TABLES AVAILABLE FROM THE AUTHORS ON REQUEST (REVIEW APPENDIX)
REVIEW APPENDIX Table 1. Changes in government subsidies and import tariffs.
Date / Headline / SourcePanel A. Subsidy decreases.
05-Jul-07 / Germany cuts solar subsidy / LexisNexis
30-May-08 / Solar stocks shine through legal cloud; Cut to German subsidy could hurt smaller firms / LexisNexis
26-Jan-12 / German minister to cut solar subsidies sooner than planned / LexisNexis
23-Feb-12 / Germany to cut solar subsidies faster than expected / Reuters
27-Jun-12 / Germany to cut solar subsidies at 52GW / LexisNexis
08-Jul-13 / Germany pulls plug on solar subsidies / LexisNexis
20-Jan-14 / Germany Energy Minister Proposes Cuts to Renewable Subsidies, Industry Reacts / RenewableEnergyWorld
27-Jun-14 / German Lawmakers Vote to Reduce Renewable-Energy Subsidies / Bloomberg
Panel B. Subsidy increases.
11-May-12 / UPDATE 1-Germany's upper house suspends solar subsidy cuts / LexisNexis
Panel C. Import tariffs.
25-Sep-12 / European solar panel makers file subsidy complaint against China / LexisNexis
07-Nov-12 / WTO complaint filed on EU solar power subsidies / LexisNexis
08-Nov-12 / EU initiates anti-subsidy investigation on solar panel imports from China / European Commission
15-Jan-13 / Importiertes Solarglas aus China – EU leitet Antidumpinguntersuchung ein / European Commission
06-May-13 / EU to Slap Tariffs on Chinese Solar Panels / Wall StreetJournal
04-Jun-13 / China solar panel duties imposed by EU / BBC News
27-Jul-13 / Amicable solution in the EU-China anti-dumping case on solar panels / European Commission
27-Nov-13 / EU to impose 42.1% anti-dumping duties on Chinese solar glass firms / PV-Tech
02-Dec-13 / EU Nations Approve Pact With China on Solar-Panel Imports / Bloomberg
10-Apr-14 / EU backs duties on Chinese solar glass imports: sources / Reuters
30-Apr-14 / China to levy duties on EU polysilicon imports / LexisNexis
15-Aug-14 / China to stop imports of solar polysilicon from US, EU, Korea / SeeNews Renewables
This table contains changes in government subsidies and import tariffs identified by a LexisNexis and Google.com web search with the key words “subsidies”, “solar” and “photovoltaic”.
REVIEW APPENDIX Table 2. (Cumulative) average abnormal returns of a competitor portfolio around bankruptcy announcements of German solar firms when confounding events are eliminated.
Day / Events / (C)AAR[%] / CRU / KOL / COR-5 / 11 / -1.25 / -1.53 / -1.15 / -1.08
-4 / 11 / 0.60 / 0.74 / 1.23 / 1.60
-3 / 11 / 1.10 / 1.35 / 1.13 / 1.51
-2 / 11 / -1.22 / -1.50 / -1.00 / -1.19
-1 / 11 / -1.19 / -1.46 / -0.94 / -0.96
0 / 11 / -0.20 / -0.24 / -0.27 / -0.37
+1 / 11 / -1.16 / -1.43 / -1.10 / -1.03
+2 / 11 / 0.47 / 0.57 / 0.27 / 0.46
+3 / 11 / -1.60 / -1.97 / * / -1.89 / * / -1.89 / *
+4 / 11 / 0.96 / 1.18 / 0.68 / 0.68
+5 / 11 / -0.42 / -0.52 / -0.31 / -0.72
[-1;+1] / 11 / -2.55 / -1.81 / * / -2.14 / ** / -1.39
[-2;+2] / 11 / -3.30 / -1.81 / * / -2.12 / ** / -1.40
[-5;+5] / 11 / -3.92 / -1.45 / -1.50 / -0.92
[-10;+10] / 11 / -4.14 / -1.11 / -1.63 / -0.94
This table shows (cumulative) average abnormal returns ((C)AARs) around bankruptcy announcements of German listed solar firms. Abnormal returns are calculated against a German version of the 4-factor model described in Fama and French (1992, 1993, 1996) and Carhart(1997). The analysis is based on a sample that excludes observations with firm-specific or subsidy- and import tariff-related confounding events in the event window. CRU is the crude-dependence adjustment test proposed by Brown and Warner (1980: 223, 253). KOL is the Kolari and Pynnonen (2010) parametric test statistic and COR the Corrado and Zivney (1992) test statistic for the rank test. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels.
REVIEW APPENDIX Table 3. (Cumulative) average abnormal returns of firms announcing bankruptcy when firms with confounding events are excluded.
Day / Firms / (C)AAR[%] / CRU / KOL / COR-5 / 7 / -2.35 / -0.71 / -0.74 / -0.72
-4 / 7 / 4.26 / 1.29 / 0.31 / -0.05
-3 / 6 / 4.93 / 1.49 / 0.65 / 0.52
-2 / 7 / -3.52 / -1.06 / -1.25 / -1.48
-1 / 7 / -16.52 / -4.99 / *** / -1.90 / * / -2.31 / **
0 / 4 / -20.37 / -6.15 / *** / -1.90 / * / -2.28 / **
1 / 7 / -3.74 / -1.13 / -0.44 / -0.49
2 / 7 / -2.73 / -0.83 / -0.11 / -0.59
3 / 7 / 4.04 / 1.22 / 0.38 / 0.26
4 / 7 / -1.54 / -0.46 / 0.03 / -0.05
5 / 7 / -0.73 / -0.22 / -0.57 / -0.88
[-1;1] / 7 / -31.89 / -5.56 / *** / -4.02 / *** / -2.91 / ***
[-2;2] / 7 / -38.15 / -5.15 / *** / -4.32 / *** / -3.08 / ***
[-5;5] / 7 / -30.24 / -2.76 / *** / -2.04 / ** / -2.29 / **
[-10;10] / 7 / -33.05 / -2.18 / ** / -0.99 / -1.71 / *
This table shows (cumulative) average abnormal returns (CAARs) around bankruptcy announcements of German solar firms. Abnormal returns are calculated against a German version of the 4-factor model described in Fama and French (1992, 1993, 1996) and Carhart(1997). CRU is the crude-dependence adjustment test proposed by Brown and Warner (1980: 223, 253). KOL is the Kolari and Pynnonen(2010) parametric test statistic and COR the Corrado and Zivney(1992) test statistic for the rank test. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. Firms with confounding events are excluded. Confounding events are identified using the Capital IQ key development and I/B/E/S earnings announcement databases.
REVIEW APPENDIX Table 4. Results of the cross-sectional analysis of information externalities when observations with confounding events are excluded.
Hyp. / Exp. / XX / [-1;+1] / XX / [-2;+2] / XX / [-5;+5] / XX / [-10;+10]LEV / 3 / - / 2.87 / 5.00 / -6.21 / -0.84 / -17.87 / *** / -14.33 / *** / -13.61 / -12.46
(0.68) / (1.12) / (-1.45) / (-0.21) / (-5.70) / (-5.34) / (-1.50) / (-1.19)
HHI / 4 / ? / -94.78 / ** / / -98.98 / / -106.22 / / -87.73 /
(-2.56) / / (-0.93) / / (-1.32) / / (-1.17) /
SIZE / ? / -1.33 / -1.43 / -1.56 / -1.12 / -2.45 / -2.54 / -3.13 / -3.22
(-1.43) / (-1.35) / (-1.23) / (-0.81) / (-1.71) / (-1.75) / (-1.24) / (-1.27)
CORR / ? / -6.42 / -10.18 / 4.66 / -7.22 / 10.19 / 5.83 / -16.65 / -7.98
(-1.06) / (-1.50) / (0.55) / (-0.98) / (0.78) / (0.31) / (-1.18) / (-0.41)
SIZEB / ? / 0.73 / *** / / 1.52 / *** / / 1.83 / *** / / 1.87 / * /
(5.64) / / (3.45) / / (3.65) / / (1.90) /
SUB / ? / -8.00 / -8.84 / -7.22 / -8.41 / -10.65 / -11.28 / -14.07 / -12.07
(-1.40) / (-1.17) / (-1.30) / (-1.17) / (-1.01) / (-0.77) / (-1.48) / (-0.94)
GCGC / + / 7.15 / 8.27 / 6.79 / 7.57 / 8.30 / 9.73 / 7.70 / 7.44
(1.66) / (1.53) / (1.62) / (1.44) / (1.08) / (1.02) / (1.22) / (1.01)
DIV / + / 0.34 / 0.69 / 2.47 / 2.53 / 2.65 / 3.62 / 3.17 / 4.07
(0.14) / (0.32) / (0.62) / (0.66) / (0.51) / (0.69) / (0.56) / (0.76)
CAR[+11;+30] / ? / 0.59 / -3.33 / -1.71 / -3.14 / -3.44 / -12.93 / -3.90 / -13.30
(0.10) / (-0.31) / (-0.26) / (-0.28) / (-0.29) / (-0.71) / (-0.33) / (-0.63)
Const. / 4.10 / -24.12 / 4.51 / -19.50 / 14.41 / -36.02 / 15.08 / -29.91
(0.61) / (-1.21) / (0.43) / (-0.85) / (1.66) / (-1.04) / (0.94) / (-0.73)
EVENT / NO / YES / NO / YES / NO / YES / NO / YES
Adj. R2 / 5.26% / 0.99% / 6.96% / 21.60% / 7.36% / 8.83% / 4.13% / 0.68%
F-Statistic / 14.99 / *** / n/a / 12.42 / *** / n/a / 11.45 / *** / n/a / 7.18 / *** / n/a
Events / 11 / 11 / 11 / 11 / 11 / 11 / 11 / 11
Firms / 20 / 20 / 20 / 20 / 20 / 20 / 20 / 20
Obs. / 74 / 74 / 74 / 74 / 74 / 74 / 74 / 74
This table shows the results for the regressions CARi[1, 2] = βo + β1LEVi + β2HHIi + β3SIZEi + β4CORRi + β5SIZEBi + β6SUBi + β7GCGCi + β8DIVi + β9CAR[+11,+30]i + εi and CARi[1, 2] = βo + β1LEVi + β2SIZEi + β3CORRi + β4SUBi + β5GCGCi + β6DIVi + β7CAR[+11,+30]i + ΣβeventEVENTi + εi. t-statistics based on heteroscedasticity consistent standard errors that are also clustered at the event level are reported in parentheses. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. The variables are defined as outlined in Table 6.
REVIEW APPENDIX Table 5. Results of the cross-sectional analysis of competition effects (CARi[1, 2]>0) when observations with confounding events are excluded.
Hyp. / Exp. / XX / [-1;+1] / XX / [-2;+2] / XX / [-5;+5] / XX / [-10;+10]LEV / 3 / - / -1.22 / 0.44 / 8.48 / 22.66 / -6.70 / -9.49 / 4.01 / 9.69
(-0.18) / (0.07) / (0.89) / (1.75) / (-1.37) / (-1.26) / (0.46) / (1.02)
HHI / 4 / + / -52.00 / / -54.89 / / -42.57 / / -46.48 /
(-1.86) / / (-0.80) / / (-0.93) / / (-1.46) /
SIZE / - / -0.06 / 0.37 / 0.31 / 3.62 / 2.43 / * / 3.15 / * / 0.99 / 1.43
(-0.06) / (0.24) / (0.17) / (1.51) / (1.94) / (2.03) / (0.49) / (0.39)
CORR / ? / 4.67 / 1.79 / 7.91 / -15.25 / 4.12 / 7.87 / 2.61 / 14.14
(0.52) / (0.12) / (0.72) / (-0.57) / (0.36) / (0.42) / (0.15) / (0.53)
SIZEB / ? / 1.55 / / 1.65 / / 1.65 / / 0.24 /
(1.02) / / (1.57) / / (1.85) / / (0.25) /
SUB / ? / -8.77 / -9.04 / -1.76 / -11.81 / 3.49 / 1.54 / -8.93 / ** / -4.16
(-1.18) / (-1.26) / (-0.25) / (-0.88) / (1.01) / (0.62) / (-2.58) / (-0.53)
GCGC / ? / 3.14 / 1.66 / 3.92 / -0.46 / -7.84 / * / -6.33 / -4.52 / -6.92
(1.39) / (0.80) / (0.79) / (-0.15) / (-2.26) / (-1.62) / (-1.07) / (-0.49)
DIV / ? / -2.26 / -0.32 / 1.36 / 2.70 / 2.24 / 4.33 / -3.02 / -3.86
(-0.66) / (-0.11) / (0.60) / (1.18) / (0.56) / (0.90) / (-0.63) / (-0.44)
CAR[+11;+30] / ? / 6.79 / 8.20 / 16.66 / 31.04 / 7.79 / 2.69 / 11.86 / 6.74
(1.06) / (0.81) / (1.30) / (1.64) / (1.39) / (0.31) / (1.22) / (0.30)
Const. / 1.26 / -1.75 / -7.17 / -35.90 / -0.77 / -7.85 / 8.02 / 0.89
(0.12) / (-0.15) / (-0.57) / (-1.72) / (-0.10) / (-1.02) / (1.09) / (0.07)
EVENT / NO / YES / NO / YES / NO / YES / NO / YES
Adj. R2 / -8.46% / 11.10% / 15.90% / 45.20% / 21.50% / 22.80% / -3.93% / -17.20%
F-Statistic / n/a / n/a / n/a / n/a / n/a / n/a / n/a / n/a
Events / 8 / 8 / 7 / 7 / 9 / 9 / 9 / 9
Firms / 15 / 15 / 14 / 14 / 15 / 15 / 14 / 14
Obs. / 26 / 26 / 23 / 23 / 30 / 30 / 29 / 29
This table shows the results for the regressions CARi[1, 2] = βo + β1LEVi + β2HHIi + β3SIZEi + β4CORRi + β5SIZEBi + β6SUBi + β7GCGCi + β8DIVi + β9CAR[+11,+30]i + εi and CARi[1, 2] = βo + β1LEVi + β2SIZEi + β3CORRi + β4SUBi + β5GCGCi + β6DIVi + β7CAR[+11,+30]i + ΣβeventEVENTi + εi. t-statistics based on heteroscedasticity consistent standard errors that are also clustered at the event level are reported in parentheses. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. The variables are defined as outlined in Table 6.
REVIEW APPENDIX Table 6. Results of the cross-sectional analysis of negative information externalities (CARi[1,2]<0) when observations with confounding events are excluded.
Hyp. / Exp. / XX / [-1;+1] / XX / [-2;+2] / XX / [-5;+5] / XX / [-10;+10]LEV / 3 / - / -1.22 / 1.97 / -12.26 / *** / -6.10 / -24.91 / ** / -17.32 / -1.97 / 1.07
(-0.31) / (0.35) / (-3.23) / (-1.39) / (-2.35) / (-1.51) / (-0.19) / (0.08)
HHI / 4 / ? / -112.15 / ** / / -133.57 / * / / -192.54 / / -123.91 /
(-2.50) / / (-2.04) / / (-1.17) / / (-1.01) /
SIZE / + / -2.32 / -2.64 / -1.65 / -1.04 / -4.94 / -5.70 / -1.82 / -1.07
(-1.63) / (-1.38) / (-1.12) / (-0.64) / (-1.59) / (-1.76) / (-0.75) / (-0.40)
CORR / ? / -8.65 / 14.49 / -11.14 / 9.27 / 5.29 / 3.94 / -13.36 / 7.49
(-0.64) / (0.52) / (-0.60) / (0.37) / (0.29) / (0.23) / (-0.51) / (0.17)
SIZEB / ? / -0.02 / / 0.83 / / 0.88 / / 0.45 /
(-0.05) / / (1.59) / / (1.47) / / (0.36) /
SUB / ? / -3.30 / -8.80 / -1.93 / -7.78 / -0.46 / -4.43 / -13.55 / -16.52
(-0.50) / (-0.93) / (-0.27) / (-0.78) / (-0.05) / (-0.29) / (-1.01) / (-0.92)
GCGC / + / 11.15 / 10.79 / 10.31 / 9.22 / 12.48 / 14.89 / 11.34 / 8.01
(1.69) / (1.67) / (1.55) / (1.30) / (1.31) / (1.21) / (1.81) / (1.03)
DIV / + / 3.90 / 5.86 / 2.44 / 4.79 / 3.29 / 14.47 / -0.35 / 3.51
(1.76) / (1.66) / (0.80) / (1.07) / (0.48) / (1.60) / (-0.05) / (0.41)
CAR[+11;+30] / ? / -0.72 / -1.82 / -1.74 / -0.99 / -7.98 / -21.31 / * / -7.01 / -6.37
(-0.12) / (-0.18) / (-0.36) / (-0.11) / (-0.76) / (-2.05) / (-0.72) / (-0.32)
Const. / 8.30 / -15.16 / 8.35 / 10.83 / 29.04 / -40.34 / * / -2.76 / -37.31
(1.09) / (-0.63) / (1.17) / (0.97) / (1.38) / (-2.00) / (-0.19) / (-0.94)
EVENT / NO / YES / NO / YES / NO / YES / NO / YES
Adj. R2 / 7.71% / 16.40% / 9.31% / 19.50% / 12.10% / 34.20% / 5.31% / 13.30%
F-Statistic / 35.23 / *** / n/a / 17.68 / *** / n/a / 1063.00 / *** / n/a / 2.60 / * / n/a
Events / 11 / 11 / 11 / 11 / 11 / 11 / 11 / 11
Firms / 19 / 19 / 19 / 19 / 18 / 18 / 18 / 18
Obs. / 48 / 48 / 51 / 51 / 44 / 44 / 45 / 45
This table shows the results for the regressions CARi[1, 2] = βo + β1LEVi + β2HHIi + β3SIZEi + β4CORRi + β5SIZEBi + β6SUBi + β7GCGCi + β8DIVi + β9CAR[+11,+30]i + εi and CARi[1, 2] = βo + β1LEVi + β2SIZEi + β3CORRi + β4SUBi + β5GCGCi + β6DIVi + β7CAR[+11,+30]i + ΣβeventEVENTi + εi. t-statistics based on heteroscedasticity consistent standard errors that are also clustered at the event level are reported in parentheses. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. The variables are defined as outlined in Table 6.
REVIEW APPENDIX Table 7. Results of the logit regression to estimate the relationship information externalities and the probability of default when observations with confounding events are excluded.
Hyp. / Exp. / XX / [-1;+1] / XX / [-2;+2] / XX / [-5;+5] / XX / [-10;+10]CAR[-;+] / 5 / - / -0.04 / -0.09 / -0.03 / -0.10 / -0.02 / * / -0.02 / -0.02 / * / -0.02
(-1.55) / (-1.35) / (-1.26) / (-0.96) / (-1.69) / (-1.52) / (-1.66) / (-0.65)
CAR[+11;+30] / - / -0.01 / -0.02 / -0.01 / -0.02 / -0.01 / -0.02 / -0.01 / -0.02
(-0.73) / (-1.03) / (-0.85) / (-1.35) / (-0.79) / (-1.11) / (-0.75) / (-0.77)
LEV / + / 4.00 / ** / 14.45 / *** / 3.53 / ** / 15.00 / *** / 3.52 / ** / 11.89 / *** / 3.60 / ** / 12.08 / ***
(-2.50) / (-3.22) / (-2.47) / (-3.19) / (-2.44) / (-4.12) / (-2.44) / (-3.78)
HHI / ? / 14.51 / * / / 16.08 / ** / / 16.21 / ** / / 16.65 / * /
(-1.71) / () / (-2.15) / () / (-2.05) / () / (-1.95) / ()
SIZE / ? / 0.34 / 0.55 / ** / 0.32 / 0.30 / 0.34 / 0.44 / 0.35 / 0.45
(-1.57) / (-2.12) / (-1.49) / (-0.85) / (-1.57) / (-1.34) / (-1.41) / (-0.89)
CORR / ? / 3.51 / -12.80 / * / 4.07 / -15.20 / * / 3.68 / -10.54 / 3.04 / -10.70 / *
(-0.76) / (-1.83) / (-0.87) / (-1.85) / (-0.87) / (-1.62) / (-0.78) / (-1.74)
SIZEB / ? / -0.05 / / -0.04 / / -0.05 / / -0.04 /
(-0.72) / () / (-0.53) / () / (-0.72) / () / (-0.58) / ()
SUB / ? / -0.30 / -3.11 / -0.23 / -4.06 / -0.05 / -2.44 / 0.04 / -2.23
(-0.26) / (-1.63) / (-0.20) / (-1.42) / (-0.04) / (-1.43) / (-0.03) / (-1.45)
GCGC / - / -0.27 / -0.10 / -0.34 / 0.10 / -0.38 / -0.27 / -0.44 / -0.25
(-0.31) / (-0.08) / (-0.40) / (-0.06) / (-0.46) / (-0.19) / (-0.47) / (-0.14)
DIV / + / -1.65 / -1.65 / -1.67 / -1.49 / -1.61 / -1.55 / -1.60 / -1.59
(-1.23) / (-0.93) / (-1.26) / (-1.14) / (-1.25) / (-0.92) / (-1.30) / (-1.01)
Const. / -7.18 / *** / -37.24 / -7.04 / *** / -37.41 / -6.93 / *** / -35.42 / -7.09 / *** / -34.19
(-4.04) / (-0.21) / (-4.23) / (-0.01) / (-4.35) / (-0.01) / (-3.87) / (-0.00)
EVENT / NO / YES / NO / YES / NO / YES / NO / YES
Pseudo R2 / 22.72% / 52.78% / 22.70% / 53.10% / 21.51% / 49.11% / 21.65% / 49.94%
Events / 11 / 11 / 11 / 11 / 11 / 11 / 11 / 11
Firms / 20 / 20 / 20 / 20 / 20 / 20 / 20 / 20
Obs. / 74 / 74 / 74 / 74 / 74 / 74 / 74 / 74
This table shows the results for the logit regressions DEFi = βo + β1CARi[1, 2] + β2CARi[+11, +30] + β3LEVi + β4HHIi + β5SIZEi + β6CORRi + β7SIZEBi+ β8SUBi + β9GCGCi + β10DIVi + εi and DEFi = βo + β1CARi[1, 2] + β2CARi[+11, +30] + β3LEVi + β4SIZEi + β5CORRi + β6SUBi + β7GCGCi + β8DIVi + ΣβeventEVENTi + εi. z-statistics based on standard errors that are also clustered at the event level are reported in parentheses. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. The variables are defined as outlined in Table 6..
REVIEW APPENDIX Table 8. (Cumulative) average abnormal returns of a competitor portfolio around bankruptcy announcements of German solar firms when portfolio returns are calculated on a value-weighted basis instead of an equal-weighted basis.
Day / Events / (C)AAR[%] / CRU / KOL / COR-5 / 15 / 0.11 / 0.18 / 0.04 / -0.29
-4 / 15 / -0.24 / -0.38 / -0.21 / -0.47
-3 / 15 / 0.24 / 0.39 / 0.55 / 0.90
-2 / 15 / -0.22 / -0.35 / -0.42 / -0.42
-1 / 15 / -0.25 / -0.39 / -0.20 / -1.13
0 / 15 / -0.95 / -1.51 / -1.29 / -1.49
+1 / 15 / -0.44 / -0.70 / -0.76 / -1.43
+2 / 15 / -0.21 / -0.33 / -0.15 / -0.86
+3 / 15 / 0.11 / 0.18 / 0.02 / -0.15
+4 / 15 / 0.18 / 0.28 / 0.26 / 0.47
+5 / 15 / 0.46 / 0.74 / 0.81 / 1.44
[-1;+1] / 15 / -1.64 / -1.50 / -1.19 / -2.31 / **
[-2;+2] / 15 / -2.07 / -1.47 / -0.79 / -2.38 / **
[-5;+5] / 15 / -1.19 / -0.57 / -0.38 / -1.08
[-10;+10] / 15 / -2.54 / -0.88 / -0.58 / -1.29
This table shows (cumulative) average abnormal returns (CAARs) around bankruptcy announcements of German solar firms. Abnormal returns are calculated against a German version of the 4-factor model described in Fama and French (1992, 1993, 1996) and Carhart(1997). CRU is the crude-dependence adjustment test proposed by Brown and Warner (1980: 223, 253). KOL is the Kolari and Pynnonen(2010) parametric test statistic and COR the Corrado and Zivney(1992) test statistic for the rank test. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels. Portfolio returns are calculated on a value-weighted basis.
REVIEW APPENDIX Table 9. (Cumulative) average abnormal returns of a competitor portfolio around bankruptcy announcements of German solar firms when the market model is applied.
Day / Events / (C)AAR[%] / CRU / KOL / COR-5 / 15 / -1.21 / -1.98 / ** / -1.86 / * / -1.93 / *
-4 / 15 / -0.13 / -0.21 / 0.01 / 0.36
-3 / 15 / 0.76 / 1.23 / 1.32 / 1.98 / **
-2 / 15 / -0.22 / -0.36 / -0.47 / -0.39
-1 / 15 / -0.79 / -1.29 / -1.03 / -1.19
0 / 15 / -0.65 / -1.07 / -1.14 / -1.33
2 / 15 / -0.10 / -0.16 / -0.20 / 0.21
3 / 15 / -1.38 / -2.25 / ** / -2.62 / *** / -2.63 / ***
4 / 15 / 0.48 / 0.79 / 0.50 / 0.41
5 / 15 / -0.32 / -0.52 / -0.36 / -0.35
[-1;1] / 15 / -2.35 / -2.21 / ** / -2.14 / ** / -2.19 / **
[-2;2] / 15 / -2.67 / -1.95 / * / -1.91 / * / -1.79 / *
[-5;5] / 15 / -4.47 / -2.20 / ** / -2.34 / ** / -1.83 / *
[-10;10] / 15 / -4.03 / -1.43 / -1.28 / -1.21
This table shows (cumulative) average abnormal returns (CAARs) around bankruptcy announcements of German solar firms. Abnormal returns are calculated against a German version of the market (single factor) model. CRU is the crude-dependence adjustment test proposed by Brown and Warner (1980: 223, 253). KOL is the Kolari and Pynnonen(2010) parametric test statistic and COR the Corrado and Zivney(1992) test statistic for the rank test. Asterisks indicate significance at the 10% [*], 5% [**] and 1% [***] levels.
1