Abstract
In this article, we consider the impact of personal contacts on the labor market outcome. Unlike previous studies, we do not assume any particular network structure or vacancies communication protocol. Instead, we state three general properties of matching functions that allow us to establish the existence and uniqueness of equilibrium and characterize the impact of social ties on the labor market. In particular, we show that a monotonically increasing matching function in socialization level is a necessary and sufficient condition for having monotonically decreasing unemployment and increasing wage and market tightness. However, the same does not apply to vacancy rate. We establish a condition under which a monotonically increasing matching function produces a vacancy rate that first increases in socialization level, but then decreases.
Acknowledgments
The author thanks Fernando Vega-Redondo, Marco Van der Leij, Andrea Galeotti, Vadym Lepetyuk, and funding from the HSE Research Laboratory for Strategic Behaviour and Institutional Design.
Appendix
Proof of Proposition 1
Using the first and second conditions, let us define the following functions:


Existence:
First, we should prove that and
implicitly define continuous relationship of u as a function of v. Note that function
is continuous in all three arguments and has following properties:

Thus, for any fixed , there exists
and it is unique. We still need to prove that
is continuous.
Let us prove by contradiction; assume that is not continuous and there exists a point of discontinuity
and
. Let us choose any sequence of points in the domain of
,
such that this sequence has limit
. Then, there exists a sequence of
such that for
,
. Note that by properties of
, function
is bounded and hence
is bounded. By the Bolzano–Weierstrass theorem, any bounded sequence has a convergent subsequence. Let us choose subsequences
such that
has a limit
. Note that
has the same limit as the original sequence, since any subsequence of convergent sequence converge to the same limit. By our assumption
. By joint continuity of
, we know that

thus we have and
since
and
satisfy condition
then both of them belong to
. However, this is possible only if
is not uniquely determined. We obtained a contradiction, since given the properties of
, we know that
is uniquely determined. Thus, we can conclude that
is jointly continuous in v and s.
The second function is continuous in all three arguments. Recall that . We know that
and it is easy to see that
and
, which is positive for any s and v as long as
. Following the same lines of proof, we can argue that
is unique and jointly continuous.
Thus, a solution obviously exists if at , the Beveridge curve is situated to the left of the labor-demand curve. We obtain
and
. So condition is
.
Now let us prove that are continuous functions in s. Let us define following function:

by joint continuity of and
we know that
is jointly continuous in v and s. Additionally, we imposed restrictions that insure that
and
for any s. Using the same line of arguments as before, we can conclude that
is continuous.
As we know, at ,
so
. Taking one of them, e.g.
we know that
is continuous and
as well, which implies that
is continuous too.
Uniqueness:
To fix the socialization level of workers s, we first prove that the Beveridge curve, u is strictly decreasing in v. We already have seen that . Using (A1), it is easy to show that
. Thus, by implicit function theorem
, and the Beveridge curve is strictly decreasing in u in the
plane.
Using (A1) and (A2), we can show that the derivative of with respect to v is negative:

Using (A1) and (A3), we can show that derivative of with respect to u is positive:

Thus by implicit function theorem . This implies that labor demand is monotonically increasing in unemployment u, which combined with the fact that the Beveridge curve is strictly decreasing, implies uniqueness of the solution for each given s.
Proof of Proposition 2
An increase in the cost of posting vacancy shifts the labor demand curve leaving the Beveridge curve unaffected. Thus, the effect of increase in
on
and
depends on the sign of
. Deriving
with respect to
we find that:

Using the fact that by the implicit function theorem we get that
and the labor demand curve shifts downward. This in turn implies that increase in
increases
and at the same time decreases
.
Proof of Proposition 3
is decreasing in s on the interval
Let us choose two values of search intensity and s s.t.
, both belonging to the interval
and compare position of curves on the
plane. Observe that for a given value u we have
, thus to equate the right-hand side v should decrease. Recall that
. This in turn implies that the Beveridge curve shifts downward on the
plane.
In Proposition 1, we have shown that . Taking derivative of
with respect to s and rearranging we obtain:

Using the implicit function theorem, we can conclude that . This implies that the labor demand curve shifts upward. Combining both results, it is easy to see that
.
By analogy if the matching function decreases on the interval then
, and thus when we move from s to
the Beveridge curve shifts upward on the
plane. The derivative
changes sign and becomes negative and thus the labor demand curve shifts downward. Combining both results, one can show that
.
decreases in s on
if
and increases otherwise
Substituting matching function into the second condition, we can rewrite it in the following way:

The second condition does not depend on s and thus is unaffected by the change in s. This implies that it should hold for all socialization levels s. Expressing the inverse of v from the second equation, we get:

Deriving the right-hand side of the expression with respect to u, we obtain:

The derivative is negative if , where
and positive otherwise. This in turn implies that the second curve is increasing on the interval
and decreases afterward. Despite this non-monotonic shape of the second curve Proposition 1 implies uniqueness of the equilibrium. There are two points of intersection, but one of them
and
is not an equilibrium because by choosing
we ensure that the vacancy rate is not zero when
.
If a matching function increases on the interval , then we have seen that an increase in s shifts the Beveridge curve downward on the
plane, leaving the second curve unaffected. This implies that the equilibrium unemployment
decreases in s, while vacancy rate
is decreasing in s if
and increasing otherwise.
If a matching function decreases on the interval , then an increase in s shifts the Beveridge curve upward on the
plane and the equilibrium unemployment
increases in s. The vacancy rate
increases in s, if
and decreases otherwise.
Proof of Proposition 4
A necessary condition
A necessary condition is that is greater than the minimal level of unemployment, which is
. Equalizing both expressions, we get
. Let us assume that the denominator of
is greater than zero, which is true when
. Thus, we get the following inequality:
![[7]](/document/doi/10.1515/bejte-2012-0021/asset/graphic/bejte-2012-0021_eq7.png)
Solving the quadratic equation, we get that if , then the expression is positive if
. If
then
should belong to interval
, which contradicts the assumption that the denominator of
is greater than zero. Thus, we get:
![[8]](/document/doi/10.1515/bejte-2012-0021/asset/graphic/bejte-2012-0021_eq8.png)
Assume now that the denominator of is less than zero and thus
. In this case, inequality [7] is reversed. Thus, if
, then the expression is negative if
or
. Taking into account the assumption that the denominator of
is negative the only interval is
. If
, then we get
or
. Taking into account assumption, we get
. Thus, we get:
![[9]](/document/doi/10.1515/bejte-2012-0021/asset/graphic/bejte-2012-0021_eq9.png)
Combining eqs [8] and [9], we get that if then condition is
, while if
then
should be greater than
, which in case
equals to
. Finally, we get that
should be greater than
.
A necessary condition
Assumption (A4) implies that equilibrium unemployment is a decreasing function in s. Thus to have non-monotonic behavior of the vacancy rate, should be greater than
. Using (A4), we substitute
into eqs [2] and [6], which gives the following system of equations:


Combining, we obtain:

Substituting , we get something negative and thus one root is positive and one is negative. Note that the expression is decreasing in
. Substituting
, we get that the expression is negative if the following holds:

Thus, a necessary condition for non-monotonic behavior of vacancy rate requires to be higher than
.
A sufficient condition
Assumption (A4) and Proposition 3 imply that unemployment is monotonically decreasing in s. Thus, the lowest possible unemployment in the labor market is and
. The last condition that we should check is that
. Depending on the assumed mechanism of vacancies transmission, the minimal unemployment level may be different for different setups. However, we know that for any setup
, when
. By Proposition 2, we know that both
and
are increasing in
, which combined with the fact that
implies that for
higher, but arbitrary close to
we will have
, which by Proposition 3 guaranties non-monotonicity of vacancy rate.
Proof of Proposition 5
To identify the effect of s on the market tightness, we express from the second equation:

This expression relates the inverse of the equilibrium market tightness to the equilibrium unemployment rate. The derivative of the right-hand side with respect to s is given by:

If a matching function is increasing on the interval , then by Proposition 3 the equilibrium unemployment level decreases in s, which implies that inline-formula>
decreases in s too and thus market tightness
increases in s. Consequently, equilibrium wage
also increases in s.
If instead a matching function decreases on the interval , then the equilibrium unemployment level increases in s, and thus both market tightness
and equilibrium wage
are decreasing in s.
Proof of Lemma 1
(A1a)is increasing and concave in v


(A1b)is increasing and concave in u

Due to the exponential term, it is difficult to identify the sign of the derivative. Let us prove first that the function is concave:

This implies that minimum of the first derivative of the matching function with respect to u lays on boards. One can easily verify that the derivative is positive on both ends: , since
and
. Thus, we can conclude that
for
.
(A2),
, and
By construction, the matching function is unemployment rate multiplied by the individual hiring probability and thus . By the model specification, the share u of vacancies goes to unemployed workers and results in immediate match. On the other hand, the share
of vacancies is received by employed workers and not all of them result in match. This happens due to the miscoordination among workers when one worker can receive multiple offers or in the case when an employed worker does not have unemployed neighbors. This implies that by construction
.
The last two properties can be easily verified by taking limits.
(A3)is decreasing in vacancies v and
is decreasing in unemployment u

trivially follows from the fact that the matching function is concave in v.

Note that .
(A4)is increasing and concave in socialization level of workers s


Substituting to the matching function, we get
.
References
Cabrales, A., A. Calvó-Armengol, and Y. Zenou. 2007. “Effort and Synergies in Network Formation,”' mimeo.Search in Google Scholar
Cahuc, P., and F. Fontaine. 2009. “On the Efficiency of Job Search with Social Networks.” Journal of Public Economic Theory 11(3):411–39.10.1111/j.1467-9779.2009.01416.xSearch in Google Scholar
Calvó-Armengol, A. 2004. “Job Contact Networks.”' Journal of Economic Theory 115(1):191–206.10.1016/S0022-0531(03)00250-3Search in Google Scholar
Calvó-Armengol, A., and M.O. Jackson. 2004. “The Effects of Social Networks on Employment and Inequality.” American Economic Review 94(3):426–54.10.1257/0002828041464542Search in Google Scholar
Calvó-Armengol, A., and Y. Zenou. 2005. “Job Matching, Social Network and Word-of-Mouth Communication.” Journal of Urban Economics 57:500–22.10.1016/j.jue.2005.01.001Search in Google Scholar
Fontaine, F. 2008. “Why Are Similar Workers Paid Differently? The Role of Social Networks.” Journal of Economic Dynamics and Control 32(12):3960–77.10.1016/j.jedc.2008.04.009Search in Google Scholar
Galenianos, M. 2012. “Hiring Through Referrals.” manuscript.10.2139/ssrn.1802812Search in Google Scholar
Galeotti, A., and L.P. Merlino. 2011. “Endogenous Job Contact Networks.” manuscript, University of Essex.Search in Google Scholar
Ioannides, Y.M., and L.D. Loury. 2004. “Job Information Networks, Neighborhood Effects, and Inequality.” Journal of Economic Literature 42(4):1056–93.10.1257/0022051043004595Search in Google Scholar
Ioannides, Y.M., and A.R. Soetevent. 2006. “Wages and Employment in a Random Social Network with Arbitrary Degree Distribution.” The American Economic Review 96:270–4.10.1257/000282806777212062Search in Google Scholar
Petrongolo, B., and C.A. Pissarides. 2001. “Looking into the Black Box: A Survey of the Matching Function.” Journal of Economic Literature 39(2):390–431.10.1257/jel.39.2.390Search in Google Scholar
Pissarides, C.A. 2000. Equilibrium Unemployment Theory, Second ed. Cambridge: MIT Press.Search in Google Scholar
Topa, G. 2011. “Labor Markets and Referrals.” In Handbook of Social Economics, edited by Jess Benhabib, Alberto Bisin and Matthew O. Jackson, 1193–1221. Amsterdam: Elsevier Science, forthcoming.10.1016/B978-0-444-53707-2.00005-0Search in Google Scholar
- 1
See, for example, Jackson and Calvó-Armengol (2004), Calvó-Armengol (2004), and Calvó-Armengol and Zenou (2005).
- 2
See, for example, Calvó-Armengol (2004), Calvó-Armengol and Zenou (2005), Ioannides and Soetevent (2006), Cahuc and Fontaine (2009), and Galeotti and Merlino (2011).
- 3
The article by Calvó-Armengol and Zenou (2005) assumes a regular network (all workers have the same number of contacts), Galeotti and Merlino (2011) employs the network formation mechanism that leads to a Poisson degree distribution, and finally Cahuc and Fontaine (2009) assumes that the network is composed of cliques (components inside which all workers are connected).
- 4
Essentially, the network structures in Galenianos (2012) and Calvó-Armengol (2004) are the same, and thus, both studies should produce the same result regarding the effect of average connectivity on the employment. However, the main difference is that in Calvó-Armengol (2004) employed contacts who hear about a vacancy redirect it at random to one of their unemployed neighbors, while in Galenianos (2012) they do it regardless of the employment status of neighbors.
©2013 by Walter de Gruyter Berlin / Boston
Articles in the same Issue
- Masthead
- Masthead
- Advances
- Dependence and Uniqueness in Bayesian Games
- Monopolistic Signal Provision†
- Multi-task Research and Research Joint Ventures
- Transparent Restrictions on Beliefs and Forward-Induction Reasoning in Games with Asymmetric Information
- A Simple Bargaining Procedure for the Myerson Value
- On the Difference between Social and Private Goods
- Optimal Use of Rewards as Commitment Device When Bidding Is Costly
- Labor Market and Search through Personal Contacts
- Contributions
- Learning, Words and Actions: Experimental Evidence on Coordination-Improving Information
- Are Trust and Reciprocity Related within Individuals?
- Optimal Contracting Model in a Social Environment and Trust-Related Psychological Costs
- Contract Bargaining with a Risk-Averse Agent
- Academia or the Private Sector? Sorting of Agents into Institutions and an Outside Sector
- Topics
- Poverty Orderings with Asymmetric Attributes
- Dictatorial Mechanisms in Constrained Combinatorial Auctions
- When Should a Monopolist Improve Quality in a Network Industry?
- On Partially Honest Nash Implementation in Private Good Economies with Restricted Domains: A Sufficient Condition
- Revenue Comparison in Asymmetric Auctions with Discrete Valuations
Articles in the same Issue
- Masthead
- Masthead
- Advances
- Dependence and Uniqueness in Bayesian Games
- Monopolistic Signal Provision†
- Multi-task Research and Research Joint Ventures
- Transparent Restrictions on Beliefs and Forward-Induction Reasoning in Games with Asymmetric Information
- A Simple Bargaining Procedure for the Myerson Value
- On the Difference between Social and Private Goods
- Optimal Use of Rewards as Commitment Device When Bidding Is Costly
- Labor Market and Search through Personal Contacts
- Contributions
- Learning, Words and Actions: Experimental Evidence on Coordination-Improving Information
- Are Trust and Reciprocity Related within Individuals?
- Optimal Contracting Model in a Social Environment and Trust-Related Psychological Costs
- Contract Bargaining with a Risk-Averse Agent
- Academia or the Private Sector? Sorting of Agents into Institutions and an Outside Sector
- Topics
- Poverty Orderings with Asymmetric Attributes
- Dictatorial Mechanisms in Constrained Combinatorial Auctions
- When Should a Monopolist Improve Quality in a Network Industry?
- On Partially Honest Nash Implementation in Private Good Economies with Restricted Domains: A Sufficient Condition
- Revenue Comparison in Asymmetric Auctions with Discrete Valuations