The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
THE CHANNEL OF MONETARY TRANSMISSION TO DEMAND:
EVIDENCE FROM THE MARKET FOR AUTOMOBILE CREDIT
Sydney Ludvigson
Federal Reserve Bank of New York
Research Paper No. 9625
August 1996
This paper is being circulated for purposes of discussion and comment only.
The contents should be regarded as preliminary and not for citation or quotation without
permission of the author. The views expressed are those of the author and do not necessarily
reflect those of the Federal Reserve Bank of New York or the Federal Reserve System.
Single copies are available on request to:
Public Information Department
Federal Reserve Bank of New York
New York, NY 10045
I
The Channel of Monetary Transmission to Demand:
Evidence from the Market for Automobile Credit
Sydney Ludvigson t
Federal Reserve Bank of New York
August 9, 1996
Abstract
In response to tight money, both consumer loans and consumption fall. In this paper, I
ask whether there is any causality running from loans to consumption by focusing on how
the composition of automobile finance between bank and nonbank sources of credit changes in
response to unanticipated innovations in monetary policy. The results indicate that contractionary monetary policy reduces the supply of bank consumer loans, which in turn produces
a decline in real consumption. The evidence is therefore supportive of a credit channel theory
of monetary transmission to aggregate consumption. Furthermore, the nature of automobile
finance is uniquely suited to identifying which of two possible sub-channels is relatively more
important, and suggests the results are more likely consistent with a bank lending channel
than with a pure balance sheet channel.
t Domestic Research - 3rd floor,
33 Liberty Street, New York, NY 10045. Phone: 212-7206810, E-mail: sydney.ludvigson@frbny.sprint.com. This paper appears as part of my Princeton
University doctoral dissertation. I am grateful to Ben Bernanke and John Y. Campbell for
their advice and many valuable discussions, to Illian Mihov for helpful comments, and to the
Alfred P. Sloan Foundation for financial support. The views expressed in the paper are those
of the author and are not necessarily reflective of views at the Federal Reserve Bank of New
York or the Federal Reserve System. Any errors and omissions are the responsibility of the
author.
1
Introduction
Is there any causal link between the availability of consumer credit and consumer demand? In bad
times, does consumer credit decline simply because demand is lower, or is the downturn exacerbated
by a reduction in the supply of consumer loans to credit-dependent households who would otherwise
maintain consumption at higher levels?
The relation of credit market conditions to the production and investment decisions of firms has
recently been a topic of extensive research (Bernanke and Blinder, 1988 and 1992, Bernanke, Gert.ler
and Gilchrist, 1994, Gertler and Gilchrist, 1993). This research emphasizes tht' importance of
endogenous changes in the firm's external finance premium in amplifying the effects of conventional
(money) channels of monetary transmission. In particular, the credit channel theory of monetary
transmission holds that recessions are worsened by the inability of credit-dependent. firms to borrow
at the levels they could in good times, either because banks decrease the supply of loans aft.er a
monetary tightening or because firm's credit-worthiness declines as their net. worth falls. 1 If firms
depend on bank loans to maintain production and investment activities and to finance inventories,
a monetary contraction will have a great.er impact on real activity than that. predict.eel by the pure
money view, according to which loan supply simply responds passively to changes in the demand
for credit. induced by variation in the cost-of-capital.
One can distinguish analogously bet.ween two possible channels of monetary transmission to
demand with the following question: Does consumer credit. simply passively respond to consumption
demand (it.self responding to variation in current. income), or does a reduct.ion in the supply of
consumer credit. by banks further retard consumption? This quest.ion is in fact t.wofolcl: 1) docs a
monetary contraction reduce the supply of consumer loans from banks, and 2) if so, does this lead
to lower consumption? Alt.hough a considerable body of literature has developed to investigate the
mechanism by which monetary policy shocks affect production and investment. of firms, lit.tie work
identifies these channels to the demand side of the economy. In this paper, I use evidence from
1 See
Bernanke and Gertler, 1995, for a general discussion of the broad credit channel.
1
the market for automobile credit to determine how unanticipated shifts in monetary policy affect
consumption.
The main problem with discriminating between the money and credit channels from a reading of
the aggregate data is that the two theories are observationally equivalent in the usual econometric
analyses. For example, in either theory, loans contract after a monetary tightening, rendering it
difficult to distinguish between supply-versus-demand induced movements in credit. Simple correlations between loans and various measures of real activity tell us something about the timing of
events, but little about the direction of causality.
One approach to this problem is suggested by Kashyap, Stein and Wilcox (1991). They study
how the ratio of business bank loans to commercial paper (a close substitute for bank credit) is
influenced by a shift in monetary policy. Their insight was if a monetary contraction leads t.o a
decline in the ratio of bank loans to the sum of commercial paper and bank loans, the supply of
bank loans must be shrinking, since presumably the demand for both types of finance should fall
in rough proportion. They found that. a monetary contract.ion was associated with a decline in this
ratio and concluded that. tight money leads to a reduction in the supply of business loans from
commercial banks, consistent. with predict.ions of the credit channel theory. 2 Thus, whereas the
money view makes no explicit. prediction about the behavior of the composition of business finance,
an implication of the credit channel theory is that. a monetary contraction will shift this composition
away from bank loans and toward non-intermediated credit.
This paper follows a research strategy very much in the spirit of Kayshap, Stein and Wilcox
(KSW) by investigating how the composition of consumer loans bet.ween bank and nonbank sources
2
Because Kashyap, Stein and Wilcox interpret their results as a reduction in the supply of bank loans, they find
the evidence supportive of, not just a credit channel, but more specifically a bank lending channel, which is thought
of as a sub-channel of the broader credit channel. Others have argued that their evidence is also consistent with an
alternate sub-channel, the balance sheet channel. In the latter, banks do not necessarily decrease the supply of loans,
but because there is a counter-cyclical demand for business credit, they may shift their loan portfolios from more
risky to less risky borrowers in bad times. Because better borrowers have access to the commercial paper market, it
is then difficult to distinguish between the two possibilities. In section 3, I argue that the evidence presented here
for automobile credit is better suited to sorting out the two possibilities, and is more supportive of a bank lending
channel than a pure balance sheet channel. Henceforth, I therefore view the empirical exercises as testing whether
the supply of loans is affected by contractionary monetary policy.
2
of finance changes after an unanticipated disturbance to monetary policy. Consumer inst.aliment
credit is issued by banks and by businesses and is given in aggregate time series data. Variation
in the composition of consumer finance after a contractionary monetary shock provides the needed
econometric identification for sorting out supply and demand movements in household credit: a fall
in the ratio of bank loans to the sum of bank and nonbank loans indicates a constrict.ion in the
supply of bank credit. Before these empirical tests can be implemented, however, two issues need
to be addressed.
The first implementation issue concerns the choice of household credit. series. To ensure that.
bank and nonbank consumer installment credit. is issued for the same pool of products, I focus my
study on loans for a particular product - automobiles. 3 Automobile credit. is ext.ended by commercial
banks or by finance companies which are usually subsidiaries of major aut.omobile manufacturers.
A second implementation issue concerns the measurement of monetary policy. In order to track
how the composition of consumer credit responds to an unanticipated shift. in monetary policy, one
needs some way of measuring the stance of monetary policy. A descriptive strategy for identifying
exogenous monetary policy shocks is advanced by Romer and Romer {1990) who interpret minutes of
the Federal Open Market. Committee to identify particular episodes of tight.er policy. Alternatively,
Bernanke and Blinder (1991) argue that. innovations in the federal funds rate are a good indicator
of central bank policy because the Federal Reserve has direct. leverage over this rate in the short
term.
Christiano, Eichenbaum and Evans (1994) compare these two ways of measuring monetary
policy. As they point out, one advantage of the Romer approach is that it requires no formal
specification of the Federal Reserve's policy rule. lt.s primary disadvantage, however, lies with
the ad-hoc identification of exogenous policy episodes (of which Romer and Romer locate six) as
deliberate and large movements in policy controlled variables. Furthermore, t.hc federal funds rate
provides more information about. the stance of monetary policy because there is a policy episode
3 This
strategy also eliminates certain types of consumer credit that are available to borrowers at their option (e.g.
credit cards) and which make it particularly difficult to distinguish supply and demand mo\'ements when observing
changes in the amount of debt outstanding.
3
for each data point in the sample. Finally the federal funds rate, unlike the Romer dates. furnishes
a quantitat ive gauge of the intensity of the policy action. For these reasons, I restrict most. of mv
calculations to those that measure monetary policy as an innovation in the federal funds rate, once
a particula r feedback rule, or reaction function, for monetary policy has been assumed.
The general strategy I employ is to use vector autoregressions to evaluate whether shocks to
Federal Reserve policy influence the composition of automobile credit, and whether variation in
this composition in turn affects automobile consumption. I test: (i) whet.her the composition of
automobile finance changes in response to an unanticip ated innovation in federal funds rate (helping
to sort out supply and demand movements in credit), and (ii) whether variation in this composition
affects sales of new automobiles (providing information on how substitut able these two forms of
finance are). In short, the results indicate that contractionary monetary policy rednces the relat.ivP
supply of bank consumer loans, which in turn produces a decline in real consumption.
The rest of this paper is organized as follows. In the next section, I briefly ontline some relevant
institutio nal details of automobile finance companies. In order t.o understa nd how finance companies
(the primary source of nonbank automobile credit) alter the extension of credit. in response t.o shifts
in monetary policy, it. is helpful to have some sense of their sources of funds. In Section 3.1 I
discuss the empirical procedure used to evaluate the possibility that a credit. diannel t.o demand
exists. Section 3.2 is devoted to documenting how t.hP composition of cons11m0r finance changes
in response to an innovation in the federal funds rate. Sect.ion 3.3 presents cvidenc<' that changes
in the composition of automobile finance affects automobile sales. and scd.ion 3 A discusses some
related issues abont. how t.o interpret the resnlts. Section 4 cond11d0.,.
2
Some Instit utiona l Aspec ts of Finan ce Comp anies
This section briefly describes some institutional aspects of finance companies' resources in order to
gain a sense of what constraints they operate under in extending funds over the bnsiness cycle. Sales
4
finance companies purchase retail installment sales contracts from automobile dealers and collect
payments directly from the consumer. 4 Finance companies also make wholesale loans to sellers in
order to help them cover the costs of high inventories in times of slow demand. Historically, finance
companies typically competed with each other by varying the rates charged on wholesale financing
rather than on retail financing, since a lower rate on retail financing commonly just reduces the
finance charges paid by the individual consumer. Today, much of the nonbank credit is extended
by a few national firms which are subsidiaries of the Big Three U.S. automobile companies (GM,
Ford and Chrysler).
The growth in finance companies from 1980 to 1990, at an average annual rate of 11.4 percent,
was initially financed by funds raised primarily in the commercial paper market.. 5 Commercial
paper issuance by finance companies grew an average of 12 percent per year during this period.
Corporate bond issuance subsequently surged leaving it the largest component of finance company
liabilities by 1990. For sales finance companies as a whole (the largest of which are auto sales finance
companies but they also include subsidiaries of major appliance manufacturers and retailers), the
largest sources of funds at the end of 1982 (as a percentage of liabilities and capital combined) were
long term debt (debt with original maturity of one year or more, 31.5%) and commercial paper
(24.9%). Bank loans made up a much smaller percentage (7.4%). 6
Similarly, for the major automobile finance companies, short term funding requirements are met
primarily through the direct sale of commercial paper. For example, General Motor Acceptance
Corporation's (GMAC) short term debt outstanding at December 31, 1983 a.mounted to $19.7
billion in the United St.at.es, about half of t.heir tot.al out.standing debt world wide. Medium and
long term not.es, including those marketed directly to investors in the U.S., account for slightly less
than half of the total indebtedness, while bank credit represented a.bout 3% of total indebtedness.
In 1989, short term debt accounted for a.bout 51 % of total indebtedness, though by 1994 it was less
4 The
information is this section is drawn from various annual reports of General Motors Acceptance Corporation,
Ford Motor Corporation, Chrysler Motor Corporation. and from Olney {1989).
5 FRBNY (Federal Reserve board of New York) Quarterly Review, summer 1992, pp. 26-27. The growth of bank
loans over this period averaged 8.4 percent annually.
6 Finance Facts Yearbook, 1984, p.61.
5
at 33%. Similarly, Ford Credit's commercial paper debt in 1992 was about 86% of its total short
term debt, itself approximately 53% of total debt.
The financial services operations of these companies, in addition to relying on loan collections and
retained earnings, depend heavily on their ability to raise funds in capital markets. Stock issuance,
surplus and undivided profits represented only 12.3% of all finance company funds in 1982. 7 Credit
ratings are therefore a key determin ant of the growth of finance company credit. When asset growth
is regressed on senior debt ratings, capital ratios, parent relationships, and demand conditions, only
the finance company's own credit rating significantly explains asset growth. 8
The credit ratings of finance companies are themselves primarily determined by the capital
strength of the parent company. Hence, financial ties to the parent company can lower the cost
of borrowing by raising the finance company's credit rating. The parent company 's senior debt
ratings are found to be more importan t than own capital ratios, asset size, and parent relationships
in determining the finance company's senior debt ratings. Rating agencies consider the capital
support the parent. company has provided in the past and their capacity to st.and behind finance
company debt. in the future. 9
The source of short-ter m finance is relevant. because consumers do not. have direct. access to nonintermediated credit. markets. Bernanke, Gertler and Gilchrist (1993) show that. there is a flight.
t.o quality in downturns, in which credit. is shifted toward high-grade borrowers with access to the
commercial paper market, and away from riskier, smaller firms. Automobile finance companies are
precisely this type of high-grade borrower. Thus some of their high-grade st.at.us may be indirectly
imparted to automobile consumers if, when faced with a decline in the supply of bank credit, they
are able to substitut e into finance company credit..
7 Finance
Facts Yearbook, 1994, p. 67.
Q~rterly Review, swnmer 1992, p. 28.
9 FRBNY Quarterly Review,
summer 1992, pp. 28-29.
8 FRBNY
6
3
Empirical Tests
3.1
Empirical procedure
Following the terminology used in KSW, I construct a variable called the mix of automobile finance.
equal to the ratio of bank automobile credit to bank automobile credit pins finance company automobile credit, all measured as the stock of credit outstanding. I then estimate a set of vector
autoregressive (VAR) models using monthly data from 1965:1 to 1994:12 and ask what the impulse responses for the mix variable and for automobile sales look like in response to innovations in
monetary policy. 10 Each VAR was estimated using 12 Jags of each variable.
As proposed by Bernanke and Blinder (1992), I nse innovations in the federal funds rate to
represent shocks to monetary policy. I make an additional assumption about the timing of information available to the Federal Reserve by placing the funds rate last in the VAR. TllP last equat.ion
in the VAR then represents a reaction function of the central bank, so that th<' innovation is the
unanticipated policy shock. The typical justification for this placement of the federal funds rate in
the VAR is the idea that it can affect other variables only with a one period lag. while the rate'
itself can respond contemporaneously given the Federal Reserve's reaction function.
The first model I estimate includes 5 variables: the log of industrial production. the log of th<'
consumer price index (CPI), the log of the commodity price index, the financing mix. and the federal
funds rate, in that order. 11 Christiano, Eichenbanm and Evans ( I 995) emphasize' tlw importance'
of including a commodity price index in VARs designed to measnr<' the effects of innovations in
monetary policy (see also Sims, 1992). As they demonstrate, this inclusion is necC'Bsary to eliminate
the price puzzle, the finding that contractionary monetary policy shifts appear to lead to a sustained
increase in the price level in VARs which do not include a commodity price index. Including the
commodity price index controls for episodes such as the oil price shock of I 974, in which the
10 The
data for the mix variable are obtained from the Federal Reserve Board of Governors in their G.19 and G.20
statistical releases. The other data are obtained from Citibase data bank.
11 Even though several of these variables may be nonstationary, I do not difference them since tho hypothesis tests
based on the VAR in levels will have standard asymptotic distributions; see Sims, Stock. and Watson (1990).
7
subsequent rise in inflation was preceded by a rise in the federal funds rate.
Next, I ask how real retail passenger car sales react to an innovation in the federal funds rate or
in the mix variable. 12 This model includes the log of industria l production, the log of the CPI and
of the commodity price index, the log of real car sales or the inventory to sales ratio, and either the
mix, or the federal funds rate, in that order.
In the next section, I discuss impulse responses using the first model, where I investigate how
monetary policy affects the composition of automobile finance. Bank loans should make up less
of the consumer's overall financing portfolio if depository institutio ns are constricting the supply
of consumer credit in response to a positive innovation in the federal funds rate. The subsequent
section uses the second VAR model and asks how the composition of automobile credit influences
real activity. If consumers cannot perfectly substitut e between the two forms of finance, then a
negative shock to the mix variable should lead to a decline in automobile consumption. Finally,
in section 3.3 I also add the finance composition variable to some standard automobile demand
equations to see if it. adds any explanatory power.
3.2
The effect of moneta ry policy on the composition of automobile credit
I begin by analyzing how the composition of automobile finance changes following periods of tight.
money. Figures I and IA plot. the financing mix (the ratio of bank automobile credit to credit
issued by automobile finance companies plus bank credit) over time. Figure I plots the simple mix
and figure IA plots the mix detrended by a quadratic time trend. The solid vertical lines in the
latter are drawn at dates Romer and Romer identify as coinciding with episodes of contracti onary
monetary policy. 13
As the figures show, there appears to be relatively little trend in the variable, though there
are some large low frequency movements. A particularly sharp drop in this ratio occurs in the
12 The
sales series is obtained from Citibase.
dates are: October 1947, September 1955, December 1968, April 1974, August 1978, and October 1979.
13 These
8
late 1970's. During the early part of that decade, depository institutions became increasingly
aggressive providers of automobile finance. After 1979, however, other, higher yielding markets for
commercial bank lending became available and decreased the amount of funds these institutions
placed in automobile credit. This resulted in a shift back to finance companies. Finance companies'
share in total automobile credit outstanding rose from 29% in 1980 to 37% by 1982. 14 This trend
appears to be reversing itself in more recent years.
Figure IA demonstrates that the detrended mix generally declines shortly after a Romer date.
Thus, if the Romer dates adequately capture periods of tight money, the figure suggests contractionary shifts in monetary policy lead to a reduction in the relative supply of bank consumer credit.
Figure 2 displays the response of the composition of automobile finance to a one standard
deviation increase in the federal funds rate, with two standard error bands computed using Monte
Carlo simulations and assuming the errors are normally distributed. There is a significant decline
in the ratio of bank loans to total automobile credit over a period of 60 months after a positive
innovation in the federal funds rate. In figures 3 and 4 I document the response of each component
of the mix variable to a positive innovation in the federal funds rate. The financing mix drops
primarily because bank loans contract. 15
The latter finding is in contrast to the results obtained in Gertler and Gilchrist (1993), where the
composition of corporate finance between bank loans and commercial paper changes primarily because commercial paper issuance rises substantially after a federal funds rat.e increase: the response
of bank loans is relatively flat. The usual explanation for the Gertlcr and Gilchrist result starts with
a countercyclical demand for short-term business credit caused by a <ledinP in thP firm's cash flow
and a need to finance growing inventories. This permits two possible interpretations of the change
14 Finance
Facts Yearbook, 1982, p. 52.
and Mihov (1995) suggest funds rate targeting as a description of Federal Reserve-Policy may not be
stable over various sub-samples. I split the full sample into three sub-samples consistent with Bernanke and Mihov:
65:1- 79:9, 79:10-94:3, 84:2-94:3. As is typical over such a small sample size, the standard error bands are quite
wide. Nonetheless, in the pre-1979 sample, the financing mix declines slightly in response to a funds rate shock,
consistent with the notion that the funds rate was a good measure of monetary policy over this period. In the other
two sub-samples, the financing composition declines, but not outside of the two standard error bands. In the last
sub-sample this is partly due to the relatively low variability of the funds rate and other variables that seems to be
a characteristic of the post 1984 period.
15 Bernanke
9
in finance composition: Either firms substitut e between loans and paper as banks constrict. loans
(as KSW suggest), or because low-grade borrowers find their balance sheet positions deteriorating.
relatively more credit is supplied to high-grade borrowers who have access to the paper market. (as
Gertler and Gilchrist suggest). In the case of automobile finance, there is no obvious analogous need
for short-ter m countercyclical finance of automobile consumption; therefore a more likely explanation for the behavior of the automobile finance mix is that consumers substitut e bet.ween the two
forms of credit, rather than lending institutio ns shifting their funds among consumers of varying
quality. The distinction between the KSW interpret ation and the Gertler and Gilchrist int.erprE>tation is often identified as two sub-channels of the broad credit channel theory; the former often
termed the bank lending channel, and the latter the balance sheet. channel. Which interpret ation
is more appropri ate has not been conclusively demonstrated, though in the next sect.ion I will arguf'
that the bank lending channel seems to better fit the facts of automobile finance.
In summary, tables 2, 3, and 4 indicate the ratio of bank auto loans to financf' company obligations declines in the face of a positive shock to the federal funds rate. This suggests that. the
correlation bet.ween consumer loans and real activity is not. simply a passivf' accommodation of loan
supply to fluctuations in demand; instead, the evidence signifies that. banks reduce the supply of
loans in response to a cont.ract.ionary policy shift..
3.3
The effects of finance composition on real consumption
The evidence present.eel in the last. sect.ion indicates that. commercial banks ma~· dP<Teasc the supply
of consumer loans in response to tight. money, and that. consumers may lw abl<' t.o substitut e into
alternative forms of finance in response. If the two forms of finarH'<' an• not perfect substitut es,
however, the overall supply of loans to credit-dependent individuals will decline, leading to a fall
in consumption. In this section, I first investigate whet.her the financing composition of consumer
credit. significantly affects automobile consumption in a VAR cont.rolling for income and the price
level. This provides a test of how substitut able finance company contracts are for bank automobile'
10
loans. Next I estimate a set of standard automobile demand equations to see if the composition
variable adds any explanatory power.
I use monthly data on retail sales of motor vehicles, seasonally adjusted and deflated, as well
as on the ratio of real passenger car sales to inventories. 16 As figure 5 shows, real automobile
sales decline in response to a federal funds rate increase, confirming a positive correlation between
tight money and a decline in consumption. Figure 6 plots the cumulative response of the log of
automobile sales to a one standard deviation increase in the mix variable. Real retail sales rise in
response to an increase in bank loans as a fraction of total credit. Approximately six months after
a positive innovation to the financing mix, auto sales significantly increase. Figure 7 shows the
response of the passenger car inventory to sales ratio to a one standard deviation increase in the
mix variable. Though this series is much more volatile than car sales alone, sales still rise relative
to inventories within 8 months after the mix rises.
In principle, another way to test whether a decline in the supply of bank loans affects real
variables is to see whether the spread between the bank interest rate and finance company interest.
rate rises in response to tight money. A credit. channel theory for consumption would predict that
there are bank dependent borrowers who can not perfectly substitute bet.ween bank and nonbank
credit, resulting in an increase in the spread. This is simply another way of asking whet.her the
decrease in bank loan supply is an additional effect of cont.ractionary monetary policy on real
variables that exists over and above conventional channels.
In practice, using the financing mix to infer the state of bank loan supply is generally more
reliable because there may be reasons why the spread does not increase even if banks do decrease
the supply of loans and even when the two forms of finance are imperfect substitutes. For example,
if banks respond to a decrease in reserves by lending to a less risky pool on average, the lower
default probability may hold down the bank lending rate so that, the spread is little affected. On
the other hand, there is not much reason to think, practically or theoretically, that the spread
would rise unless banks were constricting the supply of loans and unless alternative forms of outside
16 The
ratio is obtained from Data Stream International.
11
finance are imperfect substitutes for borrowers. Therefore, an increase in the spread in response to
tight money lends at least tentative support for the credit channel theory; in contrast, if the spread
remains unchanged or even falls, it is likely to be a less reliable indicator of bank loan supply and
financing substitutability.
With this caveat in mind, I display how the spread responds to a one standard deviation shock
in the federal funds rate in figure 8. 17 The spread significantly rises after about 2 months, lending
further support to the hypothesis that banks cut the supply of consumer loans and that consumers
find it more costly to use nonbank finance. 18
Next, I consider two aggregate automobile demand equations and ask whet.her the financing
composition adds any explanatory power. The first demand equation comes from Chow (1957).
Chow assumes that individuals choose automobile consumption by maximizing utility subject to
consumption being proportional to income. The desired stock of automobiles is a linear function of
the relative price of new automobiles and real disposable income, suggesting the following equation
for automobile demand:
(1)
where S, is sales of new automobiles, o is a constant., RP, is the relative price of new automobiles
in terms of consumption, Y, is real, disposable income, and X,_ 1 is the consumer's lagged stock of
automobiles.
Another demand equation arises from a study by Blanchard and Melino (1986). They examine
17
The interest rate data are obtained from the Federal Reserve statistical release G.19 and are annual percentage
rates. Interest rates at commercial banks are simple W1weighted averages of each bank's most common rate charged
for a 48 month new car loan during the first calendar week of the middle month of each quarter. Finance company
data are from the subsidiaries of the three major U.S. automobile manufacturers and are volume-weighted averages
covering all loans for new cars purchased during the month. Since the data for bank rates are of quarterly frequency,
I average the monthly finance company rates over the quarter for figure 8.
18 The sample mean of the spread using these rates (bank rate minus finance company rate) is -0.84 percent using
data from 1972:02 to 1995:01.
12
a whole market structure deriving both a demand and supply equation. Here I focus on their result.s
for the consumer's problem and just. extract. a demand equation from their model. In that. model.
the consumer maximizes the conditional expected value of additively separable utility from now
until infinity over consumption and the stock of automobiles, subject. to a wealth accumulation
constraint and an accumulation constraint. for the stock of cars. Aft.er linearizing t.he first. order
conditions and making some assumptions about the information set of the consumer, t.he following
demand equation arises, which is very similar to (1) 19 :
(2)
where C, is consumption, excluding automobile services. 20
Table 1 presents the results. The table reveals a puzzling out.come with respect to the relative
price, but otherwise produces some significant pat.terns with respect. to the composition of automobile finance. The first row estimates (1) itself, and the second row adds in the' current period's
financing composition as an additional explanatory variable. The current. financing composition is
strongly significant in the demand equation, as is disposable income. The price' variable which is
t.he CPI component for new cars divided by the PCE deflat.or is, oddly, strongly significant with
a positive coefficient. BM solved out the complete market structure model, so that. price was de~
termined simultaneously by supply and demand. They found however that tll<' price• equation was
in substantial contradiction with the model and speculate that prices could bC' measured badly.
Alternatively, if there are costs to adjusting prices, lagged prices should be st.at.<' variables in t.hc
19 Once
the first order conditions are linearized, the problem consists of solving a linear, expectational difference
equation. I make similar assumptions to those of Blanchard and Melino (BM) and suppos<> that the consumer's
information set contains only current and lagged variables and that consumption and prices are uncorrelated with
current and lagged disturbances to marginal utility, and are represented by a first order autoregressive structure,
making it easy to use the method of undetermined coefficients to solve for (2).
20 Data is obtained from Citibase data bank, see table 1. I follow Blanchard and l\lelino and use constant dollar
personal consumption expenditures as a measure of C,, and measure RP, as the CPI component for new cars divided
by the PCE dellator. To obtain a stock for consumer durables, I assume the initial stock is the average on expenditure
for first 15 months (67:1 • 68:3), dividing by the depreciation rate, and that subsequent values of the stock evolve
using a depreciation rate of 2% per quarter.
13
model and included in the regressions. In subsequent regressions I therefore include lags of the price
variables as well as for C, in case the first order autoregressive specification for that variable is too
restrictive.
The third, and other rows add in the rate on the three month treasury bill as a proxy for the
consumer's user cost. Again the financing composition remains a strong predict.or of automobile
sales even though the t-bill adds explanatory power to the regression. When lags oft.he relative price
are added in (fourth row), the current. period's relative price no longer has a positive coefficient. but. is
also no longer significant, and the sign on the sum of lagged price coefficients is again positive. Row
5 shows that it makes little difference whet.her consumption or disposable income is used, indicating
that there is small distinction in practice bet.ween the specifications in (1) and (2). Finally, t.hc last
two rows add in lags of the financing mix. The sum of the coefficients on the current mix and its
lags has a positive sign and is strongly significant. with or without the addition of the t.-bill rate,
while controlling for consumption, the relative price of new cars and its lags, and the lagged stock
of cars.
One obvious reason for why the relative price might. appear with a negative coefficient. in the
structural form demand equations (1) and (2) is the potential for endogeneit.y problems t.o arise.
Though the credit. channel theory in this case directly applies to consumer demand, the VAR
results suggest. t.hat, in equilibrium, final sales are affected. I at.tempt. to eliminate any potential
simultaneity bias by next. adding t.he financing composition into Blanchard and Melina's reduced
form equation for automobile sales:
(3)
where I, is the producer-dealers inventory of automobiles, Zt is automobile product.ion, W, is the
real wage, and the other variables are defined as before. 21 Blanchard and Melino obtain the reduced
21 I,
and
z, are obtained from the Citibase data bank, where the first is given as retail auto inventories of total
14
form equation by simultaneously solving the demand problem discussed above and a supply problem
for a representative automobile firm (which makes both the production and the sales decision)
that maximizes the expected present value of cash flows subject to an inventory accumulation
constraint. I also follow BM and add in dummy variables for strike dates of the Big Three automobile
manufacturers and a (quadratic) time trend.
Table 2 presents the results of estimating (3). The first row estimates (3) itself. The results are
very similar to those obtained by BM for all variables except the real wage. Here the coefficient is
significantly positive, whereas they found it to be insignificantly different. from zero. The second
row adds in the financing composition which again appears to be an important. determinant. of
automobile sales. The marginal significance level is less than 0.0001. The third row adds in the
financing composition and its lags; the sign of the sum of all coefficients on this variable is positive
and strongly significant. Finally the last two rows add in the price of fuel as a robustness check. The
price of fuel adds significant. explanatory power to the regression, but the financial fact.ors continue
to influence automobile sales at very high marginal significance levels.
In summary, the analysis in this sect.ion shows that changes in t.lw composition of consumer
finance are associated with variation in real consumption not. capt.nred by ag!!;l"egat.e income or
price variables, or by lagged stocks of automobiles or some measure of the user cost.. In addition the
reduced form estimation indicates that. financing elements influence automobile sales in equilibrium,
consistent. with the VAR results. This, along with the evidence in the previous sect.ion, suggests
that. consumers can not perfectly substitute into nonbank forms of automobile' finance when banks
reduce the supply of automobile credit. aft.er a monetary tightening. The <',·idcncc is consistent wit.h
a credit theory of monetary transmission, according to which credit. market. conditions exacerbate
the conventional effect of depressed demand brought about. by a movement. toward tight.er policy.
new passenger cars, and the second is measured as the Industrial Production Index for automobiles. W, is obtained
from Datastream International and is the average, hourly wage for production workers in the automotive industry.
15
3.4
Discussion
Two points about what the results represent are worth discussing. First, one possible interpret.at.ion
of the behavior of the mix variable in response to a positive federal funds rate shock is that, rat.her
than commercial banks cutting the supply ofloans, consumers willingly switch into finance company
credit because they offer better financing terms in bad times when automobile dealers are facing
declining demand and growing inventories.
Several considerations shed doubt on this interpretation. One is that the aut.oregressions and
OLS regressions preformed above have controlled for the most plausible conventional variables
(output or income, and interest. rates) which could impact the demand for automobiles; changes in
the composition of external finance are associated with changes in automobik consumption abovc
and beyond the component. attributable to fluctuations in income. If automobile demand declines
simply through conventional channels, the financing mix should have no additional explanatory
power. Furthermore, the scenario begs the question of why the automobile dealers or the financP
companies (the manufacturers) would choose to cut the financing cost. rat.her than cut. the list.
price directly. In fact, manufacturers usually offer a choice bet.ween better financing terms or a
cash rebate which comes directly from the manufacturer. Offering better financing terms instead
of a lower price would make sense if there was some reason t.o pricc discriminat<' bet.ween credit.dependent. and unconstrained consumers, so that. the manufacturer effect.ively offers a price cut just.
to the credit-dependent. borrowers. But finance companies would pric,• discriminat.P in this way
only if the monetary contract.ion left. them with a comparative advant.ap;c in furnishing credit.. At
the margin, there is lit.tie reason to expect manufacturers to optimally lowcr pricc only t.o crcdi !dependent. consumers unless the consumer's ability to borrow from banks has diminished. Finally,
figure 9 shows that. the interest rat.es charged by finance companies rise in rcsponse to a positive
shock in the federal funds rate (so the spread rises because bank rat.es increasc even more). Though
the interest. rate does not fully capture the financing terms of a credit. contract., at least. t.o a rough
approximation this evidence seems at. odds with the not.ion that. the shift. in the financing mix is
16
a consequence of price cutting by automobile dealers (via financing costs) designed to stimulate
sagging sales and deplete mounting inventories.
Another issue that arises is whether the results can help discriminate between the two subchannels (described above) of the broader credit channel theory. Recall the two possible interpretations of why bank loans fall relative to nonbank credit in response to a monetary contraction: one
is that banks cut the supply of loans to borrowers who cannot perfectly substitute into nonbank
forms of credit (the bank lending channel), the other is that the quality mix of borrowers changes:
because the balance sheets of borrowers deteriorates, there is a flight to quality by banks at the
same time that there is a surge in the demand for credit by high-grade borrowers who have access
to non-intermediated outside finance (the balance sheet channel). In this context however, the pure
balance sheet channel interpretation rests crucially on the institutional assumption that only highgrade borrowers have access to non-intermediated credit. This assumption is justified in the case of
firms, where commercial paper is the primary substitute for bank loans and is generally available
only to companies with out.st.anding credit ratings.
Construing the evidence presented in this paper as a pure balance sheet effect is more problematic. I obtained data on delinquency rates for· t.he automobile credit contracts of commercial banks
and finance companies. 22 The delinquency rat.es on bank loans are consistently below rates on finance company loans: In monthly dat.a from 1980:1 to 1994:12, the delinquency rate at commercial
banks averaged 1.81 percent., while that. of finance companies averaged 2.07 percent.. This evidence
does not support the balance sheet channel because the behavior of the mix cannot. be explained
by a shift to higher quality borrowers. The opposite appears to be true in the case of automobile
credit: riskier borrowers are more likely to use finance company credit.
On the other hand it is precisely because nonbank forms of business credit are accessible almost
exclusively to high grade borrowers that. is has been particularly difficult to draw a distinct.ion
bet.ween the two sub-channels in existing empirical work involving the composition of firm finance.
Staten, Michael E., and Robert W. Johnson, Ed., 1995, Household Credit Data Book, fourth edition,
West Lafayette: Credit Research Center. The rates are measured as the percent of auto loans thirty or more days
past due, seasonally adjusted.
22 Source:
17
The institutio nal nature of automobile finance provides a unique instrume nt for distinguishing
between the balance sheet and bank lending channels. Thus, at least in the case of consumer loans,
the results here provide evidence of a bank lending channel of monetary policy transmission as
distinct from a balance sheet channel.
4
Concl usion s
The importan ce of consumption in cyclical fluctuations has long been emphasized as a reason to
understa nd what determines this component of GNP in aggregate data. The goal of this paper is
to identify the channels through which aggregate consumption may be affected by unanticip ated
shifts in monetary policy. Conventional theories explain the pro-cyclical behavior of consumer
credit by assuming that loans merely passively respond to the income induced change in demand.
If instead, lending institutio ns react to tight money by decreasing the supply of consumer loans,
credit-de pendent households would be forced to cut spending by more than conventional mechanisms
imply. The bank loan supply channel predicts that. economic downturn s will be exacerba ted by the
decline in consumer credit. available from depository institutio ns, helping to understa nd why large
cyclical swings in consumption are often associated with relatively small monetary impulses.
The evidence presented in this paper is supportiv e of a credit. channel theory of monetary
transmission t.o aggregate consumption demand. The composition of automobile credit. between
banks and finance companies is significantly changed by unanticip ated shocks to monetary policy, a
finding consistent with the predict.ions of a credit channel theory, but not with conventional theories
of monetary transmission. F\irthermore, innovations in the financing mix have significant. affects
on automobile consumption not. explained by variation in income. Finally, the results are more
likely consistent with a bank lending channel than with a pme balance sheet. channel, since the
primary nonbank source of automobile credit. - finance company contracts - is generally used by
riskier borrowers. The results support the hypothesis that the correlation bet.ween consumer debt
18
outstanding and real activity is at least partially attributable to shifts in the supply of consumer
loans.
19
Refe rence s
[1] Bernanke, Ben and Alan S. Blinder, 1988, "Is it Money Or Credit, Or Both Or Neither
? Credit.
Money, and Aggregate Demand", America n Economic Review Papers and Proceedings
, 78,
2,435-439.
[2] Bernanke, Ben and Alan S. Blinder, 1992, "The Federal Funds Rate and the Channe
ls of Monetary Transmission", America n Economic Review,82, 4, 901-921.
[3] Bernanke, Ben and Mark Gertler, 1989, "Agency Costs, Net Worth, and Business Fluctua
tions",
America n Economic Review , LXXIX, 14-31.
[4] Bernanke, Ben, Mark Gertler and Simon Gilchrist, 1993 "The Financial Accelerator
and the
Flight to Quality", Review of Economic Studies (forthcoming).
[5] Bernanke, Ben and Illian Mihov, 1995, "l\leasuring Monetary Policy", NBER Working
Paper
No. 5415.
[6] Blanchard, Olivier J., and Angelo Melino, 1986, "The cyclical behavior of prices and
quantities:
The case of the automobile market.", Journal of Monetary Economics 17, 379- 407.
[7] Blanchard, Olivier, 1993, "Consumption and the Recession of 1990-1991", America
n Economic
Review Papers and Proceedings , 83, 270-274.
[8] Chow, G.C., 1957, Demand for automobiles in the United States , Amsterdam, North
Holland
Publishing company.
[9] Christiano, Lawrence, Martin Eichenbaum, and Charles Evans, 1994, "Identification
and the
Effects of Monetary Policy Shocks", Northwest.em University working paper.
[10] Eichenbaum, Martin, 1992, "Comment. on Interpreting The Macroeconomic Time Series
Facts:
The Effects of Monetary Policy", European Economic Review, 36, June, 1001-1011.
20
[11] Gertler, Mark, and Simon Gilchrist, 1993, "The Role of Credit Market Imperfections in the
Monetary Transmission Mechanism: Arguments and Evidence", Scandinavian Journal of Economics, XCV, 43-64.
[12] Hall, Robert E., 1993, "Macro Theory and the Recession of 1990-1991", American Economic
Review Papers and Proceedings , 83, 275-279.
[13] Kashyap, Anil, and Jeremy C. Stein, and David W. Wilcox, 1993, "Monetary Policy and Credit
Conditions: Evidence from the Composition of External Finance", American Economic Review,
83, 1, 78-98.
[14] Olney, Martha, 1989, "Credit as a Production Smoothing Device: ThE> Case of Automobiles,
1913-1938", The Journal of Economic History, XLIX, 2, 377-391.
[15] Sims, Christopher, 1992, "Interpreting the Macroeconomic Time Series Facts: The Effects of
Monetary Policy", European Economic Review, 36: 975-1000.
[16] Sims, Christopher, and James H. Stock and !\lark V..'. Watson, 1990. "lnfcrcnce in Linear Time
Series Models With Some Unit Roots", Econometrica, 58, 1, 113-144.
21
Bank Credit Relative to Total
0
0
0
0
0
0
0
CJ1
CJ1
CJ1
O')
O')
0
CJ1
---1
---1
co
0
0
CJ1
0
~
....I.
(D
O')
---1
(D
---1
.r:,.
rn
-
('")
0
s::
~
-
0
T
~~
-
s"
<
U)
:j
0
.,.,0 c.c-·
i::::
s
~
)::.
C:
(D
OJ
ci
-
s::
I
0
-rn
en
r-
::!J
. <D
~
-
co co
<
('")
~
~
,,
rn
-:
C'D
....
Figure 1A
Detren ded Mix
0.15
0.10 -
0.05
0.00
-1--t--r--f•--
-0.05
-0.10
-0.15
-----,-i--r --,-,-1-rT I
1948
I
I
I
I
I
I
1958
I
I
I
I
I
I
I
I
I
"1968
I
I
I
I
I
I
I
1· 1 · I
-197s
I
I
I
1--1· I
I
I
'
I
I
"1988
I
I
I
I I
Figure 2
Response of Mix to Fed Funds Rate shock and Two S.E. Bands
0.0050 --·~~··-·-····- - - .... · ..... · .. ··· · .. ··-··
·········--·-----·-········· ······-··- .. ·
, ,
0.0025 ,
'
0.0000 --~-,.c-~·,·,,:-~
I , ''
,
,
,
\
\
--- - - -
....
_,,/----
-----~~
''
-0.0025 ..
,'
\
\
'
\
\
'
-~
'
'
-0.0050 -
''
'
-0.0075
2
·13
2LI
Months
35
46
57
Figure 3
Response of Finance Co. Credit to Fed Funds Rate shock and Two S.E. Bands
2000 ~-------- .... -- - --- ------------
1500 -
1000 /
/
/
500
~~
------
I
O
I
I
I
-----------
','
''
--
,
-500 -
-1000
2
12
22
Months
32
42
52
Figure 4
Response of Bank Credit to Fed Funds Rate shock and Two S.E. Bands
1000 - , - - ~
· · · - ·--·· ·
· ·- - ·~- ··- -·- -··----· -·· · ··----·-· · ·-···--·-· -····-750 -
,
,
,
,
/
/
,
500 -
/
,
,
250 0
,
- - ,,,
,
'
, '
.....
--1---- k-''
-250 -
''
-500 -
-750
-' - '
'
''
'
'
'
,
',
''
I
,
,
,
,
/
/
/
/
-- - - - -
'
---
'
,
,
'
''
-1000 -
-1250 -··
/
..
I
21
·12
22
Months
32
42
52
,
,
Figure 5
Response of Auto Sales to Fed Funds Rate shock and Two S.E. Bands
0.02 ,-----~~----- - ---- --- --- ---------------- -~----~-- - - ----- --------------
, ,
,
0.01 -
,
I
I
/
\
/
' '
I
0.00
I
,
/
I
/
-1-~
/
A
',•----
\
I
/
I
/
, '
I
\
,_
I
-0.01 -
'
' '
'
'
/
I
,,
/
/
,
I
I
I
I
-----✓---
I
,
\
,
-0.02 -
/
I
r
/
I
-...
'
\
I
'
-0.03.,. -
,., .... .,.
'
'
I
I
\
I
\I
-0.04 -• - -
I
2
12
22
l\nonths
32
42
52
Figure 6
Response of Auto Sales to Mix shock and Two S. E. Bands
----· --- - -- ----- --------
------ ------ -·· -- --· -.. - ---
0.032 ---.--
'
✓
0.024 I
\
I
I
I
I
I
\
'
I
I
I
'
,.._,,, '
"
,,
II
0.016 -
'
,- ,
0.008 -
I
I
I
I
I
I
0.000 -1~ -(---
.,
,
,
\
\
-
I
\
'
✓
,_. ,,,
~~
''
I
I \
\
I
\
.,,
\
.....
'
\
\
/
\
,,
'
,,
I
I
\
-0.008 --
' '
' '
-- '
I
,,
I
'
'
'
-0.016
2
-12
22
Months
32
42
52
Figure 7
Response of Inventory-Sales Ratio to Mix shock and Two S.E. Bands
0.04
,,I
,,
I
I I
I I
I
I
0.02
I
I
I
I
I
I
I
I I
,,
I
I
I I
I I
,,
\
I
- I
I
I
I
/
,,
,,
I
-0.02
I
I
I II
,,
"
I I I
,,
I
,
I
\
I
I
-
___
.... _ _ _
/✓-------
/
/
\
/
'
/
I
I /
I
',
I
I/
,,
/
I
I
I \
I I
I \
I
,,
I
I
I
I
I
I I
I I
-0.08
~----/
/',
I
/
r
,,
,
'
- - - - - .... -
' ,, ,
I/
-0.06
,,.
- - - - - - -----/
I
I
-0.04
.- '
I
I ',
I
I
1 ·- !
0.00
, '
\
I
\
I
I -
I
I
I
I
I
I \
I
I
I
I
I
,,
I
I
_,
r
-0.10
2
12
22
Months
32
42
52
Figure 8
Response of Bank-F.C. Ra_te Spread to a Funds Rate shock and Two S.E. Bands
0.5 , ·
- ·-·
· · --- - -------,
,,
I
I
I
0.4
0.3
I
\
I
-- --
I
I
I
I
I
0.2
I
-----
0.1
,,
I
I
I
I
\
I
0.0
\
\
- I
\
I
\
I
I
\
\
....
--~-~
-- -·-·
I
'
I
-- - -
\
\
\
-0.1
-0.2 .
\
'
I
~
--1 ···-· -,.- -,-----·-·1··
0
T
3
I - -
I .
I
6
Quarters
-, -----,- ·--,---r--··, ----,
9
12
Figure 9
Response of F.C. Rate to Funds Rate shock and Two S.E. Bands
0.25 ·
. - - - - - --------------·
'
,
0.20
"
I
I
-
I
--
I
I
I
,
I
I
0. -15
I
,
I
'
,,
\
I
'
I
0.10
' '
I
,
I
'
--------
0.05
r, '.,
'
0.00
, I
I
-----
I
I
--- - -
--- '
,1
I
I
\
I
-0.05
I
I
\
I
\
\
I
I
I
I
I
I
\
,, I
\
I
I
'
I
-0.10
--- '
' '
I
'
-0.15
. -
2
---------
I
- - --1
I
12
22
Months
32
42
-- .
I - ---
52
. --· -·-
Table 1
Automobile Demand Equations
Dependent Variable: S,, 1968:1 to 1994:I I
y
x.,
RP
0.797
(0. I I)
(-0.001)
(0.001)
15.49
(1338)
(J.72)
0.797
(0.096)
0.002
(0.001)
17.672
(2.854)
15.77
(2.022)
0.796
(0.099)
-0.0002
(0.001)
12.288
(3,198)
16.017
(2,013)
0.831
(0.099)
-0.0004
(0.001)
-12.289
(17.64)
25.371
(17.38)
(0.000)
(0.001)
• 10.852
(16.89)
25.501
(16.78)
0.019
(0.005)
-0.012
(0.005)
FC
:!:'...FC..
19.67
18.527
(i.635)
!';.1RP-i
C
1::3;.,C-t
r
-19.46
(5.586)
-19.754
(5.56)
53.619
(23.803)
16.275
(J.706)
-0.002
(0.001)
-8.672
( 16.60)
21.564
(16.52)
0.017
(0.005)
-0.009
(0.005)
53.401
(24.399)
14.572
(2.044)
-0.002
(0.001)
·11.298
(16.75)
22.002
(16.54)
0.015
(0.005)
-0.007
(0.005)
-7.820
(5.646)
Notes: OLS estimation, standard errors in parentheses; S is auto sales, FC is the financing
composition, defined as bank loans divided by total. Y is real, personal disposable income. X is
the consumer's stock of automobiles computed assuming a quarterly depreciation rate of2 %. RP
is the Consumer Price Index for Cars divided by the implicit price deflater for personal
consumption expenditures. C is personal consumption expenditures, and r is the rate on three
month treasury bills.
Table 2
Reduced Form Equations
Dependent Variable: S~ 1968: 1 to 1994: 11
FC
10.8
(0.96)
14.7
(12.4)
10.4
(0.98)
8.2
(1.2)
9.8
(12.3)
7.7
(1.2)
w
X.1
I.1
-0.04
(0.01)
0.00
(0.05)
0.11
(0.01)
6.9
(3.3)
1.5
(3.4)
0.22
(0.03)
-0.12
(0.04)
0.00
(0.01)
-0.10
(0.05)
0.10
(0.0 I)
10.6
(2.8)
-2.5
(2.9)
0.21
(0.03)
-0.08 ·
(0.03)
0.00
(0.0 I)
-0.09
(0.05)
0.09
(0.01)
9.9
(2.8)
-1.7
(2.9)
0.22
(0.03)
-0.08
(0.03)
-0.01
(0.01)
-0.13
(0.05)
0.09
(0.01)
8.5
(2.8)
-1.3
(2.9)
0.20
(0.03)
-0.06
(0.03)
-2.6
(0.73)
-0.02
(0.01)
-0.12
(0.05)
0.09
(0.0 I)
7.7
(2.8)
-0.42
(2.87)
0.20
(0.03)
-0.06
(0.03)
-2.7
(0.72)
Notes: OLS estimation, standard errors in parentheses; S is auto sales, FC is the financing
composition defined as bank loans divided by total. X is the consumer's stock of automobiles
computed assuming a quarterly depreciation rate of 2 %. I is retail automobile inventories, Z is
the index of industrial production for autos, C is personal consumption expenditures, W is the
average hourly wage of production workers in the automobile industry, and PFWL is the
Producer Price index for fuel. Monthly strike dummies, t, t2 also included.
FEDERAL RESERVE BANK OF NEW YORK
RESEARCH PAPERS
1996
The following papers were written by economists at the Federal Reserve Bank of New York either
alone or in collaboration with outside economists. Single copies of up to six papers are available upon
request from the Public Information Department, Federal Reserve Bank of New York, 33 Liberty Street,
New York, NY 10045-0001 (212) 720-6134.
9601. Bartolini, Leonardo, and Gordon M. Bodnar. "Are Exchange Rates Excessively Volatile? And
What Does 'Excessively Volatile' Mean, Anyway?" January 1996.
9602. Lopez, Jose A. "Exchange Rate Cointegration Across Central Bank Regime Shifts."
January 1996.
9603. Wenninger, John, and Daniel Orlow. "Consumer Payments Over Open Computer Networks."
March 1996.
9604. Groshen, Erica L. "American Employer Salary Surveys and Labor Economics Research: Issues
and Contributions." March 1996.
9605.
Uctum, Merih. "European Integration and Asymmetry in the EMS." April 1996.
9606. de Kock, Gabriel S. P., and Tanya E. Ghaleb. "Has the Cost of Fighting Inflation Fallen?"
April 1996.
9607. de Kock, Gabriel S. P., and Tania Nadal-Vicens. "Capacity Utilization-Inflation Linkages: A CrossCountry Analysis." April 1996.
9608. Cantor, Richard, and Frank Packer. "Determinants and Impacts of Sovereign Credit Ratings."
April 1996.
9609. Estrella, Arturo, and Frederic S. Mishkin. "Predicting U.S. Recessions: Financial Variables as
Leading Indicators." May 1996.
9610. Antzoulatos, Angelos A. "Capital Flows and Current Account Deficits in the 1990s: Why Did Latin
American and East Asian Countries Respond Differently?" May 1996.
9611. Locke, Peter R., Asani Sarkar, and Lifan Wu. "Did the Good Guys Lose? Heterogeneous Traders
and Regulatory Restrictions on Dual Trading." May 1996.
9612. Locke, Peter R., and Asani Sarkar. "Volatility and Liquidity in Futures Markets." May 1996.
9613. Gong, Frank F., and Eli M. Remolona. ''Two Factors Along the Yield Curve." June 1996.
9614. Harris, Ethan S., and Clara Vega. "What Do Chain Store Sales Tell Us About Consume
r
Spending?" June 1996.
9615. Uctum, Merih, and Michael Wickens. "Debt and Deficit Ceilings, and Sustainability of
Fiscal
Policies: An lntertemporal Analysis." June 1996.
9616. Uctum, Merih, and Michael Aglietta. "Europe and the Maastricht Challenge." June 1996.
9617.
Laster, David, Paul Bennett, and In Sun Geoum. "Rational Bias in Macroeconomic Forecasts."
July 1996.
9618. Mahoney, James M., Chamu Sundaramurthy, and Joseph T. Mahoney. "The Effects of Corporate
Antitakeover Provisions on Long-Term Investment: Empirical Evidence." July 1996.
9619. Gong, Frank F., and Eli M. Remolona. "A Three-Factor Econometric Model of the U.S.
Term
Structure." July 1996.
9620. Nolle, Daniel E., and Rama Seth. "Do Banks Follow Their Customers Abroad?" July 1996.
9621. Mccarthy, Jonathan, and Charles Steindel. "The Relative Importance of National and Regional
Factors in the New York Metropolitan Economy." July 1996.
9622. Peristiani, S., P. Bennett, G. Monsen, R. Peach, and J. Raiff. "Effects of Househol
d
Creditworthiness on Mortage Refinancings." August 1996.
9623. Peristiani, Stavros. "Do Mergers Improve the X-Efficiency and Scale Efficiency of U.S. Banks?
Evidence from the 1980s." August 1996.
9624. Ludvigson, Sydney. "Consumption and Credit: A Model of Time-Varying Liquidity Constrain
ts."
August 1996.
9625. Ludvigson, Sydney. "The Channel of Monetary Transmission to Demand: Evidence from
the
Market for Automobile Credit." August 1996.
@FedFRASER