Research into the Effects of Social Network on Household Debt
Defaults
Dungang Zang
1
, Jiahui Chen
2
R
ural Finance, Sichuan Agricultural University,master degree,Huimin Road,Chengdu, China
2
Department of financial manegement, Sichuan Agricultural University,Huimin road, Chengdu, China
Keywords: Rural-urban migrants; Household debt defaults; Social networks.
Abstract: Based on the data from the Rural Urban Migration in China 2009, this paper explores the effects of
social networks on household debt defaults. Furthermore,we use immigrants’ communication expenses
and the ranking of their brothers and sisters to deal with the possible endogenous bias and draw some
conclusions after robustness tests. Our analysis suggests that the effects of social networks belong to
immigrants on household debt defaults are significant and positive. In addition, our study helps to
understand the mechanism of social networks to rural household debt defaults, provides valuable
policy recommendations to reduce household debt default rates.
1 FIRST SECTION
1.1 Introduction
With series governmental policies to stimulate
consumption demand, increased competition in
banking area and a variation in consuming attitude
among citizens, the scale of household debt has
drastically enlarged in our country these years
(Xinhua Guo, Bin Chen, Zhaihua Wu, 2015). Up to
2012, our country’s consumer credit remaining sum
has risen to 10.27 trillion RMB, taking 20.1 percent
of the GDP. Inevitably, the rise in household debt
leads to the growth in debt defaults, which has a
negative influence on social economy (Oleg
Lugovoy, 2002). Debt defaults represent the
difficulty in financial liquidity, besides they raise the
level of risk premium and then the cost of funds.
Most of all, defaults also result in a decrease in
consumption and savings, which will affect
macroeconomic operation.(Selcuk Caner,
Manouchehr Mokhtari, 2000)
Nevertheless, now on
the condition of “new normal”, consumption is the
main impetus on Chinese economy and there is no
doubt household debt defaults influence strongly on
consumption. Several scholars have blamed the U.S.
sub-prime crisis in 2007 for the exceeding inflation
of household debt.(McKinsey Global Insititute,
2012) .Beyond doubt, reducing household debt
defaults rates has a practical significance for
transforming and upgrading the country’s economy,
and prevents the country from financial risks. Then,
what are factors that affect the household debt
defaults? The answer is complex, especially for rural
families. Under the recent condition of financial
ystem in China, it is impractical for them to borrow
money and take loans through the existed formal
financial channels.
However, through such informal financial
channels, social networks provide rural families with
the required funds.(Kaitong Tan, 2012) Social
network is a significant concept in economic
sociology, it is an integral composition formed by
individual interactions and a stereoscopic network
nor a flat one.
According to the discussion above, based on the
trust in social networks and through informal
financial channels, rural families can obtain loans
and reduce debt defaults. However, what is the
specific connection between the family debt defaults
and social networks? To answer this question, this
article will investigate the behavior of debt defaults
in rural families from the perspective of social
networks.
1.2 Literature Review
Family debt is a financing activity between the
family and formal financial institutes or other
25
Chen J. and Zang D.
Research into the Effects of Social Network on Household Debt Defaults.
DOI: 10.5220/0006443100250030
In ISME 2016 - Information Science and Management Engineering IV (ISME 2016), pages 25-30
ISBN: 978-989-758-208-0
Copyright
c
 2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
25

economic agents. In recent years, to find out the
influenced factors of household debt defaults, a large
number of overseas researchers have made mass
research and analysis. Based on the factors of
demography, consumers with high education had
little possibility to default.(Joanna Stavins, 2000)
In contrast, the national research about household
debt defaults are scare. Professor Xinhua Dun has
pointed out unemployment, triggering events and
change of income and solvency were vital factors
incurring family debet faults.(Xinhua Guo, 2006)
Social network is a combined social relation made
up by individual relatives, schoolmates, friends,
neighbors and colleagues. Pierre Bourdieu has
indicated social network was actually a kind of
social capacity, which came from the economic
capital. Through social network, individuals
received support from cooperative funds(Pierre
Bourdieu, 1986). Professor JingJing Shan has further
pointed out the process of urbanization was also the
process of transformation in social network. Based
on the primitive relationship form region and blood,
migrant workers establish new social network,
which is a vital resource helping them obtain social
support(Jingjing Shan, 2007). Compared with the
prior migrant workers, there is an obvious decrease
in dependency and focus on rural social network,
new migrant workers however show a characteristic
of urbanize(Xin Qi, 2007) and phenomenon of
cohesive subgroups(Yike Ren, 2008).
According to the analysis above, based on the
development in several decades, the research about
household debt defaults are sophisticated in such
aboard countries. However, on account of limited
time for consumption credit, the national research is
still at the primary stage, we lack the research about
the connection between social networks and
household debt defaults. To fill this existed vacancy
in national research part, we choose the new migrant
workers as our target group and write this article
about the effects of social networks on household
debt defaults.
2 RESEARCH DESIGNE
2.1 Study Sample and Data Sources
Our data comes from the investigation of Rural
Urban Migration in China, RUMIC in 2009. This
project is initiated by researchers from Beijing
Normal University, Australian National University
and the University of Queensland and supported by
IZA and NBS. The survey samples of the project
include 5 thousand samples form urban families, 10
thousand samples from rural-urban migrations and
involved 9 provinces and 15 cities. The contents of
the questionnaires comprise basic case of family
members, income and expenditures of the families
and the condition of their education, employment
and social relationships. According to the purpose of
the research, we have choose 10 thousand samples
of rural-urban migrations from RUMIC in 2009,
deleted abnormal samples and samples lake of
relative variables, finally we gained the samples
used to positive analysis in this article.
2.2 Model Construction and
Declaration of Variables
For researching the effects of social networks on
family debt defaults, we consult relevant literature
and build our regression model as follows:
Mi=α0+α1Xi+βZi+εi
In this model, the explanatory variable Xi refers to
the ith migrant worker’s proxy variable on social
networks, the explained variable Mi refers to the ith
migrant worker’s family defaults condition. Besides,
Zi refers to the control variable of possible
effect on
family debt defaults, mainly including wage, number,
gender, nation, health, risk take, happiness, job, trust
and married ( take the married as reference group ).
Moreover, this article set up five dummy variables:
married1, married2, married3, married4, married5,
represent first marriage, remarriage, cohabitation,
divorce and death of spouse. The reason why we
choose trust as a variable, is articles of sociology
always allude trust to a vital part of social networks.
α0，α1 and β stand for coefficient of parameters to
be estimated, εi stand for error term of regression
model.
The definition and explanation of the variables are
as
table 1.
Variab
les
Description Explanation
W
repay
ment
Proxy variable of
household debt
defaults.
debt defaults / Total
income
ISME 2016 - Information Science and Management Engineering IV
26
ISME 2016 - International Conference on Information System and Management Engineering
26

repay0
1
The existence of
debt defaults
No=0；Yes=1
Wage Labor income.
Average monthly
incomes.
Helper
s
The number of
people who offered
helps to the migrate
workers in the last
year.
Include the helps of
money, job, suggestion,
and looking after baby
etc.
Numbe
r
The number of
brothers and sisters.
Gender Gender. Female=0；Male=1
Minzu Nation. Minority=0；Han=1
Marrie
d if
Get married or not. No=0；Yes=1
Marrie
d1
First marriage No=0；Yes=1
Marrie
d2
Remarriage No=0；Yes=1
Marrie
d3
Cohabitation No=0；Yes=1
Marrie
d4
Divorce No=0；Yes=1
Marrie
d5
Death of spouse No=0；Yes=1
Health Health statues
Very good=1 Good=2;
Common=3;
Bad=4;Very bad=5
Job Employment statues
Unemployment=0；
Employment=1
Risk
take
Risk bearing
Scale of 0 to 10 based
on the level of risk
Happin
ess
Sense of happiness
Very happy=1；
Happy=2；Not
happy=3；
Unfortunate=4
Trust
Trust
Scale of 1 to 4 based on
the degree of
trust.
2.3 Descriptive Statistics
According to our research, 54.5 percent of
interviewees are male, while females take up to 45.5
percent, 98.6 percent of interviewees are han, merely
come from minority. Besides for the marriage status,
only 9.8 percent of migrant workers have been
married, most of them are still single. Their health
status are generally poor and the medical cost are
always high, which could result in a increase in
household debt defaults. Take the inclination of risk
bearing as concerned, the table shows their average
inclination is neutral and the level is not high, which
indicates their low participation rate in household
financial marketing. Moreover, the average index of
happiness statue is 3.112, this low happiness statue
may result from the high economic pressure. At last,
speaking to the employment status of the migrate
workers, 89.7percent of them take upoccupation, but
few is still unemployed.
3 EMPIRICAL ANALYSIS
3.1 Analysis of Regression Result
Firstly we employ Tobit to estimate the measured
model, the regression result is as Table 2.
Research into the Effects of Social Network on Household Debt Defaults
27
Research into the Effects of Social Network on Household Debt Defaults
27

Variables (1) (2) (3) (4) (5)
Tobit Tobit Probit Probit Reg
helpers 0.573
*
**
0.499
***
0.022
*
**
0.019
*
**
0.078
**
gender 5.281
**
0.171
*
0.829
**
minzu 22.097 0.742 1.989
*
marriedif 11.332
**
*
0.431
*
**
1.701
**
*
married1 8.188
***
0.255
**
1.098
*
married2 -4.321 -0.254 -0.037
married3 6.969 0.409 -1.091
married4 14.059
*
0.479
*
1.672
married5 10.925 0.324 1.210
number 1.604
**
0.044 0.339
**
*
health -6.831
***
-0.247
***
-
0.820
**
*
risktake 0.440 0.015 0.067
j
ob 6.794 -0.389
**
-0.951
Happiness -2.948
*
-0.148
**
-0.168
Trust -3.65
**
-0.145
***
-0.38
*
Wage -0.31
**
-0.009 -0.05
**
Cons -
34.89
4
***
-12.345 -
1.22
2
***
-0.090 4.210
**
Pseudo R
2
0.004 0.03
4
0.01
0
0.072 0.033
N
1815 1664 1819 1667 1664
Table 3: Note: *, **, *** refer to reject null hypothesis
under the significance level of 10%, 5% and 1%.
The result of regression at fist to second line
showed ,whether we bring in control variables or not,
the regression index of social network is positive
under the significance level of 0.01. This indicates
social networks have a positive impact on household
debt defaults.Except for some explained variables of
social networks, we make other regression analysis
about the variables of individual feature and
household economic. The regression result at the
second line of the Tobit model indicates compared
with male the situation of debt defaults is severer
among female. However the regression index is not
outstanding in the factor of nation, illustrating nation
has no significant effect on household debt defaults.
Besides the table shows increasing labor income
adds household disposable income, meanwhile
decrease the possibility of debt defaults. Moreover,
the fifth line of the 3
rd
table the regression index of
social networks is still positive.
3.2 Process of Endogenous Character
Estimating datum equation results in biased
estimation of some key parameters, the main
problem is social networks related to error terms.
Household debt defaults, to some extent, effect
social network. Migrant workers’ social networks
and household debt defaults probably have reverse
causality, which lead to endogenous character. In
order to amend the biased error of endogenous
character, we use instrument variables: TELE and
RANK. TELE refers to migrate workers’
communication costs, RANK refers to their rank
among peers in their family. This two instrument
variables effect directly on social networks, however,
have no influence on migrate workers’ defaults. The
reason why we choose TELE as instrument variables
is communication costs are related to social
relationships. Generally, communication costs
reflect the frequency and time migrate workers
connect with others. The lager the costs are, the
more intimated connection is and the lager the scale
of social network is. We choose RANK as the
ISME 2016 - Information Science and Management Engineering IV
28
ISME 2016 - International Conference on Information System and Management Engineering
28

second instrument variable, on account of migrate
workers rank among peers in their family always
reflect the number of their brothers and sisters.
According to the diffusibility of social relationships,
lager number of sisters and brothers means lager
scale of social networks. However communication
costs and the rank have no direct influence on
household debt defaults.
Applying instrument variables, we utilize IV
Tobit, IV2SLS, IVLIML and IVGMM to analyze
data, the results are showed in the forth table. The
table’s first, third, fifth and seventh lines show,
under the significance level of 10% and 1%, the
variable coefficient of social network is positive.
Besides excluded-exogeneity estimation of Wald
rejects the null hypothesis about social network is
one exogenous explaining variable, in another word,
it’s an endogenous explaining variable. Meanwhile
we make a estimation about weak instrument
variables . The second line of the 3
rd
table indicates,
in the second stage of 2SLS, when using instrument
variables TELE and RANK, robustness of weak
instrument variables reject the null hypothesis,
which means the problem of weak instrument
variables dose not exist. The results from the forth
and sixth line is also the same with the second line.
In conclusion, the instrument variables of this article
neither have problem of weak instrument variables
nor endogenous character, besides employing TELE
and RANK to estimate the effect of new migrate
workers’ social networks on household debt defaults
is suitable and necessary.
Table 3: Test Result of Endogenous Character
Variab
les
(1) (2) (3) (4) (5) (6) (7)
IV
To
bit
First
Stage
(2SL
S)
IV
2S
LS
First
Stag
e(LI
ML)
IVL
IM
L
First
Stag
e(G
MM)
IV
G
M
M
Helper 5.5
67
*
1.4
28
*
**
1.43
1
***
1.4
62
*
**
Gender 0.5
87
0.788
**
-
0.2
57
0.78
8
**
-
0.26
0
0.78
8
***
-
0.2
87
minzu 19.
826
0.695 1.4
10
0.69
5
1.40
9
0.69
5
1.4
02
Marrie
dif
8.7
76
*
*
(4.
405
)
0.568
0.479
）
1.0
29
(0.
934
)
0.56
8
（
0.47
9）
1.02
7
（
0.93
6)
0.56
8
(0.62
0)
1.0
50
(1.
23
6)
rank 0.239
*
0.23
9
*
0.23
9
*
Cons -
50.
629
5.427
***
-
6.2
17
5.42
7
***
-
6.24
4
5.42
7
***
-
6.5
24
(31
.21
2)
(1.98
9）
(5.
619
)
(1.98
9)
0.00
0
(1.95
1)
(5.
03
7)
Pseudo
R
2
0.0
55
0.055 0.0
00
0.05
5
0.05
5
0.0
00
AR 14.34
***
14.3
4
***
6.65
***
18.6
7
***
Wald
test
4.3
7
**
6.6
5
***
151
5
9.7
2
**
N 151
5
1515 151
5
1515 1515 15
15
N
ote ：① AR refers to the examination of robustness
among weak instrument variables
② Wald test refers to the examination of exogenous
exclusion.
4 CONCLUSION
This article employs data about Chinese village-city
migrate workers in 2009 and makes a deeper
research into the effect of social networks on
household debt defaults.
Our conclusion is: social networks have a clearly
positive influence on household debt defaults,
besides the problem of endogenous character still
exists. However the result of this article is
Research into the Effects of Social Network on Household Debt Defaults
29
Research into the Effects of Social Network on Household Debt Defaults
29

inconsistent with previous ones, the reasons for this
coming from two aspects. First, the existence of
“Liability Pyramid”. Migrate workers always relieve
their borrowing and finance demand through social
networks, but on account of a decrease in excepted
income, migrate workers may not able to pay back
all their debt. If this happens, they will borrow
money once again from the third-party, which can
result a vicious circle and increase the rate of default.
Second, lack of normalization in informal lending.
Migrate workers’ loans mainly come from informal
channels, the larger scale of social networks is, the
more various informal channels they in touch with.
In case debt defaults are more widespread, because
informal channels are built by relative relationships
and lack of institutional limit. As the conclusion
showing above, this article offers 3 suggestions as
follows:
(1)Improve rural society security system.
government should increase the investment in rural
Old-Age security and New Rural Cooperative
Medical System, guaranteeing living demands of
remind peasants and reducing the costs of social
security. In this way, government will help migrate
workers reduce their economic and debt burden.
(2)Improve rural society credit system. As far as
we are concerned, government should improve rural
society credit system , build up credit evaluation
system and information sharing system.By
strengthening the link between loans and credit,
government helps perfect rural financial credit
environment and prevent migrate workers from
defaults.
(3)Improve rural financial service system.
Government should improve rural financial service
system. On the one hand, strengthen the construction
of hardware facilities,increase “Internet+agriculture”
layout and set up internet-agriculture financial
platform.
REFERENCE
Jinjin Shan, J., In Issues in Cities, 2007, (3): 59-63. ,
Variation of migrate workers’ social networks.
Joanna Stavins, J., In New England Economic
Review,2000, (2):15-30., Credit Card
Borrowing,Delinquency,and Personal Bankruptcy.
KaiTong Tan, D., A master dissertation of Hunan
university,2012, Reaserch into the effect of social
networks on migrate workers’ loan.
McKinsey Global Institute, N., Debt and
deleveraging:Uneven progress on the path to growth,
Insights & Publications, 2012, (1): 45-66.
Oleg Lugovoy, J., In Arrears in Russian Economic Trends,
2002, (2)
： 19-30., a Transition Economy:the
Mechanism of Contagion.
Pierre Bourdieu, Z., In Greenwood, New York , 1986:
241-258, The Forms of Capital.
Selcuk Caner, Manouchehr Mokhtari, J., In Economics
Quarterly, 2000, (5):27-53., Arrears and their
implications for economic performance in the Russian
federation
Xinhua Guo, D., A doctor dissertation of Huazhong
university, 2006, Household loans, default and
bankruptcy.
Xinhua GUO, Bin Chen, Wu Zhaihua, J., In Statistics and
decisions, 2015,(4):96-99., Empirical study for
connection between changing of Chinese demographic
and increase in household debt defaults
Xin Qi, J., In Journal of Jiangsu administration institute,
2007,(3):75-79., Extend and construct: New
generation of workers’ social networks.
Yike Ren, J., In Society, 2008,(5):20-40., Analysis in
cohesive subgroups of Chinese migrate workers’
social networks.
ISME 2016 - Information Science and Management Engineering IV
30
ISME 2016 - International Conference on Information System and Management Engineering
30