Embed Size (px)
Citation preview
Direction des Études et Synthèses Économiques
G 2016 / 05
MELEZE: A DSGE model for France within the Euro Area
Benoît CAMPAGNE et Aurélien POISSONNIER
Document de travail
Institut National de la Statistique et des Études Économiques
INSTITUT NATIONAL DE LA STATISTIQUE ET DES ÉTUDES ÉCONOMIQUES
Série des documents de travail de la Direction des Études et Synthèses Économiques
JUILLET 2016
The authors would like to thank Jean-Guillaume SAHUC for his fruitful discussion and advice on a first version of this paper, as well as all participants to the D2E seminar at the Insee. Model codes are available upon request.
_____________________________________________
* Département des Études Économiques - Division « Études Macroéconomiques » Timbre G220 - 15, bd Gabriel Péri - BP 100 - 92244 MALAKOFF CEDEX Crest - LMA
** Commission européenne. L’auteur était en poste à l’Insee et affilié au Crest-LMA et École Polytechnique au moment de la rédaction de ce document.
Département des Études Économiques - Timbre G201 - 15, bd Gabriel Péri - BP 100 - 92244 MALAKOFF CEDEX - France - Tél. : 33 (1) 41 17 60 68 - Fax : 33 (1) 41 17 60 45 - CEDEX - E-mail : [email protected] - Site Web Insee : http://www.insee.fr
Ces documents de travail ne reflètent pas la position de l’Insee et n'engagent que leurs auteurs. Working papers do not reflect the position of INSEE but only their author's views.
G 2016 / 05
MELEZE: A DSGE model for France within the Euro Area
Benoît CAMPAGNE* et Aurélien POISSONNIER**
2
Mélèze : une modélisation DSGE de la France au sein de la zone euro
Résumé
Mélèze (Modèle économique linéarisé d’équilibre en zone euro) est un modèle néo-Keynésien de type DSGE avec les caractéristiques suivantes : la France et le reste de la zone euro forment une union monétaire ; une fraction des ménages est non-ricardienne et consomme son revenu courant ; les entreprises sont en concurrence monopolistique sur le marché des biens, et les travailleurs le sont sur le marché du travail ; les prix et salaires sont rigides ; les biens de consommation et d’investissement sont exportés/importés librement, tandis que le travail et le capital sont immobiles.
Cet article détaille la résolution et la calibration du modèle ainsi que sa linéarisation. En particulier, nous définissons l’état stationnaire du modèle en niveau et nous explicitons l’ensemble des contraintes de long-terme pesant sur la calibration du modèle. Dans un deuxième temps, nous présentons le comportement du modèle en réponse à divers chocs transitoires.
Mots-clés : modèle DSGE, union monétaire
MELEZE: A DSGE model for France within the Euro Area
Abstract
MELEZE, standing for Modèle économique linéarisé d’équilibre en zone euro (linearised economic model of equilibrium in the euro area), is a new Keynesian DSGE model with the following characteristics: France and the rest of the Euro area form a monetary union; they are populated by infinitely lived households, of which a constant fraction is non Ricardian, consuming all of their current income; firms operate in monopolistic competition on the goods market, and so do workers on the labour market indistinctly of their financial constraints; prices and wages are sticky; consumption and investment goods can be freely exported/imported, whereas workers and installed capital cannot.
The present paper presents the resolution and the calibration of the model as well as its full linearisation. In particular, we characterize the unique steady state in levels for the real variables and explicit the induced constraints on the parametrisation. In a second part, we present the behaviour of our model with respect to standard transitory shocks.
Keywords: DSGE model, monetary union
Classification JEL : E10, F45
Contents
1 Introduction 4
2 Model 5
2.1 Goods and labour aggregations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Fiscal Authorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 Financial Intermediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.6 Monetary Authority, Prices and Inflation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.7 Market Clearing and relative prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 Calibration 30
3.1 Classification of parameters and resolution of the steady state . . . . . . . . . . . . . . . . . 30
3.2 Data, calibration and inverse inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4 Model dynamics 36
4.1 IRFs to standard shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5 Discussion of the model specification 42
5.1 Non Ricardian households, Edgeworth complementarity/substitutability of private and
public spending and their effect on consumption . . . . . . . . . . . . . . . . . . . . . . . . 42
5.2 Government spending in the utility function compared to budget rules . . . . . . . . . . . 43
A Steady state 52
B Linearisation 62
C Monetary policy in a currency union 72
D Firms and households dispersion 74
E Determination of a unique and stable steady state 76
F Other IRFs 79
3
1 Introduction
Economic policy analysis at the Insee is traditionally focused on the use of large-scale macroeconometric
models namely Mésange (Klein and Simon, 2010). The Mésange model constructed on a highly detailed
macroeconomic accounting framework in line with Quarterly National Accounts, features around 500
equations, 10% being behavioural econometric relationships. Its core long term economic structure is
neo-classical, whereas the model includes real and nominal rigidities in the short run. One main ad-
vantage of this model is its ability to replicate past observed data and to give precise and quantitative
insights for economic policy evaluations, mainly focused on fiscal evaluations.
However, Lucas’ critique (Lucas, 1976) states that the absence of a strong short-run economic struc-
ture and the absence of rational expectations might be weaknesses of such models. As such, economic
modelling has focused on alternatives and developed a new modelling stream, namely the DSGE lit-
erature. Contrary to macroeconometric models, these models focus less on giving an exact description
of observed economic data than on a strong theoretical and micro-founded structure answering Lucas’
critique. As such, they appear as an interesting parallel to macroeconometric models, as evidenced by
their large adoption today within the main national and international institutions.
This paper introduces MELEZE1, a fiscal-policy oriented new-Keynesian Dynamic Stochastic Gen-
eral Equilibrium (DSGE) model developed at the Insee. MELEZE is two-country monetary union model
aimed at representing the situation of France within the Euro area.
Compared to the Mésange model, the present model allows to focus more specifically on questions
related to rational expectations and anticipations of shocks, macroeconomic spillovers in a monetary
union, and endogenous behaviours of the government. In particular, MELEZE offers two alternative
choice of modelling for the government behaviour. A first option is the implementation of a traditional
budget rules linking today’s public consumption to past deficits (as found in the institutional DSGE
literature). As an alternative, we also develop and propose a new approach, modelling the government
as an optimizing agent maximizing households’ welfare. This second modelling directly relates to the
Ramsey policy under bounded rationality.
More generally, MELEZE’s structure compares with most standard tools developed in international
institutions and central banks, yet remains simple enough to be able to disentangle most economic re-
actions. Indeed, the construction of the model was conducted along three objectives. First, it needs to
be refined enough to give insightful supplementary policy analyses using a DSGE model rather than a
macroeconometric model, in particular with respect to fiscal behaviours. Second, it should compare to
1Modèle Économique Linéarisé d’Équilibre en Zone Euro
4
standard institutional models used at central banks and international organizations and help to bridge
the gap between Mésange and such large-scale DSGE models. Last, it should remain relatively small
compared to large institutional DSGE models to both allow for a better understanding of underlying
mechanisms and for an easier transmission of the model internally.
Along those lines, MELEZE therefore includes all traditional ingredients found in the institutional
DSGE literrature, and is largely inspired by (Smets and Wouters, 2002, 2003, 2005) or (Christiano, Eichen-
baum, and Evans, 2005).
All in all, these modelling choices allow to perform in MELEZE, all standard macroeconomic policy
evaluation exercises such as the study of permanent fiscal reforms (Campagne and Poissonnier, 2016a),
or the simulation of labour and goods markets deregulations (Campagne and Poissonnier, 2016b).
This paper is organised as follows. In Section 2, we present the model and derive the first order
conditions for all economic agents. Section 3 describes the methodology and the data used for the cali-
bration of MELEZE as a monetary union model describing the situation of France within the Eurozone.
Lastly, in order to give a few insights on the behaviour of our model, we present the short term re-
sponses through the study of impulse responses to a set of standard policy shocks (Section 4).
The Appendix gives a full description of the steady state and linearised model, as well as addi-
tional remarks on the design of monetary policy in a currency union, on the impact of wage and price
dispersion in the model and on the uniqueness and stability of the steady state.
2 Model
Non-technical outline In this paper, we develop a DSGE model comparing with standard tools devel-
oped in international institutions and central banks (Christiano et al., 2005; Smets and Wouters, 2003).
The model consists of two countries where continuums of firms and households interact on the goods,
labour and capital market. Both firms and households are considered immobile across countries.
As advocated by Mankiw (2000), we distinguish between two types of households. A fraction of
these households is Ricardian, that is not financially constrained. They hold financial asset (or debt),
own capital which they lend to firms in their country (once installed capital is assumed to become
immobile) and also own financial intermediation firms. Therefore, they receive (or pay) interests and
dividends. These Ricardian households also choose their investment each period by arbitrating between
capital and the risk free asset. Non Ricardian households on the contrary are financially constrained and
5
do not hold any asset.
Both types of households also provide labour on monopolistically competitive market. For this rea-
son, households are paid with a mark-up over their marginal disutility. Wage rigidities are added over
the cycle, and each household can only reset its wage in adequateness with his optimal consumption-
leisure arbitrage with an exogenous probability. In this framework, there is no involuntary unemploy-
ment and labour adjusts only at the intensive margin (hours worked).
Households finally consume both domestic and imported goods which are partial substitutes. For
Ricardian households, being non financially constrained allows them to smooth their consumption over
time. Non Ricardian households on the contrary can not. Once their wage level is set, their labour sup-
ply is given by firms demand, their income ensues which they consume entirely within the same quarter.
Firms produce partially substitutable goods from a standard constant returns to scale production
function. Production factors are labour and capital. Total factor productivity is exogenous and growing
at the same pace across countries. At each period firms optimize their relative demand in capital and
labour to minimize their production cost, taking the aggregate wage and gross return on capital as
given. Partial substitutability allows firm to price a mark-up over their marginal cost. Over the cycle,
with an exogenous probability each firm can reset its price to maximize its expected discounted profits,
while internalising its market power. Those price rigidities lead to a New Keynesian Phillips curve.
The modelling of governments’ behaviour departs from the fiscal and budget rules literature used
in quantitative models to endogenize tax rates and public spending in order to ensure the government’s
budget constraint. We consider here forward-looking optimizing governments. We introduce unproduc-
tive public consumption as a proxy for actual public spendings, public investment, public employment
and production of public services altogether. This consumption enters households utility function to-
gether with private consumption. Governments maximize the intertemporal households utility under
the public budget constraint which is an approximation for the exact Ramsey problem. Alternatively, we
also allow the government to behave according to a standard budget rule linking public consumption
with past deficit and output gap. Furthermore, governments collect taxes on wages, capital income, div-
idends, consumption and investment. They can distribute transfers to both types of households. They
also hold debt both at the steady state and over the cycle.
In addition to production of real goods by the firms, a union wide financial market produces finan-
cial intermediation services both for households and governments. Financial intermediaries capture on
top of the interest rate set by the central banker a fee under the form of a debt elastic spread which is
akin to fisim. There are no risk or agency issues in our model so that this fee is not to be interpreted as
6
a risk premium of any kind. In practice, these financial intermediaries ensure the closing of the model
as exposed in Schmitt-Grohé and Uribe (2003) and have a very small production compared to firms of
the real sector.
Notations As much as possible, we keep standard notations throughout this paper (C for consump-
tion, W for wage...). A superscript i ∈ 1, 2 whether on an aggregate or on a parameter refers to the
country. Subscripts are used to specify an operation related to the variable (e.g. habit on consumption or
labour), in particular Cij refers to consumption in country i of goods produced in country j. Upper-case
letters refer to aggregates while lower-case letters refer per GDP unit aggregates or sometimes when we
want to emphasize individual variables (wage, labour and output per firm or capita). Throughout, τ is
the index for a generic household and ε the index for a generic firm. R and NR superscripts relate to
Ricardian and non Ricardian households respectively. t refers to time.
A full dictionary of variables and parameters is given in Tables 3 and 4 in the Appendix.
2.1 Goods and labour aggregations
2.1.1 Aggregation of production within countries
We assume that a continuum of goods of size P is produced in the monetary union. Goods in [ 0 , pP ]
are produced in country 1, while goods in ] pP , P ] are produced in country 2. For formula generaliza-
tion, we shall denote p1 = p and p2 = 1− p.
In each country, domestic production is aggregated into a domestic good using a Dixit–Stiglitz aggre-
gator (Dixit and Stiglitz, 1977) with an elasticity of substitution specific to each country. This modelling
hypothesis is interpretable either as the technology of a perfectly competitive final good sector with sole
inputs intermediate consumption of the continuum of firms or as the relative preferences for each type
of goods of the final consumer. These hypotheses yield the following relationship between the demand
for goods produced by firm yi(ε, i) and the total demand for production of country i (Yit )i=1,2:
Y1t = Z1
(1
pP
∫ pP
0y1
t (ε)θ1−1
θ1 dε
) θ1
θ1−1= (pP)
11−θ1
(∫ pP
0y1
t (ε)θ1−1
θ1 dε
) θ1
θ1−1, (2.1)
Y2t = Z2
(1
(1− p)P
∫ P
pPy2
t (ε)θ2−1
θ2 dε
) θ2
θ2−1= ((1− p)P)
11−θ2
(∫ P
pPy2
t (ε)θ2−1
θ2 dε
) θ2
θ2−1. (2.2)
7
where θi is the elasticity of substitution between goods in country i and Zi a constant of normalisation.2
Maximising consumption under the budget constraint or alternatively, minimising the price for a
bundle unit yields the corresponding production prices:
P1t =
1Z1
(1
(pP)θ1
∫ pP
0P1
t (ε)1−θ1
dε
) 11−θ1
= (pP)1
θ1−1
(∫ pP
0P1
t (ε)1−θ1
dε
) 11−θ1
, (2.3)
P2t =
1Z2
(1
((1− p)P)θ2
∫ P
pPP2
t (ε)1−θ2
dε
) 11−θ2
= ((1− p)P)1
θ2−1
(∫ P
pPP2
t (ε)1−θ2
dε
) 11−θ2
. (2.4)
The resulting relationships between aggregated and retail prices and quantities read:
y1t (ε) =
Zθ1−11
(pP)θ1
(P1
t (ε)
P1t
)−θ1
Y1t =
(P1
t (ε)
P1t
)−θ1
Y1t
pP, (2.5)
y2t (ε) =
Zθ2−12
((1− p)P)θ2
(P2
t (ε)
P2t
)−θ2
Y2t =
(P2
t (ε)
P2t
)−θ2Y2
t(1− p)P
. (2.6)
2.1.2 Labour aggregation
The aggregation of labour in both countries is symmetric to that of goods (Dixit–Stiglitz). The population
size is set to N and a share n of households, that is households in [0,nN], live in country 1. Households
in [nN,N] live in country 2. For formula generalization, we shall denote n1 = n and n2 = 1− n. Labour
is hereafter assumed to be immobile across countries.
Household τ supplies labour li(τ, t) and demands the wage wi(τ, t) so that the total supply of labour
and average wage in country i are (Lit)i=1,2 and (Wi
t )i=1,2.
L1t = (nN)
11−θ1
w
(∫ nN
0l1t (τ)
θ1w−1θ1w dτ
) θ1w
θ1w−1
, (2.7)
L2t = ((1− n)N)
11−θ2
w
(∫ N
nNl2t (τ)
θ2w−1θ2w dτ
) θ2w
θ2w−1
. (2.8)
where θiw is the elasticity of substitution between labour types in country i.
2We take Z1 = pP and Z2 = (1− p)P to simplify the algebra. With this normalisation, if prices P1t (ε) are all equal within
country 1, then the aggregate price index is equal to the individual price P1t = P1
t (ε). In addition, each firm produces an equalshare of total output y1
t (ε) =1
pPY1
t .
8
The corresponding wages are:
W1t = (nN)
1θ1w−1
(∫ nN
0w1
t (τ)1−θ1
w dτ
) 11−θ1
w , (2.9)
W2t = ((1− n)N)
1θ2w−1
(∫ N
nNw2
t (τ)1−θ2
w dτ
) 11−θ2
w . (2.10)
The resulting relationships between aggregated and individual wage and labour supply read:
l1t (τ) =
1nN
(w1
t (τ)
W1t
)−θ1w
L1t , (2.11)
l2t (τ) =
1(1− n)N
(w2
t (τ)
W2t
)−θ2w
L2t . (2.12)
2.1.3 Aggregation of domestic and imported private consumption
In both countries, households have access to goods produced by each country; domestic and foreign
goods are partial substitutes. We derive the case for consumption, we assume that investment and
consumption goods are identical so that the same results apply to investment as well. Aggregation of
imported and domestic production along with the associated consumption price index are modelled as
follows
Cit =
Cii,t
1−αiCi
j,tαi
(1− αi)1−αiαiαi CPIi
t = Pit1−αi
Pjt
αi
(2.13, 2.14)
where Cit is the private consumption of country i and Ci
j,t is the private consumption in country i of
the aggregated goods produced in country j. αi is the import share of country i and therefore a measure
of trade openness.
Note that we have the following relationships:
Ci,t = C1i,t + C2
i,t (2.15)
Cit 6= Ci
1,t + Ci2,t but CPIi
tCit = P1
t Ci1,t + P2
t Ci2,t (2.16)
The aggregation of domestic and imported consumption in both countries yields the following rela-
tionships:
C12,t = α1
(P1
tP2
t
)1−α1
C1t C1
1,t = (1− α1)
(P2
tP1
t
)α1
C1t (2.17, 2.18)
9
C21,t = α2
(P2
tP1
t
)1−α2
C2t C2
2,t = (1− α2)
(P1
tP2
t
)α2
C2t (2.19, 2.20)
The repartition of consumptions between locally-produced goods and foreign ones depends on the
degrees of openness (conveyed by import shares αi). Imported and domestic consumption in country
1 respond to the terms of trade defined as Tt =P2
tP1
t, with elasticities α1 − 1 and α1, respectively: as
expected, the dearer are import prices in relative terms, the more households consume domestically-
produced goods.
2.2 Households
2.2.1 Consumption and investment decision of Ricardian households
In both countries, we assume that a fraction (1− µi) of households can participate to the financial mar-
kets. These households can borrow or lend money on an international market (see 2.5) and doing so
have the possibility to smooth their consumption across periods. Each household of this type (τ) max-
imises her intertemporal utility function subject to her budget constraint (determined by the recursive
law of motion of private assets). Utility is similar to Trabandt and Uhlig (2011), that is non separable,
CES in consumption with external habit formation in a multiplicative manner.3 Disutility of labour also
allows for habits.4 This functional form is compatible with long term growth (King, Plosser, and Rebelo,
2002), as under this form the disutility of labour is concave for any value of the intertemporal elasticity
of substitution of consumption, and also ensures a constant Frisch elasticity.
We also consider that households derive utility from public expenditure and therefore redefine the
consumption bundle as a combination of private and public expenditures in a Cobb-Douglas fashion.
This choice of specification is a subcase of CES aggregations as in McGrattan, Rogerson, and Wright
(1997); Bouakez and Rebei (2007); Coenen, Mohr, and Straub (2008).
In all, a Ricardian household τ solves:
maxCR,i
T (τ),FAiT(τ),I
iT(τ),K
iT(τ)
Et
∞
∑T=t
βiT−tU (CR,iT (τ), Ci
T−1)V(lR,iT (τ), Li
T−1)W(GiT , Gi
T−1) (2.21)
3As in Abel (1990); Galí (1994); Carroll, Overland, and Weil (2000); Fuhrer (2000), introducing habits allows to account forthe persistence of consumption when estimating the model. It also allows for more realistic hump-shaped response functionsfollowing shocks.
4Note however, that in the standard calibration of the model, these habits on labour are muted.
10
subject to the budget constraint
FAiT(τ) =
(RT−1 − ψ(
FAiT−1
PiT−1YiTrT−1
)
)FAi
T−1(τ) + wiT(τ)l
R,iT (τ)
− CPIiT(1 + νc,i
T )CR,iT (τ) + (1− νD,i
T )DiT(τ) + (1− νFD,i
T )FDiT(τ)
+ ΦiT(τ) + (1− νK,i
T )CPIiTrK,i
T KiT−1(τ)− CPIi
T(1 + νc,iT )Ii
T(τ)
(2.22)
and the capital accumulation equation
KiT(τ) = (1− δ)Ki
T−1(τ) + εi,IT
[1− S
(IiT(τ)
IiT−1(τ)
)]IiT(τ) (2.23)
Under the most general form, we define utility as:
U (CR,it (τ), Ci
t−1)V(lR,it (τ), Li
t−1)W(Git, Gi
t−1) =(CR,it (τ)
(Ci
t−1niN
)−hic)1−η (
Git(Gi
t−1)−hi
g
)η1−σi
c
1− σic
1− κi(1− σic)
lR,it (τ)
(Li
t−1
niN
)−hil1+σi
l
σic (2.24)
Et, βi are respectively the expectation at time t operator and the discount factor; CR,it (τ) is the con-
sumption of Ricardian agent τ in country i; σic is the inverse intertemporal elasticity of substitution. κi
is the weight assigned to labour in the utility function; σil is the inverse of the Frisch elasticity. hi
c, hig, hi
l
are the external habit formation parameters on (private and public) consumptions and labour. lR,it (τ) is
the labour supply of household τ and wit(τ) its wage.
FAit(τ) is the household’s τ asset holdings at the end of period t while FAi
t is country’s i aggregate
level of private financial assets (see the atomicity assumption explained in the following paragraph on
private asset dynamics); Rt is the interest rate set by the monetary authority in the union; ψ is an interest
premium on debt (whose function is detailed subsequently).
νc,it is the tax rate on consumption or value-added tax (VAT) through which government expenditure
is partially financed. Dit is the dividend paid by the firm to its owners taxed at rate νD,i
t , FDit are equiva-
lently the dividends paid by the financial sector taxed at rate νFD,it and Φi
t(τ) a lump-sum transfer from
the government. Finally, Kit(τ) is the capital stock of Ricardian households depreciating at rate δ and
which revenues are taxed at rate νK,it .
11
Yi corresponds to the steady state level of output, and TrT−1 = (1 + g)T−1 corresponds to the deter-
ministic trend of our model, where g is the growth rate of TFP.
In the capital accumulation equation, Iit(τ) is the investment level with an adjustment cost5 S (Ii
T(τ)/IiT−1(τ))
depending on previous period level of investment.6 As a result, households pay for the full investment
allotment IiT(τ) and a share S (Ii
T(τ)/IiT−1(τ)) is lost in the installation process. This adjustment cost miti-
gates the fluctuations of capital stock and investment in reaction to exogenous shocks. εi,IT represents an
exogenous shock to this cost sometimes found crucial to replicate the business cycle. The costate vari-
able for constraint (2.23) is defined as qtCPIit(1+ νc,i
t ) times the costate variable of the budget constraint
(2.22) βi tλt, so that qt is the market value of an additional unit of capital, that is Tobin’s marginal Q.
The Euler equation, the investment decision and Tobin’s Q for this programme are identical across
households, and under the assumption that differences in labour and consumption across Ricardian
households are of second order, we aggregate these first order conditions.7
The Euler equation describes the trade off between consumption and savings (on the financial mar-
ket):
βiEt
U ′(
CR,it+1
(1−µi)niN, Ci
t
)V(
LR,it+1
(1−µi)niN, Li
t
)W(Gi
t+1, Git)
U ′(
CR,it
(1−µi)niN, Ci
t−1
)V(
LR,it
(1−µi)niN, Li
t−1
)W(Gi
t, Git−1)
Rt − ψ
(FAi
tPi
t YiTrt
)Πc,i
t+11+νc,i
t+1
1+νc,it
= 1 (2.25)
where Πc,it+1 is the inflation of the consumption price index in country i.
5As in Smets and Wouters (2003) to Smets and Wouters (2007), this cost is introduce in order to smooth the reaction ofinvestment to shocks. Similarly, we assume that at steady state S = 0, S ′ = 0 and S ′′ > 0.
6As in most DSGE models, the investment decision is conducted by households only.7The extent of this approximation is explained by Carroll (2000) in a more general context. Within a linearised model we give
a detailed analysis of distributions in Appendix D.
12
Investment and the marginal value of capital are described by the following first order conditions:
1 =qitε
i,It
1− S
(Iit
Iit−1
)− S ′
(Iit
Iit−1
)Iit
Iit−1
+ βiEt
U ′(
CR,it+1
(1−µi)niN, Ci
t
)V(
LR,it+1
(1−µi)niN, Li
t
)W(Gi
t+1, Git)
U ′(
CR,it
(1−µi)niN, Ci
t−1
)V(
LR,it
(1−µi)niN, Li
t−1
)W(Gi
t, Git−1) qi
t+1εi,It+1S
′(
Iit+1
Iit
)(Iit+1
Iit
)2
(2.26)
qit =βiEt
U ′(
CR,it+1
(1−µi)niN, Ci
t
)V(
LR,it+1
(1−µi)niN, Li
t
)W(Gi
t+1, Git)
U ′(
CR,it
(1−µi)niN, Ci
t−1
)V(
LR,it
(1−µi)niN, Li
t−1
)W(Gi
t, Git−1)(
qit+1(1− δ) +
(1− νk,it+1)r
k,it+1
1 + νc,it+1
)(2.27)
The latter, similar to the Euler equation on consumption (2.25), describes the trade-off between invest-
ment in capital and consumption.
2.2.2 Ricardian households’ asset dynamics
By assumption, only unconstrained households can lend or borrow, their aggregate budget constraint
reads:
FAit =
(Rt−1 − ψ(
FAit−1
Pit−1YiTrt−1
)
)FAi
t−1 + WR,it LR,i
t
− CPIit(1 + νc,i
t )CR,it + (1− νD,i
T )DiT + (1− νFD,i
t )FDit
+ ΦR,it + (1− νK,i
t )CPIitr
K,it Ki
t−1 − CPIit(1 + νc,i
t )Iit
(2.28)
To make the cost of debt increase with the level of indebtedness and also ensure the stationarity
of the model (i.e. rule out unit roots to ensure the convergence of financial assets after shocks, see sec-
tion E), we include a premium on the interest rate ψ, which is akin to a transaction cost on holding
assets paid to an international financial intermediary and enforces a no-Ponzi scheme condition on the
evolution of assets (see Section 2.5). This premium depends positively on f ait =
FAit
Pit YiTrt
, which represents
the level of indebtedness of private agents in country i in real terms, Yi being the steady-state value of
output (once detrended) in country i and Trt its deterministic trend.
The premium a household faces depends on the aggregate private asset holdings of the country (or
local financial conditions), not on the household’s private personal financial position. Thus each house-
hold takes the premium as given in its consumption decision (atomicity assumption). As the model
13
will be linearised, only the value of ψ and its first derivative at the steady state will impact the model
dynamics. We specify ψ such that ψ(0) = 0 and ∂ψ(x)∂x > 0, so that both indebtedness and asset holding
incur a cost paid to the intermediary, and the value of the premium increases with debt.
If at the aggregate level, households in country i are net borrowers (i.e. FAit ≤ 0), resident households
have to pay an interest premium on their debt amounting to |ψ( f ait)|. When the country is net lender,
returns are reduced by ψ( f ait) captured by the intermediary. This mechanism is equivalent to financial
intermediation services indirectly measured (FISIM, see section 2.5).
2.2.3 Consumption decision of non Ricardian households
The remaining fraction µi of households does not have access to financial intermediation and therefore,
their consumption cannot be smoothed across periods. These non Ricardian households follow a rule-of-
thumb:
0 = wit(τ)l
it(τ) + Φi
t(τ)− CPIit(1 + νc,i
t )Cit(τ) (2.29)
on aggregate 0 = WNR,it LNR,i
t + ΦNR,it − CPIi
t(1 + νc,it )CNR,i
t (2.30)
As no assets are available to non Ricardian households, they neither hold shares in domestic firms nor
in the financial sector and hence do not receive dividends. Moreover, in the absence of precautionary
savings of these households in our model, non-Ricardian households cannot hold money as a partial
substitute to bonds or capital and are unable to smooth their consumption in time even partially. As
a consequence, our model may overestimate their reaction to shocks (Challe and Ragot). For instance
following a positive productivity shock, prices and wages being sticky, firms will lower their labour
demand. While Ricardian households can smooth their consumption by selling some assets, consump-
tion of the non Ricardian ones will drop as a direct consequence of their decrease in payroll, this lower
demand will in turn affect output negatively.
2.2.4 Labour supply decision and wage setting
As we did for consumption goods, we model labour aggregation with a Dixit–Stiglitz function. Rela-
tionships between labour and wages are therefore similar to those between consumption and prices
(see section 2.1). Unlike consumption goods, labour is considered immobile and cannot be imported or
exported. The relationship between total demand for labour and each household supply as a function
of the demanded wage reads:
lit(τ) =
1niN
(wi
t(τ)
Wit
)−θiw
Lit (2.31)
14
We assume wage stickiness à la Calvo (1983), with parameter ξ iw denoting the probability not to
adjust wages at each period. There is also partial indexation of wages on past inflation of consumption
prices according to parameter γiw and indexation on targeted inflation with parameter 1− γi
w. Wages
are also indexed on the deterministic trend of TFP.8 Households solve the following program:
maxwi t(τ),li t,T(τ)
Et
∞
∑T=t
(ξ iwβi)T−tU (Ci
T(τ), CiT−1)V
(li t,T(τ), Li
T−1
)W(Gi
T , GiT−1) (2.32)
subject to the labour demand function:
lit,T(τ) =
1niN
(wi
t,T(τ)
WiT
)−θiw
LiT , (2.33)
as well as their respective budget constraint, and the following indexation rule:
wit,T(τ) = wi
t(τ)T−1
∏k=t
(Πc,ik )γi
w(Πc,i)1−γiw
TrTTrt
= wit(τ)Γ
T−1w,t , (2.34)
where wit(τ) is the optimal wage set at time t by household τ and wi
t,T(τ) is its wage at time T when
not reset between time t and T; lit(τ) and li
t,T(τ) are the corresponding labour demands. ΓT−1w,t denotes
the indexation factor ∏T−1k=t (Π
c,ik )γi
w(Πc,i)1−γiw TrT/Trt with Πc,i the steady state inflation of CPIi.
The first order condition reads
0 =Et
∞
∑T=t
(ξ iwβi)T−t lt,T(τ)
U ′(CiT(τ), Ci
T−1)V(
li t,T(τ), LiT−1
)W(Gi
T , GiT−1)
U ′(Cit(τ), Ci
t−1)V(li t(τ), LiT−1)W(Gi
t, Git−1)[U (Ci
T(τ), CiT−1)V ′
(li t,T(τ), Li
T−1
)U ′(Ci
T(τ), CiT−1)V
(li t,T(τ), Li
T−1
) +θi
w − 1θi
w
wit(τ)Γ
T−1w,t
CPIiT(1 + νc,i
T )
] (2.35)
where one may recognize the stochastic discount factor between time t and T and between brackets, the
wedge between the ratio of the marginal utility of labour and consumption and the real wage with a
term in θiw representing the market power of households.
Note that this wage setting equation is at the individual level and therefore that the associated util-
ity function and wages depend on the individual consumption of household τ. However, as for the
Euler and investment equations, we make the standard assumption that individual dispersion can be
8These indexations are necessary to ensure that the distribution of wages does not diverge when there is non zero inflationand exogenous growth at steady state.
15
neglected (see Appendix D).
Although we can describe how wages and labour supply of resetters differ from other households,
we can not do so for consumption. With non separable utility, we are thus forced to assume that non
resetters and resetters have similar consumption within each type.9
The rest of the calculus (steady state and linearisation) is detailed in the subsequent sections, but be-
cause the consumption of the Ricardian and non Ricardian household differ, there will be two symmetric
Phillips curves for two different wages.
2.2.5 Households’ type aggregation
The introduction of two types of households results in additional aggregation rules for consumption,
labour and wages. As seen previously, the existence of two different Phillips curves for Ricardian and
non Ricardian households implies that their consumption, labour and wage patterns differ.
In terms of consumption, we now have for consumption of good j in country i and total consump-
tion:
Cij,t = CNR,i
j,t + CR,ij,t Ci
t = CNR,it + CR,i
t (2.36, 2.37)
where CR,ij,t and CNR,i
j,t respectively denote consumption of Ricardian and non Ricardian households
of good j in country i. Consumption of the two types of households is simply additive: both types
consume the same goods at the same prices with the same imported and domestic share.
Labour and wages are not directly additive, but payroll is by construction. Similarly to labour and
wage in both countries we define aggregate labours and wages for both types of households:
LNR,it = (µiniN)
11−θi
w
(∫NR,i
lt(τ)θiw−1θiw dτ
) θiw
θiw−1
, (2.38)
LR,it = ((1− µi)niN)
11−θi
w
(∫R,i
lt(τ)θiw−1θiw dτ
) θiw
θiw−1
. (2.39)
9Note that these consumptions converge to the same steady state and follow the same dynamic equations (respectively theRicardian Euler equation for consumption and the non Ricardian budget constraint).
16
The corresponding wages are:
WNR,it = (µiniN)
1θiw−1
(∫NR,i
wit(τ)
1−θiw dτ
) 11−θi
w , (2.40)
WR,it = ((1− µi)niN)
1θiw−1
(∫R,i
wit(τ)
1−θiw dτ
) 11−θi
w . (2.41)
Labour aggregation read as follows:
L1t = (nN)
11−θ1
w
(∫ nN
0lt(τ)
θ1w−1θ1w dτ
) θ1w
θ1w−1
= (nN)1
1−θ1w
(∫NR,1
lt(τ)θ1w−1θ1w dτ +
∫R,1
lt(τ)θ1w−1θ1w dτ
) θ1w
θ1w−1
=
(µi
1θ1w
(LNR,1
t
) θ1w−1θ1w + (1− µi)
1θ1w
(LR,1
t
) θ1w−1θ1w
) θ1w
θ1w−1
(2.42)
L2t =
(µi
1θ2w
(LNR,2
t
) θ2w−1θ2w + (1− µi)
1θ2w
(LR,2
t
) θ2w−1θ2w
) θ2w
θ2w−1
(2.43)
Wage aggregation reads:
Wit =
(µi(
WNR,it
)1−θiw+ (1− µi)
(WR,i
t
)1−θiw) 1
1−θiw (2.44)
In real terms, using the variable used in the linearised model, this also rewrites:
(RWi
t
)1−θiw= µi
(RWNR,i
t
)1−θiw+ (1− µi)
(RWR,i
t
)1−θiw
(2.45)
with RW the purchasing power of net wages that is deflated by the country’s VAT included CPI.
2.3 Firms
We assume an exogenous and global technological growth process in the form ζ it = ε
ζ,it (1+ g)t ζ i, where
g is the deterministic growth rate of total factor productivity, ζ the de-trended steady state level of
technology, and εζt a possibly autocorrelated stochastic productivity shock. In the rest of the paper, we
denote Trt = (1 + g)t the deterministic trend of TFP. We assume that technology can be shared and
transferred within the union, so that TFP growth is the same in both countries.10
10However, the steady state detrended level of TFP, ie. ζ i , differs across countries to take into account the initial differences inwealth across countries.
17
2.3.1 Production factors optimisation
Firms hire domestic labour at the cost Wit (1 + νw,i
t ), where νw,it is the payroll tax rate levied by the gov-
ernment on firms.11
Firms also rent capital from households at rate rk,i. In real term the rental cost of demanded capital
Kd,it (ε) is then rk,i
t Kd,it (ε) paid at time t. In nominal terms, this cost equals rk,i
t Kd,it (ε)CPIi
t : the value of
the rented capital in current e is equal to the real capital stock times its market price CPIit .
12 Note that
capital from previous period is used for production at time assuming installation delays. Therefore at
market equilibrium, we have on aggregate Kd,it = Ki
t−1 .
In each country i, firm ε produces the differentiated good yit(ε) with the following technology:
yi(ε, t) =(
ζ itL
it(ε)
)1−α (Kd,i
t (ε))α
at cost Wit (1 + νw,i
t )Lit(ε) + rk,i
t CPIitK
d,it (ε), (2.46)
where α is the share of capital costs in value added. For sake of simplicity, although it may prove em-
pirically relevant (Challe and Ragot), we do not assume that Ricardian and non Ricardian households
correspond to different types of workers. Firms hire both types of households indistinctly.
Every period, firms can reset the quantity of each production factor they use taking wage, taxes and
rental cost as exogenous. The arbitrage condition between labour and capital demand yields:
1− α
α=
Wit (1 + νw,i
t )Lit(ε)
rk,it Kd,i
t (ε)CPIit
on aggregate1− α
α=
Wit (1 + νw,i
t )Lit
rk,it Ki
t−1CPIit
(2.47)
The marginal cost of production is identical across firms and does not depend on its size:
MCit(ε) = MCi
t =1
αα(1− α)1−α
(Wi
tζ i
t(1 + νw,i
t )
)1−α (rk,i
t CPIit
)α(2.48)
RMCit =
MCit
Pit
=1
αα(1− α)1−α
(RWi
tζ i
t(1 + νc,i
t )(1 + νw,it )
)1−α (rk,i
t
)α CPIit
Pit
(2.49)
2.3.2 Price setting
The price setting follows Calvo process in each country. Firm ε can reset its price with exogenous prob-
ability (1− ξi). Producers know the relationship between their price and the demand for their product
11No taxes on labour income (social contribution, income tax) are paid by households here. The steady state is not affected bythis assumption but the reaction of wages to this tax is affected in the short-term.
12The price of capital is by convention the same as the price of investment, which is identical to the price of consumption as weassume that both goods are identical.
18
and choose their price Pit (ε) so as to maximise their expected profit under that constraint:
maxPi
t (ε)Et
∞
∑T=t
(βiξ i)T−tλiT
(Pi
t (ε)yit,T(ε)−Wi
T(1 + νw,iT )Li
t,T(ε)− rk,iT CPIi
TKd,it,T(ε)
), (2.50)
subject to yit,T(ε) =
1piP
(Pi
t,T(ε)
PiT
)−θi
YiT , (2.51)
yit,T(ε) =
(ζ i
tLit,T(ε)
)1−α (Kd,i
t,T(ε))α
, (2.52)
1− α
α=
Wit (1 + νw,i
t )Lit(ε)
rk,it Kd,i
t (ε)CPIit
, (2.53)
Pit,T(ε) = Pi
t (ε)T−1
∏k=t
(Πik)
γip(Πi)1−γi
p = Pit (ε)Γ
T−1t , (2.54)
where the Lagrange multiplier λiT is the marginal utility of a representative Ricardian households in
country i.13 yit,T(ε) is the demand for goods produced by firm ε of country i at time T when its price
was last reset at time t. γip is the parameter of price indexation on past inflation and ΓT−1
t denotes
∏T−1k=t Πi
kγi
p(Πi)1−γip . So Pi
t,T(ε) = Pit (ε)Γ
T−1t is the price of good ε of country i at time T when its price
was last reset at time t. Note that Πit is the inflation of goods produced in country i and differs from
inflation of the consumption price index CPIit , which includes inflation from imported goods as well.
Πi is the steady state value of Πit.
The first order condition reads:
0 =∞
∑T=t
(βiξ i)T−tλiT
YiT
piP
(Pi
t (ε)ΓT−1t
PiT
)−θi (Pi
t (ε)ΓT−1t − θi
θi − 1MCi
T
)(2.55)
2.3.3 Dividends distribution
Firms cannot save or invest, so they redistribute their profits to households. This distribution can be
thought of as dividends to firms owners, if negative it is similar to a recapitalisation of the firm. There-
fore, we assume that only unconstrained households, who have access to financial and investment
markets, are paid such dividends Dit.
Dit = Pi
t Yit −Wi
t (1 + νw,it )Li
t − rk,it Ki
t−1CPIit (2.56)
Dit = Pi
t Yit (1− RMCt) (2.57)
13These households own the firms, so logically their utility enters the price-setting program. This is however neutral on thelinearised Phillips curve apart from a redefinition of β when there is long term growth, a redefinition which does not depend onhouseholds type.
19
2.3.4 Aggregate production function
We assume that the production function is identical across firms. We can compute the aggregated
production function based on the definition of Yit as a function of yi
t(ε):
Yit = (piP)
11−θi
(∫ piP
0yi
t(ε)θi−1
θi dε
) θi
θi−1
=(
ζ itL
it
)1−α (Kd,i
t
)α(piP)
11−θi
∫ piP
0
( Lit(ε)
Lit
)1−α (Kd,i
t (ε)
Kd,it
)α θi−1
θi
dε
θi
θi−1(2.58)
which simplifies because of equation (2.47)
=(
ζ itL
it
)1−α (Kd,i
t
)α(piP)
11−θi
∫ piP
0
(Li
t(ε)
Lit
) θi−1θi
dε
θi
θi−1
=(
ζ itL
it
)1−α (Kd,i
t
)α∆i
t (2.59)
It follows that the production function on aggregate includes a measure of firm size dispersion.
The productivity shock could be assumed to encompass both aspects: individual productivity and size
dispersion; however the pure productivity shock appears on its own in the Phillips curve on prices
(through the marginal cost of production). Hence, following the common implicit assumption in this
literature, we assume that dispersion is stable enough for ∆t to be taken as constant, assumption verified
at first order (see Appendix D). Moreover at the steady-state, firms are all identical, that is of same size1
piP. It follows that the steady state value of the dispersion index ∆ is ∆i = 1.
2.4 Fiscal Authorities
By and large, the purpose of governments is to stimulate domestic production, labour and individual
consumption, as well as to provide with public and collective goods and services.
In the real world, fiscal policy is implemented through a large number of instruments such as
public expenditures, investment, or employment, as well as through the taxation and social contribution
system. However, this large number of instruments cannot be modelled in details and one need to resort
to simplifications.
Tax system First, in MELEZE, we assume that tax rates over consumption, labour and capital incomes
are exogenous and are discretely chosen by governments. This choice is consistent with a low variability
of apparent tax rates in the data over the calibration period.
20
Transfers Second, in the present model, governments can raise lump-sum taxes or commit to lump-
sum transfers to households. In the baseline behaviour of MELEZE, these transfers are held constant.
However, we allow for potential exogenous transfers shocks, targeted on Ricardian or/and non Ricardian
households.
Here, nominal lump-sum transfers from the government are additive Φit = ΦNR,i
t + ΦR,it , and we
chose the most simple way to distribute transfers (default setting), that is in proportion to their popula-
tion share: ΦNR,it = µiΦi
t and ΦR,it = (1− µi)Φi
t.
These transfers can easily be endogenised to reflect the redistributive policy of the welfare state
within the cycle. At steady state, any redistribution weights can be rationalized by the relative weights
assigned by the government to both types of households.
Public expenditures In the absence of public production or employment in the present model, we
capture and encompass all remaining dimensions of public intervention through public expenditures.
In order to model fiscal policy, these expenditures are decomposed between an endogenous compo-
nent answering to economic developments and an exogenous and discretionary component. In standard
DSGE models, the endogenous behaviour is either modelled as the solution to a Ramsey optimality
problem or through budget rules estimated ex ante. In the following sections, we take a closer look at
different budget rules and present an alternative to the Ramsey approach to model public expenditures.
Moreover, and to model the persistence of government expenditures (the welfare state cannot be
dramatically reshaped overnight), we introduce habits on public consumption in the households’ utility
function.
2.4.1 Budget and fiscal rules
Budget rules can be implemented in different ways all relying on the ad hoc description of governments’
spendings as a function of observable endogenous variables.
1. Exogenous public spending In closed economy when fiscal policy is not the purpose of the model,
the government behaviour can be greatly simplified: one can assume (i) the absence of taxes, (ii)
exogenous public expenditure, and (iii) public debt as a mere counterpart of private assets. In
this specific case, lump-sum transfers are assumed to balance the government budget constraint.
For instance, Negro, Schorfheide, Smets, and Wouters (2007) or Justiniano and Primiceri (2008)
21
consider a budget rule of the following form:
Git =
(1− 1
git
)Yi
t (2.60)
where git is an exogenous disturbance following the process:
log git = (1− ρi
g) log gi + ρig log gi
t−1 + σig,tε
ig,t (2.61)
However, such a rule can not be incorporated in our open economy model. As there is explicit
public debt, there must be some mechanism ensuring the existence of a steady state for public
expenditure and debt altogether in both countries.
2. Endogenous spending without distortive taxes Corsetti, Meier, and Müller (2010), in a monetary
union model, consider three fiscal instruments, namely spending, transfers and debt modelled as
follows:
Git = (1−Ψgg)Gi + ΨggGi
t−1 + Ψgy(Yit−1 −Yi, f
t−1) + ΨgdPAi
tPi
t+ εi
g,t (2.62)
Φit
Pit= Gi
(Pi
t Git
CPIitGi
)Ψtg
+ ΨtdPAi
tCPIi
t−1(2.63)
where Yi, ft−1 denotes the level of output that would prevail under flexible prices and wages. At
steady state with these specifications, there is no public debt. Public spending may be made ex-
ogenous (Ψgy = Ψgd = 0), in which case transfers adjust automatically to ensure the convergence
of public debt to its steady state. Corsetti et al. (2010) compare the previous case to a situation
where public spending is adjusted endogenously to contribute to the convergence of public debt
and may in addition be pro or contra cyclical depending on the value of Ψgy.
3. Endogenous spending with distortive taxes In models where fiscal policy is more detailed, ad-
justment can be made through tax rates. For instance, Carton and Guyon (2012) consider that the
VAT rate follows a fiscal rule which enables their monetary union model to reach a steady state:14
νc,it − νc,i = ρG(ν
c,it − νc,i) + (1− ρG)
[βBG(pai
t − pai) + βDG(pait − pai
t−1)]
(2.64)
Under this form, the VAT rate will gradually adjust to equalize the public debt to its targeted level.
14Other instruments available are social contributions, transfers or public expenditures.
22
In the European Commission model QUEST III, Ratto, Roeger, and in’t Veld (2009) consider a
much richer fiscal policy. It is assumed that the government reacts to its own measure of the output
gap, which differs from the wedge between output under staggered and flexible prices usually
considered for monetary policy analysis. This variable influences government spending, public
investment and the income tax rate. Pensions and unemployment benefit react to the population
structure and aggregate wages. A lump-sum tax ensures the convergence of public debt to a
targeted level at steady state by reacting to both debt and deficit.
Modelling governments’ behaviour through fiscal rules is a flexible approach, however it is subject
to two issues. First, budget rules postulate a behaviour rather than an objective for the government
and are therefore subject to the Lucas critique as they do not identify structural parameters for the
government’s behaviour. Second, when designing a budget rule, one should be particularly cautious
as rules may not be compatible with the existence and uniqueness of the model’s solution if long-term
solvency of the government is not properly ensured.
2.4.2 Optimizing government: a simplified approach to the Ramsey problem
Rationale The introduction of rationality in DSGE models historically and naturally lead to the defini-
tion of an optimal government behaviour as a normative benchmark, namely the Ramsey policy. Indeed,
in a internally consistent DSGE approach, governments seek to maximize the welfare of their domestic
households, and it is therefore natural to define the objective of fiscal authorities as the maximization
of the intertemporal utility of households. In the presence of rationality, this maximisation is indeed
subject to the public budget constraint but also to the full set of model constraints. In particular, when
choosing the optimal level of public expenditures Gt, the government internalizes its indirect impact on
households’ consumption and labour supply, and therefore households’ utility.
One strength of this standard Ramsey approach is its robustness to the Lucas critique as it defines
a structural behaviour consistent with the hypotheses of the model. In addition, as we introduce gov-
ernment spending in the utility function in MELEZE, this Ramsey approach appears to be even more
strongly justified.
However, solving a Ramsey problem is both analytically and numerically complex (when not infea-
sible) in large models, especially within the business cycle, as well as unrealistic as it does not embody
political choices observed in the real world that may depart from optimality. This reason underlies the
classical choice of ad hoc budget rules in DSGE models.
As an alternative to these rules, we propose a new approach based on a simplified version of the
Ramsey problem where the government still maximizes households’ utility subject to its transfers/tax
23
revenues budget constraint, however not taking into account all other constraints. Concretely, the gov-
ernment solves the Ramsey problem taking endogenous variables other than public expenditures as
given (such as CiT(τ) and Li
T(τ) here)15. As a result, such a government focuses only on the utility
derived by households through the direct action of the government rather than through second turn
effects on consumption and labour. As for budget rules, this remains inconsistent with the DSGE ap-
proach of a full knowledge of economic mechanisms by agents. However, this may also be interpreted
as a difficulty for fiscal authorities to exactly assess the impact of its policies on the economy.
Closer to the full Ramsey problem, we believe this approach to be more robust to the Lucas critique
than traditional budget rules as it partially micro-founds the behaviour of the government.
However, both approaches suffer from the same paradoxes when embedded in a general equilibrium
model solved under rational expectations. First, in order to solve for such a model, expectations of all
agents are assumed formed through the entire model. It is then paradoxical to assume that either the
government maximizes its objective under a subset of constraints or maximizes an implicit objective
through a rule defined outside the model. Second, both modelling are only simple descriptions of fiscal
authorities and do not encompass real-world phenomena such as the will of authorities to get reelected
that may induce sub-optimal behaviours.16
Program and objective of the government As the government now seeks to maximise the intertem-
poral flow of utility across all households τ, defining weights for each households (ie. the cumulative
distribution function F ), the government’s objective OGt is given by the aggregation of these intertem-
poral flows:17
OGt =
∫τ∈i
Et
∞
∑T=t
βig
T−tU (CiT(τ), Ci
T−1)V(LiT(τ), Li
T−1)W(GiT , Gi
T−1)
dF (τ)
= Et
∞
∑T=t
βig
T−tW(GiT , Gi
T−1)∫
τ∈i
U (Ci
T(τ), CiT−1)V(Li
T(τ), LiT−1)
dF (τ)︸ ︷︷ ︸
= Et
∞
∑T=t
βig
T−tW(GiT , Gi
T−1) ΩT
(2.65)
Under the reasonable assumption that the government cannot distinguish households within the same
sub-group18, denoting ωR,ig the weight on Ricardian agents, and neglecting the intra-group dispersion in
15This choice of modelling is equivalent to consider that the government behaves under bounded rationality.16See for instance, the public choice theory literature.17With separable utility, the government can restrict to maximize the intertemporal flow of utility W sep(Gi
T , GiT−1) alone.
U (CiT(τ), Ci
T−1) and V(LiT(τ), Li
T−1) terms disappear.18That is for instance, the government cannot distinguish and weight differently two Ricardian households. However, it can put
a different weight on Ricardian and non Ricardian agents.
24
labour and consumption, the weighting factor Ωt rewrites:
ΩT = niNωR,ig U
(CR,i
T(1− µi)niN
, CiT−1
)V(
LR,iT
(1− µi)niN, Li
T−1
)
+ (1−ωR,ig )niNU
(CNR,i
TµiniN
, CiT−1
)V(
LNR,iT
µiniN, Li
T−1
) (2.66)
All in all, in the most general case, the government’s program is as follows:
maxGi
T ,PAiT
Et
∞
∑T=t
βig
T−tW(GiT , Gi
T−1) ΩT(CR,iT , CNR,i
T , CiT−1, LR,i
T , LNR,iT , Li
T−1) (2.67)
withW(GiT , Gi
T−1) =
(Gi
t
(Gi
t−1
)−hig)η(1−σi
c)
(2.68)
s.t. PAit =
(Rt−1 − ψg(
PAit−1
Pit−1YiTrt−1
)
)PAi
t−1 + νw,it Wi
t Lit + νk,i
t rk,it CPIi
tKit−1
+ νc,it CPIi
t(Cit + Ii
t) + νD,it Di
t + νFD,it FDi
t − Pit Gi
t −Φit
(2.69)
where PAit denotes the nominal public assets of country i at the end of period t, and Φi
t are nominal
transfers to households.
Note that the real interest rate for governments differs from that of households because their con-
sumptions are priced differently, governments buying exclusively domestic production. Also the atom-
icity hypothesis made for households relative to the asset market does not hold for governments and
their debt premia differ (ψ versus ψg). Besides, the discount factor of the government needs not be
equal to that of households. On the one hand, the government, as an institution, is longer lived than
its citizens and for this reason could put a higher weight on future utility than households do. On the
other hand, as political entities aimed at satisfying voters and wining elections, governments may also
put a higher weight on the near future.
Solving for the previous program yields the following Euler equation for government consumption
that defines the default behaviour of fiscal authorities in MELEZE:
EtβigW1(Gi
t+1, Git)Ωt+1 + βi
gEt+1W2(Git+2, Gi
t+1)Ωt+2
W1(Git, Gi
t−1)Ωt + βigEtW2(Gi
t+1, Git)Ωt+1
Rt − ψg(
PAit
Pit YiTrt
)− PAi
tPi
t YiTrtψg′(
PAit
Pit YiTrt
)Πi
t+1= 1
(2.70)
25
2.5 Financial Intermediation
The stationarity of an open economy model is not straightforward. As explained by Schmitt-Grohé and
Uribe (2003), in a small open economy model, it can be ensured by some modelling elements, which
are usually not microfounded. The literature on monetary union model usually borrows the same so-
lutions. In our model, we microfound one of Schmitt-Grohé and Uribe’s proposals (debt elastic spreads)
and introduce a simplified international financial market.
We assume that there exists an international financial market for assets (private or public). On the
financial market, intermediaries can borrow money from the central bank of the monetary union to
finance public or private credit, and conversely borrow money from agents to deposit it at the central
bank. Through financial intermediaries, private (resp. public) agents can borrow or lend money by pay-
ing a debt premia ψ (resp. ψg). The interest rate for the exchange between the central bank and the
financial intermediary is the interest rate set by the central bank.
We assume that financial intermediaries work in perfectly competitive market. They operate a war
on prices (spreads), up to a point where they make no profit. To ensure the orthogonality of financial
intermediaries with respect to the rest of the monetary union, we assume that their unique cost is the
refinancing cost vis-à-vis the central bank. Assuming so generates no wage payment or capital and inter-
mediate consumption purchases in this branch of activity hence no transfer between the real economy
within the monetary union and financial operators located outside this union. Therefore developments
on the financial market do not affect the rest of the system. The optimisation program of financial inter-
mediaries is not needed to close our model. One could for instance assume that financial activities are
based strictly out of the monetary union, for instance in England or in Switzerland.
As a result, if households or the government in country i are net borrowers (i.e. FAit ≤ 0 or PAi
t ≤ 0 ),
this agent has to pay an interest premium on his debt amounting to |ψ( f ait)|, |ψg(pai
t)|. When the agent
is net lender, returns are reduced by this same spread captured by the intermediary. This mechanism is
equivalent to financial intermediation services (FISIM, see Figure 1). In our model, there is no explicit
risk or asymmetry of information so that the financial intermediation comes down to the spread be-
tween the refinancing rate offered by the Central Bank and the market rate set by the commercial bank
without risk, term or other premium.
Concretely, the aggregate cash needs financial intermediaries borrow from the central bank are the
opposite of all agents asset holdings:
CNt = −(FA1t + FA2
t + PA1t + PA2
t ) (2.71)
26
|PA 0
|FA
Asset
-R
Interest rate
Centralbanker
Financial in-termediary
ψg
ψ
•
•
Figure 1: Debt elastic spreads are similar to FISIM
The production of financial intermediation services associated is given by the amount of spread paid
today by agents on their stock of financial assets available at the end of previous period:
FYt = ∑i=1,2
ψ
(FAi
t−1
Pit−1YiTrt−1
)FAi
t−1 + ∑i=1,2
ψg
(PAi
t−1
Pit−1YiTrt−1
)PAi
t−1 (2.72)
As for good producing firms, we assume that financial intermediaries are owned by Ricardian house-
holds. However, ownership is transnational. We do not introduce in the model any labour or capital
for the financial intermediation industry and simply assume that all benefits are paid lump-sum to
Ricardian households (FD(1,2)t ). Moreover, these benefits are paid in proportion θ f a (c.f. section A.9).
FD1t =
θ f a
1 + θ f a FYt (2.73)
FD2t =
11 + θ f a FYt (2.74)
FD1t + FD2
t = ∑i=1,2
ψ
(FAi
t−1
Pit−1YiTrt−1
)FAi
t−1 + ∑i=1,2
ψg
(PAi
t−1
Pit−1YiTrt−1
)PAi
t−1 (2.75)
In real terms, this rewrites:
FYt
Pit YiTrt
= ∑k=1,2
ψ(FAk
t−1
Pkt−1YkTrt−1
)FAk
t−1
Pit YiTrt
+ ∑k=1,2
ψg(PAk
t−1
Pkt−1YkTrt−1
)PAk
t−1
Pit YiTrt
(2.76)
27
Denoting f dit ≡
FDit
Pit YiTrt
, we have:
f d1t =
θ f a
1 + θ f aFYt
P1t Y1Trt
=θ f a
1 + θ f a
(ψ( f a1
t−1) f a1t−1 + ψg(pa1
t−1)pa1t−1
+Tt−1
θ
[ψ( f a2
t−1) f a2t−1 + ψg(pa2
t−1)pa2t−1
]) 1Π1
t
Trt−1
Trt
(2.77)
f d2t =
11 + θ f a
FYt
P2t Y2Trt
=1
1 + θ f a
(ψ( f a2
t−1) f a2t−1 + ψg(pa2
t−1)pa2t−1
+θ
Tt−1
[ψ( f a1
t−1) f a1t−1 + ψg(pa1
t−1)pa1t−1
]) 1Π2
t
Trt−1
Trt
(2.78)
Also, to mimic the interbank overnight markets where banks clear their daily position towards the
central bank by lending or borrowing according to the refinancing rate, we assume that at each period,
the financial intermediaries clear their position towards the central bank:
CNt = −(FA1t + FA2
t + PA1t + PA2
t ) = 0 (2.79)
This zero cash needs condition can also be read as public debt being held entirely by households within
the union. In section E we show that assuming no aggregate debt in the monetary union at steady state
implies this zero cash needs constraint at all dates. This constraint ensures that the model satisfies the
Walras law, i.e. that once all markets are cleared, three out of four laws of motion of assets (public and
private in both countries) imply the fourth one. The economic property of being Walrassian implies that
the steady state is stable and the solution to the linearised model is unique.
2.6 Monetary Authority, Prices and Inflation
The central bank sets the nominal interest rate Rt common to both countries through a Taylor rule
(Taylor, 1993), where it reacts to both current inflation of the consumption price index and to the output
gap.
Rt = Rρt−1
(R∗(
Πunion,VATt
Π∗
)rπ
Yryt
)1−ρ
(2.80)
where Πunion,VATt and Yt are respectively the VAT-included average inflation of consumption in the mon-
etary union, and the total output gap of the monetary union (see Appendix C.1). R∗ is the interest-rate
28
target of the central bank and Π∗ its exogenous inflation target. rπ and ry are the Taylor rule weights
assigned to inflation and the output gap, ρ is the interest-smoothing parameter.
As there is no union-wide maximizing households embedded in the model, union aggregate price
index and aggregate output gap cannot be directly inferred. A way to bypass this issues can be to as-
sume that for instance, the aggregate price index is consumption weighted geometric average of the
national price indexes as in Eggertsson, Ferrero, and Raffo (2014).
In our model, we choose to derive approximation of the national-accounting exact definitions of
both the aggregate price index and the aggregate output gap for the monetary union. The derivations
are presented in Appendix C.1 and allow to define:
Πunion,VATt =
1
1 + (1+νC,2)CPI2C2
(1+νC,1)CPI1C1
Πc,1,VATt +
1
1 + (1+νC,1)CPI1C1
(1+νC,2)CPI2C2
Πc,2,VATt (2.81)
Yuniont =
1
1 + Tθ
Y1t
Y1 +1
1 + θT
Y2t
Y2 (2.82)
where Πunion,VAT refers to the VAT-included CPI inflation.
2.7 Market Clearing and relative prices
Every period, markets clear in quantities in both countries:
Yit = Ci
i,t + Cji,t + Ii
i,t + I ji,t + Gi
t. (2.83)
In values, this becomes:
Pit Yi
t = Pit (C
ii,t + Ii
i,t) + Pjt (C
ij,t + Ii
j,t) + Pit Gi
t + Pit (C
ji,t + I j
i,t)− Pjt (C
ij,t + Ii
j,t), (2.84)
which can also be written under the well-known form:
Pit Yi
t = CPIit(C
it + Ii
t) + Pit Gi
t + Pit Xi
t − Pjt Mi
t, (2.85)
where Xit is the exports sold to country j at the price of the domestic good. Likewise, the imports Mi
t
are bought from country j at price Pjt . Because demand for foreign goods is addressed by households
for consumption and investment, we have Mit = Ci
j,t + Iij,t = X j
t .
Moreover, the relative prices of consumption RPCit =
CPIit
Pit
with respect to production (net of taxes)
are equal to:
29
RPC1t =
(P2
tP1
t
)α1
RPC2t =
(P1
tP2
t
)α2
(2.86, 2.87)
We denote the terms of trade
Tt =P2
tP1
t(2.88)
3 Calibration
The model being derived, Appendix A details the steady states relationships between variables. Taking
into account all these relationships imposes crucial restrictions on structural parameters, endogenous
ratios to GDP, as well as on endogenous variables in level. We calibrate our quarterly model as to match
the situation of France within the Eurozone19 over the period 1995-2007, and as to stay coherent with
the traditional DSGE literature for structural parameters.
3.1 Classification of parameters and resolution of the steady state
We distinguish three types of parameters: (i) structural parameters, (ii) policy parameters and (iii) en-
dogenous parameters. Indeed, due to the large numbers of steady state restrictions to account for, some
parameters cannot be calibrated freely and are actually endogenously determined by the steady state
equations. Structural parameters are parameters (technology, preferences, etc.) deemed purely exoge-
nous, accounting for mechanisms outside of the model and not susceptible to change across simulations.
Policy parameters corresponds to discretely chosen parameters by fiscal and monetary authorities such
as the inflation target and the tax rates. Lastly, endogenous parameters are constrained by the full steady
state model and need to be solved for. In all, parameters are sorted as follows:
• Structural: n, N, µi, σic, σi
l , ηi, hic, hi
l , hig, α, ζ i, g, ξ i, θi, γi
p, ξ iw, θi
w, γiw, δ, βi, βi
g, κi, αi, ψ, ψg
• Policy: νc,i, νw,i, νk,i, νd,i, ν f d,i, Φi, Π, ry, rΠ
• Endogenous: R, pai, f ai, gyi, cyi, ¯iyi, T, θ, and other endogenous steady state values of endogenous
variables.
The number of firms p, P are mute throughout the model, the scale of production being defined by the
size of populations n, N and productivities ζ i.
Following this choice, we explicit a sequential method for the resolution of the steady state that
minimizes the number of simultaneous equations systems to solve. This resolution is implemented
under R and the main guiding lines are as follows:
1911 countries: Belgium, Germany, Ireland, Greece, Spain, Italy, Luxembourg, Netherlands, Austria, Portugal and Finland.
30
1. First, given the exogeneity of βi, βig and Π, the Euler equations of both the households and the
governments (Equations STEADY.1 and STEADY.6) define the debt to output ratios pai and f ai
as functions of the nominal interest rate R. As such, the zero cash need condition (Equation
STEADY.7) can be interpreted as a market clearing condition defining the price of bonds R de-
pending on the supply (public debt) and demand (private savings) for financial assets.
2. Second, we isolate a first system of equations composed of national market clearings (Equation
STEADY.9), the definitions of the trade balance through both financial flows (Equation STEADY.14)
and trade flows (Equation STEADY.13), the zero cash need condition (Equation STEADY.7) and
the governments’ budget constraints (Equation STEADY.12). This system allows to compute the
nominal interest rate R, the trade balance to production ratio tb, government spendings gyi, con-
sumption and investment to production ratios (cyi + iyi) ¯RPCi, as well as the relative size of nom-
inal value added across countries θ∗ = θ/T = Y1/TY2.
3. A second system of simultaneous equations corresponds to the determination of the share of (non)
Ricardian agents in consumption sic and payroll si
wl given exogenous fractions of non Ricardian
agents µi. This system consists of both the non Ricardian households budget constraints (Equation
STEADY.3) and the consumption-leisure arbitrages (Equation STEADY.4).
4. At this point, Equation STEADY.16 allows to compute the level of labour supply in both economies.
The resolution is closed by computing the level of production in both countries with use of Equa-
tion STEADY.17 and the terms of trade T = (Y1/Y2)/θ∗.
5. Others endogenous ratios are then merely computed using straightforward combinations of pre-
vious variables.
3.2 Data, calibration and inverse inference
Values for structural parameters, when possible, are chosen from the standard literature on DSGE mod-
els, these parameters being estimated either by Bayesian methods on macro data or directly on micro
data.
Based on a large literature review,20 Table 2 presents the calibration of structural parameters as well
as their sources.
20Trabandt and Uhlig (2011), Roeger, Varga, and in’t Veld (2008), Martin and Philippon (2014), Smets and Wouters (2002),Annicchiarico, Di Dio, and Felici (2013), Vogel (2012), Coenen et al. (2012), Eggertsson et al. (2014), Ratto et al. (2009), Everaertand Schule (2008), Bayoumi, Laxton, and Pesenti (2004), Hø j, Jimenez, Maher, Nicoletti, and Wise (2007), Kaplan, Violante, andWeidner (2014), Bussiere, Callegari, and Ghironi (2011), European Commission’s Quest III R&D model for France
31
DATA MELEZEEA (12) excl. FR FR EA (12) excl. FR FR
Output in 2000 (GDP)∗ 5458 1485 5314 1473Output in 2000 (VA excl Financial)∗ 4656 1278 4635 1281Output per capita average growth rate∗∗ 1,1 % 1,5 % 1,2 % 1,2 %Working age population in 2000 ∗∗∗ 110,3 25,7 110,3 25,7Hours worked per week (since 2000) 34,5 34,3 34,6 34,3Gross Op. Surplus to VA 48 % 40 % 46 % 46 %Gross wages to VA 52 % 57 % 54 % 54 %Nominal 3 month Euribor∗∗ 3,8 % - 4,0 % 4,0 %Inflation (CPI)∗∗ 2,0 % 1,6 % 2,0 % 2,0 %Private consumption to GDP ratio 57 % 55 % 58 % 58 %Public consumption to GDP ratio 19 % 23 % 22 % 23 %Investment to GDP ratio 22 % 21 % 20 % 19 %GFCF to Capital ratio - 7 % 9 % 9 %Trade balance 2 % 1 % 0 % 0 %Imports from Euro area partner† 3 % 12 % 3 % 12 %PPP (GDP, since 2002) 1,00 1,07 1,00 1,07PPP (CPI, since 2003) 1,00 1,06 1,00 1,06Public debt -51 % -37 % -51 % -38 %Private assets including firms (S1 excl. S13)∗∗ 34 % 41 % 50 % 41 %Net financial position (S2)∗∗ 17 % -3 % 1 % -3 %Tax revenue (in GDP) 40 % 44 % 37 % 40 %Implicit tax rate on consumption 20 % 20 % 20 % 20 %Consumption tax income (in GDP) 11 % 11 % 13 % 13 %Implicit tax rate on labour 38 % 39 % 38 % 39 %Labour tax income (in GDP) 21 % 22 % 18 % 18 %Capital tax income (in GDP) 8 % 10 % 7 % 8 %Transfers (in GDP) 16 % 17 % 17 % 19 %
Sources: Eurostat (National accounts, inflations, Euribor, Purchasing Power Parity (PPP), Gross Fixed Capital Formation (GFCF),population, Labour Force Survey -incl. Secondary job), Insee (Capital Stock Accounts)Data are averaged from 1995 to 2007 to exclude the crisis. Depending on availability, samples may start after 1995.EA (12) excl. FR stands for a 12-members Euro area excluding France and FR for France.∗ in billion e in current prices∗∗ annualised∗∗∗ aged from 15 to 64 in millions† share of imports from EU partners in private consumption
Table 1: Actual data for France and the Euro Area and the corresponding endogenous values at steadystate with our calibration
32
Structural parametersFrance Eurozone
Union-wideTechnology parameter α - 0.35 Consensus, ANA
Depreciation rate δ - 0.02 ConsensusCapital rigidity S - 6.17 Smets and Wouters (2005)Population size N - 135 922 100 ANA
TFP growth rate g - 0.003 ANA,Coenen et al. (2012)
Monetary policyInflation Π∗ - 1.005 Consensus, ECB
Smoothing parameter ρ - 0.86 Barthélemy, Marx, and Poissonnier (2009)Weight on inflation rπ - 1.6 Barthélemy et al. (2009)
Weight on output gap ry - 0.16 Barthélemy et al. (2009)
National specificPopulation share ni 0.19 0.81 ANA
Trade openess αi 0.15 0.04 ANASubstitutability between goods θi 6 6 Quest III
Substitutability between workers θiw 4 4 Smets and Wouters (2005), GEM, QuestIII
TFP scale factor ζ i 0.1112 0.1119 ANA, GDP targetWeight on labour disutility κi 2.77 3.30 ANA, Hours worked target
Households discount factor βi 1.0002 1.0003 ANA, Debt to GDP targetGovernment discount factor βi
g 0.998 0.997 Own calibration (see Section 3.2)Inverse risk aversion σi
c 1.13 1.13 Smets and Wouters (2005)Inverse Frisch elasticity σi
l 2 2 Smets and Wouters (2005)Weight on public consumption utility ηi 0.2 0.2 see infra.
Consumption habits hic 0.61 0.61 Smets and Wouters (2005), NAWM
Labour habits hil 0 0 Smets and Wouters (2005)
Public consumption habits hig 0.56 0.56 Smets and Wouters (2005)
Share of non-Ricardian agents µi 0.4 0.4 QuestIII, Martin and Philippon (2014)Price rigidity ξ i 0.72 0.72 Barthélemy et al. (2009)Wage rigidity ξ i
w 0.72 0.72 Smets and Wouters (2005)Price indexation γi
p 0.13 0.13 Smets and Wouters (2005)Wage indexation γi
w 0.36 0.36 Smets and Wouters (2005)Households financial premium ψslope 0.0005 0.0005 AuthorsGovernment financial premium ψ
gslope 0.0005 0.0005 Authors
Fiscal policyConsumption tax rate νc,i 20.3% 19.5% Eurostat
Labour tax rate νw,i 39.1% 37.7% EurostatCapital tax rate νk,i 21.0% 17.0% Eurostat
Transfers to GDP ratio Φi 19.4% 17.4% Eurostat
ANA stands for Annual National Accounting data. For France, data are from the Insee, whereas international comparison withinthe Eurozone is conducted based on Eurostat data.Consensus indicates a value close to a large number of standard DSGE models.
Table 2: Structural parameters33
In particular, the specification of households’ utility is a crucial determinant of the behaviour of the
model. As highlighted in Everaert and Schule (2006), the estimation and identification of these two pa-
rameters, and in particular the (Frisch) elasticity of labour supply, is very sensitive to the methodology
(micro or macro) and the sample considered.
Trabandt and Uhlig (2011) calibrate their model to an inverse Frisch elasticity of σl = 1 for France or
the EU, in line with Kimball and Shapiro (2008) for the US. They also consider an alternative based on
Cooley and Prescott (1995) with σl = 0.33 for the US. These values are in line with the business cycle
literature and close to values estimated by Bayesian methods, as for instance in the different versions
of Smets and Wouters’ model with σl = 2.4 (Smets and Wouters, 2003), σl = 2.0 (Smets and Wouters,
2005) both for the EU and σl = 1.9 (Smets and Wouters, 2007) for the US. However, micro and macro
evidence is not easily reconciled and lead to very different values of the Frisch elasticity. Bayoumi et al.
(2004) mention that micro studies give a range for σl from 3 to as large as 20. In alternative scenarios
for the GEM model, Bayoumi et al. (2004); Everaert and Schule (2006) set σl = 6 or 7 for Europe or
France specifically. We calibrate MELEZE in line with Smets and Wouters (2005) in order to work with
a medium range value of the Frisch elasticity, that is σl = 2.0.
For the inverse of the intertemporal consumption elasticity σc, the debate is less fierce and values
range from 0.5 in Bayoumi et al. (2004) for EU countries to 2 as in Trabandt and Uhlig (2011) (EU
and France). The different versions of Smets and Wouters give σc = 1.3 in Smets and Wouters (2003),
σc = 1.13 in Smets and Wouters (2005) both for the EU and σc = 1.4 in Smets and Wouters (2007) for
the US. We choose to match Smets and Wouters (2005) calibration with σic = 1.13.
Another weakly identified parameter, often estimated using Bayesian estimation methods or simply
calibrated with "expert" insights, is the share of non-Ricardian households µi. In GEM, this share is
estimated to be 35% for France and 45% in the Euro area, whereas it stands to 40% for both in QUEST
III. However, micro-studies highlight that these estimated shares might be over-evaluated as only a
few agents are strictly banned from financial markets. Indeed, a large number of agents, designated
as wealthy hand-to-mouth, do possess a large illiquid wealth, such as housing, so that their short-term
consumption is highly correlated to their current income. However, in the long-term, this conclusion
might differ as assets can be traded. Kaplan et al. (2014) compute values for the share of wealthy hand-
to-mouth agents around 20% for France. Close to Kaplan et al. (2014), Martin and Philippon (2014) focus
on the fraction of households with liquid assets representing less than 2 months of total gross income
and calibrate their model to a 46.6% share of non-Ricardian agents in France. We choose to calibrate our
model to estimated values in QUEST III.
34
Last, the calibration of the weight of public consumption in the utility is based on McGrattan et al.
(1997), Bouakez and Rebei (2007) and Coenen et al. (2008). Estimating a CES public-private consump-
tion aggregation, they identify a share of private consumption of 0.8 (1− ηi). In order to simplify the
model, we resort to a Cobb-Douglas aggregation and therefore assume an elasticity of substitution of
one.
However, this literature review is not enough to obtain a realistic calibration so that we proceed to an
inverse inference. Indeed, in order to properly match the targeted observed levels or ratios in our model,
a specific set of parameters needs to be optimized upon and cannot be set freely. For these parameters,
we verify ex-post that they remain in the range identified in the literature. Table 1 compares for main
structural levels and ratios their levels observed in the data with those generated by the model with
our calibration. Targets based on observables are computed as averages over the period 1995-2007 to
purposely exclude the crisis period.
First, the weights of leisure in the utility κi and the level of TFP in each country ζ i allow to replicate
value added levels in each country (Yi), the terms of trade (T) and hours worked (Li).
Second, the Cobb-Douglas parameter α and the market power of firms θi imply the distribution of
output between the remuneration of capital and dividends (Gross Operating Surplus) and wages. There
is however an imprecision in the data on GOS which includes mixed income, i.e. income of independent
workers. Considering this and the wide range of mark-up rates in the literature we have not introduced
differences across countries along this line.
Inflation is calibrated through the central banker’s target. The nominal interest rate is endogenous
and clears the demand and supply of financial assets union-wide.
Final demands to GDP ratio are well replicated by the model. These ratios are endogenous: public
consumption depends on the fiscal policy of both governments and balances their intertemporal budget
constraint; investment depends on other parameters through equation (STEADY.5) and is actually little
sensitive to the depreciation rate. Consumption clears the market once the trade balance is known.
The trade balance is related to the net financial asset position through aggregate national budget
constraints. It can not be sizeable without much greater imbalances in terms of asset holdings. However,
prices can differ across countries because of households preferences for domestic goods. These prefer-
ences are set such that we replicate the trade between France and its partners in the Euro Area and
prices appear higher in France than for its partners in accordance with purchasing power parities data.
35
Discount factors βi and βig are set to match observed debt-asset to GDP ratios in both countries (see
equations (STEADY.1)). First note that it is not necessary for these factors to be lower than one, this
condition applies to their tranformation βi(1+ g) (Kocherlakota, 1990) and is verified by our calibration.
Also, in a closed monetary union, it is not possible to replicate these ratios exactly as the Euro area is
indebted vis-à-vis the rest of the world. We choose to impute the discrepancy to non French Euro Area
residents who thus hold more assets in the model than in the data. In the end, the net financial asset
positions of both countries mirror each other in the model and are rather small.
Taxes are set to the implicit tax rates on consumption and labour computed by Eurostat. Tax revenues
from consumption are higher than in the data, but our models tax base does not discriminate between
consumption and investment. Tax revenues from labour are lower than in the data, but our model does
not treat the case of independent workers which may explain part of the discrepancy. As for capital, we
only tax capital incomes in the model (return on capital, dividends and financial dividends) as opposed
to capital stock. The tax rates on these three bases are supposed identical and set to approach capital tax
income without prejudice on the investment to GDP ratio. These tax rates imply a tax revenue to GDP
ratio close but lower than in the data, however, missing items in our simplified budget for the general
government may account for this discrepancy.
4 Model dynamics
We analyse the short term properties of our model through the computation of IRFs for shocks occur-
ring in France. We first plot the IRFs to standard shocks (productivity, preference, monetary policy),
then to transitory fiscal shocks (VAT and government spending).
A detailed analysis of fiscal shocks in MELEZE including permanent ones, is given in Campagne
and Poissonnier (2016a).
4.1 IRFs to standard shocks
Productivity shock In Figure 2, productivity in France is increased by one percent with autocorrela-
tion set to 0.9. This productivity shock implies an increase in output with marginal positive spillovers
to the rest of the union. Prices drop in France but not in the rest of the Euro area, so that the central
banker does not react much to this idiosyncratic shock and the real interest rate in France increases.
Relative to financial savings, capital returns also benefit from the higher productivity. Hence investment
increases to the detriment of financial savings. Private assets drop and symmetrically public debt as
well, in France but not in the rest of the Euro Area. The drop in domestic prices is favourable to the
terms of trade. However, the trade balance marginally depreciates as domestic demand is spurred while
36
y- axis in p.p. deviation from steady state
Figure 2: IRFs to a one percent productivity shock (autocorrelated)
37
demand addressed from the rest of the Euro Area is relatively unchanged. This is in part due to the
lower trade openness of the Eurozone with respect to France compared to the openness of France with
respect of the rest of the Eurozone.
In comparison with a closed economy model (Smets and Wouters, 2003, 2005), labour only temporary
decreases (the first year) as it becomes more productive. Simulations with a large weight of country one
in the monetary union show that differences are mainly due to the reaction of the central banker who
only slightly lowers its rate in comparison with a union wide productivity shock.
Preference shock In Figure 3, the discount factor in France is increased by one percent (i.e. more
patient households or a negative demand shock) with autocorrelation set to 0.9. As a consequence,
Ricardian households postpone consumption and increase their financial and physical savings. Prices
in France drop to limit the fall in final demand, so much so that the rest of the Euro Area follows
(though with delay and to a lesser extend), and the central banker sets an accommodative monetary
policy. The increase in investment more than compensates for the drop in consumption starting from
the second year after the shock. The drop in domestic demand relatively to the rest of the Euro Area is
beneficial to the trade balance, an effect magnified by the increase in the terms of trade.
Monetary policy shock In Figure 4, the monetary policy rate is increased by 100 basis points. This
large shock to monetary policy symmetrically impacts both France and the rest of the Euro Area. The
only difference stems from the financial asset holdings in both regions (and their consequences on the
trade balance and the terms of trade). Such a large tightening of monetary policy markedly depreciates
output union-wide. Prices plummet and gradually decrease both in France and in the rest of the Euro
Area, so much so that the endogenous behaviour of the central banker leads to a slight decrease in
its interest rate, partially offsetting its initial shock. Output reacts strongly negatively as the forward-
looking government cuts spendings reacting to the higher financing cost. A less adverse reaction is
observed when implementing a budget rule instead, as can be seen on Figure 9. These results are in line
with textbook simulations (Galí, 2008, Chapter 6).
Public spending Following a 1 percentage point autocorrelated increase in public spending (Figure 5),
output increases in France. With our Euler-type modelling of government consumption, this increase in
public spending is financed through public debt, mirrored by private assets. Prices marginally adjust
upwards to this demand shock and the central banker’s reaction to this idiosyncratic shock is minimal.
The demand shock is, by assumption, only addressed to domestic production and partially crowds out
investment. As a consequence, the demand addressed to the rest of the Euro Area increases but due to
price developments, the competitiveness of France deteriorates and the trade balance temporarily and
38
y- axis in p.p. deviation from steady state
Figure 3: IRFs to a one percent preference shock (autocorrelated)
39
y- axis in p.p. deviation from steady state
Figure 4: IRFs to a 100 basis points monetary policy shock
40
y- axis in p.p. deviation from steady state
Figure 5: IRFs to a one percent government spending shock (autocorrelated)
41
marginally declines.
The main difference with Smets and Wouters (2003) is the absence of crowding out on private con-
sumption: private consumption slightly increases in reaction to the shock. This is linked to the non
separability of government spendings and private consumption in the utility function and the introduc-
tion of non Ricardian households (see section 5.1 for more detailed explanation on the mechanisms at
play).
5 Discussion of the model specification
We have introduced in the model several features which require a closer look: first, non Ricardian house-
holds and Edgeworth complementarity-substitutability are two competing mechanisms to generate stronger
reaction of private expenditures to public spending shocks, second, the government maximizing house-
holds utility function departs from usual budget rules.
5.1 Non Ricardian households, Edgeworth complementarity/substitutability of pri-vate and public spending and their effect on consumption
Introducing non Ricardian households à la Campbell and Mankiw (1989) is advocated by Mankiw (2000)
to analyse fiscal policy. When it comes to fiscal multipliers these agents can generate positive reaction
of private consumption to public spending shocks in neo-Keynesian models (Galí, Vallés, and Lopez-
Salido, 2007) as their marginal propensity to consume is 1. This stylised fact is not replicated in RBC
models or standard neo-Keynesian models without non Ricardian households.
However, to replicate sizeable reactions, the share of non Ricardians may have to be set at a large value,
inconsistent with microeconomic empirical facts. Fève and Sahuc (2013) consider Edgeworth complemen-
tarity in the utility function between private and public consumption as an other mechanism to generate
a positive reaction of consumption to public spending. Their estimations of a closed economy model for
the Euro Area favour this mechanism to non Ricardian households whose share is therefore estimated to
be small (7%). With our general specification of utility, the two mechanism can be compared.
Mechanisms Without either non Ricardian households or Edgeworth complementarity-substitutability, gov-
ernment spending shocks tend to crowd out private consumption through its inflationary effect. Non
Ricardian households contrary to Ricardian ones can not save the increased income implied by higher
government expenditure: they can not postpone consumption. Hence if the fraction of non Ricardians is
sufficiently large a positive reaction of private consumption to a government spending shock. With Edge-
worth complementarity-substitutability only, government spending directly affects households marginal
42
utility and weights on all their arbitrages. However, assuming substitutability or complementarity will
imply opposite reactions.
To exemplify the implications of substitutability or complementarity in the Edgeworth mechanism,
we change the intertemporal elasticity of substitution of consumption.
With our baseline calibration (σc > 1), the marginal utility of consumption decreases with public
spendings, hence private and public consumption are Edgeworth substitute. Following a positive shock
to government spending, ceteris paribus households will have to compensate by lowering private con-
sumption to readjust the marginal utility of consumption in their different arbitrages. With an alterna-
tive calibration (σc = 0.6 < 1), private and public consumption are Edgeworth complements: following a
positive shock to government spending, the utility functional form implies that the marginal utility of
consumption increases, so that households will increase their consumption to decrease this marginal
utility in their different arbitrages.
Simulations Figure 6 exemplifies these mechanisms. We simulate a one point shock to public spend-
ing, with autocorrelation 0.9.
In our baseline scenario, we assume σc > 1, resulting also in an Edgeworth substitutability. As a result,
the two mechanisms under scrutiny work in opposite directions (Figure 6a). With non Ricardian house-
holds only, consumption reacts positively, whereas with Edgeworth substitutability only, consumption
reacts negatively upon shock.
With our calibration, the effect of non Ricardian households dominates Edgeworth substitutability. This
result crucially depends on the parameters of the utility function and the share of non Ricardians. With
neither mechanism, consumption reacts negatively upon shock (although less than with Edgeworth sub-
stitutability), with both it reacts similarly to the case with non Ricardians only.
In the case of σc < 1 (Figure 6b), assuming none of the two mechanisms generates the usual crowd-
ing out effect and a negative response of consumption upon shock. Edgeworth complementarity and non
Ricardian households both generate a (small) positive reaction of private consumption upon shock. With
both mechanisms (baseline scenario), the two effects are combined, though dominated by the non Ricar-
dian agents mechanism, and private consumption increases by one tenth of public spending deviation.
5.2 Government spending in the utility function compared to budget rules
Corsetti et al. (2010) investigate the effect of fiscal stimulus in a two country model (not in a monetary
union). Their investigation of cross border spillovers is based on a pair of budget and transfer rules
43
(a) Edgeworth substitutability σc > 1 (b) Edgeworth complementarity σc < 1
y- axis in p.p. deviation from steady state
Figure 6: Effect of non-Ricardian households and Edgeworth complementarity (or substitutability) onprivate consumption following a one percent government spending shock
endogenizing government spending and tax instruments in order to ensure long-term solvency of the
government.21 They show that the impulse responses to government spending shocks and the associ-
ated fiscal multipliers depend markedly on the combination of instruments used to ensure this long
term solvency (lump-sum taxes or spending) and on the sensitivity of these instruments to government
spending and output gap respectively.
As we introduce government spending in the utility function, we explained that the government can
be modelled as an optimizing agent with an explicit objective, and we now compare this behaviour with
Corsetti et al.’s spending rule. As they do, we consider a pro or contra cyclical rule as well as an acyclic
one (see 2.4). Figures 7, 8 and 9 present the Impulse Response Functions for productivity, preference
and monetary policy shocks.22
Two main analytical differences between the budget rule and the optimizing government are the
reaction to interest rates and the forward looking set-up of the optimizing government. Second round
effects in the Euler equation through labour and private consumption are of secondary importance. Be-
cause the optimizing government is forward looking, government spending is directly adjusted upon
shock (see for instance Figure 7). The reaction to interest rates (the cost of public debt) is also sizeable
(Figure 9).
21Corsetti et al. (2010) do not introduce a spread on public debt as in MELEZE.22Figures 10 to 12 in appendix provide complementary IRFs.
44
Despite these differences, the IRFs of output and the real interest rate are quite similar between both
types of modelling paradigms. The main differences are indeed with public consumption and public
debt. The IRFs related to the three different government spending rules (pro-cyclical, contra-cyclical and
acyclic) exhibit little differences with one another. In comparison with the optimizing government, all
three exhibit sizeable oscillations around the steady state, showing weak convergence mechanisms.23
The optimizing government shows monotonic IRFs.
All in all, these simulations show that although budget rules are a simple and flexible way to de-
scribe the government’s behaviour, one need to be careful when including such rules into a general
equilibrium model. However, differences in IRFs are confined to the government block and do not
largely influence the behaviour of other real variables across types of modelling but also across cyclical-
ity of the budget rules.
Figures 13 and 14 show similar results for a government consumption shock.
y- axis in p.p. deviation from steady state
Figure 7: IRFs to a one percent productivity shock (autocorrelated)
23Other simulations show that this behaviour is amplified if the reaction to the deviation of public debt to its target is strongerand is linked to the autocorrelation of the budget rule: when it is smaller, fluctuations fade out.
45
y- axis in p.p. deviation from steady state
Figure 8: IRFs to a one percent preference shock (autocorrelated)
46
y- axis in p.p. deviation from steady state
Figure 9: IRFs to a 100 basis points monetary policy shock
47
References
A. Abel. Asset prices under habit formation and catching up with the Joneses. The American Economic
Review, 80(2), 1990.
B. Annicchiarico, F. Di Dio, and F. Felici. Structural reforms and the potential effects on the Italian
economy. Journal of Policy Modeling, 35(1):88–109, Jan. 2013. ISSN 01618938.
J. Barthélemy, M. Marx, and A. Poissonnier. Trends and Cycles: an Historical Review of the Euro Area.
Document de travail de la Banque de France, 258(November), 2009.
T. Bayoumi, D. Laxton, and P. Pesenti. Benefits and Spillovers of Greater Competition in Europe: A
Macroeconomic Assessment. Board of Governors of the Federal Reserve System, International Finance
Discussion Papers, 803(April), 2004.
H. Bouakez and N. Rebei. Why does private consumption rise after a government spending shock?
Canadian Journal of Economics/Revue . . . , 40(3):954–979, 2007.
M. Bussiere, G. Callegari, and F. Ghironi. Estimating trade elasticities: Demand composition and the
trade collapse of 2008-09. NBER Working Paper Series, 17712(December), 2011.
G. A. Calvo. Staggered prices in a utility-maximizing framework. Journal of monetary Economics, 12(3):
383–398, 1983.
B. Campagne and A. Poissonnier. Laffer curves and fiscal multipliers: lessons from a DSGE model.
Document de travail Insee - Dese, G2016/06, 2016a.
B. Campagne and A. Poissonnier. Structural reforms in DSGE models : a plead for sensitivity analyses.
Document de travail Insee - Dese, G2016/07, 2016b.
J. Campbell and N. Mankiw. Consumption, income and interest rates: Reinterpreting the time series
evidence. NBER Macroeconomics Annual 1989, Volume . . . , 4(1989):185–216, 1989.
C. D. Carroll. Solving consumption models with multiplicative habits. Economics Letters, 68(1):67 – 77,
2000.
C. D. Carroll, J. Overland, and D. N. Weil. Saving and Growth with Habit Formation. American Economic
Review, 90(3):341 – 355, 2000.
B. Carton and T. Guyon. Désendettement en union monétaire : un modèle structurel. Economie et
statistique, 451(1):131–153, 2012. ISSN 0336-1454.
E. Challe and X. Ragot. Precautionary Saving over the Business Cycle. Economic Journal, (Forthcoming).
48
L. J. Christiano, M. Eichenbaum, and C. L. Evans. Nominal Rigidities and the Dynamic Effects of a
Shock to Monetary Policy. Journal of Political Economy, 113(1):1–45, Feb. 2005. ISSN 0022-3808.
G. Coenen, M. Mohr, and R. Straub. Fiscal consolidation in the euro area: Long-run benefits and short-
run costs. Economic Modelling, 25(5):912–932, Sept. 2008. ISSN 02649993.
G. Coenen, C. J. Erceg, C. Freedman, D. Furceri, M. Kumhof, R. Lalonde, D. Laxton, J. Lindé,
A. Mourougane, D. Muir, S. Mursula, C. de Resende, J. Roberts, W. Roeger, S. Snudden, M. Tra-
bandt, and J. in’t Veld. Effects of Fiscal Stimulus in Structural Models. American Economic Journal:
Macroeconomics, 3(4):22–68, Jan. 2012. ISSN 1945-7707.
T. Cooley and E. Prescott. Economic growth and business cycles. In Frontiers of business cycle research,
pages 1–38. 1995.
G. Corsetti, A. Meier, and G. Müller. Cross-border spillovers from fiscal stimulus. International Journal
of Central Banking, 6(March):5–37, 2010.
A. K. Dixit and J. E. Stiglitz. Monopolistic competition and optimum product diversity. The American
Economic Review, pages 297–308, 1977.
G. Eggertsson, A. Ferrero, and A. Raffo. Can Structural Reforms Help Europe ? Journal of Monetary
Economics, 61(January):2–22, 2014.
C. Erceg, D. Henderson, and A. Levin. Optimal monetary policy with staggered wage and price con-
tracts. Journal of monetary Economics, 46, 2000.
L. Everaert and W. Schule. Structural Reforms in the Euro Area: Economic Impact and Role of Synchro-
nization Across Markets and Countries. IMF Working Papers, 06(137):1, 2006. ISSN 1018-5941.
L. Everaert and W. Schule. Why It Pays to Synchronize Structural Reforms in the Euro Area Across
Markets and Countries. IMF Staff Papers, 55(2):356–366, Apr. 2008. ISSN 1020-7635.
P. Fève and J.-G. Sahuc. On the Size of the Government Spending Multiplier in the Euro Area. Toulouse
School of Economics Working papers, 396:1–31, 2013.
J. Fuhrer. Habit formation in consumption and its implications for monetary-policy models. American
Economic Review, (1996), 2000.
J. Galí. Keeping Up with the Joneses: Consumption Externalities, Portfolio Choice, and Asset Prices.
Journal of Money, Credit and Banking, 26(1):1 – 8, 1994.
J. Galí. Monetary Policy, inflation, and the Business Cycle: An introduction to the new Keynesian Framework.
Princeton University Press, 2008. ISBN 9780691133164.
49
J. Galí, J. Vallés, and J. D. Lopez-Salido. Understanding the effects of government spending on consump-
tion. Journal of the European Economic Association, 5(1):227–270, 2007.
J. Hø j, M. Jimenez, M. Maher, G. Nicoletti, and M. Wise. Product Market Competition in the OECD
Countries. OECD Economics Department Working Papers, (575), 2007.
A. Justiniano and G. E. Primiceri. The Time-Varying Volatility of Macroeconomic Fluctuations. American
Economic Review, 98(3):604–641, May 2008. ISSN 0002-8282.
G. Kaplan, G. L. Violante, and J. Weidner. The Wealthy Hand-to-Mouth. Brookings Papers on Economic
Activity, (Spring):77–153, 2014.
M. Kimball and M. Shapiro. Labor supply: Are the income and substitution effects both large or both
small? NBER Working Paper Series, 14208(July), 2008.
R. King, C. Plosser, and S. Rebelo. Production, growth and business cycles: Technical appendix. Compu-
tational Economics, pages 87–116, 2002.
C. Klein and O. Simon. Le modèle MÉSANGE réestimé en base 2000. Tome 1-Version avec volumes à
prix constants. document de travail Insee, 2010.
N. Kocherlakota. On the ’discount’ factor in growth economies. Journal of Monetary Economics, 25:43–47,
1990.
R. E. Lucas. Econometric policy evaluation: A critique. In Carnegie-Rochester conference series on public
policy, volume 1, pages 19–46. Elsevier, 1976.
N. Mankiw. The savers-spenders theory of fiscal policy. NBER Working Paper Series, 2000.
P. Martin and T. Philippon. Inspecting the Mechanism: Leverage and the Great Recession in the Euro-
zone. NBER Working Paper Series, 20572(October), 2014.
E. R. McGrattan, R. Rogerson, and R. Wright. An Equilibrium Model of the Business Cycle with House-
hold Production and Fiscal Policy. International Economic Review, 38(2):267–90, May 1997.
M. D. Negro, F. Schorfheide, F. Smets, and R. Wouters. On the fit and forecasting performance of new
keynesian models. Journal of Business and Economics Statistics, 25(2):123–143, 2007.
M. Ratto, W. Roeger, and J. in’t Veld. QUEST III : An estimated open-economy DSGE model of the euro
area with fiscal and monetary policy. Economic Modelling, 26(1):222–233, 2009. ISSN 0264-9993.
W. Roeger, J. Varga, and J. in’t Veld. Structural Reforms in the EU: A simulation-based analysis using
the QUEST model with endogenous growth. European Commission Economic Papers, 2008.
50
S. Schmitt-Grohé and M. Uribe. Closing small open economy models. Journal of International Economics,
61(1):163–185, Oct. 2003. ISSN 00221996.
F. Smets and R. Wouters. An estimated Stochastic Dynamic General Equilibrium Model of the Euro
area. Working paper series, European Central Bank, (171), 2002.
F. Smets and R. Wouters. An estimated dynamic stochastic general equilibrium model of the euro area.
Journal of the European Economic Association, 1(5):1123–1175, 2003.
F. Smets and R. Wouters. Comparing shocks and frictions in US and euro area business cycles: a Bayesian
DSGE approach. Journal of Applied Econometrics, (61), 2005.
F. Smets and R. Wouters. Shocks and frictions in US business cycles: A Bayesian DSGE approach. CEPR
Discussion Paper, (6112), 2007.
J. B. Taylor. Discretion versus policy rules in practice. In Carnegie-Rochester conference series on public
policy, volume 39, pages 195–214. Elsevier, 1993.
M. Trabandt and H. Uhlig. The Laffer curve revisited. Journal of Monetary Economics, 58(4):305–327, May
2011. ISSN 03043932.
L. Vogel. Structural reforms, fiscal consolidation and external rebalancing in monetary union: A model-
based analysis. Economic Modelling, 29(4):1286–1298, July 2012. ISSN 02649993.
51
A Steady state
We use · to denote the steady state operator. It can be defined only for detrended real variables and
inflations or relative prices as there is exogenous growth and non zero inflation.
We denote cyi = Ci/Yi, gyi = Gi/Yi and iyi = Ii/Yi, the shares of private consumption, public
expenditure and investment in each country’s GDP, and θ = Y1
Y2 , the relative size of production of
country 1 and 2.
At steady state, we assume that inflation and monetary interest rate equal the central bank’s targets.
A.1 Goods Aggregation
The different consumptions are linked as follows:
C11 = (1− α1)C1Tα1
C12 = α1C1Tα1−1 (A.1, A.2)
C21 = α2C2T1−α2
C22 = (1− α2)C2T−α2
(A.3, A.4)
Identical equalities hold for investment. Aggregation across households reads:
Ci = CNR,i + CR,i (A.5)
A.2 Households
A.2.1 Ricardian Households Euler equation
From the Euler equation of Ricardian households and governments, we have:
βi R− ψ( f ai)
Πci
= 1 (STEADY.1)
where we define βi = βi(1 + g)(1−σic)((1−η)(1−hi
c)+η(1−hig))−1 with g the exogenous growth rate of TFP.
A.2.2 Capital, Investment and Tobin’s Q
At steady state, the capital dynamics, the investment decision and Tobin’s Q are defined by:
Ii = (δ + g)Ki (A.6)
1 = βi((1− δ) +
(1− νK,i)rK,i
1 + νc,i
)(A.7)
q = 1 (A.8)
52
From the previous equation we can isolate the interest rate on capital:
rK,i =
(1βi− 1 + δ
)1 + νc,i
(1− νK,i)(STEADY.2)
A.2.3 Households’ trade-off between consumption and leisure
For households’ trade-off between consumption and leisure not to be distorted by long term growth,
the ratio of marginal utilities of labour and consumption must have the same trend as real wages. This
is possible either with non separable utility or log utility of consumption (King et al., 2002).
This arbitrage is described by the Phillips curve on wages and yields:
RWR,i LR,i
CR,i =θi
wθi
w − 1(1 + σi
l )σic
(1− σic)(1− η)
κi(1− σic)
[LR,i
(1−µi)niN
(Li
niN
)−hl]1+σi
l
1− κi(1− σic)
[LR,i
(1−µi)niN
(Li
niN
)−hl]1+σi
l(A.9)
RWNR,i LNR,i
CNR,i =θi
wθi
w − 1(1 + σi
l )σic
(1− σic)(1− η)
κi(1− σic)
[LNR,i
µiniN
(Li
niN
)−hl]1+σi
l
1− κi(1− σic)
[LNR,i
µiniN
(Li
niN
)−hl]1+σi
l(A.10)
which can be later on simplified using the share of both types of households in consumption and payroll,
the share of consumption in output and the share of payroll in output.
A.2.4 Ricardians’ and non Ricardians’ shares of consumption and payroll
We showed that consumption and payroll of Ricardians and non Ricardians add up to total consump-
tion and payroll in each country at each date. We denote the shares of non Ricardians in consumption
sic =
CNR,i
Ci , the shares in payroll siwl =
WNR,i LNR,i
Wi Li .
The Dixit-Stiglitz aggregation of labour in both countries yields:
RWR,i
RWi =
(1− si
wl1− µi
)− 1θiw−1 RWNR,i
RWi =
(si
wlµi
)− 1θiw−1
(A.11, A.12)
LR,i
Li = (1− µi)
(1− si
wl1− µi
) θiw
θiw−1 LNR,i
Li = µi
(si
wlµi
) θiw
θiw−1
(A.13, A.14)
53
The budget constraint of non Ricardians relates the two shares in the following way:
RPCicyisic =
1− α
(1 + νc,i)(1 + νw,i)
(θi − 1)θi si
wl +φNR,i
1 + νc,i (STEADY.3)
The budget constraint of Ricardians is collinear to this equation and does not help in calibrating the
shares (see E).
At steady state, the consumption-leisure arbitrages combined also relate the two shares:
(si
wl1− si
wl
) 1+σil θi
wθiw−1
=
(µi
1− µi
) (1+σil )θ
iw
θiw−1 1− si
csi
c
1 + 1−siwl
1−sic
θi−1θi
θiw−1θi
w
(1−α)(1−σic)(1−η)
(1+σil )σ
ic(1+νc,i)(1+νw,i)RPCicyi
1 + siwlsi
c
θi−1θi
θiw−1θi
w
(1−α)(1−σic)(1−η)
(1+σil )σ
ic(1+νc,i)(1+νw,i)RPCicyi
(STEADY.4)
Together, equations (STEADY.3) and (STEADY.4) define the shares of both types of households in
consumption and payroll as functions of the other parameters in the model, first of which µi. This
system must be solved numerically once other parameters are known. Note that if µi = 0, sic = si
wl = 0
is indeed a solution to this system when no transfers are paid to the (empty) population of non Ricardians.
A.3 Firms
From the production function, we have:
Yi = (ζ i Li)1−α(Ki)α∆i (A.15)
with ζ i the steady state value of detrended TFP, a scale factor to be determined, and ∆i = 1, see Ap-
pendix D. Note that due to our time to build assumption on capital, capital stock at time t− 1 is detrended
with TFP trend at time t.
From the arbitrage between capital and labour input we already have:
1− α
α=
RWi(1 + νc,i)(1 + νw,i)Li
rk,iKi (A.16)
with RWi= Wi
CPIi(1+νc,i)Trtthe purchasing power of net wages in country i.
54
From the Phillips curve on prices, at steady state the mark-up, i.e. the ratio of production price over
the marginal cost of production equals θi
θi−1 .
RPCi(
RWi
ζ i (1 + νc,i)(1 + νw,i)
)1−α (rk,i)α
= αα(1− α)1−α θi − 1θi (A.17)
RMCi=
θi − 1θi (A.18)
Also, because of constant returns to scale, we can check that the mean cost equals the marginal cost
and we find that:
RPCi RWi(1 + νc,i)(1 + νw,i)Li
Yi = (1− α)θi − 1
θi (A.19)
RPCi rk,iKi
Yi = αθi − 1
θi (A.20)
Combining the capital share with the steady state equations for capital dynamic and investment
decision (section A.2.2) gives:
αθi − 1
θi =
(1βi− (1− δ)
)1 + νc,i
1− νk,iiyi
δ + gRPCi (STEADY.5)
The firms’ dividends equation gives:
Di
PiYiTr= di = 1−
(RWi
(1 + νw,i)(1 + νc,i)Li
Yi + rk,i Ki
Yi
)RPCi (A.21)
di = 1− 11− α
RWi(1 + νw,i)(1 + νc,i)
Li
Yi RPCi (A.22)
that rewrites with help of the Phillips curve on prices and arbitrage between K and L:
di = 1 +1− θi
θi =1θi (A.23)
Firms are able to fix prices higher than their marginal cost and keep a mark-up 1θi .
55
A.4 Fiscal Authorities
With our maximizing government, the public debt level is pinned down thanks to the debt elastic spread
and from the Euler equation of governments, we obtain:
βig
R− ψ( pai)− ψ′( pai) pai
Πi= 1 (STEADY.6)
where we define βig = βi
g(1 + g)(1−σic)((1−η)(1−hi
c)+η(1−hig))−1.
A.5 Financial Intermediation
Financial intermediaries’ profits write:
f d1=
θ f a
1 + θ f aFY
P1Y1Tr(A.24)
=θ f a
1 + θ f a
(ψ( f a1
) f a1+ ψg(pa1)pa1 +
Tθ
[ψ( f a2
) f a2+ ψg(pa2)pa2
]) 1
Π1(1 + g)
(A.25)
f d2=
11 + θ f a
FYP2Y2Tr
(A.26)
=1
1 + θ f a
(ψ( f a2
) f a2+ ψg(pa2)pa2 +
θ
T
[ψ( f a1
) f a1+ ψg(pa1)pa1
]) 1
Π2(1 + g)
(A.27)
with θ f a defined by equation (A.42) below.
In addition, the zero cash need condition implies at all time and a fortiori at steady state that the net
financial position of one country neutralises the net financial position of the other:
FA1t + FA2
t + PA1t + PA2
t = 0 (A.28)
f a1+ pa1 = − f a2
+ pa2
θT (STEADY.7)
56
A.6 Monetary Authority, Prices and Inflations
Using the definition for both the relative price of consumption and the consumption price index, we
have the following relationships:
RPC1= Tα1
RPC2= T−α2
(STEADY.8)
Inflation rates are also directly related:
ΠC,1= Π11−α1
Π2α1
ΠC,2= Π21−α2
Π1α2
(A.29, A.30)
As a result, if T exist, that is if the ratio of consumption and production prices or domestic and
importation prices has a steady state value in both countries, a solution exists such that:
Π1= Π2
= ΠC,1= ΠC,2 (A.31)
We assume that such an inflation level is the central bank’s target. Finally, from the Taylor rule we have:
R = R∗, since Y = 1 and Π = Π∗ (A.32)
i.e. the central banker set the interest rate and inflation at its target level at the steady-state.
A.7 Market Clearing
At steady state, we have from the market clearing equation
Y1 = C11 + C2
1 + G1 + I11 + I2
1 (A.33)
Y2 = C22 + C1
2 + G2 + I22 + I1
2 . (A.34)
These equations imply that the market clearing condition at steady state is:
1 = (cy1 + iy1)Tα1+ gy1 + tb 1 = (cy2 + iy2)T−α2
+ gy2 − θ
Ttb (STEADY.9)
57
A.8 Budget constraints
From the budget constraint of Ricardian households, we obtain in real terms:
f ait =
FAit
Pit YiTrt
=(
Rt−1 − ψ( f ait−1)
) f ait−1
Πit(1 + g)
+ RPCit(1 + νc,i
t )RWR,i
t LR,it
YiTrt
− RPCit(1 + νc,i
t )CR,i
tYiTrt
+ (1− νFD,it ) f di
t + (1− νD,it )di
t + φR,it
− RPCit(1 + νc,i
t )Iit
YiTrt+ RPCi
t(1− νK,it )rK,i
tKi
t−1
YiTrt
(A.35)
denoting φR,it =
ΦR,it
Pit Yi . At steady state:
0 =(
R− Πi(1 + g)− ψ( f ai)) f ai
Πi(1 + g)+
1 + νc,i
Yi
[RWR,i LR,i − CR,i
]RPCi
+ (1− νFD,i) f di+ (1− νD,i)di + φR,i
+ (1− νK,i)rK,i Ki
Yi RPCi − (1 + νc,i)Ii
Yi RPCi
(A.36)
which gives the following relationship:
f ai =βi(1 + g)
1− βi(1 + g)
(1 + νc,i)RPCi
((1− sic)cyi + iyi)− 1− νD,i
θi− (1− νFD,i) ¯f di − φR,i
−(1− α)θi − 1
θi
1− siwl
1 + νw,i − α(1− νK,i)θi − 1
θi
(STEADY.10)
As for non Ricardian households, we have:
CNR,i
Yi =RWNR,i LNR,i
Yi +φNR,i
1 + νc,i1
RPCi (A.37)
RPCicyisic =
1− α
(1 + νc,i)(1 + νw,i)
(θi − 1)θi si
wl +φNR,i
1 + νc,i (STEADY.11)
58
Identically the budget constraint of governments gives in real terms:
pait =
PAit
Pit YiTrt
=(
Rt−1 − ψg(pait−1)
) pait−1
Πit(1 + g)
+ νd,it di
t + νf d,it f di
t −Gi
tYiTrt
− φit
+RPCi
tYiTrt
[(1 + νc,i
t )νw,it RWi
t Lit + νc,i
t (Cit + Ii
t) + νk,it rk,i
t Kit−1
] (A.38)
and defining pai ≡ PAk
PkYk , we obtain at steady state:
gyi + φi =
(1
βig(1 + g)
− 1 +paiψg′( pai)
Πi(1 + g)
)pai + νc,iRPCi
(cyi + iyi) +νd,i
θi + ν f d,i f di
+θi − 1
θi
(νw,i
1 + νw,i (1− α) + νk,iα
) (STEADY.12)
A.9 Trade balance
Let TBt = P1t C2
1,t + P1t I2
1,t − P2t C1
2,t − P2t I1
2,t denote the trade balance of country 1. At steady state
TBP1Y1 = tb =
C21 + I2
1Y1 −
C12 + I1
2Y1 T (A.39)
and the openness of the two countries verifies:
tb = α2 cy2 + iy2
θT1−α2 − α1(cy1 + iy1)Tα1
(STEADY.13)
Combining the three budget constraints of a country yields, after simplication of transfers and taxes,
combination of Ricardian and non Ricardian households:
FAit + PAi
t = Rt−1(FAit−1 + PAi
t−1) + TBt + FDit − ψ( f ai
t−1)FAit−1 − ψ(pai
t−1)PAit−1 (A.40)
That is in real terms:
f a1t + pa1
t =Rt−1( f a1t−1 + pa1
t−1)1
Π1t (1 + g)
+ tbt + f d1t
− ψ( f ait−1) f ai
t−11
Π1t (1 + g)
− ψ(pait−1)pai
t−11
Π1t (1 + g)
(A.41)
In words, the current account (changes in the net financial position of the country as a whole) depends
on its trade balance and net transfers with the rest of the union (interests paid on its position in the pre-
59
vious period spread included net of dividends received from the international financial intermediary).
We assume that at steady state the spread paid to the financial intermediary is offset by the dividends
received implying the following for θ f a
θ f a =θ
Tψ( f a1
) f a1+ ψg( pa1) pa1
ψ( f a2) f a2
+ ψg( pa2) pa2(A.42)
It follows at steady state that:
tb =
(1− R
Π1(1 + g)
)( f a1
+ pa1) (STEADY.14)
Given the choice of redistribution, financial dividends simplify at steady state into:
f d1=
1
Π1(1 + g)
(ψ( f a1
) f a1+ ψg(pa1)pa1
); f d
2=
1
Π2(1 + g)
(ψ( f a2
) f a2+ ψg(pa2)pa2
)(STEADY.15)
A.10 Endogenous variables in level
In the present model, growth is exogenous. In the long run, all real variables grow at the rate of TFP and
prices grow at the steady state inflation rate. Taking into account all steady state equations shows that
it is possible (and necessary) to define all endogenous variables not only in ratio to GDP but also in level.
Conveniently, for policy analysis purposes, this allows to assess how this steady state in level de-
pends on structural, technological, preference and fiscal parameters and by derivation how changes in
these parameters would influence the level of our economies.
60
Starting from the Ricardian households consumption-leisure trade-off (equation (A.10)), we get:
RWR,i LR,i
CR,i =θi
wθi
w − 1(1 + σi
l )σic
(1− σic)(1− η)
κi(1− σic)
[LR,i
(1−µi)niN
(Li
niN
)−hl]1+σi
l
1− κi(1− σic)
[LR,i
(1−µi)niN
(Li
niN
)−hl]1+σi
l(A.43)
1− siwl
1− sic
RWi Li
Ci =θi
wθi
w − 1(1 + σi
l )σic
(1− σic)(1− η)
κi[
1−siwl
1−µi
] θiw(1+σi
l )
θiw−1 (
Li)(1+σil )(1−hl)
1− κi[
1−siwl
1−µi
] θiw(1+σi
l )
θiw−1 (
Li)(1+σi
l )(1−hl)
=1− si
wl1− si
c
θi − 1θi
1− α
(1 + νc,i)(1 + νw,i)RPCicyi
(A.44)
that is
Li =
1κi
[1− µi
1− siwl
] θiw(1+σi
l )
θiw−1 BR,i
BR,i + 1
1
(1+σil )(1−hl )
(STEADY.16)
with BR,i =1−si
wl1−si
c
θiw−1θi
w
θi−1θi
(1−α)(1−σic)(1−η)
(1+σil )σ
ic(1+νc,i)(1+νw,i)RPCicyi
and κi = κi(1− σic)(niN
)−(1−hl)(1+σil )
To determine the level of output we write from the Cobb-Douglas function, simplifying with (STEADY.5):
Yi = (Ki)α(ζ Li)1−α = ζ(Ki
Yi )α
1−α Li = ζ
(iyi
δ + g
) α1−α
Li
= ζ
(α
θi − 1θi
1− νk,i
1 + νc,i (1βi− 1 + δ)−1 1
RPCi
) α1−α
Li
(A.45)
which we simplify in:
Yi = ALi (STEADY.17)
From equations (STEADY.17) and (STEADY.16) we also have the following purchasing power of
wages:
RWi= A
(1− α) θi−1θi
(1 + νc,i)(1 + νw,i)
1
RPCi (STEADY.18)
61
Finally, non Ricardians and Ricardians shares in the population and from equations (STEADY.3)
and (STEADY.4) in consumption and payroll allow to compute the steady state level of consumption,
labour supply and wage of both types of households. Similarly, from equation (STEADY.17) the steady
state levels of consumption, government spending, investment, capital and the trade balance are di-
rectly found by multiplying by cyi, gyi, iyi, iyi
δ and tb respectively. Private assets, public assets, transfers
and financial dividends in real terms can be found by multiplying Yi with f ai, pai, φi, and ¯f di. Divi-
dends from the real sector are one θith of output. Openness degrees (αi) allow to compute the share of
investment and consumption imported from the rest of the union.
B Linearisation
This appendix presents the full linearisation of the model. Equations labelled LIN are equations included
in the corresponding Dynare code (available upon request).
B.1 Goods Aggregation
Linearising the relationships in Section 2.1 gives:24
C11,t = α1Tt + C1
t C12,t = (α1 − 1)Tt + C1
t (LIN.1, LIN.2)
C21,t = (1− α2)Tt + C2
t C22,t = −α2Tt + C2
t (LIN.3, LIN.4)
Identical equalities hold for investment.
I11,t = α1Tt + I1
t I12,t = (α1 − 1)Tt + I1
t (LIN.5, LIN.6)
I21,t = (1− α2)Tt + I2
t I22,t = −α2Tt + I2
t (LIN.7, LIN.8)
24X is variable X’s log-deviation from its steady state value X.
62
B.2 Households
B.2.1 Ricardian Households Euler Equations
The Euler equation for Ricardian households is given by equation (2.25). Taking the log-linearisation
around the steady-state and simplifying the relationship yields:
0 =(
1− (1− σic)(1− η)
) [CR,i
t+1 − CR,it
]+ hi
c(1− σic)(1− η)
[Ci
t − Cit−1
]−(1− σi
c)η[
Git+1 − Gi
t
]+ hi
g(1− σic)η[
Git − Gi
t−1
]+(1 + σi
l )σicBR,i
[(LR,i
t+1 − LR,it )− hi
l(Lit − Li
t−1)]+ Πc,i
t+1 +νc,i
1 + νc,i (νc,it+1 − νc,i
t )
− 1
R− ψ( f ai)
(RRt − ψ′( f ai
) f ai f ait
)(LIN.9)
B.2.2 Tobin’s Q and investment decision
Linearising around the steady-state and simplifying the relationships yield:
0 = qit + εi,I
t − S′′(1 + g)(1 + g)2
(Iit − Ii
t−1 − βi(1 + g)( Iit+1 − Ii
t))
(LIN.10)
qit =
((1− σi
c)(1− η)− 1) [
CR,it+1 − CR,i
t
]− hi
c(1− σic)(1− η)
[Ci
t − Cit−1
]+(1− σi
c)η[
Git+1 − Gi
t
]− hi
g(1− σic)η[
Git − Gi
t−1
]−(1 + σi
l )σicBR,i
[(LR,i
t+1 − LR,it )− hi
l(Lit − Li
t−1)]
+ βi(1− δ)qit+1 + βi rK,i (1− νK,i)
(1 + νC,i)
[rK,i
t+1 −νK,i
1− νK,i νK,it+1 −
νC,i
1 + νC,i νC,it+1
](LIN.11)
B.2.3 Capital dynamics
The dynamics of capital is given by:
(1 + g)Kit = (1− δ)Ki
t−1 + (δ + g)( Iit + εi,I
t ) (LIN.12)
B.2.4 Ricardians’ and non Ricardians’ aggregation
63
Cji,t = si
cCj,NRi,t + (1− si
c)Cj,Ri,t (LIN.13)
Cit = si
cCNR,it + (1− si
c)CR,it (LIN.14)
RWit = si
wl RWNR,it + (1− si
wl)RWR,it (LIN.15)
Lit = si
wl LNR,it + (1− si
wl)LR,it (LIN.16)
LNR,it = Li
t − θiw(RW
NR,it − RW
it) (LIN.17)
LR,it = Li
t − θiw(RW
R,it − RW
it) (LIN.18)
Equations (LIN.15) to (LIN.18) are collinear. Only three of them shall be put in the linearised model.
B.3 Firms
B.3.1 Production
From Section 2.3 under the assumption that firms’ dispersion is a second order phenomenon (i.e. ∆it =
0), we have:
Y1t = (1− α)ζ1
t + (1− α)L1t + αK1
t−1, (LIN.19)
Y2t = (1− α)ζ2
t + (1− α)L2t + αK2
t−1. (LIN.20)
where ζt = εζt .
B.3.2 Capital and labour arbitrage
rK,it + Ki
t−1 = RWit +
νw,i
1− νw,i νw,it +
νC,i
1 + νC,i νC,it + Li
t (LIN.21)
B.3.3 Dividends
Dividends equal output minus the sum of wages and capital costs, namely:
Dit = Pi
t Yit −
11− α
Wit (1 + νw,i
t )Lit = Pi
t Yit (1− RMCi
t) (B.1)
Dividing both sides of the equation by Pit YiTrt:
Dit
Pit YiTrt
=Yi
tYiTrt
(1− RMCit) (B.2)
64
The linearised equation for dividends is:
dt = Yit − (θi − 1)RMC
it (LIN.22)
with RMCit = RPC
it + (1− α)
(RW
it − ζ i
t +νc,i
1 + νc,i νc,it +
νw,i
1 + νw,i νw,it
)+ αrk,i
t (LIN.23)
B.4 Fiscal Authorities
Proceeding like for households’ Euler equations, linearising equation (2.70) we get:
0 =(1− βighi
g)
(βi
g
Πi
[RRt − (2ψg′(pai) + ψg′′(pai)pai)pai pai
t
]− Πi
t+1 − ∆Git+1
)+(1− σi
c)η((1 + βi
ghig
2)∆Gi
t+1 − βighi
g∆Git+2 − hi
g∆Git
)+(1− σi
c)(1− η)hic(βi
ghig∆Ci
t+1 − ∆Cit)
+(1− σic)(1− η)
(δΩ∆CR,i
t+1 + (1− δΩ)∆CNR,it+1 − βi
ghig(δΩ∆CR,i
t+2 + (1− δΩ)∆CNR,it+2 )
)−σi
c(1 + σil )δΩBR,i
[∆LR,i
t+1 − hil∆Li
t − βighi
g(∆LR,it+2 − hi
l∆Lit+1)
]−σi
c(1 + σil )(1− δΩ)BNR,i
[∆LNR,i
t+1 − hil∆Li
t − βighi
g(∆LNR,it+2 − hi
l∆Lit+1)
]
(LIN.24)
where
δΩ =1
1 +1−ωR,i
g
ωR,ig
(si
c(1−µi)
(1−sic)µi
)(1−η)(1−σic)+σi
c(
siwl(1−µi)
(1−siwl)µ
i
) (1+σil )θ
iwσi
cθiw−1
−σic
(B.3)
We compare this equation with a budget rule similar to Corsetti et al. (2010):
Git = ΨggGi
t +Ψgy
gyi Yit +
Ψgd pai
gyi pait (LIN.25)
65
B.5 Financial intermediary
B.5.1 Dividends
Regarding the financial intermediaries, dividends write:
f d1(1 + g)Π1
( f d1t + Π1
t ) =θ f a
1 + θ f a
(ψ( f a1
) + ψ′( f a1) f a1
)f a1 f a
1t−1
+(
ψg(pa1) + ψg′(pa1)pa1)
pa1 pa1t−1
+Tθ
[ (ψ( f a2
) + ψ′( f a2) f a2
)f a2 f a
2t−1 +
(ψg(pa2) + ψg′(pa2)pa2
)pa2 pa2
t−1
+
(ψ( f a2
) f a2+ ψg(pa2)pa2
)Tt−1
](LIN.26)
f d2(1 + g)Π2
( f d2t + Π2
t ) =1
1 + θ f a
(ψ( f a2
) + ψ′( f a2) f a2
)f a2 f a
2t−1
+(
ψg(pa2) + ψg′(pa2)pa2)
pa2 pa2t−1
+θ
T
[ (ψ( f a1
) + ψ′( f a1) f a1
)f a1 f a
1t−1 +
(ψg(pa1) + ψg′(pa1)pa1
)pa1 pa1
t−1
−(
ψ( f a1) f a1
+ ψg(pa1)pa1
)Tt−1
](LIN.27)
B.5.2 Zero cash needs condition
We showed that to eliminate a unit root in the model, one should replace one of the budget constraint
of governments or Ricardian households by a zero cash needs condition on the financial intermediation
market. Once linearised, this condition reads:
0 = f a1 f a1t + pa1 pa1
t +T f a2
θ
(f a
2t + Tt
)+
T pa2
θ
(pa2
t + Tt
)(LIN.28)
66
B.6 Monetary Policy, Relative Prices, Inflations
The linearised Taylor rule reads:
Rt = ρRt−1 + (1− ρ)
(rΠ
4
t
∑k=t−3
Πunionk + ry
ˆYt
)(LIN.29)
where the weighted averages for inflation and output in the monetary union are:
ˆYuniont =
1
1 + Tθ
Y1t +
11 + θ
T
Y2t (LIN.30)
Πunion,VATt =
1
1 + (1+νC,2)T1−α1−α2 cy2(1+νC,1)θcy1
(Πc,1
t +νC,1
1 + νC,1 (νC,1t − νC,1
t−1)
)
+1
1 + (1+νC,1)θcy1
(1+νC,2)T1−α1−α2 cy2
(Πc,2
t +νC,2
1 + νC,2 (νC,2t − νC,2
t−1)
) (LIN.31)
By definition of CPIit = Pi
t1−αi
Pjt
αi
we get:
RPC1t = α1Tt (LIN.32)
RPC2t = −α2Tt (LIN.33)
Πc,1t = (1− α1)Π1
t + α1Π2t (LIN.34)
Πc,2t = (1− α2)Π2
t + α2Π1t (LIN.35)
And the terms of trade depend on inflation differentials between countries:
Tt = Tt−1 + Π2t − Π1
t (LIN.36)
B.7 Market clearing
From the market-clearing equations, we have:
Y1t =
C11
Y1 C11,t +
C21
Y1 C21,t +
G1
Y1 G1t +
I11
Y1 I11,t +
I21
Y1 I21,t (B.4)
Y2t =
C12
Y2 C12,t +
C22
Y2 C22,t +
G2
Y2 G2t +
I12
Y2 I12,t +
I22
Y2 I22,t (B.5)
67
Y1t = (1− α1)cy1Tα1
C11,t + α2cy2
1θ
T1−α2C2
1,t + gy1G1t + (1− α1)iy1Tα1
I11,t + α2iy2
1θ
T1−α2I21,t
(LIN.37)
Y2t = α1cy1θTα1−1C1
2,t + (1− α2)cy2T−α2C2
2,t + gy2G2t + α1iy1θTα1−1 I1
2,t + (1− α2)iy2T−α2I22,t
(LIN.38)
B.8 Budget constraints
B.8.1 Ricardian households
Using the steady state relationships, we linearise as follows:
f aif a
it =
f ai
βi(1 + g)
(R
R− ψ( f ai)
Rit−1 +
(1− ψ′( f ai
) f ai
R− ψ( f ai)
)f a
it−1 − Πi
t
)
+ (1− siwl)
1− α
1 + νw,iθi − 1
θi
[RPC
it +
νc,i
1 + νc,i νc,it + RW
R,it + LR,i
t
]− (1 + νi
c)RPCicyi(1− sic)
[RPC
it +
νc,i
1 + νc,i νc,it + CR,i
t
]− (1 + νi
c)RPCiiyi[
RPCit +
νc,i
1 + νc,i νc,it + Ii
t
]+ φR,iφR,i
t
+ (1− νD,i)1θi
[di
t −νD,i
1− νD,i νD,it
]+ (1− νFD,i) f d
i[
f dit −
νFD,i
1− νFD,i νFD,it
]+ (1− νK,i)α
θi − 1θi
[RPC
it + rK,i
t + Kit−1 −
νK,i
1− νK,i νK,it
]
(LIN.39)
B.8.2 Non Ricardian households
Similarly, using the steady state relationships, for the non Ricardian households, we linearise their budget
constraint as follows:
CNR,it =
(1− φNR,i
(1 + νc,i)RPCicyisic
)(RW
NR,it + LNR,i
t
)+
φNR,i
(1 + νc,i)RPCicyisic
(φNR,i
t − RPCit −
νc,i
1 + νc,i νc,it
) (LIN.40)
68
B.8.3 Governments
Similarly to non Ricardian households, linearising gives:
pai pait = pai
(1
βig(1 + g)
+ψg′( pai) pai
Πi(1 + g)
)[R
R− ψg( pai)Ri
t−1 +
(1− ψg′( pai) pai
R− ψg( pai)
)pai
t−1 − Πit
]+
νd,i
θi
(νd,i
t + dit
)+ ν f d,i f d
i(
νf d,it + ˆf d
it
)− gyiGi
t − φiφit
+
(θi − 1
θiνw,i
1 + νw,i (1− α) + νc,iRPCi(cyi + iyi) +
θi − 1θi νk,iα
)RPC
it
+θi − 1
θiνw,i
1 + νw,i (1− α)
(RW
it +
νc,i
1 + νc,i νc,it + Li
t + νw,it
)+νc,iRPCicyi
(Ci
t + νc,it
)+ νc,iRPCiiyi
(Iit + νc,i
t
)+
θi − 1θi νk,iα
(Ki
t−1 + νk,it + rk,i
t
)(LIN.41)
B.9 Phillips Curves
B.9.1 Prices
From the maximisation of the firm’s programme and the linearisation of the first-order conditions, we
have:
0 =∞
∑T=t
(βiξ i)T−tλiT
YiT
piP
(Pi
t (ε)ΓT−1t
PiT
)−θi (Pi
t (ε)ΓT−1t − θi
θi − 1MCi
T
)(B.41)
0 =∞
∑T=t
(βiξ i(1 + g))T−t
( Pit (ε)
Pit−
T
∑k=t+1
(Πi
k − γiΠik−1
)− RMC
iT
)(B.42)
We also know that Pi(ε, t) is independent from ε since all firms solve the same program, and the
Calvo process on prices yields:
Pit (ε)
Pit
=ξ i
1− ξ i
(Πi
t − γiΠit−1
). (B.43)
69
Plugging this into the equation above, and differentiating between time t and time t + 1, we obtain
a standard linearised New-Keynesian Phillips curve for prices:
Πit−γiΠi
t−1 = βi(1 + g)(
Πit+1 − γiΠi
t
)+
(1− βiξ i(1 + g))(1− ξ i)
ξ i RMCit, (LIN.42)
where inflation depends positively on past indexed inflation, future anticipated inflation, relative
prices and wages, taxes, total output in country i and negatively on productivity.
B.9.2 Wages
We recall equation (2.35):
0 = Et
∞
∑T=t
(ξ iwβi)T−t lt,T(τ)
U ′(CiT(τ), Ci
T−1)V(
li t,T(τ), LiT−1
)W(Gi
T , GiT−1)
U ′(Cit(τ), Ci
t−1)V(li t(τ), Lit−1)W(Gi
t, Git−1)[U (Ci
T(τ), CiT−1)V ′
(li t,T(τ), Li
T−1
)U ′(Ci
T(τ), CiT−1)V
(li t,T(τ), Li
T−1
) +θi
w − 1θi
w
wit(τ)Γ
T−1w,t
CPIiT(1 + νc,i
T )
]
As an example, we derive here the wage Phillips curve for Ricardian households. The first order
condition is linearised into:
0 =∞
∑T=t
(βiξ iw(1 + g))T−t
RW
R,iT +
(1 + θi
w((1 + σil )(1 + B
R,i)− 1)) (
wR,it ΓT−1
w,t
wR,iT
)
+ hil(1 + σi
l )(1 + BR,i)Li
T−1 − ((1 + σil )(1 + B
R,i)− 1)LR,iT − CR,i
T
(B.44)
Differentiating between time t and time t + 1 yields:
0 = −RWR,it −(1 + θi
w((1 + σil )(1 + B
R,i)− 1))δwR,it − hi
l(1 + σil )(1 + B
R,i)Lit−1
+((1 + σil )(1 + B
R,i)− 1)LR,it + CR,i
t
+βiξ i
w(1 + g)(1 + θiw((1 + σi
l )(1 + BR,i)− 1))
1− βiξ iw(1 + g)[
−γiwΠc,i
t + δwR,it+1 − δw
R,it + RW
R,it+1 − RW
R,it + Πc,i
t+1 +νc,i
1 + νc,i (νc,it+1 − νc,i
t )
](B.45)
70
And the Calvo process induces:
0 = ξ iw
(RW
R,it−1 − RW
R,it − (Πc,i
t − γiwΠc,i
t−1)−νc,i
1 + νc,i (νc,it − νc,i
t−1)
)+ (1− ξ i
w) ˆδwR,it (B.46)
ˆδwR,it =
ξ iw
1− ξ iw
(RW
R,it − RW
R,it−1 + (Πc,i
t − γiwΠc,i
t−1) +νc,i
1 + νc,i (νc,it − νc,i
t−1)
). (B.47)
where one may read RWR,it − RW
R,it−1 + Πc,i
t + νc,i
1+νc,i (νc,it − νc,i
t−1) more simply as nominal wage inflation
of the Ricardians in deviation from its steady state. In the end, the wage Phillips curve for the Ricardian
households writes:
RWR,it −RW
R,it−1 + (Πc,i
t − γiwΠc,i
t−1) +νc,i
1 + νc,i (νc,it − νc,i
t−1) =
βi(1 + g)(
RWR,it+1 − RW
R,it +
νc,i
1 + νc,i
(νc,i
t+1 − νc,it
)+(
Πc,it+1 − γi
wΠc,it
))+
(1− βiξ iw(1 + g))(1− ξ i
w)
ξ iw(1 + θi
w((1 + σil )(1 + BR,i)− 1))
[−RW
R,it − LR,i
t + (1 + σil )(1 + B
R,i)(LR,it − hi
l Lit−1) + CR,i
t
].
(LIN.43)
The wage Phillips curve for the non Ricardian households writes:
RWNR,it −RW
NR,it−1 + (Πc,i
t − γiwΠc,i
t−1) +νc,i
1 + νc,i (νc,it − νc,i
t−1) =
βi(1 + g)(
RWNR,it+1 − RW
NR,it +
νc,i
1 + νc,i
(νc,i
t+1 − νc,it
)+(
Πc,it+1 − γi
wΠc,it
))+
(1− βiξ iw(1 + g))(1− ξ i
w)
ξ iw(1 + θi
w((1 + σil )(1 + BNR,i)− 1))[
−RWNR,it − LNR,i
t + (1 + σil )(1 + B
NR,i)(LNR,it − hi
l Lit−1) + CNR,i
t
].
(LIN.44)
with BNR,i =si
wlsi
c
θiw−1θi
w
θi−1θi
(1−α)(1−σic)(1−η)
(1+σil )(1+νc,i)(1+νw,i)RPCicyi
so that the slope of the wage Phillips Curve very
marginally depends on the household’s type.
In all, the level of inflation on wages hinges positively on future anticipated inflation, past inflation,
taxes. The larger the discount factor βi, the more sensitive is wage inflation to inflation expectations.
The larger the probability to adjust prices 1− ξ i (i.e. the more flexible are prices and wages), the less in-
flation depends on expectations. Inflation also positively depends on labour demand and consumption:
households demand a higher wage to work more and consequently consume more.
71
C Monetary policy in a currency union
C.1 Defining aggregate indexes for the union
In the absence of exact derivations for both the union aggregate price index and the union total output
gap, we turn to the definitions of such indexes. Namely, the aggregate price index comes from :
(1 + νC,uniont )CPIunion
t Cuniont = (1 + νC,1
t )CPI1t C1
t + (1 + νC,2t )CPI2
t C2t (C.1)
therefore the VAT-included CPI inflation for the union rewrites
Πunion,VATt =
(1 + νC,uniont )CPIunion
t
(1 + νC,uniont−1 )CPIunion
t−1
=Cunion
t−1
Cuniont
C1t
C1t−1
1
1 +(1+νC,2
t−1)CPI2t−1C2
t−1
(1+νC,1t−1)CPI1
t−1C1t−1
ΠC,1,VATt +
Cuniont−1
Cuniont
C2t
C2t−1
1
1 +(1+νC,1
t−1)CPI1t−1C1
t−1
(1+νC,2t−1)CPI2
t−1C2t−1
ΠC,2,VATt
(C.2)
Our approximation relies on the choice of weights on inflations taken at their steady state values. As a
result, we obtain:
Πunion,VATt =
1
1 + (1+νC,2)CPI2C2
(1+νC,1)CPI1C1
ΠC,1,VATt +
1
1 + (1+νC,1)CPI1C1
(1+νC,2)CPI2C2
ΠC,2,VATt (C.3)
Similarly, we define the aggregate union production as follows:
Puniont Yunion
t = P1t Y1
t + P2t Y2
t (C.4)
therefor, taking the static weights at their steady state value, the union total output gap rewrites:
Yuniont =
Yuniont
Yunion =
P1t
PuniontP1
Punion
1
1 + Tθ
Y1t
Y1 +
P2t
PuniontP2
Punion
11 + θ
T
Y2t
Y2 =1
1 + Tθ
Y1t
Y1 +1
1 + θT
Y2t
Y2 (C.5)
C.2 Should inflation be net of VAT?
The ECB measures inflation with the Harmonised Index of Consumer Prices (HICP) which includes
VAT. However, changes in the tax rate seldom happen and one may expect the ECB not to react to such
72
shocks. It is then possible to define a VAT-excluded alternative:
Πuniont =
1
1 + CPI2C2
CPI1C1
ΠC,1t +
1
1 + CPI1C1
CPI2C2
ΠC,2t (C.6)
as the inflation targeted by the Central Banker. One can also easily consider other Taylor rules, with
quarter to quarter inflation, output growth, GDP or GDP growth...
C.3 Defining GDP in this context
When there are no taxes on products or subsidies on production, Yit is equal to GDP. Stricto sensu Yi
t
is actually the total value added produced in country i. In a model with taxes, in particular VAT, Yit
actually differs from GDP. In nominal GDP is:
GDPi,nominalt = Pi
t Yit + νc,i
t CPIit(C
it + Ii
t) (C.7)
Changes in the VAT rate are considered as price effects so that the growth index of GDP in real terms
can be measured by:25
IndexGDPrt =
Pit−1Yi
t + νc,it−1CPIi
t−1(Cit + Ii
t)
Pit−1Yi
t−1 + νc,it−1CPIi
t−1(Cit−1 + Ii
t−1)(C.8)
An additional issue is the production of financial intermediation services. In the present model we have
considered that they are not located in country 1 or 2. Still union wide GDP should incorporate this
production.
GDPunion,nominalt = ∑
i∈1,2Pi
t Yit + νc,i
t CPIit(C
it + Ii
t)+ ∑i∈1,2
ψ
(FAi
t−1
Pit−1YiTrt−1
)FAi
t−1 +ψg
(PAi
t−1
Pit−1YiTrt−1
)PAi
t−1
(C.9)
For financial production, the volume index is the growth rate of financial assets in real terms, thecorresponding growth index in volume is:
IndexGDPunionr
t = (C.10)
∑i∈1,2 Pit−1Yi
t + νc,it−1CPIi
t−1(Cit + Ii
t) + ∑i∈1,2 ψ
(FAi
t−2Pi
t−2Yi Trt−2
)Pi
t−2FAi
t−1Pi
t−1+ ψg
(PAi
t−2Pi
t−2Yi Trt−2
)Pi
t−2PAi
t−1Pi
t−1
∑i∈1,2 Pit−1Yi
t−1 + νc,it−1CPIi
t−1(Cit−1 + Ii
t−1) + ∑i∈1,2 ψ
(FAi
t−2Pi
t−2Yi Trt−2
)FAi
t−2 + ψg(
PAit−2
Pit−2Yi Trt−2
)PAi
t−2
25This formula corresponds to the growth rate of real GDP when measured in over the quarter overlap chained linked volumes.This technique is seldom used by national accountants but other techniques require to isolate quarters of the same year which isnot possible in the present model.
73
D Firms and households dispersion
Contrary to a flexible price model, price and wage nominal rigidities create distributional issues. As
firms and agents cannot reset their prices and wages every period, some agents will keep a constant
price (in the absence of indexation) for non-zero periods of time. As a result, at each date, a full distri-
bution of prices and wages coexists.
Working with aggregate variables, this distributional issue rarely matters and is usually invisible. It
does only matter for aggregation across firms and households. In particular, the link between an individ-
ual household consumption/labour and the aggregate consumption/labour determines the aggregate
Euler equations 2.25, investment decision 2.26, Tobin’s Q 2.27 and wage Phillips curves 2.35.
In this appendix, we show that under the reasonable assumption that the variance of dispersion in
consumption, wages, prices or labour is small, the first order conditions aggregates exactly at the first
order. As an alternative in the literature, papers such as Erceg, Henderson, and Levin (2000) resort to the
assumption of complete contingent claims markets for consumption and labour. In this case, households
can perfectly ensure against consumption and labour fluctuations, and consumption is constant across
similar households.
D.1 Labour dispersion
Recall the aggregate production (Equation 2.58) that depends on the index of labour dispersion across
firms. This index is defined as follows:
(∆i
t
) θi−1θi =
1piP
∫ piP
0
(Li(ε, t)Li
t/piP
) θi−1θi
dε︸ ︷︷ ︸H
(D.1)
74
Using a second-order Taylor expansion of H, we obtain:
H = 1 +1
piP
∫ piP
0
θi − 1θi
[Li(ε, t)Li
t/piP− 1
]− 1
2θi
[Li(ε, t)Li
t/piP− 1
]2
dε
= 1− 12piP
1θi
∫ piP
0
[Li(ε, t)Li
t/piP− 1
]2
dε
as Lit =
∫ piP
0Li(ε, t)dε
(D.2)
We notice that the first order term disappears in the Taylor expansion of the dispersion index. As
a result, it appears that labour dispersion is a purely second order phenomenon under the reasonable
assumption that Li(ε,t)Li
t/piP− 1 remains small. As we study a first-order linearised model, this assumption
of a small variance of dispersion seems rather reasonable. We can therefore neglect labour dispersion
for production aggregation.
D.2 Ricardian first order conditions
Similarly, it is possible to aggregate the Ricardians’ first order conditions at the first order, assuming that
individual consumption and labour fluctuations around the aggregate level are of small magnitude.
In the Euler equation, the investment decision and Tobin’s Q equation for the Ricardian agents, the
part of the subjective discount factor depending on τ is given by (eliminating external habits constant
across τ) :
f(
Cit+1(τ), Ci
t(τ), lit+1(τ), li
t(τ))=U ′(Ci
t+1(τ), Cit)V(lit+1(τ), Li
t)
U ′(Ci
t(τ), Cit−1)V(lit(τ), Li
t−1)
=
(Ci
t+1(τ)
Cit(τ)
)(1−σic)(1−ηi)−1(
lit+1(τ)
lit(τ)
)σic
1− κ(1− σic)[lit+1(τ)(Li
t)−hi
c
]1+σil
1− κ(1− σic)[lit(τ)(Li
t−1)−hi
c
]1+σil
σi
c−1
(D.3)
75
Conducting a Taylor expansion of this function f around the mean(
CR,it+1
(1−µi)niN, CR,i
t(1−µi)niN
,LR,i
t+1(1−µi)niN
, LR,it
(1−µi)niN
)and at the second order gives:
f(
Cit+1(τ), Ci
t(τ), lit+1(τ), li
t(τ))= f
(CR,i
t+1
(1− µi)niN,
CR,it
(1− µi)niN,
LR,it+1
(1− µi)niN,
LR,it
(1− µi)niN
)︸ ︷︷ ︸
Aggregate subjective discount factor1 +
((1− σi
c)(1− ηi)− 1)(Ci
t+1(τ)(1− µi)niN
CR,it+1
− 1
)
−((1− σi
c)(1− ηi)− 1)(Ci
t(τ)(1− µi)niN
CR,it
− 1
)
+ σic
1− σic − 1σi
c
κ(1− σic)[
LR,it+1(Li
t)−hi
c
]1+σil
1− κ(1− σic)[
LR,it+1(Li
t)−hi
c
]1+σil
(
lit+1(τ)(1− µi)niN
LR,it+1
− 1
)
− σic
1− σic − 1σi
c
κ(1− σic)[
LR,it (Li
t−1)−hi
c
]1+σil
1− κ(1− σic)[
LR,it (Li
t−1)−hi
c
]1+σil
(
lit(τ)(1− µi)niN
LR,it
− 1
)
+ o
(∥∥∥∥∥(
Ci(τ)(1− µi)niN
CR,i − 1)
,(
li(τ)(1− µi)niN
LR,i − 1))∥∥∥∥∥
2
(D.4)
As∫
R,i C(τ, t)dτ = CR,it and
∫R,i l(τ, t)dτ = LR,i
t , first order terms disappears when we aggregate the
subjective discount factor (or directly the equation) by integrating over all τ. In the end, the first order
conditions for the households is valid at the first order as long as consumption and labour dispersion
are indeed second order phenomenons, that is their respective variance is small.
The same methodology can be used to justify the use of aggregate Lagrange multipliers in both the
price and wage Calvo setting programs.
E Determination of a unique and stable steady state
E.1 Debt elastic spreads
Schmitt-Grohé and Uribe (2003) show on the case of a small open economy that a debt elastic interest
rate premium pins down a unique steady state debt level because this specification modifies the steady
state of households’ Euler equation. Following the general specification used for monetary union mod-
els, we introduce such a modification in our model for the same result: equation (STEADY.1) relates the
76
discount factor and the steady state real monetary rate with the steady state asset to GDP ratio. Put
differently, with this specification the steady state is unique in f ai and verifies equation (STEADY.1).
Moreover, the Euler equations on private consumption shows that if the spread is an increasing
function of the asset to GDP ratio, when assets are below their steady state value, consumption will
decrease today and yield an increase in assets converging to their steady state (consumption could also
increase tomorrow, a strategy which would eventually violate the no Ponzi condition).
A symmetric result holds for public assets (pai) with our maximizing government. Otherwise a bud-
get rule must incorporate a feed back loop between the level of public debt and public spending (or
another fiscal instrument).
Because the adjustment is through final demand, in terms of country i net financial position, it will
show on the trade balance rather than the net transfers to the rest of the union.
Putting together the budget constraints in each country, aggregating Ricardian and non Ricardian
households, yields:
FAit + PAi
t = (Rt−1)(FAit−1 + PAi
t−1) + TBt + FDit − Spreadsi
t (E.1)
If there were no fees for financial intermediation (FDit = Spreadsi
t = 0), any temporary imbalance of
the trade balance (TBt) would make country i’s net financial assets (FAit + PAi
t) diverge.
Convergence is nevertheless not obtain through financial transfers with the rest of the union but be-
cause imbalances of the trade balance are not simply transitory. The spread paid on assets in the Euler
equation implies that as long as the net financial position is not at steady state, either on the private or
public side, private or public consumption will compensate and the trade balance will not return to its
steady state. This is the main mechanism for convergence to the steady state.
With the introduction of financial intermediation, when the net financial position deviates from its
steady state value, fees paid to the financial market increase in both countries and so do the dividends
paid to domestic owners of financial firms. Thus, this mechanism is in first approach neutral on the
convergence to the steady state.
77
E.2 An inconvenient eigenvalue
However, to close the model, the four steady state net financial positions can not be chosen freely. More-
over, the model does not converge towards its steady state when the four asset dynamics are written.
From the model, we have the following six budget constraints:
PAit =
(Rt−1 − ψg(
PAit−1
Pit−1YiTrt−1
)
)PAi
t−1 − Pit Gi
t −Φit
+νw,it Wi
t Lit + νk,i
t rk,it CPIi
tKit−1 + νc,i
t CPIit(C
it + Ii
t) + νD,it Di
t + νFD,it FDi
t
(E.2)
FAit =
(Rt−1 − ψ(
FAit−1
Pit−1YiTrt−1
)
)FAi
t−1 − CPIit(1 + νc,i
t )CR,it − CPIi
t(1 + νc,it )Ii
t
+WR,it LR,i
t + ΦR,it + (1− νK,i
t )CPIitr
k,it Ki
t−1 + (1− νD,it )Di
t + (1− νFD,it )FDi
t
(E.3)
WNR,it LNR,i
t + ΦNR,it = CPIi
t(1 + νc,it )CNR,i
t (E.4)
Adding up these equations union-wide yields:
FA1t + FA2
t + PA1t + PA2
t =
Rt−1(FA1t−1 + FA2
t−1 + PA1t−1 + PA2
t−1)
+W1t L1
t + (1− νK,1t )CPI1
t rk,1t K1
t−1 + (1− νD,1t )D1
t + (1− νFD,1t )FD1
t + Φ1t
+W2t L2
t + (1− νK,2t )CPI2
t rk,2t K2
t−1 + (1− νD,2t )D2
t + (1− νFD,2t )FD2
t + Φ2t
−CPI1t (1 + νc,1
t )(C1t + I1
t )− CPI2t (1 + νc,2
t )(C2t + I2
t )
+νw,1t W1
t L1t + νc,1
t CPI1t (C
1t + I1
t ) + νk,1t rk,1
t CPI1t K1
t−1 + νD,1t D1
t + νFD,1t FD1
t
+νw,2t W2
t L2t + νc,2
t CPI2t (C
2t + I2
t ) + νk,2t rk,2
t CPI2t K2
t−1 + νD,2t D2
t + νFD,2t FD2
t
−P1t G1
t −Φ1t − P2
t G2t −Φ2
t
− ψ
(FA1
t−1
P1t−1Y1Trt−1
)FA1
t−1 − ψ
(FA2
t−1
P2t−1Y2Trt−1
)FA2
t−1
− ψg
(PA1
t−1
P1t−1Y1Trt−1
)PA1
t−1 − ψg
(PA2
t−1
P2t−1Y2Trt−1
)PA2
t−1
(E.5)
78
Rearranged, after simplification of taxes, transfers into:
FA1t + FA2
t + PA1t + PA2
t = Rt−1(FA1t−1 + FA2
t−1 + PA1t−1 + PA2
t−1)
+W1t (1 + νw,1
t )L1t + CPI1
t rk,1t K1
t−1 + D1t + W2
t (1 + νw,2t )L2
t + CPI2t rk,2
t K2t−1 + D2
t
−CPI1t (C
1t + I1
t )− CPI2t (C
2t + I2
t )− P1t G1
t − P2t G2
t + FD1t + FD2
t
− ψ
(FA1
t−1
P1t−1Y1Trt−1
)FA1
t−1 − ψ
(FA2
t−1
P2t−1Y2Trt−1
)FA2
t−1
− ψg
(PA1
t−1
P1t−1Y1Trt−1
)PA1
t−1 − ψg
(PA2
t−1
P2t−1Y2Trt−1
)PA2
t−1
(E.6)
We have assumed that the labour, capital and goods markets are in equilibrium, hence we can cancel out
the difference between two classic break downs of production (i.e. total revenue minus total demand in
both countries) and simplify fisims and financial dividends (union wide financial production), the sum
yields:
CNt = Rt−1CNt−1. (E.7)
Assuming the nullity of either initial or final conditions of total net asset holding in the monetary union,
i.e. CN0 or CN∞ = 0 is sufficient to have financial intermediaries clearing their position vis-à-vis the
central bank at each period.
All the asset dynamics but one and the zero cash needs constraint (2.79) at steady state is enough to
make the model Walrassian while verifying the fourth budget constraint. On the contrary, resorting
to the four asset dynamic equations would introduce a diverging dynamic in the model (verifying
CNt = Rt−1CNt−1).26 The zero cash needs constraint is thus a necessary condition with a reasonable
economic interpretation.
F Other IRFs
26A property we verified numerically.
79
y- axis in p.p. deviation from steady state
Figure 10: IRFs to a one percent productivity shock (autocorrelated)
80
y- axis in p.p. deviation from steady state
Figure 11: IRFs to a one percent preference shock (autocorrelated)
81
y- axis in p.p. deviation from steady state
Figure 12: IRFs to a 100 basis points monetary policy shock
82
y- axis in p.p. deviation from steady state
Figure 13: IRFs to a one percent government spending shock (autocorrelated)
83
y- axis in p.p. deviation from steady state
Figure 14: IRFs to a one percent government spending shock (autocorrelated)
84
Production and consumptionYi
t , yit(ε) Output in country i, resp. of firm ε
Cit, Ii
t Consumption and investment in country i
CR,it , CR,i
j,t , CNR,it , CNR,i
j,tConsumption of Ricardian (resp. non Ricardian) households in country i andof goods produced in country j
Cij,t, Ii
j,t Consumption and investment in country i of goods produced in country jCi,t, Ii,t Consumption and investment of goods produced in country j
λit, λR,i
t , λNR,it Marginal utility of consumption
Dit, Di
t(τ) Dividends received in country i or by household τ
MCit(ε), MCi
t, RMCit Marginal cost of production of firm ε, in country i and in real terms
Xit, Mi
t Exports from and imports to country iCapital market
Kit, Ki
t(τ), Kit(ε) Real capital stock in country i, of household τ or firm ε
rk,it Returns on capital
Labour market
Lit, li
t(τ), LNR,it , LR,i
tLabour supply in country i, of household τ, of non Ricardian or Ricardianhouseholds
Wit , wi
t(τ), WNR,it , WR,i
t Wages in country i, of household τ, of non Ricardian or Ricardian households
wit(τ), WR,i
t , WNR,it
Optimal wage reset in country i by household τ, Ricardian or non Ricardianhouseholds
RWit Real wage in country i, i.e. purchasing power net of taxes incl. VAT
Financial marketsFAi
t, FAit(τ) Financial assets in country i (resp. of household τ)
f ait = FAi
t/Pit YiTrt, ψ( f ai
t)Real financial asset to GDP ratio and intermediation fees paid on theseassets in country i
PAit Public assets in country i
pait, ψg(pai
t)Real public asset to GDP ratio and intermediation fees paid on these assetsin country i
FDit, FDi
t(τ) Financial dividends received in country i or by household τFYt Production of the financial sectorCNt Aggregate cash needs of the financial sector
Public sectorGi
t Public expenditure in country i
Φit, Φi
t(τ), ΦR,it , ΦNR,i
tLump-sum transfers in country i, to household τ, to Ricardian and nonRicardian households
φit, φi(τ, t), φR,i
t , φNR,it Real lump-sum public transfers to GDP ratios
νc,it , νw,i
t Tax rates on consumption and wagesνk,i
t , νd,it , ν
f d,it Tax rates on capital returns, dividends and financial dividends
Prices and inflationPi
t , Pit (ε) Production price in country i, resp. of firm ε
Pit , Pi
t (ε) Optimal price when reset in country i, resp. of firm ε
∆it Dispersion index of firm size in country i
CPIit Consumption price in country i, VAT-excluded
RPCit = CPIi
t/Pit Relative price of consumption in country i
Tt = P2t /P1
t Terms of tradeΠi
t, Πc,it Inflation of production and consumption prices in country i
ΓT−1t , ΓT−1
w,t Indexation of prices and wages if not resetMonetary policy
Yt, Πuniont Output and inflation targets of the Central Banker
Rt Monetary policy rate
GDPi,nomt , IGDPi
rt Nominal GDP and growth index of real GDP in country i
Table 3: Definition of the endogenous variables
85
Structural parametersp, P, pi Relative, absolute and national varieties numbern, N, ni Relative, absolute and national population size
g Productivity growthθi, θi
w Elasticity of substitution between goods/labour in country iαi Import share of country i
βi, βig Households and government discount factors
σic, σi
lInverse intertemporal elasticity of substitution and inverse Frisch elasticityin country i
ηi Weight of public consumption in utility from consumption in country ihi
c, hil Habit parameters on consumption and labour in country i
hig Habit parameter on government consumption in country i
κi Labour disutility weightγi
p, γiw Prices and wages indexation on past inflation in country i
1− ξ i, 1− ξ iw Calvo resetting probabilities on prices and wages in country i
µi Share of non Ricardian households in country iδ Depreciation rateα Technology parameter
θl , θ f a, θ Scale parameters in labour, financial assets and productionPolicy parameters
νc,i, νw,i Steady state tax rates on consumption and wagesνk,i, νd,i, ν f d,i Steady state tax rates on capital returns, dividends and financial dividendsφi, φR,i, φNR,i Steady state lump-sum transfers to GDP ratiosΠ∗, ρ, rπ , ry Monetary policy parameters
Endogenous parametersR Long-term interest rate
cyi, gyi, iyi Share of private/public consumption and investment in GDP in country isi
c, siwl Share of non Ricardian households in consumption and payroll in country i
Table 4: Definition of the parameters
86
G 9
001
J. F
AY
OLL
E e
t M
. F
LEU
RB
AE
Y
Acc
umul
atio
n, p
rofit
abili
té e
t end
ette
men
t des
en
trep
rises
G 9
002
H. R
OU
SS
E
Dét
ectio
n et
effe
ts d
e la
mul
ticol
inéa
rité
dans
les
mod
èles
liné
aire
s or
dina
ires
- U
n pr
olon
gem
ent
de la
réf
lexi
on d
e B
ELS
LEY
, KU
H e
t WE
LSC
H
G 9
003
P. R
ALL
E e
t J. T
OU
JAS
-BE
RN
AT
E
Inde
xatio
n de
s sa
laire
s : l
a ru
ptur
e de
198
3
G 9
004
D. G
UE
LLE
C e
t P. R
ALL
E
Com
pétit
ivité
, cro
issa
nce
et in
nova
tion
de p
rodu
it
G 9
005
P. R
ALL
E e
t J. T
OU
JAS
-BE
RN
AT
E
Les
cons
éque
nces
de
la d
ésin
dexa
tion.
Ana
lyse
da
ns u
ne m
aque
tte p
rix-s
alai
res
G 9
101
Équ
ipe
AM
AD
EU
S
Le m
odèl
e A
MA
DE
US
- P
rem
ière
par
tie -
Pré
sent
atio
n gé
néra
le
G 9
102
J.L.
BR
ILLE
T
Le m
odèl
e A
MA
DE
US
- D
euxi
ème
part
ie -
Pro
prié
tés
varia
ntie
lles
G 9
103
D. G
UE
LLE
C e
t P. R
ALL
E
End
ogen
ous
grow
th a
nd p
rodu
ct in
nova
tion
G 9
104
H
. RO
US
SE
Le
mod
èle
AM
AD
EU
S -
Tro
isiè
me
part
ie -
Le
com
mer
ce e
xtér
ieur
et l
'env
ironn
emen
t in
tern
atio
nal
G 9
105
H. R
OU
SS
E
Effe
ts d
e de
man
de e
t d'o
ffre
dans
les
résu
ltats
du
com
mer
ce e
xtér
ieur
man
ufac
turé
de
la F
ranc
e au
co
urs
des
deux
der
nièr
es d
écen
nies
G 9
106
B. C
RE
PO
N
Inno
vatio
n, ta
ille
et c
once
ntra
tion
: cau
salit
és e
t dy
nam
ique
s
G 9
107
B. A
MA
BLE
et D
. GU
ELL
EC
U
n pa
nora
ma
des
théo
ries
de la
cro
issa
nce
endo
gène
G 9
108
M. G
LAU
DE
et M
. MO
UT
AR
DIE
R
Une
éva
luat
ion
du c
oût d
irect
de
l'enf
ant d
e 19
79
à 19
89
G 9
109
P. R
ALL
E e
t alii
F
ranc
e -
Alle
mag
ne :
perf
orm
ance
s éc
onom
ique
s co
mpa
rées
G 9
110
J.L.
BR
ILLE
T
Mic
ro-D
MS
N
ON
PA
RU
G 9
111
A. M
AG
NIE
R
Effe
ts a
ccél
érat
eur
et m
ultip
licat
eur
en F
ranc
e de
puis
197
0 : q
uelq
ues
résu
ltats
em
piriq
ues
G 9
112
B. C
RE
PO
N e
t G. D
UR
EA
U
Inve
stis
sem
ent e
n re
cher
che-
déve
lopp
emen
t :
anal
yse
de c
ausa
lités
dan
s un
mod
èle
d'ac
célé
-ra
teur
gén
éral
isé
G 9
113
J.L.
BR
ILLE
T,
H.
ER
KE
L-R
OU
SS
E,
J. T
OU
JAS
-B
ER
NA
TE
"F
ranc
e-A
llem
agne
Cou
plée
s" -
Deu
x éc
onom
ies
vues
par
une
maq
uette
mac
ro-é
cono
mét
rique
G 9
201
W.J
. A
DA
MS
, B
. C
RE
PO
N,
D.
EN
CA
OU
A
Cho
ix te
chno
logi
ques
et s
trat
égie
s de
dis
suas
ion
d'en
trée
G 9
202
J. O
LIV
EIR
A-M
AR
TIN
S,
J.
TO
UJA
S-B
ER
NA
TE
Mac
ro-e
cono
mic
impo
rt fu
nctio
ns w
ith im
perf
ect
com
petit
ion
- A
n ap
plic
atio
n to
the
E.C
. Tra
de
G 9
203
I. S
TA
PIC
Le
s éc
hang
es in
tern
atio
naux
de
serv
ices
de
la
Fra
nce
dans
le c
adre
des
nég
ocia
tions
mul
tila-
téra
les
du G
AT
T
Jui
n 19
92 (
1ère
ver
sion
) N
ovem
bre
1992
(ve
rsio
n fin
ale)
G 9
204
P. S
EV
ES
TR
E
L'éc
onom
étrie
sur
don
nées
indi
vidu
elle
s-te
mpo
relle
s. U
ne n
ote
intr
oduc
tive
G 9
205
H. E
RK
EL-
RO
US
SE
Le
com
mer
ce e
xtér
ieur
et l
'env
ironn
emen
t in-
tern
atio
nal d
ans
le m
odèl
e A
MA
DE
US
(r
éest
imat
ion
1992
)
G 9
206
N. G
RE
EN
AN
et D
. GU
ELL
EC
C
oord
inat
ion
with
in th
e fir
m a
nd e
ndog
enou
s gr
owth
G 9
207
A. M
AG
NIE
R e
t J. T
OU
JAS
-BE
RN
AT
E
Tec
hnol
ogy
and
trad
e: e
mpi
rical
evi
denc
es fo
r th
e m
ajor
five
indu
stria
lized
cou
ntrie
s
G 9
208
B.
CR
EP
ON
, E
. D
UG
UE
T,
D. E
NC
AO
UA
et
P. M
OH
NE
N
Coo
pera
tive,
non
coo
pera
tive
R &
D a
nd o
ptim
al
pate
nt li
fe
G 9
209
B. C
RE
PO
N e
t E. D
UG
UE
T
Res
earc
h an
d de
velo
pmen
t, co
mpe
titio
n an
d in
nova
tion:
an
appl
icat
ion
of p
seud
o m
axim
um
likel
ihoo
d m
etho
ds to
Poi
sson
mod
els
with
he
tero
gene
ity
G 9
301
J. T
OU
JAS
-BE
RN
AT
E
Com
mer
ce in
tern
atio
nal e
t con
curr
ence
impa
r-fa
ite :
déve
lopp
emen
ts r
écen
ts e
t im
plic
atio
ns
pour
la p
oliti
que
com
mer
cial
e
G 9
302
Ch.
CA
SE
S
Dur
ées
de c
hôm
age
et c
ompo
rtem
ents
d'o
ffre
de
trav
ail :
une
rev
ue d
e la
litté
ratu
re
G 9
303
H. E
RK
EL-
RO
US
SE
U
nion
éco
nom
ique
et m
onét
aire
: le
déb
at
écon
omiq
ue
G 9
304
N. G
RE
EN
AN
- D
. GU
ELL
EC
/ G
. B
RO
US
SA
UD
IER
- L
. M
IOT
TI
Inno
vatio
n or
gani
satio
nnel
le, d
ynam
ism
e te
ch-
nolo
giqu
e et
per
form
ance
s de
s en
trep
rises
G 9
305
P. J
AIL
LAR
D
Le tr
aité
de
Maa
stric
ht :
prés
enta
tion
jurid
ique
et
hist
oriq
ue
G 9
306
J.L.
BR
ILLE
T
Mic
ro-D
MS
: pr
ésen
tatio
n et
pro
prié
tés
G 9
307
J.L.
BR
ILLE
T
Mic
ro-D
MS
- v
aria
ntes
: le
s ta
blea
ux
G 9
308
S. J
AC
OB
ZO
NE
Le
s gr
ands
rés
eaux
pub
lics
fran
çais
dan
s un
e pe
rspe
ctiv
e eu
ropé
enne
G 9
309
L. B
LOC
H -
B. C
ŒU
RE
P
rofit
abili
té d
e l'i
nves
tisse
men
t pro
duct
if et
tr
ansm
issi
on d
es c
hocs
fina
ncie
rs
G 9
310
J. B
OU
RD
IEU
- B
. C
OLI
N-S
ED
ILLO
T
Les
théo
ries
sur
la s
truc
ture
opt
imal
e du
cap
ital :
qu
elqu
es p
oint
s de
rep
ère
List
e de
s do
cumen
ts d
e tr
avail de
la
Direc
tion
des
Étu
des
et S
ynth
èses
Éco
nomique
s ii
G 9
311
J. B
OU
RD
IEU
- B
. C
OLI
N-S
ED
ILLO
T
Les
déci
sion
s de
fina
ncem
ent d
es e
ntre
pris
es
fran
çais
es :
une
éval
uatio
n em
piriq
ue d
es th
éo-
ries
de la
str
uctu
re o
ptim
ale
du c
apita
l
G 9
312
L. B
LOC
H -
B. C
ŒU
RÉ
Q
de
Tob
in m
argi
nal e
t tra
nsm
issi
on d
es c
hocs
fin
anci
ers
G 9
313
Équ
ipes
Am
adeu
s (I
NS
EE
), B
anqu
e de
Fra
nce,
M
étric
(D
P)
Pré
sent
atio
n de
s pr
oprié
tés
des
prin
cipa
ux m
o-dè
les
mac
roéc
onom
ique
s du
Ser
vice
Pub
lic
G 9
314
B. C
RE
PO
N -
E. D
UG
UE
T
Res
earc
h &
Dev
elop
men
t, co
mpe
titio
n an
d in
nova
tion
G 9
315
B. D
OR
MO
NT
Q
uelle
est
l'in
fluen
ce d
u co
ût d
u tr
avai
l sur
l'e
mpl
oi ?
G 9
316
D. B
LAN
CH
ET
- C
. BR
OU
SS
E
Deu
x ét
udes
sur
l'âg
e de
la r
etra
ite
G 9
317
D. B
LAN
CH
ET
R
épar
titio
n du
trav
ail d
ans
une
popu
latio
n hé
té-
rogè
ne :
deux
not
es
G 9
318
D. E
YS
SA
RT
IER
- N
. PO
NT
Y
AM
AD
EU
S -
an
annu
al m
acro
-eco
nom
ic m
odel
fo
r th
e m
ediu
m a
nd lo
ng te
rm
G 9
319
G.
CE
TT
E -
Ph.
CU
NÉ
O -
D.
EY
SS
AR
TIE
R -
J. G
AU
TIÉ
Le
s ef
fets
sur
l'em
ploi
d'u
n ab
aiss
emen
t du
coût
du
trav
ail d
es je
unes
G 9
401
D. B
LAN
CH
ET
Le
s st
ruct
ures
par
âge
impo
rten
t-el
les
?
G 9
402
J. G
AU
TIÉ
Le
chô
mag
e de
s je
unes
en
Fra
nce
: pro
blèm
e de
fo
rmat
ion
ou p
héno
mèn
e de
file
d'a
ttent
e ?
Que
lque
s él
émen
ts d
u dé
bat
G 9
403
P. Q
UIR
ION
Le
s dé
chet
s en
Fra
nce
: élé
men
ts s
tatis
tique
s et
éc
onom
ique
s
G 9
404
D. L
AD
IRA
Y -
M. G
RU
N-R
EH
OM
ME
Li
ssag
e pa
r m
oyen
nes
mob
iles
- Le
pro
blèm
e de
s ex
trém
ités
de s
érie
G 9
405
V. M
AIL
LAR
D
Thé
orie
et p
ratiq
ue d
e la
cor
rect
ion
des
effe
ts d
e jo
urs
ouvr
able
s
G 9
406
F.
RO
SE
NW
ALD
La
déc
isio
n d'
inve
stir
G 9
407
S. J
AC
OB
ZO
NE
Le
s ap
port
s de
l'éc
onom
ie in
dust
rielle
pou
r dé
finir
la s
trat
égie
éco
nom
ique
de
l'hôp
ital p
ublic
G 9
408
L. B
LOC
H, J
. B
OU
RD
IEU
,
B.
CO
LIN
-SE
DIL
LOT
, G
. LO
NG
UE
VIL
LE
Du
défa
ut d
e pa
iem
ent a
u dé
pôt d
e bi
lan
: les
ba
nqui
ers
face
aux
PM
E e
n di
fficu
lté
G 9
409
D. E
YS
SA
RT
IER
, P.
MA
IRE
Im
pact
s m
acro
-éco
nom
ique
s de
mes
ures
d'a
ide
au lo
gem
ent -
que
lque
s él
émen
ts d
'éva
luat
ion
G 9
410
F.
RO
SE
NW
ALD
S
uivi
con
jonc
ture
l de
l'inv
estis
sem
ent
G 9
411
C. D
EF
EU
ILLE
Y -
Ph.
QU
IRIO
N
Les
déch
ets
d'em
balla
ges
mén
ager
s : u
ne
anal
yse
écon
omiq
ue d
es p
oliti
ques
fran
çais
e et
al
lem
ande
G 9
412
J. B
OU
RD
IEU
- B
. C
ŒU
RÉ
-
B.
CO
LIN
-SE
DIL
LOT
In
vest
isse
men
t, in
cert
itude
et i
rrév
ersi
bilit
é Q
uelq
ues
déve
lopp
emen
ts r
écen
ts d
e la
théo
rie
de l'
inve
stis
sem
ent
G 9
413
B. D
OR
MO
NT
- M
. PA
UC
HE
T
L'év
alua
tion
de l'
élas
ticité
em
ploi
-sal
aire
dép
end-
elle
des
str
uctu
res
de q
ualif
icat
ion
?
G 9
414
I. K
AB
LA
Le C
hoix
de
brev
eter
une
inve
ntio
n
G 9
501
J. B
OU
RD
IEU
- B
. C
ŒU
RÉ
- B
. S
ED
ILLO
T
Irre
vers
ible
Inve
stm
ent a
nd U
ncer
tain
ty:
W
hen
is th
ere
a V
alue
of W
aitin
g?
G 9
502
L. B
LOC
H -
B. C
ŒU
RÉ
Im
perf
ectio
ns d
u m
arch
é du
cré
dit,
inve
stis
se-
men
t des
ent
repr
ises
et c
ycle
éco
nom
ique
G 9
503
D. G
OU
X -
E. M
AU
RIN
Le
s tr
ansf
orm
atio
ns d
e la
dem
ande
de
trav
ail p
ar
qual
ifica
tion
en F
ranc
e
Une
étu
de s
ur la
pér
iode
197
0-19
93
G 9
504
N. G
RE
EN
AN
T
echn
olog
ie, c
hang
emen
t org
anis
atio
nnel
, qua
-lif
icat
ions
et e
mpl
oi :
une
étud
e em
piriq
ue s
ur
l'ind
ustr
ie m
anuf
actu
rière
G 9
505
D. G
OU
X -
E. M
AU
RIN
P
ersi
stan
ce d
es h
iéra
rchi
es s
ecto
rielle
s de
sa-
laire
s: u
n ré
exam
en s
ur d
onné
es fr
ança
ises
G 9
505
D. G
OU
X -
E. M
AU
RIN
B
is
Per
sist
ence
of i
nter
-indu
stry
wag
es d
iffer
entia
ls: a
re
exam
inat
ion
on m
atch
ed w
orke
r-fir
m p
anel
dat
a
G 9
506
S. J
AC
OB
ZO
NE
Le
s lie
ns e
ntre
RM
I et c
hôm
age,
une
mis
e en
pe
rspe
ctiv
e N
ON
PA
RU
- ar
ticle
sor
ti da
ns É
cono
mie
et
Pré
visi
on n
° 12
2 (1
996)
- pa
ges
95 à
113
G 9
507
G.
CE
TT
E -
S.
MA
HF
OU
Z
Le p
arta
ge p
rimai
re d
u re
venu
C
onst
at d
escr
iptif
sur
long
ue p
ério
de
G 9
601
Ban
que
de F
ranc
e -
CE
PR
EM
AP
- D
irect
ion
de la
P
révi
sion
- É
rasm
e -
INS
EE
- O
FC
E
Str
uctu
res
et p
ropr
iété
s de
cin
q m
odèl
es m
acro
-éc
onom
ique
s fr
ança
is
G 9
602
Rap
port
d’a
ctiv
ité d
e la
DE
SE
de
l’ann
ée 1
995
G 9
603
J. B
OU
RD
IEU
- A
. DR
AZ
NIE
KS
L’
octr
oi d
e cr
édit
aux
PM
E :
une
anal
yse
à pa
rtir
d’in
form
atio
ns b
anca
ires
G 9
604
A. T
OP
IOL-
BE
NS
AÏD
Le
s im
plan
tatio
ns ja
pona
ises
en
Fra
nce
G 9
605
P.
GE
NIE
R -
S.
JAC
OB
ZO
NE
C
ompo
rtem
ents
de
prév
entio
n, c
onso
mm
atio
n d’
alco
ol e
t tab
agie
: pe
ut-o
n pa
rler
d’un
e ge
stio
n gl
obal
e du
cap
ital s
anté
?
Une
mod
élis
atio
n m
icro
écon
omét
rique
em
piriq
ue
G 9
606
C. D
OZ
- F
. LE
NG
LAR
T
Fac
tor
anal
ysis
and
uno
bser
ved
com
pone
nt
mod
els:
an
appl
icat
ion
to th
e st
udy
of F
renc
h bu
sine
ss s
urve
ys
G 9
607
N. G
RE
EN
AN
- D
. GU
ELL
EC
La
théo
rie c
oopé
rativ
e de
la fi
rme
iii
G 9
608
N. G
RE
EN
AN
- D
. GU
ELL
EC
T
echn
olog
ical
inno
vatio
n an
d em
ploy
men
t re
allo
catio
n
G 9
609
Ph.
CO
UR
- F
. R
UP
PR
EC
HT
L’
inté
grat
ion
asym
étriq
ue a
u se
in d
u co
ntin
ent
amér
icai
n : u
n es
sai d
e m
odél
isat
ion
G 9
610
S.
DU
CH
EN
E -
G.
FO
RG
EO
T -
A.
JAC
QU
OT
A
naly
se d
es é
volu
tions
réc
ente
s de
la p
rodu
cti-
vité
app
aren
te d
u tr
avai
l
G 9
611
X. B
ON
NE
T -
S. M
AH
FO
UZ
T
he in
fluen
ce o
f diff
eren
t spe
cific
atio
ns o
f wag
es-
pric
es s
pira
ls o
n th
e m
easu
re o
f the
NA
IRU
: the
ca
se o
f Fra
nce
G 9
612
PH
. C
OU
R -
E.
DU
BO
IS,
S.
MA
HF
OU
Z,
J.
PIS
AN
I-F
ER
RY
T
he c
ost o
f fis
cal r
etre
nchm
ent r
evis
ited:
how
st
rong
is th
e ev
iden
ce?
G 9
613
A. J
AC
QU
OT
Le
s fle
xion
s de
s ta
ux d
’act
ivité
son
t-el
les
seul
e-m
ent c
onjo
nctu
relle
s ?
G 9
614
ZH
AN
G Y
ingx
iang
- S
ON
G X
ueqi
ng
Lexi
que
mac
roéc
onom
ique
Fra
nçai
s-C
hino
is
G 9
701
J.L.
SC
HN
EID
ER
La
taxe
pro
fess
ionn
elle
: él
émen
ts d
e ca
drag
e éc
onom
ique
G 9
702
J.L.
SC
HN
EID
ER
T
rans
ition
et s
tabi
lité
polit
ique
d’u
n sy
stèm
e re
dist
ribut
if
G 9
703
D. G
OU
X -
E. M
AU
RIN
T
rain
or
Pay
: D
oes
it R
educ
e In
equa
litie
s to
En-
cour
age
Firm
s to
Tra
in th
eir
Wor
kers
?
G 9
704
P. G
EN
IER
D
eux
cont
ribut
ions
sur
dép
enda
nce
et é
quité
G 9
705
E. D
UG
UE
T -
N. I
UN
G
R &
D In
vest
men
t, P
aten
t Life
and
Pat
ent V
alue
A
n E
cono
met
ric A
naly
sis
at th
e F
irm L
evel
G 9
706
M.
HO
UD
EB
INE
- A
. T
OP
IOL-
BE
NS
AÏD
Le
s en
trep
rises
inte
rnat
iona
les
en F
ranc
e : u
ne
anal
yse
à pa
rtir
de d
onné
es in
divi
duel
les
G 9
707
M. H
OU
DE
BIN
E
Pol
aris
atio
n de
s ac
tivité
s et
spé
cial
isat
ion
des
dépa
rtem
ents
en
Fra
nce
G 9
708
E. D
UG
UE
T -
N. G
RE
EN
AN
Le
bia
is te
chno
logi
que
: une
ana
lyse
sur
don
nées
in
divi
duel
les
G 9
709
J.L.
BR
ILLE
T
Ana
lyzi
ng a
sm
all F
renc
h E
CM
Mod
el
G 9
710
J.L.
BR
ILLE
T
For
mal
izin
g th
e tr
ansi
tion
proc
ess:
sce
nario
s fo
r ca
pita
l acc
umul
atio
n
G 9
711
G.
FO
RG
EO
T -
J.
GA
UT
IÉ
Inse
rtio
n pr
ofes
sion
nelle
des
jeun
es e
t pro
cess
us
de d
écla
ssem
ent
G 9
712
E. D
UB
OIS
H
igh
Rea
l Int
eres
t Rat
es: t
he C
onse
quen
ce o
f a
Sav
ing
Inve
stm
ent D
iseq
uilib
rium
or
of a
n in
-su
ffici
ent C
redi
bilit
y of
Mon
etar
y A
utho
ritie
s?
G 9
713
Bila
n de
s ac
tivité
s de
la D
irect
ion
des
Étu
des
et S
ynth
èses
Éco
nom
ique
s -
1996
G 9
714
F.
LEQ
UIL
LER
D
oes
the
Fre
nch
Con
sum
er P
rice
Inde
x O
ver-
stat
e In
flatio
n?
G 9
715
X. B
ON
NE
T
Peu
t-on
met
tre
en é
vide
nce
les
rigid
ités
à la
ba
isse
des
sal
aire
s no
min
aux
?
Une
étu
de s
ur q
uelq
ues
gran
ds p
ays
de l’
OC
DE
G 9
716
N. I
UN
G -
F.
RU
PP
RE
CH
T
Pro
duct
ivité
de
la r
eche
rche
et r
ende
men
ts
d’éc
helle
dan
s le
sec
teur
pha
rmac
eutiq
ue
fran
çais
G 9
717
E. D
UG
UE
T -
I. K
AB
LA
App
ropr
iatio
n st
rate
gy a
nd th
e m
otiv
atio
ns to
use
th
e pa
tent
sys
tem
in F
ranc
e -
An
econ
omet
ric
anal
ysis
at t
he fi
rm le
vel
G 9
718
L.P
. PE
LÉ -
P. R
ALL
E
Âge
de
la r
etra
ite :
les
aspe
cts
inci
tatif
s du
rég
ime
géné
ral
G 9
719
ZH
AN
G Y
ingx
iang
- S
ON
G X
ueqi
ng
Lexi
que
mac
roéc
onom
ique
fran
çais
-chi
nois
, ch
inoi
s-fr
ança
is
G 9
720
M. H
OU
DE
BIN
E -
J.L
. SC
HN
EID
ER
M
esur
er l’
influ
ence
de
la fi
scal
ité s
ur la
loca
li-sa
tion
des
entr
epris
es
G 9
721
A. M
OU
RO
UG
AN
E
Cré
dibi
lité,
indé
pend
ance
et
polit
ique
mon
étai
re
Une
rev
ue d
e la
litté
ratu
re
G 9
722
P. A
UG
ER
AU
D -
L. B
RIO
T
Les
donn
ées
com
ptab
les
d’en
trep
rises
Le
sys
tèm
e in
term
édia
ire d
’ent
repr
ises
P
assa
ge d
es d
onné
es in
divi
duel
les
aux
donn
ées
sect
orie
lles
G 9
723
P. A
UG
ER
AU
D -
J.E
. CH
AP
RO
N
Usi
ng B
usin
ess
Acc
ount
s fo
r C
ompi
ling
Nat
iona
l A
ccou
nts:
the
Fre
nch
Exp
erie
nce
G 9
724
P. A
UG
ER
AU
D
Les
com
ptes
d’e
ntre
pris
e pa
r ac
tivité
s -
Le p
as-
sage
aux
com
ptes
- D
e la
com
ptab
ilité
d’
entr
epris
e à
la c
ompt
abili
té n
atio
nale
- A
pa
raîtr
e
G 9
801
H. M
ICH
AU
DO
N -
C. P
RIG
EN
T
Pré
sent
atio
n du
mod
èle
AM
AD
EU
S
G 9
802
J. A
CC
AR
DO
U
ne é
tude
de
com
ptab
ilité
gén
érat
ionn
elle
po
ur la
Fra
nce
en 1
996
G 9
803
X. B
ON
NE
T -
S. D
UC
HÊ
NE
A
ppor
ts e
t lim
ites
de la
mod
élis
atio
n «
Rea
l Bus
ines
s C
ycle
s »
G 9
804
C. B
AR
LET
- C
. DU
GU
ET
-
D. E
NC
AO
UA
- J
. PR
AD
EL
The
Com
mer
cial
Suc
cess
of I
nnov
atio
ns
An
econ
omet
ric a
naly
sis
at th
e fir
m le
vel i
n F
renc
h m
anuf
actu
ring
G 9
805
P. C
AH
UC
- C
h. G
IAN
ELL
A -
D
. G
OU
X -
A.
ZIL
BE
RB
ER
G
Equ
aliz
ing
Wag
e D
iffer
ence
s an
d B
arga
inin
g P
ower
- E
vide
nce
form
a P
anel
of F
renc
h F
irms
G 9
806
J. A
CC
AR
DO
- M
. JL
AS
SI
La p
rodu
ctiv
ité g
loba
le d
es fa
cteu
rs e
ntre
197
5 et
19
96
iv
G 9
807
Bila
n de
s ac
tivité
s de
la D
irect
ion
des
Étu
des
et
Syn
thès
es É
cono
miq
ues
- 19
97
G 9
808
A. M
OU
RO
UG
AN
E
Can
a C
onse
rvat
ive
Gov
erno
r C
ondu
ct a
n A
c-co
mod
ativ
e M
onet
ary
Pol
icy?
G 9
809
X. B
ON
NE
T -
E. D
UB
OIS
- L
. FA
UV
ET
A
sym
étrie
des
infla
tions
rel
ativ
es e
t men
us c
osts
: t
ests
sur
l’in
flatio
n fr
ança
ise
G 9
810
E. D
UG
UE
T -
N. I
UN
G
Sal
es a
nd A
dver
tisin
g w
ith S
pillo
vers
at t
he fi
rm
leve
l: E
stim
atio
n of
a D
ynam
ic S
truc
tura
l Mod
el
on P
anel
Dat
a
G 9
811
J.P
. BE
RT
HIE
R
Con
gest
ion
urba
ine
: un
mod
èle
de tr
afic
de
poin
te à
cou
rbe
débi
t-vi
tess
e et
dem
ande
él
astiq
ue
G 9
812
C. P
RIG
EN
T
La p
art d
es s
alai
res
dans
la v
aleu
r aj
outé
e : u
ne
appr
oche
mac
roéc
onom
ique
G 9
813
A.T
h. A
ER
TS
L’
évol
utio
n de
la p
art d
es s
alai
res
dans
la v
aleu
r aj
outé
e en
Fra
nce
reflè
te-t
-elle
les
évol
utio
ns
indi
vidu
elle
s su
r la
pér
iode
197
9-19
94 ?
G 9
814
B. S
ALA
NIÉ
G
uide
pra
tique
des
sér
ies
non-
stat
ionn
aire
s
G 9
901
S. D
UC
HÊ
NE
- A
. JA
CQ
UO
T
Une
cro
issa
nce
plus
ric
he e
n em
ploi
s de
puis
le
débu
t de
la d
écen
nie
? U
ne a
naly
se e
n co
mpa
-ra
ison
inte
rnat
iona
le
G 9
902
Ch.
CO
LIN
M
odél
isat
ion
des
carr
ière
s da
ns D
estin
ie
G 9
903
Ch.
CO
LIN
É
volu
tion
de la
dis
pers
ion
des
sala
ires
: un
essa
i de
pro
spec
tive
par
mic
rosi
mul
atio
n
G 9
904
B. C
RE
PO
N -
N. I
UN
G
Inno
vatio
n, e
mpl
oi e
t per
form
ance
s
G 9
905
B. C
RE
PO
N -
Ch.
GIA
NE
LLA
W
ages
ineq
ualit
ies
in F
ranc
e 19
69-1
992
An
appl
icat
ion
of q
uant
ile r
egre
ssio
n te
chni
ques
G 9
906
C. B
ON
NE
T -
R. M
AH
IEU
M
icro
sim
ulat
ion
tech
niqu
es a
pplie
d to
inte
r-ge
nera
tiona
l tra
nsfe
rs -
Pen
sion
s in
a d
ynam
ic
fram
ewor
k: th
e ca
se o
f Fra
nce
G 9
907
F.
RO
SE
NW
ALD
L’
impa
ct d
es c
ontr
aint
es fi
nanc
ière
s da
ns la
dé-
cisi
on d
’inve
stis
sem
ent
G 9
908
Bila
n de
s ac
tivité
s de
la D
ES
E -
199
8
G 9
909
J.P
. ZO
YE
M
Con
trat
d’in
sert
ion
et s
ortie
du
RM
I É
valu
atio
n de
s ef
fets
d’u
ne p
oliti
que
soci
ale
G 9
910
C
h. C
OLI
N -
Fl.
LEG
RO
S -
R. M
AH
IEU
B
ilans
con
trib
utifs
com
paré
s de
s ré
gim
es d
e re
trai
te d
u se
cteu
r pr
ivé
et d
e la
fonc
tion
publ
ique
G 9
911
G. L
AR
OQ
UE
- B
. SA
LAN
IÉ
Une
déc
ompo
sitio
n du
non
-em
ploi
en
Fra
nce
G 9
912
B. S
ALA
NIÉ
U
ne m
aque
tte a
naly
tique
de
long
term
e du
m
arch
é du
trav
ail
G 9
912
Ch.
GIA
NE
LLA
B
is
Une
est
imat
ion
de l’
élas
ticité
de
l’em
ploi
peu
qu
alifi
é à
son
coût
G 9
913
Div
isio
n «
Red
istr
ibut
ion
et P
oliti
ques
Soc
iale
s »
Le m
odèl
e de
mic
rosi
mul
atio
n dy
nam
ique
D
ES
TIN
IE
G 9
914
E. D
UG
UE
T
Mac
ro-c
omm
ande
s S
AS
pou
r l’é
cono
mét
rie d
es
pane
ls e
t des
var
iabl
es q
ualit
ativ
es
G 9
915
R. D
UH
AU
TO
IS
Évo
lutio
n de
s flu
x d’
empl
ois
en F
ranc
e en
tre
1990
et 1
996
: une
étu
de e
mpi
rique
à p
artir
du
fichi
er d
es b
énéf
ices
rée
ls n
orm
aux
(BR
N)
G 9
916
J.Y
. FO
UR
NIE
R
Ext
ract
ion
du c
ycle
des
affa
ires
: la
mét
hode
de
Bax
ter
et K
ing
G 9
917
B.
CR
ÉP
ON
- R
. DE
SP
LAT
Z -
J. M
AIR
ES
SE
E
stim
atin
g pr
ice
cost
mar
gins
, sca
le e
cono
mie
s an
d w
orke
rs’ b
arga
inin
g po
wer
at t
he fi
rm le
vel
G 9
918
Ch.
GIA
NE
LLA
- P
h. L
AG
AR
DE
P
rodu
ctiv
ity o
f hou
rs in
the
aggr
egat
e pr
oduc
tion
func
tion:
an
eval
uatio
n on
a p
anel
of F
renc
h fir
ms
from
the
man
ufac
turin
g se
ctor
G 9
919
S.
AU
DR
IC -
P. G
IVO
RD
- C
. PR
OS
T
Évo
lutio
n de
l’em
ploi
et d
es c
oûts
par
qua
li-fic
atio
n en
tre
1982
et 1
996
G 2
000/
01
R. M
AH
IEU
Le
s dé
term
inan
ts d
es d
épen
ses
de s
anté
: un
e ap
proc
he m
acro
écon
omiq
ue
G 2
000/
02
C. A
LLA
RD
-PR
IGE
NT
- H
. GU
ILM
EA
U -
A
. Q
UIN
ET
T
he r
eal e
xcha
nge
rate
as
the
rela
tive
pric
e of
no
ntra
bles
in te
rms
of tr
adab
les:
theo
retic
al
inve
stig
atio
n an
d em
piric
al s
tudy
on
Fre
nch
data
G 2
000/
03
J.-Y
. FO
UR
NIE
R
L’ap
prox
imat
ion
du fi
ltre
pass
e-ba
nde
prop
osée
pa
r C
hris
tiano
et F
itzge
rald
G 2
000/
04
Bila
n de
s ac
tivité
s de
la D
ES
E -
199
9
G 2
000/
05
B. C
RE
PO
N -
F.
RO
SE
NW
ALD
In
vest
isse
men
t et c
ontr
aint
es d
e fin
ance
men
t : le
po
ids
du c
ycle
U
ne e
stim
atio
n su
r do
nnée
s fr
ança
ises
G 2
000/
06
A. F
LIP
O
Les
com
port
emen
ts m
atrim
onia
ux d
e fa
it
G 2
000/
07
R. M
AH
IEU
- B
. SÉ
DIL
LOT
M
icro
sim
ulat
ions
of t
he r
etire
men
t dec
isio
n: a
su
pply
sid
e ap
proa
ch
G 2
000/
08
C. A
UD
EN
IS -
C. P
RO
ST
D
éfic
it co
njon
ctur
el :
une
pris
e en
com
pte
des
conj
onct
ures
pas
sées
G 2
000/
09
R. M
AH
IEU
- B
. SÉ
DIL
LOT
É
quiv
alen
t pat
rimon
ial d
e la
ren
te e
t sou
scrip
tion
de r
etra
ite c
ompl
émen
taire
G 2
000/
10
R. D
UH
AU
TO
IS
Ral
entis
sem
ent d
e l’i
nves
tisse
men
t : p
etite
s ou
gr
ande
s en
trep
rises
? in
dust
rie o
u te
rtia
ire ?
G 2
000/
11
G. L
AR
OQ
UE
- B
. SA
LAN
IÉ
Tem
ps p
artie
l fém
inin
et i
ncita
tions
fina
nciè
res
à l’e
mpl
oi
G20
00/1
2 C
h. G
IAN
ELL
A
Loca
l une
mpl
oym
ent a
nd w
ages
v
G20
00/1
3 B
. CR
EP
ON
- T
h. H
EC
KE
L -
Info
rmat
isat
ion
en F
ranc
e : u
ne é
valu
atio
n à
part
ir de
don
nées
indi
vidu
elle
s -
Com
pute
rizat
ion
in F
ranc
e: a
n ev
alua
tion
base
d on
indi
vidu
al c
ompa
ny d
ata
G20
01/0
1 F
. LE
QU
ILLE
R
- La
nou
velle
éco
nom
ie e
t la
mes
ure
d
e la
cro
issa
nce
du P
IB
- T
he n
ew e
cono
my
and
the
mea
sure
men
t of
GD
P g
row
th
G20
01/0
2 S
. AU
DR
IC
La r
epris
e de
la
croi
ssan
ce d
e l’e
mpl
oi p
rofit
e-t-
elle
aus
si a
ux n
on-d
iplô
més
?
G20
01/0
3 I.
BR
AU
N-L
EM
AIR
E
Évo
lutio
n et
rép
artit
ion
du s
urpl
us d
e pr
oduc
tivité
G20
01/0
4 A
. BE
AU
DU
- T
h. H
EC
KE
L Le
can
al d
u cr
édit
fonc
tionn
e-t-
il en
Eur
ope
? U
ne
étud
e de
l’h
étér
ogén
éité
de
s co
mpo
rtem
ents
d’
inve
stis
sem
ent
à pa
rtir
de
donn
ées
de
bila
n ag
régé
es
G20
01/0
5 C
. AU
DE
NIS
- P
. BIS
CO
UR
P -
N
. F
OU
RC
AD
E -
O.
LOIS
EL
Tes
ting
the
augm
ente
d S
olow
gro
wth
mod
el:
An
empi
rical
rea
sses
smen
t usi
ng p
anel
dat
a
G20
01/0
6 R
. MA
HIE
U -
B. S
ÉD
ILLO
T
Dép
art à
la r
etra
ite, i
rrév
ersi
bilit
é et
ince
rtitu
de
G20
01/0
7 B
ilan
des
activ
ités
de la
DE
SE
-
2000
G20
01/0
8 J.
Ph.
GA
UD
EM
ET
Le
s di
spos
itifs
d’
acqu
isiti
on
à tit
re
facu
ltatif
d’
annu
ités
viag
ères
de
retr
aite
G20
01/0
9 B
. CR
ÉP
ON
- C
h. G
IAN
ELL
A
Fis
calit
é, c
oût
d’us
age
du c
apita
l et
dem
ande
de
fact
eurs
: un
e an
alys
e su
r do
nnée
s in
divi
duel
les
G20
01/1
0 B
. CR
ÉP
ON
- R
. DE
SP
LAT
Z
Éva
luat
ion
des
effe
ts
des
disp
ositi
fs
d’al
lége
men
ts
de c
harg
es s
ocia
les
sur
les
bas
sala
ires
G20
01/1
1 J.
-Y. F
OU
RN
IER
C
ompa
rais
on d
es s
alai
res
des
sect
eurs
pub
lic e
t pr
ivé
G20
01/1
2 J.
-P. B
ER
TH
IER
- C
. JA
ULE
NT
R
. C
ON
VE
NE
VO
LE -
S.
PIS
AN
I U
ne
mét
hodo
logi
e de
co
mpa
rais
on
entr
e co
nsom
mat
ions
int
erm
édia
ires
de s
ourc
e fis
cale
et
de
com
ptab
ilité
nat
iona
le
G20
01/1
3 P
. BIS
CO
UR
P -
Ch.
GIA
NE
LLA
S
ubst
itutio
n an
d co
mpl
emen
tarit
y be
twee
n ca
pita
l, sk
illed
an
d le
ss
skill
ed
wor
kers
: an
an
alys
is
at
the
firm
le
vel
in
the
Fre
nch
man
ufac
turin
g in
dust
ry
G20
01/1
4 I.
RO
BE
RT
-BO
BE
E
Mod
ellin
g de
mog
raph
ic b
ehav
iour
s in
the
Fre
nch
mic
rosi
mul
atio
n m
odel
D
estin
ie:
An
anal
ysis
of
fu
ture
cha
nge
in c
ompl
eted
fert
ility
G20
01/1
5 J.
-P. Z
OY
EM
D
iagn
ostic
su
r la
pa
uvre
té
et
cale
ndrie
r de
re
venu
s :
le
cas
du
“Pan
el
euro
péen
de
s m
énag
es »
G20
01/1
6 J.
-Y.
FO
UR
NIE
R -
P.
GIV
OR
D
La
rédu
ctio
n de
s ta
ux
d’ac
tivité
au
x âg
es
extr
êmes
, une
spé
cific
ité fr
ança
ise
?
G20
01/1
7 C
. AU
DE
NIS
- P
. BIS
CO
UR
P -
N. R
IED
ING
ER
E
xist
e-t-
il un
e as
ymét
rie d
ans
la t
rans
mis
sion
du
prix
du
brut
aux
prix
des
car
bura
nts
?
G20
02/0
1 F
. M
AG
NIE
N -
J.-
L. T
AV
ER
NIE
R -
D. T
HE
SM
AR
Le
s st
atis
tique
s in
tern
atio
nale
s de
P
IB
par
habi
tant
en
st
anda
rd
de
pouv
oir
d’ac
hat
: un
e an
alys
e de
s ré
sulta
ts
G20
02/0
2 B
ilan
des
activ
ités
de la
DE
SE
- 2
001
G20
02/0
3 B
. SÉ
DIL
LOT
- E
. WA
LRA
ET
La
ces
satio
n d’
activ
ité a
u se
in d
es c
oupl
es :
y a-
t-il
inte
rdép
enda
nce
des
choi
x ?
G20
02/0
4 G
. BR
ILH
AU
LT
- R
étro
pola
tion
des
série
s de
FB
CF
et
calc
ul d
u ca
pita
l fix
e en
S
EC
-95
dans
le
s co
mpt
es
natio
naux
fran
çais
-
Ret
ropo
latio
n of
the
inv
estm
ent
serie
s (G
FC
F)
and
estim
atio
n of
fix
ed
capi
tal
stoc
ks
on
the
ES
A-9
5 ba
sis
for
the
Fre
nch
bala
nce
shee
ts
G20
02/0
5 P
. B
ISC
OU
RP
- B
. C
RÉ
PO
N -
T.
HE
CK
EL
- N
. R
IED
ING
ER
H
ow
do
firm
s re
spon
d to
ch
eape
r co
mpu
ters
? M
icro
econ
omet
ric e
vide
nce
for
Fra
nce
base
d on
a
prod
uctio
n fu
nctio
n ap
proa
ch
G20
02/0
6 C
. AU
DE
NIS
- J
. DE
RO
YO
N -
N. F
OU
RC
AD
E
L’im
pact
de
s no
uvel
les
tech
nolo
gies
de
l’i
nfor
mat
ion
et
de
la
com
mun
icat
ion
sur
l’éco
nom
ie
fran
çais
e -
un
bouc
lage
m
acro
-éc
onom
ique
G20
02/0
7 J.
BA
RD
AJI
- B
. SÉ
DIL
LOT
- E
. WA
LRA
ET
É
valu
atio
n de
tro
is r
éfor
mes
du
Rég
ime
Gén
éral
d’
assu
ranc
e vi
eille
sse
à l’a
ide
du
mod
èle
de
mic
rosi
mul
atio
n D
ES
TIN
IE
G20
02/0
8 J.
-P. B
ER
TH
IER
R
éfle
xion
s su
r le
s di
ffére
ntes
not
ions
de
volu
me
dans
les
com
ptes
nat
iona
ux :
com
ptes
aux
prix
d’
une
anné
e fix
e ou
au
x pr
ix
de
l’ann
ée
préc
éden
te, s
érie
s ch
aîné
es
G20
02/0
9 F
. H
ILD
Le
s so
ldes
d’o
pini
on r
ésum
ent-
ils a
u m
ieux
les
ré
pons
es
des
entr
epris
es
aux
enqu
êtes
de
co
njon
ctur
e ?
G20
02/1
0 I.
RO
BE
RT
-BO
BÉ
E
Les
com
port
emen
ts
dém
ogra
phiq
ues
dans
le
m
odèl
e de
m
icro
sim
ulat
ion
Des
tinie
-
Une
co
mpa
rais
on
des
estim
atio
ns
issu
es
des
enqu
êtes
Je
unes
et
C
arriè
res
1997
et
H
isto
ire
Fam
ilial
e 19
99
G20
02/1
1 J.
-P. Z
OY
EM
La
dyn
amiq
ue d
es b
as r
even
us :
une
ana
lyse
des
en
trée
s-so
rtie
s de
pau
vret
é
G20
02/1
2 F
. H
ILD
P
révi
sion
s d’
infla
tion
pour
la F
ranc
e
G20
02/1
3 M
. LE
CLA
IR
Réd
uctio
n du
tem
ps d
e tr
avai
l et
ten
sion
s su
r le
s fa
cteu
rs d
e pr
oduc
tion
G20
02/1
4 E
. WA
LRA
ET
- A
. VIN
CE
NT
-
Ana
lyse
de
la r
edis
trib
utio
n in
trag
énér
atio
nnel
le
dans
le s
ystè
me
de r
etra
ite d
es s
alar
iés
du p
rivé
- U
ne a
ppro
che
par
mic
rosi
mul
atio
n -
Intr
agen
erat
iona
l di
strib
utio
nal
anal
ysis
in
th
e fr
ench
pr
ivat
e se
ctor
pe
nsio
n sc
hem
e -
A
mic
rosi
mul
atio
n ap
proa
ch
vi
G20
02/1
5 P
. CH
ON
E -
D. L
E B
LAN
C -
I. R
OB
ER
T-B
OB
EE
O
ffre
de
trav
ail
fém
inin
e et
ga
rde
des
jeun
es
enfa
nts
G20
02/1
6 F
. MA
UR
EL
- S
. GR
EG
OIR
Le
s in
dice
s de
co
mpé
titiv
ité
des
pays
: in
ter-
prét
atio
n et
lim
ites
G20
03/0
1 N
. RIE
DIN
GE
R -
E.H
AU
VY
Le
coû
t de
dép
ollu
tion
atm
osph
ériq
ue p
our
les
entr
epris
es f
ranç
aise
s :
Une
est
imat
ion
à pa
rtir
de
donn
ées
indi
vidu
elle
s
G20
03/0
2 P
. BIS
CO
UR
P e
t F.
KR
AM
AR
Z
Cré
atio
n d’
empl
ois,
de
stru
ctio
n d’
empl
ois
et
inte
rnat
iona
lisat
ion
des
entr
epris
es
indu
strie
lles
fran
çais
es :
une
anal
yse
sur
la
pério
de
1986
-19
92
G20
03/0
3 B
ilan
des
activ
ités
de la
DE
SE
- 2
002
G20
03/0
4 P
.-O
. B
EF
FY
- J
. D
ER
OY
ON
-
N.
FO
UR
CA
DE
- S
. G
RE
GO
IR -
N.
LAÏB
-
B.
MO
NF
OR
T
Évo
lutio
ns d
émog
raph
ique
s et
cro
issa
nce
: un
e pr
ojec
tion
mac
ro-é
cono
miq
ue à
l’ho
rizon
202
0
G20
03/0
5 P
. AU
BE
RT
La
situ
atio
n de
s sa
larié
s de
plu
s de
cin
quan
te
ans
dans
le s
ecte
ur p
rivé
G20
03/0
6 P
. AU
BE
RT
- B
. CR
ÉP
ON
A
ge, s
alai
re e
t pro
duct
ivité
La
pro
duct
ivité
des
sal
arié
s dé
clin
e-t-
elle
en
fin
de c
arriè
re ?
G20
03/0
7 H
. B
AR
ON
- P
.O.
BE
FF
Y -
N.
FO
UR
CA
DE
- R
. M
AH
IEU
Le
ral
entis
sem
ent
de l
a pr
oduc
tivité
du
trav
ail
au
cour
s de
s an
nées
199
0
G20
03/0
8 P
.-O
. B
EF
FY
- B
. M
ON
FO
RT
P
atrim
oine
des
mén
ages
, dy
nam
ique
d’a
lloca
tion
et c
ompo
rtem
ent d
e co
nsom
mat
ion
G20
03/0
9 P
. BIS
CO
UR
P -
N. F
OU
RC
AD
E
Peu
t-on
m
ettr
e en
év
iden
ce
l’exi
sten
ce
de
rigid
ités
à la
ba
isse
de
s sa
laire
s à
part
ir de
do
nnée
s in
divi
duel
les
? Le
cas
de
la F
ranc
e à
la
fin d
es a
nnée
s 90
G20
03/1
0 M
. LE
CLA
IR -
P. P
ET
IT
Pré
senc
e sy
ndic
ale
dans
les
firm
es :
quel
im
pact
su
r le
s in
égal
ités
sala
riale
s en
tre
les
hom
mes
et
les
fem
mes
?
G20
03/1
1 P
.-O
. B
EF
FY
-
X.
BO
NN
ET
-
M.
DA
RR
AC
Q-
PA
RIE
S -
B.
MO
NF
OR
T
MZ
E: a
sm
all m
acro
-mod
el fo
r th
e eu
ro a
rea
G20
04/0
1 P
. AU
BE
RT
- M
. LE
CLA
IR
La
com
pétit
ivité
ex
prim
ée
dans
le
s en
quêt
es
trim
estr
ielle
s su
r la
situ
atio
n et
les
per
spec
tives
da
ns l’
indu
strie
G20
04/0
2 M
. DU
ÉE
- C
. RE
BIL
LAR
D
La
dépe
ndan
ce
des
pers
onne
s âg
ées
: un
e pr
ojec
tion
à lo
ng te
rme
G20
04/0
3 S
. RA
SP
ILLE
R -
N. R
IED
ING
ER
R
égul
atio
n en
viro
nnem
enta
le
et
choi
x de
lo
calis
atio
n de
s gr
oupe
s fr
ança
is
G20
04/0
4 A
. NA
BO
ULE
T -
S. R
AS
PIL
LER
Le
s dé
term
inan
ts d
e la
déc
isio
n d’
inve
stir
: un
e ap
proc
he
par
les
perc
eptio
ns
subj
ectiv
es
des
firm
es
G20
04/0
5 N
. RA
GA
CH
E
La d
écla
ratio
n de
s en
fant
s pa
r le
s co
uple
s no
n m
arié
s es
t-el
le fi
scal
emen
t opt
imal
e ?
G20
04/0
6 M
. DU
ÉE
L’
impa
ct d
u ch
ômag
e de
s pa
rent
s su
r le
dev
enir
scol
aire
des
enf
ants
G20
04/0
7 P
. AU
BE
RT
- E
. C
AR
OLI
- M
. RO
GE
R
New
Tec
hnol
ogie
s, W
orkp
lace
Org
anis
atio
n an
d th
e A
ge S
truc
ture
of
the
Wor
kfor
ce:
Firm
-Lev
el
Evi
denc
e
G20
04/0
8 E
. DU
GU
ET
- C
. LE
LAR
GE
Le
s br
evet
s ac
croi
ssen
t-ils
les
inc
itatio
ns p
rivée
s à
inno
ver
? U
n ex
amen
mic
roéc
onom
étriq
ue
G20
04/0
9 S
. RA
SP
ILLE
R -
P. S
ILLA
RD
A
ffilia
ting
vers
us S
ubco
ntra
ctin
g:
the
Cas
e of
Mul
tinat
iona
ls
G20
04/1
0 J.
BO
ISS
INO
T -
C. L
’AN
GE
VIN
- B
. MO
NF
OR
T
Pub
lic D
ebt
Sus
tain
abili
ty:
Som
e R
esul
ts o
n th
e F
renc
h C
ase
G20
04/1
1 S
. AN
AN
IAN
- P
. AU
BE
RT
T
rava
illeu
rs â
gés,
nou
velle
s te
chno
logi
es
et c
hang
emen
ts o
rgan
isat
ionn
els
: un
rée
xam
en
à pa
rtir
de l’
enqu
ête
« R
EP
ON
SE
»
G20
04/1
2 X
. BO
NN
ET
- H
. PO
NC
ET
S
truc
ture
s de
rev
enus
et
prop
ensi
ons
diffé
rent
es
à co
nsom
mer
-
Ver
s un
e éq
uatio
n de
co
nsom
mat
ion
des
mén
ages
pl
us
robu
ste
en
prév
isio
n po
ur la
Fra
nce
G20
04/1
3 C
. PIC
AR
T
Éva
luer
la r
enta
bilit
é de
s so
ciét
és n
on fi
nanc
ière
s
G20
04/1
4 J.
BA
RD
AJI
- B
. SÉ
DIL
LOT
- E
. WA
LRA
ET
Le
s re
trai
tes
du
sect
eur
publ
ic :
proj
ectio
ns
à l’h
oriz
on
2040
à
l’aid
e du
m
odèl
e de
m
icro
sim
ulat
ion
DE
ST
INIE
G20
05/0
1 S
. BU
FF
ET
EA
U -
P. G
OD
EF
RO
Y
Con
ditio
ns d
e dé
part
en
retr
aite
sel
on l’
âge
de f
in
d’ét
udes
: an
alys
e pr
ospe
ctiv
e po
ur
les
géné
ratio
ns 1
945
à197
4
G20
05/0
2 C
. AF
SA
- S
. BU
FF
ET
EA
U
L’év
olut
ion
de l’
activ
ité fé
min
ine
en F
ranc
e :
une
appr
oche
par
pse
udo-
pane
l
G20
05/0
3 P
. AU
BE
RT
- P
. SIL
LAR
D
Dél
ocal
isat
ions
et r
éduc
tions
d’e
ffect
ifs
dans
l’in
dust
rie fr
ança
ise
G20
05/0
4 M
. LE
CLA
IR -
S. R
OU
X
Mes
ure
et u
tilis
atio
n de
s em
ploi
s in
stab
les
da
ns le
s en
trep
rises
G20
05/0
5 C
. L’A
NG
EV
IN -
S. S
ER
RA
VA
LLE
P
erfo
rman
ces
à l’e
xpor
tatio
n de
la F
ranc
e
et d
e l’A
llem
agne
- U
ne a
naly
se p
ar s
ecte
ur e
t de
stin
atio
n gé
ogra
phiq
ue
G20
05/0
6 B
ilan
des
activ
ités
de l
a D
irect
ion
des
Étu
des
et
Syn
thès
es É
cono
miq
ues
- 20
04
G20
05/0
7 S
. RA
SP
ILLE
R
La
conc
urre
nce
fisca
le :
prin
cipa
ux
ense
igne
-m
ents
de
l’ana
lyse
éco
nom
ique
G20
05/0
8 C
. L’A
NG
EV
IN -
N. L
AÏB
É
duca
tion
et c
rois
sanc
e en
Fra
nce
et d
ans
un
pane
l de
21 p
ays
de l’
OC
DE
vii
G20
05/0
9 N
. FE
RR
AR
I P
révo
ir l’i
nves
tisse
men
t des
ent
repr
ises
U
n in
dica
teur
de
s ré
visi
ons
dans
l’e
nquê
te
de
conj
onct
ure
sur
les
inve
stis
sem
ents
da
ns
l’ind
ustr
ie.
G20
05/1
0 P
.-O
. BE
FF
Y -
C. L
’AN
GE
VIN
C
hôm
age
et b
oucl
e pr
ix-s
alai
res
: ap
port
d’u
n m
odèl
e «
qual
ifiés
/peu
qua
lifié
s »
G20
05/1
1 B
. HE
ITZ
A
tw
o-st
ates
Mar
kov-
switc
hing
mod
el o
f in
flatio
n in
F
ranc
e an
d th
e U
SA
: cr
edib
le
targ
et
VS
in
flatio
n sp
iral
G20
05/1
2 O
. BIA
U -
H. E
RK
EL-
RO
US
SE
- N
. FE
RR
AR
I R
épon
ses
indi
vidu
elle
s au
x en
quêt
es
de
conj
onct
ure
et p
révi
sion
mac
roéc
onom
ique
s :
Exe
mpl
e de
la
pr
évis
ion
de
la
prod
uctio
n m
anuf
actu
rière
G20
05/1
3 P
. AU
BE
RT
- D
. BLA
NC
HE
T -
D. B
LAU
The
lab
our
mar
ket
afte
r ag
e 50
: so
me
elem
ents
of
a F
ranc
o-A
mer
ican
com
paris
on
G20
05/1
4 D
. BLA
NC
HE
T -
T.
DE
BR
AN
D -
P
. D
OU
RG
NO
N -
P.
PO
LLE
T
L’en
quêt
e S
HA
RE
: pr
ésen
tatio
n et
pr
emie
rs
résu
ltats
de
l’édi
tion
fran
çais
e
G20
05/1
5 M
. DU
ÉE
La
m
odél
isat
ion
des
com
port
emen
ts
dém
ogra
-ph
ique
s da
ns
le
mod
èle
de
mic
rosi
mul
atio
n D
ES
TIN
IE
G20
05/1
6 H
. RA
OU
I - S
. RO
UX
É
tude
de
sim
ulat
ion
sur
la p
artic
ipat
ion
vers
ée
aux
sala
riés
par
les
entr
epris
es
G20
06/0
1 C
. B
ON
NE
T -
S.
BU
FF
ET
EA
U -
P.
GO
DE
FR
OY
D
ispa
rités
de
re
trai
te
de
droi
t di
rect
en
tre
hom
mes
et f
emm
es :
quel
les
évol
utio
ns ?
G20
06/0
2 C
. PIC
AR
T
Les
gaze
lles
en F
ranc
e
G20
06/0
3 P
. AU
BE
RT
- B
. CR
ÉP
ON
-P
. ZA
MO
RA
Le
ren
dem
ent
appa
rent
de
la f
orm
atio
n co
ntin
ue
dans
les
ent
repr
ises
: ef
fets
sur
la
prod
uctiv
ité e
t le
s sa
laire
s
G20
06/0
4 J.
-F. O
UV
RA
RD
- R
. RA
TH
ELO
T
Dem
ogra
phic
cha
nge
and
unem
ploy
men
t:
wha
t do
mac
roec
onom
etric
mod
els
pred
ict?
G20
06/0
5 D
. BLA
NC
HE
T -
J.-
F. O
UV
RA
RD
In
dica
teur
s d’
enga
gem
ents
im
plic
ites
des
syst
èmes
de
re
trai
te :
chiff
rage
s,
prop
riété
s an
alyt
ique
s et
ré
actio
ns
à de
s ch
ocs
dém
ogra
phiq
ues
type
s
G20
06/0
6 G
. B
IAU
- O
. B
IAU
- L
. R
OU
VIE
RE
N
onpa
ram
etric
For
ecas
ting
of t
he M
anuf
actu
ring
Out
put G
row
th w
ith F
irm-le
vel S
urve
y D
ata
G20
06/0
7 C
. AF
SA
- P
. GIV
OR
D
Le
rôle
de
s co
nditi
ons
de
trav
ail
dans
le
s ab
senc
es p
our
mal
adie
G20
06/0
8 P
. SIL
LAR
D -
C. L
’AN
GE
VIN
- S
. SE
RR
AV
ALL
E
Per
form
ance
s co
mpa
rées
à
l’exp
orta
tion
de
la
Fra
nce
et d
e se
s pr
inci
paux
par
tena
ires
Une
ana
lyse
str
uctu
relle
sur
12
ans
G20
06/0
9 X
. BO
UT
IN -
S. Q
UA
NT
IN
Une
m
étho
dolo
gie
d’év
alua
tion
com
ptab
le
du
coût
du
capi
tal d
es e
ntre
pris
es f
ranç
aise
s :
1984
-20
02
G20
06/1
0 C
. AF
SA
L’
estim
atio
n d’
un c
oût
impl
icite
de
la p
énib
ilité
du
trav
ail c
hez
les
trav
aille
urs
âgés
G20
06/1
1 C
. LE
LAR
GE
Le
s en
trep
rises
(in
dust
rielle
s)
fran
çais
es
sont
-el
les
à la
fron
tière
tech
nolo
giqu
e ?
G20
06/1
2 O
. B
IAU
-
N.
FE
RR
AR
I T
héor
ie d
e l’o
pini
on
Fau
t-il
pond
érer
les
répo
nses
indi
vidu
elle
s ?
G20
06/1
3 A
. KO
UB
I - S
. RO
UX
U
ne
réin
terp
réta
tion
de
la
rela
tion
entr
e pr
oduc
tivité
et
in
égal
ités
sala
riale
s da
ns
les
entr
epris
es
G20
06/1
4 R
. RA
TH
ELO
T -
P. S
ILLA
RD
T
he
impa
ct
of
loca
l ta
xes
on
plan
ts
loca
tion
deci
sion
G20
06/1
5 L.
GO
NZ
ALE
Z -
C. P
ICA
RT
D
iver
sific
atio
n, r
ecen
trag
e et
poi
ds d
es a
ctiv
ités
de s
uppo
rt d
ans
les
grou
pes
(199
3-20
00)
G20
07/0
1 D
. SR
AE
R
Allè
gem
ents
de
co
tisat
ions
pa
tron
ales
et
dy
nam
ique
sal
aria
le
G20
07/0
2 V
. ALB
OU
Y -
L. L
EQ
UIE
N
Les
rend
emen
ts n
on m
onét
aire
s de
l’é
duca
tion
: le
cas
de
la s
anté
G20
07/0
3 D
. BLA
NC
HE
T -
T.
DE
BR
AN
D
Asp
iratio
n à
la r
etra
ite,
sant
é et
sat
isfa
ctio
n au
tr
avai
l : u
ne c
ompa
rais
on e
urop
éenn
e
G20
07/0
4 M
. BA
RLE
T -
L. C
RU
SS
ON
Q
uel
impa
ct d
es v
aria
tions
du
prix
du
pétr
ole
sur
la c
rois
sanc
e fr
ança
ise
?
G20
07/0
5 C
. PIC
AR
T
Flu
x d’
empl
oi e
t de
mai
n-d’
œuv
re e
n F
ranc
e :
un
réex
amen
G20
07/0
6 V
. ALB
OU
Y -
C. T
AV
AN
M
assi
ficat
ion
et
dém
ocra
tisat
ion
de
l’ens
eign
emen
t sup
érie
ur e
n F
ranc
e
G20
07/0
7 T
. LE
BA
RB
AN
CH
ON
T
he
Cha
ngin
g re
spon
se
to
oil
pric
e sh
ocks
in
F
ranc
e: a
DS
GE
type
app
roac
h
G20
07/0
8 T
. C
HA
NE
Y -
D. S
RA
ER
- D
. TH
ES
MA
R
Col
late
ral V
alue
and
Cor
pora
te In
vest
men
t E
vide
nce
from
the
Fre
nch
Rea
l Est
ate
Mar
ket
G20
07/0
9 J.
BO
ISS
INO
T
Con
sum
ptio
n ov
er
the
Life
C
ycle
: F
acts
fo
r F
ranc
e
G20
07/1
0 C
. AF
SA
In
terp
réte
r le
s va
riabl
es
de
satis
fact
ion
: l’e
xem
ple
de la
dur
ée d
u tr
avai
l
G20
07/1
1 R
. RA
TH
ELO
T -
P. S
ILLA
RD
Z
ones
F
ranc
hes
Urb
aine
s :
quel
s ef
fets
su
r l’e
mpl
oi
sala
rié
et
les
créa
tions
d’
étab
lisse
men
ts ?
G20
07/1
2 V
. ALB
OU
Y -
B. C
RÉ
PO
N
Alé
a m
oral
en
sa
nté
: un
e év
alua
tion
dans
le
ca
dre
du m
odèl
e ca
usal
de
Rub
in
G20
08/0
1 C
. PIC
AR
T
Les
PM
E
fran
çais
es :
rent
able
s m
ais
peu
dyna
miq
ues
viii
G20
08/0
2 P
. BIS
CO
UR
P -
X. B
OU
TIN
- T
. V
ER
GÉ
T
he E
ffect
s of
Ret
ail R
egul
atio
ns o
n P
rices
E
vide
nce
form
the
Loi G
alla
nd
G20
08/0
3 Y
. BA
RB
ES
OL
- A
. BR
IAN
T
Éco
nom
ies
d’ag
glom
érat
ion
et
prod
uctiv
ité
des
entr
epris
es :
est
imat
ion
sur
donn
ées
indi
vidu
elle
s fr
ança
ises
G20
08/0
4 D
. BLA
NC
HE
T -
F. L
E G
ALL
O
Les
proj
ectio
ns
dém
ogra
phiq
ues
: pr
inci
paux
m
écan
ism
es e
t ret
our
sur
l’exp
érie
nce
fran
çais
e
G20
08/0
5 D
. BLA
NC
HE
T -
F. T
OU
TLE
MO
ND
E
Évo
lutio
ns
dém
ogra
phiq
ues
et
défo
rmat
ion
du
cycl
e de
vie
act
ive
: que
lles
rela
tions
?
G20
08/0
6 M
. BA
RLE
T -
D. B
LAN
CH
ET
- L
. CR
US
SO
N
Inte
rnat
iona
lisat
ion
et f
lux
d’em
ploi
s :
que
dit
une
appr
oche
com
ptab
le ?
G20
08/0
7 C
. LE
LAR
GE
- D
. SR
AE
R -
D. T
HE
SM
AR
E
ntre
pren
eurs
hip
and
Cre
dit
Con
stra
ints
-
Evi
denc
e fr
om
a F
renc
h Lo
an
Gua
rant
ee
Pro
gram
G20
08/0
8 X
. BO
UT
IN -
L. J
AN
IN
Are
Pric
es R
eally
Affe
cted
by
Mer
gers
?
G20
08/0
9 M
. BA
RLE
T -
A. B
RIA
NT
- L
. CR
US
SO
N
Con
cent
ratio
n gé
ogra
phiq
ue
dans
l’i
ndus
trie
m
anuf
actu
rière
et
dans
les
ser
vice
s en
Fra
nce
: un
e ap
proc
he p
ar u
n in
dica
teur
en
cont
inu
G20
08/1
0 M
. BE
FF
Y -
É. C
OU
DIN
- R
. RA
TH
ELO
T
Who
is
co
nfro
nted
to
in
secu
re
labo
r m
arke
t hi
stor
ies?
Som
e ev
iden
ce b
ased
on
the
Fre
nch
labo
r m
arke
t tra
nsiti
on
G20
08/1
1 M
. RO
GE
R -
E. W
ALR
AE
T
Soc
ial
Sec
urity
and
Wel
l-Bei
ng o
f th
e E
lder
ly:
the
Cas
e of
Fra
nce
G20
08/1
2 C
. AF
SA
A
naly
ser
les
com
posa
ntes
du
bien
-êtr
e et
de
son
évol
utio
n
Une
ap
proc
he
empi
rique
su
r do
nnée
s in
divi
duel
les
G20
08/1
3 M
. BA
RLE
T -
D. B
LAN
CH
ET
-
T.
LE B
AR
BA
NC
HO
N
Mic
rosi
mul
er le
mar
ché
du tr
avai
l : u
n pr
otot
ype
G20
09/0
1 P
.-A
. PIO
NN
IER
Le
par
tage
de
la v
aleu
r aj
outé
e en
Fra
nce,
19
49-2
007
G20
09/0
2 La
uren
t CLA
VE
L -
Chr
iste
lle M
INO
DIE
R
A
Mon
thly
In
dica
tor
of
the
Fre
nch
Bus
ines
s C
limat
e
G20
09/0
3 H
. ER
KE
L-R
OU
SS
E -
C. M
INO
DIE
R
Do
Bus
ines
s T
ende
ncy
Sur
veys
in
Indu
stry
and
S
ervi
ces
Hel
p in
For
ecas
ting
GD
P G
row
th?
A
Rea
l-Tim
e A
naly
sis
on F
renc
h D
ata
G20
09/0
4 P
. GIV
OR
D -
L. W
ILN
ER
Le
s co
ntra
ts t
empo
raire
s :
trap
pe o
u m
arch
epie
d ve
rs l’
empl
oi s
tabl
e ?
G20
09/0
5 G
. LA
LAN
NE
- P
.-A
. P
ION
NIE
R -
O.
SIM
ON
Le
par
tage
des
fru
its d
e la
cro
issa
nce
de 1
950
à 20
08 :
une
appr
oche
par
les
com
ptes
de
surp
lus
G20
09/0
6 L.
DA
VE
ZIE
S -
X.
D’H
AU
LTF
OE
UIL
LE
Fau
t-il
pond
érer
?…
O
u l’é
tern
elle
qu
estio
n de
l’é
cono
mèt
re c
onfr
onté
à d
es d
onné
es d
’enq
uête
G20
09/0
7 S
. QU
AN
TIN
- S
. RA
SP
ILLE
R -
S. S
ER
RA
VA
LLE
C
omm
erce
in
trag
roup
e,
fisca
lité
et
prix
de
tr
ansf
erts
: un
e an
alys
e su
r do
nnée
s fr
ança
ises
G20
09/0
8 M
. CLE
RC
- V
. MA
RC
US
É
last
icité
s-pr
ix d
es c
onso
mm
atio
ns é
nerg
étiq
ues
des
mén
ages
G20
09/0
9 G
. LA
LAN
NE
- E
. P
OU
LIQ
UE
N -
O.
SIM
ON
P
rix d
u pé
trol
e et
cro
issa
nce
pote
ntie
lle à
lon
g te
rme
G20
09/1
0 D
. BLA
NC
HE
T -
J. L
E C
AC
HE
UX
-
V. M
AR
CU
S
Adj
uste
d ne
t sa
ving
s an
d ot
her
appr
oach
es
to
sust
aina
bilit
y: s
ome
theo
retic
al b
ackg
roun
d
G20
09/1
1 V
. B
ELL
AM
Y -
G.
CO
NS
ALE
S -
M.
FE
SS
EA
U -
S
. LE
LA
IDIE
R -
É. R
AY
NA
UD
U
ne d
écom
posi
tion
du c
ompt
e de
s m
énag
es d
e la
co
mpt
abili
té
natio
nale
pa
r ca
tégo
rie
de
mén
age
en 2
003
G20
09/1
2 J.
BA
RD
AJI
- F
. T
ALL
ET
D
etec
ting
Eco
nom
ic
Reg
imes
in
F
ranc
e:
a Q
ualit
ativ
e M
arko
v-S
witc
hing
In
dica
tor
Usi
ng
Mix
ed F
requ
ency
Dat
a
G20
09/1
3 R
. A
EB
ER
HA
RD
T
- D
. F
OU
GÈ
RE
-
R.
RA
TH
ELO
T
Dis
crim
inat
ion
à l’e
mba
uche
: co
mm
ent
expl
oite
r le
s pr
océd
ures
de
test
ing
?
G20
09/1
4 Y
. BA
RB
ES
OL
- P
. GIV
OR
D -
S. Q
UA
NT
IN
Par
tage
de
la
va
leur
aj
outé
e,
appr
oche
pa
r do
nnée
s m
icro
écon
omiq
ues
G20
09/1
5 I.
BU
ON
O -
G. L
ALA
NN
E
The
Effe
ct o
f th
e U
rugu
ay r
ound
on
the
Inte
nsiv
e an
d E
xten
sive
Mar
gins
of T
rade
G20
10/0
1 C
. MIN
OD
IER
A
vant
ages
com
paré
s de
s sé
ries
des
prem
ière
s va
leur
s pu
blié
es
et
des
série
s de
s va
leur
s ré
visé
es -
Un
exer
cice
de
prév
isio
n en
tem
ps r
éel
de la
cro
issa
nce
trim
estr
ielle
du
PIB
en
Fra
nce
G20
10/0
2 V
. ALB
OU
Y -
L. D
AV
EZ
IES
- T
. D
EB
RA
ND
H
ealth
Exp
endi
ture
Mod
els:
a C
ompa
rison
of
Fiv
e S
peci
ficat
ions
usi
ng P
anel
Dat
a
G20
10/0
3 C
. KLE
IN -
O. S
IMO
N
Le m
odèl
e M
ÉS
AN
GE
rée
stim
é en
bas
e 20
00
Tom
e 1
– V
ersi
on a
vec
volu
mes
à p
rix c
onst
ants
G20
10/0
4 M
.-É
. CLE
RC
- É
. CO
UD
IN
L’IP
C,
miro
ir de
l’é
volu
tion
du c
oût
de l
a vi
e en
F
ranc
e ?
Ce
qu’a
ppor
te
l’ana
lyse
de
s co
urbe
s d’
Eng
el
G20
10/0
5 N
. CE
CI-
RE
NA
UD
- P
.-A
. CH
EV
ALI
ER
Le
s se
uils
de
10,
20 e
t 50
sal
arié
s :
impa
ct s
ur la
ta
ille
des
entr
epris
es fr
ança
ises
G20
10/0
6 R
. AE
BE
RH
AR
DT
- J
. PO
UG
ET
N
atio
nal
Orig
in
Diff
eren
ces
in
Wag
es
and
Hie
rarc
hica
l P
ositi
ons
- E
vide
nce
on F
renc
h F
ull-
Tim
e M
ale
Wor
kers
fro
m a
mat
ched
Em
ploy
er-
Em
ploy
ee D
atas
et
G20
10/0
7 S
. BLA
SC
O -
P. G
IVO
RD
Le
s tr
ajec
toire
s pr
ofes
sion
nelle
s en
déb
ut d
e vi
e ac
tive
: que
l im
pact
des
con
trat
s te
mpo
raire
s ?
G20
10/0
8 P
. GIV
OR
D
Mét
hode
s éc
onom
étriq
ues
pour
l’é
valu
atio
n de
po
litiq
ues
publ
ique
s
ix
G20
10/0
9 P
.-Y
. CA
BA
NN
ES
- V
. LA
PÈ
GU
E -
E
. P
OU
LIQ
UE
N -
M.
BE
FF
Y -
M.
GA
INI
Que
lle
croi
ssan
ce
de
moy
en
term
e ap
rès
la
cris
e ?
G20
10/1
0 I.
BU
ON
O -
G. L
ALA
NN
E
La r
éact
ion
des
entr
epris
es fr
ança
ises
à
la b
aiss
e de
s ta
rifs
doua
nier
s ét
rang
ers
G20
10/1
1 R
. RA
TH
ELO
T -
P. S
ILLA
RD
L’
appo
rt d
es m
étho
des
à no
yaux
pou
r m
esur
er la
co
ncen
trat
ion
géog
raph
ique
-
App
licat
ion
à la
co
ncen
trat
ion
des
imm
igré
s en
Fra
nce
de 1
968
à 19
99
G20
10/1
2 M
. B
AR
AT
ON
- M
. B
EF
FY
- D
. F
OU
GÈ
RE
U
ne é
valu
atio
n de
l’e
ffet
de l
a ré
form
e de
200
3 su
r le
s dé
part
s en
re
trai
te
- Le
ca
s de
s en
seig
nant
s du
sec
ond
degr
é pu
blic
G20
10/1
3 D
. B
LAN
CH
ET
- S
. B
UF
FE
TE
AU
- E
. C
RE
NN
ER
S
. LE
MIN
EZ
Le
m
odèl
e de
m
icro
sim
ulat
ion
Des
tinie
2
: pr
inci
pale
s ca
ract
éris
tique
s et
pre
mie
rs r
ésul
tats
G20
10/1
4 D
. BLA
NC
HE
T -
E. C
RE
NN
ER
Le
blo
c re
trai
tes
du m
odèl
e D
estin
ie 2
:
guid
e de
l’ut
ilisa
teur
G20
10/1
5 M
. B
AR
LET
-
L.
CR
US
SO
N
- S
. D
UP
UC
H
- F
. PU
EC
H
Des
se
rvic
es
écha
ngés
au
x se
rvic
es
écha
n-ge
able
s : u
ne a
pplic
atio
n su
r do
nnée
s fr
ança
ises
G20
10/1
6 M
. B
EF
FY
- T
. K
AM
ION
KA
P
ublic
-priv
ate
wag
e ga
ps:
is c
ivil-
serv
ant
hum
an
capi
tal s
ecto
r-sp
ecifi
c?
G20
10/1
7 P
.-Y
. C
AB
AN
NE
S
- H
. E
RK
EL-
RO
US
SE
-
G.
LALA
NN
E -
O.
MO
NS
O -
E.
PO
ULI
QU
EN
Le
mod
èle
Més
ange
rée
stim
é en
bas
e 20
00
Tom
e 2
- V
ersi
on a
vec
volu
mes
à p
rix c
haîn
és
G20
10/1
8 R
. AE
BE
RH
AR
DT
- L
. DA
VE
ZIE
S
Con
ditio
nal
Logi
t w
ith o
ne B
inar
y C
ovar
iate
: Li
nk
betw
een
the
Sta
tic a
nd D
ynam
ic C
ases
G20
11/0
1 T
. LE
BA
RB
AN
CH
ON
- B
. OU
RLI
AC
- O
. SIM
ON
Le
s m
arch
és d
u tr
avai
l fra
nçai
s et
am
éric
ain
face
au
x ch
ocs
conj
onct
urel
s de
s an
nées
19
86
à 20
07 :
une
mod
élis
atio
n D
SG
E
G20
11/0
2 C
. MA
RB
OT
U
ne
éval
uatio
n de
la
ré
duct
ion
d’im
pôt
pour
l’e
mpl
oi d
e sa
larié
s à
dom
icile
G20
11/0
3 L.
DA
VE
ZIE
S
Mod
èles
à
effe
ts
fixes
, à
effe
ts
aléa
toire
s,
mod
èles
mix
tes
ou m
ulti-
nive
aux
: pr
oprié
tés
et
mis
es
en
œuv
re
des
mod
élis
atio
ns
de
l’hét
érog
énéi
té d
ans
le c
as d
e do
nnée
s gr
oupé
es
G20
11/0
4 M
. RO
GE
R -
M. W
AS
ME
R
Het
erog
enei
ty
mat
ters
: la
bour
pr
oduc
tivity
di
ffere
ntia
ted
by a
ge a
nd s
kills
G20
11/0
5 J.
-C.
BR
ICO
NG
NE
- J
.-M
. F
OU
RN
IER
V
. LA
PÈ
GU
E -
O.
MO
NS
O
De
la
cris
e fin
anci
ère
à la
cr
ise
écon
omiq
ue
L’im
pact
des
per
turb
atio
ns f
inan
cièr
es d
e 20
07 e
t 20
08 s
ur la
cro
issa
nce
de s
ept p
ays
indu
stria
lisés
G20
11/0
6 P
. C
HA
RN
OZ
- É
. C
OU
DIN
- M
. G
AIN
I W
age
ineq
ualit
ies
in F
ranc
e 19
76-2
004:
a
quan
tile
regr
essi
on a
naly
sis
G20
11/0
7 M
. CLE
RC
- M
. GA
INI -
D. B
LAN
CH
ET
R
ecom
men
datio
ns
of
the
Stig
litz-
Sen
-Fito
ussi
R
epor
t: A
few
illu
stra
tions
G20
11/0
8 M
. BA
CH
ELE
T -
M. B
EF
FY
- D
. BLA
NC
HE
T
Pro
jete
r l’i
mpa
ct d
es r
éfor
mes
des
ret
raite
s su
r l’a
ctiv
ité d
es 5
5 an
s et
plu
s :
une
com
para
ison
de
troi
s m
odèl
es
G20
11/0
9 C
. LO
UV
OT
-RU
NA
VO
T
L’év
alua
tion
de
l’act
ivité
di
ssim
ulée
de
s en
tre-
pris
es s
ur l
a ba
se d
es c
ontr
ôles
fis
caux
et
son
inse
rtio
n da
ns le
s co
mpt
es n
atio
naux
G20
11/1
0 A
. SC
HR
EIB
ER
- A
. VIC
AR
D
La
tert
iaris
atio
n de
l’é
cono
mie
fr
ança
ise
et
le
rale
ntis
sem
ent
de l
a pr
oduc
tivité
ent
re 1
978
et
2008
G20
11/1
1 M
.-É
. C
LER
C -
O.
MO
NS
O -
E. P
OU
LIQ
UE
N
Les
inég
alité
s en
tre
géné
ratio
ns d
epui
s le
bab
y-bo
om
G20
11/1
2 C
. M
AR
BO
T -
D.
RO
Y
Éva
luat
ion
de l
a tr
ansf
orm
atio
n de
la
rédu
ctio
n d'
impô
t en
cré
dit
d'im
pôt
pour
l'em
ploi
de
sala
riés
à do
mic
ile e
n 20
07
G20
11/1
3 P
. GIV
OR
D -
R. R
AT
HE
LOT
- P
. SIL
LAR
D
Pla
ce-b
ased
ta
x ex
empt
ions
an
d di
spla
cem
ent
effe
cts:
A
n ev
alua
tion
of
the
Zon
es
Fra
nche
s U
rbai
nes
prog
ram
G20
11/1
4 X
. D
’HA
ULT
FO
EU
ILLE
-
P.
GIV
OR
D
- X
. B
OU
TIN
T
he E
nviro
nmen
tal
Effe
ct o
f G
reen
Tax
atio
n: t
he
Cas
e of
the
Fre
nch
“Bon
us/M
alus
”
G20
11/1
5 M
. B
AR
LET
-
M.
CLE
RC
-
M.
GA
RN
EO
-
V.
LAP
ÈG
UE
- V
. M
AR
CU
S
La
nouv
elle
ve
rsio
n du
m
odèl
e M
ZE
, m
odèl
e m
acro
écon
omét
rique
pou
r la
zon
e eu
ro
G20
11/1
6 R
. A
EB
ER
HA
RD
T -
I.
BU
ON
O -
H. F
AD
ING
ER
Le
arni
ng,
Inco
mpl
ete
Con
trac
ts
and
Exp
ort
Dyn
amic
s:
The
ory
and
Evi
denc
e fo
rm
Fre
nch
Firm
s
G20
11/1
7 C
. KE
RD
RA
IN -
V. L
AP
ÈG
UE
R
estr
ictiv
e F
isca
l Pol
icie
s in
Eur
ope:
W
hat a
re th
e Li
kely
Effe
cts?
G20
12/0
1 P
. G
IVO
RD
- S
. Q
UA
NT
IN -
C. T
RE
VIE
N
A L
ong-
Ter
m E
valu
atio
n of
the
Firs
t G
ener
atio
n of
the
Fre
nch
Urb
an E
nter
pris
e Z
ones
G20
12/0
2 N
. CE
CI-
RE
NA
UD
- V
. CO
TT
ET
P
oliti
que
sala
riale
et p
erfo
rman
ce d
es e
ntre
pris
es
G20
12/0
3 P
. FÉ
VR
IER
- L
. WIL
NE
R
Do
Con
sum
ers
Cor
rect
ly
Exp
ect
Pric
e R
educ
tions
? T
estin
g D
ynam
ic B
ehav
ior
G20
12/0
4 M
. GA
INI -
A. L
ED
UC
- A
. VIC
AR
D
Sch
ool
as a
she
lter?
Sch
ool
leav
ing-
age
and
the
busi
ness
cyc
le in
Fra
nce
G20
12/0
5 M
. GA
INI -
A. L
ED
UC
- A
. VIC
AR
D
A s
carr
ed g
ener
atio
n? F
renc
h ev
iden
ce o
n yo
ung
peop
le e
nter
ing
into
a to
ugh
labo
ur m
arke
t
G20
12/0
6 P
. AU
BE
RT
- M
. BA
CH
ELE
T
Dis
parit
és
de
mon
tant
de
pe
nsio
n et
re
dist
ribut
ion
dans
le s
ystè
me
de r
etra
ite fr
ança
is
G20
12/0
7 R
. AE
BE
RH
AR
DT
- P
GIV
OR
D -
C. M
AR
BO
T
Spi
llove
r E
ffect
of
the
Min
imum
Wag
e in
Fra
nce:
A
n U
ncon
ditio
nal Q
uant
ile R
egre
ssio
n A
ppro
ach
x
G20
12/0
8 A
. EID
ELM
AN
- F
. LA
NG
UM
IER
- A
. VIC
AR
D
Pré
lève
men
ts
oblig
atoi
res
repo
sant
su
r le
s m
énag
es :
des
can
aux
redi
strib
utifs
diff
éren
ts e
n 19
90 e
t 201
0
G20
12/0
9 O
. BA
RG
AIN
- A
. VIC
AR
D
Le
RM
I et
son
suc
cess
eur
le R
SA
déc
oura
gent
-ils
cer
tain
s je
unes
de
trav
aille
r ?
Une
ana
lyse
sur
le
s je
unes
aut
our
de 2
5 an
s
G20
12/1
0 C
. MA
RB
OT
- D
. R
OY
P
roje
ctio
ns
du
coût
de
l’A
PA
et
de
s ca
ract
éris
tique
s de
ses
bén
éfic
iaire
s à
l’hor
izon
20
40 à
l’ai
de d
u m
odèl
e D
estin
ie
G20
12/1
1 A
. MA
UR
OU
X
Le
créd
it d’
impô
t dé
dié
au
déve
lopp
emen
t du
rabl
e : u
ne é
valu
atio
n éc
onom
étriq
ue
G20
12/1
2 V
. CO
TT
ET
- S
. QU
AN
TIN
- V
. RÉ
GN
IER
C
oût
du t
rava
il et
allè
gem
ents
de
char
ges
: un
e es
timat
ion
au
nive
au
étab
lisse
men
t de
19
96
à 20
08
G20
12/1
3 X
. D
’HA
ULT
FO
EU
ILLE
-
P.
FÉ
VR
IER
-
L. W
ILN
ER
D
eman
d E
stim
atio
n in
the
Pre
senc
e of
Rev
enue
M
anag
emen
t
G20
12/1
4 D
. BLA
NC
HE
T -
S. L
E M
INE
Z
Join
t m
acro
/mic
ro e
valu
atio
ns o
f ac
crue
d-to
-dat
e pe
nsio
n lia
bilit
ies:
an
ap
plic
atio
n to
F
renc
h re
form
s
G20
13/0
1-
T.
DE
RO
YO
N -
A. M
ON
TA
UT
- P
-A P
ION
NIE
R
F13
01
Util
isat
ion
rétr
ospe
ctiv
e de
l’e
nquê
te
Em
ploi
à
une
fréq
uenc
e m
ensu
elle
: ap
port
d’
une
mod
élis
atio
n es
pace
-éta
t
G20
13/0
2-
C. T
RE
VIE
N
F13
02
Hab
iter
en
HLM
: qu
el
avan
tage
m
onét
aire
et
qu
el im
pact
sur
les
cond
ition
s de
loge
men
t ?
G20
13/0
3 A
. PO
ISS
ON
NIE
R
T
empo
ral
disa
ggre
gatio
n of
sto
ck v
aria
bles
- T
he
Cho
w-L
in m
etho
d ex
tend
ed to
dyn
amic
mod
els
G20
13/0
4 P
. GIV
OR
D -
C.
MA
RB
OT
Doe
s th
e co
st o
f ch
ild c
are
affe
ct f
emal
e la
bor
mar
ket
part
icip
atio
n? A
n ev
alua
tion
of a
Fre
nch
refo
rm o
f chi
ldca
re s
ubsi
dies
G20
13/0
5 G
. LA
ME
- M
. LE
QU
IEN
- P
.-A
. P
ION
NIE
R
In
terp
reta
tion
and
limits
of
sust
aina
bilit
y te
sts
in
publ
ic fi
nanc
e
G20
13/0
6 C
. BE
LLE
GO
- V
. DO
RT
ET
-BE
RN
AD
ET
La p
artic
ipat
ion
aux
pôle
s de
com
pétit
ivité
: qu
elle
in
cide
nce
sur
les
dépe
nses
de
R&
D e
t l’a
ctiv
ité
des
PM
E e
t E
TI ?
G20
13/0
7 P
.-Y
. CA
BA
NN
ES
- A
. MO
NT
AU
T -
P
.-A
. P
ION
NIE
R
É
valu
er
la p
rodu
ctiv
ité g
loba
le d
es f
acte
urs
en
Fra
nce
: l’a
ppor
t d’
une
mes
ure
de l
a qu
alité
du
capi
tal e
t du
trav
ail
G20
13/0
8 R
. AE
BE
RH
AR
DT
- C
. MA
RB
OT
Evo
lutio
n of
In
stab
ility
on
th
e F
renc
h La
bour
M
arke
t Dur
ing
the
Last
Thi
rty
Yea
rs
G20
13/0
9 J-
B. B
ER
NA
RD
- G
. CLÉ
AU
D
O
il pr
ice:
the
nat
ure
of t
he s
hock
s an
d th
e im
pact
on
the
Fre
nch
econ
omy
G20
13/1
0 G
. LA
ME
Was
the
re a
« G
reen
span
Con
undr
um »
in
the
Eur
o ar
ea?
G20
13/1
1 P
. CH
ON
É -
F. E
VA
IN -
L.
WIL
NE
R -
E.
YIL
MA
Z
In
trod
ucin
g ac
tivity
-bas
ed p
aym
ent
in t
he h
ospi
tal
indu
stry
: E
vide
nce
from
Fre
nch
data
G20
13/1
2 C
. GR
ISLA
IN-L
ET
RÉ
MY
Nat
ural
Dis
aste
rs: E
xpos
ure
and
Und
erin
sura
nce
G20
13/1
3 P
.-Y
. C
AB
AN
NE
S -
V.
CO
TT
ET
- Y
. D
UB
OIS
-
C.
LELA
RG
E -
M.
SIC
SIC
F
renc
h F
irms
in th
e F
ace
of th
e 20
08/2
009
Cris
is
G20
13/1
4 A
. PO
ISS
ON
NIE
R -
D. R
OY
Hou
seho
lds
Sat
ellit
e A
ccou
nt f
or F
ranc
e in
201
0.
Met
hodo
logi
cal
issu
es
on
the
asse
ssm
ent
of
dom
estic
pro
duct
ion
G20
13/1
5 G
. CLÉ
AU
D -
M. L
EM
OIN
E -
P.-
A. P
ION
NIE
R
W
hich
si
ze
and
evol
utio
n of
th
e go
vern
men
t ex
pend
iture
mul
tiplie
r in
Fra
nce
(198
0-20
10)?
G20
14/0
1 M
. BA
CH
ELE
T -
A. L
ED
UC
- A
. MA
RIN
O
Le
s bi
ogra
phie
s du
mod
èle
Des
tinie
II
: re
basa
ge
et p
roje
ctio
n
G20
14/0
2 B
. GA
RB
INT
I
L’ac
hat
de l
a ré
side
nce
prin
cipa
le e
t la
cré
atio
n d’
entr
epris
es s
ont-
ils f
avor
isés
par
les
don
atio
ns
et h
érita
ges
?
G20
14/0
3 N
. CE
CI-
RE
NA
UD
- P
. CH
AR
NO
Z -
M. G
AIN
I
Évo
lutio
n de
la v
olat
ilité
des
rev
enus
sal
aria
ux d
u se
cteu
r pr
ivé
en F
ranc
e de
puis
196
8
G20
14/0
4 P
. AU
BE
RT
Mod
alité
s d’
appl
icat
ion
des
réfo
rmes
des
ret
raite
s et
pré
visi
bilit
é du
mon
tant
de
pens
ion
G20
14/0
5 C
. GR
ISLA
IN-L
ET
RÉ
MY
- A
. KA
TO
SS
KY
The
Im
pact
of
Haz
ardo
us I
ndus
tria
l F
acili
ties
on
Hou
sing
Pric
es:
A C
ompa
rison
of
Par
amet
ric a
nd
Sem
ipar
amet
ric H
edon
ic P
rice
Mod
els
G20
14//0
6 J.
-M. D
AU
SS
IN-B
EN
ICH
OU
- A
. MA
UR
OU
X
Tur
ning
th
e he
at
up.
How
se
nsiti
ve
are
hous
ehol
ds
to
fisca
l in
cent
ives
on
en
ergy
ef
ficie
ncy
inve
stm
ents
?
G20
14/0
7 C
. LA
BO
NN
E -
G. L
AM
É
Cre
dit
Gro
wth
and
Cap
ital R
equi
rem
ents
: B
indi
ng
or N
ot?
G20
14/0
8 C
. GR
ISLA
IN-L
ET
RÉ
MY
et C
. TR
EV
IEN
T
he I
mpa
ct o
f H
ousi
ng S
ubsi
dies
on
the
Ren
tal
Sec
tor:
the
Fre
nch
Exa
mpl
e
G20
14/0
9 M
. LE
QU
IEN
et A
. MO
NT
AU
T
Cro
issa
nce
pote
ntie
lle
en
Fra
nce
et
en
zone
eu
ro :
un
tour
d’
horiz
on
des
mét
hode
s d’
estim
atio
n
G20
14/1
0 B
. GA
RB
INT
I -
P. L
AM
AR
CH
E
Les
haut
s re
venu
s ép
argn
ent-
ils d
avan
tage
?
G20
14/1
1 D
. AU
DE
NA
ER
T -
J. B
AR
DA
JI -
R. L
AR
DE
UX
-
M.
OR
AN
D -
M.
SIC
SIC
W
age
Res
ilien
ce
in
Fra
nce
sinc
e th
e G
reat
R
eces
sion
G20
14/1
2 F
. A
RN
AU
D -
J.
BO
US
SA
RD
- A
. P
OIS
SO
NN
IER
-
H. S
OU
AL
Com
putin
g ad
ditiv
e co
ntrib
utio
ns t
o gr
owth
and
ot
her
issu
es fo
r ch
ain-
linke
d qu
arte
rly a
ggre
gate
s
G20
14/1
3 H
. F
RA
ISS
E -
F.
KR
AM
AR
Z -
C.
PR
OS
T
Labo
r D
ispu
tes
and
Job
Flo
ws
xi
G20
14/1
4 P
. G
IVO
RD
-
C.
GR
ISLA
IN-L
ET
RÉ
MY
-
H.
NA
EG
ELE
H
ow
does
fu
el
taxa
tion
impa
ct
new
ca
r pu
rcha
ses?
A
n ev
alua
tion
usin
g F
renc
h co
nsum
er-le
vel d
atas
et
G20
14/1
5 P
. AU
BE
RT
- S
. RA
BA
TÉ
D
urée
pa
ssée
en
ca
rriè
re
et
duré
e de
vi
e en
re
trai
te :
que
l pa
rtag
e de
s ga
ins
d'es
péra
nce
de
vie
?
G20
15/0
1 A
. PO
ISS
ON
NIE
R
The
wal
king
dea
d E
uler
equ
atio
n A
ddre
ssin
g a
chal
leng
e to
m
onet
ary
polic
y m
odel
s
G20
15/0
2 Y
. D
UB
OIS
- A
. M
AR
INO
In
dica
teur
s de
ren
dem
ent
du s
ystè
me
de r
etra
ite
fran
çais
G20
15/0
3 T
. M
AY
ER
- C
. TR
EV
IEN
T
he
impa
cts
of
Urb
an
Pub
lic
Tra
nspo
rtat
ion:
E
vide
nce
from
the
Par
is R
egio
n
G20
15/0
4 S
.T.
LY -
A. R
IEG
ER
T
Mea
surin
g S
ocia
l Env
ironm
ent M
obili
ty
G20
15/0
5 M
. A. B
EN
HA
LIM
A -
V. H
YA
FIL
-SO
LELH
AC
M
. K
OU
BI
- C
. R
EG
AE
RT
Q
uel
est
l’im
pact
du
sy
stèm
e d’
inde
mni
satio
n m
alad
ie s
ur l
a du
rée
des
arrê
ts d
e tr
avai
l po
ur
mal
adie
?
G20
15/0
6 Y
. D
UB
OIS
- A
. M
AR
INO
D
ispa
rités
de
rend
emen
t du
sys
tèm
e de
ret
raite
da
ns
le
sect
eur
priv
é :
appr
oche
s in
terg
énér
a-tio
nnel
le e
t int
ragé
néra
tionn
elle
G20
15/0
7 B
. CA
MP
AG
NE
- V
. ALH
EN
C-G
ELA
S -
J.
-B.
BE
RN
AR
D
No
evid
ence
of f
inan
cial
acc
eler
ator
in F
ranc
e
G20
15/0
8 Q
. LA
FF
ÉT
ER
- M
. PA
K
Éla
stic
ités
des
rece
ttes
fisca
les
au
cycl
e éc
onom
ique
: é
tude
de
troi
s im
pôts
sur
la p
ério
de
1979
-201
3 en
Fra
nce
G20
15/0
9 J.
-M.
DA
US
SIN
-BE
NIC
HO
U,
S.
IDM
AC
HIC
HE
, A
. LE
DU
C e
t E
. P
OU
LIQ
UE
N
Les
déte
rmin
ants
de
l’attr
activ
ité d
e la
fon
ctio
n pu
bliq
ue d
e l’É
tat
G20
15/1
0 P
. AU
BE
RT
La
mod
ulat
ion
du m
onta
nt d
e pe
nsio
n se
lon
la
duré
e de
car
rière
et
l’âge
de
la r
etra
ite :
que
lles
disp
arité
s en
tre
assu
rés
?
G20
15/1
1 V
. DO
RT
ET
-BE
RN
AD
ET
- M
. SIC
SIC
E
ffet
des
aide
s pu
bliq
ues
sur
l’em
ploi
en
R&
D
dans
les
petit
es e
ntre
pris
es
G20
15/1
2 S
. G
EO
RG
ES
-KO
T
Ann
ual
and
lifet
ime
inci
denc
e of
the
val
ue-a
dded
ta
x in
Fra
nce
G20
15/1
3 M
. PO
ULH
ÈS
A
re
Ent
erpr
ise
Zon
es
Ben
efits
C
apita
lized
in
to
Com
mer
cial
Pro
pert
y V
alue
s? T
he F
renc
h C
ase
G20
15/1
4 J.
-B. B
ER
NA
RD
- Q
. LA
FF
ÉT
ER
E
ffet
de l’
activ
ité e
t de
s pr
ix s
ur le
rev
enu
sala
rial
des
diffé
rent
es c
atég
orie
s so
ciop
rofe
ssio
nnel
les
G20
15/1
5 C
. G
EA
Y -
M.
KO
UB
I -
G d
e LA
GA
SN
ER
IE
Pro
ject
ions
de
s dé
pens
es
de
soin
s de
vi
lle,
cons
truc
tion
d’un
mod
ule
pour
Des
tinie
G20
15/1
6 J.
BA
RD
AJI
- J
.-C
. B
RIC
ON
GN
E -
B
. C
AM
PA
GN
E -
G.
GA
ULI
ER
C
ompa
red
perf
orm
ance
s of
Fre
nch
com
pani
es
on th
e do
mes
tic a
nd fo
reig
n m
arke
ts
G20
15/1
7 C
. BE
LLÉ
GO
- R
. DE
NIJ
S
The
red
istr
ibut
ive
effe
ct o
f on
line
pira
cy o
n th
e bo
x of
fice
perf
orm
ance
of
A
mer
ican
m
ovie
s in
fo
reig
n m
arke
ts
G20
15/1
8 J.
-B. B
ER
NA
RD
- L
. BE
RT
HE
T
Fre
nch
hous
ehol
ds
finan
cial
w
ealth
: w
hich
ch
ange
s in
20
year
s?
G20
15/1
9 M
. PO
ULH
ÈS
Fe
nêtre
sur
Cou
r ou
Cha
mbr
e av
ec V
ue ?
Le
s pr
ix h
édon
ique
s de
l’im
mob
ilier
par
isie
n
G20
16/0
1 B
. G
AR
BIN
TI
- S
. G
EO
RG
ES
-KO
T
Tim
e to
sm
ell t
he r
oses
? R
isk
aver
sion
, the
tim
ing
of in
herit
ance
rec
eipt
, and
ret
irem
ent
G20
16/0
2 P
. CH
AR
NO
Z -
C. L
ELA
RG
E -
C. T
RE
VIE
N
Com
mun
icat
ion
Cos
ts
and
the
Inte
rnal
O
rgan
izat
ion
of M
ulti-
Pla
nt B
usin
esse
s: E
vide
nce
from
the
Impa
ct o
f the
Fre
nch
Hig
h-S
peed
Rai
l
G20
16/0
3 C
. B
ON
NE
T -
B.
GA
RB
INT
I -
A. S
OLA
Z
Gen
der
Ineq
ualit
y af
ter
Div
orce
: T
he F
lip S
ide
of
Mar
ital
Spe
cial
izat
ion
- E
vide
nce
from
a F
renc
h A
dmin
istr
ativ
e D
atab
ase
G20
16/0
4 D
. B
LAN
CH
ET
-
E.
CA
RO
LI
- C
. P
RO
ST
-
M.
RO
GE
R
Hea
lth c
apac
ity to
wor
k at
old
er a
ges
in F
ranc
e
G20
16/0
5 B
. C
AM
PA
GN
E -
A.
PO
ISS
ON
NIE
R
ME
LEZ
E:
A D
SG
E m
odel
for
Fra
nce
with
in t
he
Eur
o A
rea
