Individualism and Stock Price Crash Risk

ONLINE APPENDIX

Individualism and Stock Price Crash Risk

SUPPLEMENTARY MATERIAL

Note A1 / Alternative Model Specifications
Table A1 / Individualism and Crash Risk: Additional Country-Level Control Variables
Table A2 / Alternative Individualism Indexes and Crash Risk
Table A3 / Schwartz’s Indexes and Crash Risk in Non-US Firms
Table A4 / Factor Scores of GLOBE’s Indexes and Crash Risk
Table A5 / Individualism and Crash Risk: Other Robustness Checks
Table A6 / Managerial Discretionary Disclosure Ability and Crash Risk

Note A1: Alternative Model Specifications

We apply alternative specifications to the baseline regression modeland report the results in Panel D of Table A5 in this online appendix.

First, we estimate the weighted-least-squares regressions with a weight set equal to the inverse of the number of firm-year observations in each country. This method corrects for the uneven country representation in the firm-year sample.

Second, we employ the Fama-MacBeth estimation procedure to address the cross-sectional correlation in the residuals. Specifically, we estimate a separate regression in each cross-section and then take the mean of the coefficient estimates across years, where t-statistics are calculated from Newey-West standard errors adjusted for the serial dependence in the past five years.

Third, we transform each firm-level variable by taking the difference between its raw value and its median value within each country, and then enter these within-country median-adjusted firm-level variables into the regressions. This approach alleviates the endogeneity concern associated with omitted country-level variables that correlate with firm-level explanatory variables.

Fourth, we employ the hierarchical linear modeling (HLM) technique to deal with the multilevel structure of our data.The HLM cleanly partitions the variation in crash risk into what is explained by firm-level independent variables and what is explained by country-level independent variables. The HLM also rectifies the disproportionate country representation, as it estimates at both the firm and country levels and places on country-level variables a weight that is inversely related to the sample size in each country. To perform the HLM, we transform the independent variables in the baseline model following these procedures: (1) center each country-level independent variable by its grand mean and attach the suffix “_ctry”; (2) center each firm-level independent variable by its grand mean, and compute the country mean of each grand-mean-centered firm-level independent variable and attach the suffix “_ctrymean”; and (3) center each grand-mean-centered firm-level independent variable by its country mean, and add the suffix “_firmdev”. These transformed variables are used to estimate an iterative maximum likelihood random effect model. The coefficients of IDV_ctry are reported in the results.

Fifth, we correct the heteroscedasticity-robust standard errors in the baseline OLSregressions for country-level clustering, year-level clustering, two-way clustering by country and year, and two-way clustering by firm and year, respectively. These corrections reduce any potential bias associated with within-cluster correlations of the residuals.

Overall, the results show that crash risk is positively and significantly associated with individualism.

1

Table A1Individualism and crash risk: additional country-level control variables

This table presents results for the impact of individualism on crash risk by including additional country-level control variables in the baseline OLS regressions. The t-statistics based on robust standard errors clustered by firm are reported in brackets. ***, **, and *indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Dependent variable= / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT
[1] / [2] / [3] / [4] / [5] / [6] / [7] / [8] / [9] / [10] / [11] / [12]
IDV / 0.282*** / 0.143*** / 0.148*** / 0.119*** / 0.059*** / 0.077*** / 0.170*** / 0.085*** / 0.101*** / 0.133*** / 0.069*** / 0.077***
[13.78] / [14.39] / [9.32] / [8.00] / [8.05] / [6.83] / [14.68] / [15.17] / [11.76] / [10.66] / [11.24] / [8.15]
Predominant Religions:
Catholicism / -0.059*** / -0.027*** / -0.056***
[-4.71] / [-4.32] / [-5.82]
Protestantism / -0.027** / -0.012* / -0.031***
[-2.22] / [-1.95] / [-3.30]
Orthodox / 0.020 / 0.015 / -0.025
[0.93] / [1.36] / [-1.39]
Islam / 0.017 / 0.013** / -0.014
[1.33] / [2.09] / [-1.41]
Buddhism / 0.052*** / 0.030*** / 0.001
[4.08] / [4.77] / [0.13]
Legal Origins:
English / -0.008 / -0.008 / 0.005
[-0.43] / [-0.86] / [0.28]
French / -0.116*** / -0.061*** / -0.060***
[-5.94] / [-6.20] / [-3.59]
German / -0.024 / -0.015 / -0.007
[-1.24] / [-1.50] / [-0.39]
Scandinavian / -0.045** / -0.028*** / -0.003
[-2.10] / [-2.58] / [-0.17]
Governance Quality:
Governance / -0.018*** / -0.012*** / -0.004
[-3.50] / [-4.43] / [-1.00]
Daily Stock Price Limit Rules:
Price_Limit / -0.012* / -0.001 / -0.015***
[-1.70] / [-0.40] / [-2.82]
Controls / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Industry-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Year-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Adj. R2 / 0.038 / 0.043 / 0.021 / 0.038 / 0.043 / 0.022 / 0.037 / 0.042 / 0.021 / 0.037 / 0.042 / 0.021
N / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604 / 181,604

1

Table A2Alternative individualism indexes and crash risk

This table presents OLS regressionsof crash risk onalternative individualism indexes. We rescale all individualism indexes from Hofstede, Schwartz, GLOBE, and World Values Survey, as well as the first factor of the factor analysis of the above indexes, to a range 0–1. Control variables and industry- and year-fixed effects are included but suppressed. To conserve space, only the coefficients of individualism are presented. The t-statistics based on robust standard errors clustered by firm are reported in brackets. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Dependent variable= / NCSKEW / DUVOL / COUNT
[1] / [2] / [3]
Hofstede
Individualism / 0.114*** / 0.054*** / 0.074***
[15.69] / [15.70] / [13.69]
Schwartz
Affective autonomy / 0.082*** / 0.038*** / 0.062***
[5.87] / [5.66] / [5.94]
Intellectual autonomy / -0.056*** / -0.021*** / -0.030***
[-4.51] / [-3.54] / [-3.07]
Opposite embeddedness / -0.024* / -0.010 / -0.003
[-1.76] / [-1.53] / [-0.24]
GLOBE
Opposite institutional collectivism values / 0.147*** / 0.074*** / 0.098***
[15.60] / [15.71] / [12.90]
Opposite institutional collectivism practices / -0.002 / -0.002 / -0.001
[-0.27] / [-0.54] / [-0.11]
Opposite in-group collectivism values / -0.058*** / -0.025*** / -0.035***
[-5.43] / [-4.81] / [-4.22]
Opposite in-group collectivism practices / 0.213*** / 0.107*** / 0.143***
[14.48] / [14.89] / [12.82]
World Values Survey
Income inequality as incentives for individual effort / 0.066*** / 0.028** / 0.031*
[2.82] / [2.35] / [1.69]
Individualism Factor Score / 0.101*** / 0.050*** / 0.078***
[7.48] / [7.58] / [7.47]

1

Table A3 Schwartz’s indexes and crash risk in non-US firms

This table presents OLS regressions of crash risk on Schwartz’s affective autonomy, intellectual autonomy, and opposite embeddedness indexes for non-US firms. Opposite embeddedness equals the embeddedness index multiplied by −1. All the indexes are rescaled to a 0−1 range.The t-statistics based on robust standard errors clustered by firm are reported in brackets. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Dependent variable= / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT
[1] / [2] / [3] / [4] / [5] / [6] / [7] / [8] / [9]
Affective autonomy / 0.100*** / 0.046*** / 0.073***
[7.15] / [6.77] / [6.85]
Intellectual autonomy / 0.063*** / 0.037*** / 0.050***
[4.23] / [4.99] / [4.23]
Opposite embeddedness / 0.078*** / 0.038*** / 0.067***
[5.10] / [5.08] / [5.55]
Controls / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Industry-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Year-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Adj. R2 / 0.032 / 0.038 / 0.017 / 0.032 / 0.038 / 0.016 / 0.032 / 0.038 / 0.016
N / 113,909 / 113,909 / 113,909 / 113,909 / 113,909 / 113,909 / 113,909 / 113,909 / 113,909

1

Table A4Factor scores of GLOBE’s indexes and crash risk

This table presents OLS regressions of crash risk on the individualism factor scores derived from GLOBE’s indexes. The index with the prefix “opposite” is therelevant collectivism index multiplied by −1. Factor score of opposite institutional collectivism values and practices is the first factor of the factor analysis of opposite institutional collectivism values and opposite institutional collectivism practices indexes. Factor score of opposite in-group collectivism values and practices is the first factor of the factor analysis ofopposite in-group collectivism values and opposite in-group collectivism practices indexes. Factor score of opposite institutional and in-group collectivism values and practices is the first factor of all four-dimensional opposite collectivism indexes. All the factor scores are rescaled to a 0−1 range. The t-statistics based on robust standard errors clustered by firm are reported in brackets. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Dependent variable= / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT / NCSKEW / DUVOL / COUNT
[1] / [2] / [3] / [4] / [5] / [6] / [7] / [8] / [9]
Factor score of opposite institutional collectivism values and practices / 0.066*** / 0.034*** / 0.043***
[6.93] / [7.25] / [5.80]
Factor score of opposite in-group collectivism values and practices / 0.077*** / 0.045*** / 0.058***
[4.55] / [5.37] / [4.39]
Factor score of opposite institutional and in-group collectivism values and practices / 0.108*** / 0.054*** / 0.071***
[11.19] / [11.48] / [9.35]
Controls / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Industry-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Year-fixed effect / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes / Yes
Adj. R2 / 0.036 / 0.041 / 0.020 / 0.036 / 0.040 / 0.020 / 0.036 / 0.041 / 0.021
N / 179,281 / 179,281 / 179,281 / 179,281 / 179,281 / 179,281 / 179,281 / 179,281 / 179,281

1

Table A5Individualism and crash risk: other robustness checks

This table presents robustness test results for the impact of individualism on crash risk. In Panel A, we use alternative market models to compute firm-specific stock returns. In Panel B, we use a variant of the individualism index that is orthogonal to GDP per capita. In Panel C, we rerun the baseline regressions for non-US (and non-Japanese) firms and balanced panels. In Panel D, we apply alternative model specifications to the baseline regressions. Control variables and industry- and year-fixed effects are included but suppressed. To conserve space, only the coefficients of individualism are presented. The t-statistics based on cluster-robust standard errors are reported in brackets. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

Panel A: Alternative market models to calculate firm-specific stock returns
Single Market Model / 0.201*** / 0.098*** / 0.118***
[18.99] / [19.60] / [15.51]
World Market Model / 0.168*** / 0.083*** / 0.102***
[18.29] / [18.69] / [14.76]
Panel B: Orthogonalization of individualism against GDP per capita
Orthogonal IDV / 0.158*** / 0.076*** / 0.103***
[16.64] / [16.70] / [14.50]
Panel C: Alternative samples
Exclude the US / 0.111*** / 0.051*** / 0.073***
[9.66] / [9.30] / [8.45]
Exclude the US and Japan / 0.156*** / 0.075*** / 0.097***
[12.37] / [12.30] / [10.25]
Balanced Panel 2004 ̶ 2013 / 0.159*** / 0.073*** / 0.099***
[6.59] / [6.25] / [5.22]
Balanced Panel 2004 ̶ 2007 / 0.165*** / 0.078*** / 0.103***
[8.59] / [8.44] / [7.14]
Balanced Panel 2010 ̶ 2013 / 0.149*** / 0.070*** / 0.097***
[7.60] / [7.40] / [6.45]
Panel D: Alternative model specifications
Weighted Least Squares / 0.178*** / 0.090*** / 0.094***
[9.30] / [9.19] / [6.29]
Fama-MacBeth / 0.127*** / 0.062*** / 0.074***
[5.08] / [5.37] / [3.50]
Within-Country Median Adjustment / 0.075*** / 0.019*** / 0.133***
[8.29] / [4.37] / [19.23]
Hierarchical Linear Model / 0.167*** / 0.089*** / 0.098***
[3.83] / [3.68] / [3.60]
Clustering by Country / 0.148*** / 0.071*** / 0.096***
[4.84] / [4.35] / [4.20]
Clustering by Year / 0.148*** / 0.071*** / 0.096***
[5.62] / [5.06] / [5.79]
Two-Way Clustering by Country and Year / 0.148*** / 0.071*** / 0.096***
[5.09] / [4.60] / [4.33]
Two-Way Clustering by Firm and Year / 0.148*** / 0.071*** / 0.096***
[5.70] / [5.13] / [5.88]

1

Table A6 Managerial discretionary disclosure ability and crash risk

This table presents OLS regression results for the impact of managerial discretionary disclosure ability on crash risk conditional on the level of individualism. The discretionary disclosure ability is proxied by firm size in Panel A and by analyst coverage in Panel B. ANALYST is the natural logarithm of 1 plus the number of analysts who made forecasts about the firm’s earnings (data source: I/B/E/S). All time-varying independent variables are lagged by one year relative to the dependent variable. The t-statistics based on robust standard errors clustered by firm are reported in brackets. ***, **, and *indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

NCSKEW / DUVOL / COUNT
[1] / [2] / [3]
Panel A: Firm size
IDV / 0.068 / 0.060*** / 0.005
[1.54] / [2.75] / [0.15]
SIZE / 0.037*** / 0.019*** / 0.019***
[15.84] / [16.27] / [10.66]
IDV × SIZE / 0.007* / 0.001 / 0.008***
[1.89] / [0.53] / [2.88]
Controls / Yes / Yes / Yes
Industry-fixed effect / Yes / Yes / Yes
Year-fixed effect / Yes / Yes / Yes
Adj. R2 / 0.037 / 0.042 / 0.021
N / 181,604 / 181,604 / 181,604
Panel B: Analyst coverage
IDV / 0.080*** / 0.042*** / 0.053***
[6.54] / [7.22] / [5.87]
ANALYST / 0.060*** / 0.031*** / 0.034***
[14.71] / [14.87] / [10.26]
IDV × ANALYST / 0.010* / 0.001 / 0.009*
[1.68] / [0.44] / [1.81]
Controls / Yes / Yes / Yes
Industry-fixed effect / Yes / Yes / Yes
Year-fixed effect / Yes / Yes / Yes
Adj. R2 / 0.037 / 0.042 / 0.021
N / 181,604 / 181,604 / 181,604

1