Modern Economy, 2012, 3, 542-552
http://dx.doi.org/10.4236/me.2012.35071 Published Online September 2012 (http://www.SciRP.org/journal/me)
Financial Innovation, Macroeconomic Volatility and the
Great Moderation
Lorenzo Bencivelli, Andrea Zaghini
Bank of Italy, Research Department—Servizio Studi di Congiuntura e Politica Monetaria, Rome, Italy
Received April 23, 2012; revised May 29, 2012; accepted June 7, 2012
ABSTRACT
In the paper we propose an assessment of the role of financial innovation in shaping US macroeconomic dynamics. We
extend an existing model by Christiano, Eichenbaum and Evans which studied the transmission of monetary policy im-
pulses to business and corporate sector financing variables just before the Great Moderation period. By investigating the
properties of the model over a longer time span we show that in the later period a change in the monetary policy trans-
mission mechanism is likely to have occurred. In particular, we argue that the role of financial innovation has signify-
cantly altered the transmission of shocks.
Keywords: Great Moderation; Monetary Policy; Financial Innovation
1. Introduction
The Great Moderation period in the US has been broadly
investigated but it is still a matter of lively discussions.
The stylized facts are very simple and clear: the volatility
and the persistence of many macroeconomic variables
(first of all GDP and inflation) declined significantly s in c e
early 1980s. However, the reasons behind this change in
business cycle dynamics are still unclear. The economic
literature has provided three competing bu t non mutually
excluding hypotheses: the “good luck” hypothesis (the
economy was hit from less severe shocks), the “good pol -
icy” hypothesis (improved monetary policy management),
the “changes in the structure” hypothesis (modifications
in the functioning of the economy which have altered the
transmi ssi o n of monetary an d other kind of shocks).
Recently a new branch of the literature has suggested
that financial innovation may have played and important
role in influencing the business cycle dynamics of the US
economy. In particular, changes in firms’ and consumers’
behaviour, induced by significant financial improvements ,
have allowed private sector agents to better cushion them-
selves from th e impac t of in ter est-rate fluctu ation s. Wi t h in
this new framework, our paper analyses the role played
by net funds raised by the business sector. We build on
an existing model proposed by Christiano, Eichenbaum
and Evans in 1996 by extending their sample in order to
include the whole Great Moderation period. Their model
worked well in identifying monetary policy shocks and
describing the interaction among real and financial vari-
ables over a period which includes only few years of the
Great Moderation era. By adding data till 2006 and by
adequately splitting the sample we show: 1) that the mod el
is not able to describe the dynamics of US economy over
the enlarged period; 2) that the transmission of monetary
policy shocks in an early and a late sub-sample differs
significantly. In particular, the role of financial variables
seems to have changed. Even not including data from the
most recent financial crisis, in the second part of the sa m-
ple financial variab les emerge as the channel through which
the shocks pass to the real sector of the economy.
The paper is structured as follows. Section 2 proposes
a survey of the literature p aying particular attention to the
most recent contributions about the role of financial in-
novation; Section 3 introdu ces th e model by Christiano et
al. [1], Section 4 deals with the estimation results over the
two sub-samples; Section 5 provides a robustness analy-
sis; Section 6 concludes.
2. The Economic Literature
Starting from the late 1990s, a large body of the empiri-
cal literature has examined the Great Moderation era in
the US. A survey of the early contributions can be al-
ready found in Stock and Watson [2]1. More recently,
this lively literature has witnessed a acceleration due to
the employment of new econometric techniques. As al-
ready mentioned, there are three main explanations of the
declined volatility of US macroeconomic times series:
1See Bernanke and Mihov [3], Kim and Nelson [4], McConnel and
Perez-Quiros [5], Clarida et al. [6], Blanchard and Simon [7] among the
earl
y
studies.
C
opyright © 2012 SciRes. ME

L. BENCIVELLI, A. ZAGHINI 543
the “good luck”, the “good policy” and the “structural
changes” hypotheses.
The “good luck” hypothesis is based on the assumption
that macroeconomic shocks are drawn from a t i m e - v a r y in g
distribution. Over the Great Moderation years the US
economy was simply hit by less severe shocks and in
particular by smaller common international shocks (Stock
and Watson [8]; Si ms and Zha [9]).
The “good policy” explanation of the declined vola-
tile- ity is that the FED ch anged its monetary policy con-
duct improving its ability to tackle exogenous distur-
bances. By systematically responding more decisively to
fluctuations in economic conditions, a credible monetary
policy has since the early 1980s stabilized inflationary
expectations via commitment to a nominal anchor (Lubik
and Schorfheide [10]; Boivin and Giannoni [11]).
Finally, the “structural changes” hypothesis holds that
various innovations induced by technological progress or
financial innovations, might have altered the transmis-
sion mechanism of shocks as well as monetary policy
impulses allowing the private sector to better withstand
the impact of business cycle fluctuations (Giannone et al.
[12]; Galì and Gambetti [13]).
Without the pretence of being exhaustive, we report in
Table 1 a classification of the empirical contributions on
the Great Moderation with respect to the main empirical
method employed and the hypothesis backed by the results.
One of the first and still br oadly used methodology by
empirical works is the sub-sampling. The properties of
the US economy are investigated separately over two dis-
tinct periods. Even if the econometric techniques are of-
ten different (e.g., IRFs are estimated with VAR/FAVAR
or derived from small scale models) the idea is to use the
business cycle dynamics over a pre-Moderation sample
to test the changes of th e Great Moderation period. There
is a relatively large consensus on the break having oc-
curred in 1984. The early sample usually starts in mid-
late 1950s and ends in late 1983 or in any of the four
quarters of 1984. As for the closing date of the Great
Moderation period, the latest available data is commonly
employed. However, at least as inflation dynamics are
concerned, based on a review of econometric estimates
of trend inflation and surveys on inflation expectations,
Mishkin [14] argues that the process of disinflation and
the re-anchoring of long-term inflation expectations was
completed by the end of the 1990s2. Recently more so-
phisticated econometric instruments were employed to
assess the causes of the Great Moderation. In particular
(structural) time-varying coefficient VAR and time vary-
ing regimes modelling are widely used.
As for the motivations of the declined macroeconomic
volatility, from Table 1 we can see that early work s t e n d e d
to support the “good luck” hypothesis while later studies
point to the “good policy explanation” of the Great Mod-
eration. Whereas, apart from the pioneering work of Kim
and Nelson [26], the contribution pointing to a structural
change in the economic framework are the most recent,
regardless of the econometric technique employed.
Among the contributions which suggest that the struc-
tural changes witnessed by the US economy are the main
cause of the Great Moderation, there are several studies
backing the hypothesis that the change in the financial
system is the mo st important one. In particular, this branch
of the literature looks at several possible links between
the working of financial markets and the real economic
activity.
On the on e h and, th e financial accelerator and the pro-
cyclicality of the premia on credit are seen as the main
theoretical tools to assess the process through which the
financial system is supposed to transmit and amplify
economic fluctuations3. On the other hand, financial in-
novation has been proposed as a possible source of mod-
eration of business cycles. Changes in the way financial
market operates have induced structural adjustments in
firms’ and consumers’ behaviour, which in turn let the
private sector better cushioning itself from the impact of
interest-rate fluctuations and macroeconomic shocks. In
particular, the coincidence of the Great Moderation in
macroeconomic variables with an in crease in the volatile-
ity of many financial variables (financial immoderation)
Table 1. Empirical literature.
Good luck Good policy Structural changes
Sub-samples De Blas [16] Clarida et al. [6]
Lubik Schorfheide [10]
Boivin Giannoni [11 ]
Boivin et al. [17]
Canova Gambetti [18]
Gilchrist et al. [19]
Time-varying coefficients Stock Watson [8]
Prim iceri [5]
Justiniano Primiceri [20] Cecchetti et al. [21] Canova Gambetti [22]
Galì Gambetti [13]
Time-varying regimes McConnell Perez [5]
Sims Zha [9]
Korenok Radchenko [23] Kim et al. [24] Galvao Marcellino [25]
Kim Nelson [4]
2For the welfare implication of the declined inflation rate over the Great Moderation period in the US se e Calza and Zaghini [15].
3About the financial accel erator transm ission mechanism see, for instance, Gertler and Lown [26].

L. BENCIVELLI, A. ZAGHINI
544
has pushed the profession looking for possible explana-
tions of the phenomenon. The link proposed goes indeed
through the financial innovation. The underlying intuition
is that transformations occurred in the financial market
(financial innovation) have turned to opportunities for f i rm s
and households to smooth their investment and consump-
tion plans, with the result that economic agents exploited
more the financial instruments (financial immoderation),
but the fluctuations in th e main macroeconomi c aggreg at es
have moderated considerably (macroeconomic modera-
tion). For instance, a change usually suggested by the
literature to support the fin ancial innovation hypothesis is
the democratization of the market (Dynan et al. [27]), i.e.
the agents participation to the market trading without
intervention of institutional intermediaries brought about
by newly developed technologies4.
Of particular interest is the empirical approach fol-
lowed by Jermann and Quadrini [28], Gilchrist et al. [19]
and Fuentes-Albero [29]. They start from the well known
DSGE model for the US economy proposed by Smets
and Wouters [30] and introduce some financial market
frictions to account for spillovers from the financial sys-
tem to the real activity. Following the line of research
pioneered by Kiyotaki and Moore [31] and Bernanke et
al. [32] which show that frictions in the credit market
introduce a transmission mechanism which magnifies
business fluctuations, they try to prove that the process
of convergence toward a better functioning and almost
frictionless financial market can yield more moderate
fluctuations in the real variables, i.e. be the cause of the
Great Moderation5.
According to Gilchrist et al. [21], the in troduction of a
financial accelerator mechanism à la Bernanke et al. [1]
in the Smets and Wouters [34] framework , under the
hypothesis that markets are imperfect, drives a wedge
between expected return on capital and expected return
demanded by the households (the lender). The authors
show that over the period 1973-2008 this mechanism
affects significantly business cycle fluctuations. In par-
ticular, an increase in external finance premium causes
significant and protracted decline in investment spending
and in output. Furthermore, Gilchrist et al. find that the
financial stress is partly responsible for the sharp drop in
output and investments spending in the mid Seventies.
Conversely, the financial easing of the late Nineties pro-
vided a significant impetus to investments.
Looking at Flow-of-Funds data Jermann and Quadrini
[28] document the increase in volatility of firms financial
flows. Specifically, the flows of debt and equity finance-
ing in the business sector displayed much greater vari-
ability from the second half of the 1980s. Because debt
financing is negatively correlated to equity financing, t hes e
findings suggest that firms have become more flexible in
the choice of the financial structure. In their model the
driving forces of business cycles are productivity and c re di t
shocks. The former is the standard productivity shock as
in the typical real business cycle model. The latter is a
shock that affects the enforcement of debt contracts, and
therefore, the borrowing ability of firms (credit shock).
Because of financial frictions, credit shocks are transmit-
ted to the real sector of the economy through the effect
they have on the production and investment decisions of
firms. They show that credit shocks do contribute non-
negligibly to the volatility of the major macroeconomic
variables in the first sample period (1952-1983). In addi-
tion, they find that financial innovations can account for
a large decline in real macroeconomic volatility and they
can easily account for the full increase in the volatility of
the financial structure of firms in the second sample pe-
riod (1984-2006 ).
The paper by Fuentes-Albero [29] tries to reconcile the
two empirical facts of the great moderation of macro-
economic variables and the great immoderation of finan-
cial variables on the path traced by Jermann and Quadrini
[28] and Gilchrist et al. [19]. The goal being that of quan -
tifying the relative role played by financial factors in
shaping macroeconomic volatilities. The baseline model
is again the one by Smets and Wouter [30] enriched with
a financial accelerator mechanism (financial frictions).
Differently from Gilchrist et al. [19] the author allows
for two different financial shocks, one accounting for
the balance sheet channel, the other for the information
channel. The main empirical finding is a redu ction in the
average level of financial rigidities in the second sample
(1984-2006). In particular, the estimated reduction in the
size of the financial accelerator has two effects. On the
one hand, it allows the model to account for 30% of the
slowdown in the volatility o f investment and th e nominal
interest rate. On the other hand, a smaller level of finan-
cial rigidity changes the propagation mechanism of fi-
nancial shocks to the economy.
3. A Benchmark Model
In order to assess whether a change has occurred in the
monetary policy transmission mechanism which is con-
sistent with the Great Moderation timing and, in particu-
lar, if it can be related to some changes in th e way private
sector reacts to unexpected shocks, we build on the 1996
4Other transformations often quoted are the higher efficiency and speed
of the spreading of information; the quick expansion of the market for
high risk debt that has enlarged the participation to the market; the
p
hasing out of the Regulation Q, which imposed a ceiling on the inter-
est rate on deposits, with the consequence that when market rates were
to rise above this level, funds were no longer available for the lenders,
reducin
g
the amount of re so urces for borrowers.
5For a different view on the role of the US financial structure see Den
Haan and Sterk [33] who challenge the empirical evidence aboutfi na nc ia l
innovation as a possible explanatory factor of the Great Moderation.

L. BENCIVELLI, A. ZAGHINI 545
work by Christiano, Eichenbaum and Evans (CEE hence-
forth). The CEE paper is a very good starting point for
two reasons: 1) it provides an empirical framework whi ch
worked well in identifying FED monetary policy impulses;
2) it assesses the implication of a monetary policy shock
on business sector variables. In our empirical approach
we mostly focus on the analysis of the impulse response
functions (IRFs) derived from their model over different
periods. We compare the IRFs over different time spans
because there are several behavioural patterns that are
almost unanimously acknowledged to closely describe
the macroeconomic implications of a monetary policy
shock. As put by Christiano et al. [34], every model which
deals with the FED monetary policy ought to reproduce
these well measured and well accepted effects of US
monetary policy shocks. Thus if we find over a given
period that the IRFs describe different patterns with re-
spect to standard behaviours, this would suggest that the
model is no more able to describe the short- to medium-
run dynamics of the US economy. In turn this would
suggest a structural change in the fundamental working
of the US economy or more simply in a different reaction
of the economy to a monetary policy shock.
In the work of 1996, Christiano and co-authors use
two measures of monetary policy shocks (orthogonalized
shock to Fed funds rate and orthogonalized shocks to non-
borrowed reserves) in conjunction with Flow-of-Funds
data to assess the impact of monetary policy on the bor-
rowing and lending activities of different sectors of the
economy. Our first step is to check whether the implica-
tions of the CEE model for the business (and corporate)
sector are still valid over an enlarged sample which in-
cludes the whole Great Moderation period.
The CEE benchmark model is made of six variables
which enter the VAR in the following order over the pe-
riod 1960:Q1-1992:Q4: GDP, GDP deflator, a commodi ty
price index (PCOM), non-borrowed reserves (NBRD),
the Fed Funds rate (FFR), total reserves (TR). When the
Fed Funds rate is specified as the monetary policy in-
strument, the ordering of the variables in the model and
the Cholesky decomposition approach imply that the reac-
tion function of the FED is such that when deciding ab ou t
the optimal interest rate, not o nly the GDP in the curren t
quarter but also the price levels and the commo dity price
index can b e observed.
The first part of the CEE work is entirely devoted to
the assessment of the business cycle properties of the
6-variable VAR. In additio n, the authors verify the v alid-
ity of the model by adding one at time a variable whose
reaction to monetary policy has to be tested. The results
of the empir ical investig ation are in line with the exp e ct ed
textbook macroeconomic behaviours: a contractionary
shock is associated with a decline in GDP, employment,
retail sales, nonfinacial corporate profits and with an in-
crease in unemployment and inventories. The GDP price
deflator declines after two years.
In order to look at different sectors of the economy, in
the second part of their work, the authors, following the
same methodology, add a seventh variable taken from
the FoF accounts to the benchmark VAR. Their main
result is that after a contractionary monetary policy shock
net funds raised by the business sector rise for one year
(2 to 4 quarters). One possible explanation put forward
by the authors is that it is difficult for firms to quickly
alter their nomin al exp end itur es. Und er these circu mst ances ,
if a contractionary monetary policy shock leads to a fall
in firms’ receipts at the beginning of a recession and a
fall in net cash flow, say because of a fall in sales and a
rise in inventories, then we would expect their net de-
mand for funds to rise. According to this scenario, the
observed eventual decline in net funds raised by firms
reflects their ability to gradually reduce their nominal
expenditures.
As a first step of our empirical investigation, we rep li-
cate the CEE model over the original time span (1960-
1992) with the current data availability. Given the sig-
nificant data revisions occurred after the publication of
the paper (especially for the Flow-of-Funds data) it is
worth checking whether the main conclusions are still
valid. We use an increase in the Fed Funds rate as the
contractionary monetary policy shock. Consistently with
the original results, the shock determines a decline in
GDP which reaches the maximum intensity after 7-8 lags,
a decrease in the price index, which becomes significant
after 2 years, a decline in non-borrowed reserves and a
negligible effect at impact on total reserves6.
We also check the response of net funds raised by the
business sector (BNET) and by the corporate sector (CNET)
to a monetary policy shock. They are i n line with t he origin-
nal model: a rise in th e Fed Fund s rate leads to an inc r e a se
in BNET and CNET which is significant for the first 4
quarters. Again the interpretation suggested by Chris-
tiano and co-authors is still valid after the data revision.
The following step is to run the VAR model over the
entire sample 1960-2006. Figure 1 shows that while the
response of non-borrowed reserves and total reserves seems
to unfold in the expected way, it turns out that a contrac-
tionary monetary policy shock has no effect on prices, in
addition the negative reaction of GDP appear to be per-
manent. These two responses clearly suggest that the
model is not able to capture the dynamics of US econ-
omy over the whole period, in turn this is most likely due
to the effect of the Great Moderation. Our point is that
with original data till 1992 the model was still function-
ing reasonably well since just a limited number of years
within the Great Moderation was covered by the sample.
6Impulse response functions available upon request. Note that the
variable for non-
b
orrowed reserves (NBRD) enters the VAR with a
negative sign. We did so for consistency with CEE variables’ defini-
tion.

L. BENCIVELLI, A. ZAGHINI
ME
546
Figure 3 reports the impulse response functions to a
contractionary monetary policy shock for the early sam-
ple. As expected the dynamics of US economy after an
increase in the Fed Funds rate are well described by the
6-variable VAR proposed by Christiano et al. (1996),
since just few years are left out with respect to the origin-
nal sample. The monetary policy shock determines a
temporary decrease in GDP, a persistent medium-run
decline in prices (even though there is evidence of an
initial price-puzzle) a decline in non-borrowed reserves
and no effect total reserves.
Once the period is enlarged it fails to deliver the same
good results.
In order to check whether a change in the monetary
policy transmission mechanism has occurred (in particu-
lar with reference to the business sector variables), in the
next section we split the sample into an early period,
which ends before the Great Moderation, and a late pe-
riod, which includes the whole Great Moderation era.
4. The Transmission of Shocks over Time
In this section we analyse and compare the IRFs derived
from the CEE model over two sub-samples. In particular,
following the broad consen sus in the empirical literature,
we set the break date and thus the start of the “late pe-
riod” in the first quarter of 1984.
We then look at the effect of the shock in the second
half of the sample which includes the whole Great Mod-
eration period (Figure 4). The striking result is the ab-
sence of a significant response by GDP. Even though the
shock is by far smaller in the second sample than in the
first one (around one third), the rest of the IRFs are in
line with the common knowledge, only GDP dynamics
are indeed puzzling. The monetary policy intervention ha s
notwithstanding achieved the alleged target of a decline
in infl ation. It ju st seems that th e standard Keyn esian ch a n-
nel of monetary policy transmission which goes through
a decline in GDP (via a crowding out of consumption
durables and investment) is completely absent in the mos t
recent period.
The monetary policy shocks obtained from the model
estimated over the whole sample (1960Q1-2006Q4),
the early period (1960Q1-1983Q4) and the late period
(1984Q1-2006Q4) are reported in Figure 2. While the
estimated shocks over the early sample appear to follow
closely the pattern derived from the whole period, those
of the late period show a much smaller size and different
dynamics, especially during recessions. At first sight it
thus seems that the exogenous monetary policy impulses
can not be compared wit hin a unique f ramework.
Figure 1. Monetary policy shock (1960-2006)—CEE model.

L. BENCIVELLI, A. ZAGHINI 547
-3
-2
-1
0
1
2
3
4
5
60-1 62-1 64-1 66-1 68-1 70-1 72-1 74-1 76-1 78-1 80-1 82-1 84-1 86-1 88-1 90-1 92-1 94-1 96-1 98-1 00-1 02-1 04-1 06-1
CEE 1960-2006
Sample 1960-1983
Sample 1984-2006
Figure 2. Monetary policy shock (1960-2006).
Figure 3. IFRs to a monetary policy shock (1960-1983)—CEE model.
A possible suggestion of the reason behind this circu m-
stance can be found by the response of BNET (the net
funds raised by the business sector) to a Fed Funds rate
shocks. In the early period Figure 5 shows that after the
initial increase th e business variable declines significa n tl y,
while this is not true for the later period. This evidence
suggests that a possible “business channel” of the mone-
tary policy transmission that was working in the first part
of the sample is not working in the second. In the early
sample, an increase in the Fed Funds rate had an impact

L. BENCIVELLI, A. ZAGHINI
548
on firms’ financing condition which in turn, together
with other channels, affected GDP dynamics. The effect
is the one documented by Christiano et al. [1]: after a con-
tractionary monetary policy shock net funds raised by
business sector do not immediately decline because it is
difficult for firms to quickly adjust nominal expenditure,
however after few quarters firms are able to reduce their
financing needs by curbing expenditures and thus push-
ing GDP downwards.
In the most recent sample this channel seems to have
disappeared. The finding is consistent with the bran ch of
empirical literature which suggests that financial innova-
tion has induced a structural break able to mitigate busi-
ness cycle fluctuations in a way coherent with the Great
Moderation evidence. Changes in the business sector be-
haviour, induced by the sustained financial innovation of
the last two/three decades, have allowed the private sec-
tor to insulate itself from the impact of interest rate fluc-
tuations.
We also investigate the possibility that a change has
happened in the transmission of the financial shock. Fig-
ure 6 shows in the left hand side panel that in the early
period a shock to the business sector variable has no ef-
fects on real economy: GDP is unchanged, as well as prices.
Also the interest rate does not react, suggesting that finan-
cial variables were not a target of the FED policy. The
scenario changes in the second sub-sample (right hand
side panel). In fact, in the most recent period there is a
negative effect of the financial shock on real GDP, the
reaction being statistically significant and persistent.
The existence of a channel through which a financial
shock hits the real economy is even more relevant given
that we did not include in our sample the latest financial
crisis (one in which the propagation of financial turmoil
to the real economy is commonly agreed to have been
extremely strong). While firms are able to adjust to an
interest rate shock so th at the monetary policy impulse is
not transmitted to the real economy, a shock on the fi-
nancing conditi on has a direct effect on GDP .
5. Model Analysis and Robustness
In this section we propose an analysis of the con tribution
of monetary and financial shock s to the volatilit y of GDP
and prices. In addition we also present a robustness ch eck
of the results of the previous section .
Table 2 shows the contribution of a monetary policy
shock to the volatility of GDP and prices over the two
samples. In the early sample (upper panel), the percentage
of the forecast-error variance attributed to the Fed Funds
shock reaches 20 per cent after 2 years and henceforth is
stable around that share. The value is slightly smaller tha n
Figure 4. IRFs to a monetary policy shock (1984-2006)—CEE model.

L. BENCIVELLI, A. ZAGHINI 549
Figure 5. Effect on BNET of a monetary policy shock—CEE model.
Figure 6. Responses to a financial shock—CEE model.
Table 2. Monetary policy shock contribution to volatility of GDP and pr ic es.
Early sample
4 8 12 16 20 24
11.67 19.43 20.21 19.68 19.38 19.39
GDP
6.53 8.13 9.01 9.52 9.76 9.89
4.88 5.23 10.18 14.77 17.24 17.97
GDPDEF 4.49 5.90 9.12 11.46 12.33 12.53
Late sample
4 8 12 16 20 24
0.15 0.12 0.10 0.12 0.13 0.12 GDP
20.9 3.73 4.63 5.16 5.64 6.02
1.98 9.03 16.87 23.00 26.57 27.89
GDPDEF 3.21 7.71 10.87 12.32 12.73 12.57
the 29 percent after 24 quarters reported by CEE, and in
line with the findings of the empirical literature (Jang and
Ogaki [35]; Uhlig [36]; Canova and Gambetti [22]). When
the model is estimated from 1984 to 2006 (lower panel),
the contribution of the Fed Funds shock however falls to
zero at any horizon, thus confirming that monetary policy

L. BENCIVELLI, A. ZAGHINI
550
has practically no effect on real GDP in the late sample.
Even though stronger in the later sample, the contribution
of the monetary policy shock to prices’ volatility is in-
stead comparable in the two periods.
The evidence is reversed when we look at the percent-
age of the forecast-error variance attributed to the finan-
cial shock (Table 3). The most important contribution is
found in the later sample (lower panel), in which the
share of the volatility attributable to the shock is 10 per
cent after two years and grows to 19 per cent after 24
quarters. There is instead a limited impact on the vola-
tile- ity of prices in both samples.
In order to test the robustness of our results, we em-
ploy a model specification different from that of Chris-
tiano et al. [1] used in Sections 3 and 4. In particular, we
rely on a VAR model which includes the monetary ag-
gregate M2 but not non-borrowed and total reserves. We
thus introduce explicitly a money supply and a money
demand relation in the model7. The ordering of the vari-
ables in the VAR is th e fo llo w ing : GDP, GDP_d ef, PC OM,
FFR, M2.
The IRFs for the sample before the Great Moderation
(1960Q1-1983Q4) and for the sample including the Great
Moderation (1984Q1-2006Q4) show that an exogenous
increase in the Fed Funds rate determines the expected
reaction in all macroeconomic variables, including a de-
cline in M2, only in the early sample. In the late period
GDP seems again to be non affected by the monetary pol -
icy shock8. A persistent reduction in inflation is achiev ed,
but not through a GDP decline.
As in the previous section we then focus on the effects
of a financial shock: the differences in the two periods
are again significant (Figure 7). A financial shock has no
effect on GDP in the early sample (upper panel) whereas
it directly affects GDP in the most recent period (lower
panel). Thus the possibility that financial innovation has
contributed to a change in the transmission of both mo ne-
tary and financial shocks is confirmed also when looking
at a different specification of the US economy.
6. Conclusions
The paper has provided evidence of a change in the reac-
tion of the US macroeconomic variables to monetary and
financial shocks. Our findings are consistent with a broad
literature suggesting that financial innovation is at least
an important contributor to the smoothed business cycle
fluctuations in the Eighties and Nineties, period labeled
as the Great Moderation.
We started from an existing model by Christiano et al.
[1] and we show that their model which was functioning
relatively well over a period up to 1992, is not able to
deliver the same good results over a longer time span
which include all the Great Moderation period. We then
analyse separately the responses to monetary and finan-
cial shocks into two sub-samples: one ending before the
presumed start of the Great Moderation, and one include-
ing the whole Great Moderation era.
Our findings can be summarized as fo llows. First, m o n e -
tary policy impulses have had in the later period a much
weaker effect on GDP dynamics. In particular, a possible
“business channel” of the monetary policy transmission
mechanism that was working in the period before the
Great Moderation has stopped working in the most recent
period. The vigorous financial innovation of the last few
decades has most likely induced a change in the business
sector behaviour, allowing households and firms to insula te
themselves from the impact of interest-rate fluctuations.
Table 3. Financial shock contribution to volatility of GDP and prices.
Early sample
4 8 12 16 20 24
1.57 2.14 1.66 5.87 10.96 11.64
GDP
2.91 3.71 3.36 5.27 6.78 7.03
16.0 7.94 14.47 15.28 9.86 8.97
GDPDEF 2.77 7.55 11.30 11.85 10.06 9.28
Late sample
4 8 12 16 20 24
5.39 10.59 13.45 15.98 17.89 18.95
GDP
5.29 9.18 10.74 11.36 11.59 11.55
0.13 0.17 0.89 1.60 1.74 1.61
GDPDEF 1.94 3.75 5.64 6.52 6.74 6.65
7There is a broad literature supporting the introduction of a monetary aggregate in the VAR identification; see for instance Kim [37], Leeper et al.
[23], Sims and Zha [9], Boivin et al. [17].
8Results not shown but available u pon request.

L. BENCIVELLI, A. ZAGHINI 551
Figure 7. Financial shock (model 2): upper panel 1960-1983, lower panel 1984-2006.
Second, we documented a change also in the transmis-
sion of a financial shock. In the early period the financial
shock was not transmitted to the real side of th e economy,
while it significantly affects GDP dynamics in the later
sample. This evidence is even more relevant given that
our sample ends in 2006 thus not including the financial
turmoil started in the summer 2007 which had severe
spillovers on the real economy.
7. Acknowledgements
The authors would like to thank an anonymous referee
and M. Cecioni, G. Grande, S. Neri, R. Sabbatini and F.
Zollino for comments an d helpfu l sugg estions. The v iews
expressed in the paper do not necessarily reflect those of
the Bank of Italy.
REFERENCES
[1] L. J. Christiano, M. Eichenbaum and C. Evans, “The
Effects of Monetary Policy Shocks: Evidence from the
Flow of Funds,” Review of Economics and Statistics, Vol.
78, No. 1, 1996, pp. 16-34.
[2] J. Stock and M. W. Watson, “Has the Business Cycle
Changed and Why?” NBER Macroeconomics Annual,
MIT Press, Cambridge, 2002.
[3] B. S. Bernanke and I. Mihov, “Measuring Monetary Pol-
icy,” Quarterly Journal of Economics, Vol. 113, No. 3,
1998, pp. 869-902.
[4] C. J. Kim and C. R. Nelson, “Has the US Economy Be-
come More Stable? A Bayesian Approach Based on a
Markov-Switching Model of the Business Cycle,” The
Review of Economics and Statistics, Vol. 81, No. 4, 1999,
pp. 608-616.
[5] M. M. McConell and G. Perez-Quiros, “Output Fluctua-
tions in the United States: What Has Changed Since the
Early 1980s?” American Economic Review, Vol. 90, No.
5, 2000, pp. 1464-1476. doi:10.1257/aer.90.5.1464
[6] R. Clarida, J. Galí and M. Gertler, “Monetary Policy
Rules and Macroeconomic Stability: Evidence and Some
Theory,” Quarterly Journal of Economics, Vol. 105, No.
1, 2000, pp. 147-180.
[7] O. Blanchard and J. Simon, “The Long and Large Decline
in Output Volatility,” Brookings Papers on Economic Ac-
tivity, Vol. 32, No. 1, 2001, pp. 135-164.
[8] J. Stock and M. W. Watson, “Has the Business Cycle
Changed? Evidence and Explanations,” Monetary Policy
and Uncertainty, Federal Reserve Bank of Kansas City,
Kansas City, 2003, pp. 9-56.
[9] C. Sims and T. Zha, “Were There Regime Switches in US
Monetary Policy?” American Economic Review, Vol. 96,
No. 1, 2006, pp. 54-81.
doi:10.1257/000282806776157678

L. BENCIVELLI, A. ZAGHINI
552
[10] T. Lubik and F. Schorfheide, “Testing for Indeterminacy:
An Application to US Monetary Policy,” American Eco-
nomic Review, Vol. 94, No. 1, 2004, pp. 190-217.
doi:10.1257/000282804322970760
[11] J. Boivin and M. Giannoni, “Has Monetary Policy Be-
come More Effective?” Review of Economics and Statis-
tics, Vol. 88, No. 3, 2006 pp. 445-462.
doi:10.1162/rest.88.3.445
[12] D. Giannone, M. Lenza and L. Reichlin, “Explaining the
Great Moderation: It Is Not the Shocks,” Journal of the
European Economic Association, Vol. 6, No. 2-3, 2008,
pp. 621-633. doi:10.1162/JEEA.2008.6.2-3.621
[13] J. Gali and L. Gambetti, “On the Sources of the Great
Moderation,” American Economic Journal: Macroeco-
nomics, Vol. 1, 2009, pp. 26-57.
[14] F. S. Mishkin, “Inflation Dynamics,” International Fi-
nance, Vol. 10, No. 3, 2007, pp. 317-334.
doi:10.1111/j.1468-2362.2007.00205.x
[15] A. Calza and A. Zaghini, “Sectoral Money Demand and
the Great Disinflation in the US,” Journal of Money, Credit
and Banking, Vol. 42, No. 8, 2010, pp. 1663-1678.
[16] B. De Blas, “Can Financial Frictions Help Explain the
Performance of the US Fed?” The B.E. Journal of Mac-
roeconomics, Vol. 9, No. 1, 2009, 27 Pages.
[17] J. Boivin, M. T. Kiley and F. S. Mishkin, “How Has the
Monetary Transmission Mechanism Evolved over Time?”
NBER Working Paper 15879, 2010.
[18] F. Canova and L. Gambetti, “Structural Changes in the
US Economy: Is There a Role for Monetary Policy? ” Jour-
nal of Economic Dynamics and Control, Vol. 33, No. 2,
2009, pp. 477-490. doi:10.1016/j.jedc.2008.05.010
[19] S. Gilchrist, A. Ortiz and E. Zakrajšek, “Credit Risk and
the Macroeconomy: Evidence from an Estimated DSGE
Model,” Unpublished Manuscript, May 2009.
[20] A. Justiniano and G. Primiceri, “The Time Varying Vola-
tility of Macroeconomic Fluctuations,” American Eco-
nomic Review, Vol. 98, No. 3, 2006, pp. 604-641.
[21] S. G. Cecchetti, A. Flores-Lagunes and S. Krause, “Has
Monetary Policy Become More Efficient? A Cross-Coun-
try Analysis,” Economic Journal, Vol. 116, No. 511, 2006,
pp. 408-433. doi:10.1111/j.1468-0297.2006.01086.x
[22] F. Canova and L. Gambetti, “Do Expectations Matter?
The Great Modera tion Revisited,” American Economic Jour-
nal: Macroeconomics, Vol. 2, No. 3, 2010, No. 3, pp. 183-
205.
[23] O. Korenok and S. Radchenko, “The Role of Permanent
and Transitory Components in Business Cycle Volatility
Moderation,” Empirical Economics, Vol. 31, No. 1, 2006,
pp. 217-241. doi:10.1007/s00181-005-0042-5
[24] C. J. Kim, C. R. Nelson and J. Piger, “The Less-Volatile
US Economy: A Bayesian Investigation of Timing, Br-
eadth, and Potential Explanations,” Journal of Business &
Economic Statistics, Vol. 22, No. 1, 2004, pp. 80-93.
[25] A. B. Galvao and M. Marcellino, “Endogenous Monetary
Policy Regimes and the Great Moderation,” The Centre
for Economic Policy Research 2010, Discussion Paper
No. 7827.
[26] M. Gertler and C. S. Lown, “The Information in the
High-Yield Bond Spread for the Business Cycle: Evi-
dence and Some Implications,” Oxford Review on Eco-
nomic Policy, Vol. 15, No. 3, 1999, pp. 132-150.
[27] K. E. Dynan, D. W. Elmendorf and D. E. Sichel, “Can
Financial Innovation Help to Explain the Reduced Vola-
tility of Economic Activity,” Journal of Monetary Eco-
nomics, Vol. 53, No. 1, 2006.
[28] U. Jermann and V. Quadrini, “Macroeconomic Effects of
Financial Shocks,” American Economic Review, Vol. 102,
No.1, 2012, pp. 238-271.
[29] C. Fuentes-Albero, “Financial Frictions, the Financial
Immoderation, and the Great Moderation,” Mimeo, Uni-
versity of Pennsylvania, 2009.
[30] F. Smets and R. Wouters, “Shocks and Frictions in US
Business Cycles: A Bayesian DSGE Approach,” Ameri-
can Economic Review, Vol. 97, No. 3, 2007, pp. 586-606.
doi:10.1257/aer.97.3.586
[31] M. Kiyotaki and J. Moore, “Credit Cycles,” Journal of
Political Economy, Vol. 105, No. 2, 1997, pp. 211-248.
[32] B. S. Bernanke, M. Gertler and S. Gilchrist, “The Finan-
cial Accelerator in a Quantitative Business Cycle Model,”
In: J. B. Taylor and M. Woodford, Eds., The Handbook of
Macroeconomics, Elsevier, Amsterdam, 1999, pp. 1341-
1393.
[33] W. Den Haan and V. Sterk, “The Myth of Financial In-
novation and the Great Moderation,” Economic Journal,
Vol. 121, No. 553, 2011, pp. 707-739.
doi:10.1111/j.1468-0297.2010.02400.x
[34] L. Christiano, M. Eichenbaum and C. Evans, “Monetary
Policy Shocks: What Ha ve we Learned and to What End?”
In: J. Taylor and M. Woodford, Eds., Handbook of Mac-
roeconomics, North Holland, 1999, pp. 65-148.
[35] K. Jang and M. Ogaki, “The Effects of Monetary Policy
Shocks on Exchange Rate s: A Str uctural Vector Error Cor -
rection Model Approach,” Journal of the Japanese and
International Economies, Vol. 18, No. 1, 2004, pp. 99-
114. doi:10.1016/S0889-1583(03)00042-X
[36] H. Uhlig, “What Are the Effects of Monetary Policy on
Output? Results from an Agnostic Identification Proce-
dure,” Journal of Monetary Economics, Vol. 52, No. 2,
2005, pp. 381-419. doi:10.1016/j.jmoneco.2004.05.007
[37] S. Kim, “Do Monetary Policy Shocks Matter in the G-7
Countries? Using Common Identifying Assumptions
about Monetary Policy across Countries,” Journal of In-
ternational Economics, Vol. 48, No. 2, 1999, pp. 387-412.
doi:10.1016/S0022-1996(98)00052-X