to be good enough that there i no potential underpricing, financing capacity i the highet. Not only i the firm likely to generate more in the future, but financier can expect to recover what they lent through the threat of elling the fully priced aet. So they are willing to lend large amount both aet value and debt to value ratio are high. For intermediate level of pat indutry performance, indutry bidder cannot bid full value for the firm future cah flow out of their accumulated liquidity, o pledgeability of future cah flow become important to getting high outide bid and repayment. ut becaue moral hazard over pledgeability kick in, committed payment cannot be too high o a to not dicourage high pledgeability. A fall in indutry performance therefore ha a double whammy effect on financing capacity by both increaing the underpricing of the firm aet by other indutry bidder (becaue of their reduced liquidity) and alo reducing the maximum poible committed payment a a fraction of that lower value (becaue of moral hazard over pledgeability). In other word, both aet value and loan to value ratio plummet when liquidity fall. Finally, there i an additional twit becaue the incumbent liquidity generally fluctuate more with indutry performance than other indutry inider, becaue he ha paid out her cah up front to take control over the firm. Thi mean that if the indutry ha a equence of bad outcome, the incumbent would be unable to retain control in the face of higher indutry bid for the firm. Interetingly, thi will reduce moral hazard over pledgeability ince the incumbent, with no hope of retaining control, focue on getting the maximum bid for the firm. The payment that can be committed to lender will now be a higher fraction of firm reale value, even though intrinic firm value itelf i low. Indeed, thi apect of the model i again reminicent of Jenen Free Cah Flow Hypothei (Jenen (986)), where lower cah with the incumbent reduce moral hazard. Importantly, therefore, while the market value of aet fall with indutry liquidity, loan to value ratio may not decline monotonically. In the model thu far, we have hown that involuntary management turnover i higher in peritent downturn, and thi add to voluntary turnover a manager loe capability. In time of high 7

19 liquidity, there are no involuntary turnover. However, incumbent may voluntarily ell out becaue they are getting full value for the firm, and do not have enough liquidity to match indutry inider offer. So without additional pecificity about parameter, it i not poible for u to ay whether overall turnover will be higher in very liquid or very illiquid tate. Moreover, we have ignored another form of voluntary turnover when the incumbent retain ability for now, but anticipate lower ability with higher probability in the future. If incumbent can chooe the timing of their leaving the buine, they would certainly prefer to ell out when the aet i fully priced than when the aet i priced at a fraction of it fundamental value. We would thu ee voluntary control tranfer increae when aet price are high (i.e., aet are not underpriced). Thi effect of management exit through retirement would further increae turnover in time of high liquidity (peritent indutry up turn) relative to other time. We will explore thi poibility in future work..5. Date and the Prolonged Recovery Let u turn now to date and the incentive that determine how i et in period. There are two crucial difference between the date and date analyi. Firt, financier may be able win a bid at date with the purpoe of reelling the aet at date, while they will not bid at date. Second, the continuation value of the aet from date onward to the period- incumbent may be different from that to an indutry inider even though they have the ame expected capabilitie -- becaue they have different wealth, they have different expectation of retaining control at date. Contrat thi to date, when continuation value to both capable incumbent and inider ic becaue there are no i, G control contet after date. Let C H, G ( D, D ) and C ( D, D ) be the date- expected continuation value of the aet to the continuing incumbent and to indutry inider repectively, G where D and D are due on date in tate G and repectively. oth continuation value G G i hare the ame expreion,, G G G i, q CD V ( D ) ( q ) D V ( D ), but the G incumbent and indutry inider will borrow differentd, D in general. i, H, L, Let ( ), ( ), and be repectively the incumbent, the inider, and the i, outider/financier bid at date. A before, ( ) i the minimum of the incumbent ability to pay and the aet continuation value to the incumbent: ( ), (, )] [ G i, G G i, ( ) Max Min C q C D q D C D D G G, Max D D, Max D D G G Similarly, the inider will bid. 8

20 , G, G ( ) [ ( ), (, )] H, H G G H Min q C D q D C D D Max G G, Max D D, Max D D can bid up to L, G H, G G H,. Financier q ( ) ( q ) ( ). Note that thi value i alway trictly le than i, G H, G either C ( D, D ) or C ( D, D ). Intuitively, the aet i alway valued more in the hand of people who are capable of producing cah flow. Therefore, financier can only acquire the firm if neither the incumbent nor indutry inider can raie ufficient liquidity. Interetingly, the reaon indutry inider may not be able to raie a much liquidity a the financier i becaue inider uffer from moral hazard in etting pledgeability, while the financier doe not he only want to increae the ale value of the firm at date ince he can produce nothing from running it. Let max{, ( )} be the minimum bid the incumbent will face in the date- min, L, H, auction, and max{, ( )} be the maximum bid the incumbent will face. Following the max, L, H, cae analyzed in Section., we arrive at Lemma. below. The payoff function are omitted for implicity. Lemma. Let (i) (ii) (iii), If If If, Max min,, D,and. i, min, Max, max, and, D, and min, i, min,, Max, PayIC, D D, and max, min, max, min, max, and.. Proof: Along the ame line a the proof to Lemma.., hence omitted. The cae in Lemma. are thu imilar to thoe in Lemma., with ome mall difference. Cae (i) include three ubcae: (a) There i no potential underpricing at date o that ; q C C ; (b) There may be realized underpricing at date, but min, G ( ) ( ) L, H, H, o that indutry inider are contrained in how much they can raie, and thu, they are outbid by financier at date (whoe bid i unaffected by increae in pledgeability). (c) There i realized underpricing at date becaue ( ) ( ) q C C, but the L, H, min, H, max, G incumbent cannot raie the inider bid at date by etting higher becaue the inider i already able to pay for all the limited rent he can appropriate, and will not raie her date- bid. In all three ubcae, high pledgeability doe not increae the bid that the incumbent face at date, o low pledgeability i choen. 9

21 Cae (ii) and (iii) are identical to thoe in Lemma.. Cae (ii) include a imilar cenario to cae (ii) in Lemma.: the incumbent can retain control (we break tie in her favor) even though he tart with le liquidity than indutry inider, imply becaue both have ufficient liquidity (and can raie fund) to bid for all future rent when high pledgeability i choen. The intereting difference at date i that a financier may acquire control without any capacity to produce cah flow over period if ( ) ( ). Intead, he make the L, H, H, firm more pledgeable over the period. The likelihood of thi happening i particularly acute when H, indutry liquidity i low (low ) and period- moral hazard over pledgeability (choice of ) i high, o indutry inider or the incumbent cannot raie much finance (low D ). When aet move into the control of thoe who cannot produce output, economic recovery i delayed until pledgeability i retored and aet returned to the control of indutry inider, regardle of the movement of the underlying economic tate. While the hift in aet to the outider or financier i inefficient in the ene that outider cannot produce cah flow with the aet (and total urplu i not maximized), they can retore pledgeability of the firm. Anticipating retored pledgeability, and thu higher eventual acce to finance, initial bid at date - may be higher. If thee higher bid are beneficial, for example to permit a minimum quantum of invetment to be raied, then temporary outider control i contrained efficient. Outider control i alo reminicent of leveraged buyout tranaction (ee, for example, Jenen (997)), where firm in table indutrie (where moral hazard over pledgeability i high) are taken over, and the revamped management team, which i motivated by the propect of going public oon, focue on finding free cah flow that ha been eaten up either through inefficiency or miappropriated by taff (the proverbial company jet). The management team doe not really make fundamental change to the firm earning propect in the time the firm i private, but it ignificantly enhance the pledgeability of future cah flow, thu enhancing bid for the firm when it goe public. Our model ugget that the leveraged buyout i a mean to check moral hazard at a time of moderate to low liquidity, a oppoed to outright takeover, which are more likely when liquidity i higher. Example: Suppoe the parameter are a follow: G q =.8, GG q =.9, G q =., H =.7, =.,.6, =.3, C = C = C = H, =., =, =, HG,.

22 In thi example, if the incumbent in period were a high type, he loe control at date in tate G, but not in any other tate. In tate other than G, he uffer from moral hazard problem and can raie at mot D H,, which i ignificantly le than PayIC,. A a reult, an indutry inider H, cannot raie much at date in tate againt the promie of date payment and H, =.46. Moreover, ince her wealth i zero, a financier who doe not uffer from moral hazard iue in etting pledgeability can raie a much a L, =.6, and can outbid her. Therefore, if the period- incumbent ha to auction the firm either becaue required payment i high or he loe capabilitie, a financier will acquire the firm. A a reult, the recovery in cah flow i delayed, even if the tate improve..6. When Pledgeability i Choen Thu far, the incumbent et pledgeability after the tate in period i already realized (ex-pot choice). Thi reflect hort term attribute of pledgeability which can be changed rather quickly (uch a a more reputable accountant). Now let u ee what happen when the incumbent chooe pledgeability baed on the probability ditribution of the tate, before the tate for the period i known. Thi ituation repreent more durable pledgeability choice uch a the pecificity of the production technique or the internal organization of a firm, and implie rigidity of pledgeability acro future tate. The point of thi ection i to how that with tate-contingent contract, there i little difference between the two type of pledgeability, which i why we chooe the impler analyi with hort term pledgeability to illutrate our point. Durable pledgeability choice will be more important with debt contract. Figure 4 below how the timing. The incumbent make a deciion before the tate i realized baed on the probabilitie of each tate. If the cot, ε, i ufficiently mall, there i no effect on real outcome. High pledgeability i choen ex-ante if the incumbent had the incentive to chooe high pledgeability ex-pot in at leat one of the two ubequent tate. Figure 4: Timing and Deciion with ex-ante Pledgeability Choice

23 Next, we analyze what happen when i ignificantly poitive. If the incumbent had the incentive to chooe high pledgeability in both the ubequent tate when he wa making the choice ex pot obtaining knowledge of the tate, then he would chooe high pledgeability ex ante. The tatecontinent payment would be identical (becaue the payment et when choice i ex-pot give an incentive to chooe high pledgeability in both tate). However, if liquidity i o high in one of the tate (ay G) that there i no potential underpricing (or potential underpricing of le than ε), and hence there i no incentive to increae pledgeability merely anticipating the outcome in that tate, then there mut be a lower tate-contingent payment in the other tate (ay ) o that the incumbent ha ufficient rent to cover the cot of chooing high pledgeability before the tate i known 6. Thi mean the incumbent ability to raie funding will (weakly) fall relative to when pledgeability i choen ex pot if the cot ε i ignificant. If the probability of the fully liquid tate and the cot ε are both ufficiently high, the incumbent may even chooe low pledgeability ex ante, for it may not be worthwhile to lower the promied payment enough in the unlikely other tate to give her the incentive to incur cot ε. roadly, however, in the baeline cae of our model where ε i mall and contract are tate contingent, the timing of pledgeability choice i not very important, o both quickly changeable apect and durable apect of pledgeability can be imilarly incentivized. Thi i not the cae with debt contract, which we now turn to. III. Debt Contract Our analyi of tate-contingent contract erve a a building block to undertand the effect of indutry liquidity and promied payment on pledgeability choice. However, financial contract ued by mot firm are le than fully tate-contingent and are much cloer to debt contract, which pecify a contant promied payment on a given date in all tate uch that D t D. In thi ection, t we tudy uch debt contract and focu on how they can limit pledgeability. We will continue to aume that pledgeability in a period i choen before the tate in that period i realized (for reaon we will explain at the end of ection 3.) and perit over the next period we therefore focu on durable apect of pledgeability conitent with the financing cycle. For implicity, we do not add explicit friction to make debt the optimal contract, uch a cot of verifying the tate. 6 The maximal payment in tate i D Mx, a q q G G.

24 With tate-contingent contract, no one ha an interet to overpromie, given the potential damage to both incumbent and financier from inducing low pledgeability and increaing underpricing. With debt contract, however, if the incumbent mut raie a large amount initially (to fund the invetment or to outbid other to become the new incumbent), it i poible that thi amount can only be raied by making a high fixed promied payment acro tate, even if it lead to low pledgeability (which will be inefficiently low in ome tate). Section 3. formalize the analyi of period with debt contract and pledgeability et ex ante. With tate-contingent contract, we have hown that both pat and current tate affect the equilibrium outcome. With debt contract, we will how that expectation about future tate, and thu the pillover between future tate, alo affect pledgeability choice, aet allocation and financing capacitie. In Section 3., we conider period. Similar to the cae with tate-contingent contract, at date the aet can be old to a financier who ha no production capabilitie. With tatecontingent contract, the aet i old inefficiently only when indutry liquidity i very low and future moral hazard i at very high level. With debt contract, however, uch inefficiency occur even at moderate level of indutry liquidity becaue the low prior pledgeability choen in the face of debt will make it hard for current incumbent or indutry inider to raie finance. Thi further ugget why recoverie from debt-fueled, aet-price-baed expanion are low. 3. Debt Contract with Ex-ante Pledgeability Choice ecaue there i a ingle tate in period, the promied payment when contract are retricted to debt contract will be identical to that when contract are tate contingent. Next, we turn to period. The timing of event in the period i identical to Figure 4, with the exception that the promied G i, i, payment i a contant, D D V D V D repectively be the. Let, and, incumbent payoff when he chooe high and low pledgeability, given reidual required payment D. Define i, i, D V D, V D,. In the baeline model with tate- contingent contract and ex-pot pledgeability choice, the maximal promied payment D, Max, Max D if there i potential underpricing. With a contant payment and ex-ante atifie choice, the expected difference in payoff mut be non-negative to provide incentive for high pledgeability. Lemma 3. decribe D for any level of D and for low level of. Lemma 3. (i) If ( ) C, D. H, 3

25 (ii) If (iii) If C and H, ( ) C and H, ( ), D ( ) ( ) i, H, ( ) ( ), i, H, D PayIC if D D PayIC if D D. (i) (ii) (iii) Figure 5: D in Different Cae The detailed expreion for Lemma 3. are available in the Appendix. Figure 5 characterize the D in different cae. If there i no potential underpricing in tate, function D for all value of D becaue raiing pledgeability doe not change enforceable payment, while reulting in cot. If there i potential underpricing, however, and if the incumbent, Max D for D D, negative for higher value of D can outbid outider (cae (iii)), 4

26 and trictly poitive for all lower value of D. Higher committed payment depre the incentive to increae pledgeability, ince the incumbent retain control whenever he retain her ability. Finally, if there i potential underpricing and the incumbent ha no hope of retaining control once he enter an D : the incumbent ee only the upide of increaing pledgeability auction (Cae (ii)), ince the aet invariably will be old. Even for very high promied value of D above the mot the aet could be old for, the only diadvantage of chooing high pledgeability i it cot,. G With one ingle face value D D D and ex-ante pledgeability choice, there i a ingle incentive contraint acro tate. In particular, there exit a unique, IC D that atifie,, q G [ G ( D IC C )] ( q G )[ ( D IC )] uch that high pledgeability i choen if and only if the, IC face value of debt D D. In general,, IC, Max D lie in the range D GM, ax and D C. Lemma 3. Two cae, ummarized by Lemma 3., deerve pecial mention. (i) If there i no potential underpricing in tate G, D D (ii), IC, Max (the lowet poible value) a. High pledgeability i choen only if promied payment are et low. If the incumbent ha no hope of retaining control in tate in any auction after a payment, IC G, Max default, D C D (the highet poible value) a. A payment that lead to high pledgeability i alway preferred. Intuitively, in Lemma 3. (i), there i never any incentive to increae pledgeability coming from G D for all future tate G, o the incentive have to be et via tate alone. That i, poible payment. The maximum payment which till provide incentive for high pledgeability in tate i well below the mot that can be paid in tate G. In Lemma 3. (ii), in future tate, the incumbent will trictly prefer high pledgeability regardle of the ize of the promied payment (provided he recover cot D for ), becaue he know he ha to ell. That i, all promied payment. With no diincentive from tate, there will be ex-ante incentive to increae G D ). pledgeability whenever there are ex-pot incentive in tate G (i.e., whenever An important new reult i that, unlike with tate-contingent contract, D, IC, the level of debt which provide incentive for high pledgeability keeping in mind both future tate, may not be the face value that enable the incumbent to raie the mot upfront. Thi i mot eaily een when 5

27 liquidity i plentiful, a in tate G with no potential underpricing. In thi cae, the incumbent can pay ( ) C C C in tate G even if pledgeability i et low. At the ame time, the HG, anticipated high liquidity in the G tate offer no incentive for the incumbent to enhance pledgeability. Thi alo mean that the level of debt that provide incentive to et pledgeability high, D, Max, weakly le than D, and trictly le if. If the difference between C C and Max, D HG, i large, the incumbent could raie more by etting D ( ) C and inducing low pledgeability even if there i ignificant probability of the low tate occurring. The broader point i that the propect of highly liquid tate not only make feaible greater promied payment, but alo eliminate incentive to enhance pledgeability. To retore thoe incentive, keeping in mind the other tate, debt may have to be et o low that fund raied are greatly reduced omething the incumbent will not do. Note that thi can happen even if the probability of the low tate i ignificant, and even if the direct cot of enhancing pledgeability i zero. HG, More formally, if ( ) C exceed, IC HG, D, etting promied payment at ( ) C lead the incumbent to chooe low pledgeability, and pay IC, i H, HG, in tate and ( ) C in tate G. Thi can provide a larger expected payment if G HG,, IC G H, H,, IC q ( ) C D ( q ) ( ) min ( ), D. () Lemma 3.3 then decribe D, Max correponding to the cae in Lemma 3. for. Lemma 3.3 When,, the level of promied date- debt that can raie the mot,,, (i) If H i,, IC, PayIC ( ) ( ) C and ( ) C, then D D. a) If HG G G H,, PayIC q C C ( q ) ( ) D, then, Max D C C. For any PayIC,, Max promied payment D D D,. For any promied payment D D, PayIC, b) If., Max, IC, PayIC q C C ( q ) ( ) D, then D D D. G G H,, PayIC For any promied payment, Max D D,. 6

28 , ig, HG, i, H, (ii) If ( ) C, ( ) ( ), and ( ) ( ), then HG D D C D. For any promied payment, Max D D,, Max, IC G, PayIC. In ummary, high anticipated liquidity thu crowd in debt and crowd out pledgeability, etting the tage for more evere aet miallocation than in the cae with contingent contract. Interetingly, debt will not be renegotiated, before or after the tate i realized, even if renegotiation i feaible it will not be renegotiated before becaue the level of debt i et to raie the maximum amount poible even if it reult in low pledgeability, and will not be renegotiated after becaue relevant partie will not write down their claim given that pledgeability ha already been et. Interetingly, both the fixed promied debt payment acro tate, and the act of chooing pledgeability before the tate i known, have the effect of cauing a pillover between anticipated tate. Therefore, outcome are omewhat imilar even when pledgeability i choen after the future tate i fully known, but in that cae one need to explain why the level of debt payment cannot be renegotiated once the tate i known. Specifically, when the tate i known and the lender get all urplu from renegotiation, debt become equivalent to fully tate-contingent contract, in part becaue all are rik neutral. The analyi of nonrenegotiable debt when pledgeability choice i made after the tate i realized i available from the author. 3. Date Choice and Prolonged Recoverie Let u now analyze period with debt contract. With minor notational complication, the pledgeability deciion in period i imilar to that we have jut examined for period. We jut aw, however, that debt contract could induce low pledgeability becaue of cro-tate pillover. In other word, before the tate in period i realized, high liquidity in one anticipated tate (tate G) could induce both high promied debt payment at date and low pledgeability ( ). Financier may now outbid indutry inider at date in tate imply becaue the latter have little ability to pledge. Such miallocation could occur even at moderate level of liquidity in tate. Thi pillover in outcome between anticipated tate i a pecial property of debt. We explain all thi in more detail. i, With a bit of abue of notation, let C H, ( D ) and C ( D ) be the date- expected continuation value of the aet to the incumbent and indutry inider if the face value of debt i D. Unlike with and G tate-contingent contract, default i now poible on committed payment. So let D D D D be the amount that financier recover in the two tate repectively, given the face value ig C ( D )= q C D D V ( D ) ( q ) D D V ( D ) D. i, G G, G i, 7

29 H G C ( D )= q C D D V ( D ) ( q ) D D V ( D ), H, G G, G H, i, L, H, ( ),, and ( ) are then: ( ) [ ( ), ( )] i, i, G G G, Max i, Max D D Min q D D q D D C D L, G H, G G H, q ( ) ( q ) ( ) H, H, G G G H, Max Min[ q D D q D D C D, Max D D ( ) ( ), ( )]. IC Similar to date, we can define G G IC G IC D uch that q D C q D. Lemma 3.4 i analogou to Lemma 3.. The definition for.5. Lemma 3.4 and max, follow thoe in Section min, i) If o that there i no potential underpricing in tate G, D max, min, D IC, Max. IC ii) If the incumbent ha no hope of retaining control in tate, D C D. G, Max We now highlight one intereting cae which only occur in the dynamic context when contract are retricted to be debt. Lemma 3.5 a a If ( ) ( ) and H, L, H,,,, q G H G ( ) C D IC ( q G ) L mi n H ( ), D IC, then D ( ) C D D Max H, G IC, Max IC Max. For any promied payment D D D, For ufficiently high probability of tate G in period, and ufficiently high liquidity in tate G, the incumbent will et date- debt high enough that he chooe low pledgeability for period. If the realized tate i G, he repay everything. If the tate turn out to be, he i forced to ell the aet and the preet low pledgeability,, retrict the amount at which he can ell, which equal 8

30 a ale to financier. If he had et pledgeability higher, he would have old to an indutry L, inider. The additional miallocation in an indutry but tem wholly from the anticipated high liquidity in the boom, which caue the incumbent to both promie high debt payment for date, and induce her to chooe low pledgeability in period. When the expected boom (high turn out to be a but, we have a prolonged recovery due to miallocation. G q ) intead Note that the incumbent deciion i completely rational: he know the probability of a but but becaue the debt level ha to be kept very low to induce high pledgeability, he rationally ignore the conequence, even if the probability of the tate i not low. 7 y contrat, if pledgeability i elected after the tate i realized, renegotiation would eliminate thi type of miallocation ince expot, it i in the joint interet of the incumbent and financier to reduce debt face value and retore incentive for high pledgeability. Example: H, We have hown that in the cae with tate-contingent contract when =, the aet i old to an indutry outider in tate at date, ince the indutry liquidity i very low. Now we H, increae to. and how that even at thi moderate level of indutry liquidity, the aet i miallocated in the ame tate if and only if debt contract are ued. Equally important, the aet would not be miallocated even with debt contract if liquidity in the good tate were lower. Higher anticipated liquidity can induce more miallocation via the increae in debt! Conider the following parameter value: =.6, =.3, C =, C = C =, H, HG, H, ig,, =.7, =., =., =., =.5, HG, H G GG G =., =.7, =, q.9, q.9, q.5,., ) State-contingent contract Suppoe If ig, H, G i, H, = =.6, then =.9> =.5 and =.4= ig, HG, i, H, = =.3 intead, => =.5 and =.4= cae, the incumbent i able to retain control in both tate G and. Therefore, =.59 and D = D, Max, PayIC =.49. At date, financier (L type) bid up to L, i =.4. =.4. In both D = D G, Max =.75 while G, PayIC 7 The pillover between tate via debt induce lower pledgeability than would be een with tate-contingent contract. If there i ignificant underpricing in tate G and high pledgeability increae bid ufficiently to cover it cot, high pledgeability i choen with debt contract even if in tate there i no value to high pledgeability, for intance becaue the aet i old to indutry outider. Neverthele, thi i not an example of over-pledging induced by debt; even with tate-contingent contract, we would have the ame effect of a high choice of pledgeability the eemingly exceively high pledgeability in the tate, where it i of little ue, tem from the fact that the pledgeability deciion i made before the tate i known, and not from the fixed debt contract. It i in thi ene that while debt induce under-pledging, it doe not induce over-pledging. 9

31 H, indutry inider bid i, =.84 and if the tate i. ) Debt contract H, =.94 and i, =.79. Finally, the incumbent bid =.69. Therefore, the incumbent alway loe control to an indutry inider at date Now conider debt contract. Since the incumbent can retain control in both tate G and, D = HG, =.58 and the incumbent can raie more by etting C =. which raie.75 in expectation. In thi cae, the retriction to debt contract further contrain the amount that can be H, raied at date. It turn out therefore that type are unaffected: L, H, =.85 and D =.7. However, low =.75. There i no potential underpricing in tate G, GG and G. Since =.9 i very high, date- debt i et high and the incumbent et in period. If the tate i realized at the end of the period, the aet i old to an indutry outider. Such a reallocation could have been avoided if have required iuing le debt. So debt lead to low pledgeability and greater miallocation. IC, had been et at, a in the tate-contingent contract example, but that would 3) Miallocation induced by high expected liquidity Suppoe now that date- liquidity were lower, o tate G, GG and G. Therefore, D HG, PayIC, =.5. There exit potential underpricing in C D =.745 and would be choen. A a reult, the aet i correctly allocated to an indutry inider if tate i realized. So lower liquidity lead to debt with a lower face value, which in turn lead to higher pledgeability and a more productive allocation of control in downturn. More generally, anticipated liquidity operating through greater leverage caue the advere pillover between tate which puhe down pledgeability and caue miallocation. IV. Dicuion and Empirical Relevance In a boom which i likely to continue, liquidity i high and upport high debt. When borrower finance with uch high debt, however, they will chooe low pledgeability, which neverthele will be acceptable to lender who anticipate a high probability of continued high liquidity. Liquidity, aet price, and leverage follow each other up, while pledgeability fall. If the boom doe not continue, and liquidity fall, acce to finance will drop ignificantly. Outider are alo more likely to take over the firm at uch time and recoverie are likely to be low. The overhang created by liquidityinduced leverage on pledgeability cannot be renegotiated away. Importantly, all thi can occur even if the probabilitie of the downturn are not inignificant. Higher anticipated liquidity i therefore not an unmitigated bleing, and can generate wore outcome in le liquid realized tate. To the extent that government or central bank policie create anticipation of liquidity, thi i a concern that ha to be kept in mind. G q 3

32 In contrat, when lower liquidity i anticipated (in normal or bad time), a choice of low pledgeability will ignificantly reduce the amount financier expect to recover. Therefore, the competitive financial market will prevent exceive leverage (in term of it effect on pledgeability) in time of lower liquidity becaue it want incumbent to retain incentive to et high pledgeability. Importantly, all thi can occur even if the probabilitie of downturn are not inignificant. Loan contract with many covenant could be a proxy for the choice of high pledgeability. In bad to normal time we hould ee many covenant and relatively low level of leverage when freh capital tructure are choen (uch a when the firm come out of bankruptcy). In contrat, during boom we will ee higher leverage and few or no loan covenant ( covenant lite ). Indeed, Chritenen and Nikolaev (4) eparate loan covenant into thoe that are capital or balance heet related (uch a contraint on overall leverage) and thoe that are hort-term performance or incometatement related tripwire uch a return on aet ratio. Interetingly, they find that the number of balance heet related covenant in loan agreement came down ubtantially in the year of high liquidity before the financial crii in 7-8, only to rie during the crii and oon after till. In contrat, they find the number of performance related covenant remain fairly contant over the year. Arguably, thi ugget lender were more open to higher borrower leverage and lower aetrelated contraint during the boom year of high liquidity only to tighten once condition deteriorated. oom period with covenant lite loan and high leverage could alo be interpreted a an increae in the fraction of market finance (bond or covenant lite loan) a oppoed to intermediated finance (covenant-intenive bank debt). Low pledgeability could alo be caued by choice of lower accounting quality or weaker corporate governance. Computat report the auditor' opinion of the effectivene of the company' internal control over financial reporting while auditing a company' financial tatement, an opinion which i mandated by ection 44 of the Sarbane-Oxley Act. A material weakne i a deficiency, or a combination of deficiencie, in internal control over financial reporting, uch that there i a reaonable poibility that a material mitatement of the company' annual or interim financial tatement will not be prevented or detected on a timely bai. When an auditor indicate a material weakne, it ignifie a previouly undetected choice to degrade accounting quality which may not yet have influenced accounting report, and can thu erve a a meaure of low pledgeability. Figure 6 below indicate that the percentage of Computat firm with material weakne of internal control a meaure of ubtandard accounting tarted to increae in the extremely liquid period before the financial crii, fell after the onet of the crii, and tarted to increae again a central bank around the world maintained extremely liquid condition in financial market. 3

33 Figure 6 Weakne of Internal Control Note: thi figure plot the erie of percentage of firm that were reported a with weak internal control. The data i obtained a the variable AUOPIC from the Computat Annual databae. Thi dummy variable i et to for firm reporting an internal control deficiency, i.e. a material weakne in the client internal control ytem, in the retatement year and/or the two ubequent year, otherwie the variable i et to zero It may be ueful here to ee the difference between our model and the eminal work by Shleifer and Vihny (99) (henceforth SV). They focu on liquidity varying over time. SV emphaize control right excluively through aet ale while we introduce control right over cah flow through the pledgeability channel, which itelf uffer from moral hazard. Our model therefore ha different implication than Shleifer and Vihny (99). A in SV, aet migrate in our model to agent who have lower ability to manage. However, the underlying rationale i different. In SV, aet get inefficiently allocated becaue highly ability manager have le liquidity than outider. Debt, which wa created to reolve a free cah problem, ha an overhang effect which limit the amount of liquidity obtainable by indutry inider. Therefore, if financial contract were tate-contingent (or if debt could be renegotiated), the aet would never be old to outider. In our model, the aet goe to low type preciely becaue they do not uffer from the moral hazard over pledgeability and not becaue they have more liquidity. Indeed, financier are unwilling to renegotiate debt down becaue they know the aet will be old to outider who can pay more by making the aet more pledgeable even when burdened with high debt. Our paper hare imilar inight with a equence of paper by Geanakoplo (for intance, Geanakoplo ) on leverage cycle which are analogou to our financing cycle. Like u, Geanakoplo endogenize the borrowing contraint, though by a different approach. In particular, he et up a general equilibrium model in which agent have heterogeneou belief. Therefore, 3