Journal of Monetary Economics 43 (1999) 283—315

Inflation, financial markets and long-run real activity

Elisabeth Huybens *, Bruce D. Smith

Centro de Investigacio& n Econo& mica, ITAM, Camino Santa Teresa 930, Mexico D.F.10700, Mexico

Department of Economics, University of Texas, Austin, TX 78712-1173, USA

Received 25 August 1997; received in revised form 5 March 1998; accepted 6 May 1998

Abstract

Empirical evidence suggests that real activity, the volume of bank lending activity, and
the volume of trading in equity markets are strongly positively correlated. At the same
time, inflation and financial market activity are strongly negatively correlated (in the long
run), as are inflation and the real rate of return on equity. Inflation and real activity are
also negatively correlated in the long run, particularly for economies with relatively high
rates of inflation. We present a monetary growth model in which banks and secondary
capital markets play a crucial allocative function. We show that — at least under certain
configurations of parameters — the predictions of the model are consistent with these and
several other observations about inflation, finance and long-run real activity.  1999
Elsevier Science B.V. All rights reserved.

JEL classification: E5; E6; G2

Keywords: Inflation; Financial markets; Real activity

1. Introduction

Extensive empirical investigation has established the following facts about the
long-run relationships between inflation, financial market conditions, and real
activity.

* Corresponding author. E-mail: [email protected]

 We thank Jaime Calleja Alderete and Eduardo Camero Godı́nez for excellent research assist-
ance, as well as an anonymous referee for his extremely helpful comments.

0304-3932/99/$ — see front matter  1999 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 4 - 3 9 3 2 ( 9 8 ) 0 0 0 6 0 - 9
284 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

1. There is a negative long-run relationship between inflation and real economic

performance, at least at sufficiently high levels of inflation.
2. There is a positive correlation between economic performance and (a) the
volume of bank lending activity, (b) the quantity of bank liabilities issued, and
(c) the volume of trading in equity markets. (King and Levine, 1993a,b;
Levine and Zervos, 1998; Atje and Jovanovic, 1993).
3. At low-to-moderate long-run rates of inflation, there is a strong negative
correlation between inflation and (a) the volume of bank lending activity, (b)
the quantity of bank liabilities issued, and (c) the volume of trading in equity
markets. However, at higher rates of inflation these partial correlations
essentially disappear. (Boyd et al., 1996).
4. For countries with low-to-moderate rates of inflation, there is a pronounced
negative correlation between inflation and real equity returns. For countries
where inflation is sufficiently high, this correlation vanishes. (Nelson, 1976;
Fama and Schwert, 1977; Gultekin, 1983; Boyd et al., 1996).
5. As economies develop, equity markets generally become more important
relative to banks. (Gurley and Shaw, 1960; Levine, 1997).
Our purpose in this paper is to develop a theoretical model that is capable of
confronting this full set of observations. Such a model, clearly, must have several
features. First there must be a role for intermediaries, as well as equity (second-
ary capital) markets, and the level of intermediary and equity market activity
must be related to long-run real performance. Second, money must play a role in
the economy, and changes in monetary policy (that affect the long-run rate of
inflation) must be capable of affecting long-run real activity, as well as financial
market conditions.
In order to produce an economy with these features, we utilize an overlapping
generations model with production, similar to that of Diamond (1965). How-
ever, we modify Diamond’s model in several important respects. First, we allow
for the existence of multiple technologies for producing capital. Some of these
technologies are simple in nature, meaning that (a) anyone can operate them, (b)
they produce capital quickly, and (c) they are subject to no informational
asymmetries. Others are more complex, meaning that (a) only a subset of
individuals can operate them, (b) they involve a relatively lengthy gestation
period, and (c) informational asymmetries arise. More specifically, we assume

 Barro (1995) asserts that there is evidence of a negative relationship at all rates of inflation.
Bruno and Easterly (1998) and Bullard and Keating (1995) find support for the notion that this
negative relationship emerges only when rates of inflation exceed some threshold. Levine and Renelt
(1992) and Clark (1997) also question whether there is a uniformly negative relationship between
inflation and real activity independently of the prevailing rate of inflation.
 The importance of long gestation periods in capital production has been emphasized by authors
from Bohm-Bawerk (1891) to Kydland and Prescott (1982). The specific model we adopt resembles
Bencivenga et al. (1995, 1996) in the nature of the gestation lags in capital production.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 285

that capital production using long-gestation technologies is subject to a stan-

dard costly state verification (CSV) problem.
The second central feature of the analysis is that agents employing long-
gestation production technologies require some external finance in order to
operate their projects. In particular, we assume that the long-gestation techno-
logy can be operated only on a relatively large scale; a scale that requires agents
to invest more than just their own funds. In contrast, the short-gestation
technology can be operated at any scale. As a result, no external finance is
required to utilize it. However, whenever long-gestation capital production
technologies are in use, external finance is a necessity. Given the presence of the
CSV problem, this external funding will naturally be provided through inter-
mediaries (Diamond, 1984; Williamson, 1986). As a consequence, intermediary
activity will play an important role in the allocation of investment, and in the
determination of long-run real activity levels.
Third, given our assumptions on agents’ life-cycles and the gestation of certain
capital investments, agents utilizing long-lived capital production technologies
will find it necessary to sell claims to the ownership of the capital they produce
in secondary capital markets. Hence, whenever these technologies are in use,
capital formation and equity market activity must be positively related, at least
under certain technical conditions that we deduce.
These features of the economy create a role for agents to trade equity, and to
hold intermediary liabilities. In addition we will obviously assume that the
government issues a third asset that can be held by agents — fiat money. Here, as
in Diamond (1965) or Tirole (1985), we simply treat money as an additional asset
that can be held by any agent; it plays no special role in transactions. Thus, in
equilibrium, the real return on money must be equated to the real return on
competing assets. However, this feature of the model is not central to our
analysis; it would be straightforward to create a role for money in transactions
— and an additional role for banks — along the lines described by Champ et al.
(1996) or Schreft and Smith (1997). We do not follow this route for two reasons.
First, the simpler Diamond—Tirole formulation matches quite well the empirical
relationship between inflation and real equity returns. And second, the Dia-
mond—Tirole model generally gives rise to a Mundell—Tobin effect — that is to
a positive relationship between inflation and long-run output levels — in the
absence of the other features we have introduced. Thus our formulation makes it
clear that these features are essential in allowing our model to confront all of the
empirical observations described above.
Under the assumption that the supply of outside money grows at a constant,
exogenously selected rate, we provide conditions under which there will be two

 Hicks (1969) identifies the central feature of the industrial revolution to be the employment of
large scale technologies requiring the provision of external finance for their operation.
286 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

or more steady state equilibria. These steady states are differentiated by their
capital stocks and levels of real activity, as well as by the level of activity in their
financial markets. In the low-capital-stock steady state agents will utilize either
the commonly available or the long-gestation capital production technology. In
the former case clearly no financial market activity is necessary. In the high-
capital-stock steady state, on the other hand, production of capital will occur
using the long-gestation technology; hence both intermediaries and equity
markets will be active. For this steady state, we are able to state conditions
under which higher steady state levels of real activity are associated with higher
volumes of both bank lending and equity market activity. We also state condi-
tions under which equity market activity rises relative to bank lending activity
as the steady state capital stock rises.
With respect to inflation, we show that higher rates of money creation (steady
state inflation) lead to lower levels of real activity, in the high-capital-stock
steady state. They also reduce the real return on all assets, including equity, and
higher money growth is (under conditions we state) detrimental both to bank
lending activity and to the volume of trading in equity markets. Moreover, it is
also possible to state conditions under which the negative relationship between
inflation and long-run real activity becomes more pronounced at higher rates of
inflation, as many have argued is true empirically.
Finally we spend some time analyzing the properties of equilibrium dynamics,
in a neighborhood of each steady state. We are able to show that the low-
capital-stock steady state is a saddle; hence there are always equilibrium paths
that approach it. In addition, we illustrate by example that the high-capital-
stock steady state may be either a source or a saddle. If it is a saddle, it too can
potentially be approached from some combination of initial conditions. Our
examples further demonstrate the possibility that the stability properties of the
high-capital-stock steady state depend upon the rate of money growth. As the
rate of money creation rises, it can transpire that the high activity steady state is
transformed from a saddle to a source. Hence it can no longer be approached,
and if the economy converges to a steady state, it must converge to the low
activity steady state. Consequently, for economies whose inflation rate exceeds
some critical level, the only ‘relevant’ steady state may be the low activity one.
Thus economies with high enough rates of inflation can display discretely lower
long-run levels of real and financial market activity than their low inflation
counterparts. Moreover, economies in low activity steady states (those with high
rates of inflation) will display a much different correlation pattern between
inflation and financial market conditions than will economies in high activity
steady states (those with lower rates of inflation).

 Boyd and Smith (1998) and Huybens and Smith (1998) obtain some analytical results on this
point in much simpler, but related models.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 287

What accounts for the existence of multiple steady states, and for the other
findings we have described? In this economy, a steady state equilibrium has the
property that the return on loans and the return on equity must equal the
prevailing real rate of return on money. This is given, in a steady state, by the
exogenously determined rate of money growth. In the presence of the CSV
problem, the return on loans depends on two factors: the marginal product of
capital, and the quantity of internal finance provided. Thus the required rate
of return on loans can be obtained either by having a relatively low stock of
capital, along with a relatively high marginal product of capital, a low level
of income, and a low level of internal finance; or by having a relatively high stock
of capital, a correspondingly high level of internal finance, but a low marginal
product of capital. These two methods of delivering a particular rate of return
typically yield two steady state equilibria. The high activity steady state will also
have a high level of internal finance, and the higher the steady state capital stock,
the higher will be the amount of internal finance provided.
If the steady state real return is lowered by an increase in the steady state rate
of inflation, in the high activity steady state the result is a decline in the quantity
of internal finance provided. This produces the required decline in the real
return on loans. It also exacerbates the severity of the CSV problem and leads to
more extensive rationing of credit. As a result, the capital stock falls. And, under
the conditions we describe, so does the volume of financial market activity. Thus
this model can potentially account for all of the observations cited above.
The paper proceeds as follows. Sections 1 and 2 describe the environment and
the nature of trade in the model. Section 3 analyzes steady state equilibria.
Section 4 discusses steady state financial market activity and its relation to
inflation. Section 5 briefly discusses local dynamics, while Section 6 concludes.

2. The model

We examine an economy consisting of an infinite sequence of two-period

lived, overlapping generations. Each generation is identical in size and composi-
tion, and contains a continuum of agents with unit mass. Throughout, we let
t"0,1, 2 index time.
At each date a single final good is produced using a constant returns to scale
technology with capital and labor as inputs. Let K denote the time t capital
R

 Again we emphasize that this feature of the economy is not essential to our results. What is
necessary is that higher inflation lowers long-run real returns. Negative associations between
inflation and real returns on a variety of savings instruments are well-documented empirically.
 The latter point was first made by Bernanke and Gertler (1989).
 Hamid and Singh (1992) document that, as an empirical matter, countries with high income
levels also — on average — have high fractions of capital investments that are financed internally.
288 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

input, and ¸ denote the time t labor input of a representative firm. Then its final
R
output is F(K , ¸ ). We will assume that F is a CES production function with
R R
elasticity of substitution greater than 1, that is F(K, ¸),[aKM#b¸M]M, with
0(o(1. Thus F is increasing in each argument and strictly concave. In
addition, if k,K/¸ is the capital—labor ratio, and if f (k),F(k, 1)"
[akM#b]M is the intensive production function, then f '0'f  holds ∀k, and
in addition lim f (k)"R.
I
Agents are assumed to care only about old age consumption and, in addition,
all agents are risk neutral. Thus all young period income is saved.
There are potentially three assets in our economy, money and investments in
the two different technologies for converting final goods into capital. The two
capital production technologies are indexed by j"1,2. Technology j"1 is
a simple capital production technology: one unit of the final good invested at
t returns R '0 units of capital at t#1. Technology j"2 is a more complic-

ated capital production technology, which has the following properties. First,
only a fraction d3(0, 1) of the population — which we will call potential
borrowers — has access to this technology. The remaining fraction (1!d) of the
population — which we will call lenders — does not have access to the complic-
ated capital production technology. Second, the technology is indivisible: each
potential borrower has one investment project which can only be operated at
the scale q. Third, when this technology is utilized, two periods are needed to
obtain mature capital. Fourth, the return on investments in technology 2 is
random. More specifically, then, q'0 units of the final good invested in
technology 2 at t yield zq units of capital in progress (CIP) at t#1, and R zq

units of capital at t#2. The random variable z is iid (across borrowers and
periods), and is realized at t#1. We let G denote the probability distribution of
z, and assume that G has a differentiable density function g with support [0,zN ].
Let zL be the expected value of z. Finally, we assume that this technology is
subject to a standard CSV problem of the type introduced by Townsend (1979):
only the project owner can costlessly observe z, while any agent other than the
project owner can observe z only by bearing a fixed cost of c'0 units of capital
in progress (CIP).
Capital produced by the simple investment technology is a perfect substitute
for capital produced in the alternative fashion. Moreover, we assume that the
capital stock depreciates completely after being used in production.
With respect to endowments, all young agents are endowed with one unit of
labor, which is supplied inelastically, and agents are retired when old. Indi-
viduals other than the old of period zero have no endowment of capital or final

 That is, in verifying the project return, c units of CIP are used up. This assumption is responsible
for the simple form assumed by the expected return to lenders under credit rationing (see
Eq. (15)).
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 289

goods, while the initial old agents have an aggregate capital endowment of
K '0, and an aggregate endowment of capital in progress, CIP '0.
 

3. Trade

3.1. Factor markets

We assume that capital and labor are traded in competitive markets at each
date. Then, letting w denote the time t real wage rate and o the time t capital
R R
rental rate, the standard factor pricing relationships obtain:
o "f (k ), (1)
R R
w "f (k )!k f (k ),w(k ). (2)
R R R R R
Clearly w(k)'0 holds.

3.2. Credit markets

All young agents at t supply one unit of labor inelastically, earning the real
wage rate w . However, we will assume that this young period income does not
R
suffice to run a capital production project of type j"2.

Assumption 1. q'w(k ) for all ‘relevant’ values of k .

R R
Thus, potential borrowers must obtain external financing to invest in techno-
logy 2. Let b denote the amount borrowed (in real terms) at t by the operator of
R
a funded type 2 project ; clearly
b "q!w (k ). (3)
R R
We can think of this borrowing as being intermediated (Williamson, 1986).
If potential borrowers attempt to obtain external funding they do so by
announcing loan contract terms. These announced contract terms are either
accepted or rejected by intermediaries: borrowers whose terms are accepted then
receive funding and operate their projects. A loan contract consists of the
following objects. First, there is a set of project return realizations A for which
R
verification of the project return occurs at t. Verification of project returns does
not occur if z3B ,[0, zN ]!A . Second, if z3A , then it is possible to make the
R R R

 We thus abstract from stochastic state verification. In a similar context, Boyd and Smith (1994)
show that the welfare gains from stochastic monitoring are trivial when realistic parameter values
are assumed.
290 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

contractual repayment contingent on the project return. Thus if z3A we denote

R
the promised payment (per unit borrowed) by R (z). On the other hand, if z3B
R R
then the loan payment cannot meaningfully depend on the project return, and
the loan contract offers an uncontingent payment of x (per unit borrowed) for
R
all z3B . All payments specified by any contract are in real terms.
R
Loan contracts offered by borrowers are either accepted or rejected by
intermediaries who — without loss of generality — we can think of as making all
loans. Thus intermediaries take deposits, make loans, and conduct the monitor-
ing of project returns. We assume that any lender can establish an intermediary.
In equilibrium intermediaries will be perfectly diversified, earn zero profits, and
have a nonstochastic return on their portfolios.
Since agents are two-period lived, a young borrower who initiates a capital
investment of type j"2 will seek to sell his ‘immature’ capital in a secondary
market. Let u denote the price of one unit of capital in progress (CIP) at time t.
R
Intermediaries accept deposits taking the gross real return that must be paid
on them — r between t and t#1 — as given, and they act as if they can obtain any
R
desired quantity of deposits at that rate. It follows that intermediaries are willing
to accept loan contract offers yielding an expected return no less than r . Thus
R
loan contract offers must satisfy the expected return constraint

R
[R (z)b !u c]g(z)dz#x b g(z)dz5r b .
R R R> R R
R
 R R
(4)

In particular, expected repayments must at least cover the intermediary’s cost of

funds, r b , plus the real expected monitoring cost
R R

u
R>
c
R

g(z)dz.

The expected monitoring cost depends on u , of course, because c units of

R>
capital in progress are expended when project returns are verified. Finally,
project owners must have the appropriate incentives to correctly reveal when
a monitoring state has occurred. This requires that
R (z)4x , z3A . (5)
R R R
In addition, contractually specified repayments must be feasible for the bor-
rower, so that
u zq
R (z)4 R> , z3A , (6)
R b R
R

 As a result, intermediaries need not be monitored by their depositors. See Krasa and Villamil
(1992) for a consideration of intermediaries that cannot perfectly diversify risk.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 291

 
u zq
x 4inf R> . (7)
R b
X R
o
R
Eqs. (6) and (7) state that repayments never exceed the real value of the CIP
yielded by an investment project, which in state z is u zq at t#1.
R>
Borrowers announce contract terms in order to maximize their own expected
utility subject to the constraints (4)-(7). Therefore, announced loan contracts at
date t will be selected to maximize

u
R>
zL q!b
R 
R
R (z)g(z)dz!x b
R R R  g(z)dz
R

subject to these constraints.

At an optimum, borrowers offer a standard debt contract (modified for the
presence of internal finance). In particular, the borrower either repays x (princi-
R
pal plus interest) or else defaults. In the latter case the lender verifies the project
return, and retains the proceeds of the project net of monitoring costs. Formally,

Proposition 1. Suppose q'b . ¹hen the optimal contractual loan terms satisfy
R
u zq
R (z)" R> , z3A , (8)
R b R
R

 
xb
A " 0, R R , (9)
R (u q)
R>

  
u c
r" R (z)! R> g(z)dz#x g(z)dz. (10)
R R b R
R R R

The proof of Proposition 1 is standard, and we omit it here.

For future reference, substituting Eqs. (8) and (9) into Eq. (10) yields

  
u c
R (z)! R> g(z)dz#x g(z)dz
R b R
R R R


 
u c xb u q VR@RSR>O
"x ! R> G R R ! R> G(z)dz
R b (u q) b
R R> R 

 
b
,n x ; R "r . (11)
R u R
R>

 See Gale and Hellwig (1985) or Williamson (1986, 1987).

292 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

The function n gives the expected return to a lender as a function of the gross
loan rate, x , the amount of external finance required, b , and the future relative
R R
price of CIP, u .
R>
It will be useful in what follows to put some additional structure on the
function n. In particular, we will assume the following.

Assumption 2. g(z)#(c/q)g(z)50; for all z3[0,zN ].



b
Assumption 3. n 0, R '0.
 u
R>
Assumption 2 implies that n (0. Assumptions 2 and 3 imply that the

function n has the configuration depicted in Fig. 1. Evidently, given b /u ,
R R>
there is a unique value of x which maximizes the expected return that can be
R
offered. We denote this value by xL (b /u ), where the function xL is defined
R R>
implicitly by


 

  
 
b b c b b b b
n xL R ; R ,1! g xL R R !G xL R R
 u u q u (u q) u (u q)
R> R> R> R> R> R>
,0. (12)

Fig. 1. The expected return function.

 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 293

Eq. (12) and assumption 3 imply that


b b
xL R R ,g, (13)
u (u q)
R> R>
where g'0 is a constant satisfying 1!(c/q)g(g)!G(g),0. When all potential
borrowers offer the interest rate xL (b /u ), project return verification occurs iff
R R>
z3[0,g).

3.3. Credit rationing

A well known feature of the environment just described — which was originally
noted by Gale and Hellwig (1985) and Williamson (1986, 1987) — is that it
permits the existence of unfulfilled demand for credit. In particular, if all
borrowers desire to operate their projects at date t, the total (per capita) demand
for funds is dq. The total per capita supply of saving is w(k ) at t. Therefore credit
R
demand must exceed credit supply, and hence credit must be rationed, if the
following assumption holds for all t50.

Assumption 4. dq'w(k ).
R
When credit rationing exists, however, it also must be the case that


b
x "xL R . (14)
R u
R>
Eq. (14) asserts that all potential borrowers are offering the interest rate that
maximizes a prospective lender’s expected rate of return. As a result, rationed
(unfunded) potential borrowers cannot obtain credit simply by offering an
alternative set of loan contract terms, since doing so reduces the expected return
perceived by (all) lenders. Thus if assumption 4 and Eq. (14) hold at date t, credit
rationing is an equilibrium outcome. We focus here on economies where credit is
rationed at all dates.

 When credit is rationed, the probability that any project will have to be monitored, ex post, is
simply G(g). Thus the monitoring probability — and, by implication, the probability of bankruptcy
— is independent of any endogenous variables. The result is a substantial technical simplification, as
is illustrated by the relatively simple expression in Eq. (15) describing the expected return received by
a lender.
Of course credit rationing is clearly a widespread phenomenon in developing countries (Mc-
Kinnon, 1973), and there is substantial evidence of significant rationing of credit even in the
United States (Japelli, 1990). Therefore this does not seem to be an empirically unreasonable
assumption.
294 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

3.3.1. Payoffs under credit rationing

We now describe the expected payoffs received by lenders and (funded)
borrowers when credit is rationed. For lenders, Eqs. (11) and (14) imply that the
expected return on bank deposits (and loans) at t satisfies


 
b b
r "n xL R ; R
R u u
R> R>



 
 
u q b b c b b
, R> xL R R ! G xL R R
b u (u q) q u (u q)
R R> R> R> R>

  
  
VL @RSR>O u q c E
! G(z)dz , R> g! G(g)! G(z)dz . (15)
b q
 R 
In particular, the return on savings between t and t#1 is proportional to the
ratio u /b when credit rationing obtains.
R> R
It is also possible to show that the expected utility of a funded borrower at
t under credit rationing is given by


 
b b
u zL q!r b !u cG xL R R
R> R R R> u (u q)
R> R>


 
c
,u q zL ! G(g) !r b .
R> q R R
Since any potential borrower could always forego investing in his project and
deposit his income in a bank instead, all potential borrowers can guarantee
themselves the utility level r w . Thus (potential) borrowers wish to operate their
R R
projects (under credit rationing) iff


 
c
u q zL ! G(g) !r b 5r w . (16)
R> q R R R R
We now define


c

,zL ! G(g),
q
(17)


  
c E
t,q g! G(g)! G(z)dz .
q

The parameter
is the expected project yield per unit invested, net of CIP
consumed by monitoring, under credit rationing. The parameter t determines
the expected return on deposits under credit rationing, since
u u
r "t R>,t R> . (18)
R b [q!w(k )]
R R
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 295

We now observe that Eq. (16) is satisfied iff

u 5r . (19)
R> R
Eqs. (18) and (19) imply that

[q!w(k )]5t (20)
R
must hold for all t50 in order for borrowers to wish to operate their projects.
Given assumption 4, a sufficient condition for Eq. (20) to obtain is that
(1!d)
q5t. (21)

3.4. Money

The initial old at time zero are endowed with the initial per capita money
supply M 50. Thereafter, the money supply grows at the constant (gross)
\
rate p51, which the government selects once and for all at t"0. Therefore
M "pM , t5!1. (22)
R> R
We let the government have an endogenous real expenditure level of g (per
R
capita) at time t. The government budget constraint implies that
M !M
g" R R\. (23)
R p
R
Letting m ,M /p denote the per capita stock of real balances, Eqs. (22) and
R R R
(23) imply that


p!1
g" m.
R p R

4. General equilibrium conditions and steady states

In this section, we present equilibrium conditions and examine steady state

equilibria in which the same capital production technology ( j), is in use perma-
nently. We then present conditions determining the equilibrium choice of
investment technology. We begin with the case where the short-gestation capital
production technology is in use.

 The analysis would be unaltered if monetary injections ocurred via lump-sum transfers to old
lenders.
296 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

4.1. Steady state equilibria when the production technology j"1 is utilized

When the economy produces only type j"1 capital, an equilibrium in which
capital investments coexist with money at all dates must satisfy the no-arbitrage
condition
p
R f (k )" R , t50. (24)
 R> p
R>
By definition,
p m M m
R , R> R , R>, (25)
p m M pm
R> R R> R
so that Eq. (24) can be rewritten to yield
m
R f (k )" R> , t50. (26)
 R> (pm )
R
In addition, ‘sources’ and ‘uses’ of funds must be equal in equilibrium. If we let
i denote the per capita quantity of resources invested in capital production at t,
R
then an equality between sources and uses of funds requires that
w(k )"i #m , t50,
R R R
since young agents save all of their wage income. Of course,
k "R i , t50,
R> R
and therefore
k
m "w(k )! R>, t50. (27)
R R R

In steady state, Eqs. (26) and (27) reduce to
1
R f (k)" , (28)
 p
k
m"w(k)! . (29)
R

Given our assumptions on the production technology f, it is clear that a unique
monetary steady state k exists when the economy produces only type j"1
Q
capital.

4.2. Steady state equilibria when the production technology j"2 is utilized

When the economy produces only type j"2 capital, an equilibrium in which
money is valued, and in which loans to capital producers are made at all dates
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 297

requires that the returns on these two assets must be equalized,

u p
r "t R> " R , t50. (30)
R [q!w(k )] p
R R>
Furthermore, in order for young lenders at time t#1 to buy CIP at the price
u , it must be the case that, in equilibrium, these claims to capital ownership
R>
yield the same return as bank deposits or money between periods t#1 and
t#2. Therefore
R f (k ) p
 R> "r " R>, t50 (31)
u R> p
R> R>
must be satisfied as well. Eqs. (30) and (31) then imply the following equilibrium
condition
R f (k ) p p
r r "t  R> " R R>, t50. (32)
R R> [q!w(k )] p p
R R> R>
Using Eq. (25), Eq. (32) can be rewritten as
R f (k ) 1 m
t  R> " R>, t50. (33)
[q!w(k )] p m
R R
As before, it must also be the case that ‘sources’ and ‘uses’ of funds are
equated. If k denotes the fraction of potential borrowers who are funded at t,
R
then the ‘uses’ of funds in real terms at t is dqk , plus the real value of CIP
R
purchased by young agents at time t, plus real balances; that is
dqk #dqk
u #m . ‘Sources’ of funds are simply per capita savings; that is
R R\ R R
w(k ). Therefore
R
w(k )"dqk #dqk
u #m (34)
R R R\ R R
must hold at all dates. Since k "R
i "R
dqk , Eq. (34) can be rewritten
R>  R  R
as
k k
w(k )" R># R>u #m , t50. (35)
R R
R R R
 
Moreover, Eq. (31) implies that u "R pf (k )(m /m ). Substituting this re-
R  R> R R>
sult in Eq. (35) yields


m k
m "w(k )!k pf (k ) R ! R>, t50. (36)
R R R> R> m R

R> 
For an economy which produces only type j"2 capital, Eqs. (33) and (36)
describe the evolution of any equilibrium sequences +k , m , when credit is
R R
298 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

rationed. In a steady state, this dynamical system reduces to


R tf (k) 1
 " "r, (37)
[q!w(k)] p

k
m"w(k)!kpf (k)! . (38)
R


We now define the function H(k) by

f (k)
H(k), . (39)
[q!w(k)]

Then, in a steady state equilibrium where only type j"2 capital is produced, the
per capita capital stock satisfies the following condition:

1
H(k)" . (40)
R tp

It will clearly be necessary to establish some properties of the function H.
These are stated in Lemma 1.

Lemma 1. ¹he function H satisfies

(a) lim H(k)"R

I
(b) lim H(k)"R where kK ,w\(q)
IIK
(c) H(k)4(5)0 iff k4(5) f \(q), and
(d) kH(k)'(",()1 iff k'(",() f \(q).

The proof of Lemma 1 is presented in Appendix A.

Lemma 1 implies that the function H has the configuration depicted in Fig. 2,
and it is clear from this picture and Eq. (40) that there are potentially two steady
states, k and k , when the economy produces only type j"2 capital.
A A

4.3. The equilibrium choice of capital production technology

Whether type j"1 or type j"2 capital will be produced in equilibrium

depends on the respective rates of return on these alternative capital production
technologies. More precisely, we can state the following proposition.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 299

Fig. 2. The steady state capital stock with long gestation investments.

Proposition 2. In a steady state equilibrium, type j"2 capital will be produced iff
R t
f (k)[q!w(k)](  .
(R )

The proof of Proposition 2 is presented in Appendix B.

Corollary 1. Let k be defined by


R t
f (k)[q!w(k)]"  .
  (R )

¹hen type j"1 capital will be produced in steady state if k(k, while type j"2
capital will be produced in steady state if k'k. 

Proof. Clearly the expression f (k)[q!w(k)] is a decreasing function of k. In
combination with Proposition 2, this establishes the result.
We can now distinguish between two cases.

Case 1. k(f \(q). This is the situation depicted in Fig. 3. For case 1, the
following proposition is immediate from an examination of Fig. 3.
300 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

Fig. 3. Steady states — Case 1: k(f \(q).

Proposition 3. (a) Suppose that (R tp)\'H[ f \(q)]. ¹hen there are exactly

two steady state values of k, denoted by k and k in Fig. 3. If R f (k)41/p, then
    solution to
k satisfies Eq. (28). If R f (k)'1/p, then k is given by the smallest
  
Eq. (40). In each case k is  the largest solution to Eq. (40). (b) Suppose that

(R tp)\(H[ f\(q)]. ¹hen there is no monetary steady state with credit

rationing.

Fig. 4 depicts the consequences (for the steady state capital stock) of an
increase in the rate of money creation, when case 1 obtains. As is apparent from
the figure, an increase in the money growth rate (the steady state inflation rate),
increases the steady state capital stock in the low-capital stock steady state, but
decreases the steady state capital stock in the high-capital-stock steady state.
The relationship between the rate of money creation and the steady state capital
stock for case 1 is presented in Fig. 5.

Case 2. k'f \(q). This is the situation depicted in Fig. 6. For case 2, the
following proposition can be deduced from that figure.

Proposition 4. (a) Suppose that 1/p'R f ( k). ¹hen there are exactly two steady

state values of k, denoted by k and k in Fig. 6. ¹he value k satisfies Eq. (28)
  
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 301

Fig. 4. The effect of an increase in the money growth rate — Case 1.

while k is the largest solution of Eq. (40). (b) Suppose that 1/p(R f (k). ¹hen
  
there is no monetary steady state with credit rationing.

The result of increasing the money growth rate (the steady state inflation rate)
is depicted in Fig. 7. As in case 1, an increase in the rate of money creation
increases the steady state capital stock in the low-capital stock steady state, but
decreases the steady state capital stock in the high-capital-stock steady state.
The relationship between the rate of money creation and the steady state capital
stock for case 2 is presented in Fig. 8.
It remains to state conditions under which the steady state level of real
balances is positive. For steady states with the type 1 capital production
technology in use, our assumptions on f (k) imply that real balances are necessar-
ily positive - in the steady state — if R w(k)'k is satisfied. For steady state
   technology is utilized, real
equilibria where the type 2 capital production
balances will be positive necessarily if w(k )!pk f (k )'k /(
R ) holds.
    
Moreover, for steady states determined by Eq. (40) it is necessary to verify that
(i) credit is rationed, and (ii) borrowers prefer to borrow rather than lend. The
former condition will be satisfied if k (
R dq holds, while the latter will be
 
satisfied if
[q!w(k )]5t obtains.

302 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

Fig. 5. The relationship between the money growth rate and the steady state capital stock — Case 1.

From this initial analysis it is clear that our economy is capable of reproduc-
ing several of the empirical facts laid out in the introduction. First, the high-
capital-stock steady state displays a negative relationship between inflation and
real activity (see point 1 of the introduction). Moreover, in case 1 it is easy to
verify that this relationship becomes more pronounced at high rates of inflation.
Second, as is apparent from Eq. (37), the real return on equity holdings, r, is
negatively related to inflation (see point 4 of the introduction). Indeed, the real
return on equity falls one-for-one with increases in the inflation rate. This is
consistent with the large empirical literature that finds an essentially zero
correlation between inflation and nominal equity returns (Nelson, 1976; Fama
and Schwert, 1977; Gultekin, 1983; Boyd et al., 1996). And finally, a high level of
real activity is associated with a high level of internal project finance, as is true
empirically (Hamid and Singh, 1992).
What is the economic intuition behind these results? An increase in the money
growth rate, ceteris paribus, reduces the steady state return on money. Hence,
for money and other assets to be held simultaneously, the return on these assets
has to decrease as well. For steady states in which all capital is produced using
technology 1, and which therefore satisfy Eq. (28), this implies an increase in the
steady state capital stock. For steady states in which all capital is produced
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 303

Fig. 6. Steady states — Case 2: k'f \(q).

using technology 2, and which thus satisfy Eq. (40), the decrease in the return on
money likewise has to be accompanied by a decrease in the return on loans, as
well as a decrease in the return to equity holdings. Given the presence of the
CSV problem, the consequences of this observation depend on the nature of the
steady state equilibrium that obtains. In the low-capital-stock steady state,
a decrease in the rate of return on money implies an increase in the steady state
capital stock. When the capital stock increases, the level of internal finance of
investment projects rises as well, which by itself tends to mitigate the CSV
problem and increase the return on loans. However, on the downward sloping
portion of the function H(k), the higher level of internal finance fails to compen-
sate for the reduction in the marginal product of capital. Hence an increase in
the per capita capital stock leads to the required fall in the rates of return on
loans and equity holdings. In the high-capital-stock steady state the same two
effects are at work. However, on the upward sloping portion of H(k), the
consequences of a change in the level of internal finance dominate the conse-
quences of a change in the marginal product of capital. Therefore a higher
steady state rate of inflation leads to a fall in the steady state capital stock. The
implied reduction in the provision of internal finance more than offsets the effect
of the increase in the marginal product of capital, and again the rate of return on
loans and equity falls in the necessary way.
304 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

Fig. 7. The effect of an increase in the money growth rate — Case 2.

Fig. 8. The relationship between the money growth rate and the steady state capital stock — Case 2.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 305

5. Financial market activity and inflation

We now proceed to discuss the volume of financial market activity in steady

states, and to examine how this is related both to the level of real activity, and to
the rate of inflation. We are able to show that under certain technical conditions,
equity market and bank lending activity are both positively related to the level
of real activity in the high-capital-stock steady state of our economy. This
accords well with the empirical facts presented in point 2 of the introduction.
Moreover, we show that as the level of real activity increases in that steady state,
the importance of equity market activity relative to bank lending activity
increases as well, which accounts for the empirical regularity mentioned in point
5 of the introduction. Of course, when case 2 obtains, no equity market activity
or bank lending activity takes place in the low capital stock steady state.
In the next section we will also show that — at least in case 2 — the low-capital-
stock steady state is always a saddle. In addition we will provide a set of
examples with the following features. When the rate of money creation is
sufficiently low, the high-capital-stock steady state is a saddle, with a two-
dimensional stable manifold. As a result, the high activity steady state can be
approached from some combination of initial conditions. Then, over some range
of money growth rates, higher rates of money creation can — in the high activity
steady state — lead to a reduction in the level of real activity, equity market
activity and bank lending activity, and to a decline of the importance of equity
markets relative to bank lending. However, once the steady state rate of inflation
reaches some critical level, the high-capital-stock steady state is transformed
from a saddle to a source. Consequently the high activity steady state cannot be
approached, and the low-capital-stock steady state is the only economically
relevant steady state equilibrium of our economy. An increase in the money
growth rate above some critical level is therefore accompanied by a sharp
decrease in real activity, if a steady state is attained, while financial markets shut
down altogether when case 2 obtains. Further increases in the rate of money
creation then have no additional effects on the level of financial intermediation
or equity market activity. These results accord well with the evidence presented
in point 3 of the introduction. Once the rate of inflation exceeds some threshold
level, the association between further increases in inflation and financial market
activity disappears.
For the remainder of the analysis, we will focus on economies where case
2 obtains. Thus, we will henceforth adopt the following assumption.

Assumption 5. k'f \(q).


We now introduce two measures of steady state financial market activity, with
the first representing the level of equity market activity, and with the second
representing bank lending activity. Our measure of equity market activity, E,
306 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

represents the real value of CIP sold in secondary capital markets relative to the
size of the economy, that is

dkqu
E(k), .
f (k)
In effect, then, the function E(k) represents the ratio of the total value of trading
in secondary capital markets to GDP, for each possible value of the capital
stock consistent with technology 2 being in use. Since in steady state,

dkq"k/R , while u"f (k)(R /[tH(k)], an alternative expression for this

 
ratio is given by


kf (k) 1
E(k)" (k). (41)
f (k) R t

In addition, we introduce a measure of bank lending activity, B, which repres-
ents the real value of intermediated lending relative to the size of the economy for
each value of the capital stock consistent with technology 2 being in use. B is
given by
dk[q!w(k)]
B(k), .
f(k)
Of course d is the fraction of potential borrowers in the population, of whom
a fraction k actually receive credit, while each funded borrower receives a loan of
q!w(k). Hence per capita bank lending is given by dk[q!w(k)], and B(k)
expresses the volume of bank lending to the private sector relative to GDP. An
alternative expression for B(k) is obtained by noting that dk"k/(R
q), so that

kf (k) 1
B(k)" . (42)
R
qf(k) H(k)

Finally, we introduce a measure of the relative importance of equity market
versus banking activity,


E(k) R H(k)
EB(k)" "
q  . (43)
B(k) t
We can now state the following proposition.

Proposition 5. ¸et kR be defined by kRf (kR)/[q!w(kR)]"(1#o)/(1!o), and let

kRR be defined by kRRf (kRR)/[q!w(kRR)]"1/(1!o). ¹hen
(a) f \(q)(k††(k†
(b) E(k )'0 iff k (k†,
 
(c) B(k )'0 iff k (k††, and
 
(d) EB(k )'0.

The proof for Proposition 5 appears in Appendix C. Obviously, for economies
which produce only type j"1 capital, E"B"0.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 307

We now present some examples which illustrate the effect of an increase in the
steady state rate of inflation on the steady state levels of output, equity market
activity, bank lending activity and on the relative importance of equity markets
versus intermediated lending. Our examples set f (k)"(0.1025k #1.5),
q"3, g(z)"1/zN with zN "33.33, c"75.49, R "10, R "1, and in addition
 
d'0.99 holds. For these parameter values,
"10.5 and t"3. Moreover,
k"5.42, while f \(q)"5.12, so that case 2 obtains.

Example 1. For p"1.25, the low-capital-stock steady state is k "4.89 and the

high-capital-stock steady state is k "8.94. For the high-capital-stock steady state,

E"0.212, B"0.02525, and EB"8.4.

Example 2. For p"1.26, the low-capital-stock steady state is k "4.99 and the

high-capital-stock steady state is k "8.55. For the high-capital-stock steady state,

E"0.209, B"0.02517, and EB"8.3.

So, as the rate of money creation and the steady state rate of inflation increase
from 25% to 26%, real activity in the low-capital-stock steady state increases,
while output in the high-capital-stock steady state decreases. As real activity in
the high-capital-stock steady state decreases, so does the level of equity market
and bank lending activity. At the same time, the importance of equity market
activity relative to bank lending activity decreases as well.

6. Local dynamics

We now turn briefly to an analysis of the local stability properties of steady

state equilibria. This analysis permits us to formalize some of the discussion in
the previous section and — in particular — it allows us to illustrate how increases
in the rate of inflation can transform the high activity steady state from a saddle
to a source. We have already described the implications of this observation.
For equilibria with k4k, only the short-gestation capital production techno-
logy is utilized. It is then immediate that the behavior of our economy is
identical to the behavior of the Diamond (1965) model. As a result, given our
focus on case 2 economies, the low-activity steady state is necessarily a saddle
(Azariadis, 1993) (Chapter 26).
When k5k holds, on the other hand, the long-gestation capital production

technology is employed and matters are substantially more complicated. To see
this, notice that Eq. (33), and (36) describe the equilibrium evolution of the
sequences +k , m , under credit rationing. The pair of Eqs. (33) and (36) is
R R
obviously a system of two second order difference equations, which can be
alternatively represented as follows. Let

k "y , (44)
R> R
308 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

and
m "z . (45)
R> R
Then Eq. (36) can be written as

 
m
y "R
w(k )!y Rpf (y )!m , (46)
R>  R Rz R R
R
while Eq. (33) becomes

  
m
f  R
w(k )!y Rpf (y )!m
 R Rz R R
z "R tpm R , t50. (47)
R>  R [q!w(k )]
R
We now linearize the dynamical system consisting of Eqs. (44)—(47) in a neigh-
borhood of (any) steady state equilibrium (k, m, y, z). Then we have
(k !k, m !m, y !y, z !z)
R> R> R> R>
"J(k !k, m !m, y !y, z !z),
R R R R
where J is the Jacobian matrix
*k *k *k *k
R> R> R> R>
*k *m *y *z
R R R R
*m *m *m *m
R> R> R> R>
*k *m *y *z
J" R R R R ,
*y *y *y *y
R> R> R> R>
*k *m *y *z
R R R R
*z *z *z *z
R> R> R> R>
*k *m *y *z
R R R R
with all partial derivatives evaluated at the appropriate steady state. Expres-
sions for these derivatives are given in Appendix D.
The characteristic equation for J takes the form

   
*y *z *z *y
j R>!j R>!j ! R> R>
*y *z *y *z
R R R R

  
*y *z *z *y
!j R> R>! R> R>!j
*m *y *m *y
R R R R

   
*y *z *z *y
#j R> R>!j ! R> R>
*k *z *k *z
R R R R
*y *z *z *y
# R> R>! R> R>"0. (48)
*k *m *k *m
R R R R
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 309

We have not made any significant progress in providing a general character-

ization of the local stability properties of steady state equilibria when k5k.
However, we have produced a series of numerical examples that illustrate the
effect of an increase in the money growth rate on the stability properties of the
high-capital-stock steady state.
As in the previous section, our examples assume that f (k)"(0.1025 k #
1.5), q"3, g(z)"1/zN with zN "33.33, c"75.49, R "10, R "1, and that
 
d'0.99 holds. For these parameter values,
"10.5 and t"3. And, as before,
k"5.42, while f \(q)"5.12, so that case 2 obtains.

Example 3. For p"1.25, the parameters of the economy coincide with those for
Example 1, and the steady state equilibrium values of interest are described there.
In addition, at the high-capital-stock steady state, mod(j )"mod(j )"1.41,
 
and mod(j )"mod(j )"0.99, so the steady state is a saddle, with a two-
 
dimensional stable manifold. Paths approaching the steady state oscillate as they
do so.

Example 4. For p"1.30, the low-capital-stock steady state has k "5.36 and the

high-capital-stock steady state has k "6.34. For the high-capital-stock steady

state, E"0.19, B"0.02363, and EB"8.07. Moreover, at the high-capital-stock
steady state, j "!1.81, j "!1.14, and mod(j )"mod(j )"1.01, so the
   
steady state is a source.

Thus, for this set of examples, low rates of money growth result in a deter-
minate steady state. In particular, there exists a unique dynamical equilib-
rium path that approaches the high activity steady state. However, as the
money growth rate (and the steady state rate of inflation) increase, the economy
crosses a ‘threshold’ and the high-capital-stock steady state becomes a
source. The equilibrium behavior of the economy must change dramatically,
and if the economy approaches any steady state, that must obviously be
the low-capital-stock steady state. Then not only will real activity be low,
but so will financial market activity. Moreover, further increases in inflation —
at least in a case 2 economy — can have no incremental effects on the
volume of financial market activity. These predictions of the model are
quite consistent with a number of the empirical findings noted in the
introduction.
In addition, as Example 3 illustrates, dynamical equilibrium paths approach-
ing the high activity steady state can easily display endogenously arising volatil-
ity that dampens only very slowly. This is not possible here unless banks and
secondary capital markets are active. Thus, as argued by Keynes (1936) and
Friedman (1960) — and many others — the operation of the financial system can
readily give rise to endogenous fluctuations along perfect foresight equilibrium
paths.
310 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

7. Conclusions

As an empirical matter, there is a strong positive association between

measures of both bank lending activity and the volume of trading in equity
markets on the one hand and real activity on the other. In addition, inflation
and real activity are negatively correlated, particularly for economies with
relatively high rates of inflation. It is also true that inflation and the develop-
ment of the financial system are very negatively correlated, as are inflation and
real equity returns. Finally, there is some empirical evidence in favour of
thresholds: once the rate of inflation exceeds some critical level and stays there,
there are strong observed reductions in the level of real activity (‘inflation crises’,
in Bruno and Easterly’s terminology), and the empirical relationship between
inflation and financial market activity flattens substantially.
We have attempted here to produce a theoretical framework that can — at
least under some configurations of parameter values — account for these findings
and the other observations noted in the introduction. Any model capable of
doing so must contain — at a minimum — the following features. There must be
a role for banks, secondary capital markets, and money, and at least some
factors that increase the rate of inflation must also affect real activity, the
financial system, and the real rate of return on equity. The model must also
contain a mechanism explaining why matters change when the rate of inflation
exceeds some critical value.
We have produced a model that has all of these features. To do so, we have
started with a quite conventional neoclassical growth model (Diamond, 1965),
and introduced into it two technologies for producing capital. One is very
simple: it has a relatively short gestation period and anyone can operate it. The
other is more complex. Only certain people can run it, it must be operated on
a large scale, it involves a relatively lengthy gestation period for capital, and it
has attached to it a CSV problem. The combination of the CSV problem and the
long-gestation period of this technology implies that its use must be accom-
panied by banking and secondary capital market activity.
In this framework we have described conditions under which there are exactly
two steady state equilibria (with credit rationing): one with a relatively low and
one with a relatively high capital stock. In the high-capital stock steady state,
both banks and equity markets are active. The same thing may or may not be
true of the low-capital stock steady state. Moreover, we have shown that
inflation and real activity must be negatively correlated in the high activity
steady state. It can also easily happen that this negative relationship will become
more pronounced at relatively high levels of inflation. In addition, we have
stated conditions such that real activity and the volume of financial market
activity are positively correlated in the high activity steady state. When this is
the case, obviously inflation and financial market activity will be inversely
related as well.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 311

Finally, we have illustrated that the high-capital-stock steady state may be a

saddle for low rates of money growth. However, once the rate of money creation
(inflation) exceeds some critical level, the high activity steady state can be trans-
formed from a saddle to a source. Thus, thresholds can easily exist: the behavior
of the economy must differ dramatically depending on whether the steady state
rate of inflation is above or below this threshold level. This implication of the model
is again consistent with several pieces of empirical evidence described above.
In addition to explaining several established observations, our model yields
some new testable implications of its own. For instance, in a case 1 economy, the
model predicts that the correlations among inflation, financial market condi-
tions, and real activity will differ strongly across three distinct situations: (i)
k(k, (ii) k3(k, f \(q)), and (iii) k'f \(q). In a case 2 economy, these correla-
tions will change
 as k(k or k'k hold. Of course these implications of the
analysis might be difficult  to test in  practice, as the critical values k and f \(q)
might be hard to identify empirically. 
Admittedly, in order to obtain all of these results we have had to make some
strong assumptions. A particularly strong assumption has been placed on the
production technology: we have assumed that capital and labor are highly
substitutable in production (o'0). It would be interesting to derive modifications
of the analysis that would allow us to relax this sort of condition in the future.

Appendix A. Proof of Lemma 1

Part (a) of Lemma 1 is immediate from lim f (k)"R and assumption 1.

I
Part (b) is also obvious. For (c), it is easy to verify that
[ f (k)!q]
H(k)"!f (k) ,
[q!w(k)]
establishing the result. Part (d) follows from
kf (k) kf (k)
kH(k), " .
[q!w(k)] [q!f (k)#kf (k)]

Appendix B. Proof of Proposition 2

Type j"2 capital will be produced in steady state iff the internal rate of
return on investments in technology 2 exceeds that on investments in technology
1. From Eqs. (28) and (37), this condition obtains iff


f (k)
R t 'R f (k). (B.1)
 [q!w(k)] 
Rearranging terms in Eq. (B.1) establishes the result.
312 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

Appendix C. Proof of Proposition 5

(a) Given our assumption that 0(o(1, it follows that 1((1/(1!o))

(((1#o)/(1!o)), which implies 1((kRRf (kRR)/[q!w(kRR)])((kRf (kR)/
[q!w(kR)]). From Lemma 1 it is then obvious that f\(q)(kRR(kR, which
establishes part (a).
(b) Differentiating Eq. (41) and rearranging terms yields kE(k)/E(k)
"w(k)/f (k)!(1/2)kw(k)+[q!w(k)#kf (k)]/[kf (k)[q!w(k)]],. Therefore,
E(k)'0 holds iff kf (k)[q!w(k)]/f (k)'(1/2)[kw(k)/w(k)][q!w(k)#kf (k)].
Since f (k) has the CES form f (k)"[akM#b]M, with 0(o(1, it is straight-
forward to show that E(k)'0 iff kf (k)/[q!w(k)]((1#o)/(1!o)"
kRf (kR)/[q!w(kR)]. Lemma 1 then implies that kR'f\(q) is satisfied. There-
fore, since k 'f\(q) necessarily holds it follows that E(k )'0 obtains iff
 
k (kR.

(c) Differentiating Eq. (42) and rearranging terms, we obtain
kB(k)/B(k)"[w(k)/f (k)]!kw(k)/[q!w(k)]. Therefore, B(k)'0 obtains iff
[q!w(k)]/f (k)'kw(k)/w(k). Since f(k)"[akM#b]M, with 0(o(1, it is easy
to show that B(k)'0 holds iff kf (k)/[q!w(k)](1/(1!o)"kRRf (kRR)/
[q!w(kRR)]. Lemma 1 then clearly implies that kRR'f\(q) holds. In addition,
since k 'f\(q), it follows that B(k )'0 is satisfied iff k (kRR.
  
(d)H(k )'0 necessarily holds and the result is then immediate from Eq. (43).


Appendix D. Elements of the Jacobian

The elements of the Jacobian matrix are given by the following expressions:
*k
R>"0, (D.1)
*k
R
*k
R>"0, (D.2)
*m
R
*k
R>"1, (D.3)
*y
R
*k
R>"0, (D.4)
*z
R
*m
R>"0, (D.5)
*k
R
*m
R>"0, (D.6)
*m
R
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 313

*m
R>"0, (D.7)
*y
R
*m
R>"1, (D.8)
*z
R
*y
R>"R
w(k), (D.9)
*k 
R

 
*y kf (k)
R>"!R
1#p , (D.10)
*m  m
R
*y
R>"R
p[w(k)!f (k)], (D.11)
*y 
R
*y kf (k)
R>"R
p , (D.12)
*z  m
R

 
*z w(k) 1
R>"m #R
f (k) , (D.13)
*k f (k) R tp 
R 

 
*z w(k) kf (k)
R>"1#R
m 1#p , (D.14)
*m  kf (k) m
R

 
*z w(k)
R>"!R
pmf (k) 1! , (D.15)
*y  f (k)
R
*z
R>"!R
pw(k). (D.16)
*z 
R

References
Atje, R., Jovanovic, B., 1993. Stock markets and development. European Economic Review 37,
632—640.
Azariadis, C., 1993. Intertemporal Macroeconomics. Blackwell, Cambridge, Massachusetts.
Barro, R.J., 1995. Inflation and economic growth. Bank of England Quarterly Bulletin 35, 166—176.
Bencivenga, V., Smith, B.D., Starr, R., 1995. Transactions costs, technological choice, and endogen-
ous growth. Journal of Economic Theory 67, 153—177.
Bencivenga, V., Smith, B.D., Starr, R., 1996. Liquidity of secondary capital markets, allocative
efficiency, and the maturity composition of the capital stock. Economic Theory 7, 19—50.
Bernanke, B.S., Gertler, M., 1989. Agency costs, net worth, and business fluctuations. American
Economic Review 79, 14—31.
Bohm-Bawerk, E., 1891. A Positive Theory of Capital. MacMillan, London.
Boyd, J.H., Levine, R., Smith, B.D., 1996. Inflation and financial market performance. Manuscript,
Federal Reserve Bank of Minneapolis.
Boyd, J.H., Smith, B.D., 1994. How good are standard debt contracts?: Stochastic versus non-
stochastic monitoring in a costly state verification environment. Journal of Business 67, 539—561.
314 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315

Boyd, J.H., Smith, B.D., 1998. Capital market imperfections in a monetary growth model. Economic
Theory 11, 241—273.
Bruno, M., Easterly, W., 1998. Inflation crises and long-run growth. Journal of Monetary Econ-
omics, forthcoming.
Bullard, J., Keating, J., 1995. The long-run relationship between inflation and output in post-war
economies. Journal of Monetary Economics 36, 477—496.
Champ, B., Smith, B.D., Williamson, S.D., 1996. Currency elasticity and banking panics: Theory and
evidence. Canadian Journal of Economics 29, 828—864.
Clark, T.E., 1997. Cross-country evidence on long-run growth and inflation. Economic Inquiry 35,
70—81.
Diamond, D., 1984. Financial intermediation and delegated monitoring. Review of Economic
Studies 51, 393—414.
Diamond, P.A., 1965. National debt in a neoclassical growth model. American Economic Review 55,
1126—1150.
Fama, E., Schwert, G., 1977. Asset returns and inflation. Journal of Financial Economics 5,
115—146.
Friedman, M., 1960. A Program for Monetary Stability. Fordham University Press, New York.
Gale, D., Hellwig, M., 1985. Incentive-compatible debt contracts: The one-period problem. Review
of Economic Studies 52, 647—663.
Gurley, J.G., Shaw, E., 1960. Money in a Theory of Finance. The Brookings Institution, Washing-
ton, DC.
Gultekin, N., 1983. Stock market returns and inflation: Evidence from other countries. Journal of
Finance 38, 49-65.
Hamid, J., Singh, A., 1992. Corporate financial structures in developing countries. International
Finance Corporation, Technical Paper 1.
Hicks, J., 1969. A Theory of Economic History. Oxford University Press, Oxford.
Huybens, E., Smith, B.D., 1998. Financial market frictions, monetary policy and capital accumula-
tion in a small open economy. Journal of Economic Theory 81, 353—400.
Japelli, T., 1990. Who is credit-constrained in the US economy?. Quarterly Journal of Economics
105, 219—234.
Keynes, J.M., 1936. The General Theory of Employment, Interest, and Money. MacMillan, London.
Krasa, S., Villamil, A.P., 1992. Monitoring the monitor: an incentive structure for a financial
intermediary. Journal of Economic Theory 57, 197—221.
King, R.G., Levine, R., 1993a. Finance and growth: Schumpeter might be right. Quarterly Journal of
Economics 108, 717-738.
King, R.G., Levine, R., 1993b. Finance, entrepreneurship and growth: Theory and evidence. Journal
of Monetary Economics 32, 513-542.
Kydland, F., Prescott, E., 1982. Time to build and aggregate fluctuations. Econometrica 50,
1345—1370.
Levine, R., 1997. Financial development and economic growth: Views and agenda. Journal of
Economic Literature 35, 688—726.
Levine, R., Renelt, D., 1992. A sensitivity analysis of cross-country growth regessions. American
Economic Review, 82, 942—963.
Levine, R., Zervos, S., 1998 Stock markets, banks and economic growth. American Economic
Review, forthcoming.
McKinnon, R.I., 1973. Money and Capital in Economic Development. The Brookings Institution,
Washington, DC.
Nelson, C., 1976. Inflation and rates of return on common stocks. Journal of Finance 31, 471-483.
Schreft, S.L., Smith, B.D., 1997. Money, banking, and capital formation. Journal of Economic
Theory 73, 157—182.
Tirole, J., 1985. Asset bubbles and overlapping generations. Econometrica 53, 1071—1100.
 E. Huybens, B.D. Smith / Journal of Monetary Economics 43 (1999) 283–315 315

Townsend, R.M., 1979. Optimal contracts and competitive markets with costly state verification.
Journal of Economic Theory 21, 265—293.
Williamson, S.D., 1986. Costly monitoring, financial intermediation, and equilibrium credit
rationing. Journal of Monetary Economics 18, 159—179.
Williamson, S.D., 1987. Costly monitoring, loan contracts, and equilibrium credit rationing. Quar-
terly Journal of Economics 102, 135—145.