Good day. Thanks in advance for answering my question, as I'm very new to this.
I want to test for cross-sectional dependence in my panel data. My sample is N=17 and T=24. I have conducted a Breusch and Pagan LM test as I've read it is for N<T.
I am not entirely sure if I did it correctly. I run fe reg, then used xttest2.
My questions are:
1. Is it correct to use said LM test considering my sample size?
2. If yes, how do I interpret the correlation matrix of residuals? For reference, my panel data is 17 regions for the sample period 2000-2023. My y=ln_GDP per capita, X1=cargothroughput, x2=unemployment rate
3. If LM test is not applicable or maybe applicable but can be replaced by something better, what kind of test should I perform?
Correlation matrix of residuals:
| __e5 __e6 __e7 __e8 __e9 __e10
-------------+------------------------------------------------------------------
__e5 | .3185251
__e6 | .3198726 .3383375
__e7 | .1144068 .1178607 .057399
__e8 | .1911707 .2043367 .0734611 .1555141
__e9 | .1368931 .1576753 .0600229 .1093298 .110642
__e10 | .0561907 .0699871 .0399363 .0601682 .0563327 .0720236
__e11 | .1489101 .1505083 .0527082 .0892523 .0693261 .0549384
__e12 | .2476772 .261421 .0960683 .1624724 .1398158 .0765153
__e13 | .0280943 .0202184 .0194039 .024618 .0075671 .0457463
__e14 | .2193809 .2328894 .0847137 .1533144 .1168751 .0650194
__e15 | .1668182 .179338 .0704679 .1183783 .1020612 .0618981
__e16 | .111997 .1189068 .0432211 .071874 .065693 .0369161
__e17 | .3102578 .3234171 .1283312 .2107854 .1544726 .1050539
| __e11 __e12 __e13 __e14 __e15 __e16
-------------+------------------------------------------------------------------
__e11 | .1200495
__e12 | .1405499 .2347975
__e13 | .0538527 .0413157 .0751342
__e14 | .1129414 .195117 .0303609 .1737358
__e15 | .0948105 .1594248 .0316645 .1364574 .1148054
__e16 | .0622169 .10616 .0174369 .085907 .0703396 .0516206
__e17 | .1760865 .2664275 .0680071 .2375593 .1909677 .1172014
| __e17
-------------+-----------
__e17 | .353881
__e5 __e6 __e7 __e8 __e9 __e10 __e11 __e12 __e13 __e14
__e5 1.0000
__e6 0.9744 1.0000
__e7 0.8461 0.8457 1.0000
__e8 0.8589 0.8908 0.7775 1.0000
__e9 0.7292 0.8149 0.7532 0.8335 1.0000
__e10 0.3710 0.4483 0.6211 0.5685 0.6310 1.0000
__e11 0.7615 0.7468 0.6350 0.6532 0.6015 0.5908 1.0000
__e12 0.9057 0.9275 0.8275 0.8503 0.8675 0.5884 0.8372 1.0000
__e13 0.1816 0.1268 0.2955 0.2277 0.0830 0.6219 0.5670 0.3111 1.0000
__e14 0.9326 0.9606 0.8483 0.9327 0.8430 0.5812 0.7820 0.9661 0.2657 1.0000
__e15 0.8723 0.9099 0.8681 0.8859 0.9056 0.6807 0.8076 0.9710 0.3409 0.9662
__e16 0.8734 0.8997 0.7940 0.8022 0.8693 0.6054 0.7903 0.9643 0.2800 0.9071
__e17 0.9241 0.9347 0.9004 0.8985 0.7807 0.6580 0.8543 0.9243 0.4171 0.9581
__e15 __e16 __e17
__e15 1.0000
__e16 0.9137 1.0000
__e17 0.9474 0.8671 1.0000
Breusch-Pagan LM test of independence: chi2(78) = 506.671, Pr = 0.0000
Based on 11 complete observations over panel units
I want to test for cross-sectional dependence in my panel data. My sample is N=17 and T=24. I have conducted a Breusch and Pagan LM test as I've read it is for N<T.
I am not entirely sure if I did it correctly. I run fe reg, then used xttest2.
My questions are:
1. Is it correct to use said LM test considering my sample size?
2. If yes, how do I interpret the correlation matrix of residuals? For reference, my panel data is 17 regions for the sample period 2000-2023. My y=ln_GDP per capita, X1=cargothroughput, x2=unemployment rate
3. If LM test is not applicable or maybe applicable but can be replaced by something better, what kind of test should I perform?
Correlation matrix of residuals:
| __e5 __e6 __e7 __e8 __e9 __e10
-------------+------------------------------------------------------------------
__e5 | .3185251
__e6 | .3198726 .3383375
__e7 | .1144068 .1178607 .057399
__e8 | .1911707 .2043367 .0734611 .1555141
__e9 | .1368931 .1576753 .0600229 .1093298 .110642
__e10 | .0561907 .0699871 .0399363 .0601682 .0563327 .0720236
__e11 | .1489101 .1505083 .0527082 .0892523 .0693261 .0549384
__e12 | .2476772 .261421 .0960683 .1624724 .1398158 .0765153
__e13 | .0280943 .0202184 .0194039 .024618 .0075671 .0457463
__e14 | .2193809 .2328894 .0847137 .1533144 .1168751 .0650194
__e15 | .1668182 .179338 .0704679 .1183783 .1020612 .0618981
__e16 | .111997 .1189068 .0432211 .071874 .065693 .0369161
__e17 | .3102578 .3234171 .1283312 .2107854 .1544726 .1050539
| __e11 __e12 __e13 __e14 __e15 __e16
-------------+------------------------------------------------------------------
__e11 | .1200495
__e12 | .1405499 .2347975
__e13 | .0538527 .0413157 .0751342
__e14 | .1129414 .195117 .0303609 .1737358
__e15 | .0948105 .1594248 .0316645 .1364574 .1148054
__e16 | .0622169 .10616 .0174369 .085907 .0703396 .0516206
__e17 | .1760865 .2664275 .0680071 .2375593 .1909677 .1172014
| __e17
-------------+-----------
__e17 | .353881
__e5 __e6 __e7 __e8 __e9 __e10 __e11 __e12 __e13 __e14
__e5 1.0000
__e6 0.9744 1.0000
__e7 0.8461 0.8457 1.0000
__e8 0.8589 0.8908 0.7775 1.0000
__e9 0.7292 0.8149 0.7532 0.8335 1.0000
__e10 0.3710 0.4483 0.6211 0.5685 0.6310 1.0000
__e11 0.7615 0.7468 0.6350 0.6532 0.6015 0.5908 1.0000
__e12 0.9057 0.9275 0.8275 0.8503 0.8675 0.5884 0.8372 1.0000
__e13 0.1816 0.1268 0.2955 0.2277 0.0830 0.6219 0.5670 0.3111 1.0000
__e14 0.9326 0.9606 0.8483 0.9327 0.8430 0.5812 0.7820 0.9661 0.2657 1.0000
__e15 0.8723 0.9099 0.8681 0.8859 0.9056 0.6807 0.8076 0.9710 0.3409 0.9662
__e16 0.8734 0.8997 0.7940 0.8022 0.8693 0.6054 0.7903 0.9643 0.2800 0.9071
__e17 0.9241 0.9347 0.9004 0.8985 0.7807 0.6580 0.8543 0.9243 0.4171 0.9581
__e15 __e16 __e17
__e15 1.0000
__e16 0.9137 1.0000
__e17 0.9474 0.8671 1.0000
Breusch-Pagan LM test of independence: chi2(78) = 506.671, Pr = 0.0000
Based on 11 complete observations over panel units
