信任,個人所得與國家財富 - 政大學術集成

全文

(1)國立政治大學經濟學(系)研究所碩士學位論文. 信任,個人所得與國家財富 Trust, Individual Income and Nation’s Wealth. 立. 政治大. ‧. ‧ 國. 學 er. io. sit. y. Nat. al. n. v i n Ch 指導教授：莊奕琦 e n g c h i U博士研究生：吳宗暉撰. 中華民國 106 年 7 月.

(2) Contents I. Introduction ........................................................................................................1 II. The theoretical models........................................................................................4 A. The modified RBC model ..................................................................................4 B. The modified AK model ...................................................................................9 III. Data ................................................................................................................... 12 A. Measuring trust ................................................................................................ 12 B. Other data and estimation model ..................................................................... 17. 治政大 IV. Empirical analysis ............................................................................................ 23 立 A. Trust and nation’s wealth ............................................................................... 23 ‧ 國. 學. B. Trust and individual income ......................................................................... 34 V. Conclusion and extensions................................................................................ 40. ‧. References .................................................................................................................................. 43. n. al. er. io. sit. y. Nat. Appendix..................................................................................................................................... 44. Ch. engchi. i n U. v. i.

(3) Tables Table 1 Descriptive statistics for cross-country and Chinese province data............... 20 Table 2 Descriptive statistics for Chinese individual data .......................................23 Table 3 Estimation result for growth effect of trust ................................................ 24 Table 4 Estimation result for income effect of trust with cross-country data ........... 28 Table 5 Estimation result for income effect of trust with Chinese province data ..... 29 Table 6 Estimation result using CGSS individual survey data ................................. 36 Table 7 Definition for All Variables ........................................................................ 44. 政治大. Table 8 Descriptive Statistics for all variables ........................................................ 45. 立. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. ii.

(4) Trust, Individual Income and Nation’s Wealth. This paper presents two modified models in which social trust is constructed in the representative household’s budget constraint. The solutions of both models predict that trust is positively related with the overall output, while the predictions for the growth effect of trust are ambiguous. Using the cross-country data and Chinese province data, we find our result of empirical analysis suggests that trust should only. 政治大. has income effect but no growth effect. Furthermore, we find there’s an. 立. important mutually enhancing relationship between human capital and. ‧ 國. further verified using the Chinese survey data.. ‧. I.. 學. trust. On individual level, the positive income-trust relationship is. Introduction. y. Nat. sit. Ever since Kenneth Arrow (1972, p. 357) recognizing that “it can be plausibly. al. er. io. argued that much of the economic backwardness in the world can be explained. n. v i n C h reasonable support large amount of literature providing e n g c h i U for Arrow’s conjecture. by the lack of mutual confidence”, economists made considerable efforts with. Traditionally, economists tend to do analysis under the hypothesis of perfectly rational individuals and model economic transactions relying on contracts enforcing rather than mutual trust. Nevertheless, much evidence from social experiments suggests that individuals’ behavior in real world deviates from the prediction of traditional economic analysis. During the anonymous sequential prisoner’s dilemma experiment, which is the typical Nash equilibrium case, half of the first-move players choose to trust their stranger partner and three-quarters of second-move players choose to repay with the. 1.

(5) same trust (Berg, Dickhaut, McCabe, 1995; Smith, 1998). These facts reveals the fact that probably we underestimate the power of trust when trying to economically modeling our real living world in the past. Till now, much efforts has been made to improve the understanding of trust in economic activity and it is generally accepted that trust is positively related with economic performance by reducing transactional costs, providing more incentives to innovate, improving government policy efficiency and by improving the return of human capital (Putnam 1994; Porta et al. 1997; Keefer. 政治大 equilibrium heterogeneous agent growth model to determine the consequences 立. and Knack, 1997). In Zak and Knack’s work (2001), they provide a general. of different levels of trust on economic performance, and their empirical result. ‧ 國. 學. verifies the positive relationship between social trust and economic growth across countries. Apart from the analysis under the macro-level framework,. ‧. economists recently also noticed the individual trust has substantial impact on. Nat. sit. y. personal economic performance. Individual income is hump-shaped in a. io. al. er. measure of intensity of trust beliefs (Butler, Giuliano, Guiso, 2016). Highly trusting individuals has higher risk regarding being cheated while too much. n. v i n pessimistic attitude results inCtendency to give up risky h e n g c h i U profit opportunities.. Though both on the country level and individual level, trust has been tested. and confirmed to be positive related with economic performance, there’s still little research concerning the subject that to build a bridge to combine both the macro and micro level results systematically. Thus the latter blank is the focus of this paper. This paper presents a modified RBC model and a modified AK model to approach the mechanism of trust influencing the economic performance in real world. In both models, trust is introduced through a non-discriminative, risky. 2.

(6) investment market aiming to influence the economic performance. In this investment market, a fully trusted society will repay the principal and all the revenues, while a poorly trusted society indicates that large risk for households to lose all the money. We permit households to choose the degree they invest based on the known social trust level and in this way, the individual investment decision and, in turn, the performance of the economy is characterized by the trust variable. In both models, we show that people tend to invest more with higher social trust level. As a result, higher social trust level predicts both a. 政治大 However, the growth effect of trust is ambiguous with the two models’ 立 better welfare for individuals and a better overall output level for society.. model indicate positive growth effect of trust.. 學. ‧ 國. distinguishing results. The RBC model suggest no growth effect while the AK. Following the predictions of both models, in the latter part of this paper, we. ‧. provide empirical analysis to investigate the role of trust in real world economy.. Nat. sit. y. We use both Chinese province data and cross-country data to estimate the. io. al. er. macro-level relationship between trust and economic performance. The advantage to introduce Chinese province data, is that it avoid the concern of. n. v i n C which translation bias in the survey reported in cross-country U h e nareg commonly i h c research, and it also provides stronger causality implication as these provinces share a quite similar culture and thus less unidentified interferences than crosscountry data. Because the models we introduced provide insights of both growth effect and income effect of trust, we investigate both the links between trust and economic growth, income level. The result of our empirical analysis suggests that trust has only income effect and no growth effect, which is consistent with the predictions of our modified RBC model. For the individual level test, we use Chinese individual data to test both how individual and social trust affect. 3.

(7) each person’s income. Through the empirical analysis, it is verified that trust has a significantly positive impact on individual income. This paper is structured as follows. In Section II we present the two modified models and their implications. In Section III, we describe the survey data and present basic descriptive results. In Section IV we test the models’ predictions with empirical analysis both on macro and micro economic performance level, and finally, Section V concludes.. 政治大 A. The Modified RBC Model 立 II.. The Theoretical Models. Consider an economy consists of a large number of identical, price-taking firms. ‧ 國. 學. and a large number of identical, price taking, infinitely lived households. The inputs to production are capital (K), labor (L), and technology (A). Each period,. ‧. a constant growth rate 𝑔 and a disturbance constitute the growth of the. Nat. (2). sit er. al. n. (1). io. 𝑡 is. y. technology. The production function is Cobb-Douglas, and the output in period. 𝑌𝑡 = 𝐾𝑡𝛼 (𝐴𝑡 𝐿𝑡 )1−𝛼 , 0 < 𝛼 < 1,. Ch. i n U. 𝑙𝑛𝐴𝑡 = 𝐴̅ + 𝑔𝑡 + 𝐴̃𝑡 .. engchi. v. Output is divided among consumption (C), investment (I), and government purchase (G). Each period, fraction 𝛿 of capital is depreciated. Therefore, the capital stock in each period 𝑡 + 1 is (3). 𝐾𝑡+1 = 𝐾𝑡 + 𝐼𝑡 − 𝛿𝐾𝑡 .. For firms are price-taking participants in the perfect competitive market, both labor and capital are paid with their marginal products. Thus the real wage and interest rate are (4). 𝐾. 𝛼. 𝑤𝑡 = (1 − 𝛼 ) (𝐴 𝐿𝑡 ) 𝐴𝑡 , 𝑡 𝑡. 4.

(8) 𝑟𝑡 = 𝛼(. (5). 𝐴𝑡 𝐿𝑡 1−𝛼 ) 𝐾𝑡. − 𝛿.. The representative household maximizes their expected value of utility function, 𝑁. (6). −𝜌𝑡 𝑈 = ∑∞ 𝑢(𝑐𝑡 , 1 − 𝑙𝑡 ) 𝐻𝑡 . 𝑡=0 𝑒. (7). 𝑢𝑡 = 𝑙𝑛𝑐𝑡 + 𝑏𝑙𝑛(1 − 𝑙𝑡 ), 𝑏 > 0. 𝑢(∎) is the instantaneous utility function of the representative member of the household, and 𝜌 is the discount rate. 𝑐𝑡 is the consumption per member. 政治大. for the household, and 𝑙𝑡 is the labor supply per member (for simplicity, the time endowment per member is normalized to 1). 𝑁𝑡 is population and 𝐻 is. 立. the number of households in the economy; therefore. 𝑁𝑡 𝐻𝑡. is the number of. ‧ 國. 學. members of the household. For simplicity, we assume in the model that. 𝑁𝑡 𝐻𝑡. = 1,. ‧. i.e. only one member in each household. Population grows exogenously at a. Nat. ̅ + 𝑛, 𝑛 < 𝜌. 𝑙𝑛𝑁𝑡 = 𝑁. sit. (8). y. constant rate 𝑛:. n. al. er. io. Assume all households can participate in a non-discriminative bond market. i n U. v. with the help of agents，thus households can allocate their income either on. Ch. engchi. current consumption for instant utility at period 𝑡, or they can invest a fraction of their income in the bond market for future consumption. Nevertheless, it is assumed that the bond market is not risk-free and agents in the bond market may choose to cheat their clients with the probability 𝜇 (0 ≤ 𝜇 ≤ 1). We consider the probability of successful investment 1 − 𝜇 as a proxy for the society’s objective trust level and assume the representative household has the full information about the overall risk in the bond market. The intuition for this hypothesis is straightforward that other things being equal, with higher social trust, it is reasonable for households to have better opportunity to gain profit. 5.

(9) from investment and they should tend to invest more; thus the overall investment probability should reflect the social trust level. In a fully trusted society, no clients in the bond market will face the risk being cheated and thus 𝜇 = 0. And in a society everyone distrust each other, consumers entering the market will always get cheated by their agents and 𝜇 = 1. For a normal society in real life, the probability will range from 0 to 1, and the higher the cheating probability, the lower the trust level of the society. For each household facing the risk in the bond market, their expected value of the bond revenue equals to. 政治大 the bond at the beginning of period 𝑡. Therefore, as a representative household, 立. (1 − 𝜇)(1 + 𝑟𝑡 )𝑏𝑡 , where 𝑏𝑡 denotes the amount of money household paid for. the budget constraint for each period is. ‧ 國. (10). 學. (9). 𝑐𝑡 + 𝑏𝑡+1 = (1 − μ)(1 + 𝑟𝑡 )𝑏𝑡 + 𝑤𝑡 𝑙𝑡 ,. 𝑐𝑡+1 + 𝑏𝑡+2 = (1 − μ)(1 + 𝑟𝑡+1 )𝑏𝑡+1 + 𝑤𝑡+1 𝑙𝑡+1. ‧. ……. Nat. sit. y. Substitute the term 𝑏𝑡+1 in equation (9) with equation (10) and rewrite the. io. al. n. is. er. budget constraint. For one representative household, the maximization problem. Ch ∞ engchi. iv n +U 𝑏𝑙𝑛(1 − 𝑙 )],. (11). Max U = ∑𝑡=0 𝑒 −𝜌𝑡 [𝑙𝑛𝑐𝑡. (12). 𝑡 𝑠. 𝑡 ∑𝑡=0 (1−μ)𝑡(1+𝑟 = ∑𝑡=0 (1−μ)𝑡 𝑡 )𝑡. ∞. 𝑡. ∞. 𝑙. 𝑤. (1+𝑟𝑡 )𝑡. 𝑡. .. Therefore, the first order conditions for 𝑐𝑡 and 𝑙𝑡 , respectively, are (13) (14). 1 𝑐𝑡. = 𝑒 −𝜌 (1 − μ)𝐸𝑡 𝑐. 1. 𝑡+1. 𝑤𝑡 𝑐𝑡. (1 + 𝑟𝑡+1 ),. 𝑏. = 1−𝑙 . 𝑡. Equation (13) shows that for household’s optimal decision, current consumption is positively related with the cheating probability, and thus negatively related with the society’s trust level. With lower trust level and higher. 6.

(10) cheating probability, rational household will tends to avoid to invest in the bond market and to increase current consumption compared with a high trust-level society. As under the framework of RBC model that society can only accumulate capital through private investment, the implication is rather straightforward that economy will suffer from underinvestment in a low trustlevel society. The model above however cannot be further solved as it contains a mixture of linear and log-linear ingredients. To simplify the situation, we make two. 政治大 depreciation in each period. Therefore, the evolution of capital stock and the 立. changes to the model: we eliminate government purchase and assume that fully. 𝐾𝑡+1 = 𝑌𝑡 − 𝐶𝑡 , 𝑟𝑡 = 𝛼(. 𝐴𝑡 𝐿𝑡 1−𝛼 ) 𝐾𝑡. − 1.. ‧. (16). 學. (15). ‧ 國. determination of the real interest rate follows the pattern. Consider 𝑠𝑡 denotes the saving rate for a representative household. Because. y. Nat. 𝑌. Ch. i n U. 𝑐𝑡 = (1 − 𝑠𝑡 ) 𝑁𝑡 .. engchi. er. al. n. (17). io. have. sit. household can only save to invest or to consume instantly in the period 𝑡, we. 𝑡. v. Substitute both equation (16) and (17) into the household optimization condition equation (13), we have (18). 𝑌. − ln [(1 − 𝑠𝑡 ) 𝑁𝑡 ] = −ρ + 𝑙𝑛𝐸𝑡 𝑡. 𝑌 α(1−μ)( 𝑡+1⁄𝐾 ) 𝑡+1 , 𝑌 (1−𝑠𝑡+1 )( 𝑡+1⁄𝑁 ) 𝑡+1. or (19). −𝑙𝑛(1 − 𝑠𝑡 ) = −ρ + lnα + ln(1 − μ) + n − 𝑙𝑛𝐸𝑡 (1 − 𝑠𝑡+1 ) − 𝑙𝑛𝑠𝑡 .. For equation (19) has to holds for every time period 𝑡, there must be a constant saving rate 𝑠̅ that satisfies the condition. Therefore, substitute both 𝑠𝑡 and 𝑠𝑡+1 with 𝑠̅, the equation (19) becomes. 7.

(11) ln(𝑠̅) = 𝑙𝑛𝛼 + ln(1 − 𝜇) + 𝑛 − 𝜌,. (20) or. 𝑠̅ = (𝛼 + 1 − 𝜇)𝑒 𝑛−𝜌 .. (21). Now consider equation (14), as 𝑐𝑡 = 𝐶𝑡 /𝑁𝑡 = (1 − 𝑠̅)𝑌𝑡 /𝑁𝑡 , we can compute that 1−α. 1−α. 𝑙𝑡 = 𝑙 = (1−α)𝑏(1−𝑠) = (1−α)𝑏[1−(𝑙𝑛𝛼+ln(1−𝜇)+𝑛−𝜌)].. (22). Equation (21) and (22) have important implications how trust affect the consumption and labor supply of the representative household. Both equations. 政治大. indicate that the representative household tends to save more and work more. 立. with higher trust level of the society. As households’ income can only come. ‧ 國. 學. from wages and bond revenues, a society with higher trust level implies better income level for households. In turn, as only households provide investment. ‧. and labor supply for the firms, the production of the economy will benefit from higher trust level.. y. Nat. al. er. io. 𝐿𝑡 = 𝑙𝑡 𝑁𝑡 into the production function; thus. sit. To see the effect of trust on the economy directly, substitute 𝐾𝑡 = 𝑠̅𝑌𝑡−1 and. v i n C for one period and combine Backward the equation (23) h both equation (23) engchi U 𝑙𝑛𝑌𝑡 = 𝛼𝑙𝑛𝑠̅ + 𝛼𝑙𝑛𝑌𝑡−1 + (1 − 𝛼)(𝑙𝑛𝐴𝑡 + 𝑙𝑛𝑙 ̅ + 𝑙𝑛𝑁𝑡 ).. n. (23). and (24); let 𝑔𝑡𝑌 denotes the growth rate of production in period 𝑡, thus (24). 𝑙𝑛𝑌𝑡−1 = 𝛼𝑙𝑛𝑠̅ + 𝛼𝑙𝑛𝑌𝑡−2 + (1 − 𝛼)(𝑙𝑛𝐴𝑡−1 + 𝑙𝑛𝑙 ̅ + 𝑙𝑛𝑁𝑡−1 ),. (25). 𝑙𝑛𝑌𝑡 − 𝑙𝑛𝑌𝑡−1 = ln(1 + 𝑔𝑡𝑌 ) = 𝛼(𝑙𝑛𝑌𝑡−1 − 𝑙𝑛𝑌𝑡−2 ) + (1 − 𝛼)(𝐴̃𝑡 −. 𝐴̃𝑡−1 + 𝑔 + 𝑛). Equation (23) suggests that the level of output is positively related with both the saving rate and household’s labor supply, thus is positively related with a society’s trust level. However, equation (25) shows that the growth rate of the output is only affected by the growth rate of population and technology, as. 8.

(12) both the output level in period 𝑡 − 1 and 𝑡 − 2 will be certain historical information when predicting the economic performance in period 𝑡. Both equations suggest that trust has only level effect but no growth effect on the output of the economy. A society with high trust level will experience better output level and thus better welfare for the households. Nevertheless, the growth rate of the output will not be affected by the trust level of certain economy.. 政治大 The pitfall with the modified RBC model is that the result of no growth effect 立 B. The modified AK Model. of personal investment, thus our trust indicator, is ultimately determined by the. ‧ 國. 學. assumption of diminishing returns of the reproductive factors embedded in the production function. Thus it would be arbitrary to exclude the possibility of. ‧. growth effect from trust. In this part, we provide a simply modified AK model. Nat. sit. y. with household optimization that trust actually provide certain growth effect.. io. al. er. Consider an economy consists of a large number of identical, price taking, infinitely lived households. The inputs to production are period-changing. n. v i n capital (K), and technology C (A) which is assumed U h e n g c h i constant. The production function is a linear function of the capital stock that, (26). 𝑌𝑡 = 𝐴𝐾𝑡 , (A>0, A is constant).. In the economy, output is divided among consumption (C) and investment (I). Each period, faction 𝛿 of capital stock is depreciated. In contrast with the RBC model, because labor is no longer an input factor in the production function, the representative household optimize their lifelong welfare only by properly allocating their income between investment and consumption in each period. The utility function for each representative. 9.

(13) household is set as ∞. 𝑈 = ∑𝑡=0 𝛽 𝑡. (27). 𝑐𝑡1−𝛼 −1 𝑁𝑡 1−𝛼. 𝐻. .. In the utility function, 𝛽 is the discount rate and 𝑐𝑡 is the consumption per member for the household. 𝑁𝑡 is population and 𝐻 is the number of. households in the economy; therefore. 𝑁𝑡 𝐻𝑡. is the number of members of the. household. Similar as the RBC model, we assume. 𝑁𝑡 𝐻𝑡. = 1 for simplicity. Following our trust setting in the RBC model, we hence assume that. 政治大 agents with probability of 𝜇 (0 ≤ 𝜇 ≤ 1). Thus their expected total income 立 ( ). households face a risky investment market where they will be cheated by their. equals to 1 − 𝜇 [𝑦𝑡 + (1 − 𝛿)𝑘𝑡 ], where 𝑦𝑡 is the per capita income from. ‧ 國. 學. (1 − 𝛿)𝑘𝑡 is the capital stock after deducting the. investment and. deprecation. The budget constrain of the representative household for each. 𝑘𝑡+1 + 𝑐𝑡 = (1 − 𝜇)[𝑦𝑡 + (1 − 𝛿)𝑘𝑡 ].. sit. y. Nat. (28). ‧. period is. io. al. ∞. n. (29). Max 𝑈 = ∑𝑡=0 𝛽 𝑡. Ch. engchi. er. For a representative household, the maximization problem is 𝑐𝑡1−𝛼 −1. ,. i n U 1−𝛼. v. s.t 𝑘𝑡+1 + 𝑐𝑡 = (1 − 𝜇)[𝑦𝑡 + (1 − 𝛿)𝑘𝑡 ].. (30). Use the Lagrangian to solve the optimization problem, and the first order conditions for 𝑐𝑡 and 𝑘𝑡+1 , respectively, are (31). ∞. 𝑐𝑡1−𝛼 −1. ℒ = ∑𝑡=0 𝛽 𝑡 {. 1−𝛼. 𝜕ℒ. (32) (33). + 𝜆𝑡 [𝑘𝑡+1 + 𝑐𝑡 − (1 − 𝜇)(𝐴𝑘𝑡 + (1 − 𝛿)𝑘𝑡 )]}, 𝜕𝑐𝑡. 𝜕ℒ 𝜕𝑘𝑡+1. = 𝑐𝑡−𝛼 + 𝜆𝑡 = 0,. = 𝜆𝑡 − 𝜆𝑡+1 𝛽(1 − 𝜇)[𝐴 + (1 − 𝛿 )]=0.. Combing both equation (32) and (33), we have the necessary condition for household optimization that. 10.

(14) (. (34). 𝑐𝑡+1 𝛼 ) 𝑐𝑡. = 𝛽(1 − 𝜇)[𝐴 + (1 − 𝛿 )].. The result of equation (34) shows that with higher social trust, i.e. higher value of (1 − 𝜇), the share of consumption in the next period against the current consumption is higher. Because the technology level is fixed in our production function, the only way to increase future consumption for representative households is to sacrifice their current consumption and invest more. Hence, the implication of equation (34) is consistent with our findings in the modified RBC model that households tends to invest more in a society. 政治大. with high trust.. 立. For the simplicity of our analysis, we only solve for the constant growth 𝑦. ‧ 國. 學. path so we don’t need full policy functions. Let 𝑔𝑡 , 𝑔𝑡𝑐 , and 𝑔𝑡𝑘 to denote the growth rate of per capita income, consumption and investment. With the. 𝑐𝑡+1. sit. Nat. 𝑐𝑡. 1. ) − 1 = {𝛽 (1 − 𝜇)[𝐴 + (1 − 𝛿 )]}𝛼 − 1.. y. 𝑔𝑡𝑐 = (. (35). ‧. condition from equation (34), we have the consumption growth rate equals to. Because the capital stock per capita in our model is solely determined by. io. n. al. er. the investment and deprecation, we have the physical capital growth rate at 𝑔𝑡𝑘 =. (36). 𝑘𝑡+1−𝑘𝑡. C 𝑘h 𝑡. =. 𝐴𝑘𝑡 −𝑐𝑡 −𝛿𝑘𝑡. e n g c𝑘 h i 𝑡. i v𝑐 n U. = 𝐴 − 𝑘𝑡 − 𝛿. 𝑡. Notice that under the assumption of constant growth rate of capital, the value of. 𝑐𝑡 𝑘𝑡. needs to be constant. Therefore, in a stationary state growth path, 𝑦. 𝑔𝑡𝑐 = 𝑔𝑡𝑘 = 𝑔𝑡 = 𝑔∗. Combining both the equation of (35) and (36), we solve the value of (37). 𝑐𝑡 𝑘𝑡. that 𝑐𝑡 𝑘𝑡. 1. 𝑐. = 𝑘 = 𝐴 − 𝛿 − {𝛽 (1 − 𝜇)[𝐴 + (1 − 𝛿 )]}𝛼 + 1.. Hence, the growth rate of physical capital and income per capita equals to (38). 𝑦. 𝑐. 1. 𝑔𝑡𝑘 = 𝑔𝑡 = 𝐴 − 𝑘 − 𝛿 = {𝛽 (1 − 𝜇)[𝐴 + (1 − 𝛿 )]}𝛼 − 1 = 𝑔𝑡𝑐 .. 11.

(15) The result of equation (38) suggests that with higher social trust level, the accumulation of physical capital and more importantly, the economy grow faster, i.e. that trust has growth effect. As the output of the economy is a linear function of the level of physical capital under the specification of AK model, the growth effect of trust will also promise the income effect. Notice the implication of equation (38) is dramatically different from the result of our modified RBC model. The mathematical results of our modified AK model suggest that trust has both output and growth effect, while the modified RBC. 政治大. model indicates that trust should only have income effect.. 立. To sum up, the two models in which households face risky investment. ‧ 國. 學. market provide several important insights about the economic effect of trust that needs to be tested in our later empirical analysis:. ‧. a. Both model suggest that a higher trust level will encourage households. Nat. io. al. er. benefit the households with better welfare level.. sit. y. to increase private investment. In the long run, such decision will. b. Both model suggest that an economy with higher trust level, benefit. n. v i n C h will produceUa higher overall output, i.e. from higher private investment, engchi that trust should have income effect.. c. The growth effect of trust is ambiguous. The modified RBC model suggests that growth rate of the output of certain economy is not affected by the trust level in that society, while the modified AK model, on the contrary, indicates that trust has growth effect.. III.. Data. A. Measuring Trust. 12.

(16) The most important step in testing the prediction of the model is to identify a reliable measure of trust. For cross-country data, we use the measure based on the data from the sixth wave of World Value Survey (WVS) conducted from year 2010 to year 2014. The advantage to use the latest wave of the WVS data is that we can includes as many as the number of the countries in our estimation and it helps to see the general trust effect on economy in different countries, the drawback is obviously that it may mess with the causality in the equation with the time lag between trust indicator and countries’ economic. 政治大 less sensitive to time, or in other words, generalized trust is stable over time 立 performance. Nevertheless, literature in the past cautiously consider trust is. (Bjørnskov, 2007); in Keefer and Knack’s estimation (1997, p. 1267), they. ‧ 國. 學. find “ trust value for 1980 and 1990 are correlated at .91” and “changes in trust over the decade are uncorrelated with growth rates”. Thus these findings. ‧. Nat. sit. certain society in a quite large time scale.. y. imply that trust however can be viewed as a slowly-changed characteristic of. io. al. er. The question used to access the level of trust for a certain country is:” Generally speaking, would you say that most people can be trusted or that you. n. v i n C h with people”. TheUtrust indicator we use in need to be very careful in dealing engchi the paper is the percentage the respondents in each country replying “Most people can be trusted”. This trust indicator we use follows Zak and Knack’s framework when testing the trust’s growth effect in their 2001’s work, and since then many other researches concerning the subject of general trust also. pick the same trust indicator. Furthermore, Knack and Keefer (1997) provide empirical support for the validity of these data and find that values for trust is consistent with lab experiment results and case study across countries. For Chinese data, we use the measure based on the data form the Chinese. 13.

(17) General Social Survey (CGSS) conducted in year 2003, 2005, 2010 and 2013. The question used in the survey to measure trust level is that “Generally speaking, how do you trust the strangers” in the year 2003 and 2005; and for year 2010 and 2013, the question is “Generally speaking, would you say that most people can be trusted”. Different from the World Value Survey, the Chinese General Social Survey measures the intensity of individual’s trust beliefs on a scale from 1 to 5, in which “1” means that strongly distrust and “5” means strongly trust. With this trust intensity data, we are able to estimate. 政治大 mean value of individuals’ trust in each province as the trust indicator on 立. the relationship between trust and individual income. Furthermore, we use the. our theoretical models predict.. 學. ‧ 國. province-level and then test the trust effect on macro economy performance as. One problem with using survey data to represent and compute society trust. ‧. level is that the survey can only capture the subjective trust level of. Nat. sit. y. individuals. Though using mean value can predict the average trust level for. io. al. er. all individuals in the society by eliminating the impact of certain individual’s subjective deviation from the society trust, it is still necessary to verify that if. n. v i n C h trust is a good proxy the mean of subjective individual for the overall society’s engchi U objective trust level. Aiming this, we provide a Monte Carlo simulation based on a simplified trust model to test the relationship between the individuals’ subjective mean trust value and the society’s objective trust level. In order to realistically imitate the mechanism of how real social trust works, the simulation model includes the randomness of the individual’s trust, the way subjective social trust is established, and the mutually adjustment of trust levels between society and individuals. In the trust model for the purpose of simulation, we use a uniform distribution between 0 and 1 to randomly. 14.

(18) generate the initial trust level of individuals. The subjective trust level, is calculated using the individuals’ average. And the key adjustment mechanism is represented with a one-time Bernoulli trial as an approximation for the individual investment. The simulation starts with only one individual. The system first generates a random number between 0 and 1 for the individual to represent the born trust level. Thus at the start, the subjective social trust equals to the individual trust, as it is calculated as the individuals’ average in the whole society. Later the. 政治大 equals to the objective social trust given in advance. Recall the models in 立. first individual do the one-time Bernoulli trial with the success probability. Section II, as all households face a non-discriminative investment market, the. ‧ 國. 學. probability for successive investment equals to the objective trust level, or more specifically 1 − 𝜇. If the investment trial succeeds, the individual will. ‧. increase his trust but no more than the ceiling limit of the uniform distribution.. Nat. sit. y. The process of the increase will be generated by the system randomly within. io. al. er. the uniform distribution. If the investment trial fails, the individual will decrease his trust level randomly, in the same way with a result of success. n. v i n C hof the individual’sUtrust, the subjective social investment. With the adjustment engchi trust will in turn be calculated again. Following the same process, the system includes more individual for one person each time and repeat the loop continuously. Under this framework, we are able to investigate the dynamic of subjective social trust and compare it with the exogenous objective trust level. Figure 1 shows the logic of a single loop for the process,. 15.

(19) I. Individual Trust. V. Subjective Social Trust Adjustment. IV. Individual Adjustment. II. Subjecitve Social Trust. III. Bernoulli trial. 政治大. Figure 1 the Simulation Process. Figure 2 shows the simulation results with the success probability value, i.e.. 立. the objective social trust, set at 0.2, 0.6 and 0.9. The line from bottom to top. ‧ 國. 學. are the results for success values set at 0.2, 0.6 and 0.9, showing the dynamic values of subjective social trust level during the simulation. The results show. ‧. that the system equilibrium for average subjective trust level is critically. y. Nat. determined by the success level. With higher success level or social trust level,. io. sit. the mean value of individuals’ subjective trust tends to be higher than it in a. n. al. er. low social trust level society. The confusion of the result is that the average. Ch. i n U. v. subjective though accurately reflect the rank of the objective trust, the value of. engchi. the average subjective trust is not exactly the same as the objective trust given. This problem lies in the fact that for the simplicity of the simulation, the born trust level of individuals is randomly generated using the uniform distribution and not affected by the objective social trust; that is to say, only the mutual relationship between individuals and social trust is included in the simulation, but the mutual relationship between individuals is neglected. Anyway, the focus of the simulation is to test whether average subjective trust can be a proxy for the objective trust, and as long as the rank of trust is not affected, it is reasonable to use mean value of individuals’ subjective trust survey data as. 16.

(20) an approximation for the objective social trust level. Therefore, the trust data from both WVS and CGSS can be confidently used in our estimation.. Social trust=0.9 Social trust=0.6 Social trust=0.2. Socal trust=0.9. 立. 政治大. ‧. ‧ 國. 學 er. io. sit. y. Nat. Figure 2 Simulation Result for Different Social Trust Level. n. al. i n U. Notes: From bottom to top, the values of social trust are set at 0.2, 0.6 and 0.9.. Ch. engchi. v. B. Other Data and Estimation Model. To investigate the trust effect on the macro economic performance, we use both cross-country data and Chinese province data to test the predictions of our theoretical models. The cross-country data used in the paper is an unbalanced panel data covering 57 countries from year 1950 to year 2011. Except that the trust level data for each country is derived from the World Value Survey, all other data for cross-country analysis comes from the Penn World Table Database. The Chinese province data is an unbalanced panel data. 17.

(21) covering 19 provinces with the period in year 2003, year 2005, year 2010 and year 2013. Only provinces in eastern, middle and north-eastern region are included in the estimation dataset. The reason to exclude western provinces is that the economic development strategy in these provinces are highly natural resourced-oriented and government investment depended. This kind of economic development strategy deviates greatly from the generally accepted economics model, and quite different from the market-oriented economic development in which trust can make a difference, thus will surely cause. 政治大 cross-country dataset, the trust level data on province level are derived from 立. unnecessary confusion in our estimation result if included. Similar to the. the survey data, more specifically, the Chinese General Social Survey. All. ‧ 國. 學. other economic related data sources from the National Bureau of Statistics of China (NBSC).. ‧. In order to test both the growth and income effect of trust, the primary. Nat. sit. y. variables are the 5-year average GDP growth rate and the 5-year average GDP. io. al. er. per capita. The reason to use 5-year mean value for both growth rate and per capita income, is that it provides a time lag between the dependent economic. n. v i n C hindependent explanatory performance variable, and other variables. Therefore, engchi U the causality in the estimation result is clearly displayed rather than using oneyear dependent economic performance variable. Following the production form in the model, we set the basic regression equation as 𝑌𝑖,𝑡 = 𝑐 + 𝛽𝑇 𝑇𝑟𝑢𝑠𝑡𝑖,𝑡 + 𝛽𝑦 𝑦𝑖,𝑡 + 𝛽𝑝𝑜𝑝 𝑝𝑜𝑝𝑖,𝑡 + 𝛽𝐼 𝐼𝑖,𝑡 + 𝛽𝐻𝐶 𝐻𝐶𝑖,𝑡 +. 𝛽𝑇𝐹𝑃 𝑇𝑒𝑐𝑖,𝑡 + 𝜎𝑖 + 𝜏𝑖,𝑡 . To test the growth effect of trust, 𝑌𝑖,𝑡 denotes the 5-year average GDP growth rate at time 𝑡 for country or province 𝑖 depending on the dataset. 18.

(22) used. And to test the income effect of trust, 𝑌𝑖,𝑡 denotes the 5-year average GDP per capita. The variable 𝑇𝑟𝑢𝑠𝑡𝑖,𝑡 denotes the mean value of individual trust survey data at time 𝑡 for country or province 𝑖. The variable 𝑦𝑖,𝑡 , 𝑝𝑜𝑝𝑖,𝑡 , 𝐼𝑖,𝑡 , 𝐻𝐶𝑖,𝑡 , 𝑇𝑒𝑐𝑖,𝑡 denote the GDP per capita, the growth rate of population, the investment rate, the human capital indicator and the technology level. Notice that the variable of GDP per capita and the growth rate of population are only included as explanatory variable when testing the growth effect. While testing the income effect, the variable GDP per capita is. 政治大 these variables are offered or can be computed from the Penn World Table 立. the dependent variable in the estimation process. For cross-country data, all. database directly. However, for Chinese province data, no commonly used. ‧ 國. 學. human capital indicator, average population schooling years for example, is available on province level. Therefore, we use the share of the number of. ‧. college students among the overall population in each province as an. Nat. sit. y. approximation for human capital indicator. Using this proxy implies a very. io. al. er. strong assumption that all college students studying in certain province will stay and work in the same province to provide the benefit of their human. n. v i n C h the human capital capital, and thus should overestimate stock in the province. engchi U Particularly, we include the share of first industry of GDP in our Chinese province estimation. The reason is that the economic gap between agriculture and non-agriculture provinces in China is substantial, thus agriculture provinces may experience fewer productivity growth opportunities than industrial provinces in China (Demurger, 2001). Notice that for cross-country data, only one wave of WVS data is added as the trust variable, thus the panel dataset can only be estimated using the random effects model. Otherwise, the same trust variable for each country will. 19.

(23) be automatically dropped if try fixed effects model. For the Chinese province data, the result of Hausman test is 0.0036 for the base model estimating the growth effect, and 0.0105 estimating the income effect. Both values are less than the commonly used critical value 0.05. Therefore, all regression related with Chinese province data are estimated using the fixed effects model. According to the past literature results and the prediction of our model, the expected sighs of parameters are 𝛽𝑇 >, 𝑜𝑟 = 0, 𝛽𝑦 < 0, 𝛽𝑝𝑜𝑝 > 0, 𝛽𝐼 > 0, 𝛽𝐻𝐶 > 0 and 𝛽𝑇𝐹𝑃 > 0 when testing the growth effect of trust; and 𝛽𝑇 > 0,. 政治大 model predicted in the Section II, the growth effect of trust is ambiguous but it 立 𝛽𝐼 > 0, 𝛽𝐻𝐶 > 0 and 𝛽𝑇𝐹𝑃 > 0 when testing the income effect. As our. is certain that trust ought to be positively related with the output level. Labor. ‧ 國. 學. supply, physical capital, human capital and technology are direct inputs in the production function, and thus should reasonably contribute to a better. ‧. economy performance.. Nat. sit. y. Table 1 provides summary statistics for the main variables. The cross-. io. al. er. country trust level ranges from 0.028 to 0.682 with an average of 0.261. The Chinese province trust level ranges from 1.430 to 3.710 with an average of. n. v i n Ch 2.706. The difference between cross-country data and Chinese province data engchi U. comes from the different scale assigned to the survey answer. While the WVS for cross-country data only provide two answers for the respondents (0 for distrust and 1 for trust), the CGSS data provides the intensity data of individuals’ trust belief ranging from 1 to 5. The trust data cannot be compared across the two datasets, but can only be done so between the samples within the dataset themselves. Table 1 Descriptive Statistics for Cross-Country and Chinese Province Data Panel A. Cross-Country Data. 20.

(24) Variables. Mean. GDP Growth Rate. SD. Min. Max. 0.050. 0.086. -0.663. 1.608. 11485.859. 12778.201. 515.663. 134040.000. Population growth rate. 0.017. 0.018. -0.181. 0.204. Trust. 0.261. 0.183. 0.028. 0.682. Investment Rate(I/Y). 0.220. 0.089. 0.007. 0.802. Human Capital Index. 2.294. 0.597. 1.071. 3.619. TFP. 0.782. 0.505. 0.145. 7.364. GDP per capita(US dollar). Based on a sample of 2383 observations. Panel B. Chinese Province Data Variables. Mean. GDP Growth Rate. SD. Min. 政治大. Max. 0.022. 0.077. 0.174. 32314.632. 22614.331. 6375. 100105. 0.009. 0.017. -0.056. 0.058. 2.706. 0.715. 1.430. 3.710. Investment Rate(I/Y). 0.480. 0.118. Human Capital Index. 0.017. 0.007. First Industry Share. 0.110. 0.069. GDP per capita(Yuan). 立. Population growth rate. Based on a sample of 76 observations. 0.311. 0.879. 0.005. 0.036. 0.006. 0.342. Nat. y. ‧. ‧ 國. Trust. 學. 0.118. sit. Because the intensity of individual’s trust belief data are only available in. al. er. io. the CGSS database, only Chinese individual data are used to investigate the. n. v i n C h of the CGSS U estimation sources from the 4 waves e n g c h i in year 2003, 2005, 2010 trust effect on personal income. The Chinese individual data used in the. and 2013. For the reason that CGSS chooses the survey respondents randomly each wave, we can only construct a pooled dataset and use OLS method for estimation. The primary dependent variable is the personal income over the year. To eliminate the price effect on the income, we take the price level in year 2003 as the base and use consumer price index provided by the NBSC to adjust the income data in the later 3 waves of the survey. Other explanatory variable included in the estimation are the status of household registration (or Hukou),. 21.

(25) gender and schooling years of the respondent, the average working hours per week and working experience of the respondent, and most importantly, the individual trust level and the social trust level. The individual trust level uses the respondent’s answer facing the question that “Generally speaking, how do you trust the strangers” in the year 2003 and 2005; and “Generally speaking, would you say that most people can be trusted” in the year 2010 and 2013. The social trust level equals to the mean value of all the individuals’ trust in the same province and the survey wave that the respondent belongs to.. 政治大 𝑆𝑇 + 𝛽 𝐻𝐾 + 𝛽 𝐺 + 𝛽. Following the past literature, we set the basic regression model as 𝑌𝑖 = 𝛽𝑇 𝑇𝑟𝑢𝑠𝑡𝑖 + 𝛽𝑆𝑇. 立. 𝛽𝐸 𝐸𝑖 + 𝛽𝐸2 𝐸𝑖2 + 𝜀𝑖. 𝑖,𝑡. 𝐻𝐾. 𝑖. 𝑔 𝑖. 𝐸𝑑𝑢 𝐸𝑑𝑢𝑖. + 𝛽𝑊𝐻 𝑊𝐻𝑖 +. ‧ 國. 學. In the equation, 𝑌𝑖 denotes the personal income over the year for individual. ‧. 𝑖. The variable 𝑇𝑟𝑢𝑠𝑡𝑖 denotes the individual trust level while 𝑆𝑇𝑖 denotes the social trust level the respondent belongs to. The variable 𝐻𝐾𝑖 and 𝐺𝑖 are. sit. y. Nat. dummy variables and denote the respondent’s status of household registration. io. er. and gender; the value “0” implies that the respondent lives in rural area and is female, while the value “1” implies the respondent lives in urban area and is. al. n. v i n C h𝑖 , and 𝐸𝑖 denote theUrespondent’s schooling male. The variable 𝐸𝑑𝑢𝑖 , 𝑊𝐻 engchi years, the average working hours per week and working experience. The. variable 𝐸𝑖2 included in the estimation aims to capture the non-linear income effect of working experience. According to the past literature results and the prediction of our model, the expected sighs of parameters are 𝛽𝑇 > 0, 𝛽𝑆𝑇 > 0, 𝛽𝐻𝐾 > 0, 𝛽𝑔 > 0, 𝛽𝐸𝑑𝑢 > 0, 𝛽𝑊𝐻 > 0 and 𝛽𝐸 > 0, 𝛽𝐸2 < 0. It is predicted according to the model that trust has a positive effect on the individual’s income. Both past research results and evidences in China indicates that male residents live in urban area generally earn more. And. 22.

(26) because education, working experience help individual to accumulate higher human capital, the effect of both should enhance one’s ability to earn more income. Table 2 provides summary statistics for the main variables for the Chinese individual data.. Table 2 Descriptive Statistics for Chinese Individual Data Mean. SD. Min. Max. Income(Yuan). 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. 0.645. 0.478. 0. 1. Gender(0 stands for female). 0.549. 0.498. 0. Schooling Years. 10.204. 3.867. 學. 1. 0. 18.500. Average Working Hours per Week. 51.454. 14.268. 20. 98.000. Working Experience(Year). 20.799. 12.697. 0. 政治大. 立. Household Registration(1 stands for. ‧ 國. urban area). Based on a sample of 18809 observations. 50.000. sit. y. Nat. Empirical Analysis. io. er. IV.. ‧. Variables. A. Trust and Nation’s Wealth. n. al. Ch. i n U. v. Table 3 shows the result for testing the growth effect of trust. Equation (1) and. engchi. Equation (3) provide base regression result for cross-country data and Chinese province data. For cross-country data, the basic regression model performs well in predicting the dynamic of country level 5-year growth rate. The variable GDP per capita, the growth rate of population, investment rate, human capital indicator and technology indicator are all statistically significant. Nevertheless, the total 𝑅2 for both the cross-country regression equations are not ideal. The possible reason for this result is that different countries may have quite different economic growth strategy. Consider only labor supply,. 23.

(27) Table 3 Estimation Result for Growth Effect of Trust Cross-country Date. Chinese Province Data. Equation. (1). (2). (3). (4). (5). (6). Dependent variable. Growth. Growth. Growth. Growth. Growth. Solow residual. GDP per capita(log). -0.0211345***. -0.021456***. -0.1231263***. -0.1248364***. -.1231908***. (0.002648). (0.0026574). (0.0115759). (0.0147904). (.0150213). 0.2094946***. 0.211396***. 0.0482587. 0.0767829. .0769898. (0.069101). (0.0692188). (0.0985474). (0.1001783). (.1006493). 0.0554472***. 0.0540036***. 0.0148418. 0.017605. .0098945. (0.0138296). (0.0138847). (0.0192141). (0.0195354). (.022161). 0.0127934***. 0.0128063***. 2.014531**. 1.931158**. 1.793417**. (0.0040128). (0.0040412). (0.7762165). (0.7992508). (.8237765). 0.0140022***. 0.0138946***. (0.0030672). (0.0030702). -0.24466**. -0.2567873**. -.256286**. (0.0920267). (0.0987128). (.0991788). -0.0001801. .0002436. -0.0023621. (.0032459). (0.0027123). 0.0113992. io. Trust. (0.0119075) Saving/Y. y. (0.0031814). n. al. sit. Nat. First Industry Share. ‧. TFP. 學. Human Capital. er. I/Y. 立. ‧ 國. Growth of Pop. 政治大. Ch. engchi. i n U. v. -.022079 (.0294652). R-squared. 0.0395. 0.0386. 0.5092. 0.5054. 0.5821. 0.0070. Observations. 2,383. 2,383. 76. 72. 72. 109. Notes: a. As the result of Hausman test suggests, equation (1), (2), (6) are estimated using random effects model; equation (3), (4), (5) are estimated using fixed effects model. Sources: Author calculations.. 24.

(28) *** Significant at the 1 percent level. ** Significant at the 5 percent level. * Significant at the 10 percent level.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 25.

(29) investment, human capital and technology may not be sufficient to account for all factors that contributes to particular country’s growth. Thus it may not be suitable to use one simple and same growth regression to portray the dynamic of growth in 57 different countries included in our dataset. However, as the growth effect of trust is the main concern for the paper, comparing the equation (1), (2) and we see that no evidence according to the regression suggest trust contributes to positive growth for certain economy. For Chinese province data, the basic regression we used in the estimation performs as predicted with only two exceptions, the growth rate of population and the investment rate. The level of GDP per capita, the human capital indicator and the. 政治大. share of first industry in GDP are statically significant. Economy tends to grow faster. 立. with lower GDP per capita and higher human capital level; in addition, agricultural. ‧ 國. 學. provinces experience less growth compared with industrial province, which is commonly observed in China. Combining both equation (3) and (4), similar to the. ‧. result for the cross-country data, the estimated coefficient for trust is not statistically significant thus implying trust has no growth effect on province level in China.. y. Nat. sit. Therefore, both regression from cross-country data and Chinese province data are. n. al. er. io. consistent with the prediction of the modified RBC model in Section II that the. i n U. v. growth rate of the output of certain economy is not affected by the trust level in that society.. Ch. engchi. Recall that in the modified RBC model of Section II, the mathematical solution indicates that trust affect the macro economy through a constant saving rate under the assumption of no government purchase and fully depreciation. Controlling for the saving rate in the growth regression (equation 5), we see that both the trust coefficient and the saving rate coefficient remain statistically insignificant. This finding indicates that the economic growth in provinces of China is not affected by trust directly, or through saving rate indirectly. To further justify our empirical finding that trust may have no growth effect, we include the concept of Solow residual. Follow the method Chang-Tai Hsieh (1999) provided in his work, Solow residual equals to 26.

(30) ̂ − 𝑠𝑙 𝐿̂ = 𝑠𝑘 𝑟̂ + 𝑠𝑙 𝑤 𝑆𝑅 = 𝑌̂ − 𝑠𝑘 𝐾 ̂, where 𝑠𝑘 ≡ 𝑟𝐾/𝑌, 𝑠𝑙 ≡ 𝑤𝐿/𝑌 are the factor income share. Variable 𝑟 and 𝑤 denote the payments to per unit capital and labor so that 𝑌 = 𝑟𝐾 + 𝑤𝐿. Variable 𝑌̂, ̂ and 𝐿̂ denote the growth rate of output, capita and labor. Therefore, Solow 𝐾 residual is the residual that the growth rate of output after subtracting the shareweighted growth in factor quantities. In other words, Solow residual is the residual growth rate of output that cannot be explained by direct factor inputs and if trust contributes to certain economy’s growth, then trust should be significantly related with the Solow residual. As the constraint of data availability, only the Solow residual of Chinese provinces are calculated. Equation (6) provides the empirical result for the. 政治大. link between Solow residual and trust. The result of Hausman test equals to 0.9399,. 立. thus the regression is estimated using random effects model. The finding of the. ‧ 國. 學. estimation shows that no statistically significant relation exists between these two variable and thus furthermore verifies that no growth effect of trust.. ‧. Combing the empirical result both from the cross-country data and Chinese province data, we are confident with the estimation finding that trust has no growth. n. al. er. io. sit. y. Nat. effect.. i n U. v. Table 4 shows the result for testing the income effect of trust using cross-country. Ch. engchi. data. Our base model specification performs well in predicting the income per capita level on cross-country data. Similar to economic theory and past literature suggests, both investment rate, human capital and the technology level are statistically significant, positively related with the country’s income per capita level. Combing the estimation result of equation (7) and (8), it seems strangely that trust also has no income effect. However, inspired by Knack and Keefer (1997)’s estimation that their trust-growth relationship is sensitive to the choice of human capital measures, we then first test the income-trust relationship with the absence of human capital indicator in our regression. Eliminating the interference of human capital indicator (equation 9), the estimation result shows clearly that trust is strongly positive related with income per capita level. 27.

(31) Table 4 Estimation Result for Income Effect of Trust with Cross-Country Data Equation. (7). (8). (9). (10). (11). Dependent variable(Log). GDP per capita. GDP per capita. GDP per capita. Human Capital. GDP per capita. I/Y. 0.9401036***. 0.9375689***. 1.829229***. 0.9709215***. (0.0839421). (0.0839434). (0.1658596). (0.0816935). 1.272623***. 1.272192***. (0.0151541). (0.0151812). 0.4546421***. 0.454106***. (0.0159471). (0.0159452). (0.031505). (0.0155575). 0.158176. 1.53416***. (0.261632). (0.4515722). 2,383. 0.6733. 0.3414. 2,383. 2,383. io. al. y. Observations. -2.17814***. (0.4198693). (0.3213552). (0.0801758). ‧. 0.6730. 1.154551***. 0.9435113***. Nat. R-squared. Notes:. 0.43983***. ‧ 國. Trust*HC. (0.025009). 0.3106909***. 學. Trust. 立. 1.034523***. sit. TFP. 政治大. 0.1279. 0.6773. 2,668. 2,383. er. Human Capital. n. a. GDP per capita as dependent variable is taken 5-year average for clear causality purpose.. i n U. v. b. Because only one wave trust data are used to construct cross-country dataset, all equations are estimated using random effects model. Sources: Author calculations. *** Significant at the 1 percent level.. Ch. engchi. ** Significant at the 5 percent level. * Significant at the 10 percent level.. 28.

(32) Table 5 Estimation Result for Income Effect of Trust with Chinese Province Data Equation. (12). (13). (14). (15). (16). (17). Dependent variable(Log). GDP per capita. GDP per capita. GDP per capita. GDP per capita. GDP per capita. GDP per capita. I/Y. 0.5125063. 0.0463404. 0.7674266*. 0.1220121. 0.0450573. 0.0564968. (0.3869525). (0.2929024). (0.4100303). (0.3328201). (.2474157). (0.2999501). 76.9608***. 61.88246***. 79.69946***. 62.71834***. 22.49756*. 58.12971***. (12.87152). (9.626655). (12.76775). (9.849428). (11.88093). (9.690117). -4.810724**. -2.876773*. -4.295493**. -2.858322*. -2.376992*. -3.272964**. (1.827107). (1.466759). (1.823611). (1.478708). (1.243842). (1.605464). 0.258044***. 0.2516935***. -0.1267612. 0.2259023***. (0.0371792). (0.0396286). (.0902714). (0.037888). Saving/Y. Trust*HC. 0.945721*. 0.2190103. (0.5653059). (0.4449003). ‧. ‧ 國. 學. Trust. 立. (4.441325). io. Rural Population Share. y. Nat. FDI/Y. al. n. Rural Consumption. 20.19377***. -2.189576. sit. First Industry Share. 政治大. Level/Urban Level. (3.226047). er. Human Capital. Ch. engchi U. -.5221559. v ni. (0.927342) 0.8626399* (0.4401818). R-squared. 0.7162. 0.7848. 0.6696. 0.7741. 0.7544. 0.7767. Observations. 76. 72. 76. 72. 72. 72. Notes: a. GDP per capita as dependent variable is taken 5-year average for clear causality purpose.. 29.

(33) b. As the result of Hausman test suggests, all equations are estimated using fixed effects model. Sources: Author calculations. *** Significant at the 1 percent level. ** Significant at the 5 percent level. * Significant at the 10 percent level.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 30.

(34) The reason that the income-trust relationship is sensitive to human capital indicator is that human capital is strongly related with trust as equation (10) shows. Therefore, we include a new cross term 𝑇𝑟𝑢𝑠𝑡 ∗ 𝐻𝐶 which equals to the value of human capital indicator multiplied by the value of trust indicator, in the income-trust estimation. Controlling the relation between human capital and trust, the result of equation (11) provides several important implications: (a) Controlling the trust and human capital level at the observations’ mean which equals to 0.261 and 2.294, 1 percent positive change of human capital will lead to a 0.0237(2.294*1 percent *1.034523) positive income change from human capital itself, and 0.00565 (0.261*2.294*1 percent *0.94351113) positive income change from the mutual relationship between trust and. 政治大. human capital. Notice that the sum of 0.0237 and 0.00565 equals 0.02935 is larger. 立. than the result of 1 percent human capital change in equation (7) and (8) which equals. ‧ 國. 學. to 0.02919(2.294*1 percent *1.272623) and 0.02918(2.294*1 percent *1.272192). This finding suggests that trust increase the revenue for human capital. (b) Controlling. ‧. the trust and human capita level at the observations’ mean which equals to 0.261 and 2.294, 1 percent positive change of trust change will result in a 0.00568(0.261*1. y. Nat. sit. percent *2.17814) negative income change directly from trust itself, and 0.00565. n. al. er. io. (0.261*2.294*1 percent *0.94351113) positive income change from the mutual. i n U. v. relationship between trust and human capital. The sum of these two effects equals to -. Ch. engchi. 0.00003, suggesting a very small negative overall income change with 1 percent positive trust change, when the value of trust and human capital are controlled at observations’ mean. (c) The finding described in implication (b) also indicates that for trust to improve the income, the value of certain country’s human capital indicator must meet a minimum value. The threshold condition for human capital indicator according to the equation (11) is 2.309(2.17814/0.9435113), which is slightly larger than the mean human capital indicator 2.294 of our observations. It indicates that trust has a positive income effect in those countries with human capital indicator exceeding the value of 2.309. As long as a country has a human capital indicator lager than the threshold, the increase of trust will always lead to a positive income effect. On the contrary, the increase trust in a society with low accumulation of human capital, can 31.

(35) only cause economic loss. As Douglas North noticed, “the inability of societies to develop effective, low-cost enforcement of contracts is the most important source of both historical stagnation and contemporary underdevelopment in the Third World” (North, 1990, p. 54). Considering the mutual relationship between human capital and trust, it is reasonable to conclude that the Northian poverty trap can be solved by increasing both the social trust level and human capital level. As is accepted that trust can generate economic revenues by reducing the transaction cost, it is important for individuals in the society to fully understand and appreciate the value of trust for the trust-income relationship to work. More importantly, the successful transactions based on trust require both parties to strictly dripline themselves. Therefore, to overcome the. 政治大. poverty trap and realize the benefit of trust, it is essential for a society to accumulate. 立. enough human capital and increase the civilized population base who respect the. ‧ 國. 學. value of mutual trust. In the real world, it is common to observe less developed countries still stuck in poverty trap after repeated political reforms which are. ‧. supposed to have positive economic benefit. The key problem for these struggles may lies in the insufficient accumulation of human capital. Recall the economic success of. y. Nat. sit. eastern Asia, it is certain that after proper institution reforms, high accumulation of. n. al. er. io. human capital plays an important base in their economic take-offs.. Ch. engchi. i n U. v. Table 5 shows the result for testing the income effect of trust using Chinese province data. Similar to the situation in cross-country data, the base model specification as equation (12) shows performs well. Equation (13) demonstrates the result when adding the trust indicator into the base model. Because the coefficient of trust indicator is statistically significant, it is clear that income per capita level is strongly influenced by trust indicator. Equation (14) and (15) provide the estimation when saving rate is added in the regression process. Recall in Section II, that the mathematical solution for our modified RBC model implies that output level is determined by the saving rate of the economy and individual saving rate are strongly influenced by the social trust level. The estimation in equation (14) clearly demonstrates that income per capita level is 32.

(36) strongly and positively related with each province’s saving rate. Notice that the coefficient for investment rate becomes statistically significant when controlling the saving rate for province, this change indicates that personal saving is an important part for investment to improve the income level per capita. When controlling both the saving rate and trust level, the equation (15) shows that only the coefficient of trust indicator is statistically significant and the income effect of saving rate is dominated by the income effect of trust. This estimation result is consistent with our modified RBC model’s prediction that representative household’s saving rate is determined by trust level of the society. Inspired by the mutual relationship between human capital and trust, the same cross. 政治大. term is added in the Chinese provinces’ estimation. The equation (16) shows the result. 立. controlling the relationship between human capital and trust: (a) Controlling the trust. ‧ 國. 學. and human capital level at the observations’ mean which equals to 2.706 and 0.017, 1 percent positive change of human capital will lead to a 0.0038 (22.49756*1 percent. ‧. *0.017) positive income change from itself, and 0.0093 (20.19377*2.706*1 percent *0.017) positive income change from the mutual relationship. Notice the sum of the. y. Nat. sit. positive income change equals to 0.0131 which is larger than the result of 1 percent of. n. al. er. io. solely human capital change in equation (12) which equals to 0.0130. This finding is. i n U. v. consistent with the result of cross-country data that trust will increase the revenues of. Ch. engchi. human capital. (b) Controlling the trust and human capital level at the observations’ mean which equals to 2.706 and 0.017, 1 percent positive change of trust will lead to 0.0093 (20.19377*0.017*1 percent *2.706) positive income change solely from the mutual relationship. Notice the positive income change from 1 percent trust change in equation (13) equals to only 0.0070 (0.258044*1 percent *2.706). Thus it is reasonable to conclude that in China, human capital can also increase the revenues of trust. This result is slightly different from the cross-country data which require a minimum value of human capital indicator for increasing trust to have positive income change; the key difference lies in that the variable trust is not statistically significant after controlling the mutual relationship between human capital and trust in the estimation of Chinese provinces. The result may be confusing at first glance, 33.

(37) however, when referring back to the human capital indicator in our cross-country dataset, we find that after year 2003, the human capital indicator of China ranges from 2.410701 to 2.579169. Thus the solely positive income change from trust is because after year 2003, the human capital indicator in China is larger than the threshold value we calculated in the cross-country data which equals to 2.309. Therefore, the result is in fact consistent with our analysis for the cross-country data. Equation (17) provide a simple robust test for our trust indicator. As it is known that the difference between rural area and urban area is substantial and foreign investment plays an important role in some province’s economic performance, we include the rural population share, the proportion of rural individual consumption on urban. 政治大. individual consumption and the share of foreign direct investment in GDP in the. 立. estimation. Notice only the variable equals to the rural consumption level divided by. ‧ 國. 學. the urban consumption level is statistically significant, thus it suggests that the smaller the gap between rural area and urban area, the better average income level in that. ‧. province. Therefore, this result verifies the common agreement in China that to stimulate the economic development, it is essential for local government to pay their. y. Nat. sit. attention in the rural area. In addition, the estimation result also shows that the trust. n. al. er. io. indicator is rather robust with all these new variables added. During our estimation,. i n U. v. other variables are also included to test the robustness of the income-trust. Ch. engchi. relationship. All regression shows that the trust indicator in this paper is generally robust regardless which new variable is included in the regression model.. Both the estimation result of cross-country data and Chinese province data indicate that trust can only influence the economy’s income level but not the growth. An economy with higher trust level will produce higher overall output. However, the growth rate of the output of certain economy is not affected by the trust level in that society. That is, our empirical finding suggest that trust has only income effect but no growth effect, which is consistent with the predictions of the modified RBC model.. B. Trust and Individual Income 34.

(38) Recall the representative household optimization problem in Section II, the two models both predict that households earn higher income with higher social trust level. In this part, we test the income-trust relationship on individual level based on the CGSS survey data. Table 6 shows the result for the estimation. The basic model specification as equation (18) shows, indicates that male residents living in urban area tend to have higher overall income level. The estimation suggests an income gap between rural and urban residents, male and female individuals, which is commonly observable in Chinese society. Furthermore, human capital is positively related with the individual income level. The Intuition of this result is rather straightforward, individuals get. 政治大. more wage paid with higher human capital as it enhances individuals to work more. 立. efficiently. The residents living in provinces located in eastern region, especially. ‧ 國. 學. coastal provinces, have higher income level because the fact that the eastern provinces are more developed in China. Notice the sign of the coefficient of the working. ‧. experience is surprisingly negative, in addition, both the coefficients of working hours and square of working experience remain statistically insignificant. These findings. y. Nat. sit. somehow contradict with our prediction in Section III. There may be two possible. n. al. er. io. reasons for this strange estimation result: (a) The wage gap in China is rather large. i n U. v. and thus the income in the polarizing job market may not necessarily determined by. Ch. engchi. individuals’ working hours and experience. (b) The reason may also lie in the nature of the dataset that it is constructed based on the survey data. All data are self-reported by respondents and errors shall be expected. Though it may take more work to specify a clear relationship between individual income and working hours, experience in China, it is not the main concern of this paper. Equation (19) and (20) test the link between individual income and trust. In equation (19), only individual trust is included in the regression, and the estimation result demonstrate a strong positive income-trust relationship. The finding suggests that with higher individual trust level, people generally earn more than those distrust the society. This finding has two possible explanations: firstly, with higher individual trust level, people tends to feel safe when investing. Thus for people holds high trust 35.

(39) Table 6 Estimation Result using CGSS Individual Survey Data Equation. (18). (19). (20). (21). (22). (23). Dependent variable. Income. Income. Income. Income. Income. Income. Hukou (1 as urban. 0.2678224***. 0.2876917***. 0.3204453***. 0.3219156***. 0.2961358***. 0.3299439***. residents). (0.0163504). (0.0157568). (0.0143537). (0.01438). (0.015636). (0.0143416). Gender (1 as male). 0.3937577***. 0.3505052***. 0.2984744***. 0.2985042***. 0.3466457***. 0.2981885***. (0.0136098). (0.0131418). (0.0119926). (0.0119921). (0.0130023). (0.0119441). 0.1119743***. 0.1045475***. 0.0953194***. 0.0953404***. 0.103549***. 0.0950151***. (0.002136). (0.002065). (0.0018857). (0.0018857). (0.0020428). (0.0018777). Average Working Hours. 0.0005873. 0.0005191. -0.0005841. -0.0005674. 0.0004723. -0.0005877. per Week. (0.0004983). (0.0004798). (0.0004371). (0.0004372). (0.0004744). (0.0004351). Working Experience. -0.0149011***. -0.0112481***. -0.0023036. -0.0023722. -0.0111345***. -0.0024897. (0.0019052). (0.0018348). (0.0016765). (0.0016769). (0.0018163). (0.0016709). Square of Working. -6.43e-06. -0.0000105. -0.0000417. -0.0000406. -3.11e-06. -0.0000374. Experience. (0.0000399). (0.0000384). (0.000035). (0.000035). (0.000038). (0.0000349). East Region. 0.465123***. 0.4382382***. 0.392691***. 0.3947532***. 0.4482743***. 0.3955195***. (0.0136897). (0.0131891). (0.0120289). (0.0120917). (0.0130732). (0.0120141). 0.0024275. -0.0309399. Social Trust. (0.006289). Ch. engchi. 0.6574203*** (0.0105712). y. sit. er. al. n. (0.0056544). ‧. io. 0.2271343***. 學. Nat. Individual Trust. 立. ‧ 國. Schooling Years. 政治大. i n U. v. (0.0209769) 0.6251566***. .6536393***. (0.0220491). (0.0109215). Individual Trust* Social. 0.0126292*. Trust. (0.0075743). Trust Value=2. Trust Value=3. 0.2156505***. -0.0725112***. (0.0171256). (0.0164176). 0.2289155***. -0.049614***. 36.

(40) Trust Value=4. Trust Value=5. Adjusted R-squared. 0.3578. 0.4079. Observations. 18,809. 18,716. 立. Sources: Author calculations. 0.0139832. (0.0193436). (0.0219779). 0.7259889***. -0.0677873*. (0.036756). (0.036204). 0.5094. 0.4180. 0.5111. 18,716. 18,716. 18,809. 18,809. ‧. io. sit. y. Nat. n. al. er. * Significant at the 10 percent level.. 0.7913898***. 0.5093. ‧ 國. ** Significant at the 5 percent level.. (0.0192145). 學. *** Significant at the 1 percent level.. 政治大. (0.0203408). Ch. engchi. i n U. v. 37.

(41) belief, it is more likely for them to seize the profitable investment opportunity. Secondly, people with higher trust belief, are more cooperative and their trust in other people, especially strangers, makes them better team player than those individuals who hold pessimistic attitudes. Therefore, people with high trust belief tends to work more efficiently. Controlling the social trust level, however as equation (20) shows, the social trust level dominates the income effect of individual trust. The estimation suggests that as long as people live in a society with high social trust level, the income will be higher than those live in society with low trust level. The estimation also suggests that no matter what the individual’s trust belief is, it is the society’s trust level mostly matters. 政治大. for income rather than the individual’s belief. No matter what the subjective. 立. individual trust belief is, all in all, the individual economic performance is constrained. ‧ 國. 學. by the structure of the society. In a society with high trust level, individuals with pessimistic attitudes can still enjoy the benefits provided by the positive trust-income. ‧. relationship. However, individuals with strong faith in the society but living in a low trust level society, will only suffer from others’ cheating.. y. Nat. sit. So does the result of equation (20) means that individual trust will have no effect at. n. al. er. io. all on income when social trust being controlled? To further explore the mutual. i n U. v. relationship between the individual and social trust, we introduce a cross term of both. Ch. engchi. trust indicators in equation (21). With the relationship between both the trust indicators controlled, the result of equation (21) provides two interesting implications: (a) Controlling the individual trust and social trust at the observations’ mean which equals to 2.474 and 2.498, 1 percent positive change of individual trust will lead to a 0.0078(0.126292*2.474*1 percent *2.498) positive income change solely from the mutual relationship as the estimation for the individual trust indicator is not statistically significant. Thus the result indicates that even though the social trust level is the dominating factor, it is still beneficial for individuals to increase their own trust because of the mutual relationship between the social and individual trust. (b) Controlling the individual trust and social trust at the observations’ mean which equals to 2.474 and 2.498, 1 percent positive change of social trust will lead to a 38.

(42) 0.0156(0.6251566*2.498*1 percent) positive income change from itself and 0.0078(0.126292*2.498*1 percent *2.474) positive income change from the mutual relationship. Notice the combined effect equals to 0.0234 which is larger than the result of 1 percent positive change of social trust in equation (20) which equals to 0.0164(0.6574203*2.498*1 percent). Therefore, for those individuals who hold high trust believes, the positive income change from the increase of the social trust will be larger than those who distrust. In other words, individual trust will increase the income revenues of the social trust. Considering both the implications from the equation (21), all in all, it is still beneficial for individuals to trust more even when the dominating effect of social trust being considered.. 政治大. One disadvantage of using intensity data for trust in our empirical analysis, is that. 立. the difference between each answer may not be meaningfully compared. For different. ‧ 國. 學. respondents, the difference of trust level between “3” and “4” for example, may not be the same. Worrying about the disadvantage, we include dummy variables. ‧. according to the respondent’s answer in our regression. For one specific respondent, if he choose the trust value “3” in the survey, then the value of variable “Trust Value=3”. y. Nat. n. al. er. io. Value=2”,” Trust Value=4” and “Trust Value=5” will be “0”.. sit. will be “1”, and the value of other variables including “Trust Value=1”, “Trust. i n U. v. Equation (22) demonstrates the estimation result using the dummy variables of. Ch. engchi. trust. The regression shows that generally trust will positively benefit one’s income. Another important finding that equation (22) implies is that the relationship between income and individual income follows a lump-shape pattern, consistent with the finding of J. Butler, P. Giuliano and L. Guiso’s work using the European Social Survey database. The possible reason as they suggest is that “highly trusting individuals tend to assume too much social risk and to be cheated more often”, while “individuals with overly pessimistic beliefs avoid being cheated, but give up profitable opportunities” (Butler, Giuliano, Guiso, 2016, p. 1155). It might be a little confusing at first when comparing the result of equation (22) that suggests an optimal individual trust while the result of equation (19), (21) that indicate the purely income increasing effect of individual trust. However, when referring back to the individual 39.