Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Choosing Control Variables for a Multiple Regression

    Good morning,

    I am trying to estimate the effect of a firms ESG Score on its Stock Price. I have a large dataset of all companies within the S&P 500, however I am struggling to formulate my model and choosing my control variables. I have created a correlation matrix and run a panel data regression. I would like to understand the logic behind choosing the control variables and the formulation of my model.


    Below is the correlation Matrix
    sp ebit tcap current qr ltdp tate roe roa ta esgscore
    sp 1
    ebit -0.0037 1
    tcap 0.0314 -0.0057 1
    current 0.0427 0.102 0.6219 1
    qr -0.0176 0.12 0.53 0.8452 1
    ltdp -0.0707 -0.1082 -0.2524 -0.3022 -0.2953 1
    tate -0.0141 -0.0364 0.069 -0.057 -0.0565 0.1876 1
    roe -0.0121 0.0034 0.0848 -0.0132 -0.0152 0.1264 0.9258 1
    roa 0.1067 0.2606 0.495 0.4109 0.35 -0.1809 -0.0205 0.0246 1
    ta -0.0226 -0.0006 -0.1041 -0.1024 -0.0255 -0.0186 -0.0021 -0.0034 -0.0617 1
    esgscore -0.078 -0.015 -0.1114 -0.1983 -0.1048 0.1787 0.0421 0.0194 -0.0914 0.1764 1

    Below is the Panel Data Regression:
    lsp Coefficient Std. err. z P>z [95% conf. interval]
    ebit -0.1094329 0.0606259 -1.81 0.071 -0.22826 0.009392
    tcap -0.2657965 0.1900788 -1.4 0.162 -0.63834 0.106751
    current 0.0604384 0.0375471 1.61 0.107 -0.01315 0.134029
    qr -0.0597498 0.0484727 -1.23 0.218 -0.15475 0.035255
    ltdp 0.0545947 0.1191632 0.46 0.647 -0.17896 0.28815
    tate -0.0002566 0.0022631 -0.11 0.910 -0.00469 0.004179
    roe -0.0016039 0.0058455 -0.27 0.784 -0.01306 0.009853
    roa 0.4654144 0.2795494 1.66 0.096 -0.08249 1.013321
    ta 6.52E-14 5.78E-13 0.11 0.910 -1.07E-12 1.20E-12
    esgscore 0.0117663 0.0012669 9.29 0 0.009283 0.014249
    _cons 3.80707 0.1247927 30.51 0 3.562481 4.05166
    sigma_u 0.86507238
    sigma_e 0.3397378
    rho 0.86637472 (fraction of variance due to u_i)

    This shows what each Variable is:
    Variable Name Variable Label
    id Identifier
    idcode Panel Data Version of Company Identifier
    year Year
    Name Company Name
    sp Stock price
    ebit EBIT Margin - %
    tcap Total Current Assets Percentage of Total Assets
    current Current Ratio
    qr Quick Ratio
    ltdp Long Term Debt Percentage of Total Capital
    tate Total Assets to Total Equity
    roe Return on Average Common Equity - %, TTM
    roa Return on Average Total Assets - %, TTM
    ta Total Assets
    esgscore ESG Score
    esgcs ESG Combined Score
    social Social Pillar Score
    governance Governance Pillar Score
    environmental Environmental Pillar Score
    Sector Sector that the Firm Operates in
    SectorsCode Encoded Sector
    highesg When ESG Score > 70
    lsp Log of the Stock Price
    From what I understand, I want my control variables to have a low correlation with "esgscore", but have a correlation with the "lsp", which is the Log of the Stock Price.

    I would appreciate any feedback or recommendations on how to build my model!

    Kind regards,
    Derick
    Last edited by Derick Coleman; 14 Mar 2023, 05:59.
Working...
X