Trading and Liquidity with Limited Cognition


 Darrell Wiggins
 1 years ago
 Views:
Transcription
1 Trading and Liqidity with Limited Cognition Brno Biais, Johan Hombert, and PierreOlivier Weill December 11, 21 Abstract We stdy the reaction of financial markets to aggregate liqidity shocks when traders face cognition limits. While each financial instittion recovers from the shock at a random time, the trader representing the instittion observes this recovery with a delay, reflecting the time it takes to collect and process information abot positions, conterparties and risk exposre. Cognition limits lengthen the market price recovery. They also imply that traders who find that their instittion has not yet recovered from the shock place market sell orders, and then progressively by back at relatively low prices, while simltaneosly placing limit orders to sell later when the price will have recovered. This generates rond trip trades, which raise trading volme. We compare the case where algorithms enable traders to implement this strategy to that where traders can place orders only when they have completed their information processing task. Keywords: Liqidity shock, limitorders, asset pricing and liqidity, algorithmic trading, limited cognition, sticky plans J.E.L. Codes: G12, D83. We are gratefl to the editor Harald Uhlig, and for anonymos referees for insightfl comments. Many thanks, for helpfl discssions and sggestions, to Andy Atkeson, Dirk Bergemann, Darrell Dffie, Emmanel Farhi, Thierry Focalt, Xavier Gabaix, Alfred Galichon, Christian Hellwig, Hgo Hopenhayn, Vivien Lévy Garboa, Johannes Horner, Boyan Jovanovic, Ricardo Lagos, Albert Menkveld, John Moore, Stew Myers, Henri Pages, Thomas Philippon, Gary Richardson, Jean Charles Rochet, Gillame Rochetea, Ioanid Ros, Larry Samelson, Tom Sargent, Jean Tirole, Aleh Tsyvinski, Jsso Välimäki, Dimitri Vayanos, Adrien Verdelhan, and Glen Weyl; and to seminar participants at the DaphineNYSEEronext Market Microstrctre Workshop, the Eropean Smmer Symposim in Economic Theory at Gerzensee, École Polytechniqe, Stanford Gradate School of Bsiness, New York University, Northwestern University, HEC Montreal, MIT, and UCI. Palo Cotinho and Kei Kawakami provided excellent research assistance. Brno Biais benefitted from the spport of the Financial Markets and Investment Banking Vale Chain Chair sponsored by the Fédération Bancaire Française and from the Eroplace Institte of Finance, and PierreOlivier Weill from the spport of the National Science Fondation, grant SES Tolose School of Economics (CNRSCRM, IDEI), HEC Paris, University of California Los Angeles and NBER,
2 1 Introdction The perception of the intellect extends only to the few things that are accessible to it and is always very limited (Descartes, 1641) We analyze trades and price dynamics when investors face cognition limits and the market is hit by an aggregate liqidity shock. The shock indces a transient drop in the willingness and ability of financial instittions to hold assets sch as stocks or bonds. It can be triggered by changes in the characteristics of assets, e.g., certain types of instittions, sch as insrance companies or pension fnds, are reqired to sell bonds that lose their investment grade stats, or stocks that are delisted from exchanges or indices (see, e.g., Greenwood, 25). Alternatively, an aggregate shock can reflect events affecting the overall financial sitation of a category of instittions, e.g., fnds experiencing large otflows or losses (see Coval and Stafford, 27), banks incrring large losses (see Berndt, Doglas, Dffie, Fergson, and Schranz, 25), or specialists bilding extreme positions (see Comerton Forde, Hendershott, Jones, Molton, and Seasholes, 21). 1 To recover from a shock, instittions seek to nwind their positions in mltiple markets (e.g., credit defalt swaps (CDS) corporate bonds or mortgage backed secrities (MBS)). Instittions can also raise new capital, secre credit lines or strctre derivative trades to hedge their positions. These processes are complex and take time. However, once an instittion arranges enogh deals, it recovers from the liqidity shock. In this context, traders mst collect and process a large flow of information abot asset valations, market conditions and the financial stats of their own instittion. They mst obtain and aggregate information from several desks, markets and departments abot gross and net positions, the reslting risk exposre, and compliance with reglations. When traders have limited cognition, completing these tasks is challenging and reqires significant time and effort. We address the following isses: How do traders and markets cope with negative liqidity shocks? What is the eqilibrim price process after sch shocks? How are trading and prices affected by cognition limits? Do the conseqences of limited cognition vary with market mechanisms and technologies? 1 A striking example of a liqidity shock and its conseqences on instittions and market pricing is analyzed by Khandani and Lo (28). These observe that, dring the week of Agst 6 th 27, qantitative fnds sbject to margin calls and losses in credit portfolios had to rapidly nwind eqity positions. This reslted in a sharp bt transient drop in the S&P 5. Bt, by Agst 1 th 27 prices had in large part reverted. 2
3 We consider an infinite horizon, continos time market with a continm of rational, risk netral competitive financial instittions, deriving a non linear tility flow from holding divisible shares of an asset, as in Lagos and Rochetea (29) and Gârlean (29). To model the aggregate liqidity shock, we assme that at time this tility flow drops for all instittions, as in Dffie, Gârlean, and Pedersen (27) and Weill (27). Then, as time goes by, instittions progressively switch back to a high valation. More precisely, each instittion is associated with a Poisson process and switches back to highvalation at the first jmp in this process. Unconstrained efficiency wold reqire that lowvalation instittions sell to highvalation instittions. However, in or model, sch asset reallocation is delayed becase of cognition limits. In line with the rational inattention models of Reis (26a and 26b), Mankiw and Reis (22) and Gabaix and Laibson (22), we assme that each trader engages in information collection and processing for some time and, only when this task is completed, observes the crrent valation of her instittion for the asset. We refer to this observation as an information event. When this event occrs, the trader pdates her optimal asset holding plan, based on rational expectations abot ftre variables and decisions. 2 The corresponding demand, along with the market clearing condition, gives rise to eqilibrim prices. Ths, while Mankiw and Reis (22) and Gabaix and Laibson (22) analyze how inattention affects consmption, we stdy how it affects eqilibrim pricing dring liqidity shocks. In the spirit of Dffie, Gârlean, and Pedersen (25), we assme that information events, and correspondingly trading decisions, occr at Poisson arrival times. 3 Ths in or model, each instittion is exposed to two Poisson processes: one concerns changes in its valation for the asset, the other the timing of its trader s information events. For simplicity, we assme these two processes are independent. Frther, for tractability, we assme that these processes are independent across instittions. Ths, by the law of large nmbers, the aggregate state of the market changes deterministically with time. 4 Correspondingly, the eqilibrim price process is 2 Ths, in the same spirit as in Mankiw and Reis (22), traders have sticky plans bt rationally take into accont this stickiness. 3 Note however that the interpretation is different. Dffie, Gârlean, and Pedersen (25) model the time it takes traders to find a conterparty, while we model the time it takes them to collect and process information. This difference reslts in different otcomes. In Dffie, Gârlean, and Pedersen investors do not trade between two jmps of their Poisson process. In or model they do, bt based on imperfect information abot their valation for the asset. In a sense or model can be viewed as the dal of Dffie, Gârlean, and Pedersen: they assme that traders continosly observe their valation bt are infreqently in contact with the market, while we assme that traders are continosly in contact with the market bt infreqently refresh their information abot their valation. 4 We also analyze an extension of or framework where the market is sbject to recrring aggregate liqidity shocks, occrring at Poisson arrival times. While, in this more general framework, the price is stochastic, the 3
4 deterministic as well. We show eqilibrim existence and niqeness. In eqilibrim the price increases with time, reflecting that the market progressively recovers from the shock. Limits to cognition lengthen the time it takes market prices to flly recover from the shock. Yet they do not necessarily amplify the initial price drop generated by that shock. Jst after the shock, with perfect cognition, the marginal trader knows that her instittion has a low valation, while with limited cognition sch trader is imperfectly informed abot her instittion s valation, and realizes that, with some probability, it may have recovered. We also show that the eqilibrim allocation is an information constrained Pareto optimm. This is becase in or setp there are no externalities, as the holding constraints on holdings imposed by cognition limits on one trader, do not depend on the actions of other traders. While we first characterize traders optimal policies in terms of abstract holding plans, we then show how these plans can be implemented in a realistic market setting, featring an electronic order book, limit and market orders, and trading algorithms. The latter enable traders to condct programmed trades while devoting their cognitive resorces to investigating the liqidity stats of their instittion. In this context, traders who find ot their instittion is still sbject to the shock, and correspondingly has a low valation for the asset, sell a lmp of their holdings, with a market sell order. Sch traders also program their trading algorithms to then gradally by back the asset, as they expect their valation to revert pward. Simltaneosly, they sbmit limit orders to sell the asset, to be exected later when the eqilibrim price will have recovered. To the extent that they by in the early phase of the aggregate recovery, and then sell towards the end of the recovery, sch traders act as market makers. 5 The corresponding rond trip transactions reflects the traders optimal response to cognition limits. These transactions can raise trading volme above the level it wold reach nder perfect cognition. We also stdy the case where trading algorithms are not available and traders mst implement their holding plans by placing limit and market orders when their information process jmps. With increasing prices, this prevents them from bying in between jmps of their information process. When the liqidity shock is large, this constraint binds and redces the efficiency of the eqilibrim allocation. It does not necessarily amplify the price pressre of the liqidity shock, however. Since traders anticipate they will not be able to by back ntil qalitative featres of or eqilibrim are pheld. 5 In doing so they act similarly to the market makers analyzed by Grossman and Miller (1988). Note however that, while in Grossman and Miller agents are exogenosly assigned market making or market taking roles, in or model, agents endogenosly choose to spply or demand liqidity, depending on the realization of their own shocks. 4
5 their next information event, they sell less when they observe that their valation is low. Sch a redction in spply limits the selling pressre on prices. Pt differently, banning the se of algorithms cold help alleviate the initial price pressre created by the liqidity shock. Yet, this policy wold redce welfare: indeed, in or model the eqilibrim with algorithmic trading is information constrained Pareto optimal. 6 The order placement policies generated by or model are in line with several stylized facts. Irrespective of whether algorithms are available, we find that sccessive traders place limit sell orders at lower and lower prices. Sch nderctting is consistent with the empirical reslts of Biais, Hillion, and Spatt (1995), Griffiths, Smith, Trnbll, and White (2) and Elll, Holden, Jain, and Jennings (27). Frthermore, or algorithmic traders both spply and consme liqidity, by placing market and limit orders, consistent with the empirical findings of Hendershott and Riordan (21) and Brogaard (21). Brogaard also finds that algorithms i) tend not to withdraw from the market after large liqidity shocks; ii) tend to provide liqidity by prchasing the asset after large price drops; and iii) in doing so profit from price reversals. Frthermore, Kirilenko, Kyle, Samadi, and Tzn (21) find that algorithms tend to by after a rise in price and that these prchases tend to be followed by a frther increase in price. They also find that algorithmic traders engage in high freqency rond trip trades, bying the asset and then selling it to other algorithmic traders. All these findings are in line with the implications of or model. Or analysis of the dynamics of markets in which traders choose whether to place limit or market orders is related to the insightfl papers of Parlor (1998), Focalt (1999), Focalt, Kadan, and Kandel (25), Ros (29), and Goettler, Parlor, and Rajan (25, 29). However, we focs on different market frictions than they do. While these athors stdy strategic behavior and/or asymmetric information nder perfect cognition, we analyze competitive traders with symmetric information nder limited cognition. This enables s to examine how the eqilibrim interaction between the price process and order placement policies is affected by cognition limits and market instrments. The next section describes the economic environment and the eqilibrim prevailing n 6 In Section II of or spplementary appendix we analyze the case when traders can place only market orders when their information process jmps, i.e., limit orders and trading algorithms are rled ot. In sch case the price reverts to its pre shock level sooner. Indeed, when traders can place limit orders, the sell orders stored in the book exert a downward pressre on prices towards the end of the recovery. Bt, then again, the efficiency of the allocation is higher when traders can se limit orders, and even higher they can se algorithms. Indeed these market instrments enable traders to condct mtally beneficial trades that wold be infeasible with market orders only. 5
6 1 v(h, q) tility flow v(, q) qantity Figre 1: The tility flows of high (in ble) and low valation (in red) investors, when σ =.5. der nlimited cognition. Section 3 presents the eqilibrim prevailing with limited cognition. Section 4 discsses the implementation of the abstract eqilibrim holding plans with realistic market instrments sch as limit and market orders and trading algorithms. Section 5 concldes. Proofs not given in the text are fond in the appendix, and a spplementary appendix offers additional information abot the model, along with proofs and analyses. 2 The economic environment 2.1 Assets and agents Time is continos and rns forever. A probability space (Ω, F, P)is fixed,as well as an information filtration satisfying the sal conditions (Protter, 199). 7 There is an asset in positive spply s (, 1) and the economy is poplated by a [, 1]continm of infinitelylived agents that we call financial instittions (fnds, banks, insrers, etc.) disconting the ftre at the same rate r>. Each instittion can be in one of two states. Either it derives a high tility flow ( θ = h ) from holding any qantity q oftheasset,oritderivesalowtilityflow( θ = ), as illstrated in Figre 1. For high valation instittions, the tility flow per nit of time is v(h,q) = q, for all q 1, and v(h,q) = 1, for all q > 1. For low valation instittions, it is v(, q) = q δ q1+σ, for all q 1, and v(, q) = 1 δ/(1 + σ), for all q > 1+σ 1.8 The two parameters δ (, 1] and σ > captretheeffectofthelowstateontilityflows. The 7 To simplify the exposition, for most stated eqalities or ineqalities between stochastic processes, we sppress the almost srely qalifier as well as the corresponding prodct measre over times and events. 8 The short selling constraint is withot loss of generality in the following sense. If we extend the tility fnctions to q< in any way sch that they remain concave, then the eqilibrim otcomes we characterize are naffected. In particlar, q< never arises. 6
7 parameter δ controls the level of tility: the greater is δ, theloweristhemarginaltilityflow of low valation instittions. The parameter σ, on the other hand, controls the crvatre of low valation instittions tility flows. The greater is σ, the less willing sch instittions are to hold extreme asset positions. 9 Becase of this concavity, it is efficient to spread holdings among low valation instittions. This is similar to risk sharing between risk averse agents, and as shown below will imply that eqilibrim holdings take a rich set of vales. 1 This is in line with Lagos and Rochetea (29) and Gârlean (29). Note that, even in the σ limit, low valation investors tility flow is redced, by a factor 1 δ, btinthatcasethe tility flow is piecewise linear. 11 The difference between the high and the low states can be interpreted as a holding cost, or a capital charge or shadow cost associated with positions. As discssed in the introdction and in Dffie, Gârlean, and Pedersen (27) a variety of instittional factors can generate sch costs, e.g., reglatory constraints on holdings (see, e.g., Greenwood, 25), need for cash (see Coval and Stafford, 27, Berndt et al., 25), position limits (see Comerton Forde et al., 21, or Hendershott and Seasholes, 27), or tax considerations. Lastly, in addition to deriving tility from the asset, instittions can prodce (or consme) a nonstorable nméraire good at constant marginal cost (tility) normalized to Liqidity shock To model liqidity shocks we follow Dffie, Gârlean, and Pedersen (27) and Weill (27). Before the shock, each instittion is in the high valation state, θ = h, andholdss shares of the asset. Bt, at time zero, the liqidity shock hits all the instittions, and they make a switch to low valation, θ =. Note, however, that the shock is transient. As discssed in the introdction, to cope with the shock, instittions seek to nwind positions, raise capital, secre credit lines, or hedge positions. All this process is complex and takes time. However, once the instittion is able to arrange enogh deals, it recovers from the liqidity shock. To captre the recovery process we assme that, for each instittion, there is a random time at which it reverts 9 The crvatre of low valation tilities contrasts with the constant positive marginal tility high valation instittions have for q<1. One cold have introdced sch crvatres for high valations as well, as in Lagos and Rochetea (29) or Gârlean (29) at the cost of redced tractability, withot qalitatively altering or reslts. 1 Note also that the holding costs of low valation instittions are homothetic. This reslts in homogenos asset demand and, as will become clear later, this facilitates aggregation. 11 For the σ limit, see or spplementary appendix, (Biais, Hombert, and Weill, 21b), Section III. 7
8 to the high valation state, θ = h, andthenremainsthereforever. Forsimplicity,weassme that recovery times are exponentially distribted, with parameter γ, and independent across investors. Hence, by the law of large nmbers, the measre µ ht of high valation investors at time t is eqal to the probability of high tility at that time conditional on low tility at time zero. 12 Ths µ ht =1 e γt, (1) and we denote by T s the time at which the mass of high tility instittions eqals the spply of the asset, i.e., µ hts = s. (2) 2.3 Eqilibrim withot cognition limits Consider the benchmark case where instittions continosly observe their θ t. To find the competitive eqilibrim, it is convenient to solve first for efficient asset allocations, and then find the price path that decentralizes these efficient allocations in a competitive eqilibrim. 13 In the efficient allocation, for t>t s, all assets are held by high valation instittions, and all marginal tilities are eqalized. Indeed, with an (average) asset holding eqal to s/µ ht < 1, the marginal tility is 1 for high valation instittions, while with zero asset holdings marginal tility is v q (, ) = 1 for low valation instittions. In contrast, for t T s,wehaveµ ht s, and each high valation instittion holds one nit of the asset while the residal spply, s µ ht, is held by low valation instittions. The asset holding per low valation instittion is ths: q t = s µ ht 1 µ ht. (3) This is an optimal allocation becase all high valation instittions are at the corner of their tility fnction: redcing their holdings wold create a tility loss of 1, while increasing their holdings wold create zero tility. Low valation instittions, on the other hand, have holdings 12 For simplicity and brevity, we do not formally prove how the law of large nmbers applies to or context. To establish the reslt precisely, one wold have to follow Sn (26), who relies on constrcting an appropriate measre for the prodct of the agent space and the event space. 13 Note that, with qasi linear tilities and nlimited cognition, in all Pareto efficient allocation of assets and nméraire goods, the asset allocation maximizes, at each time, the eqally weighted sm of the instittions tility flows for the asset, sbject to feasibility. 8
9 in [, 1), so their marginal tility is strictly positive and less than 1. For t T s,assoonasaninstittionswitchesfromθ = to θ = h, itsholdingsjmpfromq t to 1, while as long as its valation remains low, it holds q t,givenin(3),whichsmoothlydeclines with time. This decline reflects that, as time goes by, more and more instittions recover from the shock, switch to θ = h and increase their holdings. As a reslt, the remaining low valation instittions are left with less shares to hold. Eqilibrim prices reflect the cross section of valations across instittions. In or setting, by the law of large nmbers, there is no aggregate ncertainty and this cross section is deterministic. Hence, the price also is deterministic. For t T s,itiseqaltothepresentvaleofa low valation instittion s marginal tility flow: p t = t e r(z t) v q (, q z ) dz, where q z is given in (3). Taking derivatives with respect to t, wefindthatthepricesolvesthe Ordinary Differential Eqation (ODE): v q (, q t )=rp t ṗ t ξ t. (4) The lefthand side of (4) is the instittion s marginal tility flow over [t, t + dt]. The righthand side is the opportnity cost of holding the asset: the cost of bying a share of the asset at time t and reselling it at t + dt, i.e., the time vale of money, rp t,minsthecapitalgain,ṗ t. Finally, when t T s, v q (, q z )=v q (, ) = 1 and the price is p t =1/r. Ths, the price increases deterministically towards 1/r, astheholdingsoflowvalation instittions go to zero and their marginal tility increases. Instittions do not immediately bid p this predictable price increase becase the demand for the asset bilds p slowly: on the intensive margin, high valation instittions derive no tility flow if they hold more than one nit; and, on the extensive margin, the recovery from the aggregate liqidity shock occrs progressively as instittions switch back to high tility flows. Ths, there are limits to arbitrage in or model, in line with the empirical evidence on the predictable patterns of price drops and reversals arond liqidity shocks. 14 Throghot this paper we illstrate or reslts with nmerical comptations based on the parameter vales shown in Table 1. We take the discont rate to be r =.5, in line with Dffie, 14 See, e.g., for short lived shocks the empirical findings of Hendershott and Seasholes (27), Hendershott and Menkveld (21) and Khandani and Lo (28). 9
10 Panel A: Proportion of high valation investors 1 µ ht s.5 T s Panel B: Price 1 perfect cognition limited cognition time (days) Figre 2: Poplation of high valation investors (Panel A) and price dynamics when σ =.3 (Panel B). T s T f Gârlean, and Pedersen (27). We select the liqidity shock parameters to match empirical observations from large eqity markets. Hendershott and Seasholes (27) and Hendershott and Menkveld (21) find liqidity price pressre effects of the order of 1 to 2 basis points, with dration ranging from 5 to 2 days. Dring the liqidity event described in Khandani and Lo (28), the price pressre sbsided in abot 4 trading days. Adopting the convention that there are 25 trading days per year, setting γ to 25 means that an instittion takes on average 1 days to switch to high valation. Setting the asset spply s to.59 then implies that with nlimited cognition the time it takes the market to recover from the liqidity shock (T s ) is approximately 9 days, as illstrated in Figre 2, Panel A. Lastly, for these parameter vales, setting δ =1impliesthattheinitialpricepressregeneratedbytheliqidityshockis between 1 and 2 basis points, as illstrated in Figre 2, Panel B Dffie, Gârlean and Pedersen (27) provide a nmerical analysis of liqidity shocks in over the conter markets. They choose parameters to match stylized facts from illiqid corporate bond markets. Becase we focs on more liqid electronic exchanges, we chose parameter vales different from theirs. For example in their analysis the price takes one year to recover while in ors it takes less than two weeks. 1
11 Table 1: Parameter vales Parameter Vale Intensity of information event ρ 25 Asset spply s.59 Recovery intensity γ 25 Discont rate r.5 Utility cost δ 1 Crvatre of tility flow σ {.3,.5, 1.5} 3 Eqilibrim with limited cognition We now trn to the case where agents have limited cognition. In the first sbsection below, we present or assmptions on cognition limits. In the second sbsection we solve for eqilibrim. To do so, we follow several standard steps: we first compte the vale fnction of the traders, then we maximize this fnction to pin down demands, finally we sbstite demands into the market clearing condition and we obtain the eqilibrim price. In the third sbsection we present the properties of the eqilibrim, regarding welfare, holding plans, trading volme, and prices. In the forth sbsection we show that or qalitative reslts are pheld in an extension of the model where recrring preference shocks lead to stochastic eqilibrim prices. 3.1 Assmptions Limited cognition Each instittion is represented in the market by one trader. 16 To determine optimal asset holdings, the trader mst analyze the liqidity stats of her instittion, θ t.thistaskiscognitively challenging. As mentioned in the previos section, to recover from the shock, the instittion engages in several financial transactions in a variety of markets, some of them complex, opaqe and not compterized. Evalating the liqidity stats of the instittion reqires collecting, analyzing and aggregating information abot the reslting positions. Or key assmption is that, becase of limited cognition and information processing constraints, the trader cannot continosly and immediately observe the liqidity stats of her instittion. 17 Instead, we assme there is a conting process N t sch that the trader observes θ t at each jmp of N t (and only 16 For simplicity we abstract from agency isses and assme that the trader maximizes the intertemporal expected tility of the instittion. 17 Reglators have recently emphasized the difficlty of devising an integrated measrement of all relevant risk exposres within a financial instittion (see Basel Committee on Banking Spervision, 29). Academic research has also nderscored the difficlties associated with the aggregation of information dispersed in mltiple departments of a financial instittion (see Vayanos, 23). 11
12 then). 18 At the jmps of her information process N t the trader sbmits a new optimal trading plan, based on rational expectations abot {N,θ : t}, andherftredecisions. This is in line with the rational inattention model of Mankiw and Reis (22). For simplicity, the traders information event processes, N t, are assmed to be Poisson distribted, with intensity ρ, andindependentfromeachotheraswellasfromthetimesatwhichinstittionsemergefrom the liqidity shock Conditions on asset holding plans and prices When an information event occrs at time t>, a trader designs an pdated asset holding plan, q t,,forallsbseqenttimes t ntil the next information event. Formally, denoting D = {(t, ) R 2 + : t }, weletanasset holding plan be a bonded and measrable stochastic process q : D Ω R + (t,, ω) q t, (ω), satisfying the following two conditions: Condition 1. For each t, the stochastic process (t, ω) q t, (ω) is F t predictable, where {F t } t is the filtration generated by N t and θ t. Condition 2. For each (t, ω), the fnction q t, (ω) has bonded variations. Condition 1 means that the plan designed at time t, q t,,candependonlyonthetrader s timet information abot her instittion: the history of her informationevent conting process, and of her instittion tility stats process p to, bt not inclding, time t. 2 Condition 2 is a technical reglarity condition ensring that the present vale of payments associated with q t, 18 The time between jmps creates delays in obtaining fresh information abot θ t, which can be interpreted as the time it takes the risk management nit or head of strategy to aggregate all relevant information and disseminate it to the traders. 19 For simplicity, we don t index the information processes of the different traders by sbscripts specific to each trader. Rather we se the same generic notation, N t, for all traders. 2 We add the not inclding qalifier becase the asset holding plans are assmed to be F t predictable instead of F t measrable. This predictability assmption is standard for dynamic optimization problems involving decisions at Poisson arrival times (see Chapter VII of Brémad, 1981). For mch of this paper, however, we need not distingish between F t predicability and F t measrability. This is becase the probability that the trader type switches exactly at the same time an information event occrs is of second order. Therefore, adding or removing the type information accring exactly at information events leads to almost srely identical optimal trading decisions. 12
13 is well defined. To simplify notations, from this point on we sppress the explicit dependence of asset holding plans on ω. At this stage of the analysis, we assme that traders have access to a sfficiently rich men of market instrments to implement any holding plan satisfying Conditions 1 and 2. We examine implementation in Section 4, where we analyze which market instrments are needed to implement eqilibrim holding plans, and the eqilibrim that arises when the men of market instrment is not sfficiently rich. The last technical condition concerns the price path: Condition 3 (Wellbehaved price path). The price path is bonded, deterministic and continosly differentiable (C 1 ). As in the nbonded cognition case, becase there is no aggregate ncertainty, it is natral to focs on deterministic price paths. Frther, in the environment that we consider the eqilibrim price mst be continos, as formally shown in or spplementary appendix (see Biais, Hombert, and Weill, 21b, Section VI). The economic intition is as follows. If the price jmps at time t, alltraderswhoexperienceaninformationeventshortlybeforet wold want to arbitrage the jmp: they wold find it optimal to by an infinite qantity of asset and resell these assets jst after the jmp. This wold contradict market clearing. Finally, the condition that the price be bonded is imposed to rle ot bbbles (see Lagos, Rochetea, and Weill, 27, for a proof that bbbles cannot arise in a closely related environment). 3.2 Eqilibrim Intertemporal payoffs For any time, let τ denote the time of the last information event before, withthe convention that τ =ifnoinformationeventoccrred. Correspondinglyq, represents the holdings of a trader who had no information event by time and ths no opportnity to pdate her holding plan. Given that all traders start with the same holdings at time zero, we have q, = q, = s for all. The trader s objective is to maximize the intertemporal expected vale of tility flows, net of the cost of bying and selling assets. With the above notations in mind, this can be written 13
14 as: E e v(θ r,q τ,)d p dq τ,, (5) where v(θ,q τ,) d is the tility enjoyed, and p dq τ, is the cost of asset prchases dring [, + d], given the holding plan chosen at τ,thelastinformationeventbefore. Given or distribtional assmptions for the type and information processes, and given technical Conditions 1 to 3 we can rewrite this objective eqivalently as: Lemma 1. The intertemporal payoffs associated with the holding plan q t, is, p to a constant: V (q) =E where ξ = rp ṗ. e rt t e (r+ρ)( t) E t [v(θ,q t, )] ξ q t, d ρdt, (6) The interpretation of eqation (6) is as follows. The oter expectation sign takes expectation over all time t histories. The ρdt term in the oter integral is the probability that an information event occrs dring [t, t + dt]. Conditional on the time t history and on an information event occrring dring [t, t + dt], the inner integral is the disconted expected tility of the holding plan ntil the next information event. At each point in time this involves the difference between a trader s time t expected valation for the asset, E t [v(θ,q t, )], and the opportnity cost of holding that asset, ξ. This is similar to the reslt in Lagos and Rochetea (29) that an investor s demand depends on his crrent marginal tility from holding the asset as well as his expected marginal tility in the ftre. Finally, the discont factor applied to time is adjsted by the probability e ρ( t) that the next information event occrs after Market clearing In all what follows we focs on the case where all traders choose the same holding plan, which is natral given that traders are ex ante identical. 21 Of corse, while traders choose ex ante the same holding plan, ex post they realize different histories of N t and θ t,andhencedifferent asset holdings. The market clearing condition reqires that, at each date, the crosssectional average asset holding be eqal to s, thepercapitaassetspply. Bythelawoflargenmbers,andgiven 21 By ex ante identical we mean that traders start with the same asset holdings and have identically distribted processes for information event and tility stats. 14
15 ex ante identical traders, the crosssectional average asset holding is eqal to the expected asset holding of a representative trader. Hence, the market clearing condition can be written: E [q τ,] =s. (7) for all. Integrating against the distribtion of τ,andkeepinginmindthatq, = s, leads to or next lemma: Lemma 2. The time market clearing condition, (7), writes: ρe (1 ρ( t) µ ht )E [q t, θ t = ]+µ ht E [q t, θ t = h] s dt =. (8) This lemma states that the aggregate net demand of traders who experienced at least one information event is eqal to zero. The first mltiplicative term in the integrand of (8), ρe ρ( t),isthedensityoftime t traders, i.e., traders whose last information event occrred at time t (,]. The first and second terms in the crly bracket are the gross demands of time t low and high valation traders respectively. The last term in the crly bracket is their gross spply. It is eqal to s becase information events arrive at random, which implies that the average holding of time t traders jst prior to their information event eqals the poplation average Eqilibrim existence and niqeness We define an eqilibrim to be a pair (q, p) sbjecttoconditions1,2and3andschthat:i) given the price path, the asset holding plans maximize V (q) givenin(6),andii)theholding plans are sch that the market clearing condition (8) holds at all times. In this sbsection we first present, in Lemmas 4 to 6, properties of holding plans implied by i) and ii). Then, based on these properties we obtain or first proposition, which states the niqeness and existence of eqilibrim and gives the eqation for the corresponding price. Going back to the vale V (q), in eqation (6), and bearing in mind that a trader can choose any fnction q t, sbject to Conditions 1 and 2, it is clear that the trader intertemporal problem redces to pointwise optimization. That is, a trader whose last information event occrred at time t chooses her asset holding at time, q t,, in order to maximize the difference 15
16 between her expected valation for the asset and the corresponding holding cost: E t [v(θ,q t, )] ξ q t,. (9) Now, for all traders, tilities are strictly increasing for q t, < 1 and constant for q t, 1. So, if one trader finds it optimal to hold strictly more than one nit at time, thenitmstbethat ξ, implying that all other traders find it optimal to hold more than one nit. Inspecting eqation (8), one sees that in that case the market cannot clear since s<1. We conclde that: Lemma 3. In eqilibrim, ξ > and q t, [, 1] for all traders. To obtain frther insights on holding plans, consider first a time t high valation trader, i.e., a trader who finds ot at time t that θ t = h. Schtraderknowsthathervalationforthe asset will stay high forever. Hence E t [v(θ,q t, )] = v(h, q t, ), t. (1) Next, consider a time t low valation trader, i.e. a trader who finds ot at time t that θ t =. Thistraderanticipatesthathertilitystatswillremainlowbytime with probability (1 µ h )/(1 µ ht ). Hence: E t [v(θ,q t, )] = q t, δ 1 µ h qt, 1+σ 1 µ ht 1+σ, q t, [, 1] (11) Comparing (1) and (11), one sees that, for all asset holdings in (, 1), high valation traders have a niformly higher marginal tility than low valation traders. Now let S ρe ρ( t) (s µ ht ) dt, (12) the gross asset spply broght by all traders mins the maximm (nit) demand of high valation traders, integrating across all traders with at least one information event. Given the definition of S, and based on the above ranking of marginal tilities, one sees that the economy can be in one of two regimes. The first regime arises if S > : in that case, since q t, 1 for high valation traders, market clearing implies that q t, > for some low valation trader. Bt then all high valation traders mst hold one nit, since they have niformly higher marginal tility for holdings in [, 1]. The second regime arises if S <. In this case market clearing implies that some high valation trader mst find it optimal to hold strictly less than 16
17 one share: sch trader either strictly prefer to hold zero share, or is indifferent between any holding in [, 1]. Bt since high valation traders have niformly higher marginal tility for holdings in [, 1], this implies that all low valation traders hold zero shares. Smmarizing: Lemma 4. Let T f be the niqe strictly positive soltion of S =. Then: if (,T f ) then, for all t (,], θ t = h implies q t, =1; if [T f, ) then, for all t (,], θ t = implies q t, =. Next, consider the demand of high valation traders when >T f. We know from the previos lemma that low valation traders hold no asset. Ths, high valation traders mst hold the entire asset spply. Moreover, since S <, the marketclearing condition implies that some high valation traders mst hold strictly less than one share. Keeping in mind that high valation traders have the same linear tility flow over [, 1], this implies they mst be indifferent between any holding in [, 1]. Ths we can state the following lemma. Lemma 5. For all >T f, the average asset holding of a high valation trader is ρe ρ( t) sdt ρe ρ( t) µ ht dt, bt the distribtion of asset holdings across high valation traders is indeterminate. Now trn to the demand of low valation traders when <T f. Taking firstorder conditions when θ t = in (9), we obtain, given q t, [, 1]: q t, = ifξ 1 (13) q t, =1 ifξ 1 δ 1 µ h (14) 1 µ ht q t, =(1 µ ht ) 1/σ Q if ξ 1 δ 1 µ 1/σ h 1 ξ, 1, where Q. (15) 1 µ ht δ(1 µ h ) Eqation (13) states that low valation traders hold zero nit if the opportnity cost of holding the asset is greater than 1, their highest possible marginal tility, which arises when q =. Eqation (14) states that low valation traders hold one nit if the opportnity cost of holding the asset is below the lowest possible marginal tility, which arises when q =1. Lastly,eqation (15) pins down a low valation trader s holdings in the intermediate interior case by eqating to the derivative of (11) with respect to q t,. 17
18 As discssed above, prior to time T f the holdings of some low valation traders mst be strictly greater than zero: ths, holdings are determined by either (14) or (15) and ξ >. By the definition of Q, ξ 1 δ(1 µ h )/(1 µ ht )ifandonlyif(1 µ ht ) 1/σ Q 1. Hence, the asset demand defined by (14) and (15) can be written as q t, =min{(1 µ ht ) 1/σ Q, 1}. (16) Sbstitting the demand from (16) into the marketclearing condition (8) and sing the definition of S in (12), the following lemma obtains. Lemma 6. If (,T f ), then for all t (,], θ t = implies q t, =min{(1 µ ht ) 1/σ Q, 1} where: (1 µ ht )min{(1 µ ht ) 1/σ Q, 1}ρe ρ( t) dt = S. (17) Eqation (17) is a oneeqationinonenknown for Q that is shown in the proof appendix to have a niqe soltion. Taken together, Lemmas 4 throgh 6 imply: Proposition 1. There exists an eqilibrim. The eqilibrim asset allocation is niqe p to the distribtion of asset holdings across high valation traders after T f, and is characterized by Lemma 46. The eqilibrim price path is niqe, is increasing ntil T f, constant thereafter, and solves the following ODE: (,T f ): rp ṗ =1 δ(1 µ h )Q σ (18) [T f, ) : p = 1 r. (19) As in the perfect cognition case, the price deterministically increases ntil it reaches 1/r. One difference is that, while nder perfect cognition this recovery occrs at time T s (defined in eqation (2)), with limited cognition it occrs at the later time T f >T s. For <T f,the time low valation traders are the marginal investors, and the eqilibrim price is sch that their marginal valation is eqal to the opportnity cost of holding the asset, as stated by (18). For >T f, the entire spply is held by high valation investors. Ths the eqilibrim price is eqal to the present vale of their tility flow, as stated by (19). 22 This proposition is illstrated 22 They mst be indifferent between trading or not. This indifference condition implies that 1 rp +ṗ =. And, p =1/r is the only bonded and positive soltion of this ODE. 18
19 in Figre 2, Panel B, which plots the eqilibrim price nder limited cognition. Note that for this nmerical analysis we set the intensity of information events ρ to 25, which means that traders observe refreshed information on θ on average once a day. This is a plasible freqency, given the time it takes to collect and aggregate information across desks, departments and markets in a financial instittion. 3.3 Eqilibrim properties In this sbsection we discss the properties of the eqilibrim price and trades and compare them to their nbonded cognition conterparts Welfare To stdy welfare we define an asset holding plan to be feasible if it satisfies Conditions 1 and 2 as well as the resorce constraint, which is eqivalent to the marketclearing condition (7). Frthermore, we say that an asset holding plan q Pareto dominates some other holding plan q if it is possible to generate a Pareto improvement by switching from q to q while making time zero transfers among traders. Becase tilities are qasi linear, q Pareto dominates q only if W (q) >W(q ), where if and W (q) =E e r v(θ τ,q τ,) d. (2) The next proposition states that in or model the first welfare theorem holds: Proposition 2. The holding plan arising in the eqilibrim characterized in Proposition 1 maximizes W (q) among all feasible holding plans. This proposition reflects that, in or setp, there are no externalities, in that the holdings constraints imposed by limited cognition for one agent (and expressed in conditions 1 and 2) do not depend on the actions of other agents. These constraints translate into simple restrictions on the commodity space (conditions 1 and 2), allowing s to apply the standard proof of the first welfare theorem (see MasColell, Whinston, and Green, 1995, Chapter 16) Holdings As set forth in eqation (16), for a trader observing at t that her valation is low, the optimal holdings at time >tare (weakly) increasing in Q.Relyingonthemarketclearingcondition, 19
20 the next proposition spells ot the properties of Q. Proposition 3. The fnction Q is continos, sch that Q + = s and Q Tf s σ 1+σ =. Moreover, if (21) Q is strictly decreasing with time. Otherwise, it is hmpshaped. The economic intition is as follows. At time + the mass of traders with high valation is negligible. Therefore low valation traders mst absorb the entire spply. Hence, Q + = s. At time T f high valation traders absorb the entire spply. Hence, Q Tf =. When the per capita spply of assets affected by the shock s is low, so that (21) holds, the incoming flow of high valation traders reaching a decision at a given point in time is always large enogh to accommodate the spply from low valation traders. Correspondingly, in eqilibrim low valation traders sell a lmp of their assets when they reach a decision, then smoothly nwind their inventory ntil the next information event. In contrast, when s is so large that (21) fails to hold, the liqidity shock is more severe. Hence, shortly after the initial aggregate shock, the inflow of high valation traders is not large enogh to absorb the sales of low valation traders who crrently reach a decision. In eqilibrim, some of these sales are absorbed by early low valation traders who reached adecisionattimet<and have not had another information event. Indeed, these early low valation traders anticipate that, as time goes by, their instittion is more likely to have recovered. Ths, their expected valation (in the absence of an information event) increases with time and they find it optimal to by if their tility is not too concave, i.e., if σ is not too high. Correspondingly, near time zero, Q is increasing, as depicted in Figre 3 for σ =.3 and.5. Combining Lemma 4, Lemma 5, Lemma 6 and Proposition 3, one obtains a fll characterization of the eqilibrim holdings process, which can be compared to its conterpart in the nlimited cognition case. When cognition is not limited, as long as an instittion has not recovered from the shock, its holdings decline smoothly, and, as soon as it recovers, its holdings jmp to 1. Trading histories are qite different with limited cognition. First an instittion s holdings remain constant ntil its trader s first information event. Then, if at her first information event the trader learns that her instittion has a low valation, the trader sells a lmp. After that, if (21) does not hold, the trader progressively bys back, then eventally sells ot 2
Bonds with Embedded Options and Options on Bonds
FIXEDINCOME SECURITIES Chapter 14 Bonds with Embedded Options and Options on Bonds Callable and Ptable Bonds Instittional Aspects Valation Convertible Bonds Instittional Aspects Valation Options on Bonds
More informationHerzfeld s Outlook: Seasonal Factors Provide Opportunities in ClosedEnd Funds
VIRTUS HERZFELD FUND Herzfeld s Otlook: Seasonal Factors Provide Opportnities in ClosedEnd Fnds When it comes to investing in closedend fnds, a comprehensive nderstanding of the inefficiencies of the
More informationInterDealer Trading in Financial Markets*
S. Viswanathan Dke University James J. D. Wang Hong Kong University of Science and Technology InterDealer Trading in Financial Markets* I. Introdction Trading between dealers who act as market makers
More informationResource Pricing and Provisioning Strategies in Cloud Systems: A Stackelberg Game Approach
Resorce Pricing and Provisioning Strategies in Clod Systems: A Stackelberg Game Approach Valeria Cardellini, Valerio di Valerio and Francesco Lo Presti Talk Otline Backgrond and Motivation Provisioning
More informationGUIDELINE. Guideline for the Selection of Engineering Services
GUIDELINE Gideline for the Selection of Engineering Services 1998 Mission Statement: To govern the engineering profession while enhancing engineering practice and enhancing engineering cltre Pblished by
More informationCandidate: Shawn Mullane. Date: 04/02/2012
Shipping and Receiving Specialist / Inventory Control Assessment Report Shawn Mllane 04/02/2012 www.resorceassociates.com To Improve Prodctivity Throgh People. Shawn Mllane 04/02/2012 Prepared For: NAME
More informationUsing GPU to Compute Options and Derivatives
Introdction Algorithmic Trading has created an increasing demand for high performance compting soltions within financial organizations. The actors of portfolio management and ris assessment have the obligation
More informationRollup/down options
Rollp/down options INTRODUCTION Rollp and rolldown options are options whose strike is atomatically reset pwards or downwards while transformed into pandot or downandot options. More specifically,
More informationDesigning and Deploying File Servers
C H A P T E R 2 Designing and Deploying File Servers File servers rnning the Microsoft Windows Server 2003 operating system are ideal for providing access to files for sers in medim and large organizations.
More informationEquilibrium of Forces Acting at a Point
Eqilibrim of orces Acting at a Point Eqilibrim of orces Acting at a Point Prelab Qestions 1. What is the definition of eqilibrim? Can an object be moving and still be in eqilibrim? Explain.. or this lab,
More informationPurposefully Engineered HighPerforming Income Protection
The Intelligent Choice for Disability Income Insrance Prposeflly Engineered HighPerforming Income Protection Keeping Income strong We engineer or disability income prodcts with featres that deliver benefits
More informationSingleYear and MultiYear Insurance Policies in a Competitive Market
SingleYear and MltiYear Insrance Policies in a Competitive Market Pal R. Kleindorfer INSEAD, France Howard Knrether The Wharton School University of Pennsylvania Chieh OYang City University of Hong
More information10 Evaluating the Help Desk
10 Evalating the Help Desk The tre measre of any society is not what it knows bt what it does with what it knows. Warren Bennis Key Findings Help desk metrics having to do with demand and with problem
More informationThe Intelligent Choice for Basic Disability Income Protection
The Intelligent Choice for Basic Disability Income Protection provider Pls Limited Keeping Income strong We prposeflly engineer or basic disability income prodct to provide benefitrich featres delivering
More informationCandidate: Charles Parker. Date: 01/29/2015
Software Developer / Programmer Assessment Report 01/29/2015 www.resorceassociates.com To Improve Prodctivity Throgh People. Janary 29, 2015 01/29/2015 The following pages represent a report based on the
More informationThe Boutique Premium. Do Boutique Investment Managers Create Value? AMG White Paper June 2015 1
The Botiqe Premim Do Botiqe Investment Managers Create Vale? AMG White Paper Jne 2015 1 Exective Smmary Botiqe active investment managers have otperformed both nonbotiqe peers and indices over the last
More informationResearch on Pricing Policy of Ebusiness Supply Chain Based on Bertrand and Stackelberg Game
International Jornal of Grid and Distribted Compting Vol. 9, No. 5 (06), pp.0 http://dx.doi.org/0.457/ijgdc.06.9.5.8 Research on Pricing Policy of Ebsiness Spply Chain Based on Bertrand and Stackelberg
More informationChapter 3. 2. Consider an economy described by the following equations: Y = 5,000 G = 1,000
Chapter C evel Qestions. Imagine that the prodction of fishing lres is governed by the prodction fnction: y.7 where y represents the nmber of lres created per hor and represents the nmber of workers employed
More informationCorporate performance: What do investors want to know? Innovate your way to clearer financial reporting
www.pwc.com Corporate performance: What do investors want to know? Innovate yor way to clearer financial reporting October 2014 PwC I Innovate yor way to clearer financial reporting t 1 Contents Introdction
More informationDESTINATION ASSURED CONTACT US. Products for Life
DESTINATION ASSURED CONTACT US For more information abot any of the services in this brochre, call 18007484302, visit or website at www.mac.com or stop by the branch nearest yo. LR2011 Federally insred
More informationOptimal Trust Network Analysis with Subjective Logic
The Second International Conference on Emerging Secrity Information, Systems and Technologies Optimal Trst Network Analysis with Sbjective Logic Adn Jøsang UNIK Gradate Center, University of Oslo Norway
More information11 Success of the Help Desk: Assessing Outcomes
11 Sccess of the Help Desk: Assessing Otcomes I dread sccess... I like a state of continal becoming, with a goal in front and not behind. George Bernard Shaw Key Findings Respondents help desks tend to
More informationOptimal Personalized Filtering Against SpearPhishing Attacks
Optimal Personalized Filtering Against SpearPhishing Attacks Aron Laszka and Yevgeniy Vorobeychik and Xenofon Kotsokos Institte for Software Integrated Systems Department of Electrical Engineering and
More informationThe Intelligent Choice for Disability Income Protection
The Intelligent Choice for Disability Income Protection provider Pls Keeping Income strong We prposeflly engineer or disability income prodct with featres that deliver benefits sooner and contine paying
More informationEffective governance to support medical revalidation
Effective governance to spport medical revalidation A handbook for boards and governing bodies This docment sets ot a view of the core elements of effective local governance of the systems that spport
More informationSpectrum Balancing for DSL with Restrictions on Maximum Transmit PSD
Spectrm Balancing for DSL with Restrictions on Maximm Transmit PSD Driton Statovci, Tomas Nordström, and Rickard Nilsson Telecommnications Research Center Vienna (ftw.), DonaCityStraße 1, A1220 Vienna,
More informationMSc and MA in Finance and Investment online Study an online MSc and MA in Finance and Investment awarded by UNINETTUNO and Geneva Business School
MSc and MA in Finance and Investment online Stdy an online awarded by UNINETTUNO and Geneva Bsiness School Awarded by Geneva Bsiness School Geneva Barcelona Moscow Class profile The connects yo with stdents
More informationMotorola Reinvents its Supplier Negotiation Process Using Emptoris and Saves $600 Million. An Emptoris Case Study. Emptoris, Inc. www.emptoris.
Motorola Reinvents its Spplier Negotiation Process Using Emptoris and Saves $600 Million An Emptoris Case Stdy Emptoris, Inc. www.emptoris.com VIII03/3/05 Exective Smmary With the disastros telecommnication
More informationStock Market Liquidity and MacroLiquidity Shocks: Evidence from the 20072009 Financial Crisis
Stock Market Liqidity and MacroLiqidity Shocks: Evidence from the 20072009 Financial Crisis Chris Florackis *, Alexandros Kontonikas and Alexandros Kostakis Abstract We develop an empirical framework
More informationKentucky Deferred Compensation (KDC) Program Summary
Kentcky Deferred Compensation (KDC) Program Smmary Smmary and Highlights of the Kentcky Deferred Compensation (KDC) Program Simple. Smart. For yo. For life. 457 Plan 401(k) Plan Roth 401(k) Deemed Roth
More informationEvery manufacturer is confronted with the problem
HOW MANY PARTS TO MAKE AT ONCE FORD W. HARRIS Prodction Engineer Reprinted from Factory, The Magazine of Management, Volme 10, Nmber 2, Febrary 1913, pp. 135136, 152 Interest on capital tied p in wages,
More informationDeploying Network Load Balancing
C H A P T E R 9 Deploying Network Load Balancing After completing the design for the applications and services in yor Network Load Balancing clster, yo are ready to deploy the clster rnning the Microsoft
More informationBorrowing for College. Table of contents. A guide to federal loans for higher education
Borrowing for College A gide to federal loans for higher edcation Table of contents Edcation loan basics 2 Applying for edcation loans 3 Repaying edcation loans 3 Controlling edcation loan debt 5 Glossary
More informationStock Market Liquidity and MacroLiquidity Shocks: Evidence from the 20072009 Financial Crisis
Stock Market Liqidity and MacroLiqidity Shocks: Evidence from the 20072009 Financial Crisis Chris Florackis *, Alexandros Kontonikas and Alexandros Kostakis Abstract We develop an empirical framework
More informationDesigning an Authentication Strategy
C H A P T E R 1 4 Designing an Athentication Strategy Most organizations need to spport seamless access to the network for mltiple types of sers, sch as workers in offices, employees who are traveling,
More information9 Setting a Course: Goals for the Help Desk
IT Help Desk in Higher Edcation ECAR Research Stdy 8, 2007 9 Setting a Corse: Goals for the Help Desk First say to yorself what yo wold be; and then do what yo have to do. Epictets Key Findings Majorities
More informationPosition paper smart city. economics. a multisided approach to financing the smart city. Your business technologists.
Position paper smart city economics a mltisided approach to financing the smart city Yor bsiness technologists. Powering progress From idea to reality The hman race is becoming increasingly rbanised so
More information8 Service Level Agreements
8 Service Level Agreements Every organization of men, be it social or political, ltimately relies on man s capacity for making promises and keeping them. Hannah Arendt Key Findings Only abot 20 percent
More informationCloser Look at ACOs. Making the Most of Accountable Care Organizations (ACOs): What Advocates Need to Know
Closer Look at ACOs A series of briefs designed to help advocates nderstand the basics of Accontable Care Organizations (ACOs) and their potential for improving patient care. From Families USA Updated
More information7 Help Desk Tools. Key Findings. The Automated Help Desk
7 Help Desk Tools Or Age of Anxiety is, in great part, the reslt of trying to do today s jobs with yesterday s tools. Marshall McLhan Key Findings Help desk atomation featres are common and are sally part
More informationA Spare Part Inventory Management Model for Better Maintenance of Intelligent Transportation Systems
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 A Spare Part Inventory Management Model for Better Maintenance of Intelligent
More informationASAND: Asynchronous Slot Assignment and Neighbor Discovery Protocol for Wireless Networks
ASAND: Asynchronos Slot Assignment and Neighbor Discovery Protocol for Wireless Networks Fikret Sivrikaya, Costas Bsch, Malik MagdonIsmail, Bülent Yener Compter Science Department, Rensselaer Polytechnic
More informationSickness Absence in the UK: 19842002
Sickness Absence in the UK: 19842002 Tim Barmby (Universy of Drham) Marco Ecolani (Universy of Birmingham) John Treble (Universy of Wales Swansea) Paper prepared for presentation at The Economic Concil
More informationSEGREGATED ACCOUNTS COMPANIES ACE CAPABILITIES: AN OVERVIEW
SEGREGATED ACCOUNTS COMPANIES CAPABILITIES: AN OVERVIEW SIMPLICITY OUT OF COMPLEXITY SEGREGATED ACCOUNTS CAPABILITIES Managing yor own risks jst got simpler. In recent years, increasing reglation has led
More informationOn the urbanization of poverty
On the rbanization of poverty Martin Ravallion 1 Development Research Grop, World Bank 1818 H Street NW, Washington DC, USA Febrary 001; revised Jly 001 Abstract: Conditions are identified nder which the
More informationIntroduction to HBase Schema Design
Introdction to HBase Schema Design Amandeep Khrana Amandeep Khrana is a Soltions Architect at Clodera and works on bilding soltions sing the Hadoop stack. He is also a coathor of HBase in Action. Prior
More informationMake the College Connection
Make the College Connection A college planning gide for stdents and their parents Table of contents The compelling case for college 2 Selecting a college 3 Paying for college 5 Tips for meeting college
More informationFINANCIAL FITNESS SELECTING A CREDIT CARD. Fact Sheet
FINANCIAL FITNESS Fact Sheet Janary 1998 FL/FF02 SELECTING A CREDIT CARD Liz Gorham, Ph.D., AFC Assistant Professor and Family Resorce Management Specialist, Utah State University Marsha A. Goetting,
More informationDirect Loan Basics & Entrance Counseling Guide. For Graduate and Professional Student Direct PLUS Loan Borrowers
Direct Loan Basics & Entrance Conseling Gide For Gradate and Professional Stdent Direct PLUS Loan Borrowers DIRECT LOAN BASICS & ENTRANCE COUNSELING GUIDE For Gradate and Professional Stdent Direct PLUS
More informationPlanning and Implementing An Optimized Private Cloud
W H I T E PA P E R Intelligent HPC Management Planning and Implementing An Optimized Private Clod Creating a Clod Environment That Maximizes Yor ROI Planning and Implementing An Optimized Private Clod
More informationPeriodized Training for the Strength/Power Athlete
Periodized Training for the /Power Athlete Jay R. Hoffman, PhD, FACSM, CSCS *D The se of periodized training has been reported to go back as far as the ancient Olympic games. Its basic premise is that
More informationCloser Look at ACOs. Putting the Accountability in Accountable Care Organizations: Payment and Quality Measurements. Introduction
Closer Look at ACOs A series of briefs designed to help advocates nderstand the basics of Accontable Care Organizations (ACOs) and their potential for improving patient care. From Families USA Janary 2012
More informationCandidate: Kevin Taylor. Date: 04/02/2012
Systems Analyst / Network Administrator Assessment Report 04/02/2012 www.resorceassociates.com To Improve Prodctivity Throgh People. 04/02/2012 Prepared For: Resorce Associates Prepared by: John Lonsbry,
More informationDIRECT TAX LAWS Taxability of Capital Gains on Byback of Shares  Debate ignites after AAR s rling in RST s case BACKGROUND 1. Recently, the Athority for Advance Rlings ( AAR ) in the case of RST, In
More informationeconstor zbw www.econstor.eu
econstor www.econstor.e Der OpenAccessPblikationsserver der ZBW LeibnizInformationszentrm Wirtschaft The Open Access Pblication Server of the ZBW Leibniz Information Centre for Economics MacMinn, Richard;
More informationPlanning a Managed Environment
C H A P T E R 1 Planning a Managed Environment Many organizations are moving towards a highly managed compting environment based on a configration management infrastrctre that is designed to redce the
More informationCandidate: Suzanne Maxwell. Date: 09/19/2012
Medical Coder / Billing Clerk Assessment Report Szanne Maxwell 09/19/2012 www.resorceassociates.com Szanne Maxwell 09/19/2012 Prepared For: NAME Prepared by: John Lonsbry, Ph.D. & Lcy Gibson, Ph.D., Licensed
More informationCandidate: Kyle Jarnigan. Date: 04/02/2012
Cstomer Service Manager Assessment Report 04/02/2012 www.resorceassociates.com To Improve Prodctivity Throgh People. Cstomer Service Manager Assessment Report 04/02/2012 Prepared For: NAME Prepared by:
More informationCERCLA Section 108(b) Financial Responsibility. A public webinar hosted by the United States Environmental Protection Agency May 17, 2016
CERCLA Section 108(b) Financial Responsibility A pblic webinar hosted by the United States Environmental Protection Agency May 17, 2016 Presentation Overview Backgrond: CERCLA Section 108(b) Financial
More informationOpening the Door to Your New Home
Opening the Door to Yor New Home A Gide to Bying and Financing. Contents Navigating Yor Way to Home Ownership...1 Getting Started...3 Finding Yor Home...9 Finalizing Yor Financing...12 Final Closing...13
More informationRegular Specifications of Resource Requirements for Embedded Control Software
Reglar Specifications of Resorce Reqirements for Embedded Control Software Rajeev Alr and Gera Weiss University of Pennsylvania Abstract For embedded control systems a schedle for the allocation of resorces
More informationCRM Customer Relationship Management. Customer Relationship Management
CRM Cstomer Relationship Management Farley Beaton Virginia Department of Taxation Discssion Areas TAX/AMS Partnership Project Backgrond Cstomer Relationship Management Secre Messaging Lessons Learned 2
More informationTowers Watson Manager Research
Towers Watson Manager Research How we se fnd performance data Harald Eggerstedt 13. März 212 212 Towers Watson. All rights reserved. Manager selection at Towers Watson The goal is to find managers that
More informationHigh Availability for Microsoft SQL Server Using DoubleTake 4.x
High Availability for Microsoft SQL Server Using DobleTake 4.x High Availability for Microsoft SQL Server Using DobleTake 4.x pblished April 2000 NSI and DobleTake are registered trademarks of Network
More informationHigh Availability for Internet Information Server Using DoubleTake 4.x
High Availability for Internet Information Server Using DobleTake 4.x High Availability for Internet Information Server Using DobleTake 4.x pblished April 2000 NSI and DobleTake are registered trademarks
More informationCloser Look at ACOs. Designing ConsumerFriendly Beneficiary Assignment and Notification Processes for Accountable Care Organizations
Closer Look at ACOs A series of briefs designed to help advocates nderstand the basics of Accontable Care Organizations (ACOs) and their potential for improving patient care. From Families USA Janary 2012
More informationWHITE PAPER. Filter Bandwidth Definition of the WaveShaper Sseries Programmable Optical Processor
WHITE PAPER Filter andwidth Definition of the WaveShaper Sseries 1 Introdction The WaveShaper family of s allow creation of sercstomized filter profiles over the C or L band, providing a flexible tool
More informationMarket Impact and Optimal Equity Trade Scheduling
Market Impact and Optimal Eqity Trade Schedling Dan dibartolomeo Northfield Information Services, Inc. Colmbia University Math Finance Seminar September 2007 Presentation Otline Brief review of the optimal
More informationIsilon OneFS. Version 7.1. Backup and recovery guide
Isilon OneFS Version 7.1 Backp and recovery gide Copyright 20132014 EMC Corporation. All rights reserved. Pblished in USA. Pblished March, 2014 EMC believes the information in this pblication is accrate
More informationPlanning an Active Directory Deployment Project
C H A P T E R 1 Planning an Active Directory Deployment Project When yo deploy the Microsoft Windows Server 2003 Active Directory directory service in yor environment, yo can take advantage of the centralized,
More informationRoth 401(k) and Roth 403(b) Accounts: Pay Me Now or Pay Me Later Why a Roth Election Should Be Part of Your Plan Now
Reprinted with permission from the Society of FSP. Reprodction prohibited withot pblisher's written permission. Roth 401(k) and Roth 403(b) Acconts: Why a Roth Election Shold Be Part of Yor Plan Now by
More informationOn the Equivalence Between a Commonly Used Correlation Coefficient and a Least Squares Function
On the Eqivalence Between a Commonly Used Correlation Coefficient and a Least Sqares Fnction Diane C. Jamrog George N. Phillips, Jr. Yin Zhang Janary 2003 (revised October 2003) Abstract Many objective
More informationCandidate: Cassandra Emery. Date: 04/02/2012
Market Analyst Assessment Report 04/02/2012 www.resorceassociates.com To Improve Prodctivity Throgh People. 04/02/2012 Prepared For: Resorce Associates Prepared by: John Lonsbry, Ph.D. & Lcy Gibson, Ph.D.,
More information3.2 The Stress Function Method
Section.. The Stress Fnction Method An effective wa of dealing with man two dimensional problems is to introdce a new nknown, the Air stress fnction, an idea broght to s b George Air in 186. The stresses
More informationHSBC Internet Banking. Combined Product Disclosure Statement and Supplementary Product Disclosure Statement
HSBC Internet Banking Combined Prodct Disclosre Statement and Spplementary Prodct Disclosre Statement AN IMPORTANT MESSAGE FOR HSBC CUSTOMERS NOTICE OF CHANGE For HSBC Internet Banking Combined Prodct
More informationwww.macu.com LH2011 Products for Life
www.mac.com LH2011 Prodcts for Life 2 6 8 10 12 13 prchase Qick 14 insrance There s nothing like the feeling of bying yor own home. Yo finally have yor own space, and yor money is going toward something
More informationHorses and Rabbits? Optimal Dynamic Capital Structure from Shareholder and Manager Perspectives
Horses and Rabbits? Optimal Dynamic Capital trctre from hareholder and Manager Perspectives Nengji J University of Maryland Robert Parrino University of exas at Astin Allen M. Poteshman University of Illinois
More informationThe Institute Of Commercial Management. Prospectus. Start Your Career Here! www.icm.ac.uk info@icm.ac.uk
The Institte Of Commercial Management Prospects Start Yor Career Here! www.icm.ac.k info@icm.ac.k The fondation of every state is the edcation of it s yoth. Diogenes Laertis Welcome... Althogh we are
More informationCONTACT US. The Financial ABCs for Raising a Family
The Financial ABCs for Raising a Family CONTACT US For more information abot any of the in this brochre, call 18007484302, visit or at www.mac.com or stop by the branch nearest yo. Federally insred
More informationForm M1 Report for Multiple Employer Welfare Arrangements (MEWAs) and Certain Entities Claiming Exception (ECEs)
U.S. Department of Labor Employee Benefits Secrity Administration Room N5511 200 Constittion Avene, NW Washington, DC 20210 P450 Form M1 Report for Mltiple Employer Welfare Arrangements (MEWAs) and Certain
More information3 Building Blocks Of Optimized Price & Promotion Strategies
3 Bilding Blocks Of Optimized Price & Promotion Strategies Boosting Brand Loyalty And Profits With Visal Analytics Sponsored by Ebook Table of contents Introdction... 3 Consolidate And Analyze Data From
More informationOn a Generalized Graph Coloring/Batch Scheduling Problem
Reglar Papers On a Generalized Graph Coloring/Batch Schedling Problem Giorgio Lcarelli 1, Ioannis Milis Dept. of Informatics, Athens University of Economics and Bsiness, 104 34, Athens, Greece, {glc, milis}@aeb.gr
More informationInvestment under Uncertainty and Regulation of New Access Networks
Discssion Paper o. 00 Investment nder Uncertainty and Reglation of ew Access etworks Roman Inderst and Martin Peitz Discssion Paper o. 00 Investment nder Uncertainty and Reglation of ew Access etworks
More informationEnabling Advanced Windows Server 2003 Active Directory Features
C H A P T E R 5 Enabling Advanced Windows Server 2003 Active Directory Featres The Microsoft Windows Server 2003 Active Directory directory service enables yo to introdce advanced featres into yor environment
More informationKEYS TO BEING AN EFFECTIVE WORKPLACE PERSONAL ASSISTANT
5 KEYS TO BEING AN EFFECTIVE WORKPLACE PERSONAL ASSISTANT by: John Barrett Personal assistants (PAs) and their ability to effectively provide essential spports at the workplace are extremely important
More information20142015 SUGGESTED LENDER LIST FOR PRIVATE LOAN OPTIONS AVAILABLE TO INTERNATIONAL GRADUATE STUDENTS
The chart below otlines the lender contact information and the lender description of the fee strctre, interest rates, and borrower benefits associated with each lender s private loan prodct for loans disbrsed
More informationThe Inevitability of MarketWide Underpricing of Mortgage Default Risk
The Inevitability of MarketWide Underpricin of Mortae Defalt Risk Forthcomin Real Estate Economics Andrey Pavlov Simon Fraser University 8888 University Dr., Brnaby, BC V5A 1S6, Canada Email: apavlov@sf.ca,
More informationA Contemporary Approach
BORICP01.doc  1 Second Edition Edcational Psychology A Contemporary Approach Gary D. Borich The University of Texas at Astin Martin L. Tombari University of Denver (This pblication may be reprodced for
More informationA Message from the CEO
October 2007 A Message from the CEO For the past 53 years, we have been recognized by or logo, name, competitive rates, and qality member service. All these have been vital to the strength of or organization.
More informationThe Good Governance Standard for Public Services
The Good Governance Standard for Pblic Services The Independent Commission for Good Governance in Pblic Services The Independent Commission for Good Governance in Pblic Services, chaired by Sir Alan Langlands,
More informationCompensation Approaches for Farfield Speaker Identification
Compensation Approaches for Farfield Speaer Identification Qin Jin, Kshitiz Kmar, Tanja Schltz, and Richard Stern Carnegie Mellon University, USA {qjin,shitiz,tanja,rms}@cs.cm.ed Abstract While speaer
More informationReview Finite Element Modeling
Strctral Dynamic Modeling Techniqes & Modal Analysis Methods [ M ] [ K ] n [ M ] n a [ K ] a [ ω ] [ E a ] Review Finite Element Modeling Peter Avitabile Mechanical Engineering Department University of
More informationA guide to safety recalls in the used vehicle industry GUIDE
A gide to safety recalls in the sed vehicle indstry GUIDE Definitions Aftermarket parts means any prodct manfactred to be fitted to a vehicle after it has left the vehicle manfactrer s prodction line.
More information1. Introduction. Keywords Monte Carlo Simulation, Risk Analysis, Expected Earnings. H. I. Okagbue, S. O. Edeki *, A. A. Opanuga
American Jornal of omptational and Applied Mathematics 2014, 4(5): 161166 DOI: 105923/jajcam2014040502 A Monte arlo Simlation Approach in Assessing Risk and Uncertainty Involved in Estimating the Expected
More informationFacilities. Car Parking and Permit Allocation Policy
Facilities Car Parking and Permit Allocation Policy Facilities Car Parking and Permit Allocation Policy Contents Page 1 Introdction....................................................2 2.0 Application
More informationBuilding Trust How Banks are Attracting and Retaining Business Clients With Institutional Money Fund Portals
Bilding Trst How Banks are Attracting and Retaining Bsiness Clients With Instittional Money Fnd Portals By George Hagerman, Fonder and CEO, CacheMatrix Holdings, LLC C ompetitive pressres are driving innovation
More informationSME Business. Solutions
SME Bsiness Soltions For more information, please contact SME Banking: Bahrain World Trade Center, East Tower 15th Floor Tel: 17 13 3383 Fax: 17 13 1043 Email: sme@bmi.com.bh ABOUT BMI BANK We re a yong,
More informationFaceTrust: Assessing the Credibility of Online Personas via Social Networks
FaceTrst: Assessing the Credibility of Online Personas via Social Networks Michael Sirivianos Kyngbaek Kim Xiaowei Yang Dke University University of California, Irvine Dke University msirivia@cs.dke.ed
More informationCollege Planning 101: your seat is saved
: yor seat is saved www.mac.com LC2011 Prodcts for Life 2 7 11 paying for Sddenly yo re off on yor own. Yo have a 1 whole bnch of new financial responsibilities paying for tition, books, bills, room and
More information