Effects of 10-K readability on institutional blockholder monitoring of risk management

ABSTRACT

Using the Plain Writing Act (PWA) of October 2010 as a source of exogenous variation in 10-K readability, we examine whether improvements in 10-K reports enhance institutional investors’ monitoring of firm risk management. The results indicate that the negative association between monitoring institutional ownership and systematic risk is stronger when firms provide more readable 10-K reports. The impact of readability improvement on the relation between institutional monitoring and risk is more pronounced after the PWA. This finding confirms our hypothesis that improved readability provides institutional investors with more complete firm information with which to conduct their monitoring efforts, thus inducing more effective risk management.

KEYWORDS:

• 10-K report readability
• institutional monitoring
• risk management
• plain writing act

JEL CLASSIFICATION:

• G30
• G39
• M40

Acknowledgement

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2020R1G1A1100261).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Appendix provides further details on index calculations and descriptions of control variables.

² We conduct two sets of additional robustness tests on our main findings in Table 2. First, we decompose the sample by firm size into two using the median value of Logasset and re-estimate the models in Table 2. We find that the moderating role of readability on the relationship between institutional blockholder monitoring and risk management is more pronounced in small-sized firms, thus implying that reliable sources of firm information are more valuable for firms in less transparent information environments. Next, we use a two-stage least squares (2SLS) estimation method to re-estimate the models in Table 2. We address the endogenous variation of 10-K report readability by employing the 2SLS method. The results of our instrumental variable approach are as follows. In the first stage, we use the average of readability by year and industry to mitigate concerns about possible estimation biases caused by important omitted variables. The averages of readability variables are uncorrelated with the error terms and highly correlated with the firm-level readability variables. The F-statistic in the first-stage 2SLS regression is 633.74, which is far above the cut-off criterion (F-statistics = 10). The first-stage t-statistic is 25.17, indicating that the instrument is sufficiently correlated with the key variables of interest. When we regress adjusted beta on the predicted value of readability variables in the second stage of the 2SLS regression, we see that the coefficients of interaction terms are still negative and significant, thus corroborating the results reported in Table 2. Though the results are unreported for brevity, they are available upon request.

³ The Bog index data are publicly available at https://kelley.iu.edu/bpm/activities/bogindex.html.