Description
The purpose of this paper is to study the impact of the 2007 Italian severance payment
reform on the cost and the access to credit for small- and medium-sized enterprises (SMEs).
Journal of Financial Economic Policy
An analysis of the effects of the severance payment reform on credit to Italian SMEs
Riccardo Calcagno Roman Kraeussl Chiara Monticone
Article information:
To cite this document:
Riccardo Calcagno Roman Kraeussl Chiara Monticone, (2011),"An analysis of the effects of the severance
payment reform on credit to Italian SMEs", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 243 - 261
Permanent link to this document:
http://dx.doi.org/10.1108/17576381111152227
Downloaded on: 24 January 2016, At: 21:42 (PT)
References: this document contains references to 33 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 204 times since 2011*
Users who downloaded this article also downloaded:
Sameeksha Desai, J ohan Eklund, Andreas Högberg, (2011),"Pro-market reforms and allocation
of capital in India", J ournal of Financial Economic Policy, Vol. 3 Iss 2 pp. 123-139 http://
dx.doi.org/10.1108/17576381111133606
Ignacio Lozano, Karen Rodríguez, (2011),"Assessing the macroeconomic effects of fiscal
policy in Colombia", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 206-228 http://
dx.doi.org/10.1108/17576381111152209
Marco A. Espinosa-Vega, J uan Solé, (2011),"Cross-border financial surveillance: a
network perspective", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 182-205 http://
dx.doi.org/10.1108/17576381111152191
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
An analysis of the effects
of the severance payment reform
on credit to Italian SMEs
Riccardo Calcagno
Department of Economics, Finance and Control, EM Lyon Business School,
Ecully, France and
CeRP/Collegio Carlo Alberto, Torino, Italy
Roman Kraeussl
Department of Finance and Financial Sector Management,
VU University Amsterdam, Amsterdam, The Netherlands, and
Chiara Monticone
CeRP/Collegio Carlo Alberto, Torino, Italy
Abstract
Purpose – The purpose of this paper is to study the impact of the 2007 Italian severance payment
reform on the cost and the access to credit for small- and medium-sized enterprises (SMEs).
Design/methodology/approach – The authors study the implications of the reform adapting the
theoretical credit-rationing model of Holmstrom and Tirole, then estimate the capital out?ows due to
the reform and, using the theoretical prediction, assess its impact using mathematical simulations.
Findings – The authors predict that the reform may cause severe credit constraints to SMEs which
cannot pledge enough collateral in order to obtain credit. The most direct consequences are to reduce in
the long run the amount of liquid assets available to Italian ?rms, and to reduce their aggregate
investment in a more than proportional way, due to access to credit restrictions. However, it will not
increase the cost of intermediated ?nance, ceteris paribus.
Practical implications – The fact that the reform restricts access to credit, but does not increase the
cost of debt, has important policy consequences, as public interventions subsidizing credit through a
constant cost of debt may be ineffective.
Originality/value – While the topic has been analyzed in several respects (e.g. workers’
participation to the reform, cost of an access to credit subsidy, etc.), no other study proposed an
integrated view of these effects with a rigorous micro-economic approach.
Keywords Italy, Legislation, Severance payment, Moral hazard, Credit constraints,
Small to medium-sized enterprises
Paper type Research paper
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – G31, G32, G38
This research has been supported by the Center for Research on Pensions and Welfare Policies
(CeRP), at Collegio Carlo Alberto. The authors wishto thank Stefan Arping, Onorato Castellino, Elsa
Fornero, Giovanna Nicodano, as well as participants to seminars at Collegio Carlo Alberto, VU
University Amsterdam, 2007 Conference on Macroeconomics and International Finance at the
University of Crete for their useful comments. The authors owe many thanks to Fondazione Rodolfo
Debenedetti (FRDB) for providing access to the TFR survey, to CERIS-CNR for providing access to
AIDA, and to Simone Ceccarelli (Covip), Angelo Pace (Bank of Italy), Alessandro Cappellini,
Vilma Marchese and Martin Marchese (Intesa-SanPaolo) for their assistance inthe collection of data.
Severance
payment
reform
243
Journal of Financial Economic Policy
Vol. 3 No. 3, 2011
pp. 243-261
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576381111152227
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
1. Introduction
Berger and Udell (1998) observe that small, private businesses are “acutely
informationally opaque” and argue that this is a major impediment for them to
access the public capital markets. Small ?rms are forced to rely mostly on internal cash
resources and on intermediated ?nance, because intermediaries have some advantages
on direct lending through the debt markets. They can offer small- and medium-sized
enterprises (SMEs)-speci?c contracts that alleviate the acute adverse selection
problem (Bester, 1985) and they can monitor ?rms activity reducing moral hazard
(Diamond, 1984).
A recent reform of the system of severance indemnities in Italy is likely to affect the
amount of liquid assets of Italian SMEs, hence their main source of internal ?nance. This
law, included in the Financial Budget Law for 2007, allows employees to choose from
July 2007 on whether they want the future ?ows of their severance indemnity fund
(“Trattamento di Fine Rapporto”, TFR) to be invested in pension funds, instead of being
kept in the ?rm, as it was previously done. Before the reform, a quota approximately
equal to one-month pay was retained by the ?rm and paid to the worker at the moment
he/she would leave (voluntarily or not) his current employer[1]. Fromthe point of viewof
the ?rm, these funds were very similar to long-term corporate debt, although with two
major differences. The funds were being backed by a state insurance scheme in case of
insolvency of the ?rm, and by receiving a remuneration determined by the Italian law[2].
Since the rate of return paid on the TFR was lower than the risk-free rate of treasury
bond in most of the years (at least before the introduction of the Euro), the TFR ?ows
traditionally provided cheap ?nance for Italian ?rms.
The reform of July 2007 allows workers to transfer their future TFR ?ows in a
pension fund[3]. Thus, Italian SMEs will lose (part) of this cheap and liquid liability, in
a proportion related to the amount of workers who decide to invest their TFR outside
the ?rm. The objective of this paper is to determine the long-run effects of this reform
on the access and the cost of credit for small- and medium-sized ?rms.
There is a quite large empirical evidence that the debt capacity of Italian SMEs is
positively correlated with the dimension of the assets they can pledge as collateral.
Sapienza (1997) disentangles the determinants of the demand and the supply of credit and
concludes that loans supplyis signi?cantlycorrelatedwithassets tangibility. Guiso(2003)
shows that size is a major determinant of the probability of success in obtaining as much
bank ?nance as needed: ?rms with less than approximately 30 employees are twice less
likely to have ?nancial debt than bigger ?rms, and this probability is strongly and
positively correlated with the quota of tangible assets over total assets. From these
observations, he concludes that “?rms’ ability to pledge collateral strengthens their
capacity to borrow, in particular when bank-?rm relationships are not well established”.
These empirical ?ndings justify the use of the model of credit rationing proposed by
Holmstrom and Tirole (1997) (HT in the following) in order to study the impact of the
reform of severance payments on SMEs. Since it is well-documented that Italian SMEs
have almost no access to public debt, we use a modi?ed version of HT assuming that
the supply of capital to ?rms is due only to ?nancial intermediaries (in particular
banks) who typically perform an activity of costly monitoring ?rms’ decisions.
We show that in the long run the lower amount of liquid assets available to each
?rm makes the credit constraints for the smaller ?rms more severe, reducing
their total investment; in turn, this decreases their demand of external ?nance,
JFEP
3,3
244
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and the interest rate. The main intuition of this result is that since debt capacity is
increasing in the amount of liquid assets invested by the ?rm itself, reducing the latter
also reduces the former, so that, SMEs will be able to make less investments on
aggregate. As an extension of the HT framework, we consider the case in which banks
can optimally choose their monitoring level, and we conclude that in this extended
model version both the effects produced in the long run on the investment level and the
interest rate are exacerbated. Our theoretical framework allows us to quantify the
reduction in investment SMEs will suffer as a function of the predicted out?ow of
severance indemnity. Our main theoretical prediction is that in the long run the ?ow of
newinvestments will decrease by more than proportionally to the TFR out?ow. For our
empirical analysis, we have collected data from the Bank of Italy about the total credit
granted to Italian ?rms, divided by class of the total amount of credit each client has
received. Assuming that most of the SMEs receive a total annual credit which is less
than e125,000, we obtain an upper bound of their total bank borrowing. From Italian
National Statistical Of?ce (ISTAT) data showing the distribution of private sector
employees and earnings across ?rms’ size (de?ned by the number of employees), we
recover the annual ?ow of TFR in 2007. Based on assumptions on the macroeconomic
development of employment and wages, we then project TFR ?ows for the period
2008-2010 and assess the magnitude of TFR funds that Italian ?rms would lose as a
result of the reform.
Finally, we use data from Guiso (2003) to obtain the average leverage of SMEs
towards ?nancial institutions. Our simulations predict that the decrease of
SMEs investment due to the reform will be equal to 130-147 percent of the out?ow
of TFR in the long run. Our theoretical model also forecasts that in equilibrium the
loan rate will not increase in the long run; empirically, we forecast it may slightly
decrease – ceteris paribus – for an amount inversely proportional to the elasticity of
the supply of bank capital.
Our results suggest that in order to reduce the negative impact of the reform on
SMEs investment, the Italian Government should not subsidize the cost of future bank
loans to ?rms (as it is doing now), but instead should protect the access to credit of
?rms who are loosing a quota of their liquid funds.
The remainder of this paper is organized as follows. Section 2 presents the basic
theoretical framework for the analysis, which is a simpli?ed version of HT (1997),
while Section 3 discusses the effects of the policy change predicted by the model.
Section 4 presents the data and our estimates for the future out?ows of TFR, together
with a quantitative assessment of the main effects of the reform on the future
investment and the loan rate. Section 5 concludes.
2. The basic model
Our theoretical framework for the analysis of the effects of the reform of the Italian
severance payment system is a simpli?ed version of the model in HT (1997). In our
version of the model, there are two sets of agents: (small) ?rms and ?nancial
intermediaries. They operate on two periods: in the ?rst period a ?nancial contract
between each single ?rm and a competitive intermediary (a bank in the following) is
signed, and ?rms invest; in the second period the returns of investment are realized and
are distributed. Each ?rm has an initial amount of capital A
0
, representing the market
value of all assets that can be pledged as a collateral to the ?nancial contract
Severance
payment
reform
245
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
(a loan in the following). The distribution of assets across ?rms is described by the
c.d.f. F(A), and the aggregate amount of ?rm capital is denoted by K
f
¼
R
AdFðAÞ. All
banks are identical in any relevant respect for the analysis.
Each ?rm is endowed with the same investment projects: a good project G and a bad
project W. The ?rm can undertake any of them paying an initial amount I . A
0
, which
represents the scale of the project. The ?rm with own internal capital A
0
needs then to
borrow (at least) I 2 A
0
from the bank in order to undertake the investment at level I.
Investing I in the good project generates at t ¼ 2 a veri?able return, equal either to 0
(with probability 1 2 p
H
) or to R(I) (with probability p
H
), while the investment I in the
bad project pays 0 with probability 1 2 p
L
and R(I) with probability p
L
, with p
L
, p
H
.
In each ?rm, a risk-neutral entrepreneur selects his preferred project. Moral hazard
between the entrepreneur and the investor is formalized as in HT: for any unit of
investment in the bad project, the entrepreneur enjoys private bene?ts B. The choice of
the good project by the entrepreneur reduces his private bene?ts to zero.
Following HT, the gross rate of return on bank loans is denoted by b. On the public
capital markets, the risk-free rate is normalized to zero. As in HT, we assume that
b $ 1 in order to make economically pro?table for banks to invest in the SMEs’
projects. All individuals are risk neutral, and due to portfolio optimization, the
aggregate supply of credit to SMEs is (weakly) increasing with b: the higher b, the
more restricted is the set of alternative projects in the economy which provide at least
the same expected return as the ones undertaken by SMEs. Thus, the higher b, the
more capital banks will be willing to invest in SMEs’ projects.
Only the good project dominates the investment in public capital markets while the
bad project has a negative NPV:
p
H
RðI Þ 2I . 0 . p
L
RðI Þ þ BI 2I ð1Þ
Given equation (1), a necessary condition for the ?rm getting external ?nance is the
entrepreneur chooses the good project:
p
H
R
f
ðI Þ $ p
L
R
f
ðI Þ þ BI ð2Þ
where R
f
(I)[4] is the share of the project returns paid to the ?rm. Under equation (2), each
SMEchooses the amount of internal assets A # A
0
to invest, knowing that the expected
return for the bank is bounded by the quota of the proceeds that is paid to the ?rmR
f
, in
order to let the entrepreneur choose project G. Finally, we assume that the total return of
the project in case of success is linear in the initial investment, R(I) ¼ RI. Under these
assumptions, it is obvious that the ?rm will invest all its internal funds in the project,
A ¼ A
0
, but it will never borrow more than it needs to implement a given investment
level I. Indeed, the internal rate of return of an investment in the good project for a ?rmis
higher than b[5] and, for any ?xed I, the ?rm pays a gross return of b $ 1 on external
funds (while the opportunity cost of internal funds is equal to 1).
The decision problem of the risk-neutral entrepreneur is then simply given as:
I
max p
H
R
f
s:t: p
H
ðRI 2R
f
Þ ¼ bðI 2A
0
Þ ð3Þ
p
H
R
f
$ p
L
R
f
þ BI ð4Þ
JFEP
3,3
246
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where the ?rst constraint ensures that the bank earns a gross return of b on its
investment I 2 A
0
and the second constraint ensures that the entrepreneur chooses the
good project (incentive constraint).
Substituting for equation (3), we can rewrite the incentive constraint (4) as:
A
0
$ I 1 2
p
H
b
R 2
B
Dp
We assume that the expression within brackets is positive which is equivalent to
specify that the NPV of the good project at a cost of capital of b is positive. We can then
rewrite the problem as:
I
max p
H
RI 2bðI 2A
0
Þ
s:t: I #
A
0
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
Since we have assumed that the internal rate of return of a good project is higher than
b, ?rms invest the highest possible I given the initial liquid assets equal to A
0
:
I ¼
A
0
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
ð5Þ
A form of “investment multiplier” of internal cash is at work here: each unit invested
by the ?rm allows to attract more than one unit of external capital to invest in the good
project. Indeed, A
0
increases the total investment I by:
1
½1 2ð p
H
=bÞðR 2ðB=DpÞÞ? . 1
:
The ?nal payoff is split between R
f
, which goes to the ?rm, and R
b
, paid to the bank, in
such a way that it is possible to remunerate the bank capital exactly at the gross
discount rate b, and to let the ?rm earn a higher rate on its initial investment A
0
.
To ?nd the equilibrium on the market for intermediated capital, we solve for the rate
b
*
that equalizes demand and supply of bank capital. The aggregate demand of credit
is given by:
DðbÞ ¼
Z
ðI 2A
0
ÞdFðA
0
Þ ¼
1
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
21
Z
A
0
dFðA
0
Þ ð6Þ
while the supply of bank capital is exogenously given and equal to K
b
(b), where
K
0
b
¼ ›K
b
ðbÞ=›b . 0: The aggregate demand of intermediated capital is monotone
decreasing in b[6].
Let K
f
¼
R
A
0
dFðA
0
Þ be the aggregate internal funds of ?rms. The equilibrium on
the (intermediated) capital market is:
1
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
21
K
f
¼ K
b
ðbÞ
EðR
b
ÞðK
f
þ K
b
ðbÞÞ ¼ bK
b
ðbÞ ð7Þ
where we denote for simplicity p
H
ðR 2B=DpÞ ¼ EðR
b
Þ:
Severance
payment
reform
247
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The market clearing is obtained varying the return b (hence the aggregate supply of
capital K
b
(b) as well), while K
f
is ?xed. One can interpret the equilibrium condition as
follows: if b is “too” high, the supply of capital banks are willing to invest in SMEs is
higher than the demand; reducing b increases the maximum investment I a ?rm with
initial liquid assets A
0
can sustain, by equation (5), thus in turn increases the demand
D(b), and reduces the supply K
b
(b) until the equilibrium is reached. Notice ?nally that
the aggregate level of investment, K
f
þ K
b
ðbÞ; depends only on the aggregate level of
?rm capital K
f
, since K
b
(b) adjusts to guarantee the equilibrium on the market.
3. The effects of the reform: the model predictions
The reform approved by the Italian Parliament allows each employee in the private
sector to invest his future ?ows of severance indemnities (TFR), which the ?rm was
managing up to 2007 as its own liabilities, in pension fund schemes. Thus, if we assume
that the status quo is stationary in the sense that the in?ows of severance payments in
every period is equal to the out?ows, the reform is going to reduce the in?ow without
affecting the out?ow of TFR. Each ?rm will then suffer a reduction in the amount of
liquid assets. In terms of the model presented above, the most direct consequence of this
reformis a decline of A
0
(with respect to the case of no reform) fromJuly 2007 on. The key
question of interest is then how such a decrease in all ?rms’ internal funds affects the
equilibrium cost of capital, demand of credit and investment.
From equation (5), it is clear that a reduction of one unit in A
0
decreases I by more
than one unit. As a consequence, less bank capital can be attracted due to the binding
participation constraint of the bank:
I
b
¼ p
H
R 2
B
Dp
I
b
:
This in turn reduces the aggregate demand of intermediated capital. We can show
that as a consequence of this “multiplier” effect, the reduction of aggregate
investment at equilibrium is higher than the reduction of the aggregate amount of
liquid assets K
f
.
Proposition 1. If all ?rms’ internal liquid funds A
0
marginally decrease, causing a
decrease of the aggregate internal funds K
f
, then the cost of bank capital for the ?rms
will decrease as well as the aggregate investment in the ?rms (good) projects.
Moreover, the latter will reduce more than proportionally than K
f
:
db
dK
f
¼
EðR
b
Þ
K
*
b
þ ðb
*
2EðR
b
ÞÞK
0
b
ðb
*
Þ
$ 0 ð8Þ
dI
dK
f
. 1 ð9Þ
Proof. A decrease in each ?rm A
0
is going to reduce the ?rm investment, see
equation (5). However, the aggregate reduction in K
f
will also have an effect on the
equilibrium cost of capital b, by equation (7)[7]. Let us rewrite equation (5) as:
I ¼ ð1 þfðbÞÞA
0
ð10Þ
JFEP
3,3
248
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where:
fðbÞ ¼
ER
b
=b
1 2ðER
b
=bÞ
. 0 iff b . ER
b
:
Aggregating equation (10) across ?rms one obtains:
K
b
þ K
f
¼ ð1 þfðbÞÞK
f
K
b
¼ fðbÞK
f
Differentiating the equilibrium condition (7) w.r.to K
f
we obtain:
db
dK
f
K
b
ðbðK
f
ÞÞ þb
dK
b
dK
f
¼ ER
b
1 þ
dK
b
dK
f
›b
›K
f
K
b
ðbÞ þb
›K
b
›b
›b
›K
f
¼ ER
b
1 þ
›K
b
›b
›b
›K
f
›b
›K
f
K
b
þ
›K
b
›b
ðb 2ER
b
Þ
¼ ER
b
›b
›K
f
¼
ER
b
K
b
þ K
0
b
ðb 2ER
b
Þ
$ 0
We nowmove to compute explicitly the change inaggregate investment, 1 þ ðdK
b
=dK
f
Þ
due to the variation dK
f
. From K
b
¼ fðbðK
f
ÞÞK
f
one obtains:
dK
b
dK
f
¼
›K
b
›b
›b
›K
f
þ
›K
b
›K
f
¼ K
f
f
0
ðbÞ
›b
›K
f
þfðbÞ
where:
f
0
ðbÞ ¼ 2
ER
b
ðb 2ER
b
Þ
, 0:
Substituting for f
0
(b) and f(b) into the previous expression leads to:
dK
b
dK
f
¼ K
f
›b
›K
f
2
ER
b
ðb 2ER
b
Þ
2
þ
ER
b
=b
1 2ER
b
=b
thus :
dK
f
þ dK
b
dK
f
¼ K
f
1 2
›b
›K
f
ER
b
ðb 2ER
b
Þ
2
þ
ER
b
=b
1 2ER
b
=b
¼
ER
b
b 2ER
b
2
ER
b
ðb 2ER
b
Þ
2
K
f
›b
›K
f
þ K
f
¼ fðbÞ 2fðbÞK
f
1
b 2ER
b
›b
›K
f
þ K
f
Severance
payment
reform
249
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and substituting for ›b/›K
f
results in:
dK
f
þ dK
b
dK
f
¼ fðbÞ 1 2K
f
1
b 2ER
b
ER
b
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2
ER
b
b 2ER
b
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
2d
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
þdfðbÞ
þ K
f
¼ df
2
ðbÞ þ K
f
. K
f
where:
d ¼
K
f
K
b
2
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
. 0:
A
We can make three important observations. First, the proof of the proposition shows
our main difference with the theoretical ?ndings in HT. Their comparative statics
result for ›(K
b
þ K
f
)/›K
f
holds only if one considers ›b
*
=›K
f
¼ 0: In proposition 1,
we explicitly consider this indirect effect on the equilibrium loan rate.
Second, notice that at a ?rm level the same result as proposition 1 holds for the ?rm
investment only if all ?rms suffer the same proportional loss in A
0
. In our setup,
however, it is not crucial whether the reduction in A
0
is different across ?rms, since
every ?rm has the same production function. Thus, for the aggregate level of
investment what matters is only the total amount of TFR funds that is transferred from
the ?rms[8].
Finally, to estimate relations (8)-(9), we will use the fact that in the model
EðR
b
Þ=b
*
¼ I
b
=I , while the elasticity of the supply of capital to SMEs K
0
b
=ðK
*
b
=b
*
Þ
must be calibrated on real data.
We nowcheck whether the role of banks as ?rms monitors (Diamond, 1984) may alter
the predictions obtained before. For simplicity, we consider monitoring as essential for
receiving bank ?nance (because SMEs do not have access to uninformed ?nance in our
model), but in their role of delegated monitors, banks also suffer of a moral hazard
problem towards the depositors (Diamond, 1984; Chiesa, 2001). Thus, in our model,
contrarily to HT, the choice of monitoring by the bank is endogenous.
We restrict our analysis to an exclusive bank-?rm relationship[9] We ?rst show that
the monitoring intensity of banks increases with the ratio of own capital to the total
invested capital (as in Chiesa (2001), Carletti (2004), Carletti et al. (2005), among others).
To illustrate this point, let us write the pro?t function of a representative bank as:
P
bank
¼ E½max{
~
R
b
I 2r
D
D; 0}? 2
cI
2
m
2
ð11Þ
JFEP
3,3
250
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where:
~
R
b
¼
R 2
B
Dp
with proba: p
H
ðmÞ
0 with proba: 1 2p
H
ðmÞ
(
given that, to provide the entrepreneur with the right incentives, the maximum payoff
obtained by the bank is R 2 (B/Dp) for each unit invested. The probability of success
of the good project is increasing in the level of monitoring effort by the bank. We denote
with r
D
the rate of return paid on deposits, and with D the amount of deposits raised by
the bank. The cost of monitoring activity is linear in the dimension of the project
(the investment I) and convex in the monitoring intensity m[10].
The rationale of this formalization is that by performing a high-monitoring activity
the bank can improve the expected cash ?ow of the project (increasing the probability
of success).
Rewriting equation (11) gives:
P
bank
¼ p
H
ðmÞðR
b
I 2ðr
D
2SÞDÞ 2
cI
2
m
2
¼ p
H
ðmÞðR
b
ðA
0
þ E þ DÞ 2ðr
D
2SÞDÞ 2
cI
2
m
2
where S ¼ ð1 2p
H
ðmÞÞr
D
represents the per-unit expected shortfall on the deposit
contract, E is the bank own equity capital, and c is the unit monitoring cost for the
bank. Solving for the optimal monitoring intensity gives:
m
max
Y
bank
ð12Þ
m
*
¼
Dp
c
R
b
2r
D
D
I
ð13Þ
which is decreasing in D/I because of the moral hazard towards depositors. Since the
deposit rate is set before monitoring is decided, the bank always has an incentive to
increase her pro?t by increasing the expected shortfall, thus reducing monitoring
ex post. The higher the ratio of external capital, D/I, the lower the incentive to provide
monitoring.
The next step veri?es that proposition 1 holds also in this new setup. Turning to the
equilibrium condition (7), the lower monitoring activity affects the term EðR
b
ðmÞÞ that
is now equal to p
H
ðmÞR
b
2ðc=2Þm
2
: This term is increasing in m since by f.o.c. of
equation (12), DR
b
2cm
*
¼ 2ð›S=›mÞD . 0: Thus, a reduction on m, caused by the
lower A
0
, reduces in turn E(R
b
(m)) causing an even stronger decrease in K
b
(b) at
equilibrium[11]. The effect on aggregate investment described in proposition 1 is then
exacerbated when we allow banks to monitor the ?rms activity, choosing their optimal
monitoring intensity.
4. An estimate of the capital out?ows due to the reform
In order to make predictions about the change of the aggregate credit granted by banks
and about the loan rate obtained by SMEs, we have to quantitatively determine two
Severance
payment
reform
251
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
elements exogenous to our theoretical analysis. First, we have to predict the out?ow
of funds from SMEs due to the reform. Second, we need to assess the reaction of the
bank credit to SMEs following a collateral squeeze. Here, we concentrate on the ?rst
aspect.
The TFR consists in a fraction of gross earnings set aside by the employers[12] and
paid back to workers when their working relationship ends (for any reason, including
retirement). The amount set aside every year sums up to the already existing stock
which is capitalized at a given rate[13].
The severance pay reform, ?rst announced and approved in 2004 but modi?ed
afterwards within the Budget Law of 2007, does not concern the TFR stock already
accumulated within the ?rms, but only its future ?ows. More precisely, from July 2007
on workers have to decide whether to invest their current and future TFR ?ows in
private pension funds or not. If they do not make an explicit choice, their TFR ?ows
will be automatically diverted into pre-speci?ed pension funds, according to a
mechanism where no choice is equivalent to a tacit consent. In case they explicitly state
that they do not want to join any pension fund, their TFR will either be accumulated to
National Institute for Social Security Payments (INPS) (if they work in ?rms with
50 þ employees[14]), or it will remain at the ?rm’s disposal (for ?rms with less than
50 employees).
Even though the reform has been implemented in July 2007, data about its effects
are still scarce. The reform caused TFR ?ows amounting to e3.2 billion to be invested
in pension funds during 2007, which is 1.5 billion more than in 2006 (Covip, 2008b).
About 30 percent of private sector employees were members of a pension fund in 2007,
compared to about 25 percent in 2006 (Covip, 2008a).
A broader view on the reform is provided by an ad hoc survey carried out by
Eurisko for Anima Research Lab on more than 1,000 private sector employees who
were asked how they allocated their TFR ?ows (the data are described in greater detail
in Tito and Zingales (2008)). Table I reports some summary statistics about workers’
choices as of July 1, 2007, and allows to infer the actual destination of TFR ?ows.
Table I shows workers’ choices about their TFR (?rst panel) and how these choices
translated in terms of TFR allocation (second panel). The most preferred choice – both
,50 50 þ Total
“Raw” private employees’ choices
Pension fund 10.09 41.48 22.65
Firm 77.50 47.67 65.57
Did not choose 12.4 10.85 11.78
Total 100.0 100.0 100.0
TFR destinations
Pension fund 22.49 52.33 34.43
Firm 77.50 0.00 46.51
INPS 0.00 47.67 19.06
Total 100.0 100.0 100.0
Notes: Percentages in the bottom panel are computed recalling that not choosing is equivalent to
putting the TFR into pension funds, and that in ?rms with 50 þ employees TFR ?ows not going to
pension funds are diverted to INPS; all data weighted using FRDB (2008) weights
Source: Own elaboration on FRBD (2008)
Table I.
Workers’ choice and TFR
destination, by ?rm size
(percentages)
JFEP
3,3
252
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in small- and medium-large ?rms – is to leave the funds at the employer’s disposal.
Analogously, the percentage of workers who do not choose any speci?c investment for
their TFR is quite similar across ?rm size. However, a remarkable difference between
the two groups is given by those who actively decided: only 10 percent out of small
?rms’ workers decided to join a pension fund, while this percentage increases to 41.5 in
larger ?rms. Using other answers from the same survey, Tito and Zingales (2008)
suggest that workers trust more their own ?rm than INPS, so, when faced with the
choice between pension funds and INPS they tend to choose the former more often than
when they have to decide between pension funds and ?rm[15].
The idea that workers in ?rms with 50 þ employees chose pension funds more
often than workers in smaller ones is con?rmed by Covip data on the participation to
occupational funds: the membership rate of private sector employees is 12 percent in
?rms with less than 50 workers and 42 percent in medium-large ones (Covip, 2008b).
In order to get an insight on the out?ows of TFR from Italian SMEs, we ?rst focus
on the forecasts of the total yearly TFR, which represent the (maximum) total amount
of money that can be diverted from ?rms due to the reform[16]. Table II shows TFR
?ows in 2007, by ?rm size. Various ISTAT sources have been employed to assess
the magnitude of 2007 TFR[17]. As Table II shows, small ?rms make up for more than
half private sector employees. However, since earnings are on average higher in larger
?rms most TFR is accumulated in ?rms with more than 50 employees.
The amount of TFR amount shown in Table II is then used to compute TFR
destinations in 2007, applying the participation rates to the reform reported in Table I,
and assuming that all choices concerning TFR destination were taken as of July 1, 2007.
We compute the total amount of TFR?ows that were, respectively, invested into pension
funds, collected by INPS, or remained within the ?rms in 2007. The results of Table III
include howmuch TFRhas been “lost” by ?rms, that is the sumof TFRpaid to INPS and
to pension funds[18].
Afewremarks are needed regarding the results of Table III. First, the ?owto pension
funds and INPSin 2007 is relatively small because the reformdoes not apply to the whole
year. Therefore, ?rms will “lose” relativelymore funds inthe future. Second, most TFRis
lost by medium-large ?rms both because they accumulate most TFR and because they
lose this source of ?nancing entirely (either to INPS or to pension funds). On the contrary
SMEs (,50), lose a smaller amount (i.e. only the ?ows diverted to pension funds).
Although results in Table III are obtained under admittedly strong assumptions,
they are closely in line with aggregate data from other sources. First, Covip (2008b)
reports that the amount of TFR ?ows conveyed in pension funds in 2007 was equal to
e3.2 billion. The fact that we obtain a slightly higher value might be because some
workers are allowed to invest in pension funds only a fraction of their TFR. Second,
Private sector employees
(number)
Avg. gross earnings per employee
(e)
Wage bill
(e)
Total TFR
(e)
,50 6,014,429 18.737 112,691,294 7,786,968
50 þ 5,662,249 26.950 152,597,242 10,544,469
Total 11,676,678 22.528 265,288,537 18,331,438
Source: ISTAT (2007, 2008)
Table II.
Macro data on earnings
and TFR (2007)
Severance
payment
reform
253
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
according to ISTAT (2008), the total amount of TFR lost by ?rms in 2007 amounted
to e6.8 billion and the TFR ?ows going into ?rms[19] were about 13.9 billion in the
same year. Both ?gures are not too distant from what we ?nd (respectively 6.1 and
e12.2 billion).
Table IV shows projections of TFR ?ows that ?rms are going to “lose” in the future,
according to various hypotheses about macroeconomic scenarios and the participation
to the reform. We chose three growth rates for the total TFR (2.5, 3.5 and 4.5 percent)
re?ecting employment and wages growth[20] and two alternative scenarios for
workers’ participation in pension funds. In the ?rst, we assumed workers are going to
choose their future TFR destination with the same frequencies as of July 2007. In the
second, we assume an increasing participation to private pension funds (50 percent of
TFR in pension funds by 2010).
As we can see, TFR lost from 2008 on roughly doubles with respect to 2007 and – as
already noted – every year medium-large ?rms lose a remarkably higher amount than
small ones (which lose up to e4.4 billion in 2010) with respect to the case of no reform.
Moreover, the amount of TFR lost by ?rms is obviously increasing with higher TFR
growth rates and in the scenario where workers’ participation to pension funds
increases.
In the following, we derive some ?rst indications on the potential impact of TFR
out?ows on the loan rate paid by ?rms.
The main theoretical predictions of the model are contained in equations (8) and (9).
Thus, we will calibrate the main parameters of the model on real data.
The bank’s participation constraint reads as: I
b
¼ EðR
b
ÞI =b; that is
EðR
b
Þ ¼ bI
b
=I . We measure I
b
/I with the ratio between ?nancial debt over total
capital observed on the capital structure of ?rms. The reason for our choice is the
following: if we assume that the optimal capital structure is quite stable across time,
the leverage ratio we observe should correspond to the optimal ratio between the ?ow
of external ?nance and the total capital ?ow. Guiso (2003) measures this ratio
for Italian manufacturing ?rms from a survey of over 4,000 ?rms (mostly small
,50 50 þ Total
Second semester 2007
Pension funds 875,645 2,758,960 3,634,605
Firm 3,017,450 – 3,017,450
INPS – 2,513,274 2,513,274
Total 3,893,484 5,272,235 9,165,719
Total 2007
Pension funds 875,645 2,758,960 3,634,605
Firm 6,910,934 5,272,235 12,183,169
INPS – 2,513,274 2,513,274
Total 7,786,968 10,544,469 18,331,438
Total TFR “lost” by ?rms 875,645 5,272,235 6,147,879
Notes: To compute data in this table, we use TFR amounts from the last column of Table II – halved
to re?ect our assumption that choices are made as of July 1, 2007 – and apply to them the percentages
indicated in the second panel of Table I, indicating TFR destinations; TFR “lost” by ?rms equals the
sum of TFR paid to pension funds and INPS
Source: Own elaboration on ISTAT (2007, 2008) and FRDB (2008)
Table III.
TFR destination, by ?rm
size (e)
JFEP
3,3
254
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and medium sized) conducted in 1999 by Mediocredito Centrale, obtaining a median
value of 23.1 percent and an average value of 32.2 percent. We use these two statistics
as an estimate of I
b
/I.
Furthermore, we obtain a measure of the total bank credit to ?rms from the
“Bollettino Statistico” of the Bank of Italy: in this report, we can observe the total credit
granted frombanks (and other ?nancial institutions) to all clients, classi?ed by loan size.
We obtain from this document, for each class, both the number of credits and the total
amount lent by all banks and ?nancial institutions in Italy. We use this information to
estimate the total capital K
*
b
banks lend to ?rms at equilibrium before the reform.
Reasonably, small ?rms obtain on average credits of relatively small amounts, hence we
refer to the credit classes with less than e250,000. Since SMEs are probably not the only
clients obtaining such a loan (households mortgages for investments in real estate
represent for sure an important quota of these loans), we consider the numbers reported
by Bank of Italy as an upper bound of the credits granted to SMEs.
To compute the impact of the reform on the loan rate paid by SMEs who still have
access to bank credit pre- and post-reform, we need the equilibrium loan rate
pre-reform, and the elasticity of the banks supply of capital to the loan rate. We obtain
Same choices as in 2007 Increasing participation to PF
2007 2008 2009 2010 2007 2008 2009 2010
TFR growth: 2.5 percent
Pension funds 3.6 7.5 7.6 7.8 3.6 9.0 10.8 12.7
Firm 12.2 6.2 6.3 6.5 12.2 5.5 4.8 4.2
INPS 2.5 5.2 5.3 5.4 2.5 4.3 3.6 2.8
Total 18.3 18.8 19.3 19.7 18.3 18.8 19.3 19.7
Lost by ?rms, of which 6.1 12.6 12.9 13.2 6.1 13.3 14.4 15.5
, 50 0.9 1.8 1.8 1.9 0.9 2.5 3.3 4.2
50 þ 5.3 10.8 11.1 11.4 5.3 10.8 11.1 11.4
TFR growth: 3.5 percent
Pension funds 3.6 7.5 7.8 8.1 3.6 9.1 11.0 13.1
Firm 12.2 6.2 6.5 6.7 12.2 5.5 4.9 4.3
INPS 2.5 5.2 5.4 5.6 2.5 4.4 3.7 2.9
Total 18.3 19.0 19.6 20.3 18.3 19.0 19.6 20.3
Lost by ?rms, of which 6.1 12.7 13.2 13.6 6.1 13.5 14.7 16.0
, 50 0.9 1.8 1.9 1.9 0.9 2.6 3.4 4.3
50 þ 5.3 10.9 11.3 11.7 5.3 10.9 11.3 11.7
TFR growth: 4.5 percent
Pension funds 3.6 7.6 7.9 8.3 3.6 9.2 11.2 13.5
Firm 12.2 6.3 6.6 6.9 12.2 5.6 5.0 4.4
INPS 2.5 5.3 5.5 5.7 2.5 4.4 3.7 3.0
Total 18.3 19.2 20.0 20.9 18.3 19.2 20.0 20.9
Lost by ?rms, of which 6.1 12.8 13.4 14.0 6.1 13.6 15.0 16.5
, 50 0.9 1.8 1.9 2.0 0.9 2.6 3.5 4.4
50 þ 5.3 11.0 11.5 12.0 5.3 11.0 11.5 12.0
Notes: 2007 amounts are taken from Table III and are then projected according to a TFR growth
hypothesis (2.5, 3.5 and 4.5 percent) and a participation hypothesis; in the scenario “same choices as in
2007”, we assumed workers will choose their TFR destination with the same frequencies as in 2007; in
the scenario “increasing participation to PF”, we assume that 50 percent of TFR will be paid to pension
funds by 2010
Table IV.
TFR projection
2007-2010, by destination
(billion e)
Severance
payment
reform
255
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
information about the average loan rate currently paid by SMEs from the survey
by Capitalia (2005), indicating an average rate of 7.4 percent for the period 2001-2003
across all size classes. We use the lower and the upper bounds (5.4 and 8.2 percent) in
this same survey as two limit scenarios.
Finally, to estimate the elasticity of the loan supply with respect to the loan rate we
rely on Huelsewig et al. (2005) who use aggregate data to estimate the response of bank
loans to a monetary policy shock. As a robustness check, we also use the estimates in
King (1986), and we notice that the sensitivity of the change in the loan rate to this
elasticity parameter is extremely low. Finally, the impact of the reform on the future
investment is independent of this parameter (see equation (9)). Our main conclusions
are collected in Table V.
Table V indicates that the impact of the out?ow of TFR from the balance sheet of
SMEs on the loan rate is, in any of the scenarios, negligible (a maximum reduction of
far less than one basis point in any possible scenario). Indeed, the model forecasts that
the reduced amount of assets forces some ?rms to lose their access to credit, reducing
the aggregate demand of loans that banks can accept. The impact on the loan rate is
proportional to the out?ow of TFR, but in any case we see that this amount is too low
to produce any important macroeconomic effect on b.
However, according to equations (9) and (5), the reform will produce an important
reduction of SMEs investment. In particular, the forecasted reduction is more than
proportional to the out?ow of the TFR, and it is larger the larger the quota of
investment actually ?nanced by the banks (i.e. the higher the leverage ratio of SMEs,
according to our assumptions). We assess that this reduction can be about
130-147 percent of the future out?ow of TFR, as an effect of the “multiplier”
1=ð1 2I
b
=I Þ. To obtain this result we proxy I
b
/I with the amount of bank ?nancing on
total ?rms assets estimated by Guiso (2003) for Italian SMEs. The lower and the upper
bounds of our interval correspond, respectively, to Guiso’s highest and lowest
estimates of the ratio between bank ?nancement and total ?rm assets.
We conclude then that if the reform will have a negative impact on the SMEs, this
will be due to their reduced access to bank credit, and not to more expensive loans for
“Same as 2007”
“Increased
particip.”
Bank ?nancing
a
Interest rate
b
Loans supply elasticity
c
3.5% 4.5% 3.5% 4.5%
23.1 5.4 0.14 0.000090 0.000091 0.000127 0.000128
0.65 0.000066 0.000067 0.000093 0.000094
8.2 0.14 0.000137 0.000138 0.000192 0.000194
0.65 0.000101 0.000102 0.000142 0.000143
32.2 5.4 0.14 0.000127 0.000128 0.000178 0.000180
0.65 0.000096 0.000097 0.000136 0.000137
8.2 0.14 0.000193 0.000194 0.000271 0.000274
0.65 0.000146 0.000148 0.000206 0.000208
Notes: By rearranging equation (8), we obtain db=dK
f
¼ bðI
b
=I Þ=K
*
b
ð1 þ ðK
0
b
=K
*
b
Þbð1 2ðI
b
=I ÞÞÞ
where I
b
/I is “bank ?nancing”, b is “interest rate”, and K
0
b
=K
*
b
b is the “loans supply elasticity”; the
“same as 2007” and “increased participation” scenarios are those of Table IV
Source:
a
Guiso (2003),
b
Lower and upper bounds from Capitalia (2003),
c
Huelsewig et al. (2005) and
King (1986)
Table V.
Decrease in the loan rate
(percentage points) due to
the reform
JFEP
3,3
256
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
those ?rms who will continue to access bank credit. Some of the SMEs will be forced
to ask more ?nancing, and due to the risk of moral hazard, banks will reject their
requests; this in turn will decrease aggregate investment. A useful policy intervention
would then subsidize the access to credit for ?rms. Just subsidizing loans that are in
any case already granted by banks to some of the SMEs, as the government decided to
do in the 2007 law, is not going to reduce the inef?ciency.
5. Conclusions
The lack of external capital is often quoted as one of the main impediments for SMEs to
grow. In the presence of a moral hazard problem between the borrowing ?rm and
external ?nanciers, the debt capacity of the former depends on the amount of collateral
SMEs can pledge to the lender (Berger and Udell, 1995; HT, 1997), and the amount of
liquid assets they invest in the new projects. In our paper, we study the effects of the
2007 government reform of the system of severance indemnities currently in use for
Italian employees on the cost and the access to credit for Italian SMEs. The most direct
consequence of the reform is to reduce the amount of liquid assets of Italian ?rms, with
respect to the situation pre-reform. Some empirical literature (Guiso, 2003; Bianco,
1997; Sapienza, 1997) provides evidence that Italian ?rms are credit constrained if they
operate below a certain assets threshold, suggesting that the liquid net worth is one of
the main determinants for their debt capacity. The Italian severance payments reform
produces then a structural break on the dynamics of SMEs liquid assets which fully
satis?es the assumptions of the debt capacity theory mentioned above. We use the
model of HT (1997) to make predictions about the effects of such a reform on the
aggregate investments of SMEs, their access to credit, and the loan rate. In order to
gain insight on the reform effects, we performed an estimate of the future out?ows
from the severance indemnities fund. As the reform reaches its steady-state from 2008
on, the overall annual out?ow amounts to over e10 billion. However, most of the
out?ow affects medium-large ?rms, while SMEs suffer a relatively smaller loss.
Using the annual out?ows computed under different scenarios, we then provided
some estimates of the effects on the loan rate and investments according to our
theoretical model. First, the impact of the out?ow of TFR from the balance sheet of
SMEs on the cost of intermediated ?nance is likely to be, in any of the scenarios,
negligible. Indeed, the model forecasts that the reduced amount of assets forces some
?rms to lose their access to credit, reducing the aggregate demand of loans that banks
can accept, but this amount is too low to produce any important macroeconomic effect
on b. Second, we showed that the proposed reform will reduce in the long run the
aggregate investment by SMEs at a rate that is more than proportional to the out?ows
of TFR funds: this reduction is higher the higher the current leverage ratio of ?rms.
Notes
1. Every year, the employer must set aside in the TFR a quota equal to 1/13,5 (7.41 percent) of
the employee’s gross annual salary. Actually, only 6.91 percent is accumulated each year, as
0.5 percent is paid to INPS.
2. The TFR is capitalized annually at a rate equal to 1.5 percent plus three-fourth of the
in?ation rate measured in the previous year by the ISTAT.
3. More precisely, the law distinguishes between ?rms with less than 50 employees, and those
with 50 employees or more: for those working in the former who choose not to invest their
Severance
payment
reform
257
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
TFR in pension funds, it will remain within the ?rm; for those who work in larger ?rms, the
same choice will imply that the TFR indemnities are automatically deposited within INPS.
4. From now on, for easiness of notation, we will drop the functional form R
f
(I) to just R
f
.
5. Investing in a good project, a ?rm earns:
p
H
R
f
¼ p
H
RI 2bðI 2AÞ ¼ ð p
H
R 2bÞI þbA ¼ p
H
R 2b
À Á
I þbA
0
which is higher than what the ?rm would get investing all her liquid assets A
0
at rate b.
6. Denoting p
H
ðR 2ðB=DpÞÞ ¼ EðR
b
Þ we have:
›ðK
f
=ð1 2ðEðR
b
Þ=bÞÞÞðEðR
b
Þ=bÞ
›b
¼
EðR
b
ÞK
f
EðR
b
Þ 2b
1
b 2EðR
b
Þ
, 0:
7. The comparative statics result described in HT for ð›ðK
b
þ K
f
ÞÞ=›K
f
holds only if one
considers ›b
*
=›K
f
¼ 0:
8. When the technology is the same for all ?rms affected by the reform and it has constant
returns to scale, it is not important which ?rm actually loses most due to the exit of TFR
funds and which one is affected less. They all have, in a way, the “same” production function.
9. We assume here, that SMEs will not change the number of bank relationships following the
reform. Rajan (1992) and Petersen and Rajan (1995) show that a unique bank relationship
helps the access to credit, but it can be more expensive in the long run (a “hold up” problem).
However, Detragiache et al. (2000) ?nd that Italian ?rms typically maintain multiple
relations with banks. In principle, we cannot exclude then that, if the ?rms access to credit
will decrease in the future, some of the smallest ?rms will react by restricting themselves to a
unique relation with a bank.
10. We choose such a formalization for the costs of monitoring since without moral hazard
between bank and depositors the optimal monitoring intensity would be independent of I.
11. More formally, we can rewrite the aggregated equilibrium condition as:
bK
b
ðbÞ ¼ EðR
b
ðmÞÞðK
f
þ K
b
Þ
and differentiate by dK
f
. We obtain:
db
dK
f
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
À Á
¼ EðR
b
ðmÞÞ þ
›EðR
b
ðmÞÞ
›m
›m
›K
f
ðK
f
þ K
b
Þ
db
dK
f
¼
EðR
b
ðmÞÞ þ ðð›EðR
b
ðmÞÞÞ=›mÞð›m=›K
f
ÞðK
f
þ K
b
Þ
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
.
EðR
b
ðmÞÞ
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
for any given level of m, since monitor intensity m increases when total investment K
f
þ K
b
is higher (due to the fact that the ratio D/I reduces for higher I) and since
ð›EðR
b
ðmÞÞ=›mÞ . 0.
12. The aggregate ?ows accruing each year are calculated as:
FlowTFR
i;t
¼ 6:91%
*
W
tfrði;tÞ
where W
tfr(i,t)
is the aggregate gross wage received by employees.
JFEP
3,3
258
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
13. The TFR stock at the end of each year is equal to:
TFR
i;t
¼ ð1 þ r
t
ÞðTFR
i;t21
2TFRLIQ
i;t
Þ þ FlowTFR
i;t
where TFRLIQ
i,t
is the amount of TFR liquidated in year t due to the exit of the employee
from the ?rm, and r
t
is the rate at which TFR is capitalized annually, that is 1.5 percent plus
three-fourth of the in?ation rate annually measured by ISTAT.
14. That is medium and large ?rms according to the European classi?cation (European
Commission, 2003).
15. This happens even though TFR ?ows into INPS are going to receive exactly the same
treatment as within the ?rm in terms of capitalization and workers are not going to
experience any practical difference.
16. In doing this, we assume that the reform is not affecting the labor market conditions, i.e. is
not going to provoke a change in the structure of employment (between SMEs and large
?rms) nor is going to affect the duration of employment. Also, we do not try to estimate the
out?ows of TFR due to workers changing employer or leaving the labor market, which in
reality affect the availability of funds for each ?rm (Fugazza and Teppa, 2005). Our approach
to the labor market conditions is admittedly strong, since it reduces to an estimate of future
in?ows to the TFR based only on the total employment level and the average remuneration.
17. The number of private sector employees up to 2006, by ?rm size, is from ISTAT (2008). The
2007 ?gure has been obtained applying the growth rate of dependent employees as from
ISTAT (2008). The average gross earnings per dependent employee, by ?rm size, up to 2005
comes from ISTAT (2007). Growth rates up to 2007 are from ISTAT (2008). The wage bill
has been obtained by multiplying the number of dependent employees by their average
salary. Total TFR corresponds to 6.91 percent of the wage bill.
18. We assume that workers’ choices about TFR destination – as displayed in the second panel
of Table I – only apply to TFR ?ows in the second semester.
19. We have adjusted the ?gures in order to make them comparable with those of Table II, by
subtracting agricultural/?shery in accordance with the de?nition used in ISTAT (2008).
20. Overall TFR grew at an average rate of 3.6 percent in the 2000-2006 period (ISTAT, 2008).
References
Berger, A.N. and Udell, G.F. (1995), “Relationship lending and lines of credit in small ?rm
?nance”, Journal of Business, Vol. 68 No. 3, pp. 351-81.
Berger, A.N. and Udell, G.F. (1998), “The economics of small business ?nance: the roles of private
equity and debt markets in the ?nancial growth cycle”, Journal of Banking & Finance,
Vol. 22, pp. 613-73.
Bester, H. (1985), “Screening versus rationing in credit markets with imperfect information”,
American Economic Review, Vol. 75, pp. 850-5.
Bianco, M. (1997), “Vincoli ?nanziari e scelte reali delle imprese italiane: gli effetti di una
relazione stabile con una banca”, in Angeloni, I., Conti, V. and Passacantando, F. (Eds),
Le banche e il ?nanziamento delle imprese, Ente per gli Studi Monetari e Finanziari
“Luigi Einaudi”, Rome, pp. 33-60.
Capitalia (2005), “Indagine sulle imprese italiane – Rapporto sul sistema produttivo e sulla
politica industriale”, Osservatorio sulle piccole e medie imprese, Ottobre, Capitalia, Roma.
Carletti, E. (2004), “The structure of relationship lending, endogenous monitoring and loan rates”,
Journal of Financial Intermediation, Vol. 13, pp. 58-86.
Severance
payment
reform
259
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Carletti, E., Cerasi, V. and Daltung, S. (2005), “Multiple-bank lending: diversi?cation and
free-riding in monitoring”, mimeo, Center for Financial Studies, University of Frankfurt,
Frankfurt.
Chiesa, G. (2001), “Incentive-based lending capacity, competition and regulation in banking”,
Journal of Financial Intermediation, Vol. 10, pp. 28-53.
Covip (2008a), La previdenza complementare nel 2007. Dati provvisori, Covip, Roma.
Covip (2008b), Relazione per l’anno 2007, Covip, Roma.
Detragiache, E., Garella, P. and Guiso, L. (2000), “Multiple vs. single banking relationship: theory
and evidence”, Journal of Finance, Vol. 55, pp. 1133-61.
Diamond, D. (1984), “Financial intermediation and delegated monitoring”, Review of Economic
Studies, Vol. 51, pp. 393-414.
European Commission (2003), “Commission recommendation of 6 May 2003 concerning the
de?nition of micro, small and medium-sized enterprises (2003/361/EC)”, Of?cial Journal of
the European Union, May 20.
FRDB (2008), The TFR Survey, available at: www.frdb.org/scheda.php?id¼12&doc_pk¼11026
Fugazza, C. and Teppa, F. (2005), “An empirical assessment of the Italian severance payment”,
WP CeRP No. 38/05.
Guiso, L. (2003), “Small business ?nance in Italy”, EIB Papers, Vol. 7 No. 2.
Holmstrom, B. and Tirole, J. (1997), “Financial intermediation, loanable funds, and the real
sector”, Quarterly Journal of Economics, Vol. 112 No. 3, pp. 663-91.
Huelsewig, O., Mayer, E. and Wollmershaeuser, T. (2005), “Bank loan supply and monetary
policy transmission in Germany: an assessment based on matching impulses responses”,
CESInfo WP No. 1380, University of Munich, Munich.
ISTAT (2007), “Struttura e competitivita´ del sistema delle imprese industriali e dei servizi 2005”,
Statistiche in breve, available at: www.istat.it/salastampa/comunicati/non_calendario/
20071029_00/
ISTAT (2008), “Conti Economici Nazionali, 1970-2007”, available at: www.istat.it/dati/dataset/
20080328_00/
King, S. (1986), “Monetary transmission: through bank loans or bank liabilities?”, Journal of
Money, Credit and Banking, Vol. 18 No. 3, pp. 290-303.
Petersen, M. and Rajan, R. (1995), “The effect of credit market competition on lending
relationships”, Quarterly Journal of Economics, Vol. 110, pp. 407-43.
Rajan, R. (1992), “Insiders and outsiders: the choice between informed and arm’s-lenght debt”,
Journal of Finance, Vol. 47, pp. 1367-400.
Sapienza, P. (1997), “Le scelte di ?nanziamento delle imprese italiane”, in Angeloni, I., Conti, V.
and Passacantando, F. (Eds), Le banche e il ?nanziamento delle imprese, Ente per gli Studi
Monetari e Finanziari “Luigi Einaudi”, Rome, pp. 61-94.
Tito, B. and Zingales, L. (2008), “Chi ha paura dei fondi pensione?”, Anima FinLab, Numero 1,
available at: www.animasgr.it/ANIMA/IT/AnimaFinLab/
Further reading
Bardazzi, R. and Pazienza, M.G. (2006), “La riforma del TFR e il costo per le imprese minori:
un’analisi di microsimulazione”, Politica Economica, Vol. 22 No. 1, pp. 5-50.
Broecker, T. (1990), “Credit-worthiness and interbank competition”, Econometrica, Vol. 58 No. 2,
pp. 429-52.
JFEP
3,3
260
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Freixas, X. and Rochet, J.C. (1998), Microeconomics of Banking, MIT Press, Cambridge, MA.
Garibaldi, P. and Pacelli, L. (2004), “Do larger severance payments increase individual job
duration?”, CEPR Discussion Papers No. 4607.
Istituto di Studi e Analisi Economica (2005), Rapporto ISAE, Finanza pubblica e redistribuzione,
Istituto di Studi e Analisi Economica, Rome, Ottobre.
Palermo, G. and Valentini, M. (2000), “Il Fondo di Trattamento di Fine Rapporto e la struttura
?nanziaria delle imprese manifatturiere”, WP n. 12 02/2000 Banca di Roma.
Pammolli, F. and Salerno, N.C. (2006), “Le imprese e il ?nanziamento del pilastro previdenziale
private”, Nota CERM 02/2006.
Petersen, M. and Rajan, R. (1994), “The bene?t of lending relationships: evidence from small
business data”, Journal of Finance, Vol. 49, pp. 1367-400.
Corresponding author
Riccardo Calcagno can be contacted at: [email protected]
Severance
payment
reform
261
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_976841243.pdf
The purpose of this paper is to study the impact of the 2007 Italian severance payment
reform on the cost and the access to credit for small- and medium-sized enterprises (SMEs).
Journal of Financial Economic Policy
An analysis of the effects of the severance payment reform on credit to Italian SMEs
Riccardo Calcagno Roman Kraeussl Chiara Monticone
Article information:
To cite this document:
Riccardo Calcagno Roman Kraeussl Chiara Monticone, (2011),"An analysis of the effects of the severance
payment reform on credit to Italian SMEs", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 243 - 261
Permanent link to this document:
http://dx.doi.org/10.1108/17576381111152227
Downloaded on: 24 January 2016, At: 21:42 (PT)
References: this document contains references to 33 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 204 times since 2011*
Users who downloaded this article also downloaded:
Sameeksha Desai, J ohan Eklund, Andreas Högberg, (2011),"Pro-market reforms and allocation
of capital in India", J ournal of Financial Economic Policy, Vol. 3 Iss 2 pp. 123-139 http://
dx.doi.org/10.1108/17576381111133606
Ignacio Lozano, Karen Rodríguez, (2011),"Assessing the macroeconomic effects of fiscal
policy in Colombia", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 206-228 http://
dx.doi.org/10.1108/17576381111152209
Marco A. Espinosa-Vega, J uan Solé, (2011),"Cross-border financial surveillance: a
network perspective", J ournal of Financial Economic Policy, Vol. 3 Iss 3 pp. 182-205 http://
dx.doi.org/10.1108/17576381111152191
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
An analysis of the effects
of the severance payment reform
on credit to Italian SMEs
Riccardo Calcagno
Department of Economics, Finance and Control, EM Lyon Business School,
Ecully, France and
CeRP/Collegio Carlo Alberto, Torino, Italy
Roman Kraeussl
Department of Finance and Financial Sector Management,
VU University Amsterdam, Amsterdam, The Netherlands, and
Chiara Monticone
CeRP/Collegio Carlo Alberto, Torino, Italy
Abstract
Purpose – The purpose of this paper is to study the impact of the 2007 Italian severance payment
reform on the cost and the access to credit for small- and medium-sized enterprises (SMEs).
Design/methodology/approach – The authors study the implications of the reform adapting the
theoretical credit-rationing model of Holmstrom and Tirole, then estimate the capital out?ows due to
the reform and, using the theoretical prediction, assess its impact using mathematical simulations.
Findings – The authors predict that the reform may cause severe credit constraints to SMEs which
cannot pledge enough collateral in order to obtain credit. The most direct consequences are to reduce in
the long run the amount of liquid assets available to Italian ?rms, and to reduce their aggregate
investment in a more than proportional way, due to access to credit restrictions. However, it will not
increase the cost of intermediated ?nance, ceteris paribus.
Practical implications – The fact that the reform restricts access to credit, but does not increase the
cost of debt, has important policy consequences, as public interventions subsidizing credit through a
constant cost of debt may be ineffective.
Originality/value – While the topic has been analyzed in several respects (e.g. workers’
participation to the reform, cost of an access to credit subsidy, etc.), no other study proposed an
integrated view of these effects with a rigorous micro-economic approach.
Keywords Italy, Legislation, Severance payment, Moral hazard, Credit constraints,
Small to medium-sized enterprises
Paper type Research paper
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – G31, G32, G38
This research has been supported by the Center for Research on Pensions and Welfare Policies
(CeRP), at Collegio Carlo Alberto. The authors wishto thank Stefan Arping, Onorato Castellino, Elsa
Fornero, Giovanna Nicodano, as well as participants to seminars at Collegio Carlo Alberto, VU
University Amsterdam, 2007 Conference on Macroeconomics and International Finance at the
University of Crete for their useful comments. The authors owe many thanks to Fondazione Rodolfo
Debenedetti (FRDB) for providing access to the TFR survey, to CERIS-CNR for providing access to
AIDA, and to Simone Ceccarelli (Covip), Angelo Pace (Bank of Italy), Alessandro Cappellini,
Vilma Marchese and Martin Marchese (Intesa-SanPaolo) for their assistance inthe collection of data.
Severance
payment
reform
243
Journal of Financial Economic Policy
Vol. 3 No. 3, 2011
pp. 243-261
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576381111152227
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
1. Introduction
Berger and Udell (1998) observe that small, private businesses are “acutely
informationally opaque” and argue that this is a major impediment for them to
access the public capital markets. Small ?rms are forced to rely mostly on internal cash
resources and on intermediated ?nance, because intermediaries have some advantages
on direct lending through the debt markets. They can offer small- and medium-sized
enterprises (SMEs)-speci?c contracts that alleviate the acute adverse selection
problem (Bester, 1985) and they can monitor ?rms activity reducing moral hazard
(Diamond, 1984).
A recent reform of the system of severance indemnities in Italy is likely to affect the
amount of liquid assets of Italian SMEs, hence their main source of internal ?nance. This
law, included in the Financial Budget Law for 2007, allows employees to choose from
July 2007 on whether they want the future ?ows of their severance indemnity fund
(“Trattamento di Fine Rapporto”, TFR) to be invested in pension funds, instead of being
kept in the ?rm, as it was previously done. Before the reform, a quota approximately
equal to one-month pay was retained by the ?rm and paid to the worker at the moment
he/she would leave (voluntarily or not) his current employer[1]. Fromthe point of viewof
the ?rm, these funds were very similar to long-term corporate debt, although with two
major differences. The funds were being backed by a state insurance scheme in case of
insolvency of the ?rm, and by receiving a remuneration determined by the Italian law[2].
Since the rate of return paid on the TFR was lower than the risk-free rate of treasury
bond in most of the years (at least before the introduction of the Euro), the TFR ?ows
traditionally provided cheap ?nance for Italian ?rms.
The reform of July 2007 allows workers to transfer their future TFR ?ows in a
pension fund[3]. Thus, Italian SMEs will lose (part) of this cheap and liquid liability, in
a proportion related to the amount of workers who decide to invest their TFR outside
the ?rm. The objective of this paper is to determine the long-run effects of this reform
on the access and the cost of credit for small- and medium-sized ?rms.
There is a quite large empirical evidence that the debt capacity of Italian SMEs is
positively correlated with the dimension of the assets they can pledge as collateral.
Sapienza (1997) disentangles the determinants of the demand and the supply of credit and
concludes that loans supplyis signi?cantlycorrelatedwithassets tangibility. Guiso(2003)
shows that size is a major determinant of the probability of success in obtaining as much
bank ?nance as needed: ?rms with less than approximately 30 employees are twice less
likely to have ?nancial debt than bigger ?rms, and this probability is strongly and
positively correlated with the quota of tangible assets over total assets. From these
observations, he concludes that “?rms’ ability to pledge collateral strengthens their
capacity to borrow, in particular when bank-?rm relationships are not well established”.
These empirical ?ndings justify the use of the model of credit rationing proposed by
Holmstrom and Tirole (1997) (HT in the following) in order to study the impact of the
reform of severance payments on SMEs. Since it is well-documented that Italian SMEs
have almost no access to public debt, we use a modi?ed version of HT assuming that
the supply of capital to ?rms is due only to ?nancial intermediaries (in particular
banks) who typically perform an activity of costly monitoring ?rms’ decisions.
We show that in the long run the lower amount of liquid assets available to each
?rm makes the credit constraints for the smaller ?rms more severe, reducing
their total investment; in turn, this decreases their demand of external ?nance,
JFEP
3,3
244
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and the interest rate. The main intuition of this result is that since debt capacity is
increasing in the amount of liquid assets invested by the ?rm itself, reducing the latter
also reduces the former, so that, SMEs will be able to make less investments on
aggregate. As an extension of the HT framework, we consider the case in which banks
can optimally choose their monitoring level, and we conclude that in this extended
model version both the effects produced in the long run on the investment level and the
interest rate are exacerbated. Our theoretical framework allows us to quantify the
reduction in investment SMEs will suffer as a function of the predicted out?ow of
severance indemnity. Our main theoretical prediction is that in the long run the ?ow of
newinvestments will decrease by more than proportionally to the TFR out?ow. For our
empirical analysis, we have collected data from the Bank of Italy about the total credit
granted to Italian ?rms, divided by class of the total amount of credit each client has
received. Assuming that most of the SMEs receive a total annual credit which is less
than e125,000, we obtain an upper bound of their total bank borrowing. From Italian
National Statistical Of?ce (ISTAT) data showing the distribution of private sector
employees and earnings across ?rms’ size (de?ned by the number of employees), we
recover the annual ?ow of TFR in 2007. Based on assumptions on the macroeconomic
development of employment and wages, we then project TFR ?ows for the period
2008-2010 and assess the magnitude of TFR funds that Italian ?rms would lose as a
result of the reform.
Finally, we use data from Guiso (2003) to obtain the average leverage of SMEs
towards ?nancial institutions. Our simulations predict that the decrease of
SMEs investment due to the reform will be equal to 130-147 percent of the out?ow
of TFR in the long run. Our theoretical model also forecasts that in equilibrium the
loan rate will not increase in the long run; empirically, we forecast it may slightly
decrease – ceteris paribus – for an amount inversely proportional to the elasticity of
the supply of bank capital.
Our results suggest that in order to reduce the negative impact of the reform on
SMEs investment, the Italian Government should not subsidize the cost of future bank
loans to ?rms (as it is doing now), but instead should protect the access to credit of
?rms who are loosing a quota of their liquid funds.
The remainder of this paper is organized as follows. Section 2 presents the basic
theoretical framework for the analysis, which is a simpli?ed version of HT (1997),
while Section 3 discusses the effects of the policy change predicted by the model.
Section 4 presents the data and our estimates for the future out?ows of TFR, together
with a quantitative assessment of the main effects of the reform on the future
investment and the loan rate. Section 5 concludes.
2. The basic model
Our theoretical framework for the analysis of the effects of the reform of the Italian
severance payment system is a simpli?ed version of the model in HT (1997). In our
version of the model, there are two sets of agents: (small) ?rms and ?nancial
intermediaries. They operate on two periods: in the ?rst period a ?nancial contract
between each single ?rm and a competitive intermediary (a bank in the following) is
signed, and ?rms invest; in the second period the returns of investment are realized and
are distributed. Each ?rm has an initial amount of capital A
0
, representing the market
value of all assets that can be pledged as a collateral to the ?nancial contract
Severance
payment
reform
245
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
(a loan in the following). The distribution of assets across ?rms is described by the
c.d.f. F(A), and the aggregate amount of ?rm capital is denoted by K
f
¼
R
AdFðAÞ. All
banks are identical in any relevant respect for the analysis.
Each ?rm is endowed with the same investment projects: a good project G and a bad
project W. The ?rm can undertake any of them paying an initial amount I . A
0
, which
represents the scale of the project. The ?rm with own internal capital A
0
needs then to
borrow (at least) I 2 A
0
from the bank in order to undertake the investment at level I.
Investing I in the good project generates at t ¼ 2 a veri?able return, equal either to 0
(with probability 1 2 p
H
) or to R(I) (with probability p
H
), while the investment I in the
bad project pays 0 with probability 1 2 p
L
and R(I) with probability p
L
, with p
L
, p
H
.
In each ?rm, a risk-neutral entrepreneur selects his preferred project. Moral hazard
between the entrepreneur and the investor is formalized as in HT: for any unit of
investment in the bad project, the entrepreneur enjoys private bene?ts B. The choice of
the good project by the entrepreneur reduces his private bene?ts to zero.
Following HT, the gross rate of return on bank loans is denoted by b. On the public
capital markets, the risk-free rate is normalized to zero. As in HT, we assume that
b $ 1 in order to make economically pro?table for banks to invest in the SMEs’
projects. All individuals are risk neutral, and due to portfolio optimization, the
aggregate supply of credit to SMEs is (weakly) increasing with b: the higher b, the
more restricted is the set of alternative projects in the economy which provide at least
the same expected return as the ones undertaken by SMEs. Thus, the higher b, the
more capital banks will be willing to invest in SMEs’ projects.
Only the good project dominates the investment in public capital markets while the
bad project has a negative NPV:
p
H
RðI Þ 2I . 0 . p
L
RðI Þ þ BI 2I ð1Þ
Given equation (1), a necessary condition for the ?rm getting external ?nance is the
entrepreneur chooses the good project:
p
H
R
f
ðI Þ $ p
L
R
f
ðI Þ þ BI ð2Þ
where R
f
(I)[4] is the share of the project returns paid to the ?rm. Under equation (2), each
SMEchooses the amount of internal assets A # A
0
to invest, knowing that the expected
return for the bank is bounded by the quota of the proceeds that is paid to the ?rmR
f
, in
order to let the entrepreneur choose project G. Finally, we assume that the total return of
the project in case of success is linear in the initial investment, R(I) ¼ RI. Under these
assumptions, it is obvious that the ?rm will invest all its internal funds in the project,
A ¼ A
0
, but it will never borrow more than it needs to implement a given investment
level I. Indeed, the internal rate of return of an investment in the good project for a ?rmis
higher than b[5] and, for any ?xed I, the ?rm pays a gross return of b $ 1 on external
funds (while the opportunity cost of internal funds is equal to 1).
The decision problem of the risk-neutral entrepreneur is then simply given as:
I
max p
H
R
f
s:t: p
H
ðRI 2R
f
Þ ¼ bðI 2A
0
Þ ð3Þ
p
H
R
f
$ p
L
R
f
þ BI ð4Þ
JFEP
3,3
246
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where the ?rst constraint ensures that the bank earns a gross return of b on its
investment I 2 A
0
and the second constraint ensures that the entrepreneur chooses the
good project (incentive constraint).
Substituting for equation (3), we can rewrite the incentive constraint (4) as:
A
0
$ I 1 2
p
H
b
R 2
B
Dp
We assume that the expression within brackets is positive which is equivalent to
specify that the NPV of the good project at a cost of capital of b is positive. We can then
rewrite the problem as:
I
max p
H
RI 2bðI 2A
0
Þ
s:t: I #
A
0
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
Since we have assumed that the internal rate of return of a good project is higher than
b, ?rms invest the highest possible I given the initial liquid assets equal to A
0
:
I ¼
A
0
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
ð5Þ
A form of “investment multiplier” of internal cash is at work here: each unit invested
by the ?rm allows to attract more than one unit of external capital to invest in the good
project. Indeed, A
0
increases the total investment I by:
1
½1 2ð p
H
=bÞðR 2ðB=DpÞÞ? . 1
:
The ?nal payoff is split between R
f
, which goes to the ?rm, and R
b
, paid to the bank, in
such a way that it is possible to remunerate the bank capital exactly at the gross
discount rate b, and to let the ?rm earn a higher rate on its initial investment A
0
.
To ?nd the equilibrium on the market for intermediated capital, we solve for the rate
b
*
that equalizes demand and supply of bank capital. The aggregate demand of credit
is given by:
DðbÞ ¼
Z
ðI 2A
0
ÞdFðA
0
Þ ¼
1
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
21
Z
A
0
dFðA
0
Þ ð6Þ
while the supply of bank capital is exogenously given and equal to K
b
(b), where
K
0
b
¼ ›K
b
ðbÞ=›b . 0: The aggregate demand of intermediated capital is monotone
decreasing in b[6].
Let K
f
¼
R
A
0
dFðA
0
Þ be the aggregate internal funds of ?rms. The equilibrium on
the (intermediated) capital market is:
1
1 2ð p
H
=bÞðR 2ðB=DpÞÞ
21
K
f
¼ K
b
ðbÞ
EðR
b
ÞðK
f
þ K
b
ðbÞÞ ¼ bK
b
ðbÞ ð7Þ
where we denote for simplicity p
H
ðR 2B=DpÞ ¼ EðR
b
Þ:
Severance
payment
reform
247
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The market clearing is obtained varying the return b (hence the aggregate supply of
capital K
b
(b) as well), while K
f
is ?xed. One can interpret the equilibrium condition as
follows: if b is “too” high, the supply of capital banks are willing to invest in SMEs is
higher than the demand; reducing b increases the maximum investment I a ?rm with
initial liquid assets A
0
can sustain, by equation (5), thus in turn increases the demand
D(b), and reduces the supply K
b
(b) until the equilibrium is reached. Notice ?nally that
the aggregate level of investment, K
f
þ K
b
ðbÞ; depends only on the aggregate level of
?rm capital K
f
, since K
b
(b) adjusts to guarantee the equilibrium on the market.
3. The effects of the reform: the model predictions
The reform approved by the Italian Parliament allows each employee in the private
sector to invest his future ?ows of severance indemnities (TFR), which the ?rm was
managing up to 2007 as its own liabilities, in pension fund schemes. Thus, if we assume
that the status quo is stationary in the sense that the in?ows of severance payments in
every period is equal to the out?ows, the reform is going to reduce the in?ow without
affecting the out?ow of TFR. Each ?rm will then suffer a reduction in the amount of
liquid assets. In terms of the model presented above, the most direct consequence of this
reformis a decline of A
0
(with respect to the case of no reform) fromJuly 2007 on. The key
question of interest is then how such a decrease in all ?rms’ internal funds affects the
equilibrium cost of capital, demand of credit and investment.
From equation (5), it is clear that a reduction of one unit in A
0
decreases I by more
than one unit. As a consequence, less bank capital can be attracted due to the binding
participation constraint of the bank:
I
b
¼ p
H
R 2
B
Dp
I
b
:
This in turn reduces the aggregate demand of intermediated capital. We can show
that as a consequence of this “multiplier” effect, the reduction of aggregate
investment at equilibrium is higher than the reduction of the aggregate amount of
liquid assets K
f
.
Proposition 1. If all ?rms’ internal liquid funds A
0
marginally decrease, causing a
decrease of the aggregate internal funds K
f
, then the cost of bank capital for the ?rms
will decrease as well as the aggregate investment in the ?rms (good) projects.
Moreover, the latter will reduce more than proportionally than K
f
:
db
dK
f
¼
EðR
b
Þ
K
*
b
þ ðb
*
2EðR
b
ÞÞK
0
b
ðb
*
Þ
$ 0 ð8Þ
dI
dK
f
. 1 ð9Þ
Proof. A decrease in each ?rm A
0
is going to reduce the ?rm investment, see
equation (5). However, the aggregate reduction in K
f
will also have an effect on the
equilibrium cost of capital b, by equation (7)[7]. Let us rewrite equation (5) as:
I ¼ ð1 þfðbÞÞA
0
ð10Þ
JFEP
3,3
248
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where:
fðbÞ ¼
ER
b
=b
1 2ðER
b
=bÞ
. 0 iff b . ER
b
:
Aggregating equation (10) across ?rms one obtains:
K
b
þ K
f
¼ ð1 þfðbÞÞK
f
K
b
¼ fðbÞK
f
Differentiating the equilibrium condition (7) w.r.to K
f
we obtain:
db
dK
f
K
b
ðbðK
f
ÞÞ þb
dK
b
dK
f
¼ ER
b
1 þ
dK
b
dK
f
›b
›K
f
K
b
ðbÞ þb
›K
b
›b
›b
›K
f
¼ ER
b
1 þ
›K
b
›b
›b
›K
f
›b
›K
f
K
b
þ
›K
b
›b
ðb 2ER
b
Þ
¼ ER
b
›b
›K
f
¼
ER
b
K
b
þ K
0
b
ðb 2ER
b
Þ
$ 0
We nowmove to compute explicitly the change inaggregate investment, 1 þ ðdK
b
=dK
f
Þ
due to the variation dK
f
. From K
b
¼ fðbðK
f
ÞÞK
f
one obtains:
dK
b
dK
f
¼
›K
b
›b
›b
›K
f
þ
›K
b
›K
f
¼ K
f
f
0
ðbÞ
›b
›K
f
þfðbÞ
where:
f
0
ðbÞ ¼ 2
ER
b
ðb 2ER
b
Þ
, 0:
Substituting for f
0
(b) and f(b) into the previous expression leads to:
dK
b
dK
f
¼ K
f
›b
›K
f
2
ER
b
ðb 2ER
b
Þ
2
þ
ER
b
=b
1 2ER
b
=b
thus :
dK
f
þ dK
b
dK
f
¼ K
f
1 2
›b
›K
f
ER
b
ðb 2ER
b
Þ
2
þ
ER
b
=b
1 2ER
b
=b
¼
ER
b
b 2ER
b
2
ER
b
ðb 2ER
b
Þ
2
K
f
›b
›K
f
þ K
f
¼ fðbÞ 2fðbÞK
f
1
b 2ER
b
›b
›K
f
þ K
f
Severance
payment
reform
249
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and substituting for ›b/›K
f
results in:
dK
f
þ dK
b
dK
f
¼ fðbÞ 1 2K
f
1
b 2ER
b
ER
b
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2
ER
b
b 2ER
b
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
2d
þ K
f
¼ fðbÞ 1 2fðbÞ
K
f
K
b
þdfðbÞ
þ K
f
¼ df
2
ðbÞ þ K
f
. K
f
where:
d ¼
K
f
K
b
2
K
f
K
b
þ K
0
b
ðb 2ER
b
Þ
. 0:
A
We can make three important observations. First, the proof of the proposition shows
our main difference with the theoretical ?ndings in HT. Their comparative statics
result for ›(K
b
þ K
f
)/›K
f
holds only if one considers ›b
*
=›K
f
¼ 0: In proposition 1,
we explicitly consider this indirect effect on the equilibrium loan rate.
Second, notice that at a ?rm level the same result as proposition 1 holds for the ?rm
investment only if all ?rms suffer the same proportional loss in A
0
. In our setup,
however, it is not crucial whether the reduction in A
0
is different across ?rms, since
every ?rm has the same production function. Thus, for the aggregate level of
investment what matters is only the total amount of TFR funds that is transferred from
the ?rms[8].
Finally, to estimate relations (8)-(9), we will use the fact that in the model
EðR
b
Þ=b
*
¼ I
b
=I , while the elasticity of the supply of capital to SMEs K
0
b
=ðK
*
b
=b
*
Þ
must be calibrated on real data.
We nowcheck whether the role of banks as ?rms monitors (Diamond, 1984) may alter
the predictions obtained before. For simplicity, we consider monitoring as essential for
receiving bank ?nance (because SMEs do not have access to uninformed ?nance in our
model), but in their role of delegated monitors, banks also suffer of a moral hazard
problem towards the depositors (Diamond, 1984; Chiesa, 2001). Thus, in our model,
contrarily to HT, the choice of monitoring by the bank is endogenous.
We restrict our analysis to an exclusive bank-?rm relationship[9] We ?rst show that
the monitoring intensity of banks increases with the ratio of own capital to the total
invested capital (as in Chiesa (2001), Carletti (2004), Carletti et al. (2005), among others).
To illustrate this point, let us write the pro?t function of a representative bank as:
P
bank
¼ E½max{
~
R
b
I 2r
D
D; 0}? 2
cI
2
m
2
ð11Þ
JFEP
3,3
250
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
where:
~
R
b
¼
R 2
B
Dp
with proba: p
H
ðmÞ
0 with proba: 1 2p
H
ðmÞ
(
given that, to provide the entrepreneur with the right incentives, the maximum payoff
obtained by the bank is R 2 (B/Dp) for each unit invested. The probability of success
of the good project is increasing in the level of monitoring effort by the bank. We denote
with r
D
the rate of return paid on deposits, and with D the amount of deposits raised by
the bank. The cost of monitoring activity is linear in the dimension of the project
(the investment I) and convex in the monitoring intensity m[10].
The rationale of this formalization is that by performing a high-monitoring activity
the bank can improve the expected cash ?ow of the project (increasing the probability
of success).
Rewriting equation (11) gives:
P
bank
¼ p
H
ðmÞðR
b
I 2ðr
D
2SÞDÞ 2
cI
2
m
2
¼ p
H
ðmÞðR
b
ðA
0
þ E þ DÞ 2ðr
D
2SÞDÞ 2
cI
2
m
2
where S ¼ ð1 2p
H
ðmÞÞr
D
represents the per-unit expected shortfall on the deposit
contract, E is the bank own equity capital, and c is the unit monitoring cost for the
bank. Solving for the optimal monitoring intensity gives:
m
max
Y
bank
ð12Þ
m
*
¼
Dp
c
R
b
2r
D
D
I
ð13Þ
which is decreasing in D/I because of the moral hazard towards depositors. Since the
deposit rate is set before monitoring is decided, the bank always has an incentive to
increase her pro?t by increasing the expected shortfall, thus reducing monitoring
ex post. The higher the ratio of external capital, D/I, the lower the incentive to provide
monitoring.
The next step veri?es that proposition 1 holds also in this new setup. Turning to the
equilibrium condition (7), the lower monitoring activity affects the term EðR
b
ðmÞÞ that
is now equal to p
H
ðmÞR
b
2ðc=2Þm
2
: This term is increasing in m since by f.o.c. of
equation (12), DR
b
2cm
*
¼ 2ð›S=›mÞD . 0: Thus, a reduction on m, caused by the
lower A
0
, reduces in turn E(R
b
(m)) causing an even stronger decrease in K
b
(b) at
equilibrium[11]. The effect on aggregate investment described in proposition 1 is then
exacerbated when we allow banks to monitor the ?rms activity, choosing their optimal
monitoring intensity.
4. An estimate of the capital out?ows due to the reform
In order to make predictions about the change of the aggregate credit granted by banks
and about the loan rate obtained by SMEs, we have to quantitatively determine two
Severance
payment
reform
251
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
elements exogenous to our theoretical analysis. First, we have to predict the out?ow
of funds from SMEs due to the reform. Second, we need to assess the reaction of the
bank credit to SMEs following a collateral squeeze. Here, we concentrate on the ?rst
aspect.
The TFR consists in a fraction of gross earnings set aside by the employers[12] and
paid back to workers when their working relationship ends (for any reason, including
retirement). The amount set aside every year sums up to the already existing stock
which is capitalized at a given rate[13].
The severance pay reform, ?rst announced and approved in 2004 but modi?ed
afterwards within the Budget Law of 2007, does not concern the TFR stock already
accumulated within the ?rms, but only its future ?ows. More precisely, from July 2007
on workers have to decide whether to invest their current and future TFR ?ows in
private pension funds or not. If they do not make an explicit choice, their TFR ?ows
will be automatically diverted into pre-speci?ed pension funds, according to a
mechanism where no choice is equivalent to a tacit consent. In case they explicitly state
that they do not want to join any pension fund, their TFR will either be accumulated to
National Institute for Social Security Payments (INPS) (if they work in ?rms with
50 þ employees[14]), or it will remain at the ?rm’s disposal (for ?rms with less than
50 employees).
Even though the reform has been implemented in July 2007, data about its effects
are still scarce. The reform caused TFR ?ows amounting to e3.2 billion to be invested
in pension funds during 2007, which is 1.5 billion more than in 2006 (Covip, 2008b).
About 30 percent of private sector employees were members of a pension fund in 2007,
compared to about 25 percent in 2006 (Covip, 2008a).
A broader view on the reform is provided by an ad hoc survey carried out by
Eurisko for Anima Research Lab on more than 1,000 private sector employees who
were asked how they allocated their TFR ?ows (the data are described in greater detail
in Tito and Zingales (2008)). Table I reports some summary statistics about workers’
choices as of July 1, 2007, and allows to infer the actual destination of TFR ?ows.
Table I shows workers’ choices about their TFR (?rst panel) and how these choices
translated in terms of TFR allocation (second panel). The most preferred choice – both
,50 50 þ Total
“Raw” private employees’ choices
Pension fund 10.09 41.48 22.65
Firm 77.50 47.67 65.57
Did not choose 12.4 10.85 11.78
Total 100.0 100.0 100.0
TFR destinations
Pension fund 22.49 52.33 34.43
Firm 77.50 0.00 46.51
INPS 0.00 47.67 19.06
Total 100.0 100.0 100.0
Notes: Percentages in the bottom panel are computed recalling that not choosing is equivalent to
putting the TFR into pension funds, and that in ?rms with 50 þ employees TFR ?ows not going to
pension funds are diverted to INPS; all data weighted using FRDB (2008) weights
Source: Own elaboration on FRBD (2008)
Table I.
Workers’ choice and TFR
destination, by ?rm size
(percentages)
JFEP
3,3
252
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
in small- and medium-large ?rms – is to leave the funds at the employer’s disposal.
Analogously, the percentage of workers who do not choose any speci?c investment for
their TFR is quite similar across ?rm size. However, a remarkable difference between
the two groups is given by those who actively decided: only 10 percent out of small
?rms’ workers decided to join a pension fund, while this percentage increases to 41.5 in
larger ?rms. Using other answers from the same survey, Tito and Zingales (2008)
suggest that workers trust more their own ?rm than INPS, so, when faced with the
choice between pension funds and INPS they tend to choose the former more often than
when they have to decide between pension funds and ?rm[15].
The idea that workers in ?rms with 50 þ employees chose pension funds more
often than workers in smaller ones is con?rmed by Covip data on the participation to
occupational funds: the membership rate of private sector employees is 12 percent in
?rms with less than 50 workers and 42 percent in medium-large ones (Covip, 2008b).
In order to get an insight on the out?ows of TFR from Italian SMEs, we ?rst focus
on the forecasts of the total yearly TFR, which represent the (maximum) total amount
of money that can be diverted from ?rms due to the reform[16]. Table II shows TFR
?ows in 2007, by ?rm size. Various ISTAT sources have been employed to assess
the magnitude of 2007 TFR[17]. As Table II shows, small ?rms make up for more than
half private sector employees. However, since earnings are on average higher in larger
?rms most TFR is accumulated in ?rms with more than 50 employees.
The amount of TFR amount shown in Table II is then used to compute TFR
destinations in 2007, applying the participation rates to the reform reported in Table I,
and assuming that all choices concerning TFR destination were taken as of July 1, 2007.
We compute the total amount of TFR?ows that were, respectively, invested into pension
funds, collected by INPS, or remained within the ?rms in 2007. The results of Table III
include howmuch TFRhas been “lost” by ?rms, that is the sumof TFRpaid to INPS and
to pension funds[18].
Afewremarks are needed regarding the results of Table III. First, the ?owto pension
funds and INPSin 2007 is relatively small because the reformdoes not apply to the whole
year. Therefore, ?rms will “lose” relativelymore funds inthe future. Second, most TFRis
lost by medium-large ?rms both because they accumulate most TFR and because they
lose this source of ?nancing entirely (either to INPS or to pension funds). On the contrary
SMEs (,50), lose a smaller amount (i.e. only the ?ows diverted to pension funds).
Although results in Table III are obtained under admittedly strong assumptions,
they are closely in line with aggregate data from other sources. First, Covip (2008b)
reports that the amount of TFR ?ows conveyed in pension funds in 2007 was equal to
e3.2 billion. The fact that we obtain a slightly higher value might be because some
workers are allowed to invest in pension funds only a fraction of their TFR. Second,
Private sector employees
(number)
Avg. gross earnings per employee
(e)
Wage bill
(e)
Total TFR
(e)
,50 6,014,429 18.737 112,691,294 7,786,968
50 þ 5,662,249 26.950 152,597,242 10,544,469
Total 11,676,678 22.528 265,288,537 18,331,438
Source: ISTAT (2007, 2008)
Table II.
Macro data on earnings
and TFR (2007)
Severance
payment
reform
253
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
according to ISTAT (2008), the total amount of TFR lost by ?rms in 2007 amounted
to e6.8 billion and the TFR ?ows going into ?rms[19] were about 13.9 billion in the
same year. Both ?gures are not too distant from what we ?nd (respectively 6.1 and
e12.2 billion).
Table IV shows projections of TFR ?ows that ?rms are going to “lose” in the future,
according to various hypotheses about macroeconomic scenarios and the participation
to the reform. We chose three growth rates for the total TFR (2.5, 3.5 and 4.5 percent)
re?ecting employment and wages growth[20] and two alternative scenarios for
workers’ participation in pension funds. In the ?rst, we assumed workers are going to
choose their future TFR destination with the same frequencies as of July 2007. In the
second, we assume an increasing participation to private pension funds (50 percent of
TFR in pension funds by 2010).
As we can see, TFR lost from 2008 on roughly doubles with respect to 2007 and – as
already noted – every year medium-large ?rms lose a remarkably higher amount than
small ones (which lose up to e4.4 billion in 2010) with respect to the case of no reform.
Moreover, the amount of TFR lost by ?rms is obviously increasing with higher TFR
growth rates and in the scenario where workers’ participation to pension funds
increases.
In the following, we derive some ?rst indications on the potential impact of TFR
out?ows on the loan rate paid by ?rms.
The main theoretical predictions of the model are contained in equations (8) and (9).
Thus, we will calibrate the main parameters of the model on real data.
The bank’s participation constraint reads as: I
b
¼ EðR
b
ÞI =b; that is
EðR
b
Þ ¼ bI
b
=I . We measure I
b
/I with the ratio between ?nancial debt over total
capital observed on the capital structure of ?rms. The reason for our choice is the
following: if we assume that the optimal capital structure is quite stable across time,
the leverage ratio we observe should correspond to the optimal ratio between the ?ow
of external ?nance and the total capital ?ow. Guiso (2003) measures this ratio
for Italian manufacturing ?rms from a survey of over 4,000 ?rms (mostly small
,50 50 þ Total
Second semester 2007
Pension funds 875,645 2,758,960 3,634,605
Firm 3,017,450 – 3,017,450
INPS – 2,513,274 2,513,274
Total 3,893,484 5,272,235 9,165,719
Total 2007
Pension funds 875,645 2,758,960 3,634,605
Firm 6,910,934 5,272,235 12,183,169
INPS – 2,513,274 2,513,274
Total 7,786,968 10,544,469 18,331,438
Total TFR “lost” by ?rms 875,645 5,272,235 6,147,879
Notes: To compute data in this table, we use TFR amounts from the last column of Table II – halved
to re?ect our assumption that choices are made as of July 1, 2007 – and apply to them the percentages
indicated in the second panel of Table I, indicating TFR destinations; TFR “lost” by ?rms equals the
sum of TFR paid to pension funds and INPS
Source: Own elaboration on ISTAT (2007, 2008) and FRDB (2008)
Table III.
TFR destination, by ?rm
size (e)
JFEP
3,3
254
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
and medium sized) conducted in 1999 by Mediocredito Centrale, obtaining a median
value of 23.1 percent and an average value of 32.2 percent. We use these two statistics
as an estimate of I
b
/I.
Furthermore, we obtain a measure of the total bank credit to ?rms from the
“Bollettino Statistico” of the Bank of Italy: in this report, we can observe the total credit
granted frombanks (and other ?nancial institutions) to all clients, classi?ed by loan size.
We obtain from this document, for each class, both the number of credits and the total
amount lent by all banks and ?nancial institutions in Italy. We use this information to
estimate the total capital K
*
b
banks lend to ?rms at equilibrium before the reform.
Reasonably, small ?rms obtain on average credits of relatively small amounts, hence we
refer to the credit classes with less than e250,000. Since SMEs are probably not the only
clients obtaining such a loan (households mortgages for investments in real estate
represent for sure an important quota of these loans), we consider the numbers reported
by Bank of Italy as an upper bound of the credits granted to SMEs.
To compute the impact of the reform on the loan rate paid by SMEs who still have
access to bank credit pre- and post-reform, we need the equilibrium loan rate
pre-reform, and the elasticity of the banks supply of capital to the loan rate. We obtain
Same choices as in 2007 Increasing participation to PF
2007 2008 2009 2010 2007 2008 2009 2010
TFR growth: 2.5 percent
Pension funds 3.6 7.5 7.6 7.8 3.6 9.0 10.8 12.7
Firm 12.2 6.2 6.3 6.5 12.2 5.5 4.8 4.2
INPS 2.5 5.2 5.3 5.4 2.5 4.3 3.6 2.8
Total 18.3 18.8 19.3 19.7 18.3 18.8 19.3 19.7
Lost by ?rms, of which 6.1 12.6 12.9 13.2 6.1 13.3 14.4 15.5
, 50 0.9 1.8 1.8 1.9 0.9 2.5 3.3 4.2
50 þ 5.3 10.8 11.1 11.4 5.3 10.8 11.1 11.4
TFR growth: 3.5 percent
Pension funds 3.6 7.5 7.8 8.1 3.6 9.1 11.0 13.1
Firm 12.2 6.2 6.5 6.7 12.2 5.5 4.9 4.3
INPS 2.5 5.2 5.4 5.6 2.5 4.4 3.7 2.9
Total 18.3 19.0 19.6 20.3 18.3 19.0 19.6 20.3
Lost by ?rms, of which 6.1 12.7 13.2 13.6 6.1 13.5 14.7 16.0
, 50 0.9 1.8 1.9 1.9 0.9 2.6 3.4 4.3
50 þ 5.3 10.9 11.3 11.7 5.3 10.9 11.3 11.7
TFR growth: 4.5 percent
Pension funds 3.6 7.6 7.9 8.3 3.6 9.2 11.2 13.5
Firm 12.2 6.3 6.6 6.9 12.2 5.6 5.0 4.4
INPS 2.5 5.3 5.5 5.7 2.5 4.4 3.7 3.0
Total 18.3 19.2 20.0 20.9 18.3 19.2 20.0 20.9
Lost by ?rms, of which 6.1 12.8 13.4 14.0 6.1 13.6 15.0 16.5
, 50 0.9 1.8 1.9 2.0 0.9 2.6 3.5 4.4
50 þ 5.3 11.0 11.5 12.0 5.3 11.0 11.5 12.0
Notes: 2007 amounts are taken from Table III and are then projected according to a TFR growth
hypothesis (2.5, 3.5 and 4.5 percent) and a participation hypothesis; in the scenario “same choices as in
2007”, we assumed workers will choose their TFR destination with the same frequencies as in 2007; in
the scenario “increasing participation to PF”, we assume that 50 percent of TFR will be paid to pension
funds by 2010
Table IV.
TFR projection
2007-2010, by destination
(billion e)
Severance
payment
reform
255
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
information about the average loan rate currently paid by SMEs from the survey
by Capitalia (2005), indicating an average rate of 7.4 percent for the period 2001-2003
across all size classes. We use the lower and the upper bounds (5.4 and 8.2 percent) in
this same survey as two limit scenarios.
Finally, to estimate the elasticity of the loan supply with respect to the loan rate we
rely on Huelsewig et al. (2005) who use aggregate data to estimate the response of bank
loans to a monetary policy shock. As a robustness check, we also use the estimates in
King (1986), and we notice that the sensitivity of the change in the loan rate to this
elasticity parameter is extremely low. Finally, the impact of the reform on the future
investment is independent of this parameter (see equation (9)). Our main conclusions
are collected in Table V.
Table V indicates that the impact of the out?ow of TFR from the balance sheet of
SMEs on the loan rate is, in any of the scenarios, negligible (a maximum reduction of
far less than one basis point in any possible scenario). Indeed, the model forecasts that
the reduced amount of assets forces some ?rms to lose their access to credit, reducing
the aggregate demand of loans that banks can accept. The impact on the loan rate is
proportional to the out?ow of TFR, but in any case we see that this amount is too low
to produce any important macroeconomic effect on b.
However, according to equations (9) and (5), the reform will produce an important
reduction of SMEs investment. In particular, the forecasted reduction is more than
proportional to the out?ow of the TFR, and it is larger the larger the quota of
investment actually ?nanced by the banks (i.e. the higher the leverage ratio of SMEs,
according to our assumptions). We assess that this reduction can be about
130-147 percent of the future out?ow of TFR, as an effect of the “multiplier”
1=ð1 2I
b
=I Þ. To obtain this result we proxy I
b
/I with the amount of bank ?nancing on
total ?rms assets estimated by Guiso (2003) for Italian SMEs. The lower and the upper
bounds of our interval correspond, respectively, to Guiso’s highest and lowest
estimates of the ratio between bank ?nancement and total ?rm assets.
We conclude then that if the reform will have a negative impact on the SMEs, this
will be due to their reduced access to bank credit, and not to more expensive loans for
“Same as 2007”
“Increased
particip.”
Bank ?nancing
a
Interest rate
b
Loans supply elasticity
c
3.5% 4.5% 3.5% 4.5%
23.1 5.4 0.14 0.000090 0.000091 0.000127 0.000128
0.65 0.000066 0.000067 0.000093 0.000094
8.2 0.14 0.000137 0.000138 0.000192 0.000194
0.65 0.000101 0.000102 0.000142 0.000143
32.2 5.4 0.14 0.000127 0.000128 0.000178 0.000180
0.65 0.000096 0.000097 0.000136 0.000137
8.2 0.14 0.000193 0.000194 0.000271 0.000274
0.65 0.000146 0.000148 0.000206 0.000208
Notes: By rearranging equation (8), we obtain db=dK
f
¼ bðI
b
=I Þ=K
*
b
ð1 þ ðK
0
b
=K
*
b
Þbð1 2ðI
b
=I ÞÞÞ
where I
b
/I is “bank ?nancing”, b is “interest rate”, and K
0
b
=K
*
b
b is the “loans supply elasticity”; the
“same as 2007” and “increased participation” scenarios are those of Table IV
Source:
a
Guiso (2003),
b
Lower and upper bounds from Capitalia (2003),
c
Huelsewig et al. (2005) and
King (1986)
Table V.
Decrease in the loan rate
(percentage points) due to
the reform
JFEP
3,3
256
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
those ?rms who will continue to access bank credit. Some of the SMEs will be forced
to ask more ?nancing, and due to the risk of moral hazard, banks will reject their
requests; this in turn will decrease aggregate investment. A useful policy intervention
would then subsidize the access to credit for ?rms. Just subsidizing loans that are in
any case already granted by banks to some of the SMEs, as the government decided to
do in the 2007 law, is not going to reduce the inef?ciency.
5. Conclusions
The lack of external capital is often quoted as one of the main impediments for SMEs to
grow. In the presence of a moral hazard problem between the borrowing ?rm and
external ?nanciers, the debt capacity of the former depends on the amount of collateral
SMEs can pledge to the lender (Berger and Udell, 1995; HT, 1997), and the amount of
liquid assets they invest in the new projects. In our paper, we study the effects of the
2007 government reform of the system of severance indemnities currently in use for
Italian employees on the cost and the access to credit for Italian SMEs. The most direct
consequence of the reform is to reduce the amount of liquid assets of Italian ?rms, with
respect to the situation pre-reform. Some empirical literature (Guiso, 2003; Bianco,
1997; Sapienza, 1997) provides evidence that Italian ?rms are credit constrained if they
operate below a certain assets threshold, suggesting that the liquid net worth is one of
the main determinants for their debt capacity. The Italian severance payments reform
produces then a structural break on the dynamics of SMEs liquid assets which fully
satis?es the assumptions of the debt capacity theory mentioned above. We use the
model of HT (1997) to make predictions about the effects of such a reform on the
aggregate investments of SMEs, their access to credit, and the loan rate. In order to
gain insight on the reform effects, we performed an estimate of the future out?ows
from the severance indemnities fund. As the reform reaches its steady-state from 2008
on, the overall annual out?ow amounts to over e10 billion. However, most of the
out?ow affects medium-large ?rms, while SMEs suffer a relatively smaller loss.
Using the annual out?ows computed under different scenarios, we then provided
some estimates of the effects on the loan rate and investments according to our
theoretical model. First, the impact of the out?ow of TFR from the balance sheet of
SMEs on the cost of intermediated ?nance is likely to be, in any of the scenarios,
negligible. Indeed, the model forecasts that the reduced amount of assets forces some
?rms to lose their access to credit, reducing the aggregate demand of loans that banks
can accept, but this amount is too low to produce any important macroeconomic effect
on b. Second, we showed that the proposed reform will reduce in the long run the
aggregate investment by SMEs at a rate that is more than proportional to the out?ows
of TFR funds: this reduction is higher the higher the current leverage ratio of ?rms.
Notes
1. Every year, the employer must set aside in the TFR a quota equal to 1/13,5 (7.41 percent) of
the employee’s gross annual salary. Actually, only 6.91 percent is accumulated each year, as
0.5 percent is paid to INPS.
2. The TFR is capitalized annually at a rate equal to 1.5 percent plus three-fourth of the
in?ation rate measured in the previous year by the ISTAT.
3. More precisely, the law distinguishes between ?rms with less than 50 employees, and those
with 50 employees or more: for those working in the former who choose not to invest their
Severance
payment
reform
257
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
TFR in pension funds, it will remain within the ?rm; for those who work in larger ?rms, the
same choice will imply that the TFR indemnities are automatically deposited within INPS.
4. From now on, for easiness of notation, we will drop the functional form R
f
(I) to just R
f
.
5. Investing in a good project, a ?rm earns:
p
H
R
f
¼ p
H
RI 2bðI 2AÞ ¼ ð p
H
R 2bÞI þbA ¼ p
H
R 2b
À Á
I þbA
0
which is higher than what the ?rm would get investing all her liquid assets A
0
at rate b.
6. Denoting p
H
ðR 2ðB=DpÞÞ ¼ EðR
b
Þ we have:
›ðK
f
=ð1 2ðEðR
b
Þ=bÞÞÞðEðR
b
Þ=bÞ
›b
¼
EðR
b
ÞK
f
EðR
b
Þ 2b
1
b 2EðR
b
Þ
, 0:
7. The comparative statics result described in HT for ð›ðK
b
þ K
f
ÞÞ=›K
f
holds only if one
considers ›b
*
=›K
f
¼ 0:
8. When the technology is the same for all ?rms affected by the reform and it has constant
returns to scale, it is not important which ?rm actually loses most due to the exit of TFR
funds and which one is affected less. They all have, in a way, the “same” production function.
9. We assume here, that SMEs will not change the number of bank relationships following the
reform. Rajan (1992) and Petersen and Rajan (1995) show that a unique bank relationship
helps the access to credit, but it can be more expensive in the long run (a “hold up” problem).
However, Detragiache et al. (2000) ?nd that Italian ?rms typically maintain multiple
relations with banks. In principle, we cannot exclude then that, if the ?rms access to credit
will decrease in the future, some of the smallest ?rms will react by restricting themselves to a
unique relation with a bank.
10. We choose such a formalization for the costs of monitoring since without moral hazard
between bank and depositors the optimal monitoring intensity would be independent of I.
11. More formally, we can rewrite the aggregated equilibrium condition as:
bK
b
ðbÞ ¼ EðR
b
ðmÞÞðK
f
þ K
b
Þ
and differentiate by dK
f
. We obtain:
db
dK
f
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
À Á
¼ EðR
b
ðmÞÞ þ
›EðR
b
ðmÞÞ
›m
›m
›K
f
ðK
f
þ K
b
Þ
db
dK
f
¼
EðR
b
ðmÞÞ þ ðð›EðR
b
ðmÞÞÞ=›mÞð›m=›K
f
ÞðK
f
þ K
b
Þ
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
.
EðR
b
ðmÞÞ
K
b
ðbÞ þ K
0
b
ðbÞðb 2EðR
b
ðmÞÞÞ
for any given level of m, since monitor intensity m increases when total investment K
f
þ K
b
is higher (due to the fact that the ratio D/I reduces for higher I) and since
ð›EðR
b
ðmÞÞ=›mÞ . 0.
12. The aggregate ?ows accruing each year are calculated as:
FlowTFR
i;t
¼ 6:91%
*
W
tfrði;tÞ
where W
tfr(i,t)
is the aggregate gross wage received by employees.
JFEP
3,3
258
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
13. The TFR stock at the end of each year is equal to:
TFR
i;t
¼ ð1 þ r
t
ÞðTFR
i;t21
2TFRLIQ
i;t
Þ þ FlowTFR
i;t
where TFRLIQ
i,t
is the amount of TFR liquidated in year t due to the exit of the employee
from the ?rm, and r
t
is the rate at which TFR is capitalized annually, that is 1.5 percent plus
three-fourth of the in?ation rate annually measured by ISTAT.
14. That is medium and large ?rms according to the European classi?cation (European
Commission, 2003).
15. This happens even though TFR ?ows into INPS are going to receive exactly the same
treatment as within the ?rm in terms of capitalization and workers are not going to
experience any practical difference.
16. In doing this, we assume that the reform is not affecting the labor market conditions, i.e. is
not going to provoke a change in the structure of employment (between SMEs and large
?rms) nor is going to affect the duration of employment. Also, we do not try to estimate the
out?ows of TFR due to workers changing employer or leaving the labor market, which in
reality affect the availability of funds for each ?rm (Fugazza and Teppa, 2005). Our approach
to the labor market conditions is admittedly strong, since it reduces to an estimate of future
in?ows to the TFR based only on the total employment level and the average remuneration.
17. The number of private sector employees up to 2006, by ?rm size, is from ISTAT (2008). The
2007 ?gure has been obtained applying the growth rate of dependent employees as from
ISTAT (2008). The average gross earnings per dependent employee, by ?rm size, up to 2005
comes from ISTAT (2007). Growth rates up to 2007 are from ISTAT (2008). The wage bill
has been obtained by multiplying the number of dependent employees by their average
salary. Total TFR corresponds to 6.91 percent of the wage bill.
18. We assume that workers’ choices about TFR destination – as displayed in the second panel
of Table I – only apply to TFR ?ows in the second semester.
19. We have adjusted the ?gures in order to make them comparable with those of Table II, by
subtracting agricultural/?shery in accordance with the de?nition used in ISTAT (2008).
20. Overall TFR grew at an average rate of 3.6 percent in the 2000-2006 period (ISTAT, 2008).
References
Berger, A.N. and Udell, G.F. (1995), “Relationship lending and lines of credit in small ?rm
?nance”, Journal of Business, Vol. 68 No. 3, pp. 351-81.
Berger, A.N. and Udell, G.F. (1998), “The economics of small business ?nance: the roles of private
equity and debt markets in the ?nancial growth cycle”, Journal of Banking & Finance,
Vol. 22, pp. 613-73.
Bester, H. (1985), “Screening versus rationing in credit markets with imperfect information”,
American Economic Review, Vol. 75, pp. 850-5.
Bianco, M. (1997), “Vincoli ?nanziari e scelte reali delle imprese italiane: gli effetti di una
relazione stabile con una banca”, in Angeloni, I., Conti, V. and Passacantando, F. (Eds),
Le banche e il ?nanziamento delle imprese, Ente per gli Studi Monetari e Finanziari
“Luigi Einaudi”, Rome, pp. 33-60.
Capitalia (2005), “Indagine sulle imprese italiane – Rapporto sul sistema produttivo e sulla
politica industriale”, Osservatorio sulle piccole e medie imprese, Ottobre, Capitalia, Roma.
Carletti, E. (2004), “The structure of relationship lending, endogenous monitoring and loan rates”,
Journal of Financial Intermediation, Vol. 13, pp. 58-86.
Severance
payment
reform
259
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Carletti, E., Cerasi, V. and Daltung, S. (2005), “Multiple-bank lending: diversi?cation and
free-riding in monitoring”, mimeo, Center for Financial Studies, University of Frankfurt,
Frankfurt.
Chiesa, G. (2001), “Incentive-based lending capacity, competition and regulation in banking”,
Journal of Financial Intermediation, Vol. 10, pp. 28-53.
Covip (2008a), La previdenza complementare nel 2007. Dati provvisori, Covip, Roma.
Covip (2008b), Relazione per l’anno 2007, Covip, Roma.
Detragiache, E., Garella, P. and Guiso, L. (2000), “Multiple vs. single banking relationship: theory
and evidence”, Journal of Finance, Vol. 55, pp. 1133-61.
Diamond, D. (1984), “Financial intermediation and delegated monitoring”, Review of Economic
Studies, Vol. 51, pp. 393-414.
European Commission (2003), “Commission recommendation of 6 May 2003 concerning the
de?nition of micro, small and medium-sized enterprises (2003/361/EC)”, Of?cial Journal of
the European Union, May 20.
FRDB (2008), The TFR Survey, available at: www.frdb.org/scheda.php?id¼12&doc_pk¼11026
Fugazza, C. and Teppa, F. (2005), “An empirical assessment of the Italian severance payment”,
WP CeRP No. 38/05.
Guiso, L. (2003), “Small business ?nance in Italy”, EIB Papers, Vol. 7 No. 2.
Holmstrom, B. and Tirole, J. (1997), “Financial intermediation, loanable funds, and the real
sector”, Quarterly Journal of Economics, Vol. 112 No. 3, pp. 663-91.
Huelsewig, O., Mayer, E. and Wollmershaeuser, T. (2005), “Bank loan supply and monetary
policy transmission in Germany: an assessment based on matching impulses responses”,
CESInfo WP No. 1380, University of Munich, Munich.
ISTAT (2007), “Struttura e competitivita´ del sistema delle imprese industriali e dei servizi 2005”,
Statistiche in breve, available at: www.istat.it/salastampa/comunicati/non_calendario/
20071029_00/
ISTAT (2008), “Conti Economici Nazionali, 1970-2007”, available at: www.istat.it/dati/dataset/
20080328_00/
King, S. (1986), “Monetary transmission: through bank loans or bank liabilities?”, Journal of
Money, Credit and Banking, Vol. 18 No. 3, pp. 290-303.
Petersen, M. and Rajan, R. (1995), “The effect of credit market competition on lending
relationships”, Quarterly Journal of Economics, Vol. 110, pp. 407-43.
Rajan, R. (1992), “Insiders and outsiders: the choice between informed and arm’s-lenght debt”,
Journal of Finance, Vol. 47, pp. 1367-400.
Sapienza, P. (1997), “Le scelte di ?nanziamento delle imprese italiane”, in Angeloni, I., Conti, V.
and Passacantando, F. (Eds), Le banche e il ?nanziamento delle imprese, Ente per gli Studi
Monetari e Finanziari “Luigi Einaudi”, Rome, pp. 61-94.
Tito, B. and Zingales, L. (2008), “Chi ha paura dei fondi pensione?”, Anima FinLab, Numero 1,
available at: www.animasgr.it/ANIMA/IT/AnimaFinLab/
Further reading
Bardazzi, R. and Pazienza, M.G. (2006), “La riforma del TFR e il costo per le imprese minori:
un’analisi di microsimulazione”, Politica Economica, Vol. 22 No. 1, pp. 5-50.
Broecker, T. (1990), “Credit-worthiness and interbank competition”, Econometrica, Vol. 58 No. 2,
pp. 429-52.
JFEP
3,3
260
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Freixas, X. and Rochet, J.C. (1998), Microeconomics of Banking, MIT Press, Cambridge, MA.
Garibaldi, P. and Pacelli, L. (2004), “Do larger severance payments increase individual job
duration?”, CEPR Discussion Papers No. 4607.
Istituto di Studi e Analisi Economica (2005), Rapporto ISAE, Finanza pubblica e redistribuzione,
Istituto di Studi e Analisi Economica, Rome, Ottobre.
Palermo, G. and Valentini, M. (2000), “Il Fondo di Trattamento di Fine Rapporto e la struttura
?nanziaria delle imprese manifatturiere”, WP n. 12 02/2000 Banca di Roma.
Pammolli, F. and Salerno, N.C. (2006), “Le imprese e il ?nanziamento del pilastro previdenziale
private”, Nota CERM 02/2006.
Petersen, M. and Rajan, R. (1994), “The bene?t of lending relationships: evidence from small
business data”, Journal of Finance, Vol. 49, pp. 1367-400.
Corresponding author
Riccardo Calcagno can be contacted at: [email protected]
Severance
payment
reform
261
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
4
2
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_976841243.pdf