B. Trust and individual income
V. Conclusion and extensions
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
dataset itself, we find that the mean and standard deviation of individual trust is 2.474 and 1.175 which indicates that most observations in the dataset hold the trust believes less than 4. Notice that in the result of equation (22), the estimation parameters for individual trust variables are increasing before the indicators reaching the trust level of 4. Therefore, it is reasonable for the results of equation (19) and (21) to suggest increasing income effect of individual trust as the constraint of OLS estimation that it mainly captures the dynamic of means rather than those extreme values.
Equation (23) add the social trust level in the estimation compared with equation (22), and shows again that the dominating effect of social trust over individual trust.
With different trust variable used, the estimation result of equation (19) and (20), equation (22) and (23) verifies each other.
To conclude the estimation findings of individual data, trust will positively
influence one’s income level. Furthermore, it is the social trust rather than individual trust that dominate the individual’s income level. That is, for individuals living in trusting society, their welfare is generally better off than those living in distrusting society.
V. Conclusion and Extensions
Both the modified RBC model and AK model in this paper present a society with infinitely-lived representative household facing a non-discriminative investment market in which the investment risk is determined by the objective society trust level.
The solutions to the representative household’s optimization problem in both models imply that the individual investment level is positively influenced by the social trust level. Along the positive trust-investment relationship on individual level, with different production function specification, our two models, however, predict distinguishing trust effects on the macro economic performance. For the modified RBC model, with two assumptions that no government purchase and fully
depreciation, the model is solvable and the mathematical solutions indicate that the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
solutions for the constant growth path predict that, on the contrary, trust has both income effect and growth effect.
One thing about the prediction of our two theoretical models, however, needs careful caution. To some extent, the prediction of our model is pre-determined by the model specification itself. As the social trust variable is constructed in the household’s budget constrain, the variable can only affect the individual investment in our models.
Thus, in the modified RBC model, with the assumption of diminishing returns of reproductive factors, the social trust in the model has no growth effect but income effect. Nevertheless, in the AK model, without the constraint of diminishing returns, trust then have both income and growth effect. The two models aim to approach the mechanism of trust but much is sacrificed for the sake of the simplicity for analysis.
Though neither of the model in this paper should be seen as the true model, they shall offer avenues to think about the various ingredients that are behind the dynamic between trust and economic performance in real world.
Using the cross-country data and Chinese province data, our empirical result
demonstrates only income effect and no growth effect of trust which is consistent with the prediction of our modified RBC model. This result is different from the P. Zak and S. Knack’s work (2001), and the possible reason is that they mainly use the OLS method because the constraint of data availability while we are able to construct our panel dataset. Furthermore, the estimation result of cross-country data and Chinese province data verifies each other well in this paper and it shall improve the credibility of our empirical result. In addition, the CGSS survey data is used to investigate the link between individual income and trust. The empirical result implies that personal income is positively related with both individual trust and social trust, while the social trust plays a dominating role.
Apart from the trust-income effect, another important finding in our empirical analysis is that for trust to generate positive revenue in economy, there’s a threshold of human capital for society to meet. A society without enough civilized individuals
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
trust-income relationship can generate positive revenues as people can discipline themselves and not to break promises. Therefore, to maximize the positive revenues of trust, it is important for society to invest in human capital. Only then, the trust-income effect can benefit the economy.
Several extensions for our work can be performed and should be interesting to undertake. For the model part, most importantly, ours is constructed based on a system whose trust level is given exogenously in advance. In the real world, the trust level of a society should be closely related with the penalty for cheating, the
development of the social institutions, and so on. Therefore, it would be beneficial to consider the alternatives that trust is determined endogenously and in turn, affect the economic system. Secondly, only social trust is included in our theoretical model as a budget constraint for the representative household while individual trust is not. The extensions can include individual trust as another important variable into the
individual’s decision-making process and thus may affect the economy’s performance.
Along this line, further extensions should consider other channels that social trust could influence the individual’s decision-making process. In our model, because the social trust affects the expected value of bond, therefore the investment is affected by the social trust level. In the real world, trust not only affect the investment decision but also other aspects, therefore other mechanism of trust to influence economic performance should be considered. Finally, we only use the newest wave of World Value Survey to construct our cross-country dataset, an alternative approach is to organize all waves of WVS data and place different country-level trust data along the time. This approach shall provide a more accurate estimation when testing the trust effect on economic performance.
All in all, the mechanism that trust affect the real world economic performance are much complicated than what we present in this paper and further work shall be done in the future.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
References
Arrow, K. J. 1972. “Gifts and exchanges”. Philosophy & Public Affairs, 343-362.
Berg, J., Dickhaut, J., & McCabe, K. 1995. “Trust, reciprocity, and social history”.
Games and economic behavior, 10(1), 122-142.
Bjørnskov, C. 2007. “Determinants of generalized trust: A cross-country comparison”.
Public choice, 130(1), 1-21.
Butler, J. V., Giuliano, P., & Guiso, L. 2016. “The right amount of trust”. Journal of the European Economic Association, 14(5), 1155-1180.
Coleman, J. S. 1988. “Social capital in the creation of human capital”. American journal of sociology, 94, S95-S120.
Demurger, S. 2001. “Infrastructure development and economic growth: an
explanation for regional disparities in China”. Journal of Comparative economics, 29(1), 95-117.
Hsieh, C. T. 1999. “Productivity growth and factor prices in East Asia”. The American Economic Review, 89(2), 133-138.
Knack, S., & Keefer, P. 1995. “Institutions and economic performance: cross‐country tests using alternative institutional measures”. Economics & Politics, 7(3), 207-227.
Knack, S., & Keefer, P. 1997. “Does social capital have an economic payoff? A cross-country investigation”. The Quarterly journal of economics, 112(4), 1251-1288.
North, D. C. 1990. “Institutions, institutional change and economic performance”.
Cambridge university press.
Porta, R. L., Lopez-De-Silane, F., Shleifer, A., & Vishny, R. W. 1996. “Trust in large organizations” (No. w5864). National Bureau of Economic Research.
Putnam, R. D., Leonardi, R., & Nanetti, R. Y. 1994. “Making democracy work: Civic traditions in modern Italy”. Princeton university press.
Romer, D. 2012. “Advanced macroeconomics”. McGraw-Hill Company. 4th edition.
Smith, V. L. 1998. “The two faces of Adam Smith”. Southern economic journal, 2-19.
Zak, P. J., & Knack, S. 2001. “Trust and growth”. The economic journal, 111(470),
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Appendix:
A. Variable Definitions
Table 7 Definition for All Variables Panel A. Cross-Country Data and Chinese Province Data
Variables Definitions
Growth 5-year average GDP Growth rate for certain country or province.
GDP per capita Log value of GDP per capita level for certain country or province. Log value of 5-year average GDP capita if the variable is the dependent variable. US dollar for cross-country data and Yuan for Chinese data.
Growth of Pop The growth rate of population for certain country or province.
Trust The social trust level calculated using WVS data or CGSS data.
I/Y The investment rate for certain country or province. The variable equals to the share of investment of GDP for certain country or province.
Human Capital The human capital indicator. Variable for cross-country data sources from the Penn world table. Particularly, for Chinese province data, the variable is calculated using the share of the number of college students among the overall population in each province.
TFP The total-factor productivity sources from the Penn World Table.
First Industry Share The share of first industry in the GDP for certain country or province.
Saving/Y The saving rate equals to the share of overall saving against the level of GDP in certain province.
Trust*HC The cross term equals to the level of trust multiplied by the level of human capital.
Solow Residual The variable is the residual that the growth rate of output after subtracting the share-weighted growth in factor quantities.
FDI/Y The variable equals to the share of foreign direct investment against the level of GDP.
Rural Population Share The variable equals to the rural population divided by the overall population in certain province.
Rural Consumption Level/Urban Level
The variables equals to the level of average rural population’s consumption divided by the level of average urban population’s consumption.
Panel B. Chinese Individual Data
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Income The log value of individual’s yearly income.
Individual Trust The reported trust value for the respondent. The variable ranges from “1” to “5”.
Social Trust The variable equals to the mean value of all the individuals’ trust in the same province and the survey wave that the respondent belongs to.
Hukou The status of household registration. Dummy variable that “0” stands for rural residents and “1” stands for urban residents.
Gender Dummy variable that “0” stands for female residents and “1” stands for male residents.
Schooling Year The education level for the respondent.
Average Working Hours per Week The average working hours per week for the respondent.
Working Experience The variable equals to the value of years since the respondent’s first job.
Square of Working Experience The square value of the respondent’s working experience since the first job.
East Region Dummy variable that “0” stands for residents not living in eastern provinces and “1” stands for those living in eastern provinces.
Individual Trust* Social Trust The cross term equals to the level of individual trust multiplied by the level of social trust.
Trust Value=”2”, ”3”, ”4”, ”5” Dummy variable that represents the respondent’s individual trust level.
B. Descriptive Statistics for All Variables
Table 8 Descriptive Statistics for all variables Panel A. Cross-Country Data
Variable Mean SD Min Max
Growth 0.050 0.086 -0.663 1.608
GDP per capita(US dollar) 11485.859 12778.201 515.663 134040.000
Growth of Pop 0.017 0.018 -0.181 0.204
Trust 0.261 0.183 0.028 0.682
I/Y 0.220 0.089 0.007 0.802
Human Capital 2.294 0.597 1.071 3.619
TFP 0.782 0.505 0.145 7.364
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Panel B Chinese Province Data
Variable Mean SD Min Max
Growth 0.118 0.022 0.077 0.174
GDP per capita(Yuan) 32314.632 22614.331 6375 100105
Growth of Pop 0.009 0.017 -0.056 0.058
Trust 2.706 0.715 1.430 3.710
I/Y 0.480 0.118 0.311 0.879
Human Capital 0.017 0.007 0.005 0.036
First Industry Share 0.110 0.069 0.006 0.342
Solow Residual 0.071 0.021 0.011 0.149
Trust*HC 0.048 0.025 0.012 0.113
FDI/Y 0.727 0.157 0.492 1.205
Rural Population Share 0.014 0.012 0.001 0.060
Rural Consumption Level/Urban Level
0.634 0.197 0.120 0.900
Panel C. Chinese Individual Data
Variable Mean SD Min Max
Income 19738.744 30652.840 500 800000
Individual Trust 2.474 1.174 1 5.000
Social Trust 2.498 0.730 1.430 3.960
Hukou 0.645 0.478 0 1
Gender 0.549 0.498 0 1
Schooling Year 10.204 3.867 0 18.500
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Working Experience 20.799 12.697 0 50.000
East Region 0.442 0.497 0 1
Individual Trust* Social Trust 6.676 4.555 1.430 19.800
Trust=2 0.334 0.472 0 1
Trust=3 0.170 0.375 0 1
Trust=4 0.218 0.413 0 1
Trust=5 0.035 0.183 0 1
Sources: Author calculations
‧
C. SAS code for trust simulation Data sim1(keep=tr1);
plot tr1*t tr2*t tr3*t/
overlay;
run;