Table 5 Heterogeneity analysis.

From: The rationality of the punishment ladder: evidence from a quasi-natural experiment in China

Variable

(1)

(2)

(3)

(4)

(5)

(6)

Term of imprisonment

Number of fines imposed

Linear

Quadratic

Cubic

Linear

Quadratic

Cubic

Panel A: Eastern region

Whether the number of pieces of information is more than 50,000

8.777**

8.822**

8.885**

15.15**

14.29*

13.17*

(4.218)

(4.130)

(4.002)

(7.367)

(7.228)

(7.019)

Number of pieces of information

0.0311

−0.000594

−0.000773

−12.00**

−1.135**

−0.138**

(3.052)

(0.298)

(0.038)

(5.331)

(0.522)

(0.067)

Constant term

−0.718

−0.562

−0.492

35.7

5.322

−4.573

(16.620)

(11.030)

(9.799)

(29.020)

(19.310)

(17.190)

Observations

320

320

320

320

320

320

R2

0.503

0.503

0.503

0.279

0.276

0.272

Panel B: Central and Western region

Whether the number of pieces of information is more than 50,000

3.76

3.876

4.065

−110

−131.9

−154

(8.162)

(8.137)

(7.995)

(256.700)

(255.900)

(251.300)

Number of pieces of information

3.155

0.31

0.0394

−12.19

0.619

0.332

(5.940)

(0.600)

(0.079)

(186.800)

(18.860)

(2.466)

Constant term

−6.73

1.268

4.046

−31.4

−98.42

−114.7

(32.460)

(22.020)

(19.710)

(1020.700)

(692.300)

(619.300)

Observations

152

152

152

152

152

152

R2

0.608

0.608

0.607

0.316

0.316

0.317

  1. The control variables in this table include the logarithm of profit amount, tribunal size, principal criminal, accessary criminal, first offender, surrender, admittance of guilt and acceptance of punishment, presence of a defender, gender, educational attainment, unemployed status, the logarithm of per-capita GDP and the logarithm of R&D expenditure per province. The values shown in parentheses are robust standard errors of the estimates, *, **, and *** indicate 10%, 5%, and 1% significance levels, respectively.