Secondary Markets, Financial Development, and Economic Growth Burak R. Uras∗ Tilburg University European Banking Center

Abstract Advocates of financial development claim that the availability of finance is vital for economic growth. Although there is ample empirical support for this argument, there is non-negligible evidence for counterproductive consequences of financial deepening as well. In this paper, I analyze the long-run interactions between secondary financial markets, financial development and economic growth and argue that the finance & growth nexus is sensitive to the structure of secondary markets in a society. In investigating the validity of this hypothesis, I study a dynamic general equilibrium model where entrepreneurial long-term investment is the engine of growth in an environment characterized by financial market imperfections. Financial frictions in the model are twofold: (1) Financial intermediation is limited because financing long-term investment requires costly enforcement of entrepreneur’s loan repayment; and, (2) Financiers are short-lived and therefore long-term capital is traded in a secondary market. The key finding of the paper shows that if the secondary market price of capital rises with financial development, reducing the cost of intermediation suppresses economic growth under fairly general conditions. Otherwise, financial development and growth are positively correlated. I present microfoundations for financial intermediation costs that induce the secondary market price of capital to grow with financial depth and draw policy conclusions.

Keywords: microfoundations of financial frictions, long-term investment, secondary markets, and economic growth. JEL Classification Numbers: E44, G2, O16, 047.

∗

Tilburg University, Department of Economics Room K322B 5000 LE Tilburg,The Netherlands. [email protected].

1

E-mail:

1

Introduction

The large literature that focuses on long-run implications of financial deepening provide theoretical foundations to explain why access to finance is important for macroeconomic development1 . However, a number of empirical studies are skeptical about the growth effects of financial development. For example, Pagano and Jappelli (1994), Demetriades and Luintel (2001), Castro et al. (2004) and Masten et al. (2008) show that, depending on country specific context, financial deepening might be counterproductive for the aggregate economy. The aim of this paper is to provide a theoretical foundation and uncover the potential non-monotone growth effects of financial development in a dynamic general equilibrium model with secondary capital markets. In the model, long-term investment is the engine of economic growth where financial frictions constrain the capacity to invest long-term. Specifically, (1) financial intermediation is constrained because there is costly enforcement of loan repayment, and (2) financiers are short-lived and therefore, long-term capital needs to be traded in a secondary market. An important general equilibrium result from this set-up is the endogenous existence of a non-monotonicity between financial frictions and steady-state output: Lowering financial intermediation costs is not always associated with a rise in economic development. This key result hinges on the behavior of equilibrium security prices in the secondary market. Reducing loan enforcement costs stimulates macroeconomic performance to the extent it mitigates the secondary market price of capital. If the secondary market price of capital rises with financial development, reducing the cost of enforcement suppresses aggregate long-term investment and might become counterproductive. I provide microfoundations that characterize the behavior of prices in the secondary market with respect to enforcement costs and draw policy conclusions concerning the growth effects of financial development. In investigating the effects of secondary market trading on finance & growth nexus, I study an overlapping generations model with occupation choice and production. In this model, two-period lived economic agents, in the first period of the life-cycle, endogenously choose an occupation as a worker or as an entrepreneur to manage a 1-period short-term production plant, or as an entrepreneur to adopt a 2-period long-term production technology. The occupation choice determines whether an agent becomes a financier (if the occupation selection is a worker or a short-term producer) or a borrower (if the occupation is the long-term producer). Long-term investment determines the long-run economic performance through endogenous productivity growth. Aggregate long-term output of today determines the productivity of short-term and long-term production plants of the future through intergenerational productivity spillovers. Since long-term investment takes two periods to complete and each financier can save at most for one period, long-term financial assets (financial returns from long-term investment) are traded in a secondary market. There is a primary and a secondary financier of each long-term asset. Primary financiers extend long-term loans (capital) for long-term entrepreneurial production and receive financial claims against next period’s output realization. Secondary financiers purchase these financial claims from primary financiers and collect capital returns from entrepreneurs upon the finalization of long-term production. 1

Greenwood and Jovanovic (1990), Bencivenga and Smith (1991), Marcet and Marimon (1992), Banerjee and Newman (1993), Acemoglu and Zillibotti (1997), Aghion and Bolton (1997), Azariadis and Kaas (2007), Antunes et al. (2008), and Aghion et al. (2010).

2

Financial intermediation frictions in the form of costly enforcement of loan repayment raise the cost of finance above the fundamental value of capital services. I study three financial regimes. Financial Structure 1: Primary financiers enforce loan repayment; Financial Structure 2: Secondary financiers enforce loan repayment; and Financial Structure 3: A combination of primary and secondary financiers enforce loan repayment. The delegation of enforcement to a particular financier group could be related to the securitization design or the regulatory framework in the society. Specifically, if each primary financier desires to sell a “bundle” of securities that promise returns from a large collection of long-term projects, then the institutional set-up might require the loan enforcement of primary financiers. On the other hand, if each primary financier sells a “single” security with promised returns from a single long-term project only, then it is more likely to have secondary financiers enforcing the loan repayment. The delegation of enforcement in each case is expected to lower the social costs of financial transactions within the respective environments: In the former financial regime where “bundled securities” are traded, secondary financiers as collectors of financial returns would find it hard evaluate the quality of the complex financial instruments that they buy unless primary financiers provide credible enforcement mechanisms that guarantee repayment in the future. In the second financial regime, complex financial instruments are absent at secondary market trade. Therefore, secondary financiers as capital return collectors arise as natural candidates for the enforcement of loan repayment. Regulatory framework might also assign exogenously set enforcement roles to a particular group of financiers in the economy. The key insights developed from each financial regime are as follows. Financial Structure 1: Primary financiers incur the cost of loan enforcement, and the equilibrium secondary market price of capital equals to the primary financier’s gross capital return and can be expressed as R(1 + φ): R is the steady-state rate of capital return to be determined endogenously, and φ is the capital price wedge implied by the costly loan enforcement. The rate of capital return for secondary financiers equals to R. With this model specification, a permanent reduction in capital price wedge φ raises the equilibrium rate of return R and suppresses R(1 + φ). The decrease in the cost of loan enforcement in turn increases the stock of financial assets since the rate of return from financing rises; and at the same time the amount of investable funds paid to each primary financier declines because the secondary market price of capital decreases with financial development. The former channel causes an expansion in the “financial sector” and fosters financial deepening in the economy and the latter channel stimulates the “allocative efficiency”. Both channels activate a rise in “real long-term investment” and aggregate productivity and promote financial deepening to be growth enhancing. This analytical result confirms with the standard argument about the positive growth effects of financial development presented by Bencivenga and Smith (1991), Aghion and Bolton (1997), Azariadis and Kaas (2007) and many others. Financial Structure 2: Secondary financiers enforce loan repayment, which implies that the secondary market price of capital equals R; whereas, the compensation of secondary financiers is R(1+φ). In this financial regime, lowering the cost of intermediation φ, increases the stock of financial assets as in Financial Structure 1, and raises the investable funds that needs to be paid to each primary financier since the real rate of financial return R rises with financial development. The net effect of financial development on aggregate long-term investment under fairly general model conditions is negative: Lowering φ reduces the real investment in long-term projects. Since long-term project investment is the engine of economic growth there exists an optimum level of financial repression that maximizes the steady-state per-capita output in the economy.

3

There is empirical support for growth reducing financial development experiences. For example, Pagano and Jappelli (1994) show that financial deregulation and the implied decline in cost of borrowing have contributed to the decline in economic growth rates during the 1980s in OECD countries and argue for potential optimality of financial constraints. Another important result from the model emerges when we compare the level of steady-state output under financial structure 1 against the level of steady-state output with financial structure 2. Holding everything else constant, shifting the delegation of enforcement from secondary financiers to primary financiers lowers the long-term project output, simply because the price of long-term claims rises as primary financiers become more involved at loan repayment enforcement. Since long-term investment is the engine of the productivity growth, the steady-state productivity of an economy governed by financial-structure-2 is higher than the steady-state productivity of an economy with financialstructure-1. This result suggests that if the introduction of complex financial instruments such as bundled asset-backed securities shift the enforcement of loan contracts from secondary financiers to primary financiers, macroeconomic performance might deteriorate. The insights from financial structures 1 and 2 reveals that microfoundations of financial intermediation costs are important to understand the implications of financial development for macroeconomic performance. To investigate the contribution of economic development in understanding the financedevelopment nexus, I also analyze an interim case (Financial Structure 3) with heterogenous primary financiers. In this extension, an exogenously set η fraction of all long-term projects’ loan repayment is enforced by primary financiers and the remaining fraction by secondary financiers. Comparative statics in this interim case indicate an interaction between economic development and the potential growth-reducing effects of financial development. Specifically, in high income countries as an initial condition a relatively larger share of primary financiers need to be enforcing loan repayment so that reducing intermediation costs can promote economic growth. This result implies that in high income countries financial development is more likely to reduce growth rates. Empirical findings from past research confirm with this result as well. For instance, Castro et al. (2004) measure costly loan repayment with “limited investor protection” and document non-linear growth effects of investor protection. An important empirical result from their work shows that investor protection and economic growth exhibit a negative relationship for the sample of high income countries. The theoretical results from this paper are important for development policy as well. For developing countries, there are cases where well-intended financial development policies reduced economic growth rates, such as in Latin American countries during 1990s, whereas similar financial development experiences promoted economic growth in other countries, e.g. Turkey during 1980s. My theoretical results in this paper draw attention to the characteristics of the financial structure that policy makers might need to pay attention in order to avoid policies with potential counterproductive consequences. Related Literature: The paper contributes to the literature on Finance and Macroeconomic Development. Some of the important studies in this literature are Greenwood and Jovanovic (1990), Bencivenga and Smith (1991), Banerjee and Newman (1993), Antunes et al. (2008), and Aghion et al. (2010).2 This paper is essentially related to the literature that identifies non-monotone real effects of financial sector development. Some important studies in this area of research are Bencivenga et al. (1995 and 1996), Deidda and Fatouh (2002), Castro et al. (2004), and Uras (2012). In these studies, the 2

There is also a vast empirical literature focusing on the association between the level of financial development and economic growth. See for example Levine (1993 and 1997), Beck, Levine, and Loayza (2000), Christopoulos and Tsionas (2004), Castro et al. (2004) and Beck et al. (2008).

4

non-monotonicity between finance and economic development arises if quantities of some particular macroeconomic variable, such as the steady-state capital stock (Bencivenga et al. (1995 and 1996)) or the population weight of firms with an access to a long-term production technology (Uras (2012)), do not reach a critical level. Different from the approach of these studies, in the current paper, the nonmonotonicity between finance and economic development depends not only on the stage of economic development but also on the structure of the financial sector and the underlying microfoundations that explain the cost of financial intermediation. To the best of my knowledge the current study is the first in pointing out this specific issue about the potential counterproductivity of financial development. The rest of the paper is organized as follows. Section 2 introduces the basic model environment. Sections 3, 4 and 5 compare the growth implications of financial development across three financial regimes and draw policy conclusions. Section 6 concludes the paper.

2

The Model Environment

I study an Overlapping Generations model (OLG) with production and occupation choice. Time is indexed with t and continues forever. There is a single good in the economy, which is produced by entrepreneurial projects. Each period a continuum of risk-neutral agents with measure M enter the economy. Agents live for two periods, denoted as young and old. All young agents are ex-ante identical who at the beginning of their life-cycle endogenously specialize in one of the following three occupations: Workers, Entrepreneurs with a Short-term Investment Project, and Entrepreneurs with a Long-term Investment Project. The endogenous population measures of workers, short-term producers and long-term producers in any given period t are denoted by Mω,t , Me,t , and ML,t . Workers supply labor for short-term entrepreneurial projects in the first-period of the life-cycle and receive real wages wt . The real wage will be determined endogenously. Workers have no labor endowment in the second period of their life-cycle, so that they have to save to finance the old-period consumption. Entrepreneurs with a short-term investment project employ labor to produce the consumption good using a technology that takes one period to complete: Short-term production, for each lt quantity of labor employed, returns s(lt ) units of the consumption good at the end of the first period of an entrepreneur’s life-cycle. Assumption 1. st (lt ) = AS,t ltβ , with 0 < β < 1. Assumption 1 simply implies a span of control (decreasing returns to scale) with respect to the labor employed in short-term projects. As,t is a productivity parameter. The process that governs the time-trend of As,t will be delineated below. Similar to workers, entrepreneurs of short-term investment projects manage production plants when young and retire when old. Entrepreneurs with a long-term investment project operate production plants that require a long-term capital investment. Long-term production takes two-periods to complete, and for each kt units of capital employed in period t, it returns `t+1 (kt ) units of the consumption good at the end of 5

the second period. As in the short-term production, `t+1 (kt ) exhibits decreasing returns to scale. Assumption 2. `t+1 (kt ) = AL,t ktα , with 0 < α < 1. The capital good depreciates completely during production process of a long-term project, and unlike workers and short-term producers, long-term entrepreneurs manage projects during both periods of their life-cycle. Productivity. The productivity of short-term and long-term projects in any given period t is a function of aggregate output produced by long-term entrepreneurial Rprojects in period t − 1. I denote the aggregate output produced by long-term projects by Wt−1 ≡ ML,t−1 `i,t−1 di , and formalize the productivity process in the economy with the following assumption: Assumption 3. For J ∈ {S, L}, AJ,t = gJ (Wt−1 ) with

∂gJ (W ) ∂W

> 0.

This assumption states that there are inter-generational knowledge spillovers in the form of productivity enhancing investment: Long-term output of today determines the aggregate productivity of the future. With this time-trend of productivity, one can think of the long-term entrepreneurial projects as an investment opportunity that incorporates R&D or human capital development. Productivity enhancing long-term investment of this form is a stylized feature of endogenous growth models with overlapping generations as in Aghion et al. (2010). Financial Market Transactions. Long-term entrepreneurial capital can be financed only if longterm financial claims can be traded (rolled over) in a secondary capital market between the financiers of two consecutive generations. In any given period t the sequence of events, that also characterize financial market transactions and the need for a secondary capital market, is as the following: 1. Young agents select an occupation from the set {Worker, Short-term entrepreneur, Long-term entrepreneur }. 2. Short-term project entrepreneurs hire labor from workers (young), and Long-term project entrepreneurs borrow capital from financiers (old). 3. The production output from short-term and long-term projects are collected. 4. Capital and labor gets paid their returns: Old capital providers (financiers) and young workers are paid. 5. Secondary capital market opens. Short-term producers and workers (young financiers) purchase financial claims against period t + 1 long-term investment returns. Specifically, old-financier, who have extended long-term capital to young entrepreneurs at the beginning of the period t, sell securities to young financiers. 6. Agents consume. The timing of events imply that there is a primary and a secondary financier of a long-term investment project: Primary financiers extend long-term loans (capital) for entrepreneurial projects in period t and receive financial claims against next period’s (t + 1) output realization. Secondary financiers purchase

6

these financial claims from primary financiers in t and collect capital returns from entrepreneurs upon the finalization of long-term projects in period t + 1. Financial Frictions. I study the effects of financial intermediation costs on steady-state output and consumption. Specifically, the two-period capital cost for a long-term entrepreneur equals RL = Rq(1 + φ),

(1)

where R and q are endogenously determined (net) private returns to capital that accrue to primary (R) and secondary (q) financiers, and φ ≥ 0 is the cost of financial intermediation, as emphasized in Demirguc and Huizinga (1999) and Antunes et al. (2007). The novel feature of the current model is that the financial intermediation costs are micro-founded. Financiers incur a utility loss (cost of enforcement) that is proportional to the size of the private capital returns from financing long-term projects which in turn increases the cost of finance for entrepreneurs (RL ) by a capital price wedge as in (1). The central remark that I aim to deliver is the microfoundations of financial repression are important to understand the nexus between financial development and economic growth. To analyze the microfoundations of financial frictions, I investigate financial development experiments in three different regimes. Financial Structure 1: Primary financiers enforce loan repayment and suffer a utility loss at extending capital to long-term entrepreneurs that increases RL for long-term entrepreneurs. Financial Structure 2: Secondary financiers enforce loan repayment, and therefore, suffer a welfare loss when purchasing financial claims from primary financiers which again increases RL for long-term entrepreneurs. Financial Structure 3: A combination of primary and secondary financiers enforce loan repayment. The delegation of enforcement to a particular financier group could be motivated by the securitization design in the society. The trade of complex financial instruments, e.g. bundled long-term project returns, might require the involvement of primary financiers at loan enforcement, whereas enforcement of secondary financiers is expected to minimize financial transaction costs if simple long-term financial assets are traded in the secondary market. As I will delineate in the consecutive analysis, the nexus between financial development and long-run economic performance is highly sensitive to the mode of financial structure a country exhibits: While with Financial Structure 1 reducing financial intermediation costs is always growth promoting, in Financial Structure 2 financial development is likely to be counterproductive. In financial structure 3 the growth effects of financial development is a function of the total number of primary financiers who are allocated at enforcing the loan repayment. Preferences. The life-time utility from consumption is specified as U = cy + β i co ,

(2)

where cy is the youth-period consumption and co is the old period consumption of an agent. The discount parameter βi is specific to an agent’s occupation and his role in the financial structure. 1 , 1} with φ ≥ 0; and , Specifically, βi ∈ { 1+φ 1. βL = 1 for all long-term entrepreneurs (subscript L stands for long-term) in all financial regimes. 2. In financial structure 1, β1 = and 2 for secondary.

1 1+φ

and β2 = 1 where subscript-1 stands for primary financiers

7

3. In financial structure 2, β1 = 1 and β2 =

1 . 1+φ

4. In financial structure 3, an η fraction of all long-term projects require the enforcement of primary financiers whereas 1 − η fraction are enforced by secondary financiers. Therefore, for η fraction of 1 all primary financiers in a cohort β1e = 1+φ whereas for the remaining subset of primary financiers ne β1 = 1. With this utility specification, I incorporate an effort cost (equivalent to loss of utility from consumption) associated with investing for the second period of a financier’s life-cycle. Such enforcement costs can be rationalized with costly “monitoring” of entrepreneurial projects. In equilibrium the discount parameter φ will generate a capital price wedge as in (1). The risk-neutrality implied by the linear utility specification is not important for the qualitative messages that I highlight in this paper. Incorporating risk-aversion, for instance in the form of logarithmic preferences, would not alter the theoretical contribution.

3

Financial Structure 1: Costly Primary Finance

Financial intermediation costs are incurred by primary financiers. Therefore, primary financiers discount the second period consumption by 1/1 + φ, whereas the secondary financiers value the second period consumption as much as the first period consumption. I allow any convex combination of the two investment options (as a primary financier or as a secondary financier) for a particular saver. That means, a saver can be a primary financier for some long-term projects and at the same time invest as a secondary financier to some others. Suppose pt is the price of a one-unit long-term financial claim in the perfectly competitive secondary market that returns RtL in period t + 1. The private capital return of a primary financier net of cost pt . Suppose qt+1 denotes the private capital of loan enforcement is denoted by Rt and equals Rt = 1+φ returns for a secondary financier from a one-unit financial claim purchased in the secondary market. Lemma 1 summarizes the conditions that must hold between R, p and q to guarantee the existence of an equilibrium with long-term investment.

Lemma 1 (i) An equilibrium with long-term capital investment exists if and only if Rt+1 = qt+1 ≥ 1. (ii) The unit cost of two-period finance for a borrower equals to Rt Rt+1 (1 + φ). Proof Denote the two-period cost of finance for a borrower by RtL . Given the secondary market price RL of a long-term claim (pt ), qt+1 = ptt . The no-arbitrage condition RtL pt+1 = , pt 1+φ |{z} | {z }

=qt+1

=Rt+1

should hold between primary and secondary financier’s capital returns; otherwise, all financiers strictly prefer to be either a primary financier or a secondary financier, and under either case long-term finance 8

cannot become available for entrepreneurs. Therefore, RtL = Rt Rt+1 (1+φ) holds at interior equilibrium where long-term investment exists. Finally Rt+1 = qt+1 ≥ 1 should hold. If not, linear life-time utility implies that all consumption will take place in the first period of the life-cycle which implies zero long-term investment in the aggregate.  An important feature of this economy is that in equilibrium the unit capital return for a financier (primary or secondary) equals to R. Therefore, variations in financial intermediation costs affect the life-time value from being a saver only through general equilibrium adjustments in R. In the consecutive analysis I analyze the behavior of R with respect to φ and study the growth implications of financial development where financial development will be defined as a permanent reduction in cost of enforcement, φ.

3.1

Optimizing Behavior

Workers. Risk neutrality implies that workers save the entire wage compensation wt from the youth period by financing long-term projects as long as the returns from project finance are large enough (Rt+1 ≥ 1). For Rt+1 ≥ 1 a worker’s lifetime utility equals to Vω,t = Rt+1 wt . Short-term producers. Short-term entrepreneurs maximize: max π(lt ) = s(lt ) − wlt . lt

Denote the optimum quantity of labor employed as a function of the equilibrium wage rate by lt∗ (wt ). The implied entrepreneurial profit in equilibrium from short-term investment is π(lt∗ (wt )). Due to risk neutral preferences short-term producers also save the entirety of their profits by lending π(lt∗ (wt )) units of the consumption good to finance long-term projects if Rt+1 ≥ 1. Provided Rt+1 ≥ 1 a short-term entrepreneur’s lifetime utility is given by Ve,t = Rt+1 π(lt∗ (wt )). Long-term producers. An entrepreneur with a long-term project born in period t maximizes: AL,t ktα − Rt Rt+1 (1 + φ)ktα . A long-term entrepreneur’s optimal capital demand kt∗ and the associated life-time value function are given by 1  1−α αA L,t , kt∗ = Rt Rt+1 (1 + φ) α   1−α αAL,t VL (kt ) = (1 − α)AL,t . Rt Rt+1 (1 + φ)



The financial friction (φ) has a direct impact (as well as an indirect price effect) on a long-term project entrepreneur’s lifetime value. For constant R the higher φ the lower is the lifetime value from being an entrepreneur.

9

3.2

Equilibrium

Definition The dynamic general equilibrium of the economy is characterized by an infinite stream t=∞ ), one-period rate of capital returns ({Rt }t=∞ of wage rates ({wt }t=0 t=0 ) and security prices in the sect=∞ ondary capital market ({pt }t=0 ) at which 1. Agents optimize their life-time utility, 2. Life-time returns from the three occupation choices are equalized, such that V (w ) = V (l ) = VL,t (kt ), | ω,t t {z e,t t}

(3)

≡VS,t

1 3. The life-time value from being a primary financier (VS,t ) is equal to the life-time value of a 2 secondary financier (VS,t ) 1 2 VS,t = VS,t (4)

4. Labor and capital markets clear Z Z

s li,t (wt )di

Z

d lj,t (wt )dj Me,t Mω,t Z Z d ki,t (Rt , Rt+1 )di = wj,t dj +

=

ML,t

Mω,t

(5) πj,t (wt )dj

(6)

Me,t

where Mω,t , Me,t , and ML,t are equilibrium measures of workers, short-term entrepreneurs and long-term entrepreneurs respectively.

The model is recursive. This means, labor market clearance determines the equilibrium wage rate wt , and wt in turn pins down Rt+1 that clears the capital market in the next period.

Lemma 2 A unique stationary equilibrium exists.

Proof Denote the steady-state wage rate and the one-period rate of capital by w and R respectively. In a steady-state equilibrium, endogenously determined fractions of the population select to become workers, short-term producers and long-term producers. The conditions that need to hold for the existence of an equilibrium where positive measures of each type of agent exist in the economy is given by equation (1). The first equality at (1) can be re-written as: Rw = Rπ(w) |{z} | {z } =Vω

(7)

=Ve

Equation (7) implies that in a steady-state equilibrium w = π(w): Workers and short-term producers share the returns from short-term technology output with equal shares, equalizing the life-time values across two occupations such that Vω = Ve ≡ VS . It proves to be useful to define MS = Mω + Me where MS is the aggregate measure of savers in a cohort. 10

Using the production function specified at assumption 2, the short-term producer’s profit function
β S 1−β , where AS is the steady-state level of short-term project is given by π(w) = (1 − β)AS βA w productivity. Since in equilibrium π(w) = w holds: 1

w = AS β(1 − β) 1−β . Given w, the labor market clearance sets the relative population weights between workers (Mω ) and short-term producers (Me ), and the second part of equality (1) pins down the general equilibrium rate of private capital (financial) return R. The rate R in turn determines the invariant distribution between savers and borrowers (long-term entrepreneurs) in the economy. The saver-borrower distribution in the economy settles the steady-state short-term (AS ) and long-term (AL ) productivity of entrepreneurial projects.  In a stationary equilibrium the cost of finance for a long-term entrepreneur is stated as: RL (φ) = R2 (1 + φ). The price of a long-term claim in the secondary market satisfies: p(φ) = R(1 + φ). We can express the occupation choice indifference between being a saver (VS ) and a long-term entrepreneur (VL ) as the following:  wR = (1 − α)AL |{z} VS |

αAL 2 R (1 + φ) {z

α  1−α

.

(8)

}

VL

Solving for R as a function of φ, one can derive: 1 1+α

R(φ) = AL



1−α w

  1−α 1+α

α 1+φ

α  1+α

.

(9)

R(φ) is a decreasing function of φ. This equilibrium property implies that the life-time value from being a financier (saver) increases with financial development. Financial development makes the “finance-sector” attractive. The rise in the rate of return on capital as financial frictions contract is a standard feature of general equilibrium models of financial development, such as Aiyagari (1994) and Angeletos and Calvet (2006). Given R(φ), the two-period cost of finance RL can be expressed as a function of φ: 1−α

RL (φ) = R2 (1 + φ) ∝ (1 + φ) 1+α

(10)

As it can be observed at expression (10) the cost of long-term finance declines with financial development, and stimulates the value from being a long-term entrepreneur. Since agents are ex-ante identical the life-time value from long-term entrepreneurship increases by the same rate at which financier’s lifetime value rises following a reduction in φ.

11

The secondary market price of a long-term claim, p, can be solved as a function of φ as well: 1

p(φ) = R(1 + φ) ∝ (1 + φ) 1+α .

(11)

The price of long-term claims in the secondary market increases with φ. That p is an increasing function of φ is important to understand the long-run effects of financial development. In steady-state a constant amount of per-financier capital is invested in long-term entrepreneurial projects. Denote this steady-state amount by x. The aggregate capital invested in long-term entrepreneurial projects then equals to Ms x. The aggregate compensation of primary savers in the stationary equilibrium should then be equal to MS xp(φ) = MS xR(1 + φ). Since the aggregate capital in the economy equals to MS w, we can derive x as w w = . (12) x= 1+p 1 + R(1 + φ) Since the price of long-term claims increases with φ, x rises with financial development. To understand whether the aggregate capital invested in long-term projects rises with financial development we need to derive the equilibrium measure of savers in the economy (MS ) as a function of φ. The capital market clearance condition can be stated as: 1  1−α  w αAL MS = ML 1 + R(1 + φ) R2 (1 + φ) 1  1−α  αAL VL = ML ⇒ MS R(1 + R(1 + φ)) R2 (1 + φ) α 1   1−α     1−α αAL αAL 1 ⇒ MS (1 − α)AL = ML R2 (1 + φ) R(1 + R(1 + φ)) R2 (1 + φ) Solving for the equilibrium ratio of savers relative to long-term entrepreneurs (MS /ML ) yields: MS α 1 + R(1 + φ) = ML 1 − α R(1 + φ)

(13)

The equation (13) shows that the MS /ML ratio is a decreasing function of φ. Financial development stimulates aggregate capital available for long-term projects (MS x) by 2 channels: (1) The total number of savers (MS ) in the economy rises, and (2) the per-financier amount of savings allocated to long-term projects (x) increases. Steady-state Output and Consumption. I measure the long-run performance of the economy by steady-state consumption (and output) in per-capita. Denoting c as the consumption per-capita in the steady-state (and per-capita output by y), from equation (8), we can derive: y = c = wR, = (AL w)

1 1+α

(1 − α)

1−α 1+α



α 1+φ

α  1+α

.

(14)

Steady-state consumption is a function of AL and w. Productivity of long-term projects and, AL , and the real wage rate, w, are both determined by the aggregate output produced by long-term entrepreneurial projects (W ) which can be derived as:  α w MS W = ML AL , 1 + R(1 + φ) ML  α w 1 + R(1 + φ) α = ML AL , 1 + R(1 + φ) R(1 + φ) 1 − α  α w α = ML AL . R(1 + φ) 1 − α 12

Using z ≡

1−α α

and MS = M − ML at (13): zR(1 + φ) ML = , M − ML 1 + R(1 + φ) zR(1 + φ) ⇒ ML = M. 1 + (1 + z)R(1 + φ)

Now W as a function of φ and R can be stated as, W =

[R(1 + φ)]1−α AL z 1−α [w]α M, 1 + (1 + z)R(1 + φ)

(15)

or as a function of the secondary market price of capital p: W =

3.3

p1−α AL z 1−α [w]α M. 1 + (1 + z)p

(16)

Financial Development and Economic Performance

The output produced by long-term projects determines the steady-state per-capita consumption as the steady-state productivity depends on the aggregate output produced by long-term entrepreneurial projects. Although financial development is expansionary for the aggregate capital invested in longterm projects, it also reduces the total number of long-term entrepreneurs in the economy. Therefore, one needs to check whether the aggregate output produced by long-term projects does in fact rise with financial development in this “benchmark” specification. I define financial development as an event that occurs at the beginning of a period τ (at stage-0 of the flow of events presented in section 2) and reduces the cost of financial intermediation φ permanently. Comparative statics. Differentiating (16) with respect to φ, and denoting the first partial derivative of p by p0 (φ), we can show that lowering the intermediation friction φ improves the steady-state aggregate output produced by long-term projects if and only if: (1 − α) [p(φ)]−α p0 (φ)[1 + (1 + z)p(φ)] < (1 + z) [p(φ)]1−α p0 (φ)

(17)

Using the definition of z in (17), the condition (17) reduces to (1 − α)α(1 + (1 + z)p(φ)) < p(φ), ⇒ (1 − α) < p(φ), ⇒ (1 − α) < R(1 + φ)

(18)

The rate of capital return, R, must exceed 1 to induce risk-neutral agents to invest their first period income in long-term projects. Therefore, as long as an equilibrium with long-term investment exists, characterized by R ≥ 1, the condition at (18) holds for all parameter values: Reducing the intermediation costs always stimulates the steady-state output from long-term projects. Proposition 3.1 The inequality at (18) is satisfied for all parameter values that support an interior equilibrium. Lowering financial intermediation frictions stimulates the aggregate output generated by long-term projects. 13

Transitory Dynamics. A permanent reduction in φ in a period τ stimulates the aggregate output of long-term projects in the consecutive period (Wτ +1 ). Since the productivity of short-term and long-term projects are time-dependent as characterized at assumption 3, a rise in current output from long-term projects translates into a higher project productivity in the future. Therefore, comparing the steady-state values of productivity and output for long-term projects (AL , W ) against the productivity and output in period τ + 1 (AL,τ +1 , Wτ +1 ) shows that AL > AL,τ +1 and W > Wτ +1 . These results imply that future generations benefit from financial development more than the current generation who invests in institutions that promote financial intermediation. Going back to equation (14): The current generation (cohort born in period τ ) clearly benefits from financial development since for constant w and AL , consumption in period in τ + 1 rises as φ declines. However, the steadystate per-capita consumption, c, is strictly larger than the per-capita consumption in period τ + 1, cτ +1 , because the rise in W stimulates AL and w in the long-run.

FIGURES 1 AND 2 ABOUT HERE

3.4

Discussion

The current economic framework resembles a two-sector model characterized by the interactions between the “financial-sector” that generates the capital input necessary for the “real investment-sector”. When intermediation frictions are large the economy exhibits an inefficiently undersized finance-sector because the rate of capital return is too small. Financial development stimulates the rates of return on capital causing an expansion in the size of the finance-sector, and an increase in capital available for the real production sector. Figures 1 and 2 summarize the results from section 3. Figure 2 draws the net capital returns (R) and aggregate output produced by long-term projects (W ) in steady-state, and figure 3 draws the secondary market price of long-term claims (p) together with W . p is decreasing in φ, and so is W in p; therefore, a decline in cost of enforcement stimulates steady-state output produced by long-term projects. The key result that shows the steady-state consumption is a decreasing function of φ has an intuitive interpretation. Permanent reductions in φ generate three effects on the macroeconomy: (1) The contraction in capital price wedge φ increases the aggregate capital stock in steady-state: The population share of financiers (MS ) rises as φ decreases. (2) The contraction in φ also mitigates the fraction of investable funds that needs to be paid to primary financiers in secondary market transactions: p is an increasing function of φ, and therefore the smaller φ the larger is the fraction of the physical capital stock that can be allocated to long-term entrepreneurial production. (3) The output produced by long-term projects rises with the expansion of the capital stock which stimulates the productivity of short-term and long-term projects in the economy. The channel 1 fosters the capital deepening whereas the latter two channels stimulate the total factor productivity by improving the allocative efficiency of capital (channel 2) and as a result stimulating inter-generational productivity spillovers (channel 3). Institutional development (lowering cost of financial intermediation φ permanently) stimulates the steady-state aggregate product and consumption through all three channels. 14

4

Financial Structure 2: Costly Secondary Finance

Consider now a financial structure where secondary financiers incur the effort cost of financing longterm projects. The preferences specified in section 2 imply that the secondary financiers discount the capital returns that will be collected from long-term financial claims purchased in the secondary market. Denoting pt again as the secondary market price of long-term financial claims, the private capital RtL . As in section 3, Rt return of a secondary financier net of effort costs is computed as qt+1 = pt (1+φ) denotes the private capital returns for a primary financier where with Financial Structure 2 Rt = pt . The following lemma shows that RtL = Rt Rt+1 (1 + φ) continues to hold in equilibrium. Lemma 3 The unit cost of two-period finance satisfies RtL = Rt Rt+1 (1 + φ). Proof The no-arbitrage condition RtL = pt+1 , |{z} p (1 + φ) | t {z } =Rt+1 =qt+1

should hold between primary and secondary financier’s capital returns as in Financial Structure 1; otherwise, all financiers strictly prefer to be either a primary financier or a secondary financier, and with either case long-term finance cannot become available for entrepreneurs. Therefore, RtL = Rt Rt+1 (1 + φ). 

4.1

Optimizing Behavior and Equilibrium

The optimizing behavior for three occupations and the existence result for the stationary equilibrium remain the same as in section 3. The occupation indifference condition between being a saver (worker or short-term producer) and a long-term entrepreneur solves the real rate of capital return, yielding the same steady-state expressions for R(φ) 1 1+α



R(φ) = AL

1−α w

 1−α  1+α

α 1+φ

α  1+α

,

and for the cost of finance RL (φ) 1−α

RL (φ) ∝ (1 + φ) 1+α , as in section 3. Different from the financial structure analyzed in section 3, in this case the secondary financiers incur the cost of long-term project finance; therefore p = R in equilibrium with 1 1+α

p(φ) = AL



1−α w

 1−α  1+α

α 1+φ

α  1+α

,

which implies that the secondary market price of capital increases with the level of financial development. Denoting the per-saver capital invested in entrepreneurial projects by x, the aggregate capital 15

stock invested in long-term entrepreneurial projects equals to Ms x which implies that the aggregate compensation of primary savers must equal to MS xp = MS xR. We can derive w w = , (19) x= 1+p 1+R and observe that x decreases as cost enforcement φ decreases. Each primary financier provides a relatively smaller fraction of his savings for long-term projects as the financial sector develops. This is a result highlighted also by Bencivenga, Smith and Starr (1996): If long-term assets need to be rolled over in a secondary market, reducing the cost of financial transactions lowers long-term investment. Different from their work, this paper shows that the delegation of loan enforcement is central to understand the effects of financial intermediation costs on long-term investment. We can express the capital market clearance condition as: w = ML MS 1+R



αA 2 R (1 + φ)

1  1−α

,

1  1−α  αA VL = ML , ⇒ MS R(1 + R) R2 (1 + φ) α  1   1−α   1−α  αA 1 αA ⇒ MS (1 − α)A . = ML R2 (1 + φ) R(1 + R) R2 (1 + φ)

Solving for MS /ML ratio yields: MS α 1+R = . (20) ML 1 − α R(1 + φ) The equation (20) shows that the MS /ML ratio (similar to the Financial Structure 1 from section 3) increases as φ declines permanently. Steady-state Output and Consumption. The steady-state expression for per-capita consumption remains as before y = c = wR = (AL w)

1 1+α

(1 − α)

1−α 1+α



α 1+φ

α  1+α

(21)

The aggregate output produced by long-term projects, which in the long-run determines AL and w can be derived as the following:  α w MS W = ML AL 1 + R ML  α w 1+R α = ML AL 1 + R R(1 + φ) 1 − α  α w α = ML AL . R(1 + φ) 1 − α Using z ≡

1−α α

and MS = M − ML at (20): ML zR(1 + φ) = , M − ML 1+R zR(1 + φ) ⇒ ML = M. 1 + R + zR(1 + φ) 16

Then W as a function of R reduces to: W =

[R(1 + φ)]1−α AL z 1−α [w]α M, 1 + R + zR(1 + φ)

(22)

W =

[p(1 + φ)]1−α AL z 1−α [w]α M. 1 + R + zp(1 + φ)

(23)

or in terms of R and p:

An important implication of this model can be observed by comparing the steady-state aggregate output from long-term projects that we derived at (23) (with p = R), against the aggregate longterm project output at (16) (with p = R(1 + φ)): Holding everything else constant, shifting the delegation of enforcement from secondary financiers to primary financiers, or in other words changing the financial structure of the economy such that p1 (= R) goes up to p2 (= R(1 + φ)) lowers the long-term project output. Financial structure 2 exhibits a long-term intensive composition of output compared to financial structure 1. Since long-term investment is the engine of the productivity growth in this economy, the steady-state productivity of an economy governed by financial structure 2 (p = R) is higher than the steady-state productivity of an economy with financial structure 1 (p = R(1 + φ)). I summarize this important result with the following proposition.

Proposition 4.1 Ceteris paribus, a financial regime that delegates the enforcement of loan repayment to primary financiers lower aggregate productivity compared to a regime where secondary financiers enforce loan repayment.

Referring back to the previous discussion on securitization and the delegation of enforcement; proposition 4.1 suggests that if the introduction of complex financial instruments such as bundled asset-backed securities over the last two decades shifted the enforcement of loan contracts from secondary financiers to primary financiers, this financial innovation might have slowed down the economic growth rates in countries with large volumes of secondary market transactions.

4.2

Financial Development and Economic Performance

I consider again a financial development exercise at the beginning of a period τ that reduces the cost of intermediation φ permanently. As in section 3, I will first study the effects of financial development on steady-state output produced by long-term projects and then draw conclusions for the per-capita consumption in steady-state equilibrium. Long-term Projects. In order to analyze the effects of a permanent reduction in φ on steady-state < 0 if and only if: output, I differentiate the expression at (22) with respect to φ, and show that ∂W ∂φ

⇒

[(1 − α)R−α (1 + φ)1−α R0 + R1−α (1 + φ)−α ][1 + R + zR(1 + φ)] < [1 + z(1 + φ)R0 + zR][R1−α (1 + φ)1−α ] (1 − α)R−α (1 + φ)1−α R0 + (1 − α)R1−α (1 + φ)1−α R0 + z(1 − α)R1−α (1 + φ)2−α R0 +R1−α (1 + φ)−α + R2−α (1 + φ)−α + zR2−α (1 + φ)1−α < R1−α (1 + φ)1−α R0 + z(1 + φ)2−α R1−α R0 + zR2−α (1 + φ)1−α . 17

In this inequality, R0 represents the first partial derivative of R with respect to φ. The inequality simplifies to: (1 − α)R−α (1 + φ)1−α R0 + R1−α (1 + φ)−α (1 + R) < αR1−α (1 + φ)1−α R0 + (1 − α)R1−α (1 + φ)2−α R0 | {z } | {z } | {z } | {z } ≡X

>0

|X| (as defined at nequality (24)), it shows that the inequality (24) holds only if R < R∗ where R∗ is a critical level of R with R∗ < 1. Therefore, the inequality (24) does not hold for any parameter values that satisfy the interior equilibrium condition (R > 1). 0

Proposition 4.2 A permanent reduction in cost of intermediation φ lowers the steady-state output produced by long-term projects for all parameter values as long as R > 1.

There are two counteracting forces that generate the result at proposition 4.1. (1) The permanent reduction in φ increases the aggregate capital stock in steady-state: The population share of financiers and hence the aggregate size of the financier-sector rises (MS ) as φ decreases. (2) However, the contraction in φ raises the fraction of investable funds that needs to be paid to primary financiers in secondary market transactions: R, and p are increasing in φ, and therefore the smaller φ is, the lower is the fraction of the physical capital stock that is allocated at long-term entrepreneurial production. The former channel expands the capital deepening in the economy and stimulates the steady-state output produced by long-term projects whereas the latter channel lowers the total factor productivity by distorting the allocative efficiency of capital and retards the steady-state output from long-term projects. The reduced form effect of financial development depends on the relative dominance between the two channels. In the current framework the TFP channel dominates the capital deepening channel and thus reducing financial frictions in the form of capital price wedges lowers the output produced by long-term projects. Financial development has two effects on steady-state per-capita consumption: (1) The direct (positive) impact that expands the “financial-sector” and (2) the indirect (negative) impact that contracts the “real investment” from long-term projects, and suppresses AL and w in steady-state. The net effect of financial development on steady-state consumption depends on the relative weights between the two channels. The level of intermediation frictions that optimize the steady-state consumption can be found by maximizing (21) with respect to φ subject to (22) and the project productivity processes gJ (W ) described at assumption 3 for both J ∈ {S, L}. Specifically, defining gJ0 as the first partial derivative of gJ (.), the level of financial repression that maximizes the steady-state consumption (φ∗ ) solves:    0 ∂W 1 gS gL + gS gL0 =α . (25) gS gL ∂φ 1 + φ∗ Transitory Dynamics. Following a permanent reduction of φ in period τ , AL and w do not adjust until period τ + 2. Therefore, per-capita consumption in period τ + 1 rises. Since AL,t and wt steadily adjust downwards with perpetual contractions in Wt , the per-capita consumption in period τ + 1 exceeds the steady-state per-capita consumption. This is a result contrasting with the transitory dynamics derived in section 3. Furthermore, depending on the functional form assigned for g(W ) and 18

the parameter values of aJ and µJ for J ∈ {S, L}, per-capita consumption in the new steady-state following a permanent reduction in the cost of enforcement φ could be lower than the per-capita consumption in the old steady-state. Figure 3 summarizes the equilibrium implications of financial development on steady-state output from long-term projects in an economy characterized by a financial regime with costly secondary finance.

FIGURE 3 ABOUT HERE

4.3

Discussion

Comparing the results from sections 3 and 4 shows that understanding the microfoundations of intermediation frictions are important for uncovering the effects of financial development on steady-state macroeconomic performance. This analytical conclusion has important empirical implications and also relevance for financial market policies. Cross-country empirical evidence suggests that financial development and economic growth are not always positively associated. For example, Bandiera et al. (2000) suggest that the effects of the domestic financial liberalization on economic growth are mixed. The authors argue that for 1970-1994 time period, the relationship between banking deregulation and economic performance is negative and significant in Korea and Mexico3 , whereas it is positive and significant in Turkey and Ghana. Pagano and Jappelli (1994) show that borrowing limits and economic growth displayed a negative correlation during 1980s in OECD countries, and Bayoumi (1993) found similar results for the United Kingdom for the same time interval. For developing countries, there are cases where well-intended financial development policies reduced economic growth. For example, the Latin American countries experienced secular banking deregulation experiences during 1990s, which according to Fostel and Geanakoplos (2008), generated an undersupply of market liquidity and undermined economic performance. The theoretical results presented in sections 3 and 4 offer an understanding for the conditions that policy makers should pay attention in order to avoid financial market policies with potential counterproductive consequences, especially when long-term investment and liquid secondary markets are essential for real economic performance.

5

Financial Structure 3: Heterogeneous Primary Financiers

When all primary financiers are compensated with the capital price wedge as in financial structure 1, that is consistent with a secondary market price of long-term capital, financial development stimulates economic growth. When all secondary financiers are paid the capital price wedge, financial development could reduce the economic development. Since these two extreme regimes lead to a bifurcation concerning the effects of finance on economic development, it would be interesting to study an interim case. 3

Similar conclusions are reached for Korea by Demetriades and Luintel (2001) and Castro et al. (2004).

19

In this section, I study an extension where an η fraction of all long-term projects require the enforcement of primary financiers whereas 1−η fraction are enforced by secondary financiers. This assumption implies that the η fraction of primary financiers from a cohort are compensated with the capital price wedge when settling secondary market payments. I will derive a critical level of η ∗ , and show that financial development stimulates steady-state aggregate output produced by long-term projects only if η ≥ η ∗ . I will also analyze the behavior of η ∗ with respect to equilibrium the wage rate, and draw some conclusions concerning the effects of stage of economic development on finance-development nexus.

5.1

Equilibrium

Denote the steady-state per-saver capital invested in production again by x. The aggregate capital invested in long-term production equals to Ms x which implies that the aggregate compensation of primary savers must equal to MS xR(1 + ηφ). We can derive x as a function of R and φ: x=

w . 1 + R(1 + ηφ)

The capital market clearance provides ML =

zR(1 + φ) . 1 + (1 + z)R(1 + ηφ)

(26)

Using the expression for ML in aggregate production, the steady-state aggregate output produced by long-term projects can be stated as [R(1 + φ)]1−α AL z 1−α [w]α M. W = 1 + (1 + z)R(1 + ηφ)

(27)

Note that, the real rate of interest R is determined uniquely by the endogenous occupation choice as in ˜ ˜ sections 3 and 4. Therefore, ∂R/∂φ < 0 continues to hold. Define R(φ) = R(1 + φ) where ∂ R/∂φ > 0. Differentiating W with respect to φ we can show that lowering φ would stimulate steady-state aggregate long-term project output only if:   h i1−α h i−α 0 0 ˜ (φ) (1 + z) R(φ) ˜ ˜ ˜ (1 − η)(1 + z) (R (φ)φ + R(φ)) < R − (1 − α) R(φ) [1 + (1 + z)R(φ)] (28)

Proposition 5.1 There exists a critical η ∗ < 1 such that if and only if η > η∗, the inequality at (28) is satisfied, and financial development stimulates steady-state aggregate long-term project output.

With financial structure 1, financial development promotes economic growth whereas with financial financial structure 2 financial development could be growth-reducing. Proposition 5.1 shows that when the financial regime is characterized as a combination of the two, the hybrid regime should resemble 20

the financial structure 1 as much as possible in order financial development and economic growth to be positively related. Using inequality (28) we can study the behavior of η ∗ with respect to the steady-state wage rate w.
1−α ˜ 0 (φ) are proportional to 1 1+α . Therefore, it is useful to re-write (28) Note that, R(φ), R0 (φ) and R w as:  0  h i1−α h i−α R (φ)φ + R(φ) ˜ ˜ ˜ (1 − η)(1 + z) ≤ (1 + z) R(φ) − (1 − α) R(φ) [1 + (1 + z)R(φ)] (29) ˜ 0 (φ) R Proposition 5.2 The threshold η ∗ is an increasing function of w.

Proof The left hand-side of (29) is constant in wage rate w whereas the right hand side is a decreasing function of w. Hence, the threshold η ∗ that satisfies (29) with an equality, increases as w rises. 

5.2

Discussion

The steady-state wage rate w is larger in an economy that is at a higher stage of economic development. To this end, proposition 5.1 provides an important insight concerning the influence of economic development on how finance and economic development might be related: In richer economies, the fraction of primary financiers who incur the financial intermediation costs must be relatively large compared to a low income country such that financial development could promote aggregate output produced by long-term projects. This implies that in high-income countries the chances of financial development policies to be counterproductive is relatively higher compared to low income countries. This result matches with empirical findings that point out a potential non-monotone relationship between finance and economic development: For instance, Castro et al. (2004) measure costly external finance by “limited investor protection” and present empirical evidence for the non-linear growth effects of financial development. An important empirical result from their work shows that investor protection and economic growth exhibit a negative relationship especially for high income countries. Similar conclusions are also reached by Reinhardt and Tokatlidis (2005). The authors show that the economic growth rates in low income countries benefit relatively more compared to growth experiences in relatively high income countries following the implementation of domestic financial deregulation policies.

6

Conclusion

I studied the interactions between financial intermediation costs, secondary market trading and economic development in an overlapping generations model of occupation choice. The key analytical finding from the paper is the endogenous existence of a non-monotone effect of loan enforcement costs on steady-state output. I showed that a large size of secondary market is necessary but not a sufficient condition for the non-monotone real effects of financial development. I characterized the behavior of secondary market prices of financial claims and the microfoundations consistent with such behavior that determine the implications of financial development for economic growth.

21

The key contribution of the paper assigns a central role to the delegation of loan enforcement in explaining the finance & development nexus. Specifically, the results from sections 3 and 4 show that holding everything else constant, a financial regime that delegates the enforcement of loan repayment to primary financiers (financial structure 1) is less productive from a macroeconomic point of view relative to a regime where secondary financiers enforce loan repayment (financial structure 2). Furthermore, comparing the growth rates, under financial structure 1 lowering financial intermediation costs (cost of enforcement) raises aggregate long-term investment and stimulates productivity growth whereas with financial structure 2 financial development reduces aggregate long-term investment and lowers productivity growth. Finally, as presented in section 5, countries with a high level of economic development are the most likely candidates to suffer from counterproductive financial development experiences. The theoretical results presented in this paper provide empirically testable predictions and important policy conclusions. The regulatory framework as well as the complexity of financial instruments sold in secondary markets are expected to dictate whether primary or secondary financiers of long-term assets enforce repayment. The conclusions from my analysis suggests that policy makers need to pay attention to the effects of financial innovation that shifts the enforcement of loan repayment to primary financiers in order to avoid banking sector policies with potential counterproductive consequences.

22

References 1. Aghion, P., Angeletos, G.-M., Banerjee, A., Manova, K., 2010. Volatility and growth: Credit constraints and composition of investment. Journal of Monetary Economics, 57 (3). 2. Aiyagari, S.R., 1994. Uninsured Idiosyncratic Risk and Aggregate Saving. Quarterly Journal of Economics, 109 (3). 3. Angeletos, G.M., and Calvet, L.E., 2007. Idiosyncratic production risk, growth and the business cycle. Journal of Monetary Economics, 53 (6). 4. Antunes, A., Cavalcanti, T., and Villamil, A., 2008. The effect of financial repression and enforcement on entrepreneurship and economic development. Journal of Monetary Economics, 55 (2). 5. Bandiera, G., Caprio, P., Honohan, and Schiantarelli, 2000. Does Financial Reform Raise or Reduce Saving? Review of Economics and Statistics, 82 (2). 6. Banerjee, A. and Newman, A., 1993. Occupational Choice and Process of Development. Journal of Political Economy 101 (2). 7. Bayoumi, T., 1993. Financial Deregulation and Household Saving. Economic Journal, 103. 8. Beck, T., Levine, R., and Loayza, N., 2000. Financial Intermediation and Growth: Causality and Causes. Journal of Monetary Economics, 46 (1). 9. Beck, T. and Levine, R., 2002. Industry Growth and Capital Allocation: Does Having a Marketor Bank-Based System Matter? Journal of Financial Economics, 64. 10. Beck, T. and Levine, R., 2002. Stock Markets, Banks, and Growth: Panel Evidence Journal of Banking and Finance, 28. 11. Beck, T., Demirguc-Kunt, A., Laeven., L., and Levine, R, 2008. Finance, Firm Size, and Growth. Journal of Money, Credit and Banking, 40(7). 12. Bencivenga, V. and Smith, B., 1991. Financial Intermediation and Endogenous Growth, Review of Economic Studies, 58 (2). 13. Bencivenga, V., Smith, B., and Starr, R.M, 1995. Transactions Costs, Technological Change and Endogenous Growth, Journal of Economic Theory, 67. 14. Bencivenga, V., Smith, B., and Starr, R.M, 1996. Liquidity of Secondary Markets: Allocative Efficiency and the Maturity Composition of the Capital Srock, Economic Theory, 7 (1). 15. Castro, R., Clementi, G.L., and MacDonald, G., 2004. Investor Protection, Optimal Incentives, And Economic Growth. Quarterly Journal of Economics, 119 (3). 16. Christopoulos, D.K. and Tsionas, E.G., 2004. Financial Development and Economic Growth: Evidence from Panel Unit Root and Cointegration Tests. Journal of Development Economics, 73 (1). 17. Deidda, L., and Fatouh, B., 2002. Non-linearity between finance and growth. Economic Letters. 18. Demirguc-Kunt, A., and Huizinga, H., 1999. Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence. The World Bank Economic Review, 13 (2). 23

19. Demirguc-Kunt, A. and Maksimovic, V., 1999. Institutions, financial markets, and firm debt maturity. Journal of Financial Economics, 54. 20. Demetriades, P.O. and Luintel, K.B., 2001. Financial restraints in the South Korean miracle. Journal of Development Economics, 64 (2). 21. Diamond, D. and Dybvig P., 1983. Bank Runs, Deposit Insurance, and Liquidity. Journal of Political Economy, 91 (3). 22. Greenwood, J. and Jovanovic, B., 1990. Financial Development, Growth, and the Distribution of Income. Journal of Political Economy, 98 (5). 23. Geanakoplos, J. and Fostel, A., 2008. Collateral Restrictions and Liquidity Under-Supply: A Simple Model. Economic Theory, 35. 24. Levine, R., 1993. Finance and Growth: Schumpeter might be right. Quarterly Journal of Economics, 108 (3). 25. Levine, R., 1993. Financial Development and Economic Growth: Views and Agenda. Journal of Economic Literature, 35 (2). 26. Masten, A.B., Coricelli F., Masten, I., 2008. Non-linear growth effects of financial development: Does financial integration matter? Journal of International Money and Finance. 27 (2). 27. Matsuyama, K., 2004. Financial Market Globalization, Symmetry-Breaking, and Endogenous Inequality of Nations. Econometrica, 72 (3). 28. Pagano, M. and Jappelli, T., 1994. Saving, Growth and Liquidity Constraints. Quarterly Journal of Economics, 109 (1). 29. Romer, P., 1990. Endogeneous Technological Change. Journal of Political Economy, 98. 30. Uras, R.B., 2012. Investment Composition and Productivity with Heterogeneous Entrepreneurs. Working Paper.

24

W

W2 W1

W (R, φ) a function of R.

R(φ1 )R(φ2 )

... ... ... ... ... ... ... ... ... ... ... .... ... .... .... ... ....... .... ....... .... ........ .... ......... .... .......... ............ ..... ............... ...... .................. ...... .......................... ....... ............... ....... ........ ........ ......... .......... ........... ............ ............. ............... ................. .........

6

When φ1 decreases to φ2 , R(φ) shifts right from R1 to R2 ,

W (R, φ2 )

W (R, φ1 )

W (R, φ) shifts up to W (R, φ2 ), W1 increases to W2

-

R1

R

R2

Figure 1. Financial Structure 1: Costly Primary Finance

W 6

W2 W1

W (p(φ)) is decreasing in p(φ),

p(φ2 ) p(φ1 )

... ... ... ... ... ... ... .... .... .... .... ..... ...... ...... ....... ....... ........ ........ ......... .......... ........... ............ ............. ............... ................. .........

p(φ) is decreasing in φ. ⇒ A decline in φ stimulates W .

W (p(φ1 )) -

p2

R

p1

Figure 2. Financial Structure 2: Costly Secondary Finance

W 6

W1 W2

W (R, φ) ≡ W (R, p),

R(φ1 )R(φ2 )

... ... ... ... ... ... ... ... ... ... ... .... ... .... .... ..... .. .... .... ........ .. .... ..... ........... . ...... ...... ............. ....... . ....... ............... . ........ ........ .................. . ......... .......... ...................... . ........... ............ ............................ .. ............. ............... ........................ ................. .........

W (R, p) is decreasing in both arguments. ∂p/∂φ < 0 but ∂p/∂φ > 0. “R” channel dominates “p” channel,

W (R, φ2 ) W (R, φ1 ) -

R1

R2

R

⇒ As φ1 declines to φ2 , W1 decreases to W2 .

Figure 3. Financial Structure 2: Costly Secondary Finance

25