SWEDISH SCHOOL OF ECONOMICS AND BUSINESS ADMINISTRATION THE YRJÖ JAHNSSON WORKING PAPER SERIES IN INDUSTRIAL ECONOMICS 2 (2003) - PDF

Description
SWEDISH SCHOOL OF ECONOMICS AND BUSINESS ADMINISTRATION THE YRJÖ JAHNSSON WORKING PAPER SERIES IN INDUSTRIAL ECONOMICS (3) Thomas Gehrig & Rune Stenbacka SCREENING CYCLES 3 Swedish School of Economics

Please download to get full document.

View again

of 27
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Instruction manuals

Publish on:

Views: 45 | Pages: 27

Extension: PDF | Download: 0

Share
Transcript
SWEDISH SCHOOL OF ECONOMICS AND BUSINESS ADMINISTRATION THE YRJÖ JAHNSSON WORKING PAPER SERIES IN INDUSTRIAL ECONOMICS (3) Thomas Gehrig & Rune Stenbacka SCREENING CYCLES 3 Swedish School of Economics and Business Administration Thomas Gehrig & Rune Stenbacka Distributor: Library Swedish School of Economics and Business Administration P.O.Box Helsinki Finland Phone: , Fax: SHS intressebyrå IB (Oy Casa Security Ab), Helsingfors 3 ISBN SCREENING CYCLES * Thomas Gehrig University of Freiburg and CEPR, London Rune Stenbacka Swedish School of Economics, Helsinki May 3, 3 We demonstrate how endogenous information acquisition by financiers creates investment cycles when competing financiers undertake their screening decisions in an uncoordinated way, thereby highlighting the role of intertemporal screening externalities induced by screening competition as a structural source of instability. We show that uncoordinated screening behavior of competing financiers may be not only the source of an important financial multiplier, but also an independent source of fluctuations inducing investment cycles. The screening cycle mechanism is robust to generalizations along many dimensions. * We are grateful for the comments by seminar participants at the Swedish School of Economics, London School of Economics, ESSEC Paris, University of Frankfurt, Ente Luigi Einaudi, Fondation Banque de France, Deutsche Bundesbank, Mannheim University and the EEA meeting in Venice. We are particularly grateful to Elena Carletti, Hans Gersbach, Luigi Guiso, Martin Hellwig, Heinz Herrman, Dima Leshchinski, Andrei Sarychev and Hannu Vartiainen for constructive comments and suggestions. Philip Jung and Almira Buzaushina provided excellent research assistance. Financial support of the Academy of Finland, the DAAD, the Deutsche Forschungsgemeinschaft (DFG), the Hanken Foundation and the Yrjö Jahnsson Foundation is gratefully acknowledged. - Correspondence: Thomas Gehrig: University of Freiburg, D-7985 Freiburg, Germany, Rune Stenbacka: Swedish School of Economics, P.O.Box 479, 11 Helsinki, Finland, I. Introduction At the doorstep of the information economy many observers have repeatedly emphasized the critical importance of innovation and human capital as promotors of sustained firm profitability as well as economic and job growth. In most cases, however, the transformation of innovative ideas into viable business ventures imposes considerable challenges to the financial system. Namely, due to the inalienability of human capital, innovative activities and R&D-projects typically require unsecured funding, in particular in cases of start-ups. Many modern technology leaders like Amazon, Apple, Cisco, e- Bay, Genentech, Intel, Microsoft and Sun Microsystems are evidence for the success of the American financial system in channeling unsecured initial funding to promising highquality start-ups. In fact, all of these firms started with the financial support of venture capital firms. Interestingly, and possibly surprisingly, the funding activities of innovative startup companies are typically accompanied by a large degree of cyclicality. Hence, industry experts such as W. Sahlmann claim: Cycles are inevitable but not necessarily bad, provided that the players anticipate them and respond accordingly.... As more and more capital chases a limited set of solid opportunities, it inevitably leads to what our forebears called a 'Tragedy of Commons'--too many cows feeding on the same pasture. When it does happen, I suspect we will be shocked, even though the inevitability of such cycles is clear. (cited in Jacobs, 1999). How can we make economic sense of such an experience? What is the mechanism that renders cycles in innovative activity inevitable? In the present study we provide a theory of unsecured funding, which under certain conditions generates funding cycles by necessity. Hence, we argue that decentralized screening may be the key to explain such cyclicality, which we refer to as screening cycles. Screening cycles arise, because uncoordinated financiers cannot identify applicants who have previously been rejected by competitors. The rejection decision of a financier typically imposes a negative externality on rival financiers, since a rejected applicant will typically submit his application to another financier. Accordingly, the total pool of applicants is adversely selected compared to the original pool of incoming projects. Thus the screening activity of financiers necessarily generates a pool-worsening externality, which can affect the whole funding industry. By cherry picking the best innovators in a narrowly specified market segment financiers leave an adversely selected pool for further screening by competitors at later rounds of evaluation. Thus after a period of intensive screening this activity will by necessity face diminishing returns in the immediately subsequent period(s). This screening externality will reduce the monitoring incentives for the whole industry until the pool of applicants has recovered to a sufficient extent through the entry of new innovators. When the negative externality is sufficiently strong there might be a phase when screening is no longer profitable. Hence financiers prefer to wait until the pool of projects has improved to a sufficient extent by the arrival of new projects. After a phase of inactivity this pool improvement ultimately triggers positive screening incentives in some later period leading the industry into cycling between states of high and low screening activities, thereby constituting the basis for screening-induced investment cycles. In our theory screening cycles emerge endogenously and by necessity unless the economy is in a stationary state where the optimal funding strategy is to constantly fund all projects without screening or to constantly grant no finance at all. The screening cycles result from uncoordinated search for creditworthy projects. In a more coordinated setting, i.e. in a cartel or under information sharing, endogenous cycles cannot occur. Screening cycles occur under conditions where the pools of applicants are adversely selected to a sufficient degree and where screening costs are significant. These conditions are most likely to be met in the venture capital industry. Bengtsson et al. () provide evidence about substantial costs of screening for a particular venture capital fund. They report that the acceptance rate in their sample ranges from to 5 percent 1, and that the screening process typically involves several rounds of increasing screening intensity by highly qualified experts. Thus screening typically involves significant delay in the order of several weeks and even months until approval of a single successful project. It also comprises significant opportunity costs in terms of expert salaries. The predictions of our theory seem to accord well with the cyclical features of the venture capital industry in the US. We provide novel evidence for a number of high tech 1 These numbers were communicated by P. Strömberg in a seminar at the Swedish School of Economics, Helsinki. 3 industries, which exhibit high frequency funding cycles with typical cycle durations of 3 to 6 months. Cycles of such a high frequency have not yet been documented nor been explained so far. Models addressing financial accelerator effects emphasize mechanisms whereby adverse shocks to the economy are endogenously amplified and propagated by credit market imperfections. These models are surveyed within a dynamic general equilibrium framework by, for example, Bernanke, Gertler and Gilchrist (). On an intuitive level already Fisher (1933) discussed how credit constraints propagate the effects of shocks on aggregate output and asset prices. According to Fisher, the more the private sector places emphasis on solving its debt problem the deeper the economy will be caught in a debt trap. In an influential recent article Kiyotaki and Moore (1997) constructed a model of a dynamic economy where borrowers' credit limits are affected by the prices of the collateralized assets. Their analysis shows how the dynamic interaction between credit limits and asset prices will constitute an important transmission mechanism whereby shocks to the economy persist, amplify and spill over across different sectors. In contrast to these theories focusing on how asset price fluctuations are amplified and propagated by credit market imperfections our theory does not require the existence of exogenous stochastic shocks. In the present analysis we propose a new mechanism, which is able to generate large, persistent and asymmetric fluctuations of economic activity in an otherwise stationary environment. This mechanism builds on endogeneous screening investments by specialized lenders engaged in repeated non-cooperative competition. Our mechanism does not require the existence of exogenous random shocks. We provide conditions for the existence of screening cycles in an otherwise completely stationary environment. While it is in general difficult to empirically identify investment cycles on an aggregate level, Gehrig and Stenbacka (b) document cycles of the length of only few quarters for numerous high-tech sectors for venture financing in the period of These cycles seem particularly prevalent in biotechnology, electronics, financial services, healthcare and medical services and consumer products. Moreover, and interestingly, in a number of prominent high technology segments the numbers of In the case of exogenous random shocks our mechanism will amplify these shocks and generate persistence similarly to the mechanisms surveyed so far. 4 venture financing deals do not seem to be directly affected by the stock market bubble in The empirical findings reported by Gehrig and Stenbacka (b) complement scant earlier evidence. Based on annual data Lerner et al. () find low frequency cycles in the U.S. biotechnological industry between Likewise Gompers and Lerner (1999, Fig. 1.1) find weak evidence of low-frequency cycles for aggregate U.S. venture capital data from Our model can also be viewed as a contribution to the literature on the relationship between banks' incentives for ex ante monitoring and lending market structure. The existing literature focusing on this relationship within the framework of a static context, for example, Gehrig (1998), Shaffer (1998), and Kanniainen and Stenbacka (), has shown that competition tends to undermine the incentives to avoid project-specific classification errors. In this respect the present paper emphasizes an additional mechanism. Uncoordinated screening by competing banks generates a poolworsening external effect whereby competition opens up a probability of entering a phase of inactivity, where no projects are funded. Furthermore, if the pool-worsening effect is not strong enough to induce inactivity, it will nevertheless increase the lending rate relative to that which a static banking oligopoly would charge. Our model makes it possible to characterize the nature of the screening-induced investment cycles. We find that these cycles are affected by the number of competing financiers, the growth rate of the economy in the sense of the size of the newly born generation of projectholders relative to the size of the incumbent vintage of entrepreneurs as well as on the uncertainty generated by imperfections in the screening technology. Our analysis proceeds as follows. Section II presents the basic framework. Section III analyses a coordinated funding industry operating in the absence of competition. Section IV presents the central result and demonstrates how competition among duopoly financiers gives rise to dynamic instability and cycles in screening and investments. In section V we provide novel evidence on high frequency cyclicality in the US venture capital industry. Section VI outlines generalizations and policy implications and Section VII concludes. 3 See also Shoar (, p. Cycles are no News to the Venture Capital Industry ) for more recent aggregate data of the whole venture capital industry. 5 II. A Model with Costly Screening Let us now present a stylised model of unsecured lending in a dynamic adverse selection framework. So consider a pool of risky projects. Each project requires an entrepreneurial idea and one unit of funding. The projectholders are equipped with an entrepreneurial idea but do not hold any capital of their own, nor do they have access to outside equity. Thus, we assume that the projects will have to be fully financed by outside financiers such as banks or venture capital firms. Financiers have access to a competitive deposit market. Their opportunity cost of capital equals the (safe) interest rate r. For subsequent use we let R = 1+ r. Both types of agents, the financiers as well as the entrepreneurs discount future payments at the same rate of δ = 1 R. In section III we will analyse the case of a single financier, while in section IV we will allow for two (or more) competing financiers. Entrepreneurs can be of two types. Either they have a potentially valuable idea and control a project of type G (good) or their idea is fundamentally flawed, in which case we denote the project type as B (bad). We assume that a project of type G has a success probability π as well as an associated return under success, R G, satisfying π 1. Type-G projects yield a zero return only under failure, but the success R G probability of this project type is sufficiently high so as to justify funding. Projects of type-b are assumed to always generate a zero return. We call entrepreneurs of type G creditworthy, whereas type-b entrepreneurs are not. Further, as we primarily focus on R&D-projects we assume these to exhibit a depreciation factor δ ( δ 1 ). This feature captures the return-significance of timing for high-tech products. It could also be given the interpretation that productmarket competitors might be able to imitate the idea incorporated in the project and that this happens at the probability rate δ. The funding industry operates with an infinite time horizon t=,1,.... Each period new potential projects enter the banking market. Denote the mass of entering projects in period t by η t and the proportion of profitable (good) projects by λ t 1. In principle, the size of the pool of new projects as well as its composition may vary over 6 time in the business cycle. Since our concern is to analyze how the conduct of financiers engaged in repeated competition may generate cycles, we will be largely concerned with a stationary pool of new projects so as to actually bias the model against cycling. 4 Thus, we assume η t = η and λ t = λ for t=1,,.... Financiers cannot directly distinguish type-g from type-b applicants. However, they have access to a screening technology. We assume that access to this screening technology is costly and imperfect. Hence type-i and type-ii errors will be made. With α we denote the conditional probability that a truly good project is mis-classified (type-i error), whereas β denotes the conditional probability that a truly bad project passes the credit test (type-ii error). The conditional probability that a project of the pool of new projects classified as creditworthy is truly of type G can be calculated as τ ( α, β ) λ (1 α) =. (1) λ (1 α) + (1 λ) β Therefore, λ τ ( α, β ) 1 and τ ( α, β ) 1 as we approach perfect screening (i.e. α, β ). Furthermore, it can immediately be verified that τ ( α, β ) is a decreasing function of both α and β. The project-specific monitoring expenditures are summarized by a parameter c . 5 If the financier makes use of the screening technology it is optimal to grant credit to those projects classified as G, while denying finance to those classified as B. It is worth stressing that the resource cost c may be quite significant. It is meant to measure, for example, the opportunity costs of experts evaluating projects. Moreover, in addition to the financial cost of screening there is a cost of delay because screening requires time. In line with evidence documented in the venture capital industry we assume that screening requires time. Specifically, we assume that screening requires one time 4 Clearly, while we predominantly concentrate on the endogenous generation of cycles, our analysis has implications for the amplification of exogenous shocks. 5 Hence, in the baseline model we do not allow banks to strategically finetune the quality of information. See Gehrig (1998) for a model with a strategic choice of information quality. 7 period. 6 Bengtsson et al. () demonstrate that successful applications in the venture capital industry typically require a screening period ranging from a couple of weeks up to a few months. In each period t financiers typically face a pool of project applications consisting of new entrants and, in addition, applicants that have been rejected by some rival financier in some earlier period. The statistical properties of this pool depend among other things on whether banks recall earlier applications, on the extent to which the funding industry adopts information sharing and on the classification errors prevalent in the screening technology. We assume perfect recall on the side of the financiers. Hence a rejected applicant will direct future funding applications to the rival financier and leave the pool of applicants when the set of financiers is exhausted. Moreover, we assume that financiers do not share information about earlier screening results. Accordingly, the pool of applications for a given financier consists of a random allocation of the new vintage of projects and a share of opportunistic applications of formerly rejected entrepreneurs. With a perfectly accurate screening technology type-b entrepreneurs would have no chance to pass a credit test and the pool of funded projects would exclusively consist of creditworthy projects. However, with imperfect credit tests B-type entrepreneurs enjoy an additional positive probability of funding due to banks classification errors. Thus, in the presence of screening imperfections each competing financier increases the chances of funding for bad entrepreneurs. On the other hand, the screened pool of project applications will be less adversely selected than under a regime of almost perfect screening, since with screening imperfections some good entrepreneurs may have been rejected earlier due to the α -errors. In each period t, financiers need to decide about their screening and funding activities. They can grant funds without screening in order to economize on the screening costs, they can provide screened funding only, or they can remain inactive altogether. After a decision has been made to finance a project the terms of funding have to be negotiated. We assume that banks submit take-it-or-leave-it offers. These offers, though, can be made conditional on the screening history of the borrower. Thus, the 6 In fact, this can be seen as our definition of a period within the context of the present model. However, we want to emphasize that our model focuses on qualitative aspects and the quantitative aspects of this interpretation should not be taken too literally. 8 financier will typically take into account the information transmitted if a borrower can produce evidence of a lending offer from a rival financier. 7 Optimal lending offers to entrepreneurs who pass the credit test are typically designed so as to keep the borrower indifferent between accepting and waiting to solicit another evaluation from a rival lender in some future period. 8 Since only successful projects of type G generate a positive cash-flow and since borrowers are protected by limited liability, the negotiations about repayments to the financiers may usefully be interpreted as negotiations about lending rates. Note, however, that we could easily reinterpret these n
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks