individual perception of accessible social capital and attitude to thrift

14
Individual perception of accessible social capital and attitude to thrift Meng-Na Xu Ming-Lin Wang Received: 23 July 2013 / Accepted: 1 February 2014 Ó Springer Science+Business Media New York 2014 Abstract We empirically assess the effects of individual perception of accessible social capital on individual attitude to thrift in this paper, based on the view that social capital serves as an informal insurance. We look at individual-level data from the World Values Survey. After controlling for demographic characteristics as well as wave and country fixed effects, we find that individual perception of accessible social capital does matter. One who is more generally trusting or has a broader social network tends to value thrift less. Keywords Thrift Á Social capital Á Perception Á Precautionary saving Á Generalized trust JEL Classification D01 Á D14 Á J1 1 Introduction Household saving largely relies on individual saving behavior. However, from a micro perspective, individual saving behavior can be regarded as having two main sources, namely, one’s attitude to thrift (whether one values thrift or not) as well as one’s ability to save, which are not well distinguished in many researches. To neglect or mix these two sources leads to confusion in both theoretical and practical fields. And since one’s ability to save is important only when his/her attitude to thrift does not block the saving intention, individual attitude to thrift is noteworthy. M.-N. Xu (&) Á M.-L. Wang School of Economics and Management, Hangzhou Normal University, No. 58, Haishu Rd., Cangqian, Yuhang District, Hangzhou 311121, Zhejiang Province, China e-mail: [email protected] M.-L. Wang e-mail: [email protected] 123 Rev Econ Household DOI 10.1007/s11150-014-9240-2

Upload: ming-lin

Post on 23-Dec-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Individual perception of accessible social capital and attitude to thrift

Individual perception of accessible social capitaland attitude to thrift

Meng-Na Xu • Ming-Lin Wang

Received: 23 July 2013 / Accepted: 1 February 2014

� Springer Science+Business Media New York 2014

Abstract We empirically assess the effects of individual perception of accessible

social capital on individual attitude to thrift in this paper, based on the view that

social capital serves as an informal insurance. We look at individual-level data from

the World Values Survey. After controlling for demographic characteristics as well

as wave and country fixed effects, we find that individual perception of accessible

social capital does matter. One who is more generally trusting or has a broader

social network tends to value thrift less.

Keywords Thrift � Social capital � Perception � Precautionary saving �Generalized trust

JEL Classification D01 � D14 � J1

1 Introduction

Household saving largely relies on individual saving behavior. However, from a

micro perspective, individual saving behavior can be regarded as having two main

sources, namely, one’s attitude to thrift (whether one values thrift or not) as well as

one’s ability to save, which are not well distinguished in many researches. To

neglect or mix these two sources leads to confusion in both theoretical and practical

fields. And since one’s ability to save is important only when his/her attitude to

thrift does not block the saving intention, individual attitude to thrift is noteworthy.

M.-N. Xu (&) � M.-L. Wang

School of Economics and Management, Hangzhou Normal University, No. 58, Haishu Rd.,

Cangqian, Yuhang District, Hangzhou 311121, Zhejiang Province, China

e-mail: [email protected]

M.-L. Wang

e-mail: [email protected]

123

Rev Econ Household

DOI 10.1007/s11150-014-9240-2

Page 2: Individual perception of accessible social capital and attitude to thrift

Guiso et al. (2006) also provide a suggestive evidence that the proportion of

individuals who value thrift in each country is positively related to national saving.

Therefore, we intend to explore the factors that affect individual attitude to thrift,

especially the function of individual perception of accessible social capital.

There is a large literature surrounding the factors that cause different levels of

household saving. Economic environment and public policies have attracted much

attention (Wilcox 1991; Berry et al. 2009). Researchers believe that a sound social

security system can weaken individual saving motive, which is often linked with

discussions on precautionary saving (Kantor and Fishback 1996; Chou et al.

2003). Demographics are also noticed as important factors. Loayza et al. (2000)

find that the level and growth rate of real per capita income play an important role

in the rise of private saving rates. Bloom et al. (2002) explore the effects of health

and longevity. As for psychological factors, emotional stability, autonomy, and

extraversion are found to be robust predictors of saving and borrowing behaviors

(Nyhus and Webley 2001). Duflo and Saez (2003) show that individual saving

decisions are likely to be affected by the choices of others in their peer group.

Nonetheless, most of the researches paid little attention to individual attitude to

thrift, or that they did not distinguish between attitude to thrift and saving

behavior. It makes the mechanism that causes different levels of household saving

remains unclear. In fact, scholars have argued on whether the Americans have

sufficient savings. A widespread ‘‘buy first, worry later’’ attitude is cited. It is

warned that the typical American household spends too much and saves too little

and therefore is quite vulnerable to a serious financial crisis.1 This concern is not

without reason. Then, what affects individual attitude to thrift? Several researches

started to explore the links between different economic and socio-demographic

factors and individual preference for thrift (Guiso et al. 2006; Anderson and

Nevitte 2006). However, they did not take individual perception of accessible

social capital into consideration.

The concept of social capital has been applied in the studies at different levels.

On the one hand, social capital is often evaluated from the collective perspective

and refers to characteristics of communities, workplaces, or regions (Putnam 1995a,

b; Schafft and Brown 2000; Paxton 2002). On the other hand, social capital is also

conceptualized as a pool of resources for the individual in particular (Erickson 1996;

Portes 2000; Agarwal et al. 2011). In this paper, we relate social capital to

individual attitude to thrift and concentrate on individual-level analyses. The

intuition is that social capital serves as a substitution of the formal social insurance

and we intend to work within the framework of precautionary saving. Precautionary

saving is often linked with insurance, but seldom with informal insurance such as

social capital. Nevertheless, it is plausible that people who think themselves as

having more accessible social capital would value thrift less. And regarding

individual attitude to thrift, the subjective perception of social capital is more

important than the real accessible social capital.

1 See ‘‘Many in U.S. Think Lottery, Not Savings, Is Path to Security’’, Wall Street Journal—Eastern

Edition, 10/29/1999, 234 (85), pA6.

M.-N. Xu, M.-L. Wang

123

Page 3: Individual perception of accessible social capital and attitude to thrift

Therefore, we intend to use individual level data from the World Values Survey

(hereafter referred to as WVS) to examine how individual perception of accessible

social capital relates to attitude to thrift.2 The results conduce to a better

understanding of the formation of individual attitude to thrift and contribute to the

literature on household saving by providing new evidences from a micro and

psychological perspective.

The rest of the paper is organized as follows. Theoretical consideration and

hypotheses are offered in Sect. 2. Section 3 focuses on data and empirical strategy.

In Sect. 4, we present the descriptive statistics, empirical results and discussion.

Finally, Sect. 5 concludes.

2 Individual perception of accessible social capital and hypotheses

We assume that social capital serves as a substitution of the formal social insurance.

Accordingly, individuals who regard themselves as having more accessible social

capital tend to value thrift less for insurance. Therefore, individual perception is of

great importance. But how can we measure individual perception of accessible

social capital? To answer this question, we should first note that the subjective

perception of social capital is different from the real accessible social capital.

Individual perception of accessible social capital can be influenced by both intrinsic

values and social capital investment. In this paper, we intend to focus on generalized

trust and social network. They depict two channels and function simultaneously.

The influence of the availability of informal insurance on individual attitude to

thrift is unlikely to be affected by the needs to help others out. Although some

people may want to save for others even strangers, it is plausible that this motive is

not very strong compared with the saving motive for themselves and family (it is

notable that family is a limited group). Moreover, when a person believes most

people can be trusted for informal insurance, he/she does not mean that most people

can be trusted for him/her solely excluding other people. Thus, we believe that one’s

perception of accessible social capital will largely influence his/her attitude to thrift.

Individual perception of accessible social capital reflects one’s real social capital,

yet not entirely. One who is more generally trusting or has a broader social network

tends to have higher perception of accessible social capital. On the one hand, there

is a common claim that individual social capital is linked with one’s social network

(Erickson 1996; Van Der Gaag and Snijders 2005; Son and Lin 2008). Individuals

can get access to their real social capital by assessing social network. On the other

hand, generalized trust is also important in the formation of individual perception.

Generalized trust gives individuals more confidence in getting help from other

people. It can be regarded as an factor that differentiates the subjective perception

and the real accessible social capital.

Generalized trust has not received much attention in the studies on household

saving. It is regarded as one of the most important attitudinal elements of social

capital and often interpreted in the literature as a measure for generalized morality

2 Available at http://www.worldvaluessurvey.org.

Individual perception of accessible social capital

123

Page 4: Individual perception of accessible social capital and attitude to thrift

(Putnam 1993; Rothstein and Uslaner 2005; Tabellini 2008). Platteau (1994) points

out that generalized morality is ‘‘embedded in the historically-determined cultural

endowment of a society’’. Generalized trust can be seen in the similar way (Guiso

et al. 2004; Tabellini 2008). It is often found to be positive and nurturing, for which

the application of trust is not limited to a small network (Knack and Keefer 1997;

Zak and Knack 2001; Tabellini 2010). In our view, people who are more generally

trusting often expect to have more accessible social capital. Therefore, generalized

trust should reduce the incentives to do precautionary saving.

Based on the above consideration, we propose that individuals with higher

perception of accessible social capital would value thrift less. As mentioned before,

individual perception of accessible social capital can be influenced by both

generalized trust and individual social network. Accordingly, we have the following

testable hypotheses: (1) individuals who are more generally trusting would value

thrift less; (2) individuals who have broader social networks would value thrift less.

3 Data and empirical strategy

We use variables obtained from the WVS which is a compilation of national surveys

concerning a wide variety of topics. The surveys cover five waves (1981–1984,

1989–1993, 1994–1999, 1999–2004, 2005–2007).3 In this paper, we are concerned

only with the data of the latest two waves. The sample consists of 75 countries/

regions.

For the dependent variable, we rely on a survey that focuses on the values

transmitted from parents to children. The variable is based on the following question

from the WVS: ‘‘Here is a list of qualities that children can be encouraged to learn at

home. Which, if any, do you consider to be especially important? Please choose up

to five.’’ The answer reflects the respondent’s attitude to thrift. If one would like to

teach children to value thrift, he/she must think highly of thrift in the first place.

Attitude to thrift is defined as 1 if the respondent is more favorable to the quality

‘‘thrift saving money and things’’ while the value 0 indicates ‘‘not mentioned’’ (the

alternative qualities in the survey question are good manners, politeness and

neatness, independence, hard work, honesty, feeling of responsibility, patience,

imagination, tolerance and respect for other people, leadership, self-control,

determination perseverance, religious faith, unselfishness, obedience, loyalty).

People in different countries have very different attitudes to thrift. The proportion of

individuals who value thrift in each selected country has been presented using the

data of the latest wave (2005–2007), which can help us get a general idea. As we

can see from Fig. 1, among these countries, China has the highest proportion of

individuals who value thrift while UK has the lowest. The proportions of individuals

favorable to thrift are above 0.5 in India and Germany. Meanwhile, we also present

average gross saving rates of the year 2005–2007 in Fig. 1 using World Bank Open

Data. The proportion of individuals favorable to thrift appears to be largely linked

with national saving.

3 A new wave of surveys during 2010–2014 is yet to come.

M.-N. Xu, M.-L. Wang

123

Page 5: Individual perception of accessible social capital and attitude to thrift

For individual perception of accessible social capital, we need to assess both

generalized trust and social network. Individuals who are more generally trusting

would expect more accessible social capital, and thus would value thrift less. The

variable Trust relies on the following question from the WVS: ‘‘Generally speaking,

would you say that most people can be trusted or that you need to be very careful in

dealing with people?’’ Trust is recoded so that the value 1 means the respondent is

more generally trusting and 0 otherwise. To measure one’s social network, we use

Voluntary groups that relies on the following question: ‘‘Please look carefully at the

following list of voluntary organisations and activities and say… which, if any, do

you belong to?’’ Voluntary organizations can be considered as networks (Son and

Lin 2008). Participation in voluntary organizations and activities conduces to the

expansion of individual social network. We divide the answers of the question into

three categories. The first category indicates that the respondent does not participate

in any voluntary organization or activity and Voluntary groups takes the value 1.

The second and third categories are constructed according to the average level of

voluntary organizations and activities answered by the respondents, and the variable

takes the value 2 or 3 respectively. Accordingly, a higher number is associated with

‘‘more voluntary organizations and activities’’. The correlations between Trust and

Voluntary groups, Voluntary groups and Attitude to thrift are 0.0275 and -0.0576

respectively. They are all significant at the 10 % significance level. The correlation

between Trust and Attitude to thrift is -0.0055 and it appears to be statistically

insignificant.

Apart from the explanatory variables, the effects of weakness of family ties are

controlled. It has been pointed out that intra-family transfers of resources are

Fig. 1 Proportion of individuals who value thrift and gross saving rates

Individual perception of accessible social capital

123

Page 6: Individual perception of accessible social capital and attitude to thrift

substantial not only in developing countries (Cigno et al. 2006). We look at three

questions in the WVS for Weakness of family ties (Alesina and Giuliano 2010). The

first question assesses how important family is in one’s life. The second question

asks if one must always love and respect his/her parents, regardless of the qualities

and faults of them. The third question asks if the respondent approves that ‘‘Parents’

duty is to do their best for their children even at the expense of their own well-

being’’. The first principal component of variables corresponding to these three

questions is used in the regressions.

Demographic characteristics are also controlled in the regressions, like age,

gender, income level, education level, etc. The variable Age/10 refers to the

respondent’s age divided by ten. Gender takes the value 0 for male and 1 for female.

We recoded the variable for ‘‘Marital status’’ in the WVS so that Married takes the

value 1 if the respondent is married and 0 means unmarried. We also look at ‘‘Scale

of incomes’’ and divide the answers into two categories according to the levels of

income. The variable Low income takes the value 1 if the respondent belongs to the

‘‘low income’’ category and 0 otherwise. ‘‘Highest educational level attained’’ gives

us information about the respondent’s education level. A higher number refers to a

higher education level. Meanwhile, we expect health to be influential and ‘‘State of

health (subjective)’’ is controlled. The question is ‘‘All in all, how would you

describe your state of health these days?’’ From 1 to 5, a higher number indicates

poorer health. And we recoded the variable for ‘‘How many children do you have’’

in the WVS so that Child takes the value 1 if the respondent has one child or more

and 0 otherwise. Moreover, the effects of several main religious denominations are

controlled, including Catholic, Protestant, Orthodox, Jews, Muslim, Hindu, and

Buddhist. Other refers to other religious denominations besides the ones that have

been listed.

The descriptive statistics of main variables are shown in Table 1.

With regard to cross-country and within-country analyses, the former does not

seem to be proper in this research. Many country-level factors may have influences

on individual attitude to thrift. For instance, better insurance market can make one

less favorable to thrift. Countries differ in many aspects that can contain omitted

variables. Thus, we concentrate on the within-country evidence and control for

country fixed effects to remove the impact of the national institutions and policies.

While drawing inference from within-country data, we must pay attention to the

endogeneity problem as there can be factors affect individual attitude to thrift as

well as generalized trust. We intend to use instrumental variables (IVs). Two

linguistic rules are notable. The first concerns the use of pronoun drop. Some

languages, like Korean, make the use of subjective pronouns optional to the

speakers. Other languages, like English, forbid pronoun drop in the conversation.

Another important linguistic rule concerns the use of multiple types of second-

person pronoun. Some languages differentiate between second-person pronouns,

like T and V types in many Indo-European languages. Others do not have multiple

types of second-person pronoun. According to Kashima and Kashima (1998),

allowing pronoun drop is often associated with a lower level of Individualism. Licht

et al. (2007) use a variable on pronoun drop as an instrument for emphases on

autonomy versus embeddedness. Meanwhile, the use of different second-person

M.-N. Xu, M.-L. Wang

123

Page 7: Individual perception of accessible social capital and attitude to thrift

pronouns is often linked with a higher level of power distance (Kashima and Kashima

1998; Tabellini 2008). These two grammatical rules reflect people’s conception

regarding differentiation between the individual and the context, and also interper-

sonal relationships. Tabellini (2008) uses the linguistic rules as instruments for

generalized trust. We intend to employ the same features of individual ‘‘language at

home’’ in the WVS for instruments. The IVs are to be used in the within-country

regressions, with the precondition that language in different regions can be different

and that language at home and language used in public occasions may differ. The

linguistic rules are valid instruments as they are formed in the long time of history and

unlikely to be correlated with other determinants of an individual’s current attitude to

thrift besides generalized trust. Also, language evolves slowly over time and we do

not need to worry about the problem of reverse causality.

4 Results and discussion

This section presents the results of the regressions using individual level data.

Firstly, as our dependent variable is categorical, we use OLS and probit regressions.

Table 2 shows the influence of individual perception of accessible social capital on

attitude to thrift. Secondly, we turn to IV estimates including the Two-Stage least

squares (2SLS) regression and the IV probit regression. Tables 3 and 4 report the

results.

Table 1 Descriptive statistics

Variables Obs. Mean SD Min Max

Attitude to thrift 32,446 0.3778 0.4849 0 1

Trust 32,446 0.2381 0.4259 0 1

Voluntary groups 32,446 1.8632 0.8050 1 3

Age/10 32,446 4.0059 1.5350 1.5 9.9

Gender 32,446 0.5061 0.5000 0 1

Education level 32,446 4.4736 2.2988 1 8

Low income 32,446 0.5427 0.4982 0 1

Poor health 32,446 2.1814 0.8605 1 5

Married 32,446 0.6031 0.4893 0 1

Child 32,446 0.7400 0.4386 0 1

Catholic 32,446 0.2594 0.4383 0 1

Protestant 32,446 0.0896 0.2856 0 1

Orthodox 32,446 0.0989 0.2986 0 1

Jews 32,446 0.0043 0.0653 0 1

Muslim 32,446 0.1314 0.3379 0 1

Hindu 32,446 0.0477 0.2131 0 1

Buddhist 32,446 0.0298 0.1700 0 1

Other 32,446 0.0865 0.2812 0 1

Individual perception of accessible social capital

123

Page 8: Individual perception of accessible social capital and attitude to thrift

Table 2 Individual perception of accessible social capital and attitude to thrift

(1) OLS (2) Probit (3) Probit (4) Probit (5) Probit

Trust -0.0296***

(0.0062)

-0.0892***

(0.0184)

-0.0330***

(0.0067)

-0.0336***

(0.0067)

Voluntary groups -0.0196***

(0.0042)

-0.0540***

(0.0117)

-0.0202***

(0.0044)

-0.0206***

(0.0044)

Age/10 0.0104***

(0.0021)

0.0310***

(0.0060)

0.0116***

(0.0022)

0.0112***

(0.0022)

0.0112***

(0.0022)

Gender 0.0010

(0.0052)

0.0008

(0.0151)

0.0003

(0.0056)

0.0018

(0.0056)

0.0003

(0.0056)

Education level -0.0151***

(0.0014)

-0.0430***

(0.0039)

-0.0161***

(0.0015)

-0.0171***

(0.0014)

-0.0161***

(0.0015)

Low income 0.0104*

(0.0058)

0.0318*

(0.0170)

0.0119*

(0.0063)

0.0136**

(0.0063)

0.0124**

(0.0063)

Poor health 0.0066**

(0.0033)

0.0181*

(0.0094)

0.0068*

(0.0035)

0.0068*

(0.0035)

0.0076**

(0.0035)

Married 0.0109*

(0.0065)

0.0323*

(0.0191)

0.0120*

(0.0071)

0.0121*

(0.0071)

0.0118*

(0.0071)

Child 0.0057

(0.0077)

0.0185

(0.0229)

0.0069

(0.0085)

0.0072

(0.0085)

0.0072

(0.0085)

Weakness of family ties -0.0057**

(0.0027)

-0.0175**

(0.0081)

-0.0065**

(0.0030)

-0.0070**

(0.0030)

-0.0065**

(0.0030)

Catholic -0.0062

(0.0092)

-0.0220

(0.0272)

-0.0082

(0.0101)

-0.0113

(0.0101)

-0.0085

(0.0101)

Protestant -0.0213*

(0.0114)

-0.0711**

(0.0351)

-0.0262**

(0.0128)

-0.0312**

(0.0127)

-0.0271**

(0.0128)

Orthodox -0.0139

(0.0162)

-0.0431

(0.0436)

-0.0160

(0.0161)

-0.0161

(0.0160)

-0.0155

(0.0161)

Jews -0.0265

(0.0382)

-0.0905

(0.1176)

-0.0332

(0.0423)

-0.0370

(0.0421)

-0.0329

(0.0424)

Muslim 0.0126

(0.0137)

0.0277

(0.0373)

0.0104

(0.0140)

0.0083

(0.0140)

0.0093

(0.0140)

Hindu 0.0282

(0.0218)

0.0731

(0.0580)

0.0276

(0.0221)

0.0268

(0.0221)

0.0264

(0.0221)

Buddhist 0.0004

(0.0163)

0.0072

(0.0510)

0.0027

(0.0191)

-0.0009

(0.0190)

0.0029

(0.0191)

Other -0.0020

(0.0115)

-0.0081

(0.0334)

-0.0030

(0.0124)

-0.0087

(0.0123)

-0.0036

(0.0124)

Number of obs 32446

0.1156

32446

0.0955

32446

0.0955

32446

0.0950

32446

0.0950R2/Pseudo R2

Robust standard errors are in parentheses. Column (3)–(5) report the marginal effects calculated at the

means of the variables

*** Significant at the 1 % level; ** Significant at the 5 % level; * Significant at the 10 % level. All

regressions control for wave and country fixed effects

M.-N. Xu, M.-L. Wang

123

Page 9: Individual perception of accessible social capital and attitude to thrift

Table 3 The first stage

regression of generalized trust

on IVs regarding the use of

multiple types of second-person

pronoun (Sediffer) and whether

it is operative to forbid pronoun

drop (Nprondr): IV probit

Robust standard errors are in

parentheses. The symbols *, **,

*** indicate that the coefficient

is significant, respectively, at the

10, 5, and 1 % level. All

regressions control for wave and

country fixed effects

Trust Trust

Nprondr -0.0355

(0.0404)

Sediffer -0.1517***

(0.0236)

-0.1464***

(0.0224)

Voluntary groups 0.0189***

(0.0049)

0.0189***

(0.0049)

Age/10 0.0132***

(0.0025)

0.0131***

(0.0025)

Gender -0.0018

(0.0062)

-0.0020

(0.0062)

Education level 0.0052***

(0.0016)

0.0052***

(0.0016)

Low income -0.0191***

(0.0072)

-0.0191***

(0.0072)

Poor health -0.0320***

(0.0040)

-0.0319***

(0.0040)

Married 0.0161**

(0.0075)

0.0163**

(0.0075)

Child -0.0014

(0.0087)

-0.0014

(0.0087)

Weakness of family ties 0.0043

(0.0033)

0.0043

(0.0033)

Catholic 0.0122

(0.0111)

0.0117

(0.0111)

Protestant 0.0223

(0.0152)

0.0222

(0.0152)

Orthodox 0.0055

(0.0280)

0.0056

(0.0280)

Jews 0.0462

(0.0430)

0.0459

(0.0430)

Muslim -0.0044

(0.0194)

-0.0043

(0.0194)

Hindu 0.0354

(0.0259)

0.0353

(0.0259)

Buddhist -0.0397

(0.0293)

-0.0395

(0.0294)

Other -0.0085

(0.0144)

-0.0088

(0.0144)

Constant 0.1836*

(0.0954)

0.1785*

(0.0951)

Number of obs 17619 17619

Individual perception of accessible social capital

123

Page 10: Individual perception of accessible social capital and attitude to thrift

Table 4 Results of the IV estimates

2SLS IV Probit

Coefficient Marginal effect

Trust -0.2951*

(0.1584)

-0.8605**

(0.4053)

-0.2763**

(0.1111)

Voluntary groups -0.0152**

(0.0063)

-0.0415**

(0.0189)

-0.0153**

(0.0069)

Age/10 0.0179***

(0.0035)

0.0510***

(0.0084)

0.0188***

(0.0032)

Gender 0.0016

(0.0072)

0.0017

(0.0202)

0.0006

(0.0075)

Education level -0.0135***

(0.0020)

-0.0369***

(0.0068)

-0.0136***

(0.0024)

Low income 0.0019

(0.0089)

0.0038

(0.0251)

0.0014

(0.0092)

Poor health -0.0024

(0.0068)

-0.0084

(0.0189)

-0.0031

(0.0070)

Married 0.0212**

(0.0091)

0.0577**

(0.0249)

0.0212**

(0.0092)

Child -0.0108

(0.0102)

-0.0295

(0.0292)

-0.0109

(0.0108)

Weakness of family ties -0.0001

(0.0037)

-0.0005

(0.0106)

-0.0002

(0.0039)

Catholic 0.0044

(0.0121)

0.0093

(0.0348)

0.0034

(0.0128)

Protestant 0.0233

(0.0176)

0.0638

(0.0492)

0.0238

(0.0185)

Orthodox 0.0180

(0.0359)

0.0408

(0.0920)

0.0151

(0.0344)

Jews 0.0175

(0.0434)

0.0366

(0.1263)

0.0136

(0.0472)

Muslim 0.0223

(0.0260)

0.0504

(0.0689)

0.0187

(0.0258)

Hindu -0.0140

(0.0325)

-0.0404

(0.0889)

-0.0148

(0.0322)

Buddhist -0.0012

(0.0308)

-0.0170

(0.0787)

-0.0062

(0.0288)

Other 0.0090

(0.0164)

0.0225

(0.0468)

0.0083

(0.0173)

M.-N. Xu, M.-L. Wang

123

Page 11: Individual perception of accessible social capital and attitude to thrift

From the results in Table 2, we can see that one who is more generally trusting or

has a broader social network would value thrift less, which accords with our

hypotheses. The coefficients on Trust and Voluntary groups are all negative and

statistically significant at the 1 % significance level. The standardized coefficients

are -0.0260 and -0.0326 respectively for Trust and Voluntary groups. Moreover,

the aged people tend to value thrift more. It is reasonable that aged people have

learned or been forced to treasure thrift. As can be anticipated, individuals with

lower income levels or poorer health are more favorable to thrift. People tend to

value thrift more when they are in need of money. In addition, individuals with

higher education levels tend to value thrift less. The reason may be that people with

higher education have a much larger chance to find a job with good pay. And it is

not surprising that the married people would value thrift more. Meanwhile, the

coefficients on Weakness of family ties are negative and statistically significant at

the 5 % level. Individuals with stronger family ties appear to be more favorable to

thrift. It is reasonable as strong family ties can foster narrow social networks. It also

indicates that stronger family ties may lead to wishing to have more insurance

considering the family. The effects of other independent variables turn out to be

insignificant except for one religion. Although religious denominations are

controlled in all regressions, it is not our intention to discuss religions in this paper.

To address the endogeneity problem, we intend to use IV estimates. Linguistic

rules have been presented to be associated with specific conceptions of people

(Tabellini 2008; Kashima and Kashima 1998). As we focus on the within-country

evidence, ‘‘Language at home’’ in the WVS is noteworthy for generalized trust. It

relies on the question: ‘‘What language do you normally speak at home?’’ Language

evolves slowly and reflects the inherited culture. One’s language at home indicates

the culture that has the largest influence on him/her, especially in the case of

immigrants and multi-language countries. For the instruments, following Tabellini

(2008), we define Nprondr as 1 if it is operative to forbid first-person pronoun drop

in spoken language and 0 otherwise, expecting it to be positively correlated with the

indicator of generalized trust. Furthermore, since the existence of multiple types of

second-person pronoun is often linked with social hierarchy that deemphasizes

generalized trust, we define Sediffer as 1 if multiple types of second-person pronoun

can be used and 0 otherwise. The language data are from Kashima and Kashima

Table 4 continued

2SLS IV Probit

Coefficient Marginal effect

Number of obs 17619 17619 17619

R-squared 0.0515

Robust standard errors are in parentheses. The marginal effects calculated at the means of the variables

are presented

*** Significant at the 1 % level; ** Significant at the 5 % level; * Significant at the 10 % level. All

regressions control for wave and country fixed effects. The centered R-squared is presented

Individual perception of accessible social capital

123

Page 12: Individual perception of accessible social capital and attitude to thrift

(1998) as well as Kashima and Kashima (2005). The first stage regression is

reported in Table 3.

For the first stage regression, we first project Trust on Nprondr and Sediffer. It

turns out that Nprondr does not have significant effects and the result of Hansen J

statistic is not very promising. Then we project Trust on Sediffer. The coefficient on

Sediffer is -0.1464 and statistically significant at the 1 % level, which accords with

our analyses. The results of the IV estimates are displayed in Table 4. The Anderson

canon. corr. LR statistic as part of 2SLS regression is 50.09 with a p value of 0.00.

The Cragg-Donald F statistic is 50.04 which exceeds the critical values of Stock and

Yogo (2002). They all confirm that our instrument is appropriate.

According to Table 4, Trust has a negative coefficient statistically significant at

the 5 % level in the IV probit equation. The coefficient on Voluntary groups is also

negative and significant at the 5 % level. The results accord with our hypotheses.

Additionally, Age/10 has a positive coefficient statistically significant at the 1 %

level. The aged people do value thrift more. The coefficient on Education level is

negative and significant at the 1 % significance level while the coefficient on

Married is positive and significant at the 5 % level. It means that individuals with

lower education and those who are married tend to value thrift more. The effects of

other independent variables turn out to be insignificant.

The signs of the coefficients statistically significant in the IV probit regression

remain unchanged compared with those of the probit regressions. The marginal

effect of Trust in Table 4 is -0.2763. It is relatively larger in absolute value than the

marginal effect using probit estimate. There is a remarkable difference in the size of

the marginal effects. It suggests that generalized trust might be measured with error,

which goes away when Trust is projected on linguistic rules. An alternative

explanation is that the linguistic rules affect individual attitude to thrift also through

other channels instead of generalized trust. For instance, perhaps the linguistic rules

affect individual confidence in the formal institutions or ‘‘self-expression’’ values,

which in turn affects individual attitude to thrift. However, one can hardly have

confidence in the formal institutions of society if he/she only trusts a limited group

of people (Fukuyama 1995; Rothstein and Uslaner 2005). Similarly, generalized

trust is often seen as a proxy for ‘‘self-expression’’ values (Inglehart and Baker

2000; Inglehart et al. 2002). It is plausible that the IV estimates capture the

component of generalized trust that is driven by the cultural legacy of history which

might also be seen within confidence in the formal institutions of society or ‘‘self-

expression’’ values. On the whole, the results do not invalidate our hypotheses.

Individuals with higher perception of accessible social capital appear to be less

favorable to thrift. The null hypothesis of Wald test of exogeneity of trust is

significantly rejected (Prob [v2 = 0.0915).

5 Conclusion

Individual attitude to thrift is an important factor that has macro effects, like impacts

on national savings and economic growth. It differs greatly in different people, even

within the same country. Some people do not value thrift much and would rather

M.-N. Xu, M.-L. Wang

123

Page 13: Individual perception of accessible social capital and attitude to thrift

borrow money whereas others think highly of thrift. Where does this difference

come from?

In this paper we have investigated whether and how one’s subjective perception

of accessible social capital affects his/her attitude to thrift, assuming that social

capital serves as an informal insurance. We estimate individual perception of social

capital on generalized trust and individual social network. Generalized trust can be

regarded as an factor that differentiates the subjective perception and the real

accessible social capital. Even if two persons have the same accessible social

capital, they can still have diverse perceptions. One’s social network is measured by

participation in voluntary organizations and activities. The variables are based on

answers from the World Values Survey. The results confirm that individuals with

higher perception of accessible social capital would value thrift less. The IV

estimates also suggest that our results are robust.

The results provided in this paper show how individual perception of accessible

social capital functions in the formation of different attitudes to thrift. Anyone who

seeks to influence individual attitude to thrift for economic purposes must pay

attention to generalized trust as well as social network. And as an indication for

future researches, one’s subjective perception matters a lot for individual behavior,

which does not necessarily accord with the real world. We can better understand

economic phenomena without treating a person like a ‘‘black box’’ that outputs

certain reaction purely based on input information that is all ‘‘real’’. Diverse

perceptions create different reactions. Individual subjective perception might be

delusional, but the reactions based on these perceptions have real influences. The

factors involved in the formation of individual perception can reveal deeper causes

behind economic phenomena. It may be helpful to pay more attention to individual

subjective perception with regard to household saving or in other economic fields.

Acknowledgments The authors wish to thank Alberto Alesina, Mao-Liang Ye, and Sheng-Chun Zhou

for helpful comments and suggestions. We gratefully acknowledge financial supports from Hangzhou

Normal University (No. 2012QDW106).

References

Agarwal, S., Chomsisengphet, S., & Liu, C. (2011). Consumer bankruptcy and default: The role of

individual social capital. Journal of Economic Psychology, 32, 632–650.

Alesina, A., & Giuliano, P. (2010). The power of the family. Journal of Economic Growth, 15, 93–125.

Anderson, C. L., & Nevitte, N. (2006). Teach your children well: Values of thrift and saving. Journal of

Economic Psychology, 27, 247–261.

Berry, S., Waldron, M., & Williams, R. (2009). Household saving. Bank of England Quarterly Bulletin,

49(3), 191–201.

Bloom, D. E., Canning, D., Graham, B. (2002). Longevity and life cycle savings. NBER Working Paper,

No. 8808.

Chou, S. Y., Liu, J. T., & Hammitt, J. K. (2003). National health insurance and precautionary saving:

Evidence from Taiwan. Journal of Public Economics, 87, 1873–1894.

Cigno, A., Giannelli, G. C., Rosati, F. C., & Vuri, D. (2006). Is there such a thing as a family constitution?

A test based on credit rationing. Review of Economics of the Household, 4(3), 183–204.

Duflo, E., & Saez, E. (2003). The role of information and social interactions in retirement plan decisions:

Evidence from a randomized experiment. Quarterly Journal of Economics, 118(3), 815–842.

Erickson, B. H. (1996). Culture, class, and connections. American Journal of Sociology, 102, 217–251.

Individual perception of accessible social capital

123

Page 14: Individual perception of accessible social capital and attitude to thrift

Fukuyama, F. (1995). Trust: The social virtues and the creation of prosperity. New York: Free Press.

Guiso, L., Sapienza, P., & Zingales, L. (2004). The role of social capital in financial development. The

American Economic Review, 94(3), 526–556.

Guiso, L., Sapienza, P., & Zingales, L. (2006). Does culture affect economic outcomes? Journal of

Economic Perspectives, 20(2), 23–48.

Inglehart, R., & Baker, W. E. (2000). Modernization, cultural change, and the persistence of traditional

values. American Sociological Review, 65, 19–51.

Inglehart, R., Norris, P., & Welzel, C. (2002). Gender equality and democracy. Comparative Sociology, 1,

321–346.

Kantor, S. E., & Fishback, P. V. (1996). Precautionary saving, insurance, and the origins of workers’

compensation. Journal of Political Economy, 104(2), 419–442.

Kashima, E. S., & Kashima, Y. (1998). Culture and language: The case of cultural dimensions and

personal pronoun use. Journal of Cross-Cultural Psychology, 29(3), 461–486.

Kashima, E. S., & Kashima, Y. (2005). Erratum to Kashima and Kashima (1998) and reiteration. Journal

of Cross-Cultural Psychology, 36(3), 396–400.

Knack, S., & Keefer, P. (1997). Does social capital have an economic payoff? A cross-country

investigation. Quarterly Journal of Economics, 112(4), 1251–1288.

Licht, A. N., Goldschmidt, C., & Schwartz, S. H. (2007). Culture rules: The foundations of the rule of law

and other norms of governance. Journal of Comparative Economics, 35, 659–688.

Loayza, N., Schmidt-Hebbel, K., & Serven, L. (2000). What drives private saving across the world?

Review of Economics and Statistics, 82(2), 165–181.

Nyhus, E. K., & Webley, P. (2001). The role of personality in household saving and borrowing behaviour.

European Journal of Personality, 15, S85–S103.

Paxton, P. (2002). Social capital and democracy: an interdependent relationship. American Sociological

Review, 67, 254–277.

Platteau, J. (1994). Behind the market stage where real societies exist - Part I and II. Journal of

Development Studies, 30(3), 533-577, 753-817.

Portes, A. (2000). The two meanings of social capital. Sociological Forum, 15(1), 1–12.

Putnam, R. (1993). Making democracy work: Civic traditions in modern Italy. Princeton: Princeton

University Press.

Putnam, R. (1995a). Bowling alone: America’s declining social capital. Journal of Democracy, 6, 65–78.

Putnam, R. (1995b). Tuning in, tuning out: the strange disappearance of social capital in America. PS:

Political Science and Politics, 28, 664–683.

Rothstein, B., & Uslaner, E. (2005). All for all: Equality, corruption, and social trust. World Politics,

58(1), 41–72.

Schafft, K. A., & Brown, D. L. (2000). Social capital and grassroots development: The case of Roma self-

governance in Hungary. Social Problems, 47, 201–219.

Son, J., & Lin, N. (2008). Social capital and civic action: A network-based approach. Social Science

Research, 37, 330–349.

Stock, J. H., Yogo, M. (2002). Testing for weak instruments in linear IV regression. NBER Technical

Working Paper No. 284.

Tabellini, G. (2008). Institutions and culture. Journal of the European Economic Association, 6, 255–294.

Tabellini, G. (2010). Culture and institutions: Economic development in the regions of Europe. Journal of

the European Economic Association, 8(4), 677–716.

Van Der Gaag, M., & Snijders, T. (2005). The resource generator: Social capital quantification with

concrete items. Social Networks, 27, 1–29.

Wilcox, D. W. (1991). Household spending and saving: Measurement, trends, and analysis. Federal

Reserve Bulletin, 77(1), 1–17.

Zak, P. J., & Knack, S. (2001). Trust and growth. The Economic Journal, 111(470), 295–321.

M.-N. Xu, M.-L. Wang

123