%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%                                                                     %
%              Scientific Word   Wrap/Unwrap  Version 2.5             %
%              Scientific Word   Wrap/Unwrap  Version 3.0             %
%                                                                     %
% If you are separating the files in this message by hand, you will   %
% need to identify the file type and place it in the appropriate      %
% directory.  The possible types are: Document, DocAssoc, Other,      %
% Macro, Style, Graphic, PastedPict, and PlotPict. Extract files      %
% tagged as Document, DocAssoc, or Other into your TeX source file    %
% directory.  Macro files go into your TeX macros directory. Style    %
% files are used by Scientific Word and do not need to be extracted.  %
% Graphic, PastedPict, and PlotPict files should be placed in a       %
% graphics directory.                                                 %
%                                                                     %
% Graphic files need to be converted from the text format (this is    %
% done for e-mail compatability) to the original 8-bit binary format. %
%                                                                     %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%                                                                     %
% Files included:                                                     %
%                                                                     %
% "/document/Sudden_stops.tex", Document, 30145, 1/6/2005, 21:27:41, ""%
%                                                                     %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%% Start /document/Sudden_stops.tex %%%%%%%%%%%%%%%%%%


\documentclass[leqno,fleqn]{article}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\usepackage{srav5}

%TCIDATA{TCIstyle=frb/frb.lat,srav5,sr}

%TCIDATA{OutputFilter=LATEX.DLL}
%TCIDATA{Version=4.10.0.2363}
%TCIDATA{Created=Monday, December 13, 2004 11:46:20}
%TCIDATA{LastRevised=Thursday, January 06, 2005 15:27:41}
%TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
%TCIDATA{<META NAME="DocumentShell" CONTENT="FRB\Staff Report Style A (v5) [endnotes]">}
%TCIDATA{Language=American English}
%TCIDATA{CSTFile=sr.cst}

\input{tcilatex}
\input{lefttag.sty}

\begin{document}


%TCIMACRO{\TeXButton{TeX field}{\thispagestyle{empty}}}%
%BeginExpansion
\thispagestyle{empty}%
%EndExpansion
%TCIMACRO{%
%\TeXButton{TeX field}{\setstretch{1.0}
%\normalsize}}%
%BeginExpansion
\setstretch{1.0}
\normalsize%
%EndExpansion

\noindent \textbf{V. V. Chari}

\noindent 612-204-5518

\noindent chari@res.mpls.frb.fed.us

\bigskip

\noindent \textbf{Patrick J. Kehoe}

\noindent 612-204-5525

\noindent pkehoe@res.mpls.frb.fed.us

\bigskip

\noindent \textbf{Ellen R. McGrattan}

\noindent 612-204-5523

\noindent erm@mcgrattan.mpls.frb.fed.us

%TCIMACRO{\TeXButton{TeX field}{\vspace{.5in}}}%
%BeginExpansion
\vspace{.5in}%
%EndExpansion

\noindent \emph{All authors:}

\bigskip

\noindent Research Department

\noindent Federal Reserve Bank of Minneapolis

\noindent 90 Hennepin Avenue

\noindent Minneapolis, MN 55401-1804

\bigskip

\noindent Fax: 612-204-5515

%TCIMACRO{\TeXButton{TeX field}{\vspace{.5in}}}%
%BeginExpansion
\vspace{.5in}%
%EndExpansion

\noindent \textit{\textbf{Title of Session at Which Paper Presented:}}

\bigskip

\noindent New Perspectives on Reputation and Debt

%TCIMACRO{\TeXButton{TeX field}{\vspace{.75in}}}%
%BeginExpansion
\vspace{.75in}%
%EndExpansion

\noindent \textit{\textbf{Name and Version of Word-Processing Program Used
to Prepare Document:}}

\bigskip

\noindent Scientific Word 5.0 for text

\noindent Microsoft Word (Windows XP) for figure text

\noindent Adobe Acrobat (Figures)

\bigskip

\noindent Hard copies of figures, suitable for reproduction, coming by
overnight mail.

\newpage

%TCIMACRO{\TeXButton{TeX field}{\setcounter{page}{1}}}%
%BeginExpansion
\setcounter{page}{1}%
%EndExpansion
%TCIMACRO{%
%\TeXButton{TeX field}{\setstretch{2.0}
%\normalsize}}%
%BeginExpansion
\setstretch{2.0}
\normalsize%
%EndExpansion

\begin{center}
{\large \textbf{Sudden Stops and Output Drops}}

\emph{By} \textsc{V. V. Chari, Patrick J. Kehoe, and Ellen R. McGrattan}*

\bigskip
\end{center}

Recent financial crises in emerging markets have included two features:
abrupt declines in capital inflows, commonly known as \emph{sudden stops}
(Guillermo Calvo, 1998), and large declines in output. Here we ask whether
theory predicts that these two features are related; do sudden stops
necessarily lead to output drops? We ask this in a standard equilibrium
model in which sudden stops are generated by an abrupt tightening of a
country's collateral on foreign borrowing. Theory's answer is no; sudden
stops, by themselves, do not lead to decreases in output, but rather to
increases. To generate an output drop during a financial crisis, the model
must include other frictions which have negative effects on output that are
large enough to overwhelm the positive effect of the sudden stop.

We begin by setting up a standard model of a small open economy in which
foreign borrowing is subject to a collateral constraint. We view
fluctuations in this collateral constraint as arising from fluctuations in a
country's reputation. In the model, the country's budget constraint implies
that an abrupt decrease in capital inflows produces an abrupt increase in
net exports. Following our earlier approach (in V. V. Chari, Patrick Kehoe,
and Ellen McGrattan, 2004), we show that the equilibrium outcomes in the
small open economy are equivalent to those of a closed-economy prototype
growth model of the kind widely used in the business cycle literature. In
particular, we show that a rise in net exports in the small open economy
corresponds to a rise in government consumption in the prototype model. It
is well known that an increase in government consumption produces an
increase in output in models like our prototype growth model. A sudden stop
that produces an increase in net exports in the small open economy thus also
leads to an increase in output. We demonstrate this quantitatively with data
from Mexico in the mid-1990s.

In three other studies, researchers have built small open-economy business
cycle models in which sudden stops lead to output drops. In these studies,
however, output drops because of other frictions that overwhelm the direct
effect on output from sudden stops. In all three studies, firms must borrow
in advance to pay for inputs to production. In Pablo Neumeyer and Fabrizio
Perri (2004), firms must borrow to pay for a fraction of the wage bill,
while in Lawrence Christiano, Christopher Gust, and Jorge Roldos (2004) and
Enrique Mendoza (2004), firms must borrow to pay for foreign intermediate
inputs. This payment-in-advance requirement, by itself, does not introduce a
friction because firms can simply borrow at the market interest rate to make
the payments. In Neumeyer and Perri (2004) and Mendoza (2004), the key
friction is that firms are effectively required to put the funds in a
non--interest-bearing escrow account. In Neumeyer and Perri, this
requirement introduces a wedge between the marginal product of labor and the
marginal rate of substitution between leisure and consumption. In Mendoza,
this requirement produces a shock to total factor productivity. In
Christiano, Gust, and Roldos, the payment-in-advance requirement interacts
with the collateral constraint to produce a shock to total factor
productivity. Here we demonstrate in a version of our model that the output
drops in these studies are not due to the sudden stops alone.

\bigskip

\noindent {\large \textbf{I. Collateral Constraints}}

Here we develop a model with a collateral constraint facing a small open
economy. We show that equilibrium allocations in this original model
coincide with those in a closed-economy prototype growth model with shocks
to government consumption. Specifically, a decline in net exports resulting
from a tightening of the collateral constraint corresponds to a rise in
government consumption in the prototype growth model.

\bigskip

\noindent \textbf{A. Original Economy}

Consider the following model of a small open economy embedded in a world
economy with a single homogenous good in each period. The economy
experiences one of finitely many events $s_{t}$, which index the shocks. We
denote by $s^{t}=(s_{0},\ldots ,s_{t})$ the history of events up through and
including period $t$. The probability, as of period 0, of any particular
history $s^{t}$ is $\pi (s^{t})$. The initial realization $s_{0}$ is given.

The representative consumer in this economy has preferences 
\begin{equation}
\sum \beta ^{t}\pi (s^{t})U(c(s^{t}),l(s^{t})),  \label{preferences}
\end{equation}%
where $c(s^{t})$ and $l(s^{t})$ denote consumption and labor. The country's
budget constraint is 
\begin{equation}
c(s^{t})+b(s^{t})+k(s^{t})\leq F(k(s^{t-1}),l(s^{t}))+(1-\delta
)k(s^{t-1})+\sum_{s^{t+1}|s^{t}}q(s^{t+1})b(s^{t+1}),  \label{budget}
\end{equation}%
where $b(s^{t+1})$ denotes the amount of state-contingent debt, or borrowing
from the rest of the world by the country in period $t$, $q(s^{t+1})$
denotes the corresponding state-contingent price, and $k(s^{t})$ denotes the
capital stock chosen in period $t$ for use in period $t+1$. The country
faces a collateral constraint on borrowing of%
\begin{equation}
b(s^{t+1})\leq V\left( s_{t+1}\right) ,  \label{collateral}
\end{equation}%
where the maximal amount of borrowing $V\left( s_{t+1}\right) $ depends on
the shock in period $t+1.$ To avoid Ponzi schemes, we assume that $V\left(
s_{t+1}\right) $ is uniformly bounded above. The constraint (\ref{collateral}%
) implies that having more collateral allows the country to borrow more. We
interpret shocks to this collateral constraint as arising from changes in
the relationship between the country and international financial markets (or
the country's financial reputation).

The government of this country maximizes the utility of the representative
consumer (\ref{preferences}) subject to the country's budget constraint (\ref%
{budget}) and collateral constraint (\ref{collateral}). The world gross
interest rate is constant and equal to $R$ across both states and time.
Arbitrage requires that, in equilibrium, 
\begin{equation}
q(s^{t+1})=\frac{\pi (s^{t+1}|s^{t})}{R}.  \label{q}
\end{equation}%
An \textit{equilibrium for the original economy} is a set of allocations $%
\left( c(s^{t}),k(s^{t}),l(s^{t}),b(s^{t+1})\right) $ and prices $\left(
q(s^{t})\right) $ such that these allocations solve the government's problem
and the prices satisfy the arbitrage condition (\ref{q}).

In this economy, a \textit{sudden stop} is defined as an abrupt increase in
net exports. Here, net exports are clearly $%
F(k(s^{t-1}),l(s^{t}))-[k(s^{t})-(1-\delta )k(s^{t-1})]-c(s^{t}),$ which
from (\ref{budget}) equals%
\begin{equation}
b(s^{t})-\sum_{s^{t+1}|s^{t}}q(s^{t+1})b(s^{t+1}).  \label{borrowing}
\end{equation}%
Thus, a sudden stop is equivalently defined as an abrupt decrease in new
borrowing. From (\ref{collateral}), we know that a sudden drop in $%
V(s_{t+1}) $ leads to a drop in $b(s^{t+1})$ and, when the collateral
constraint is binding, to a sudden stop.

To understand how the collateral constraint affects the equilibrium,
consider the first-order conditions of the government's problem. Let $\beta
^{t+1}\pi (s^{t+1})\mu (s^{t+1})$ be the multiplier on the collateral
constraint, so that $\mu (s^{t+1})$ is positive when the collateral
constraint is binding and zero when it is not. Then the first-order
conditions imply that%
\begin{equation}
-\frac{U_{l}(s^{t})}{U_{c}(s^{t})}=F_{l}(s^{t}),  \label{labor}
\end{equation}%
\begin{equation}
U_{c}(s^{t})=\sum_{s^{t+1}|s^{t}}\beta \pi (s^{t+1}|s^{t})U_{c}(s^{t+1}) 
\left[ F_{k}(s^{t+1})+1-\delta \right] ,\text{ }  \label{Euler}
\end{equation}%
and 
\begin{equation}
U_{c}(s^{t})=\beta R[U_{c}(s^{t+1})+\mu (s^{t+1})].  \label{mu}
\end{equation}%
Notice that the collateral constraint does not distort either the
first-order condition (\ref{labor}), governing labor supply, or the
intertemporal Euler equation (\ref{Euler}), governing capital accumulation.
As can be seen from (\ref{mu}), the collateral constraint affects only the
intertemporal marginal rate of substitution in consumption.

\bigskip

\noindent \textbf{B. Associated Prototype Economy}

Now consider a closed-economy prototype model with an exogenous stochastic
variable, government consumption\textit{\ }$g(s^{t}),$ which we call the 
\textit{government consumption wedge}. In this economy, consumers maximize (%
\ref{preferences}) subject to the budget constraint%
\begin{equation}
c(s^{t})+k(s^{t})\leq w(s^{t})l(s^{t})+\left[ r(s^{t})+1-\delta \right]
k(s^{t-1})+T(s^{t}),  \label{prototype budget}
\end{equation}%
where $w(s^{t}),$ $r(s^{t}),$ and $T(s^{t})$ are the wage rate, the capital
rental rate, and the lump-sum transfers. In each state $s^{t},$ firms choose 
$k$ and $l$ to maximize $F(k,l)-r(s^{t})k-w(s^{t})l$. The government's
budget constraint is 
\begin{equation}
g(s^{t})+T(s^{t})=0.  \label{government budget}
\end{equation}%
The resource constraint for this economy is 
\begin{equation}
c(s^{t})+g(s^{t})+k(s^{t})=F\left( k(s^{t-1}),l(s^{t})\right) +(1-\delta
)k(s^{t-1}).  \label{resource}
\end{equation}%
An \textit{equilibrium of the prototype economy} consists of allocations $%
\left( c(s^{t}),k(s^{t}),l(s^{t}),g(s^{t}),T(s^{t})\right) $ and prices $%
\left( w(s^{t}),r(s^{t})\right) $ such that these allocations are optimal
for consumers and firms and the resource constraint is satisfied.

The following proposition shows that the government consumption wedge in the
prototype economy consists of net exports in the original economy:

\bigskip

\noindent \textit{Proposition} 1.\textit{\ }Consider an equilibrium of $%
\left( c(s^{t}),k(s^{t}),l(s^{t}),b(s^{t+1})\right) $ and $\left(
q(s^{t})\right) $ for the original economy. Let the government consumption
wedge be%
\begin{equation}
g(s^{t})=F\left( k(s^{t-1}),l(s^{t})\right) -\left[ k(s^{t})-(1-\delta
)k(s^{t-1})\right] -c(s^{t}),  \label{g}
\end{equation}%
let the wage and capital rental rates be $w(s^{t})=F_{l}(s^{t})$ and $%
r(s^{t})=F_{k}(s^{t}),$ and let the lump-sum transfers $T(s^{t})$ be defined
by (\ref{government budget}). Then the allocations $\left(
c(s^{t}),k(s^{t}),l(s^{t}),g(s^{t}),T(s^{t})\right) $ and the prices $\left(
w(s^{t}),r(s^{t})\right) $ are an equilibrium for the prototype economy.
\bigskip

\textit{Proof}. The first-order conditions for the prototype economy are (%
\ref{labor}) and (\ref{Euler}). Under the construction of the government
consumption wedge in (\ref{g}), the resource constraint (\ref{resource}) is
equal to (\ref{budget}). The proposition then follows. $Q.E.D.$ \bigskip

Now consider a sudden stop, as in the original economy. In the prototype
economy, this sudden stop manifests itself as an abrupt increase in the
government consumption wedge. As is well known from the business cycle
literature, an increase in government consumption by itself leads to an
increase in labor and an increase in output. (See, for example, S. Rao
Aiyagari, Lawrence Christiano, and Martin Eichenbaum, 1992.) Thus, in this
economy, a sudden stop does not generate an output drop; it generates an
output rise.

Note that for simplicity we have abstracted from any government consumption
in the original economy. If we let the original economy have government
consumption$,$ then the government consumption wedge in the prototype
economy is the sum of government consumption and net exports in the original
economy.

Note also that given an equilibrium in the prototype economy, we can
construct the associated equilibrium in the original economy if $%
U_{c}(s^{t})/\beta R-U_{c}(s^{t+1})$ is nonnegative for all $s^{t+1}$ for
some choice of $R>1.$ This $R$ serves as the world interest rate in the
original economy, and at such an $R$, the multiplier on the collateral
constraint is nonnegative. In this constructed equilibrium, the value of the
initial debt in the original economy is set equal to the present discounted
value of government consumption in the prototype economy. The debt in the
original economy at state $s^{t}$ is, of course, the present discounted
value of net exports and, hence, in the prototype economy corresponds to the
present discounted value of future government consumption.

Proposition 1 is closely related to a proposition in our earlier work
(Chari, Kehoe, and McGrattan, 2004). In that proposition, we consider an
original economy with no collateral constraint but with a fluctuating world
interest rate. We establish a similar equivalence proposition there. In the
original economy, fluctuations in the world interest rate lead to
fluctuations in net exports. In the prototype economy, these fluctuations in
net exports show up as fluctuations in the government consumption wedge.

\bigskip

\noindent \textbf{C. Extensions to Uncontingent Asset Markets}

Consider now a version of the original economy in which we replace the
state-contingent debt with uncontingent debt. Here the budget constraint
becomes%
\begin{equation}
c(s^{t})+b(s^{t-1})+k(s^{t})\leq F(k(s^{t-1}),l(s^{t}))+(1-\delta
)k(s^{t-1})+q(s^{t})b(s^{t}),
\end{equation}%
where now $b(s^{t})$ denotes the amount of state-uncontingent debt owed to
the rest of the world by the country in period $t$ and $q(s^{t})$ denotes
the corresponding price. The collateral constraint then becomes%
\begin{equation}
b(s^{t})\leq V\left( s_{t+1}\right) \text{ for all }s_{t+1}.
\end{equation}%
Here $q(s^{t})=1/R.$ The first-order conditions associated with the problem
are (\ref{labor}), (\ref{Euler}), and 
\[
U_{c}(s^{t})=\beta R\sum_{s^{t+1}|s^{t}}\pi (s^{t+1}|s^{t})\left[
U_{c}(s^{t+1})+\mu (s^{t+1})\right] . 
\]%
With this setup, the analog of Proposition 1 immediately applies. In
particular, in the prototype economy, fluctuations in the government
consumption wedge play the same role as fluctuations in net exports in the
original economy.

\bigskip

\noindent {\large \textbf{II. A Quantitative Analysis of a Sudden Stop}}

Here we use the logic of Proposition 1 to examine the quantitative effects
of a well-known sudden stop episode, in Mexico in the mid-1990s.

In Figure 1, we plot Mexican data on real net exports and the government
consumption wedge (the sum of real government consumption and real net
exports) between the last quarters of 1994 and 1996. We normalize both
series by the level of real GDP in 1994:4. The figure shows an abrupt and
dramatic increase in the government consumption wedge in the first quarter
of 1995 and suggests that this increase was due almost entirely to the sharp
rise in net exports. That is, Mexico experienced a sudden stop.

We study the quantitative effects of this sudden stop in a prototype growth
model. We use a version of the model in Chari, Kehoe, and McGrattan (2004),
in which the stochastic processes for the government consumption wedge, the
labor wedge, the efficiency wedge, and the investment wedge are estimated
using the model and Mexican data from 1980:1 to 2003:4. (For details, see
Chari, Kehoe, and McGrattan, 2005.)

To assess the effects of the sudden stop, we feed the realized values of the
Mexican government consumption wedge over the period from 1994:4 to 1996:4
into our prototype growth model, holding the values of the other Mexican
wedges fixed at their 1994:4 levels. Figure 2 shows the result: Given that
historical Mexican data, the model predicts that the sudden stop, by itself,
leads to a small increase in output. Figure 2 also shows what actually
happened to Mexican real GDP in 1995: it fell sharply. This analysis clearly
demonstrates that the sharp output drop is not due to the sudden stop alone.

\bigskip

\noindent {\large \textbf{III. Other Frictions}}

We have shown that in our simple model with a collateral constraint on
foreign borrowing, sudden stops do not lead to output drops. Now we consider
adding other frictions to the simple model and ask if sudden stops,
interacting with these frictions, lead to output drops. We first consider an
economy with an endogenous collateral constraint and show that this
constraint corresponds to a subsidy to investment in the prototype growth
model and, therefore, to an increase in output. We then consider the role of
an advance-payment constraint. We show that such a constraint can lead to
output drops from sudden stops only when coupled with yet other frictions.

\bigskip

\noindent \textbf{A. Investment Wedges from Endogenous Collateral Constraints%
}

Consider a version of the original economy with one change, that the
collateral constraint (\ref{collateral}) is replaced by an endogenous
collateral constraint:

\begin{equation}
b(s^{t+1})\leq V\left( k(s^{t}),s_{t+1}\right) .  \label{collateral 2}
\end{equation}%
Here the maximal amount that can be borrowed, $V\left(
k(s^{t}),s_{t+1}\right) ,$ depends on the capital stock chosen in period $t$
and the shock in period $t+1.$ We assume that $V(k(s^{t}),s_{t+1})$ is
strictly increasing in $k,$ so that having more collateral allows\ a country
to borrow more, and that $V\left( k(s^{t}),s_{t+1}\right) $ is uniformly
bounded above, to avoid Ponzi schemes. This formulation is similar to that
of Nobuhiro Kiyotaki and John Moore (1997) and is motivated by the idea that
a portion of the capital stock can effectively be seized by foreign lenders
in the event of default; hence, foreign lenders will not lend more than the
value of the seizable portion.

An equilibrium in this \textit{endogenous collateral-constraint economy} is
defined as before. The first-order conditions here are the same as in the
original economy except that the intertemporal Euler equation (\ref{Euler})
is replaced by 
\begin{equation}
U_{c}(s^{t})=\sum_{s^{t+1}|s^{t}}\beta \pi (s^{t+1}|s^{t})U_{c}(s^{t+1}) 
\left[ (F_{k}(s^{t+1})-\delta )+1\right] +\sum_{s^{t+1}|s^{t}}\beta \pi
(s^{t+1}|s^{t})\mu (s^{t+1})V_{k}(s^{t+1}).  \label{Euler 2}
\end{equation}

The associated prototype economy has an \textit{investment wedge} $1-\tau
_{k}(s^{t})$, resembling a tax on capital income, along with the government
consumption wedge. The consumers' budget constraint is now 
\begin{equation}
c(s^{t})+k(s^{t})\leq w(s^{t})l(s^{t})+\left[ \left( 1-\tau
_{k}(s^{t})\right) \left( r(s^{t})-\delta \right) +1\right]
k(s^{t-1})+T(s^{t}).  \label{budget 2}
\end{equation}%
The government's budget constraint is now 
\begin{equation}
g(s^{t})+T(s^{t})=\tau _{k}(s^{t})\left( r(s^{t})-\delta \right) k(s^{t-1}),
\label{government budget 2}
\end{equation}%
and the resource constraint is unaffected. An equilibrium of the prototype
economy is defined as before.

The following proposition shows that the endogenous collateral constraint
manifests itself as an investment wedge in the associated prototype economy:
\bigskip\ 

\noindent \textit{Proposition} 2.\textit{\ }Consider an equilibrium for the
endogenous collateral-constraint economy. Let the investment wedge be 
\begin{equation}
\tau _{k}(s^{t+1})=-\frac{\mu (s^{t+1})V_{k}(s^{t+1})}{U_{c}(s^{t+1})\left(
F_{k}(s^{t+1})-\delta \right) },  \label{tau}
\end{equation}%
let $T(s^{t})$ be defined by (\ref{government budget 2}), and let the rest
of the variables be as in Proposition 1. Then the resulting allocations,
prices, and wedges are an equilibrium for the associated prototype economy.
\bigskip

The proof of the proposition follows immediately from a comparison of the
first-order conditions, budget constraints, and resource constraints of the
two economies.

Note that if the endogenous collateral-constraint economy is in the
neighborhood of a steady state, then $F_{k}(s^{t+1})-\delta $ is positive,
so that in this neighborhood, a binding collateral constraint ($\mu
(s^{t+1})>0)$ corresponds to a subsidy to capital accumulation ($\tau
_{k}(s^{t+1})<0).$ The intuition for this result is that if the collateral
constraint in the original economy is binding in period $t+1,$ then capital
accumulation in period $t$ helps to relax the collateral constraint in
period $t+1$ and, hence, provides an additional benefit beyond that from the
marginal product of capital.

Consider a sudden stop in the endogenous collateral-constraint economy
generated by a tightening of the collateral constraint. Proposition 2 shows
that such a sudden stop corresponds to a subsidy to capital accumulation
along with an increase in government consumption. A well-known result in the
business cycle literature is that a subsidy to capital accumulation
stimulates investment and output. We have already argued that an increase in
government consumption also stimulates output. Thus, in the endogenous
collateral-constraint economy, sudden stops do not lead to output drops.
\bigskip

\noindent \textbf{B. Labor and Efficiency Wedges from an Advance-Payment
Constraint}

The literature on sudden stops has introduced various forms of constraints
involving payments in advance and has shown that sudden stops interacting
with these constraints and other frictions can produce output drops.
Neumeyer and Perri (2004) introduce advance-payment constraints on wages,
together with an escrow provision, and Christiano, Gust, and Roldos (2004)
and Mendoza (2004) introduce them on payments of intermediate goods.
Christiano, Gust, and Roldos subject these payments to collateral
constraints. Mendoza has both an escrow provision and a collateral
constraint. Here we introduce an advance-payment constraint on wages and
show that alone it does not produce simultaneous sudden stops and output
drops.

In the business cycle literature, it has been widely argued that requiring
firms to pay workers in advance of production introduces a wedge between the
marginal product of labor and the marginal rate of substitution between
leisure and consumption; we have called this a \textit{labor wedge} (Chari,
Kehoe, and McGrattan, 2004)\textit{. }Here we argue that such a requirement,
by itself, does not introduce a labor wedge, but it does so when coupled
with a requirement that firms escrow future wages in non--interest-bearing
accounts.

Consider the following deterministic version of our economy with an
advance-payment constraint together with an escrow provision. In this
economy, in period $t-1,$ a firm must escrow with the government its wage
bill $w_{t}l_{t}$ due in period $t.$ To do so, the firm borrows $w_{t}l_{t}$
at $t-1$ from foreign lenders at the world (gross) interest rate $R_{t}$ and
escrows these funds with the government. (Note that the firm's cash flows in
period $t-1$ with respect to these transactions net out to zero.) In period $%
t,$ the firm uses the escrowed funds to pay its workers, and it must repay
the foreign lenders $R_{t}w_{t}l_{t}.$ The firm's problem is to maximize
profits at $t$ given by 
\[
F(k_{t},l_{t})-R_{t}w_{t}l_{t}-r_{t}k_{t}. 
\]%
The government rebates the interest on escrowed funds as a lump-sum transfer
to consumers. Consumers maximize utility $\sum \beta ^{t}U(c_{t},l_{t})$
subject to the budget constraint 
\[
c_{t}+R_{t}b_{t}+k_{t}\leq w_{t}l_{t}+[r_{t}+(1-\delta
)]k_{t}+b_{t+1}+T_{t}. 
\]%
The resulting first-order conditions imply that 
\[
-\frac{U_{lt}}{U_{ct}}=\frac{F_{lt}}{R_{t}}. 
\]%
This economy is equivalent to a prototype economy with tax rate on labor $%
\tau _{lt}$ equal to $1-(1/R_{t}).$ The advance-payment constraint with an
escrow provision induces a labor wedge of $1-\tau _{lt}=1/R_{t}.$

To see that an advance-payment constraint, by itself, does not introduce a
labor wedge, let the government pay interest on the escrow accounts at rate $%
R_{t}.$ Here the firms borrow $w_{t}l_{t}/R_{t}$ at $t-1,$ their problem is
to maximize $F(k_{t},l_{t})-w_{t}l_{t}-r_{t}k_{t},$ and there is no labor
wedge.

It is easy to show that an advance-payment constraint on intermediate goods,
by itself, does not distort decisions. When coupled with other frictions,
such as subjecting advance payments to collateral constraints, these
constraints on intermediate goods manifest themselves as efficiency wedges
that resemble shocks to total factor productivity.

\bigskip

\noindent {\large \textbf{IV. Conclusion}}

We have shown theoretically and empirically that in standard equilibrium
models, sudden stops of capital inflows lead to increases, not drops, in
output. The key frictions that generate output drops in the existing
literature on sudden stops are subtle ones for which so far there is little
direct evidence. Finding that evidence is the challenge for future research.

\newpage

\begin{center}
\textbf{REFERENCES}
\end{center}

\begin{description}
\item \textbf{Aiyagari, S. Rao; Christiano, Lawrence J. and Eichenbaum,
Martin.} \textquotedblleft The Output, Employment, and Interest Rate Effects
of Government Consumption.\textquotedblright\ \textit{Journal of Monetary
Economics,} October 1992, \emph{30}(1), pp. 73--86.

\item \textbf{Calvo, Guillermo A.\ }\textquotedblleft Capital Flows and
Capital-Market Crises: The Simple Economics of Sudden
Stops.\textquotedblright\ \textit{Journal of Applied Economics}, November
1998, \emph{1}(1), pp. 35--54.

\item \textbf{Chari, V.V.; Kehoe, Patrick J. and McGrattan, Ellen R.}
\textquotedblleft Business Cycle Accounting.\textquotedblright\ Research
Department Staff Report 328, Federal Reserve Bank of Minneapolis, 2004.

\item \textbf{Chari, V.V.; Kehoe, Patrick J. and McGrattan, Ellen R.}
\textquotedblleft Sudden Stops and Output Drops.\textquotedblright\ Research
Department Staff Report, Federal Reserve Bank of Minneapolis, 2005.

\item \textbf{Christiano, Lawrence J.; Gust, Christopher and Roldos, Jorge.}
\textquotedblleft Monetary Policy in a Financial Crisis.\textquotedblright\ 
\textit{Journal of Economic Theory,} November 2004, \emph{119}(1), pp.
64--103.

\item \textbf{Kiyotaki, Nobuhiro and Moore, John.} \textquotedblleft Credit
Cycles.\textquotedblright\ \textit{Journal of Political Economy}, April
1997, \emph{105}(2), pp. 211--48.

\item \textbf{Mendoza, Enrique G.} \textquotedblleft\ `Sudden Stops' in an
Equilibrium Business Cycle Model with Credit Constraints: A Fisherian
Deflation of Tobin's q.\textquotedblright\ Manuscript, University of
Maryland, 2004.

\item \textbf{Neumeyer, Pablo A. and Perri, Fabrizio.} \textquotedblleft
Business Cycles in Emerging Economies: The Role of Interest
Rates.\textquotedblright\ National Bureau of Economic Research (Cambridge,
MA) Working Paper 10387, 2004.
\end{description}

\newpage

*Chari, University of Minnesota and Federal Reserve Bank of Minneapolis
(Research Department, 90 Hennepin Avenue, Minneapolis, MN 55401-1804);
Kehoe, Federal Reserve Bank of Minneapolis and University of Minnesota; and
McGrattan, Federal Reserve Bank of Minneapolis and University of Minnesota.
The authors thank the National Science Foundation for its support and Kathy
Rolfe for excellent editorial assistance. Any views expressed here are those
of the authors and not necessarily those of the Federal Reserve Bank of
Minneapolis or the Federal Reserve System.

\end{document}

%%%%%%%%%%%%%%%%%%%% End /document/Sudden_stops.tex %%%%%%%%%%%%%%%%%%%