1994 Insider Trading and the Managerial Choice Among Risky Projects

7/30/2019 1994 Insider Trading and the Managerial Choice Among Risky Projects

1/14

JOURNA L OF FINA NCIAL A ND QUANTITATIVE ANALYSIS VO L 29 , NO 1, MARCH 1994

Insider Trading and the Managerial Choice amongRisky ProjectsLucian Arye B ebchuk a nd Chaim Fershtman*

AbstractThe concern of this paper is with the effects of insider trading on ex ante managerialbehavior. Specifically, the paper focuses on how insider trading affects insiders' choiceamo ng investment projects. Oth er things equal, insider trading leads insiders to choos eriskier investment projects, because increased volatility of results enables insiders t o mak egreater trading profits if they learn these results in advance of the market. Th is effect mightbe beneficial, however, because insiders' risk aversion pulls them toward a conservativeinvestment policy. Insiders' choices of projects are identified and comp ared w ith insidertrading and those without such trading. Using these results, the conditions under whichinsider trading increases or decreases corporate value by affecting the choice of projectswith uncertain returns are identified.

I. IntroductionThe managers of a corporation may well wish to buy or sell shares of their

company. Th e legal rules of the United S tates, as well as those of other advancedma rket econ om ies, substantially limit-but do not proh ibit or prevent-such trad-ing by corporate insiders. Th e extent, if any, to which such trading by insiders isharm ful and sho uld be constrained has long been the subject of active and intensepublic debate.O ne possibly impo rtant effect of insider trading is on the ex ante managem entdecisions of insiders. Given the general recognition that manage rs cann ot alwaysbe induced to act in the shareholders' interest, it is natural to ask whether thepossibility of trading leads insiders to make m anagem ent decisions that are closerto, or further away from, the value-maximizing decisions. While the law review

*Bebchuk, Haward Law School, Harvard University, Boston, MA 02163 and the National Bu-reau of Economic Research. Fershtman, Department of Economics, Tel-Aviv University, Tel-Aviv69978, Israel. The authors thank Howard Chang and Christine Jolls for helpful research assistance,Patricia Weber, JFQA Managing Editor Jonathan Karpoff, and JFQA Referee and Associate EditorRobert Heinkel for helpful comments and suggestions. Financial support for Bebchuk's work has beenprovided by the National Science Foundation and by the John M. Olin Foundation. Partial financialsupport for Fershtman's work has been provided by funds granted to the Foerder Institute for EconomicResearch by the Nur Moshe Fund.


2/14

2 Journal of Financial and Quantitative Analysisliterature is full of informal assertions a nd speculations con cernin g this que stion,'the eco nom ic literature has thus f ar devoted little attention to it.

Mo st of the substantial work d one on insider trading in recent years has beendevoted to modeling the insider trading process itself. Re searc hers have examinedhow the possession of inside information enables insiders to make profits andhow this information eventually comes to be reflected in the market price.2 Onlya few papers allow the insiders to make corporate as well as trading decisions.Dye (1984) considers whether shareholders can draw useful information fromthe man agers' trades, assuming that these trades are observable. Giam marino,Heinkel, and Hollifield (1992) examine a model in which managers who makecorporate decisions are allowed to trade and thus have incentives to mislead themarket through corporate announcements; they demonstrate that, in some cases,managers manipulate corporate actions in an opportunistic fashion. Bagnoli andKhan na (19 91) develop the intuition that anonym ous managerial insider tradingeliminates the incentives to truthful revelation of inform ation . Fischer (199 2)considers insider trading by individuals who obtain inside information about aparticu lar firm as a result of providing som e services to the firm. Su ch insidertrading may aggravate existing agency problems and hence justify a completeprohibition of su ch agent trading.This paper analyzes the effect of insider trading o n manag ers' choice amo ngrisky investment pro jec ts3 In particular, we com pare the project choices thatinsiders make under contracts that allow insider trading with those they makeund er contracts that prohibit such trading. Th e main point of the paper is to showthat insider trading is not necessarily harmful and can be a part of the optimalcompensation scheme.

Und er contracts that prohibit insider trading, man agers' choice am ong riskyinvestment projects may b e inefficient f or a familiar reason. Un like the sharehold-ers, who can diversify, ma nage rs' attitude toward a risky project's results is likelyto be characterized by a significant degree of risk aversion. To the extent that themana gers' salary is made depen dent upon the firm's profits (to induce m anagerialeffort), the managers will be too conservative: they may cho ose a safer projecteven if it offers a lower ex pecte d return.

Und er contracts that allow insider trading, managers look m ore favorably onrisky projects. T he reason fo r this is that, to the extent that ma nag ers learn ahead ofthe market how uncertainty is resolved, greater uncertainty enables them to makegreate r trading profits. Th us , the possibility of insider trading will induce themanagers to accept some desirable risky projects that would be rejected withoutinsider trading. An alternative interpretation is that the existen ce of insider trading

'See , for exam ple, Carlton and Fischel (1983), Easterbrook (1985), and Scott (1980). The m ainclaim o f this literature is that insider trading should not necessarily be prohibited a s allowing managersto insider trade can sometimes be a part of the managerial optimal com pensation schem e. Our papercan thus be viewed a s a formalization a nd analysis of the claims in the law review literature.2 ~ e e ,or example, Glosten and Milgrom (1985), Kyle (1985), Laffont and Maskin (1990), andMirman and Samuelson (1989). Recent papers by Ausubel (1990 ), Manove (1989), and Fishmanand Hagerty (1992) have analyzed cenain important ex ante effects of insider trading (on investmentdecisions and information acquisition), but these papers abstract from the agency problems on whichour project focu ses.3 ~ e b c h u knd Fershtman (1992) analyze the effect of insider trading on mana gers' effort level, andBebchuk and Fershtman (1991 ) focus on effect on mana gers' incentives to "waste" c orporate value.


3/14

Bebchuk and Fershtman 3creates an option for the mana gers, the value of which is enha nced by increasedvolatility of the underlying cash flow to the firm.Allowing insider trading , however, also involves certa in disadvantages. First,the desire to increas e trading profits might lead the ma nagers to prefer a very riskyproject even if it offers a lower expected return than a saf er alternative. Further-mo re, compensa ting manag ers with trading profits of uncertain size increases therisk that they bear an d thus requires an increase in their total expected c omp ensa-tion. Th e paper identifies the conditions under w hich these effects domina te thebenefit from allowing insider trading an d the conditions under w hich the converseis true.It is imp ortant to note that on e can mimic the man agerial comp ensation thatallows for insider trading with com pensation s that prohibit such trading but requiremanagers to hold risky securities that give returns identical to the insider tradingreturns. Th us , the main interpretation of ou r result is not the fact that managers canbe ind uced to ch oose risky investment optimally, but the observation that insidertrading is not necessarily harmful and can even be used to correct some of thedistortion created by the manag er's risk aversion.

In assessing the importance and relevance of conclusions about insiders'behavior in the presence of insider trading, it is important to recognize that theworld in which we live features a significant amo unt of such trading. Th e lawdoe s not totally prohibit insider trading. Th e law includes an absolute prohibitionon s uch trading only wh en the insider buys an d then sells (or sells and then buys)within a six-month period. W hen insiders do not go in and out of the company 'sstock during a six-m onth period, the law co nstrains their trading only when it canbe show n to be based on "material" inside information. Becau se insiders' motivefor trading is often not o bservab le or not verifiable, they often can openly mak eprofitable trades (as the eviden ce indeed indicates, see for exam ple, Jaffe (1974 )and Seyhun (1986)). Furtherm ore, insiders may hide the trading itself; muchtrading by insiders m ay w ell go undetected.Th e amo unt of insiders' trading profits is a function of both the strictness ofthe legal and corporate arrangem ents governing such trading and the expend itureson enforce men t. T hus, findings on the consequ ences of insider trading have bothnormative and positive implications. From a normative perspective, they are rele-vant for determ ining how much (if any) insider trading should be curtailed. Froma positive perspective, given that m uch insider trading takes place at present, suchconclu sions are necessary for a full understanding of actual insider behavior underthe existing legal regim e.

II. The ModelTh e sequenc e of events in the mo del is as follows. In period 0, the firm is

forme d and the manage rial contract is specified. In period 1, the managers cho osean investment project. In period 2, the managers get advan ce information abou t theproject's results a nd, if the man agerial contract allows them to d o so, they use thisprivate information to participate profitably in the trading that takes place duringthis period. In period 3, the final period, the project's results are realized. Ou rassum ptions concerning each of these elem ents of the mo del are described below.


4/14

4 Journal of Financial and Quantitative AnalysisPeriod 0: Contract Specification. A firm is formed, and the managers and theshareholders spe ci fy a contract. T he contract provides the managers w ith a salarythat increases in the firm's final output. W e focus on schem es that are linear in thisoutput, denoted b y w . ~hu s, the contract specifies some S and ct. E [0,11, andthe salary scheme is C ( W )= S + a W . In addition, the contract spec ifies whetherthe managers are allowed to trade in the firm's shares. W e shall re fer to contractsprohibiting insider trading as NT contracts and to those allowing such trading asIT contracts . In the case o f an IT contract, we will den ote by I7 the insider tradingprofits managers will make. The initial value o f the firm is denoted by Vo andis endogenously determ ined, depending on the managerial contract and the set o fpossible projects available to managers. The main question to be considered in thispaper is how NT and IT contracts d if fe r n their e ff ec t on managers' subsequentproject cho ice and the initial value Vo.Period 1:Project Choice. The managers choose an investment project. The choiceis between a project that will produce final output W = Wo + 8 ("project zero"),where Q is a random variable such that E ( Q )= 0 , and a riskier project, denoted b y( r R , R ) , that produces final output W ( r R , R )= WO+ Y R + ( 1 + e R ) Q . R is assumed tobe positive , which m akes the project r iskier, wh ile r~ may be positive or negative.For simplicity, we assum e that Q can be eitherm or -m , each w ith probability112, for some m > 0. This assum ption will s im plif y our calculations considerably,allowing us to focus our analysis on the e ffe ct s o f insider trading.Th e shareholders can observe the managers' choice o f project, and the firm 'svalue in period 1 will reflect this observa tion. The managerial choice is not ve rifi-able by courts, how ever , and for this reason the managerial com pensation cannotbe made contingent on it.

The value o f the firm's shares in this period is V1. W e assume that the twoprojects between which the managers choose in period 1 are known in period 0.Given this, the managers' choice in period 1 will be anticipated in period 0 andthus reflected in Vo. Therefore,V1 = Vo.Period 2: Trading. At the beginning o f this period, the managers (but not others)learn Q and thus know the final value o f the firm , V'. Subsequently, trading inthe firm's shares takes place with the following participants: liquidity-m otivatedsellers, a (z ero-p rofit)market maker (specialist)who sets the price, and the insiders,i f their contract so allows. Th e liquidity-motivated sellers are some o f the initialshareholders who cannot defer realizing the value o f their shares until the finalperiod; each o f the initial shareholders is assumed to face the same probabilityo f needing to liquidate his hold ings during this trading period. For our purposes,there is no reason to model the trading process, as it has already been analyzedin detail by Kyle (1 98 5) and by Glosten and Milgrom (1985),and we can rely ontheir conclusions. Th ese authors have shown that the insiders will be able to makesome profits, but will not be able to capture the full gap between the pretrading

4 ~ eocus on linear schem es for the sake of tractability. For a similar assumption in a similarcontext, see Holmstrom and Tirole (1990). For an analysis of the conditions under which linearcontracts are optima l, see Holmstrom and Milgrom (198 7).


5/14

Bebchuk and Fershtman 5value and the final value of the firm.5 C learly, if there are no insider trading profits,then NT and IT contracts are equivalent.

Relying o n the conclusions of the existing insider trading models, we assum ethat the insiders can capture som e fraction of the difference betwee n the pretradingvalues V1 and the final value Vf. Specifically, we ass um e that the insiders' tradingprofits are 17 = PIVf - V1 for some P E [O , We let the parameter /3 beexogenously given. The comp arison between the two types of contract will depend,am ong oth er things, on the value of P.Of course, the insiders' trading profits all com e at the expen se of the liquidity-motivated sellers, as the market m aker is assumed to mak e zero-expected profit.Thu s, because each of the initial shareholders faces the sam e ex ante probabilityof having to liquidate his holdings in period 2, the initial shareholders expect tobear the costs of the insiders' trading profits as much as they expect to bear thecosts of other elements of the insiders' compensation scheme.Period 3: Realization of Output. The final output W s realized, and the salaryS + aW s paid to the ma nagers. Th e final value of the shares is thus Vf = W -C(W) ( 1 - a)W- S.Th e curtain now goes down and the firm dissolves.A. The Managerial Labor Market Cons traint

In designing the contract, the following aspects about the managers mustbe taken into account. Since only shareholders can spread the risk amo ng manyfirms, we assume that, while the shareholders are risk neutral, the managers arerisk averse. Specifically, we assum e that man age rs' exp ected utility from randomreturn x is given by

where o(x)s the stand ard deviation of x, nd 0 < y < 1 is a given constant.'We further assume that the manag erial labor market is perfectly com petitive.

Let (? be the man agers' reservation wage. The managers' participation constraint5 ~ h i l ehe rich structure of these models can not be summ arize d briefly, the following points maybe made a bout the basic conclusions described in the text. First, the insiders will be able to mak esome profit because, at least initially, the market maker will not be able to tell for sure whether theinsiders are selling or buying. (The supp ly of shares from liquidity-motivated sellers is assu med tobe a random variable, and the market maker is assumed to obse rve only the net aggregate of ordersfrom liquidity-motivated sellers and insiders; consequently, while the market maker can draw someinferences about the direction in which the insiders are likely to be trading, he canno t know for sur e.)Second, one reason why the insiders can capture only part of the gap between the pretrading value andthe final value of the firm is that, as the insiders trade more shares, their information w ill be increasinglyreflected in the market price. We assum e that the manage rs trade anonymo usly; if they do not, thentheir position co uld be used to convey their private information (see al so John and La ng (1991)).6 ~ ehus assume that insiders are equally able to sell and buy share s and can capture the sam efraction of the gap betw een pretrade a nd final value when the final value is high a s when the final valueis low. Allowing for different /3s for the two case s would com plica te our calculations considerably. Ifonly on e direction of inside trading is possible, then our analysis will collapse, but if insiders' abilityto trade is similar (but not identical) in good and bad states, then we expect that there would not bea change in the nature of our results and that there would be values of for which an IT contract issuperior.7y > 0 is implied by the managers' risk aversion. y < 1 is necessary to guarantee that if theriskier project dom inate s project ze ro, i.e., if W o + rR - 1 + ~ ) m W o + m, then the riskier projectwill be chosen by the managers.


6/14

6 Journal of Financial an d Q uantitative Analysisrequires that their compen sation under the man agerial contract gives them expectedutility greater than or equ al to C.B. The First-Best

Th e initial shareholders wish to m aximiz e the firm's initial value, V o ,whichis the firm's expected output minus the managers' total expected compensation.Clearly, if the managerial actions were all observable and verifiable, the initialshareholders w ould direct the mana gers to choo se the riskier project if and only ifrR > 0 , and would provide the ma nagers a salary equal to c. Consequently, thefirst-best initial value is V: = W o+ max(0, r R )- c.Ill. Project Cho ice and Corporate Value under NT ContractsA. Insiders' Project Choice

Let us first examine the manage rs' choice under a given NT co ntract ( S ,a ) . Inthe absence of trading profits, the manag ers' total comp ensation will be C ( W ( r ,E ) ) ,where ( r , ) is the project chosen by mana gem ent. Now ob serve that

Thus, the ma nagers' expected utility is

Proposition I . Given any NT contract ( S ,a ) , he man agers w ill choose the riskierproject ( r R , R ) iff

Proof. Th e manage rs will choose the riskier project as long as U ( C ( W ( r R ,R ) ) ) >U ( C ( W ( 0 , O ) ) ) . hus, using ( 4 ) , he riskier project is ch osen iff( 6 ) S + a ( W o + r R ) - a ( l + ~ ~ ) m S + a W o - a m .Simplifying ( 6 )yields the condition in (5).

Note that c N T (c R)s strictly positive for every a > 0. That is, the man agersmay not choose the riskier project even if it offers a higher expected return thanproject zero; this will happen if the rise in the ma nagers' expected comp ensationis not sufficient to com pensate them for the higher risk associated with the project( r ~ .R ) .B. Corporate Value

Having identified the man age rs' cho ice, let us now consid er the firm's initialvalue. Assum ing that share holders can ma ke the participation constraint bindingand project ( r , ) is chosen,(7) SN T = C - a (W O+ Y ) + y a ( 1 + ~ ) m .


7/14

Bebchuk and Fershtman 7Thus, given that ma nagers are provided with an NT contract ( S N T , ) ,which impliesthat the initial value of the firm, as a function of the chos en project, is

where

The first expression in (8) is the value of the project and the initial value is thisproject value minus the managers' reservation wage, C, and the additional com-pensation given to them for the risk they are bearing. Notice that V l may be lowerthan the first-best initial value in two w ays. First, the ma nag ers' choice of projectma y be inefficient. Second , because m anagers must be c om pensated for the riskthey bear, the expected com pensation that must be given to them e xceeds C by therequired risk premium (the last term on the right side of ( 8 ) ) .

Note that, because managers must be compensated for the risk imposed onthem, shareholders may no longer prefer the riskier project (with r ~ > 0 ) overproject zero. Specifically:Proposition 2 . i) Given that ma nagers are to get an NT contract ( S N T , ) , share-holders prefer the riskier project ( r R , R ) iff

ii) There are projects ( r R , R ) hat the manag ers will not choose even though theyare preferred by the shareholders.Proof. As shareholders wish to m aximize the initial value V 1 ,hey prefer the riskierproject ( r R , R ) as long as V l r R, R ) > V 1 0,O). Using ( 8 ) , his cond ition impliesthat the riskier project is preferred by shareholders as long as Wo+r ~-C- a ( 1 +eR)m> Wo- C - a m , wh ich, in turn, implies that sh areholders prefer th e riskyproject if and only if r ~ > y a m c ~ .Using (S), as long as a < 1 , rCT(ER)< 1 N T ( f R ) for every ER. Thus, a project( r R , R ) ,uch that rCT(eR)< r ~ < I . K ~ ( E ~ ) , ill not be chosen by managers byProposition 1, even thou gh i of Proposition 2 implies that shareh olders prefer it toproject 0.

Intuitively, whe n r ~ < rCT(cR)he risk prem ium that sha reholders must pro-vide exceeds the benefits to them of the managers' choice of the riskier project.No te that for every given ( S ,a) , uch that a < 1, the shareholders prefer the cho iceof the riskier project wh enever r ~> 0 ; however, whe n the contract includes auto-matic compensation for extra risk, i.e., specifies ( S N T , )where S N T s defined by(7), shareholders prefer the riskier project only when r ~ > r i T ( e R ) .IV. Project Choice and Corporate Value under IT ContractsA. Insiders' Project Choice

Under an IT contra ct, the profit from insider trading is a random variable thatdepends o n the chosen project and the initial value V 1 ( r , ) .Denoting by X ( r , E )


8/14

8 Journal of Financial and Quantitative Analysisthe total compe nsation managers receive if they choose the project ( r , ) , we obtainthat

where U ( r , ) is the insider trading profit. Th is profit can be written U ( r , ) as itdepend s on the chosen project an d on the initial value, which is also a function of( r , 1 .

Under an IT contract, mana gers' com pensation (and thus their project choice)depends on V 1 , s their trading profits depend on V l V 1 , n turn, depends on share-holders' expectation s as to how the man agers will ma ke their project choice. Giventhe expected m anagerial cho ice between (0 ,O)and ( r R , R ) , a (rational expectations)equilibrium is a value V ; and a managerial cho ice ( r * ,E * ) uch that: the choice of(I-*, E * ) is optimal for the ma nagers given V ; ; and , given the choice ( r * , * ) ,V ; isthe expected value of the final output net of the m anagers' total compen sation.Proposition 3 . Given an IT contract, mana gers will choose the risky project ( r R , R )iff

Proof. Th e man agers' project choice is determ ined by comparing U ( X ( r R , R ) ) andU ( X ( 0 , O ) ) .T he expected profits from insider trading are8( 1 2 ) E ( U ( r , ) ) = E ( P V'(r, E , 0) - V I r , 11 )

= 1 / 2 P [ ( l - a ) ( W o + r + ( l + ~ ) r n ) - S - V I ( ~ , E ) ]+ 1 / 2 P [ V l r , ) - 1 - a ) W o + r - 1 + ~ ) m ) S ]

Given the expected insider trading profit, one can calcula te the initial value, whichis the expected final output m inus the expected m anagerial com pensation, i.e.,

Using ( 1 2 )and ( l 3 ) , he manage rs' expected compen sation and standard deviationof their com pensation are

Substituting ( 1 4 a ) and ( 1 4 b ) nto ( 1 ) yields the managers' expected utility,( 1 5 ) U ( X ( r , ) ) = S + a (W o + r ) + P ( l - a ) ( l + ~ ) m- 7 [ a ( l + ~ ) mP2(1 - a ) ( l + ~ ) m ]

8 ~ nalculating the expected profit fro m insider trading, we co nside r the case in which (I - ) ( W o+r - 1 + e)m) - < VI r,E)< (1 - ) ( W O+ r + ( 1 + ~ ) m ) S. V1(r,) in (13) satisfies this inequality.


9/14

Bebchuk and Fershtman 9Th e riskier project (r R ,E R )s chosen only if U(X(rR ,ER))> U(X(0,O)). Using(15), the riskier project yields higher expe cted utility iff

(16) S + Q ( W o + r R ) + P ( 1 a ) ( l + t R ) m- 7 [ ~ ( l R ) m + P 2 ( 1 Q ) ( l + f R ) m ]> S + aW o + P ( l - ) m - 7 [ a m + g 2 ( 1- )m ] .

Rearranging (16) yields the condition in (1 1).Rem arks. i) Note now that for a < 1, a higher /3 implies a 1 0 w e r ~ ~ ~ ( c ~ ) .ntuitively,one can se e (in Equation (12)) that the expected gap betwee n the pretrading valueand the final value of the firm is positively correlated with the variability of theproject, i.e., the expected gap goes up with E. Thus, an increase in the fraction4 of the gap captured by the m anag ers imp roves the relative attractiveness of theriskier project. Thu s, manag ers will be willing to accept a project with a lowerexpected return, given that they can benefit m ore from exploiting the variability ofthe project.ii) Note that if cI T (t R )< 0, he possibility of insider trading might induce m anag ersto accept projects that yield a negative expected return. Using (1 I) , cIT (tR )< 0iff y < (1 - ) P ( P y + l ) / a .B. Corporate Value

Having identified the managers' choice under a given IT contract, let uscons ider the firm's initial value. Shar ehold ers can mak e the participation constraintbinding by fixing the salary SITas follows,(17) SIT = C - a (WO+ r ) - P(1 - ) ( l + t)m+ ym [ a i l + c) + f12(1- 0)(1 + t)] .

For such S IT , he initial value is(18) Vl(r , c) = Wo + r - C - ym [ a ( l + c) + g 2 ( 1- ) ( l + t)] ,

1. = l . ~ , = t~ i f r R >where r = 0, c = 0 if rR < cIT c R ) .Th e value V1 may be low er than the first-best initial value in two ways. First,

the ma nager s' c hoice of project may be inefficient. Seco nd, because managersmust be compensated for the risk they bear, the expected compensation that mustbe given to them exceeds C by the required risk premium. Th is risk premium isequa l to the last term o n the right side of (18) and, as will be discussed later, it isgreater than the risk prem ium required in the case of an NT contract w ith the sam ea .Proposition 4 . i) Given that managers are to get an IT contract that makes theparticipation constraint binding, shareholders prefer the riskier project (rR, R) oproject ze ro iff(19) TR > r;T (E R ) E ym [a + P2(1 - ) ] c ~ .


10/14

10 Journal of F inancial an d Quantitative Analysisii) The re are two possible types of "conflict" between share hold ers and managers:it is poss ible that the man age rs prefer the riskier project w hile sh areh olde rs preferproject zero, and it is also possible that the managers prefer project zero whileshare hold ers prefer the riskier project.Proof. i) Shareholders prefer the riskier project ( r R , R )as long as V 1 rR , R ) >V 1 ( O , ). Using ( 1 8 ) , his con dition im plies that the riskier project is preferred byshareholders as long as(20) Wo+ rR - - - -p?a 1 + E R ) + p2(1 - 0 ) 1 + E R ) ] >Wo- C - ym [ a + p2(1 - a ) ]which yields after simplifications the condition in (19).ii) If ( 1 + a ) ? p 2 > 7 , -0,hen Y ~ * ~ ( E R ) . [ T ( E ~ ) . hus, every project ( r R , R ) suchthat c I T ( f R ) < 1-R < Y ; T ( E R ) will be accepted by m anagers against the sharehold ers'will. O n the other hand, if ( 1 + a)-yf12< y a - 8,hen r;T(eR) < c I T ( c R ) , hichimplies that there are projects ( r R , R ) uch that rTT(eR)< r ~ < c I T ( c R )hat are notaccepted by managers even though shareholders wish them to acce pt,V. Comparison of IT and NT Contracts

We will first compare an NT contract with the salary scheme ( S N T , ) andan IT contract with the salary schem e ( S I T , ) .That is, both contracts specify thesame a and specify S such that ma nagers get exactly their reservation wage. Ourfirst comparison formalizes the basic intuition that the possibility of exploitinginside information create s an option for managers and cause s them to increase therisk of projects they accept, offsetting the risk aversion they otherwise exhibit inproject selection.Proposition 5. i) If the riskier project is chosen by ma nagers under th e N T contract( S N T , ) , hen it is also chosen und er the IT contract ( S I T , ) .ii) There are risky projects that are chosen under the I T contract while the sa meprojects are rejected by m anagers with the NT contract.Proof. i) A project ( r R , R ) s chosen by the managers under the N T contract iffr ~ > > N T ( ~ R ) ,he same project is chosen under the IT contract iff r ~ > ~ ~ ~ ( 6 ~ ) .Using (5) and ( 1 1 ) yields

Thus, if project ( r R , R ) s chosen by m anagers under the NT contract, i.e., r ~ >c N T ( e R ) ,hen it is also chosen by managers under the IT contract ( rR> e I T ( e R ) ) .ii) Any project ( r R , R )such that < r ~ < c N T ( e R )s chosen under the ITcontract, but not under the NT contract,

Propositions 1 and 2 imply that un der an NT contract there is only one type ofmanagement-shareholder conflict. Since cN T (cR) r G T (eR) ,managers m ay rejectprojects that shareholders wish to accept, but managers never accept projects that


11/14

Bebch uk and Fershtman 1 1shareholders wish to reject. Under an IT contract, however, there are two typesof conflict. U sing Propositions 3 and 4 and comparing r,, and r&, managers mayreject projects that shareholders wish to accept but may also accept projects thatshareholders wish they would reject.

In both types of contra cts, we let the fixed salary be su ch that the m anag ers'exp ected utility eq uals their reservation w age. Sinc e ma nage rs are risk averse, thisdoe s not imply that the two typ es of contracts involve the same expected mon etarypaym ent. Specifically,Proposition 6 . Given a project (rR, R), nsiders' total expected compensation isalways higher under the IT contract than under the NT contract.Proof. From (17), it is evident that the risk premium under the IT contract isym[n( l + E ) + $ ' ( I - a ) ( l + ) I . Meanwhile, from (7) we learn that the riskpremium under the NT contract is ~ n ( 1e)m. Thus, for a given a < 1 and e, therisk prem ium under the IT contract is greater than the risk premium under the NTcontract.

An important corollary of Proposition 6 is that if both IT and N T co ntractsproduce the sam e managerial choice, then the N T contract is superior as it implieslower managerial com pensation and thus higher initial value.

Th us, it remains to com pare the two contracts w hen they yield different ma n-agerial decisions. Proposition 5 indicates that when the riskier project is chosenunder an N T contract, it will also be chosen under an IT contract. Therefore,the only possible difference in managerial decisions is that the riskier project isrejected under an NT contract and accepted under an IT contract. Exam ining sucha case yields th e following result.Proposition 7. Given a project (rR, R), he IT co ntract yields a higher initial valuethan the NT contract if the following conditions hold,

i) C N T ( E R ) > 1 . ~> Y I T ( C R )ii) rR > -ymatR + ymb'2(1 - a ) ( l + e ~ ) .Proof. Cond ition i guarantees that (rR, R) s accepted under the I T contract andrejected under the NT contract. Using Equations (8) nd (18) o com pare the initialvalues under the two types of contract, it is evident that Condition ii guarantees

that the IT contract yields a higher initial value. aProposition 7 summ arizes one of the main conclusions of this paper. As

Proposition 5 indicates, letting managers engage in insider trading leads them tocho ose riskier projects. Bu t such incentives are costly: as Proposition 6 indicates,the IT contract involves higher expected managerial compensation, which lowersinitial value. T he overall effectiveness of the IT contract depe nds on both theseeffects and, as Proposition 7 indicates, when Conditions i and ii hold, the ITcontract yields a higher initial value than the NT contract.


12/14

12 Journal of Financial and Quantitative AnalysisVI. Optimal Contracts

In the previous section, we com pared N T and IT contracts with the same a.However, it is possible that shareholders will choose one a for the NT contractand another a for the IT contract. We will now consider and com pare the optimalNT and I T contracts when shareholders choose both a and S optimally.

We would like to note, however, that the analysis below is c a m e d out under thesimplifying assump tion that the managerial choice is between two risky projectsand share hold ers are aware of the characteristics of these projects. If shareholdersknow only that the risky projects are taken from some given distribution, then awill be set to maxim ize the value of V1 veraged over all project features, takinginto account the ma nagers' optim al acceptlreject cho ice.A. The Optimal NT Contract

Equation (8) indicates that V1 r , ) s a decreasing function of a. Th e greater ais, the greater is the risk m ana gers have to bear, and thu s the greater is the expectedcom pensation they get. Thu s, it is not optimal for shareholders to increase abeyond som e exogenously given minimal level of Q, and the optimal N T contractis ( S N T , ) .6. The Optima l IT Contract

Using (14) o differentiate a with respect to a yields that dalda = ( 1 +~ )m ( l 132) > 0. Thus, an increase in a increases the risk tha t manag ers bear,which imp lies higher expected managerial comp ensation. In our model, therefore,the only reason to provide m anagers w ith a contract with a > is to induce themto change their choice of project. From ( 1 ) , it is evident that Y ~ ~ ( E ~ )ncreaseswith a. Thus, increasing a beyond is optimal only when the riskier project isaccepted with a = Q and sh areholders prefer that the m anagers reject the riskierproject.9 In order to induc e manag ers to reject this project, shareho lders shouldse t a = aIT(rR, R ) , where aIT(rR,R) is such that the adjusted cIT(eR) s equal tor ~ . ny increase of a beyond this level is not optimal, as it does not affect theman agers' choice of project, but increases their expected com pensation. Using(1 ) , we obtain that

Providing a com pensation schem e with aIT(rR,R ) nduces managers to changetheir behavior and , in particular, to reject a project that sh areh olde rs wish them toreject. But, at the sam e time, it increases the risk that they bear and incr eases , con -sequently, their expected com pensation. Such a contract is optimal only ifVl(r , )is greater with aIT(rR,R) ,which implies that the riskier project is rejected, than

9 ~ o t elso that a higher a implies a higher r;T because it also increases the risk premium thatmanagers get.


13/14

Bebchuk and Fershtrnan 13with Q, which implies that the riskier project is accepted. Specifically, the contractwith CVIT s optimal only if

It remains to see how optimal contracting affects the comparison between NTand I T contracts:i) < cN T (fR ) ven when the two types of contract specify differentvalues at a . Thus, we can extend Proposition 5 to optimal contracts and claimthat if the riskier project is accepted with the optimal NT contract, then it is alsoaccepted with the optimal IT contract.ii) From (17), the managerial risk prem ium under the IT contract is am (1 +

) [ a + p 2 (1 - a ) ] whereas, using (7), this risk premium under the NT contract isy a ( 1 + ~ ) m .When the optimal IT contract is with a = CY Proposition 6 applies.When the optimal IT contract is with CVIT> Q, then, since the risk premium isincreasing with a , the IT contract involves an even larger risk prem ium. Hence,Proposition 6 can also be extended to optimal contracts: the total expected com pen -sation under the optimal IT contract is higher than under the optimal N T contract.

iii) Propo sition (7) discusses situations in which the IT contract yields high erinitial value than the NT contract. This proposition can be also extended to optimalcontracts by considering contracts with a = Q and keeping in mind that lettingshareholders chang e a can only increase the initial value.VII. Conclusion

Insider trading has been shown to have a significant effect on insiders' choiceamo ng investment projects with uncertain returns. In particular, it requires refiningthe familiar proposition that managers' risk ave rsion leads them to choose a moreconservative investment policy than shareholders desire. In the presence of insidertrading, increased u ncertainty presents insiders not only with costs , but also withbenefits, as uncertainty enable s them to m ake greater trading profits. This effectmitigates, and may even outweigh, the conservative tendency arising from insid-ers' risk aversion. Whether the presence of insider trading improves or worsensinsiders' project choices de pen ds on conditions that our analysis has identified.Th e presence of insider trading has also been shown to have an effect on thestructure and overall level of insiders' com pensation. First, with insider trading,the expected value of insiders' total com pensation m ust be higher to com pensateinsiders for the fact that they are partly paid in trading profits that are of uncertainsize. Second, under some circumstances, the presence of insider trading leads toa relative increase in the degree to which insiders' salary depends on the firm'sresults.

More generally, the analysis of this paper suggests that the extent to whichinsiders may trade in their firm's shares has considerable effects on the agencyproblem in corporations. Th us, an understanding of these effects is necessary forboth i) designing the corporate and legal arrangem ents governing insider trading,and ii) forming an accurate picture of the agency problem in corporations. Wehave sought in this paper to contribute to the understanding of these effects.


14/14

14 Journ al of Financial an d Quantitative AnalysisReferencesAusubel, L. "Insider Trading in a Rational Expectations Economy." American Economic Review, 80(Dec. 1990), 1022-1041.Bagno li, M., and N. Khan na. "Insider Trading in Financial Signalling Mode ls." W orking Paper, Univ.of Michigan (1991).Be bch uk, L., and C. Fershtman . "The Effects of Insider Trading on Inside rs' Level of Effort." EuropeanJournal of Political Econ omics, 9 (Dec. 1993) , 4 6 9 4 8 1.. "The Effects of Insider Trading on Insiders' Reactions to Opportunities to'Waste' Corporate Value." NBER Technical Working Paper No. 95 (1992).Carlton, D., and D . Fischel. "The Regulation of Insider Trading." Stanford Law Review, 35 (Sept.1983), 857-895.Dye, R. "Inside Trading and Incentives." Journal of Business, 57 (July 1984), 295-313.Easterbrook, F. H . "Insider Trading as an Agency Problem." In Principles and Agents: The Structureof Business, J. W. Pratt and R. J. Zeckhause r, eds. Bo ston , MA : Harvard Business Scho ol Press(1985 ), 81-100.Fischer, P. E. "Optimal Contracting and Insider Trading Restrictions." The Journal of Finance, 47,(June 1992) , 673494.Fishman, M ., and K. Hagerty. "Insider Trading and the Efficiency of Stock Prices." Rand Journal ofEconomics, 23 (Spring 1992), 10 61 22 .Giamm arino R.; R. Heinkel; an d B. Hollifield. "Corporate D ecisions and Anonym ous Trading."Working Paper, Univ. of British Columbia (1992).Glo sten , L . R ., and P. R. Milgrom . "Bid, Ask and Transaction Prices in a Specialist Market withHeterogeneously Informed Traders." Journal of Financial Economics, 14 (March 1985 ), 71-100.Holm strom, B., and P. Milgrom . "Agg regation and Linearity in the Provision of Intertemporal Incen -tives." Econometrica, 55 (March 1987), 303-328.Holmstrom, B., and J. Tirole. "Corporate Control and the Monitoring Role of the Stock Market."M.I.T. and Yale, Mimeo (1990 ).Jaffe, J. "Special Inform ation and Insider Trading ." Journal ofBusiness, 47 (July 1974) , 4 1 M 2 8 .John, K ., and L. H. P. Lan g. "Insider Trading around Dividend Annou ncemen t: Theory and Evidence."Journal of Finance, 56 (Sept. 1991), 1361-1389.Kyle, A. "Continuous Auctions and Insider Trading." Econometrica, 53 (N ov. 19 85), 1315-1335.Laffont, J. J., and E. S. Mask in. "The Efficient Market Hypothesis and Insider Trading on the Stock

Market." Journal ofPolitica1 Economy, 98 (Feb. 1 990 ), 70-93.Manove, M. "The Harm in Insider Trading." Quarterly Journal of Economics, 104 (Nov. 1989),823-845.Mirman, L. J., and L. Samuelson. "Information and Equilibrium with Inside Traders." EconomicJournal, 99 (Conference 1989 ), 152-167.Scott, K. "Insider Trading: Rule lob-5, Disclosure, and Corporate Privacy." Journal of Legal Studies,9 (July 1980), 801-818.Seyhun , H . N. "Insiders' Profits, Costs of Trading, and Market Efficiency." Journal of FinancialEconomics, 16 (June 1986 ), 189-212.

1994 Insider Trading and the Managerial Choice Among Risky Projects

Documents

Transcript of 1994 Insider Trading and the Managerial Choice Among Risky Projects