Announcement

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

  • Insignificant variable results in Fixed Effects regression

    Hi!
    I´m working on a thesis on onset of civil war using panel data. While most of my control variables (GDP per capita, population, peace years) are stastically significant, my main independent variables are mostly not, except for food price shocks. Do you have any suggestions for improving the model or increasing the explanatory power of thesee variables (according to the theory they should be significant).

    code:
    xtreg conflict_onset disputed_election_index leder_dod z_ln_total_deaths z_ln_total_population z_ln_GDP_per_capita food_price_real_lag e_pt_coup_lag z_ln_GDP_per_capita_growth inv_peace v2x_polyarchy_lag i.year, fe


    Variables
    Variable | Obs Mean Std. dev. Min Max
    -------------+---------------------------------------------------------
    conflict_o~t | 8,164 .0249878 .1560972 0 1
    disputed_e~x | 7,839 -7.45e-10 .8340335 -1.591862 2.160036
    leder_dod | 9,212 .0167173 .1282172 0 1
    z_ln_total~s | 5,429 -6.20e-09 1 -1.571417 4.741596
    z_ln_total~n | 9,236 3.69e-10 1 -3.127967 3.133524
    -------------+---------------------------------------------------------
    z_ln_GDP_p~a | 9,236 -1.90e-09 1 -2.899817 2.0317
    food~eal_lag | 9,273 93.251 16.44187 67.21377 138.0701
    e_pt_coup_~g | 9,151 .0215277 .1524872 0 2
    z_ln_GDP_p~h | 7,426 6.89e-10 1 -7.232816 3.896336
    inv_peace | 9,496 .1821606 .3467801 0 1
    -------------+---------------------------------------------------------
    v2x_polyar~g | 9,227 .4266807 .2872424 .007 .923


    Results:

    Fixed-effects (within) regression Number of obs = 3,192
    Group variable: gwno Number of groups = 162

    R-squared: Obs per group:
    Within = 0.0335 min = 1
    Between = 0.0387 avg = 19.7
    Overall = 0.0062 max = 55

    F(67, 2963) = 1.53
    corr(u_i, Xb) = -0.8510 Prob > F = 0.0037

    conflict_onset Coefficient Std. err. t P>|t| [95% conf. interval]
    disputed_election_index -.0053183 .0119455 -0.45 0.656 -.0287407 .0181041
    leder_dod -.0347662 .0223004 -1.56 0.119 -.0784921 .0089597
    z_ln_total_deaths -.0001766 .0037739 -0.05 0.963 -.0075763 .0072231
    z_ln_total_population .0977481 .0335138 2.92 0.004 .0320354 .1634608
    z_ln_GDP_per_capita .0213272 .0085101 2.51 0.012 .0046408 .0380135
    food_price_real_lag .0128425 .0054108 2.37 0.018 .0022333 .0234518
    e_pt_coup_lag .014516 .0270241 0.54 0.591 -.0384719 .067504
    z_ln_GDP_per_capita_growth -.0022677 .0033216 -0.68 0.495 -.0087806 .0042451
    inv_peace -.104289 .0171007 -6.10 0.000 -.1378194 -.0707585
    v2x_polyarchy_lag -.0383662 .0386587 -0.99 0.321 -.1141669 .0374345
    year
    1963 -.0029834 .0444522 -0.07 0.946 -.0901438 .0841769
    1964 -.1164471 .0627312 -1.86 0.064 -.2394483 .0065541
    1965 -.1137012 .0763563 -1.49 0.137 -.2634181 .0360156
    1966 -.1255914 .0763897 -1.64 0.100 -.2753736 .0241909
    1967 -.1028639 .0645792 -1.59 0.111 -.2294885 .0237607
    1968 -.0752557 .0544727 -1.38 0.167 -.1820638 .0315523
    1969 .0100853 .0399413 0.25 0.801 -.0682303 .0884008
    1970 -.014779 .0384094 -0.38 0.700 -.0900908 .0605329
    1971 .0544263 .0375035 1.45 0.147 -.0191093 .127962
    1972 -.0389418 .0407334 -0.96 0.339 -.1188104 .0409269
    1973 -.00064 .0388509 -0.02 0.987 -.0768176 .0755376
    1974 -.2184557 .1084697 -2.01 0.044 -.4311392 -.0057721
    1975 -.5295297 .2196248 -2.41 0.016 -.9601624 -.098897
    1976 -.4021842 .1890427 -2.13 0.033 -.7728524 -.0315159
    1977 -.2036733 .0877422 -2.32 0.020 -.3757152 -.0316314
    1978 -.0582704 .0524241 -1.11 0.266 -.1610618 .0445209
    1979 .0117652 .0336904 0.35 0.727 -.0542937 .0778242
    1980 .0109281 .0336811 0.32 0.746 -.0551126 .0769688
    1981 -.0303249 .0367464 -0.83 0.409 -.1023759 .0417262
    1982 .0761319 .0325814 2.34 0.020 .0122474 .1400164
    1983 .1222691 .0589195 2.08 0.038 .0067418 .2377964
    1984 .2340701 .0686553 3.41 0.001 .0994532 .368687
    1985 .1054685 .0420211 2.51 0.012 .0230751 .187862
    1986 .1747711 .0798727 2.19 0.029 .0181594 .3313827
    1987 .3306279 .1449007 2.28 0.023 .0465116 .6147442
    1988 .3852219 .1666338 2.31 0.021 .0584921 .7119517
    1989 .3732147 .1450464 2.57 0.010 .0888129 .6576165
    1990 .332429 .1228861 2.71 0.007 .0914782 .5733798
    1991 .2923688 .1155705 2.53 0.011 .0657622 .5189754
    1992 .2993054 .118347 2.53 0.011 .0672547 .5313561
    1993 .2995918 .1132808 2.64 0.008 .0774747 .5217089
    1994 .347457 .1387539 2.50 0.012 .0753933 .6195207
    1995 .217139 .0946375 2.29 0.022 .0315772 .4027009
    1996 .1960504 .077406 2.53 0.011 .0442755 .3478254
    1997 .1598932 .063291 2.53 0.012 .0357945 .2839919
    1998 .2055158 .0833604 2.47 0.014 .0420656 .368966
    1999 .2479137 .1013347 2.45 0.014 .0492202 .4466072
    2000 .3779155 .1552967 2.43 0.015 .0734151 .6824159
    2001 .4006379 .1622591 2.47 0.014 .082486 .7187897
    2002 .3066162 .1361491 2.25 0.024 .0396599 .5735726
    2003 .3316954 .1450035 2.29 0.022 .0473775 .6160132
    2004 .3443332 .1318477 2.61 0.009 .085811 .6028555
    2005 .2639382 .1075798 2.45 0.014 .0529996 .4748769
    2006 .2294455 .1095035 2.10 0.036 .0147349 .444156
    2007 .2072453 .0885911 2.34 0.019 .0335391 .3809516
    2008 -.0535676 .0255129 -2.10 0.036 -.1035923 -.0035428
    2009 -.2233261 .0986595 -2.26 0.024 -.4167741 -.029878
    2010 .0055734 .0280286 0.20 0.842 -.0493842 .0605309
    2011 -.1086807 .0594356 -1.83 0.068 -.22522 .0078585
    2012 -.3020991 .1209255 -2.50 0.013 -.5392057 -.0649925
    2013 -.1944404 .0828145 -2.35 0.019 -.3568202 -.0320607
    2014 -.1576051 .0731557 -2.15 0.031 -.3010462 -.014164
    2015 -.1187245 .0566931 -2.09 0.036 -.2298863 -.0075626
    2016 .0214926 .0249807 0.86 0.390 -.0274886 .0704738
    2017 -.032742 .024018 -1.36 0.173 -.0798357 .0143517
    2018 -.0567261 .0312713 -1.81 0.070 -.1180419 .0045896
    2019 .027666 .0266159 1.04 0.299 -.0245216 .0798536
    2020 0 (omitted)
    _cons -1.241057 .5303056 -2.34 0.019 -2.280862 -.2012524

    Thank you!
    Sandra

  • #2
    Sandra:
    welcome to this forum.
    1) your R_sq within is very low. I would revise my model specification.
    2) with 162 panels, you should switch to -vce(cluster panelid)- standard errors.
    Kind regards,
    Carlo
    (Stata 19.0)

    Comment

    Working...
    X