Appendix
Table A.1: The Levin–Lin–Chu(LLC) bias adjusted t-statistic for panel unit-root test
Bias adjusted t*Variable / Includes trend / without trend
n1_9 / -43.729*** / -32.122***
n10_19 / -62.683*** / -42.005***
n20_49 / -49.390*** / -42.559***
n1_49 / -39.238*** / -32.888***
n50_99 / -56.850*** / -44.811***
n100_more / -52.842*** / -37.855***
est / -40.256*** / -32.715***
Source: Authors
Table A.2: Summary Statistics, Metro
Variable / Obs / Mean / Std. Dev / Min / Maxn1_9 / # of est. with 1-9 employees / 4,124.491 / 9,207.869 / 32 / 189,561
n10_19 / # of est. with 10-19 employees / 719.517 / 1,516.323 / 2 / 28,782
n20_49 / # of est. with 20-49 employees / 500.177 / 1,114.648 / 0 / 21,323
n1_49 / # of est. with 1-49 employees / 5,306.601 / 11,767.770 / 36 / 239,666
n50_99 / # of est. with 50-99 employees / 173.658 / 406.816 / 0 / 7,626
n100_more / # of est. with 100 and more employees / 144.525 / 353.341 / 0 / 6,204
est / # of establishments / 5,619.334 / 12,499.670 / 38 / 253,004
loan / SBA disaster loan, $ / 85,498.680 / 1,352,659.000 / 0 / 126,000,000
l_loan_est / log of the SBA disaster loan per establishment / 0.368 / 1.263 / 0 / 9.401
loan_est / SBA disaster loan per establishment, $ / 26.175 / 301.241 / 0 / 12,105.130
l_income / Log of income / 10.342 / 0.215 / 9.661 / 11.593
HI_c / HHI index / 2,253.330 / 1,446.236 / 418.228 / 10,000
Population / Population size in 1000s / 225.8793 / 484.2631 / 1.571 / 9,826.044
l_wage_manufacture / log of manufacturing wage rate / 9.295 / 0.283 / 7.167 / 10.679
rain_ano / Precipitation deviation from its long run average / 0.06001 / 1.1050 / -6.6154 / 8.9934
l_totaldamage / log total SHELDUS damages per capita / 0.4435 / 1.0823 / 0 / 10.1873
Notes: sample contains 12,781county by year observations; total number of metro counties 1,093
Table A.3: Summary Statistics, Urban
Variable / Obs / Mean / Std. Dev / Min / Maxn1_9 / # of est. with 1-9 employees / 586.824 / 444.226 / 57 / 4,209
n10_19 / # of est. with 10-19 employees / 95.603 / 75.722 / 5 / 620
n20_49 / # of est. with 20-49 employees / 56.857 / 47.811 / 1 / 391
n1_49 / # of est. with 1-49 employees / 737.875 / 561.838 / 69 / 5,163
n50_99 / # of est. with 50-99 employees / 16.963 / 15.382 / 0 / 117
n100_more / # of est. with 100 and more employees / 13.827 / 12.577 / 0 / 101
est / # of establishments / 768.311 / 585.617 / 72 / 5,359
loan / SBA disaster loan, $ / 27,220.430 / 386,336.300 / 0 / 21,300,000
l_loan_est / log of the SBA disaster loan per establishment / 0.286 / 1.203 / 0 / 11
loan_est / SBA disaster loan per establishment, $ / 35.407 / 567.923 / 0 / 52,454.760
l_income / Log of income / 10.193 / 0.173 / 9.347 / 11.662
HI_c / HHI index / 2,724.020 / 1,176.945 / 714.933 / 10,000
Population / Population size in 1000s / 31.939 / 22.0415 / 3.327 / 190.175
l_wage_manufacture / log of manufacturing wage rate / 9.094 / 0.251 / 7.208 / 10.803
rain_ano / Precipitation deviation from its long run average / 0.0384 / 1.1262 / -11.4844 / 24.1616
l_totaldamage / log total SHELDUS damages per capita / 0.5262 / 1.2228 / 0 / 9.4657
Notes: sample contains 14,733county by year observations; total number of urban counties 1,275
Table A.4: Summary Statistics, Rural
Variable / Obs / Mean / Std. Dev / Min / Maxn1_9 / # of est. with 1-9 employees / 183.126 / 129.232 / 24 / 1,310
n10_19 / # of est. with 10-19 employees / 23.761 / 16.689 / 0 / 130
n20_49 / # of est. with 20-49 employees / 11.827 / 9.123 / 0 / 71
n1_49 / # of est. with 1-49 employees / 218.317 / 152.414 / 25 / 1,482
n50_99 / # of est. with 50-99 employees / 3.451 / 2.922 / 0 / 21
n100_more / # of est. with 100 and more employees / 2.443 / 2.519 / 0 / 17
est / # of establishments / 224.170 / 155.777 / 25 / 1,504
loan / SBA disaster loan, $ / 10,182.610 / 222,091.100 / 0 / 12,800,000
l_loan_est / log of the SBA disaster loan per establishment / 0.237 / 1.161 / 0 / 10
loan_est / SBA disaster loan per establishment, $ / 38.871 / 614.023 / 0 / 28,707.370
l_income / Log of income / 10.180 / 0.206 / 9.606 / 11.107
HI_c / HHI index / 4,402.188 / 1,968.164 / 932.975 / 10,000
Population / Population size in 1000s / 9.7621 / 5.8263 / 1.022 / 38.215
l_wage_manufacture / log of manufacturing wage rate / 8.911 / 0.327 / 7.436 / 10.789
rain_ano / Precipitation deviation from its long run average / 0.1009 / 1.1093 / -7.4346 / 11.8719
l_totaldamage / log total SHELDUS damages per capita / 0.6012 / 1.4027 / 0 / 10.2625
Notes: sample contains 4,458county by year observations; total number of urban counties 467.
Table A.5. Metro Sample, Small Establishment Categories
n1_9 / n10_19 / n20_49l_loan_est / 0.6752 / 0.3476*** / 0.2858
(0.619231) / (0.1315) / (0.2427)
HI_c / -0.0011*** / -0.0002* / -0.0004***
(0.0004) / (0.0001) / (0.0001)
population / 0.0148*** / 0.0057** / -0.0069***
(0.0041) / (0.0026) / (0.0019)
l_income / 12.7912*** / 2.8177*** / 2.6907**
(4.7589) / (0.9076) / (1.0454)
l_wage_manuf_rate / -5.2040** / -0.5135 / -1.8976***
(2.0304) / (0.3571) / (0.5149)
l_totaldamage / -0.1632 / -0.0977*** / -0.0710
(0.1457) / (0.0332) / (0.0575)
rain_ano / -0.1074* / -0.0321** / -0.0312
(0.0629) / (0.0133) / (0.0211)
L1rain_ano / -0.2194 / -0.0712*** / -0.0625
(0.1356) / (0.0237) / (0.0435)
L2rain_ano / -0.1466* / -0.0425*** / -0.0376*
(0.0774) / (0.0140) / (0.0224)
L3rain_ano / -0.1491*** / -0.0288*** / -0.0260**
(0.0506) / (0.0095) / (0.0131)
L4rain_ano / -0.0353 / -0.0143 / -0.0087
(0.0309) / (0.0099) / (0.0128)
L5rain_ano / 0.0231 / -0.0100 / 0.0125
(0.0285) / (0.0075) / (0.0124)
L.y / 0.9223*** / 0.7967*** / 1.3049***
(0.0216) / (0.0932) / (0.0962)
N / 12,781 / 12,781 / 12,781
Sargan (39) / 5,880.27*** / 1,662.14*** / 1,977.95***
Hansen (39) / 33.32 / 30.49 / 33.85
Hansen P-value / 0.73 / 0.83 / 0.70
AR(1) / -2.32** / -2.87*** / -2.03**
AR(2) / -0.79 / -0.96 / -0.76
AR(3) / -2.37** / -1.60 / -0.53
AR(4) / -1.23 / 1.63 / 0.42
Notes: Dependent variables are scaled by 100. Column headings correspond to total numbers of different establishment categories that serve as dependent variables in the GMM model. L.y represents the lag dependent variable. Instruments correspond to all available lags starting from the lag 5. Includes county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.6. Urban Sample, Small Establishment Categories
n1_9 / n10_19 / n20_49l_loan_est / 0.1067** / -0.0104 / 0.0305***
(0.0436) / (0.0116) / (0.0102)
HI_c / -0.00006** / 0.0000 / -0.000009
(0.00003) / (0.000006) / (0.000006)
population / 0.0020 / 0.0019** / 0.0011
(0.0013) / (0.0008) / (0.0007)
l_income / 0.4913*** / 0.0824** / 0.0920***
(0.1134) / (0.0368) / (0.0312)
l_wage_manuf_rate / -0.1311*** / -0.0016 / -0.0075
(0.0328) / (0.0074) / (0.0074)
l_totaldamage / -0.0185** / 0.0024 / -0.0053***
(0.0082) / (0.0022) / (0.0019)
rain_ano / -0.0107*** / 0.0011 / -0.0011
(0.0032) / (0.0010) / (0.0008)
L1rain_ano / -0.0081* / 0.0018 / -0.0025**
(0.0043) / (0.0013) / (0.0011)
L2rain_ano / -0.0069** / 0.0011 / -0.0009
(0.0027) / (0.0008) / (0.0006)
L3rain_ano / -0.0081*** / -0.0015** / -0.0009
(0.0023) / (0.0007) / (0.0006)
L4rain_ano / -0.0039 / -0.0021*** / -0.0015**
(0.0025) / (0.0008) / (0.0006)
L5rain_ano / -0.0011 / 0.0006 / 0.0011
(0.0024) / (0.0008) / (0.0007)
L.y / 0.9879*** / 0.9401*** / 0.9182***
(0.0064) / (0.0262) / (0.0367)
N / 14,733 / 14,733 / 14,733
Sargan (39) / 384.12*** / 116.40*** / 116.52***
Hansen (39) / 27.44 / 29.79 / 24.75
Hansen P-value / 0.55 / 0.42 / 0.69
AR(1) / -8.07*** / -17.37*** / -13.00***
AR(2) / 0.84 / 4.60*** / 2.93***
AR(3) / -0.52 / 0.60 / 1.08
AR(4) / -0.54 / 0.07 / -0.45
Notes: Dependent variables are scaled by 100. Column headings correspond to total numbers of different establishment categories that serve as dependent variables in the GMM model. L.y represents the lag dependent variable. Instruments correspond to all available lags starting from the lag 5. Includes county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.7. Rural Sample, Small Establishment Categories
n1_9 / n10_19 / n20_49l_loan_est / 0.0169 / -0.0034 / -0.0052
(0.0153) / (0.0055) / (0.0035)
HI_c / -0.000003 / -0.000001 / 0.0000
(0.00001) / (0.000004) / (0.000002)
population / 0.0204*** / 0.0037*** / 0.0026*
(0.0054) / (0.0014) / (0.0014)
l_income / 0.6213*** / 0.0961*** / 0.0627**
(0.1391) / (0.0366) / (0.0302)
l_wage_manuf_rate / -0.0552*** / -0.0072* / -0.0028
(0.0186) / (0.0042) / (0.0026)
l_totaldamage / -0.0001 / 0.0015 / 0.0011
(0.0034) / (0.0012) / (0.0008)
rain_ano / -0.0056** / -0.0014* / 0.00007
(0.0023) / (0.0008) / (0.0005)
L1rain_ano / -0.0005 / 0.0007 / 0.0007
(0.0021) / (0.0007) / (0.0005)
L2rain_ano / -0.0027 / 0.0013* / 0.0004
(0.0021) / (0.0008) / (0.0004)
L3rain_ano / -0.0062** / -0.0014* / -0.0004
(0.0026) / (0.0008) / (0.0005)
L4rain_ano / -0.0066*** / -0.0005 / -0.0003
(0.0025) / (0.0008) / (0.0005)
L5rain_ano / -0.0008 / 0.0011 / 0.0002
(0.0021) / (0.0008) / (0.0005)
L.y / 0.9263*** / 0.8850*** / 0.8002***
(0.0200) / (0.0573) / (0.1083)
N / 4,458 / 4,458 / 4,458
Sargan (39) / 205.51*** / 38.29 / 56.72**
Hansen (39) / 48.85 / 41.16 / 39.40
Hansen P-value / 0.13 / 0.38 / 0.45
AR(1) / -9.35*** / -10.14*** / -8.19***
AR(2) / 0.07 / 4.30*** / 3.72***
AR(3) / 0.01 / -1.11 / 0.54
AR(4) / -0.36 / 1.10 / -1.00
Notes: Dependent variables are scaled by 100. Column headings correspond to total numbers of different establishment categories that serve as dependent variables in the GMM model. L.y represents the lag dependent variable. Instruments correspond to all available lags starting from the lag 5. Includes county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.8: Summary Statistics, Coastal
Variable / Obs / Mean / Std. Dev / Min / Maxn1_9 / # of est. with 1-9 employees / 4,629.535 / 11,338.42 / 57 / 189,561
n10_19 / # of est. with 10-19 employees / 758.233 / 1,805.107 / 3 / 28,782
n20_49 / # of est. with 20-49 employees / 519.662 / 1,318.027 / 0 / 21,323
n1_49 / # of est. with 1-49 employees / 5,803.935 / 14,309.790 / 55 / 239,666
n50_99 / # of est. with 50-99 employees / 179.898 / 478.294 / 0 / 7,626
n100_more / # of est. with 100 and more employees / 150.502 / 413.245 / 0 / 6,204
est / # of establishments / 6,128.678 / 15,170.740 / 57 / 253,004
l_loan_est / SBA disaster loan, $ / 0.355 / 1.211 / 0 / 9
loan / log of the SBA disaster loan per establishment / 93,659.200 / 1,650,134.000 / 0 / 126,000,000
loan_est / SBA disaster loan per establishment, $ / 19.029 / 199.813 / 0 / 10,699.370
l_income / Log of income / 10.327 / 0.257 / 9.347 / 11.593
HI_c / HHI index / 2,532.844 / 1,646.105 / 539.482 / 10,000
Population / Population size in 1000s / 246.8195 / 593.2031 / 1.14 / 9,826.044
l_wage_manufacture / log of manufacturing wage rate / 9.298 / 0.306 / 7.629 / 10.605
rain_ano / Precipitation deviation from its long run average / 0.0591 / 1.1383 / -11.4844 / 24.1616
l_totaldamage / log total SHELDUS damages per capita / 0.3971 / 1.0241 / 0 / 7.6197
Notes: sample contains 7,129 county by year observations; total number of coastal counties619
Table A.9. Coastal Sample, Small Establishment Categories
n1_9 / n10_19 / n20_49l_loan_est / 0.1445 / 0.0701 / 0.2239**
(0.3032) / (0.0690) / (0.0980)
HI_c / 0.0010 / -0.0002 / -0.0004*
(0.0012) / (0.0001) / (0.0002)
population / 0.0309*** / 0.0047 / -0.0068**
(0.0095) / (0.0030) / (0.0027)
l_income / 4.6641 / 1.9059*** / 1.5674**
(3.4392) / (0.7111) / (0.7814)
l_wage_manuf_rate / 0.1699 / -0.4094 / -1.5631***
(2.1396) / (0.3684) / (0.5344)
l_totaldamage / -0.0040 / -0.0393*** / -0.0443**
(0.0367) / (0.0127) / (0.0180)
rain_ano / 0.0049 / -0.0082 / -0.0122
(0.0402) / (0.01100) / (0.0135)
L1rain_ano / -0.0326 / -0.0249 / -0.0411*
(0.0640) / (0.0159) / (0.0219)
L2rain_ano / -0.0079 / -0.0161 / -0.0181
(0.0500) / (0.0114) / (0.0152)
L3rain_ano / -0.0271 / -0.0231* / -0.0221
(0.0630) / (0.0127) / (0.0157)
L4rain_ano / 0.0210 / -0.0308** / -0.0196
(0.0625) / (0.0133) / (0.0140)
L5rain_ano / 0.0335 / -0.0297*** / -0.0047
(0.0449) / (0.0098) / (0.0111)
L.y / 0.8464*** / 0.8311*** / 1.2979***
(0.0541) / (0.1090) / (0.1333)
N / 7,129 / 7,129 / 7,129
Sargan (39) / 330.24*** / 220.02*** / 143.23***
Hansen (39) / 6.07 / 2.80 / 6.37
Hansen P-value / 0.81 / 0.99 / 0.78
AR(1) / -4.22*** / -1.40 / -3.64***
AR(2) / -0.80 / 0.18 / -0.59
AR(3) / -1.82* / -0.75 / -0.48
AR(4) / -0.74 / -0.70 / -0.54
Notes: Dependent variables are scaled by 100. Column headings correspond to total numbers of different establishment categories that serve as dependent variables in the GMM model. L.y represents the lag dependent variable. Instruments correspond to all available lags starting from the lag 5. Includes county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.10 Different log-transformation
n1_49ln(loan+1) / n1_49
ln(+ ) / n1_49
ln(loan+0.0001) / n1_49
ln(loan+100)
l_loan_est / 1.3469*** / 0.8486*** / 0.5812*** / 5.1057**
(0.5150) / (0.3087) / (0.2056) / (2.3237)
HI_c / -0.0021*** / -0.0021*** / -0.0022*** / -0.0021**
(0.0008) / (0.0008) / (0.0008) / (0.0008)
Population / 0.0165*** / 0.0161*** / 0.0157*** / 0.0167***
(0.0057) / (0.0057) / (0.0055) / (0.0058)
l_income / 14.0559*** / 14.3233*** / 14.4710*** / 14.1710***
(4.2869) / (4.3975) / (4.4971) / (4.4069)
l_wage_manuf_rate / -5.7909*** / -6.0179*** / -6.1911*** / -5.7060**
(2.2339) / (2.3108) / (2.3836) / (2.2522)
l_totaldamage / -0.2776*** / -0.2762*** / -0.2894*** / -0.3398**
(0.1052) / (0.1002) / (0.1024) / (0.1503)
rain_ano / -0.1455*** / -0.1542*** / -0.1636*** / -0.1366***
(0.0512) / (0.0526) / (0.0553) / (0.0517)
L1rain_ano / -0.2103*** / -0.2316*** / -0.2551*** / -0.1861**
(0.0761) / (0.0799) / (0.0861) / (0.0751)
L2rain_ano / -0.1152*** / -0.1251*** / -0.1356*** / -0.1061**
(0.0445) / (0.0464) / (0.0494) / (0.0440)
L3rain_ano / -0.0797*** / -0.0821*** / -0.0837*** / -0.0783***
(0.0278) / (0.0290) / (0.0301) / (0.0278)
L4rain_ano / -0.0041 / -0.0035 / -0.0017 / -0.0058
(0.0238) / (0.0246) / (0.0256) / (0.0249)
L5rain_ano / 0.0658** / 0.0673** / 0.0697** / 0.0723**
(0.0305) / (0.0308) / (0.0318) / (0.0347)
L.n1_49 / 0.9194*** / 0.9204*** / 0.9218*** / 0.9185***
(0.0246) / (0.0242) / (0.0238) / (0.0249)
N / 31,982 / 31,982 / 31,982 / 31,982
Sargan (29) / 5,285.48*** / 4,550.12*** / 3,743.54*** / 4,724.06***
Hansen (29) / 24.98 / 26.12 / 27.88 / 26.58
Hansen P-value / 0.68 / 0.62 / 0.52 / 0.59
Notes: Dependent variables are number of business establishments of 1-49 employment size and are log transformed using different logarithmic transformations. Column headings correspond to differences in log transformations. L.n1_49 represents the lag dependent variable. Instruments used in GMM models correspond to all available lags starting from the lag 5. All models include county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.11. Different log-transformation, total establishments
estln(loan+1) / est
ln(+ ) / est
ln(loan+0.0001) / est
ln(loan+100)
l_loan_est / 1.3383*** / 0.8581*** / 0.5946*** / 4.6116**
(0.4938) / (0.3008) / (0.2021) / (2.1996)
HI_c / -0.0023*** / -0.0024*** / -0.0024*** / -0.0023***
(0.0009) / (0.0009) / (0.0009) / (0.0009)
population / 0.017792** / 0.0175** / 0.0171** / 0.0179***
(0.0069) / (0.0069) / (0.0068) / (0.0069)
l_income / 15.7151*** / 16.0271*** / 16.1924*** / 15.7813***
(4.7971) / (4.9335) / (5.0451) / (4.8971)
l_wage_manuf_rate / -6.3610*** / -6.6095*** / -6.7946*** / -6.2457**
(2.4387) / (2.5255) / (2.6031) / (2.4471)
l_totaldamage / -0.2773*** / -0.2806*** / -0.2973*** / -0.3094**
(0.1017) / (0.0984) / (0.1016) / (0.1425)
rain_ano / -0.1510*** / -0.1610*** / -0.1717*** / -0.1369***
(0.0529) / (0.0546) / (0.0575) / (0.0524)
L1rain_ano / -0.2157*** / -0.2400*** / -0.2662*** / -0.1806**
(0.0756) / (0.0799) / (0.0866) / (0.0733)
L2rain_ano / -0.1224*** / -0.1335*** / -0.1451*** / -0.1088**
(0.0461) / (0.0483) / (0.0513) / (0.0447)
L3rain_ano / -0.0874*** / -0.0899*** / -0.0915*** / -0.0866***
(0.0306) / (0.0319) / (0.0330) / (0.0304)
L4rain_ano / -0.0096 / -0.0086 / -0.0065 / -0.0135
(0.0260) / (0.0269) / (0.0278) / (0.0267)
L5rain_ano / 0.0731** / 0.0752** / 0.0780** / 0.0768**
(0.0324) / (0.0329) / (0.0340) / (0.0358)
L.est / 0.9165*** / 0.9172*** / 0.9183*** / 0.9161***
(0.0281) / (0.0279) / (0.0275) / (0.0283)
N / 31,982 / 31,982 / 31,982 / 31,982
Sargan (29) / 6,179.84*** / 5,230.76*** / 4,257.29*** / 6,255.07***
Hansen (29) / 23.84 / 26.67 / 27.75 / 25.01
Hansen P-value / 0.74 / 0.59 / 0.53 / 0.68
Notes: Dependent variables are number of business establishments and are log transformed using different logarithmic transformations. Column headings correspond to differences in log transformations. L.est represents the lag dependent variable. Instruments used in GMM models correspond to all available lags starting from the lag 5. All models include county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.12. Model Comparison; total of all business establishments (est)
Anderson and Hsiao GMM / Two-Step feasible GMM / Arellano and Bond GMM / Blundell and BondSystem GMM / OLS / Panel FE
l_loan_est / 0.0221*** / 0.5018** / 1.3419*** / 1.3383*** / -0.0123** / 0.0045
(0.0074) / (0.2149) / (0.3990) / (0.4938) / (0.0062) / (0.0052)
HI_c / -0.00004 / -0.0008*** / 0.0017* / -0.0023*** / 0.000007 / -0.00006***
(0.00002) / (0.0002) / (0.0010) / (0.0009) / (0.000006) / (0.00002)
population / 0.06375** / 0.0181 / -0.0907** / 0.0178** / 0.0043*** / 0.0143***
(0.0288) / (0.0131) / (0.0437) / (0.0069) / (0.0009) / (0.0030)
l_income / 2.92096*** / 3.2291** / 26.2658*** / 15.7151*** / 0.5812*** / 3.0789***
(0.5750) / (1.4529) / (3.9307) / (4.7971) / (0.0853) / (0.2925)
l_wage_manuf_rate / 0.0244 / -1.3077** / -0.2873* / -6.3610*** / -0.2162*** / -0.0219
(0.0543) / (0.5261) / (0.1538) / (2.4387) / (0.0553) / (0.0376)
l_totaldamage / -0.0004 / -0.0979** / -0.2241*** / -0.2773*** / 0.0023 / -0.0016
(0.0035) / (0.0431) / (0.0708) / (0.1017) / (0.0046) / (0.0042)
rain_ano / -0.0133** / -0.0531** / -0.0285* / -0.1510*** / -0.0040 / 0.0086
(0.0058) / (0.0206) / (0.0160) / (0.0529) / (0.0066) / (0.0059)
L1rain_ano / 0.0221** / -0.0813*** / -0.1050*** / -0.2157*** / 0.0213*** / 0.0292***
(0.0089) / (0.0284) / (0.0332) / (0.0756) / (0.0076) / (0.0073)
L2rain_ano / 0.0502*** / -0.0500*** / 0.0020 / -0.1224*** / 0.0111 / 0.0208***
(0.0091) / (0.0184) / (0.0154) / (0.0461) / (0.0069) / (0.0079)
L3rain_ano / 0.0051 / -0.0545*** / 0.0026 / -0.0874*** / -0.0421*** / -0.0311***
(0.0143) / (0.0168) / (0.0193) / (0.0306) / (0.0063) / (0.0071)
L4rain_ano / -0.0018 / -0.0104 / 0.0324 / -0.0096 / -0.0248*** / -0.0268***
(0.0123) / (0.0174) / (0.0225) / (0.0260) / (0.0078) / (0.0076)
L5rain_ano / -0.0367*** / 0.0123 / 0.1074*** / 0.0731** / -0.0098 / -0.0159***
(0.0124) / (0.0148) / (0.02467) / (0.0324) / (0.0074) / (0.0059)
L.y / 0.5697*** / 0.9191*** / 0.8862*** / 0.9165*** / 0.9895*** / 0.8427***
(0.1793) / (0.0552) / (0.0442) / (0.0281) / (0.0031) / (0.0100)
N / 28,881 / 31,982 / 28,863 / 31,982 / 31,982 / 31,982
R2 / 0.46 / 0.99 / 0.91
Sargan / 3,061.33*** / 3,623.19*** / 6,179.84***
Hansen / 22.93* / 37.76 / 23.84
Hansen P-value / 0.09 / 0.30 / 0.74
Notes: Dependent variables are total number of business establishments and are scaled by 100. Column headings correspond to different models. L.y represents the lag dependent variable. Instruments used in GMM models correspond to all available lags starting from the lag 5. All models except for the OLS model include county fixed effects. All models include year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.13. Model Comparison; business establishment size 1-49
Anderson and Hsiao GMM / Two-Step feasible GMM / Arellano and Bond GMM / Blundell and BondSystem GMM / OLS / Panel FE
l_loan_est / 0.0167** / 0.3799** / 1.1365*** / 1.3469*** / -0.0120** / 0.0044
(0.0067) / (0.1697) / (0.3409) / (0.5150) / (0.0057) / (0.0047)
HI_c / -0.00004 / -0.0008*** / 0.0014 / -0.0021*** / 0.000006 / -0.000047***
(0.00002) / (0.0002) / (0.0009) / (0.0008) / (0.000006) / (0.000018)
Population / 0.0458*** / 0.0191* / -0.0666* / 0.0165*** / 0.0041*** / 0.0143***
(0.0176) / (0.0106) / (0.0354) / (0.0057) / (0.0008) / (0.0028)
l_income / 2.1710*** / 3.3150*** / 24.4985*** / 14.0559*** / 0.5032*** / 2.7309***
(0.3804) / (1.2655) / (3.9119) / (4.2869) / (0.0754) / (0.2599)
l_wage_manuf_rate / 0.0607 / -1.3159*** / -0.2836** / -5.7909*** / -0.2138*** / -0.0266
(0.0384) / (0.4572) / (0.1375) / (2.2339) / (0.0537) / (0.0340)
l_totaldamage / -0.0001 / -0.0746** / -0.1889*** / -0.2776*** / 0.0022 / -0.0012
(0.0031) / (0.0345) / (0.0602) / (0.1052) / (0.0043) / (0.0039)
rain_ano / -0.0154*** / -0.0488*** / -0.0297** / -0.1455*** / -0.0064 / 0.0061
(0.0053) / (0.0176) / (0.0143) / (0.0512) / (0.0063) / (0.0056)
L1rain_ano / 0.0227*** / -0.0683*** / -0.0915*** / -0.2103*** / 0.0185*** / 0.0261***
(0.0073) / (0.0231) / (0.0297) / (0.0761) / (0.0068) / (0.0066)
L2rain_ano / 0.0392*** / -0.0432*** / 0.0029 / -0.1152*** / 0.0109* / 0.02016***
(0.0079) / (0.0157) / (0.0135) / (0.0445) / (0.0063) / (0.0072)
L3rain_ano / -0.0072 / -0.0510*** / -0.0025 / -0.0797*** / -0.0400*** / -0.0300***
(0.0111) / (0.0147) / (0.0159) / (0.0278) / (0.0062) / (0.0068)
L4rain_ano / -0.0093 / -0.0131 / 0.0255 / -0.0041 / -0.0211*** / -0.0232***
(0.0091) / (0.0151) / (0.0191) / (0.0239) / (0.0070) / (0.0067)
L5rain_ano / -0.04673*** / 0.0074 / 0.0922*** / 0.0658** / -0.0111* / -0.0172***
(0.0099) / (0.0127) / (0.0224) / (0.0305) / (0.0067) / (0.0056)
L.y / 0.6861*** / 0.9130*** / 0.8550*** / 0.9193*** / 0.9899*** / 0.8506***
(0.1181) / (0.0468) / (0.0416) / (0.0246) / (0.0029) / (0.0100)
N / 28,881 / 31,982 / 28,863 / 31,982 / 31,982 / 31,982
R2 / 0.42 / 0.99 / 0.92
Sargan / 3,375.59*** / 3,848.50*** / 5,285.47***
Hansen / 18.79 / 38.13 / 24.98
Hansen P-value / 0.22 / 0.29 / 0.68
Notes: Dependent variables are number of business establishments of 1-49 employment size and are scaled by 100. Column headings correspond to different models. L.y represents the lag dependent variable. Instruments used in GMM models correspond to all available lags starting from the lag 5. All models except for the OLS model include county fixed effects. All models include year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
Table A.14. Different lags used as instruments in GMM estimation
n1_49 / n1_49 / n1_49 / est / est / estlag 5-7
(1) / lag 5-8
(2) / lag 5-10
(3) / lag 5-7
(4) / lag 5-8
(5) / lag 5-10
(6)
l_loan_est / 1.5735*** / 0.6757*** / 0.8274*** / 2.2789*** / 0.7211*** / 0.8549***
(0.2351) / (0.1408) / (0.1496) / (0.3221) / (0.1512) / (0.1595)
HI_c / -0.0009 / -0.0015** / -0.0026*** / -0.0008 / -0.0018** / -0.0029***
(0.0006) / (0.0007) / (0.0010) / (0.0007) / (0.0008) / (0.0011)
population / 0.0376*** / 0.0432*** / 0.0161*** / 0.0437** / 0.0486*** / 0.0171**
(0.0138) / (0.0124) / (0.0057) / (0.0181) / (0.0146) / (0.0069)
l_income / 7.4367** / 12.4547*** / 16.9844*** / 6.7266 / 14.8278*** / 19.2210***
(3.0803) / (3.8319) / (5.3565) / (4.1281) / (4.7837) / (6.0770)
l_wage_manuf_rate / -2.3071 / -3.9073* / -7.0248** / -1.9586 / -4.5930* / -7.8560**
(1.8802) / (2.0829) / (2.7603) / (2.4439) / (2.3962) / (3.0767)
l_totaldamage / -0.3166** / -0.1400 / -0.1740* / -0.4575** / -0.1511 / -0.1812*
(0.1510) / (0.1347) / (0.0955) / (0.1779) / (0.1287) / (0.0948)
rain_ano / -0.1111** / -0.0982* / -0.1374*** / -0.1354** / -0.1081** / -0.1479***
(0.0523) / (0.0506) / (0.0508) / (0.0631) / (0.0524) / (0.0537)
L1rain_ano / -0.1926** / -0.1267 / -0.1728** / -0.2600** / -0.1409* / -0.1858***
(0.0943) / (0.0867) / (0.0706) / (0.1109) / (0.0839) / (0.0715)
L2rain_ano / -0.0884* / -0.0746* / -0.1069** / -0.1125** / -0.0860* / -0.1182**
(0.0457) / (0.0436) / (0.0446) / (0.0548) / (0.0449) / (0.0475)
L3rain_ano / -0.0559*** / -0.0768*** / -0.0918*** / -0.0517** / -0.0864*** / -0.1014***
(0.0192) / (0.0235) / (0.0329) / (0.0238) / (0.0280) / (0.0369)
L4rain_ano / 0.0017 / -0.0219 / -0.0164 / 0.0125 / -0.0282 / -0.0216
(0.0221) / (0.0245) / (0.0273) / (0.0259) / (0.0270) / (0.0308)
L5rain_ano / 0.0376 / 0.0387 / 0.0679** / 0.0481 / 0.0482 / 0.0780**
(0.0301) / (0.0298) / (0.0329) / (0.0371) / (0.0320) / (0.0359)
L.y / 0.8361*** / 0.8067*** / 0.9182*** / 0.8210*** / 0.7927*** / 0.9150***
(0.0558) / (0.0495) / (0.0244) / (0.0693) / (0.0563) / (0.0280)
N / 31,982 / 31,982 / 31,982 / 31,982 / 31,982 / 31,982
Sargan / 1,053.74*** / 4,924.01*** / 7,171.37*** / 628.98*** / 5,136.78*** / 7,650.46***
Hansen / 7.77 / 18.54 / 22.26 / 6.7 / 19.47 / 22.43
Hansen P-value / 0.46 / 0.14 / 0.5 / 0.57 / 0.11 / 0.49
Notes: Dependent variables in columns (1)-(3) are number of business establishments of1-49 size category; Dependent variable in columns (4)-(6) correspond to the total number of business establishments. Column headings also include lags used in GMM estimation. L.y represents the lag dependent variable. All models include county and year fixed effects. Robust standard errors are reported in parenthesis and are clustered by county. * p<0.1; ** p<0.05; *** p<0.01
1