CEO Values Index - CVI Construction Approach

How to measure and quantify CEOs' ESG-oriented values and construct the CVI?

Given that the development of personal values is a rather complex process and cannot be comprehensively and convincingly reflected by a single factor, we employ the index construction approach and follow five steps (as shown in the right figure) to measure CEOs' ESG-oriented values with multi-dimensional approach and construct our CVI.

What is the index construction approach?

The index construction approach is a commonly used method to comprehensively consider multiple key indicators in order to construct an effective index for measuring overall performance or trends in a particular field.

This approach is widely utilized in diverse research domains, including social sciences, economics, and health science, and has also been introduced in some corporate studies as well.

Step 1. Select a bunch of categories and factors related to CEOs' ESG-oriented values development.

CVI Measurement Categories and Factors

In the first step, we first refer to existing research in psychology, social sciences, and other disciplines, and summarize the key factors influencing the development of personal values into five categories: lifestyle, cultural background, life experiences, familiar elements, and external environmental exposure.

Then, based on the identified five categories, we further select eleven factors related to the development of CEOs' ESG-oriented values to represent these categories.

Detailed assumptions of all selected CVI measurement factors:

Factor 1. Plant-based diet habit: CEOs who follow a plant-based diet habit are more likely to develop stronger ESG-oriented values.

Factor 2. Social media use habit: CEOs who are more active on social media and post more ESG-related information are more likely to have stronger ESG-oriented values.

Factor 3. Pre-adulthood relocation experience: CEOs who have experienced more significant relocations (such as across subnational divisions and international moves) before adulthood are more likely to develop stronger ESG-oriented values.

Factor 4. Health struggle experience: CEOs who have experienced health struggle experiences or have witnessed and accompanied close relatives through health struggle experiences are more likely to develop stronger ESG-oriented values.

Factor 5. Charitable engagement experience: CEOs who have more previous and current charitable engagement experiences are more likely to have stronger ESG-oriented values.

Factor 6. Birthplace cultural background: Based on Hofstede's Cultural Dimension Theory, CEOs from countries characterized by collectivism, masculinity, high uncertainty avoidance, low power distance, long-term orientation, and restraint are more likely to develop stronger ESG-oriented values.

Factor 7. Children: CEOs with children are more likely to develop stronger ESG-oriented values compared to those without children, but this enhancement effect diminishes as the number of children increases. Furthermore, when focusing on the gender of the children, CEOs with more daughters are more likely to develop stronger ESG-oriented values.

Factor 8. Siblings: CEOs with siblings are more likely to develop stronger ESG-oriented values compared to those without siblings, but this enhancement effect diminishes as the number of siblings increases. Furthermore, when focusing on the gender of the siblings, CEOs with more female siblings are more likely to develop stronger ESG-oriented values.

Factor 9. Spouse: CEOs with spouses are more likely to have stronger ESG-oriented values compared to those without spouses.

Factor 10. Residential environment: CEOs living in more degraded and safer areas are more likely to develop stronger ESG-oriented values.

Factor 11. Workplace environment: Based on Leadership in Energy and Environmental Design (LEED) scoring system, CEOs working in more sustainable and greener environments are more likely to develop stronger ESG-oriented values.

Step 2. Use suitable measurement variables to quantify the selected factors.

To quantify the nine selected ESG-oriented values measurement factors and cover the information mentioned in our previous hypotheses, we used a total of 19 variables as CVI measurement variables. The figure below provides the overall relational framework for all the variables, factors, and categories used to measure CVI in our study. The table below shows the definitions of all CVI measurement variables.

Step 3. Data collection for all CVI measurement variables.

We manually collected CVI-based information for 756 CEOs from 449 unique S&P 500 companies (excluding financial and utility firms) between 2013 and 2022. Detailed information about how we collected the CVI-based data can be found here.

Step 4. Address potential concerns in the data and prepare them for the index construction.

There are three challenges rooting in the CVI measurement variables we selected:

Challenge 1: Different CVI measurement variables have different value ranges.

↘ Contribution degree of different measurement variable to the final CVI would be different.

Challenge 2: Different CVI measurement variables have different impact on ESG-oriented values development.

↘ Contribution approach of different measurement variable to the final CVI would be different.

Challenge 3: Some CVI measurement variables are time invariant.

↘ Fluctuations of the final constructed CVI may be not obvious.

To address all three concerns at the same time, we apply the quartile processing approach to all CVI measurement variables. Briefly, we quartile each CVI measurement variable on an annual basis, assigning values ranging from 1 to 4 or from 4 to 1, contingent upon whether the variable is positively or negatively associated with ESG-oriented values development.

Moreover, when converting some CVI measurement variables into quartiles, we adjust the approach to accommodate the uniqueness of different variables.

First, for categorical variables, namely Twitter_Engagement, CEO_Plant, CEO_Sick, CEO_Spouse, and CEO_Relocation, except for CEO_Plant, CEO_Spouse, and Twitter_Engagement, the other two variables are integers ranging from 0 to 4 and thus do not require quartile conversion. For CEO_Plant and Twitter_Engagement, we replace the original values 1 and 2, with 2 and 4, while retaining the value of 0 for omnivorous CEOs or CEOs without personal Twitter account. For CEO_Spouse, we replace the original values 1 with 4, while retaining the value of 0 for unmarried CEOs.
Second, for family environment-based variables, since CEOs with children, daughters, siblings, and female siblings often exhibit stronger ESG-oriented values than those without these family relationships, we assign a value of 0 to CEOs without these family relationships. Subsequently, we apply the standard quartile conversion procedure specifically to the remaining non-zero data.
Third, for the birthplace cultural background variables, due to the high proportion of CEOs born in the United States, we performed a binary conversion to balance the distribution of the variable after processing.

Step 5. Design the appropriate index construction approach and get the final CVI.

Since most of the CVI-based information disclosure is entirely voluntary by the CEOs, each of our CVI measurement variables has some missing values. To minimize the impact of these missing values on the construction of the CVI, we employ a three-layer balancing process.

First, we balance the contribution of each variable to each factor by summing the non-missing CVI measurement variables under each factor and dividing by the number of non-missing CVI measurement variables. Then, using a similar method, we balance the contribution of each factor to each category, and finally, the contribution of each category to the CVI.

Below is the complete formula for CVI construction. The CVI scores for our research sample can be found here.

Page updated

Google Sites

Report abuse