Ask the Scholar

Document scope · 1 page
doc
Scholar
Ask about this object, its catalog metadata, its source description, or the page inventory. For page-specific OCR and visual context, open one of the page chats.

Scholar Source Context

Document identity
localId
225358443
label
CO2 [Carbon Dioxide] Emissions - Technical Paper
core
doc
dtoType
document
pageCount
1
Source metadata
Source extras
naId
225358443
levelOfDescription
fileUnit
otherTitles
42-t-4739519-20171095F-012-003-2019
recordType
description
ocrSource
nara-archive
Single page context
seq
1
pageIndex
0
type
document
mediaId
7a75aeff0e73f5dc
ocrText
FOIA Number: 2017-1095-F FOIA MARKER This is not a textual record. This is used as an administrative marker by the William J. Clinton Presidential Library Staff. Collection/Record Group: Clinton Presidential Records Subgroup/Office of Origin: Council of Economic Advisers Series/Staff Member: Judson Jaffe Subseries: OA/ID Number: 20746 FolderID: Folder Title: CO2 [Carbon Dioxide] Emmissions - Technical Paper Stack: Row: Section: Shelf: Position: S 20 6 3 3 Tech1.log 151099 / > Detrend the series > * * / sort year gen dntek = nonhitek - 1.02808*nonhitek[_n - 1] (1 missing value generated) gen dco2 = co2 [_n] - 1.007727*co2 n - 1] (1 missing value generated) gen tntek = 1.02808*nonhitek[_n-1] (1 missing value generated) gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) gen cheat = htdgyear [ n] - htdgyear [_n-1] (1 missing value generated) gen ccool = cldgyear [_n] - cldgyear [_n-1] (1 missing value generated) gen cfoss = ffppcomr [n] - ffppcomr [_n-1] (1 missing value generated) gen cres = eleprrer [ n] - eleprrer [_n-1] (1 missing value generated) gen ccom = eleprcor [ n] - eleprcor [_n-1] (1 missing value generated) gen ccoal = ffppcoar [n] - ffppcoar[_n-1] (1 missing value generated) gen cgas = ffppngr [_n] - ffppngr [_n-1] (1 missing value generated) gen coil = ffppoilr [n] - ffppoilr[_n-1] (1 missing value generated) gen cnumref = numrefin [n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap [n] - refincap [_n-1] (1 missing value generated) gen cnetsum = netsumca [ n] - netsumca{_n-1} (1 missing value generated) gen cextra = extralos [n] - extralos[_n-1] (1 missing value generated) 1 Tech1.log 151099 gen cpdctu = cpvtyund[_n] - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur[_n] - cpvtysur[_n-1] (2 missing values generated) gen cnumele = numeleut [_n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal[_n] - propcoal[_n-1] (1 missing value generated) gen cresl = cres [_n-1] (2 missing values generated) gen ccoml = ccom [_n-1] (2 missing values generated) gen cfoss1 = cfoss[_n-1] (2 missing values generated) gen ccoall = ccoal[_n-1] (2 missing values generated) gen cgasl = cgas [_n-1] (2 missing values generated) gen coill = coil [_n-1] (2 missing values generated) /******* > *** Fossil Index > / / * Current Foss Only * / reg dco2 dntek cheat ccool cfoss if year < 1998 Source I SS df MS Number of obs = 37 F( 4, 32) = 13.65 Model 27064.2851 4 6766.07127 Prob > F = 0.0000 Residual 15857.0642 32 495.533256 R-squared = 0.6306 + Adj R-squared = 0.5844 Total 42921.3493 36 1192.2597 Root MSE = 22.261 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2737161 .0402009 6.809 0.000 .1918295 .3556026 cheat .0270028 .0178131 1.516 0.139 -.0092813 .063287 ccool .0193422 .0370667 0.522 0.605 -.0561601 .0948445 cfoss -10.13689 11.87851 -0.853 0.400 -34.33263 14.05884 cons 9.628127 3.663471 2.628 0.013 2.165881 17.09037 2 Tech1.log 151099 reg dco2 dntek cheat ccool cfoss (dntek cheat ccool cnumre crefcap) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 + F( 4, 32) = 11.44 Model 24009.4166 4 6002.35414 Prob > F = 0.0000 Residual 18911.9327 32 590.997897 R-squared = 0.5594 Adj R-squared = 0.5043 Total 42921.3493 36 1192.2597 Root MSE = 24.31 dco2 Coef. Std. Err. t P>|t [95% Conf. Interval] cfoss 19.3563 25.19825 0.768 0.448 -31.97085 70.68345 dntek .2950022 .0465891 6.332 0.000 .2001033 .3899012 cheat .0278568 .0194635 1.431 0.162 -.0117891 .0675026 ccool .0300333 .0412304 0.728 0.472 -.0539503 .1140169 cons 9.317951 4.007268 2.325 0.027 1.155413 17.48049 predict pdco2 HT = 142.71 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1475.818 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool cfoss1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 15.24 Model 28388.0878 4 7097.02194 Prob > F = 0.0000 Residual 14435.0364 31 465.646336 R-squared Il 0.6629 Adj R-squared = 0.6194 Total 42823.1242 35 1223.51783 Root MSE = 21.579 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2498562 .0412819 6.052 0.000 .1656612 .3340513 cheat .0251607 .0175264 1.436 0.161 -.0105847 .0609061 ccool .0179313 .0359694 0.499 0.622 -.0554288 .0912913 cfoss1 -23.79871 12.16326 -1.957 0.059 -48.60584 1.008411 cons 9.850245 3.600215 2.736 0.010 2.507557 17.19293 3 Tech1.log 151099 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1484.811 / * Current and lagged foss * / drop pdco2 pco2 reg dco2 dntek cheat ccool cfoss1 cfoss if year < 1998 Source I SS df MS Number of obs = 36 F( 5, 30) = 11.83 Model 28412.9204 5 5682.58407 Prob > F = 0.0000 Residual 14410.2038 30 480.340127 R-squared = 0.6635 Adj R-squared = 0.6074 Total 42823.1242 35 1223.51783 Root MSE = 21.917 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .249136 .0420477 5.925 0.000 .1632632 .3350088 cheat .0251523 .0178009 1.413 0.168 -.0112019 .0615064 ccool .0171 .036715 0.466 0.645 -.0578821 .092082 cfossl -22.78568 13.13255 -1.735 0.093 -49.60593 4.034561 cfoss -2.827392 12.4351 -0.227 0.822 -28.22326 22.56848 cons 9.869592 3.657568 2.698 0.011 2.399842 17.33934 reg dco2 dntek cheat ccool cfoss1 cfoss (dntek cheat ccool cfoss1 cnumre cref > cap) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 9.69 Model 24795.5748 5 4959.11496 Prob > F = 0.0000 Residual 18027.5494 30 600.918312 R-squared = 0.5790 Adj R-squared = 0.5089 Total 42823.1242 35 1223.51783 Root MSE = 24.514 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cfoss 31.2974 28.41318 1.102 0.279 -26.73006 89.32486 dntek .2578286 .0474516 5.434 0.000 .1609195 .3547378 cheat .0252546 .0199103 1.268 0.214 .0154076 .0659168 ccool .0271332 .0417066 0.651 0.520 -.058043 .1123094 cfoss1 -35.01231 17.16275 -2.040 0.050 -70.06333 .0386982 cons 9.636083 4.094475 2.353 0.025 1.274049 17.99812 4 Tech1.log 151099 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1475.143 / > *** Coal Prices > * / / * Current Coal Only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoal if year < 1998 Source SS df MS Number of obs = 37 + F( 4, 32) = 13:61 Model 27030.0231 4 6757.50577 Prob > F = 0.0000 Residual 15891.3262 32 496.603943 R-squared = 0.6298 Adj R-squared = 0.5835 Total 42921.3493 36 1192.2597 Root MSE Il 22.285 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2716014 .0410022 6.624 0.000 .1880827 .3551201 cheat .0276114 .0178333 1.548 0.131 -.0087138 .0639365 ccool .0161517 .0378151 0.427 0.672 -.0608751 .0931786 ccoal -22.30923 27.50882 -0.811 0.423 -78.34287 33.7244 cons 9.476006 3.665723 2.585 0.015 2.009172 16.94284 reg dco2 dntek cheat ccool ccoal (dntek cheat ccool cpdctu cpdcts) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 5.16 Model -646.664551 4 -161.666138 Prob > F = 0.0025 Residual 43568.0138 32 1361.50043 R-squared = Adj R-squared = Total 42921.3493 36 1192.2597 Root MSE = 36.899 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccoal -227.6731 226.4482 -1.005 0.322 -688.9329 233.5868 dntek .1847882 .1157666 1.596 0.120 -.0510207 .4205971 5 Tech1.log 151099 cheat .0305115 .0296937 1.028 0.312 -.0299726 .0909957 ccool -.0470433 .0926269 -0.508 0.615 -.2357182 .1416316 cons 9.057044 6.086494 1.488 0.147 -3.340738 21.45483 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1477.428 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoall if year < 1998 Source SS df MS Number of obs = 36 + F( 4, 31) = 12.76 Model 26645.4082 4 6661.35205 Prob > F = 0.0000 Residual 16177.716 31 521.861806 R-squared = 0.6222 Adj R-squared = 0.5735 Total 42823.1242 35 1223.51783 Root MSE = 22.844 dco2 Coef. Std. Err. t P>Itl [95% Conf. Interval] + dntek .2789835 .0409795 6.808 0.000 .1954053 .3625618 cheat .0275711 .0185617 1.485 0.148 -.0102858 .065428 ccool .0225673 .0380051 0.594 0.557 -.0549445 .1000792 ccoall -7.471242 26.99979 -0.277 0.784 -62.53767 47.59519 cons 9.528624 3.807614 2.503 0.018 1.762943 17.2943 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1482.501 / * Current and lagged foss * / drop pdco2 pco2 6 Tech1.log 151099 reg dco2 dntek cheat ccool ccoall ccoal if year < 1998 Source SS df MS Number of obs = 36 + F( 5, 30) = 10.17 Model 26933.3741 5 5386.67483 Prob > F = 0.0000 Residual 15889.75 30 529.658334 R-squared = 0.6289 Adj R-squared = 0.5671 Total 42823.1242 35 1223.51783 Root MSE = 23.014 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2715152 .0425088 6.387 0.000 .1847007 3583297 cheat .0274464 .0187006 1.468 0.153 .0107453 .0656382 ccool .0159155 .0393364 0.405 0.689 -.0644201 .0962511 ccoall .0277121 29.03983 0.001 0.999 -59.27954 59.33496 ccoal -22.38282 30.35583 -0.737 0.467 -84.37768 39.61205 cons 9.509775 3.836036 2.479 0.019 1.675544 17.34401 reg dco2 dntek cheat ccool ccoall ccoal (dntek cheat ccool ccoall cpdctu cpdc > ts) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 + F( 5, 30) = 1.87 Model -46578.2366 5 -9315.64733 Prob > F = 0.1293 Residual 89401.3608 30 2980.04536 R-squared = Adj R-squared = Total 42823.1242 35 1223.51783 Root MSE II 54.59 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccoal 335.2374 522.3383 0.642 0.526 -731.5198 1401.995 dntek .3908402 .1999126 1.955 0.060 -.0174357 . 7991161 cheat .0294381 .0444512 0.662 0.513 -.0613435 .1202196 ccool .1221956 .1798474 0.679 0.502 -.2451018 . 4894931 ccoall -119.7864 186.5149 -0.642 0.526 -500.7006 261.1278 cons 9.810921 9.109477 1.077 0.290 -8.793114 28.41496 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1489.691 / ******* > *** Oil Prices 7 Tech1.log 151099 > * / / * Current Coal Only * / drop pdco2 pco2 reg dco2 dntek cheat ccool coil if year < 1998 Source SS df MS Number of obs = 37 + F( 4, 32) = 13.36 Model 26847.0568 4 6711.76421 Prob > F = 0.0000 Residual 16074.2924 32 502.321639 R-squared = 0.6255 Adj R-squared = 0.5787 Total 42921.3493 36 1192.2597 Root MSE = 22.413 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2782483 .0398855 6.976 0.000 .1970042 .3594923 cheat .0260073 .0180927 1.437 0.160 - -.0108462 .0628609 ccool .0204306 .0373812 0.547 0.588 -.0557123 .0965736 coil -2.499771 4.674564 -0.535 0.597 -12.02155 7.022004 cons 9.548796 3.686687 2.590 0.014 2.039262 17.05833 reg dco2 dntek cheat ccool coil (dntek cheat ccool cnumre crefcap) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 11.73 Model 24454.7477 4 6113.68693 Prob > F = 0.0000 Residual 18466.6015 32 577.081298 R-squared = 0.5698 Adj R-squared = 0.5160 Total 42921.3493 36 1192.2597 Root MSE = 24.023 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] coil 7.701606 9.589022 0.803 0.428 -11.83059 27.2338 dntek .2896092 .0437095 6.626 0.000 .2005759 .3786425 cheat .0312676 .0198453 1.576 0.125 -.0091559 .0716911 ccool .0309845 .0409495 0.757 0.455 -.0524269 .1143959 cons 9.437479 3.952523 2.338 0.023 1.386453 17.48851 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1473.783 8 Tech1.log 151099 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool coill if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 13.70 Model 27354.4663 4 6838.61657 Prob > F = 0.0000 Residual 15468.6579 31 498.988964 R-squared = 0.6388 Adj R-squared = 0.5922 Total 42823.1242 35 1223.51783 Root MSE = 22.338 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval ] dntek .2550741 .0447506 5.700 0.000 .1638046 .3463436 cheat .0256393 .0181565 1.412 0.168 -.0113911 .0626696 ccool .0192192 .0372641 0.516 0.610 -.0567815 .09522 coill -6.366348 5.196248 -1.225 0.230 -16.96416 4.231469 cons 9.693438 3.725393 2.602 0.014 2.095449 17.29143 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1485.126 / * Current and lagged foss * / drop pdco2 pco2 reg dco2 dntek cheat ccool coill coil if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 10.67 Model 27411.2483 5 5482.24966 Prob > F = 0.0000 Residual 15411.8759 30 513.729195 R-squared = 0.6401 Adj R-squared = 0.5801 Total 42823.1242 35 1223.51783 Root MSE = 22.666 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2544965 .04544 5.601 0.000 .1616956 .3472974 cheat .0248547 .0185732 1.338 0.191 -.0130768 .0627863 ccool .017695 .0380875 0.465 0.646 -.06009 .09548 coill -6.068622 5.34795 -1.135 0.265 -16.99059 4.853349 9 Tech1.log 151099 coil -1.595205 4.798197 -0.332 0.742 -11.39443 8.20402 cons 9.710924 3.780383 2.569 0.015 1.990352 17.4315 reg dco2 dntek cheat ccool coill coil (dntek cheat ccool coill cnumre crefcap > ) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 + F( 5, 30) = 9.00 Model 24176.381 5 4835.27619 Prob > F = 0.0000 Residual 18646.7432 30 621.558107 R-squared = 0.5646 + Adj R-squared = 0.4920 Total 42823.1242 35 1223.51783 Root MSE = 24.931 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] coil 10.44515 10.37356 1.007 0.322 -10.74049 31.63079 dntek .258856 .0500863 5.168 0.000 .1565661 .3611459 cheat .0307761 .0208964 1.473 0.151 -.0119 .0734523 ccool .0291996 .0427546 0.683 0.500 -.0581169 .1165161 coill -8.315811 6.114072 -1.360 0.184 -20.80241 4.170788 cons 9.578943 4.159392 2.303 0.028 1.084331 18.07355 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1474.094 /******* > *** Residential Electricity > ******* / / * Current only A. / drop pdco2 pco2 reg dco2 dntek cheat ccool cres if year < 1998 Source I SS df MS Number of obs = 37 F( 4, 32) = 22.48 Model 31654.7076 4 7913.67691 Prob > F = 0.0000 Residual 11266.6416 32 352.082551 R-squared II 0.7375 Adj R-squared = 0.7047 Total 42921.3493 36 1192.2597 Root MSE Il 18.764 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] 10 Tech1.log 151099 dntek .2120832 .0378695 5.600 0.000 .1349455 .2892208 cheat .0254613 .0150202 1.695 0.100 -.0051338 .0560565 ccool .0124035 .0311614 0.398 0.693 -.0510703 .0758773 cres -39.59006 10.5572 -3.750 0.001 -61.09438 -18.08574 cons 5.571467 3.261015 1.709 0.097 -1.071004 12.21394 reg dco2 dntek cheat ccool cres (dntek cheat ccool cnetsum cnumele cprpcoal) > if year < 1998 (2SLS) Source I SS df MS Number of obs = 37 F( 4, 32) = 19.47 Model 30490.9378 4 7622.73445 Prob > F II 0.0000 Residual 12430.4115 32 388.450359 R-squared II 0.7104 Adj R-squared = 0.6742 Total 42921.3493 36 1192.2597 Root MSE = 19.709 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -58.78383 19.43911 -3.024 0.005 -98.37999 -19.18766 dntek .1786558 .0485325 3.681 0.001 .0797984 .2775132 cheat .0245717 .0157942 1.556 0.130 -.0076001 .0567435 ccool .0072581 .03301 0.220 0.827 -.059981 .0744971 cons 3.656433 3.777601 0.968 0.340 -4.038288 11.35115 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1484.023 / * Lagged only * / drop pdco2 pco2 reg dco2 dntek cheat ccool cresl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 22.59 Model 31883.8434 4 7970.96085 Prob > F = 0.0000 Residual 10939.2808 31 352.880025 R-squared = 0.7445 Adj R-squared II 0.7116 Total 42823.1242 35 1223.51783 Root MSE = 18.785 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2618714 .0335244 7.811 0.000 .1934979 3302448 11 Tech1.log 151099 cheat .0329325 .0153008 2.152 0.039 .0017263 .0641387 ccool .0250683 .0312384 0.802 0.428 -.0386428 .0887795 cresl -36.31314 9.389152 -3.868 0.001 -55.46244 -17.16384 cons 6.021033 3.26017 1.847 0.074 -.6281266 12.67019 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1484.231 / * current and lagged prices * / drop pdco2 pco2 reg dco2 dntek cheat ccool cresl cres if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 21.74 Model 33560.9051 5 6712.18101 Prob > F = 0.0000 Residual 9262.21912 30 308.740637 R-squared Il 0.7837 Adj R-squared = 0.7477 Total 42823.1242 35 1223.51783 Root MSE = 17.571 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2220629 .0357078 6.219 0.000 .1491379 .2949879 cheat .029858 .0143726 2.077 0.046 .0005052 .0592107 ccool .0172297 .0294124 0.586 0.562 -.0428385 .0772978 cresl -25.35373 9.96197 -2.545 0.016 -45.69879 -5.008675 cres -26.1429 11.21698 -2.331 0.027 -49.05104 -3.23476 cons 4.502406 3.118301 1.444 0.159 -1.866014 10.87083 reg dco2 dntek cheat ccool cres1 cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 20.35 Model 33204.5064 5 6640.90127 Prob > F II 0.0000 Residual 9618.61781 30 320.620594 R-squared = 0.7754 Adj R-squared Il 0.7380 Total 42823.1242 35 1223.51783 Root MSE = 17.906 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -38.19457 25.0231 -1.526 0.137 -89.29855 12.90941 12 Tech1.log 151099 dntek .2037115 .0497293 4.096 0.000 .1021506 .3052723 cheat .0284406 .0148786 1.912 0.066 -.0019455 .0588268 ccool .0136161 .0307071 0.443 0.661 - -.0490961 .0763283 cresl -20.30153 13.78901 -1.472 0.151 -48.46244 7.859374 cons 3.802331 3.430736 1.108 0.277 -3.204166 10.80883 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1484.589 / > *** Commercial Electricity > / / * Current only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccom if year < 1998 Source SS df MS Number of obs = 37 + F( 4, 32) = 24.44 Model 32337.9362 4 8084.48405 Prob > F = 0.0000 Residual 10583.4131 32 330.731659 R-squared = 0.7534 Adj R-squared = 0.7226 Total 42921.3493 36 1192.2597 Root MSE = 18.186 dco2 Coef. Std. Err. t P>Itl [95% Conf. Interval] + dntek .2143291 .0359269 5.966 0.000 .1411484 .2875098 cheat .0296766 .0145613 2.038 0.050 .0000161 .0593371 ccool .0113524 .0302095 0.376 0.710 -.0501823 .0728871 ccom -42.53594 10.30541 -4.128 0.000 -63.52737 -21.54451 cons 5.525065 3.14398 1.757 0.088 -.879013 11.92914 reg dco2 dntek cheat ccool ccom (dntek cheat ccool cnetsum cnumele cprpcoal) > if year < 1998 (2SLS) Source SS df MS Number of obs = 37 + F( 4, 32) = 21.64 Model 31884.1979 4 7971.04947 Prob > F = 0.0000 Residual 11037.1514 32 344.910981 R-squared = 0.7429 Adj R-squared = 0.7107 Total 42921.3493 36 1192.2597 Root MSE = 18.572 13 Tech1.log 151099 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccom -54.60657 18.05595 -3.024 0.005 -91.38535 -17.82779 dntek .1954005 .0433063 4.512 0.000 .1071884 .2836126 cheat .0303521 .0148929 2.038 0.050 .0000163 .0606878 ccool .0080423 .0311115 0.259 0.798 -.0553297 .0714144 cons 4.390971 3.494084 1.257 0.218 -2.726245 11.50819 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1486.484 / * Lagged only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccom1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 18.32 Model 30094.7825 4 7523.69563 Prob > F = 0.0000 Residual 12728.3417 31 410.591667 R-squared = 0.7028 Adj R-squared II 0.6644 Total 42823.1242 35 1223.51783 Root MSE = 20.263 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2595669 .0365143 7.109 0.000 .1850955 .3340384 cheat .0333822 .0165628 2.015 0.053 -.0003979 .0671624 ccool .0235372 .0336915 0.699 0.490 -.0451772 .0922515 ccoml -30.70544 10.53293 -2.915 0.007 -52.18749 -9.223391 cons 6.652871 3.519086 1.891 0.068 I -.5243513 13.83009 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1480.233 14 Tech1.log 151099 / * current and lagged prices * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoml ccom if year < 1998 Source SS df MS Number of obs = 36 F( ( 5, 30) = 19.95 Model 32921.1425 5 6584.2285 Prob > F II 0.0000 Residual 9901.98166 30 330.066055 R-squared Il 0.7688 Adj R-squared = 0.7302 Total 42823.1242 35 1223.51783 Root MSE = 18.168 . dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2162079 .0359355 6.017 0.000 .1428178 .2895979 cheat .0321365 .0148562 2.163 0.039 .001796 .062477 ccool .0136472 .0303961 0.449 0.657 -.0484298 .0757243 ccoml -15.46589 10.78453 -1.434 0.162 -37.49085 6.559061 ccom -34.40684 11.75795 -2.926 0.006 -58.41978 -10.39391 cons 4.874915 3.213156 1.517 0.140 -1.687225 11.43705 reg dco2 dntek cheat ccool ccoml ccom (dntek cheat ccool ccom1 cnetsum cnumel > e cprpcoal) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 + F( 5, 30) = 18.50 Model 32785.008 5 6557.0016 Prob > F II 0.0000 Residual 10038.1162 30 334.603873 R-squared II 0.7656 + Adj R-squared = 0.7265 Total 42823.1242 35 1223.51783 Root MSE = 18.292 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccom -41.95803 26.2134 -1.601 0.120 -95.49294 11.57688 dntek .2066919 .0466667 4.429 0.000 .1113859 .301998 cheat .0318631 .014982 2.127 0.042 .0012659 .0624604 ccool .0114767 .031334 0.366 0.717 -.0525158 .0754692 ccom1 -12.12131 15.00713 -0.808 0.426 -42.76996 18.52735 cons 4.484711 3.453538 1.299 0.204 -2.568353 11.53778 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1484 1484.769 15 Tech1.log 151099 log close 16 Hiteck V-A Nou- Hitek Growth (Billious) (Billion1992) 1953:3 1957:3 1960:2 1969:4 7.269 3385.3 1973:4 9.556 3937.5 K 3.850 (69-75) 1980:1 22.15 4656.85 2.378 (73-81) 1981:3 28.73 4724.3 2.808 (81-90\ 1990:3 80.71 6061.39 CO2 1953 684.9 2.180 1953-1990 =>1.843%/year 1857 746.6 1.811 1960 787.9 3.940 1969 1115.6 3.454 1973 1277.9 -0.216 1981 1256 0.773 (990 1346.1 Possible CO2 Growth ,ats: 1981-1990 0.773 % 1953-1990 1.843% 75% GDP (81-90) 2.1790 75% NH-GDP 2.119 Dates of Peaks Jul 1953 Aug 1957 Ap, 1960 Dec 1969 Nov 1973 Jou 1980 Jul 1981 Jul 1990 Choined $ 1992 real GDP (billious) Quouter Value Growth Rate 1953:3 1887.4 2.631 1957:3 2094.0 2.904 1960:2 2265.5 N 4.342 1969:4 3392.6 3.857 1973:4 3947.1 1980:1 4679.0 2.426 1981:3 ) 4753.0 2.820 1990:3 6142.1 Tech3.log 061099 / > *** Detrend the series > * / sort year gen dntek = nonhitek [_n] - 1.02808 *nonhitek - 1] (1 missing value generated) gen tntek = 1.02808 nonhitek [_n-1] (1 missing value generated) gen cheat = htdgyear [_n] - htdgyear[_n-1] (1 missing value generated) gen ccool = cldgyear [_n] - cldgyear[_n-1] (1 missing value generated) gen cfoss = ffppcomr [_n] - ffppcomr [_n-1] (1 missing value generated) gen celec = eleprrer [_n] - eleprrer [_n-1] (12 missing values generated) gen cnumref = numrefin [n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [n] - netsumca[_n-1] (1 missing value generated) gen cextra = extralos [_n] - extralos [_n-1] (1 missing value generated) gen cpdctu = cpvtyund [n] - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur [n] - cpvtysur [_n-1] (2 missing values generated) gen cnumele = numeleut [n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal[_n] - propcoal [_n-1] (1 missing value generated) gen celecl = celec[_n-1] (13 missing values generated) gen cfossl = cfoss[_n-1] 1 Tech3.log 061099 (2 missing values generated) / * 1 * / gen tco2 = co2 [_n-1] (1 missing value generated) gen dco2 = co2 [_n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 21.19 Model 30487.8661 4 7621.96653 Prob > F If 0.0000 Residual 11148.4168 31 359.62635 R-squared = 0.7322 Adj R-squared = 0.6977 Total 41636.2829 35 1189.60808 Root MSE = 18.964 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2565274 .0338437 7.580 0.000 .1875027 .325552 cheat .0323265 .0154464 2.093 0.045 .0008234 .0638296 ccool .0240597 .0315356 0.763 0.451 - -.0402575 .088377 celecl -35.23996 9.478494 -3.718 0.001 -54.57148 -15.90844 cons 15.42274 3.291186 4.686 0.000 8.710317 22.13515 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1482.271 2.670949 1479.6 / * 1.0025 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.0025 * co2 [n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 2 Tech3.log 061099 Source SS df MS Number of obs = 36 F( 4, 31) = 21.66 Model 30935.9298 4 7733.98246 Prob > F = 0.0000 Residual 11070.3422 31 357.107812 R-squared = 0.7365 Adj R-squared = 0.7025 Total 42006.272 35 1200.1792 Root MSE = 18.897 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2582568 .033725 7.658 0.000 .1894743 .3270393 cheat .0325223 .0153922 2.113 0.043 .0011297 .0639148 ccool .0243854 .031425 0.776 0.444 -.0397063 .088477 celecl -35.58695 9.445245 -3.768 0.001 -54.85065 -16.32324 cons 12.38091 3.279642 3.775 0.001 5.692039 19.06979 - predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1482.905 - .3943477 1483.299 / * 1.005 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.005*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 22.11 Model 31387.3978 4 7846.84945 Prob > F = 0.0000 Residual 11002.282 31 354.912323 R-squared = 0.7404 Adj R-squared = 0.7070 Total 42389.6798 35 1211.13371 Root MSE = 18.839 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2599864 .0336211 7.733 0.000 .1914157 .3285572 cheat .0327182 .0153448 2.132 0.041 .0014222 .0640141 ccool .0247109 .0313282 0.789 0.436 -.0391834 .0886052 3 Tech3.log 061099 celec1 I -35.9339 9.416166 -3.816 0.001 -55.1383 -16.72951 cons 9.339088 3.269544 2.856 0.008 2.670808 16.00737 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1483.538 -3.459741 1486.998 / * 1.0075 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.0075*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 22.55 Model 31842.2195 4 7960.55487 Prob > F = 0.0000 Residual 10944.2143 31 353.039172 R-squared = 0.7442 + Adj R-squared = 0.7112 Total 42786.4338 35 1222.46954 Root MSE = 18.789 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .261716 .0335323 7.805 0.000 .1933264 .3301055 cheat .032914 .0153043 2.151 0.039 .0017008 .0641273 ccool .0250365 .0312454 0.801 0.429 -.038689 .088762 celec1 -36.28086 9.391285 -3.863 0.001 -55.43451 -17.12721 cons 6.297263 3.260905 1.931 0.063 -.3533966 12.94792 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 4 Tech3.log 061099 year co2 pco2 pdco2 tco2 50. 1998 1484 1484.172 -6.525086 1490.697 / * 1.01 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.01*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 22.97 Model 32300.4537 4 8075.11342 Prob > F = 0.0000 Residual 10896.1744 31 351.489496 R-squared = 0.7478 + Adj R-squared = 0.7152 Total 43196.6281 35 1234.18937 Root MSE = 18.748 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2634457 .0334586 7.874 0.000 .1952064 .331685 cheat .0331099 .0152706 2.168 0.038 .0019652 .0642546 ccool .025362 .0311768 0.813 0.422 -.0382235 .0889475 celecl -36.62782 9.370651 -3.909 0.000 -55.73939 -17.51625 cons 3.255445 3.25374 1.001 0.325 -3.380602 9.891492 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1484.806 -9.590449 1494.396 / * 1.015 - * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.015*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) 5 Tech3.log 061099 reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 23.78 Model 33226.9219 4 8306.73046 Prob > F = 0.0000 Residual 10830.1123 31 349.358461 R-squared = 0.7542 Adj R-squared = 0.7225 Total 44057.0341 35 1258.7724 Root MSE = 18.691 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2669046 .033357 8.001 0.000 . .1988725 .3349367 cheat .0335016 .0152243 2.201 0.035 .0024515 .0645517 ccool .0260131 .0310821 0.837 0.409 - -.0373793 .0894056 celecl -37.32168 9.342201 -3.995 0.000 -56.37523 -18.26814 cons -2.828188 3.243862 -0.872 0.390 -9.444088 3.787711 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1486.073 -15.72112 1501.794 / * 1.02 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.02*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.51 Model 34166.9319 4 8541.73298 Prob > F = 0.0000 Residual 10804.0909 31 348.51906 R-squared = 0.7598 Adj R-squared = 0.7288 Total 44971.0228 35 1284.88637 Root MSE = 18.669 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval ] dntek 2703637 .0333169 8.115 0.000 .2024134 .3383141 cheat .0338932 .015206 2.229 0.033 .0028804 .064906 6 Tech3.log 061099 ccool .0266642 .0310448 0.859 0.397 -.036652 .0899804 celecl -38.0156 9.330971 -4.074 0.000 -57.04624 -18.98496 cons -8.911834 3.239962 -2.751 0.010 -15.51978 -2.303887 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1487.34 -21.8518 1509.192 / * 1.025 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.025*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 25.16 Model 35120.4497 4 8780.11243 Prob > F = 0.0000 Residual 10818.1052 31 348.971134 R-squared = 0.7645 Adj R-squared = 0.7341 Total 45938.5549 35 1312.53014 Root MSE = 18.681 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2738227 .0333385 8.213 0.000 .2058283 .3418171 cheat .0342849 .0152158 2.253 0.031 .0032521 .0653178 ccool .0273154 .0310649 0.879 0.386 -.0360419 .0906727 celecl -38.70957 9.337021 -4.146 0.000 -57.75255 -19.66659 cons -14.99549 3.242063 -4.625 0.000 -21.60772 -8.38326 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 7 Tech3.log 061099 year co2 pco2 pdco2 tco2 50. 1998 1484 1488.607 -27.9825 1516.59 / * 1.03 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.03*co2 [_n-1] (1 missing value generated) gen dco2 = co2 [_n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 25.72 Model 36087.49 4 9021.8725 Prob > F = 0.0000 Residual 10872.1551 31 350.71468 R-squared = 0.7685 Adj R-squared = 0.7386 Total 46959.6451 35 1341.70415 Root MSE = 18.727 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .277282 .0334217 8.296 0.000 .209118 .3454461 cheat .0346766 .0152538 2.273 0.030 .0035663 .0657869 ccool .0279666 .0311424 0.898 0.376 -.0355488 .0914819 celecl -39.40348 9.360317 -4.210 0.000 -58.49397 -20.31299 cons -21.07913 3.250152 -6.486 0.000 -27.70786 -14.4504 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1489.875 -34.11317 1523.988 / * 1.035 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.035*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) 8 Tech3.log 061099 reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 26.20 Model 37067.9516 4 9266.98789 Prob > F = 0.0000 Residual 10966.2552 31 353.750167 R-squared = 0.7717 Adj R-squared = 0.7422 Total 48034.2067 35 1372.40591 Root MSE = 18.808 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .280741 .033566 8.364 0.000 .2122826 .3491993 cheat .0350684 .0153197 2.289 0.029 .0038237 .066313 ccool .0286177 .0312769 0.915 0.367 -.0351719 .0924074 celecl -40.09737 9.400737 -4.265 0.000 -59.2703 -20.92444 cons -27.16277 3.264187 -8.321 0.000 -33.82012 -20.50541 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1491.142 . -40.24387 1531.386 / * 1.04 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.04*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 26.57 Model 38061.9623 4 9515.49057 Prob > F = 0.0000 Residual 11100.3842 31 358.076911 R-squared = 0.7742 + Adj R-squared = 0.7451 Total - 49162.3465 35 1404.63847 Root MSE = 18.923 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ] + dntek I .2842001 .0337707 8.416 0.000 .2153243 .3530758 9 Tech3.log 061099 cheat .0354601 .0154131 2.301 0.028 .0040249 .0668952 ccool .0292689 .0314676 0.930 0.359 -.0349097 .0934474 celec1 -40.7913 9.458053 -4.313 0.000 -60.08112 -21.50147 cons -33.24643 3.284089 -10.123 0.000 -39.94437 -26.54848 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1492.409 -46.37459 1538.784 / * 1.045 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.045*co2[ n-1] (1 missing value generated) gen dco2 = co2 [_n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 26.86 Model 39069.4451 4 9767.36128 Prob > F = 0.0000 Residual 11274.5624 31 363.695561 R-squared = 0.7760 Adj R-squared = 0.7472 Total I 50344.0075 35 1438.40022 Root MSE = 19.071 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2876592 .0340346 8.452 0.000 .2182451 3570732 cheat .0358517 .0155335 2.308 0.028 .0041709 .0675326 ccool .02992 .0317135 0.943 0.353 -.0347601 .0946001 celecl -41.48521 9.531968 -4.352 0.000 -60.92578 -22.04463 cons -39.33006 3.309754 -11.883 0.000 -46.08035 -32.57977 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year 10 Tech3.log 061099 list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1493.677 -52.50525 1546.182 /* 1.05 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.05*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [_n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 27.04 Model 40090.4755 4 10022.6189 Prob > F = 0.0000 Residual 11488.7676 31 370.605405 R-squared = 0.7773 Adj R-squared = 0.7485 Total 51579.243 35 1473.69266 Root MSE = 19.251 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek . 2911184 .0343564 8.473 0.000 .2210481 .3611887 cheat .0362435 .0156804 2.311 0.028 .0042631 .0682238 ccool .0305712 .0320133 0.955 0.347 -.0347205 .0958628 celecl -42.17914 9.622091 -4.384 0.000 -61.80352 -22.55475 cons -45.41371 3.341047 -13.593 0.000 -52.22782 -38.5996 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1494.944 -58.63597 1553.58 / * 1.055 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.055*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) 11 Tech3.log 061099 reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 27.14 Model 41124.9411 4 10281.2353 Prob > F = 0.0000 Residual 11743.0141 31 378.806907 R-squared = 0.7779 Adj R-squared = 0.7492 Total 52867.9552 35 1510.51301 Root MSE = 19.463 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .2945774 .0347345 8.481 0.000 .223736 .3654188 cheat .0366352 .0158529 2.311 0.028 .0043029 .0689674 ccool .0312222 .0323656 0.965 0.342 -.0347879 .0972324 celecl -42.87306 9.727977 -4.407 0.000 -62.7134 -23.03272 cons -51.49735 3.377813 -15.246 0.000 -58.38645 -44.60826 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1496.211 -64.76665 1560.978 /* 1.06 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.06*co2[_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 27.15 Model 42172.8941 4 10543.2235 Prob > F + = 0.0000 Residual 12037.3103 31 388.300331 R-squared = 0.7780 Adj R-squared = 0.7493 Total 54210.2044 35 1548.86298 Root MSE = 19.705 dco2 | Coef. Std. Err. t P>It| [95% Conf. Interval] + 12 Tech3.log 061099 dntek .2980364 .035167 8.475 0.000 .2263128 .3697601 cheat .0370268 .0160504 2.307 0.028 .0042919 .0697617 ccool .0318734 .0327687 0.973 0.338 -.0349588 .0987056 celecl -43.56694 9.849121 -4.423 0.000 -63.65435 -23.47952 cons -57.58099 3.419878 -16.837 0.000 -64.55587 -50.6061 - predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1497.479 -70.89732 1568.376 / * 1.065 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.065*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 27.08 Model 43234.3634 4 10808.5908 Prob > F = 0.0000 Residual 12371.6276 31 399.084762 R-squared = 0.7775 + Adj R-squared = 0.7488 Total 55605.991 35 1588.7426 Root MSE = 19.977 dco2 Coef. Std. Err. t P>|t| I [95% Conf. Interval] dntek .3014954 .035652 8.457 0.000 .2287826 .3742082 cheat .0374186 .0162717 2.300 0.028 .0042322 .070605 ccool .0325246 .0332206 0.979 0.335 -.0352293 .1002785 celec1 -44.26088 9.984956 -4.433 0.000 -64.62533 -23.89643 cons -63.66465 3.467043 -18.363 0.000 -70.73573 -56.59357 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year 13 Tech3.log 061099 list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1498.746 -77.02805 1575.774 / * 1.07 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.07*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 26.94 Model 44309.3824 4 11077.3456 Prob > F = 0.0000 Residual 12745.9942 31 411.161103 R-squared II 0.7766 Adj R-squared = 0.7478 Total 57055.3766 35 1630.15362 Root MSE = 20.277 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .3049547 .0361874 8.427 0.000 .23115 .3787595 cheat .0378102 .0165161 2.289 0.029 .0041255 .071495 ccool .0331757 .0337195 0.984 0.333 - -.0355957 .1019471 celecl -44.95481 10.1349 -4.436 0.000 -65.62508 -24.28454 cons -69.74829 3.519109 -19.820 0.000 -76.92556 -62.57102 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1500.013 -83.15871 1583.172 / * 1.075 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.075*co2 [n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 14 Tech3.log 061099 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 26.73 Model 45397.8136 4 11349.4534 Prob > F = 0.0000 Residual 13160.4247 31 424.529828 R-squared = 0.7753 Adj R-squared = 0.7463 Total 58558.2383 35 1673.09252 Root MSE = 20.604 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3084137 .036771 8.387 0.000 .2334187 .3834087 cheat .0382019 .0167824 2.276 0.030 .0039739 .0724299 ccool .0338268 .0342633 0.987 0.331 - -.0360536 .1037073 celecl -45.6487 10.29835 -4.433 0.000 -66.65233 -24.64508 cons -75.83194 3.575862 -21.207 0.000 -83.12496 -68.53892 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1501.281 -89.28941 1590.57 / * 1.08 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.08*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 + F( 4, 31) = 26.47 Model 46499.7699 4 11624.9425 Prob > F = 0.0000 Residual I 13614.8701 31 439.189357 R-squared = 0.7735 Adj R-squared = 0.7443 Total 60114.64 35 1717.56114 Root MSE Il 20.957 dco2 I Coef. Std. Err. t P>|t| [95% Conf. Interval] 15 Tech3.log 061099 dntek .3118727 .0374005 8.339 0.000 2355939 .3881516 cheat .0385936 .0170697 2.261 0.031 .0037797 .0734076 ccool .0344779 .0348499 0.989 0.330 -.0365988 .1055547 celecl -46.34262 10.47465 -4.424 0.000 -67.70581 -24.97943 cons -81.91558 3.637078 -22.522 0.000 -89.33344 -74.49771 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1502.548 -95.42011 1597.968 /* 1.085 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.085*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 26.15 Model 47615.2361 4 11903.809 Prob > F If 0.0000 Residual 14109.3557 31 455.140507 R-squared II 0.7714 Adj R-squared = 0.7419 Total 61724.5918 35 1763.55977 Root MSE = 21.334 dco2 Coef. Std. Err. t P>Itl [95% Conf. Interval] dntek .3153318 .0380736 8.282 0.000 .2376801 .3929836 cheat .0389853 .017377 2.244 0.032 .0035448 .0744259 ccool .0351292 .0354771 0.990 0.330 -.0372268 .1074852 celecl -47.03655 10.66317 -4.411 0.000 -68.78422 -25.28887 cons -87.99921 3.702537 -23.767 0.000 -95.55059 -80.44784 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) 16 Tech3.log 061099 sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1503.815 -101.5508 1605.366 / * 1.09 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.09*co2[ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 + F( 4, 31) = 25.80 Model 48744.2128 4 12186.0532 Prob > F II 0.0000 Residual 14643.879 31 472.383192 R-squared Il 0.7690 Adj R-squared Il 0.7392 Total 63388.0918 35 1811.08834 Root MSE = 21.734 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek . 3187909 .0387881 8.219 0.000 .239682 .3978999 cheat .0393771 .0177031 2.224 0.034 .0032715 .0754827 ccool .0357802 .0361429 0.990 0.330 -.0379336 .1094941 celecl -47.73049 10.86328 -4.394 0.000 -69.88629 -25.5747 cons -94.08287 3.772019 -24.942 0.000 -101.776 -86.38978 predict pdco2 (13 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 50. 1998 1484 1505.082 -107.6815 1612.764 log close 17 Tech2.log 141099 / > *** Detrend using series as a function of time and time squared > / sort year gen t = year - 1959 gen t2 = t^2 reg nonhitek t t2 if year > 1959 & year < 1998 Source I SS df MS Number of obs = 38 F( 2, 35) = 3373.05 Model 70241913.8 2 35120956.9 Prob > F = 0.0000 Residual 364428.018 35 10412.2291 R-squared = 0.9948 Adj R-squared = 0.9945 Total 70606341.9 37 1908279.51 Root MSE = 102.04 nonhitek I Coef. Std. Err. t P>|t| [95% Conf. Interval] + t 108.6737 6.195223 17.542 0.000 96.09674 121.2507 t2 .3910398 .1540643 2.538 0.016 .0782728 .7038069 cons 2175.67 52.39272 41.526 0.000 2069.307 2282.033 pred tntek reg co2 t t2 if year > 1959 & year < 1998 Source SS df MS Number of obs = 38 F( 2, 35) = 141.13 Model 1173434.20 2 586717.098 Prob > F = 0.0000 Residual 145503.137 35 4157.23248 R-squared = 0.8897 + Adj R-squared = 0.8834 Total 1318937.33 37 35646.9549 Root MSE = 64.477 co2 Coef. Std. Err. t P>|t| [95% Conf. Interval] t 33.95598 3.914598 8.674 0.000 26.00892 41.90304 t2 - -.4776876 .0973492 -4.907 0.000 .6753169 - -.2800584 cons 776.02 33.10558 23.441 0.000 708.8121 843.2279 pred tco2 gen dntek = nonhitek - tntek gen dco2 = co2 - tco2 1 Tech2.log 141099 dntek .3249791 .1333193 2.438 0.021 .0527047 .5972535 cheat .025129 .0518355 0.485 0.631 -.0807331 .1309912 ccool .0122978 .1024298 0.120 0.905 -.1968917 .2214873 celecl 16.67516 42.00755 0.397 0.694 -69.11569 102.466 cons 3.680069 11.4322 0.322 0.750 -19.6676 27.02774 predict pdco2 (13 missing values generated) gen res = dco2 - pdco2 (13 missing values generated) sort year gen resl = res [_n-1] (14 missing values generated) reg res resl, noconstant Source I SS df MS Number of obs = 36 + F( 1, 35) = 170.75 Model 101449.498 1 101449.498 Prob > F = 0.0000 Residual 20795.2708 35 594.150595 R-squared = 0.8299 Adj R-squared = 0.8250 Total 122244.769 36 3395.68803 Root MSE = 24.375 res Coef. Std. Err. t P>|t| [95% Conf. Interval] resl .9551859 .073099 13.067 0.000 .8067871 1.103585 predict reshat (14 missing values generated) gen pco2 = tco2 + pdco2 (13 missing values generated) gen co2hat = tco2 + pdco2 + reshat (14 missing values generated) sort year list year co2 pco2 co2hat if year > 1959 year co2 pco2 co2hat 12. 1960 787.9 13. 1961 793.7 14. 1962 826.6 861.1526 15. 1963 858.6 882.0838 849.0797 16. 1964 895.2 914.7574 892.326 17. 1965 934.5 969.9254 951.2444 18. 1966 985.7 1017.398 983.5597 3 Tech2.log 141099 19. 1967 1011.5 1027.084 996.8068 20. 1968 1069.2 1074.772 1059.887 21. 1969 1115.6 1087.439 1082.116 22. 1970 1149 1069.444 1096.343 23. 1971 1162.5 1097.739 1173.73 24. 1972 1223.7 1156.702 1218.561 25. 1973 1277.9 1195.143 1259.139 26. 1974 1233.1 1187.084 1266.132 27. 1975 1197 1173.222 1217.176 28. 1976 1269.5 1206.286 1228.998 29. 1977 1307.4 1240.279 1300.66 30. 1978 1320.9 1297.921 1362.034 31. 1979 1337.7 1291.802 1313.752 32. 1980 1289.4 1275.659 1319.5 33. 1981 1256 1284.841 1297.966 34. 1982 1189.1 1225.455 1197.907 35. 1983 1182.5 1249.35 1214.624 36. 1984 1245.2 1309.361 1245.507 37. 1985 1244.6 1338.548 1277.262 38. 1986 1245.9 1348.877 1259.14 39. 1987 1286.6 1371.523 1273.161 40. 1988 1348.6 1409.235 1328.117 41. 1989 1360.8 1427.823 1369.905 42. 1990 1346.1 1395.736 1331.717 43. 1991 1330.8 1359.754 1312.342 44. 1992 1352.1 1367.159 1339.503 45. 1993 1379.8 1374.64 1360.255 46. 1994 1398.4 1377.574 1382.504 47. 1995 1411.7 1368.92 1388.812 48. 1996 1460.6 1373.958 1414.821 49. 1997 1479.6 1374.753 1457.512 50. 1998 1484 1373.336 1473.484 / * Get the standard errors correct * / drop pco2 pdco2 reg dco2 dntek cheat ccool celec1 cnetsum cnumele cprpcoal if year > 1961 & y > ear < 1998 Source SS df MS Number of obs = 36 F( 7, 28) = 2.00 Model 47587.6488 7 6798.23554 Prob > F II 0.0903 Residual 94978.2503 28 3392.08037 R-squared II 0.3338 Adj R-squared II 0.1672 Total 142565.899 35 4073.3114 Root MSE II 58.242 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2556151 .134049 1.907 0.067 -.0189719 .5302021 cheat .01618 .0569826 0.284 0.779 -.1005436 .1329035 ccool -.0196871 .1108926 -0.178 0.860 -.2468403 .2074661 celecl 15.9527 35.29283 0.452 0.655 -56.34137 88.24678 cnetsum 1.974694 1.136384 1.738 0.093 -.3530843 4.302471 cnumele .2231086 .3302875 0.675 0.505 -.4534547 .8996719 4 Tech2.log 141099 cprpcoal -1.098215 740.5721 -0.001 0.999 -1518.091 1515.895 cons -40.90295 23.1374 -1.768 0.088 -88.29775 6.49186 predict jdco2 (13 missing values generated) reg celec dntek cheat ccool celecl cnetsum cnumele cprpcoal if year > 1961 & > year < 1998 Source SS df MS Number of obs = 36 F( 7, 28) = 3.75 Model 2.00272606 7 .286103723 Prob > F = 0.0055 Residual 2.13727558 28 .076331271 R-squared = 0.4838 Adj R-squared = 0.3547 Total 4.14000164 35 .118285761 Root MSE = .27628 celec Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .000102 .0006359 0.160 0.874 -.0012006 .0014045 cheat .0002875 .0002703 1.064 0.297 -.0002662 .0008412 ccool .0005813 .000526 1.105 0.279 -.0004962 .0016589 celecl .2379135 .1674189 1.421 0.166 -.1050286 .5808557 cnetsum .0131577 .0053907 2.441 0.021 .0021154 .0242 cnumele -.0047249 .0015668 -3.016 0.005 -.0079343 -.0015154 cprpcoal 4.046307 3.51306 1.152 0.259 -3.149869 11.24248 cons .0293055 .1097571 0.267 0.791 -.1955218 .2541327 predict jcelec (13 missing values generated) newey jdco2 dntek cheat ccool celecl jcelec if year > 1961 & year < 1998, t (y > ear) lag (1) Regression with Newey-West standard errors Number of obs = 36 maximum lag : 1 F( 5, 30) = 16.78 Prob > F = 0.0000 Newey-West jdco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3249791 .0651013 4.992 0.000 .1920245 .4579337 cheat .025129 .0196122 1.281 0.210 -.0149243 .0651824 ccool .0122978 .0312907 0.393 0.697 -.0516064 .076202 celecl 16.67517 16.87041 0.988 0.331 -17.77882 51.12915 jcelec 10.35778 28.35256 0.365 0.717 -47.54587 68.26143 cons 3.680069 6.307196 0.583 0.564 -9.200944 16.56108 log close 5 Tech4.log 141099 / > *** Detrend the series > * / sort year gen dntek = nonhitek [_n] - 1.02808 *nonhitek - 1] (1 missing value generated) gen dco2 = co2[n] - 1.007727*co2[_n - 1] (1 missing value generated) gen tntek = 1.02808*nonhitek[_n-1] (1 missing value generated) gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) gen cheat = htdgyear [_n] - htdgyear [_n-1] (1 missing value generated) gen ccool = cldgyear[_n] - cldgyear[_n-1] (1 missing value generated) gen cfoss = ffppcomr [_n] - ffppcomr[_n-1] (1 missing value generated) gen cres = eleprrer [_n] - eleprrer{_n-1} (1 missing value generated) gen ccom = eleprcor[ n] - eleprcor[_n-1] (1 missing value generated) gen ccoal = ffppcoar [ n] - ffppcoar[_n-1] (1 missing value generated) gen cgas = ffppngr[ n] - ffppngr[_n-1] (1 missing value generated) gen coil = ffppoilr[ n] - ffppoilr[_n-1] (1 missing value generated) gen cnumref = numrefin - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap [n] - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [n] - netsumca[_n-1] (1 missing value generated) gen cextra = extralos [_n] - extralos[_n-1] (1 missing value generated) 1 Tech4.log 141099 gen cpdctu = cpvtyund[_n] - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur [_n] - cpvtysur [_n-1] (2 missing values generated) gen cnumele = numeleut [ n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal [ n] - propcoal [_n-1] (1 missing value generated) gen cresl = cres [_n-1] (2 missing values generated) gen ccoml = ccom [ n-1] (2 missing values generated) gen cfossl = cfoss[_n-1] (2 missing values generated) gen ccoall = ccoal [ n-1] (2 missing values generated) gen cgasl = cgas [_n-1] (2 missing values generated) gen coill = coil [ n-1] (2 missing values generated) / > Residential Electricity > * / reg dco2 dntek cheat ccool cres1 cres (dntek cheat ccool cres1 cnetsum cnumel > e cprpcoal) if year < 1998 & year > 1960 (2SLS) Source I SS df MS Number of obs = 36 F( 5, 30) = 20.35 Model 33204.5064 5 6640.90127 Prob > F II 0.0000 Residual 9618.61781 30 320.620594 R-squared = 0.7754 Adj R-squared = 0.7380 Total 42823.1242 35 1223.51783 Root MSE = 17.906 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -38.19457 25.0231 -1.526 0.137 -89.29855 12.90941 dntek .2037115 .0497293 4.096 0.000 .1021506 .3052723 cheat .0284406 .0148786 1.912 0.066 -.0019455 .0588268 ccool .0136161 .0307071 0.443 0.661 -.0490961 .0763283 cresl -20.30153 13.78901 -1.472 0.151 -48.46244 7.859374 2 Tech4.log 141099 cons I 3.802331 3.430736 1.108 0.277 -3.204166 10.80883 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) test cres cresl (1) cres = 0.0 ( 2) cresl = 0.0 F( 2, 30) = 9.40 Prob > F = 0.0007 gen res = abs (co2 - pco2) (2 missing values generated) gen pres = abs (co2 - pco2) / co2 (2 missing values generated) sort year egen sres = sum (res) if year > 1961 & year < 1998 (3 missing values generated) egen spres = sum (pres) if year > 1961 & year < 1998 (3 missing values generated) replace sres = sres/36 (36 real changes made) replace spres = spres/36 (36 real changes made) list sres spres if year == 1997 sres spres 6. 12.53774 .0103448 sort year list year co2 pco2 year co2 pco2 1. 1960 787.9 2. 1961 793.7 3. 1962 826.6 833.2172 4. 1963 858.6 868.6913 5. 1964 895.2 896.0519 3 Tech4.log 141099 6. 1965 934.5 953.1162 7. 1966 985.7 1010.504 8. 1967 1011.5 1012.588 9. 1968 1069.2 1056.961 10. 1969 1115.6 1119.938 11. 1970 1149 1137.622 12. 1971 1162.5 1169.539 13. 1972 1223.7 1197.881 14. 1973 1277.9 1253.107 15. 1974 1233.1 1228.765 16. 1975 1197 1195.57 17. 1976 1269.5 1231.474 18. 1977 1307.4 1286.804 19. 1978 1320.9 1353.797 20. 1979 1337.7 1335.69 21. 1980 1289.4 1297.376 22. 1981 1256 1261.197 23. 1982 1189.1 1200.138 24. 1983 1182.5 1205.921 25. 1984 1245.2 1242.179 26. 1985 1244.6 1279.089 27. 1986 1245.9 1261.297 28. 1987 1286.6 1276.935 29. 1988 1348.6 1333.497 30. 1989 1360.8 1380.504 31. 1990 1346.1 1348.214 32. 1991 1330.8 1327.44 33. 1992 1352.1 1345.84 34. 1993 1379.8 1371.45 35. 1994 1398.4 1396.819 36. 1995 1411.7 1406.316 37. 1996 1460.6 1431.828 38. 1997 1479.6 1476.042 39. 1998 1484 1484.589 / > *** Summary Stats > / su dco2 dntek cheat ccool cfoss cres ccom ccoal cgas coil cnumref crefcap cne > tsum cextra cpdctu cpdcts cnumele cprpcoal if year < 1998 Variable Obs Mean Std. Dev. Min Max dco2 37 9.487894 34.52911 -76.60513 63.25078 dntek 37 .4687974 94.94416 -235.7253 192.0416 cheat 37 -4.918919 216.0386 -710 353 ccool 37 -1.351351 104.9924 -291 256 cfoss 37 .0108108 .3211125 -1.15 .8200002 cres 37 -.1 .3391166 -.7000008 1 ccom 37 -.0945946 .3299377 -.6000004 1 ccoal 37 -.0018919 .1437946 -.13 .74 cgas 37 .0362162 .2346079 -.71 .49 coil 37 .0143243 .816056 -2.61 2.22 cnumref 37 -3.918919 10.48644 -43 14 crefcap 37 .1516216 .4648567 -1.029999 1.16 4 Tech4.log 141099 cnetsum 37 14.72432 9.749998 2 43.79999 cextra 37 .0113514 .0297361 -.04 .05 cpdctu 37 -.472973 3.286539 -19.91 .26 cpdcts 37 -.3635135 3.328039 -19.95 .8099995 cnumele 37 63.97297 42.49607 -54 137 cprpcoal 37 .0010541 .0144497 - -.0297088 .0360969 log close 5 1/7/00 Hi-Tek V-A GDP Nou-Hitek (Billinus 19925) ( " ) ( " \ 1981Q3 28.7287 4591.93 4563.20 1990Q3 80.7081 6112.29 6 031.58 GDP (1996 chaived Billious) 1992 1 1981 Q3 6643.9 5030.0 1996 = 1.0954 1990Q3 6643.9 6695.4 Tsend Non- Wtek GDP = 3.148% SAAR 1992# 1992 NHTK2 (1996$) (1992) 5/24/00 19812023 H.Tek VA DP GDP Non-Hikk 1981 a3 28.25 5056.8 4616.86 4588.61 1990Q3 79.45 6719.4 6134.81 6055.36 = 3.13% Trend. After GDP revisions. Tech1.log 1/7/00 /******** > *** Detrend the series > * / sort year gen dntek = nonhitek [_n] - 1.03148 *nonhitek - 1] (1 missing value generated) gen dco2 = co2[_n] - 1.007727* co2 [_n - 1] (1 missing value generated) gen tntek = 1.03148*nonhitek[_n-1] (1 missing value generated) gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) gen cheat = htdgyear [_n] - htdgyear [_n-1] (1 missing value generated) gen ccool = cldgyear [_n] - cldgyear[_n-1] (1 missing value generated) gen cfoss = ffppcomr [n] - ffppcomr [_n-1] (1 missing value generated) gen cres = eleprrer [_n] - eleprrer[_n-1] (1 missing value generated) gen ccom = eleprcor [n] - eleprcor[_n-1] (1 missing value generated) gen ccoal = ffppcoar [_n] - ffppcoar[_n-1] (1 missing value generated) gen cgas = ffppngr [_n] - ffppngr[_n-1] (1 missing value generated) gen coil = ffppoilr[_n] - ffppoilr[_n-1] (1 missing value generated) gen cnumref = numrefin [n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap [ n] - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [n] - netsumca[_n-1] (1 missing value generated) gen cextra = extralos [_n] - extralos[_n-1] (1 missing value generated) 1 Tech1.log 1/7/00 . gen cpdctu = cpvtyund[_n - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur[_n] - cpvtysur[_n-1] (2 missing values generated) gen cnumele = numeleut [n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal[_n] - propcoal [_n-1] (1 missing value generated) gen cresl = cres [_n-1] (2 missing values generated) gen ccoml = ccom [_n-1] (2 missing values generated) gen cfoss1 = cfoss [_n-1] (2 missing values generated) gen ccoall = ccoal [_n-1] (2 missing values generated) - gen cgasl = cgas [ n-1] (2 missing values generated) gen coill = coil [n-1] (2 missing values generated) /******* > *** Fossil Index > ******* / / * Current Foss Only * / reg dco2 dntek cheat ccool cfoss if year < 1998 Source I SS df MS Number of obs = 37 F( 4, 32) = 13.02 Model I 26578.2509 4 6644.56272 Prob > F = 0.0000 Residual 16335.8787 32 510.49621 R-squared = 0.6193 Adj R-squared = 0.5718 Total 42914.1296 36 1192.05916 Root MSE = 22.594 dco2 I Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .2950632 .0444559 6.637 0.000 .2045095 .3856169 cheat .0295247 .0180842 1.633 0.112 -.0073116 .066361 ccool .0267653 .0377781 0.708 0.484 - -.0501861 .1037168 cfoss -8.444157 12.12046 -0.697 0.491 -33.13272 16.2444 cons 10.76143 3.720339 2.893 0.007 3.18335 18.33951 2 Tech1.log 1/7/00 reg dco2 dntek cheat ccool cfoss (dntek cheat ccool cnumre crefcap) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 11.15 Model 23761.1072 4 5940.2768 Prob > F = 0.0000 Residual 19153.0224 32 598.53195 R-squared = 0.5537 Adj R-squared = 0.4979 Total 42914.1296 36 1192.05916 Root MSE = 24.465 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cfoss 20.02845 25.45074 0.787 0.437 -31.81302 71.86991 dntek .319671 .0516945 6.184 0.000 .2143728 .4249693 cheat .030556 .0195974 1.559 0.129 -.0093627 .0704747 ccool .0378712 .041781 0.906 0.371 -.0472338 .1229762 cons 10.55627 4.031438 2.618 0.013 2.344496 18.76804 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1466.833 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool cfoss1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 14.05 Model 27597.0144 4 6899.25359 Prob > F = 0.0000 Residual 15218.6723 31 490.924914 R-squared II 0.6446 Adj R-squared = 0.5987 Total 42815.6867 35 1223.30533 Root MSE = 22.157 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2694691 .0467843 5.760 0.000 .1740518 .3648864 cheat .0278148 .0180136 1.544 0.133 -.0089242 .0645538 ccool .0247151 .0371063 0.666 0.510 -.0509636 .1003938 cfossl -21.19378 12.71484 -1.667 0.106 -47.12586 4.738296 cons 10.83366 3.69625 2.931 0.006 3.295114 18.37222 3 Tech1.log 1/7/00 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1476.503 / * Current and lagged foss * / drop pdco2 pco2 reg dco2 dntek cheat ccool cfoss1 cfoss if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 10.90 Model 27611.7716 5 5522.35431 Prob > F = 0.0000 Residual 15203.9152 30 506.797172 R-squared II 0.6449 Adj R-squared = 0.5857 Total 42815.6867 35 1223.30533 Root MSE = 22.512 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2687645 .0477136 5.633 0.000 .1713203 .3662088 cheat .0278002 .0183027 1.519 0.139 -.009579 .0651793 ccool .0240458 .0379048 0.634 0.531 - -.0533661 .1014578 cfoss1 -20.42994 13.67229 -1.494 0.146 -48.35247 7.492595 cfoss -2.181591 12.78464 -0.171 0.866 -28.29131 23.92813 cons 10.84626 3.756252 2.888 0.007 3.174971 18.51755 reg dco2 dntek cheat ccool cfoss1 cfoss (dntek cheat ccool cfoss1 cnumre cref > cap) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 9.20 Model 24383.553 5 4876.71059 Prob > F = 0.0000 Residual 18432.1337 30 614.404458 R-squared !! 0.5695 Adj R-squared = 0.4978 Total 42815.6867 35 1223.30533 Root MSE = 24.787 . dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cfoss 30.08498 28.67437 1.049 0.302 -28.47588 88.64585 dntek .2791855 .0531514 5.253 0.000 .1706359 .3877351 cheat .0280164 .020153 1.390 0.175 -.0131415 .0691744 ccool .0339446 .0424332 0.800 0.430 -.0527156 .1206047 cfoss1 -31.72751 17.41058 -1.822 0.078 -67.28467 3.82964 cons 10.65995 4.138365 2.576 0.015 2.208279 19.11162 4 Tech1.log 1/7/00 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1467.56 / > *** Coal Prices > * / / * Current Coal Only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoal if year < 1998 Source SS df MS Number of obs = 37 F( 4, 32) = 13.00 Model 26568.114 4 6642.0285 Prob > F = 0.0000 Residual 16346.0156 32 510.812988 R-squared = 0.6191 Adj R-squared = 0.5715 Total 42914.1296 36 1192.05916 Root MSE II 22.601 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .293008 .0453409 6.462 0.000 .2006517 .3853644 cheat .0300217 .0180866 1.660 0.107 -.0068196 .066863 ccool .0238948 .0385666 0.620 0.540 -.0546627 .1024524 ccoal -19.14481 28.06847 -0.682 0.500 -76.31841 38.0288 cons 10.62559 3.722092 2.855 0.007 3.043942 18.20725 reg dco2 dntek cheat ccool ccoal (dntek cheat ccool cpdctu cpdcts) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 5.10 Model -365.074834 4 -91.2687085 Prob > F = 0.0027 Residual 43279.2044 32 1352.47514 R-squared = Adj R-squared = . Total 42914.1296 36 1192.05916 Root MSE = 36.776 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccoal -222.9577 230.2258 -0.968 0.340 -691.9122 245.9968 dntek I .1934355 .1326507 1.458 0.155 -.0767651 .4636361 5 Tech1.log 1/7/00 cheat .0320567 .0295162 1.086 0.286 -.0280658 .0921792 ccool -.0417282 .0960038 -0.435 0.667 -.2372816 .1538253 cons 9.827239 6.120645 1.606 0.118 -2.640107 22.29458 - predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1466.537 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoall if year < 1998 Source SS df MS Number of obs = 36 + F( 4, 31) = 12.26 Model 26235.1467 4 6558.78667 Prob > F = 0.0000 Residual 16580.54 31 534.85613 R-squared = 0.6127 + Adj R-squared = 0.5628 Total 42815.6867 35 1223.30533 Root MSE II 23.127 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .3018354 .0452654 6.668 0.000 .209516 .3941549 cheat .030036 .0187908 1.598 0.120 -.0082881 .06836 ccool .030127 .0386135 0.780 0.441 - -.0486257 .1088797 ccoall -1.732004 27.50936 -0.063 0.950 -57.83772 54.37371 cons 10.67016 3.857248 2.766 0.009 2.803248 18.53707 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1473.632 / * Current and lagged foss * / drop pdco2 pco2 6 Tech1.log 1/7/00 reg dco2 dntek cheat ccool ccoall ccoal if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 9.73 Model 26485.9716 5 5297.19431 Prob > F = 0.0000 Residual 16329.7152 30 544.323839 R-squared = 0.6186 Adj R-squared = 0.5550 Total 42815.6867 35 1223.30533 Root MSE = 23.331 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .293909 .0471336 6.236 0.000 .1976494 .3901686 cheat .0298546 .0189582 1.575 0.126 -.0088633 .0685725 ccool .0236718 .0400977 0.590 0.559 -.0582186 .1055622 ccoall 5.093006 29.5169 0.173 0.864 -55.18855 65.37456 ccoal -20.94069 30.84852 -0.679 0.502 -83.94177 42.06039 cons 10.62306 3.891856 2.730 0.011 2.67483 18.57129 reg dco2 dntek cheat ccool ccoall ccoal (dntek cheat ccool ccoall cpdctu cpdc > ts) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 1.36 Model -80248.9922 5 -16049.7984 Prob > F = 0.2679 Residual 123064.679 30 4102.15596 R-squared = . Adj R-squared = - Total 42815.6867 35 1223.30533 Root MSE = 64.048 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccoal 411.0347 655.8316 0.627 0.536 -928.3521 1750.421 dntek . 4574188 .2780996 1.645 0.110 -.1105363 1.025374 cheat .0335955 .0523484 0.642 0.526 -.0733142 .1405053 ccool .1568329 .2287072 0.686 0.498 -.3102496 .6239153 ccoall -135.6969 226.9204 -0.598 0.554 -599.1302 327.7365 cons 11.59459 10.78366 1.075 0.291 -10.42858 33.61777 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1486.937 /******* > *** Oil Prices 7 Tech1.log 1/7/00 > ******* / / * Current Coal Only * / drop pdco2 pco2 reg dco2 dntek cheat ccool coil if year < 1998 Source SS df MS Number of obs = 37 F( 4, 32) = 12.84 Model 26440.9343 4 6610.23358 Prob > F = 0.0000 Residual 16473.1953 32 514.787353 R-squared = 0.6161 Adj R-squared = 0.5682 Total 42914.1296 36 1192.05916 Root MSE = 22.689 dco2 Coef. Std. Err. t P>|t [95% Conf. Interval] dntek .299496 .043824 6.834 0.000 .2102294 .3887626 cheat .0286743 .0183254 1.565 0.127 -.0086534 .0660021 ccool .0277012 .0379809 0.729 0.471 -.0496635 .1050658 coil -2.195324 4.739168 -0.463 0.646 -11.84869 7.458046 cons 10.71355 3.735018 2.868 0.007 3.105564 18.32153 reg dco2 dntek cheat ccool coil (dntek cheat ccool cnumre crefcap) if year < > 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 11.44 Model 24246.2102 4 6061.55255 Prob > F = 0.0000 Residual 18667.9194 32 583.372482 R-squared = 0.5650 Adj R-squared = 0.5106 Total 42914.1296 36 1192.05916 Root MSE = 24.153 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] coil 7.590067 9.65071 0.786 0.437 -12.06779 27.24792 dntek .3122672 .0478718 6.523 0.000 .2147556 .4097788 cheat .0338279 .0199834 1.693 0.100 -.006877 .0745328 ccool .038211 .0413862 0.923 0.363 -.04609 .122512 cons 10.65578 3.976345 2.680 0.012 2.55623 18.75533 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1465.158 8 Tech1.log 1/7/00 / * Lagged foss only * / drop pdco2 pco2 reg dco2 dntek cheat ccool coill if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 12.87 Model 26725.727 4 6681.43175 Prob > F = 0.0000 Residual 16089.9597 31 519.030958 R-squared = 0.6242 Adj R-squared = 0.5757 Total 42815.6867 35 1223.30533 Root MSE = 22.782 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2774027 .0506027 5.482 0.000 .1741979 .3806075 cheat .0284117 .0185431 1.532 0.136 -.0094072 .0662305 ccool .0263869 .038206 0.691 0.495 -.0515348 .1043087 coill -5.280549 5.419812 -0.974 0.337 -16.33433 5.77323 cons 10.71304 3.799519 2.820 0.008 2.963871 18.46221 predict pdco2 (2 missing values generated) # gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1476.443 / * Current and lagged foss * / drop pdco2 pco2 reg dco2 dntek cheat ccool coill coil if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 10.02 Model 26775.1056 5 5355.02113 Prob > F = 0.0000 Residual 16040.5811 30 534.686036 R-squared = 0.6254 Adj R-squared = 0.5629 Total 42815.6867 35 1223.30533 Root MSE = 23.123 dco2 Coef. Std. Err. t P>|t| I [95% Conf. Interval] dntek .2767223 .0514089 5.383 0.000 .1717313 .3817133 cheat .0276717 .0189775 1.458 0.155 -.0110855 .066429 ccool .0249385 .0390697 0.638 0.528 -.0548525 .1047296 coill -5.010263 5.57238 -0.899 0.376 -16.39058 6.370055 9 Tech1.log 1/7/00 coil -1.487905 4.896153 -0.304 0.763 -11.48718 8.511374 cons 10.72705 3.856669 2.781 0.009 2.850679 18.60342 reg dco2 dntek cheat ccool coill coil (dntek cheat ccool coill cnumre crefcap > ) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 + F( 5, 30) = 8.67 Model 23923.6235 5 4784.7247 Prob > F = 0.0000 Residual 18892.0632 30 629.735441 R-squared = 0.5588 + Adj R-squared = 0.4852 Total I 42815.6867 35 1223.30533 Root MSE = 25.095 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + coil 9.818926 10.42164 0.942 0.354 -11.46491 31.10276 dntek .2818929 .055942 5.039 0.000 .1676442 .3961416 cheat .0332944 .0210723 1.580 0.125 -.009741 .0763298 ccool .0359451 .0432893 0.830 0.413 -.0524634 .1243536 coill -7.064215 6.26288 -1.128 0.268 -19.85472 5.726293 cons 10.6206 4.1863 2.537 0.017 2.071039 19.17017 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1466.255 /******* > *** Residential Electricity > ******* / / * Current only * / drop pdco2 pco2 reg dco2 dntek cheat ccool cres if year < 1998 Source SS df MS Number of obs = 37 F( 4, 32) = 21.04 Model 31091.3447 4 7772.83618 Prob > F = 0.0000 Residual 11822.7849 32 369.462028 R-squared = 0.7245 Adj R-squared = 0.6901 Total 42914.1296 36 1192.05916 Root MSE = 19.221 dco2 I Coef. Std. Err. t P>|t| [95% Conf. Interval] 10 Tech1.log 1/7/00 dntek .2260852 .0424549 5.325 0.000 .1396075 .3125629 cheat .027363 .0153955 1.777 0.085 -.0039965 .0587226 ccool .0175285 .0320766 0.546 0.589 -.0478093 .0828664 cres -39.20205 10.9207 -3.590 0.001 -61.44679 -16.95732 cons 6.495383 3.374005 1.925 0.063 -.3772403 13.36801 reg dco2 dntek cheat ccool cres (dntek cheat ccool cnetsum cnumele cprpcoal) > if year < 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 18.27 Model 29905.5283 4 7476.38207 Prob > F = 0.0000 Residual 13008.6013 32 406.518791 R-squared = 0.6969 Adj R-squared = 0.6590 Total 42914.1296 36 1192.05916 Root MSE = 20.162 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -58.7668 20.38959 -2.882 0.007 -100.299 -17.23457 dntek .1880178 .05532 3.399 0.002 .0753346 .3007009 cheat .0261315 .016184 1.615 0.116 -.0068341 .0590972 ccool .0112749 .034076 0.331 0.743 -.0581356 .0806854 cons 4.396674 3.974864 1.106 0.277 -3.699859 12.49321 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1477.643 / * Lagged only * / drop pdco2 pco2 reg dco2 dntek cheat ccool cresl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 21.35 Model 31411.6358 4 7852.90894 Prob > F = 0.0000 Residual 11404.0509 31 367.872611 R-squared = 0.7336 Adj R-squared = 0.6993 Total 42815.6867 35 1223.30533 Root MSE = 19.18 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2810719 .0371414 7.568 0.000 .2053216 .3568222 11 Tech1.log 1/7/00 cheat .0354186 .0156214 2.267 0.030 .0035587 .0672786 ccool .0317545 .0319954 0.992 0.329 -.0335006 .0970096 cresl -35.99991 9.594969 -3.752 0.001 -55.56898 -16.43084 cons 7.110895 3.336568 2.131 0.041 .3059192 13.91587 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1475.724 / * current and lagged prices * / drop pdco2 pco2 reg dco2 dntek cheat ccool cresl cres if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 20.16 Model 32994.5459 5 6598.90918 Prob > F = 0.0000 Residual 9821.1408 30 327.37136 R-squared = 0.7706 Adj R-squared = 0.7324 Total 42815.6867 35 1223.30533 Root MSE = 18.093 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2373973 .0402753 5.894 0.000 .1551442 .3196505 cheat .0320169 .0148174 2.161 0.039 .0017558 .0622779 ccool .0228764 .0304517 0.751 0.458 -.0393142 .0850669 cresl -25.36774 10.26191 -2.472 0.019 -46.32535 -4.410121 cres -25.63882 11.65978 -2.199 0.036 -49.45127 -1.826379 cons 5.448693 3.237042 1.683 0.103 -1.162228 12.05961 reg dco2 dntek cheat ccool cresl cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 19.01 Model 32698.8207 5 6539.76415 Prob > F = 0.0000 Residual 10116.866 30 337.228866 R-squared = 0.7637 Adj R-squared = 0.7243 Total 42815.6867 35 1223.30533 Root MSE = 18.364 dco2 I Coef. Std. Err. t P>|t| I [95% Conf. Interval] cres I -36.72072 26.67595 -1.377 0.179 -91.20027 17.75883 12 Tech1.log 1/7/00 dntek .2185198 .0577016 3.787 0.001 .1006774 .3363623 cheat .0305465 .0153697 1.987 0.056 -.0008426 .0619356 ccool .019039 .0319962 0.595 0.556 -.0463061 .084384 cresl -20.77218 14.37943 -1.445 0.159 -50.1389 8.594546 cons 4.730238 3.632671 1.302 0.203 -2.688665 12.14914 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1477.432 / > *** Commercial Electricity > ******* / / * Current only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccom if year < 1998 Source SS df MS Number of obs = 37 F( 4, 32) = 22.67 Model 31719.926 4 7929.9815 Prob > F = 0.0000 Residual 11194.2036 32 349.818863 R-squared Il 0.7391 Adj R-squared = 0.7065 Total 42914.1296 36 1192.05916 Root MSE = 18.703 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2283654 .0404288 5.649 0.000 .1460147 .3107161 cheat .0315505 .0149721 2.107 0.043 .0010534 .0620476 ccool .0165476 .0312172 0.530 0.600 -.0470399 .080135 ccom -42.00009 10.70038 -3.925 0.000 -63.79605 -20.20412 cons 6.46953 3.262096 1.983 0.056 -.1751425 13.1142 reg dco2 dntek cheat ccool ccom (dntek cheat ccool cnetsum cnumele cprpcoal) > if year < 1998 (2SLS) Source SS df MS Number of obs = 37 F( 4, 32) = 20.12 Model 31298.6293 4 7824.65732 Prob > F = 0.0000 Residual 11615.5003 32 362.984385 R-squared il 0.7293 Adj R-squared = 0.6955 Total 42914.1296 36 1192.05916 Root MSE = 19.052 13 Tech1.log 1/7/00 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ccom -53.74289 19.09151 -2.815 0.008 -92.63102 -14.85476 dntek .2076769 .049584 4.188 0.000 .1066776 .3086762 cheat .0320314 .0152647 2.098 0.044 .0009382 .0631246 ccool . 0127699 .0321965 0.397 0.694 -.0528123 .0783521 cons 5.28657 3.678994 1.437 0.160 -2.207297 12.78044 predict pdco2 (1 missing value generated) gen pco2 = tco2 + pdco2 (1 missing value generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1479.097 / * Lagged only * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccoml if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 17.02 Model 29418.7547 4 7354.68868 Prob > F = 0.0000 Residual 13396.932 31 432.159097 R-squared = 0.6871 Adj R-squared = 0.6467 Total 42815.6867 35 1223.30533 Root MSE = 20.788 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .277966 .0407883 6.815 0.000 .1947778 .3611542 cheat .0356403 .0169871 2.098 0.044 .000995 .0702857 ccool .0300232 .0346759 0.866 0.393 -.0406987 .1007451 ccom1 -29.46616 10.85278 -2.715 0.011 -51.60056 -7.33177 cons 7.818191 3.622846 2.158 0.039 .4293482 15.20703 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1471.796 14 Tech1.log 1/7/00 / * current and lagged prices * / drop pdco2 pco2 reg dco2 dntek cheat ccool ccom1 ccom if year < 1998 Source SS df MS Number of obs = 36 F( 5, 30) = 18.25 Model 32221.9527 5 6444.39053 Prob > F = 0.0000 Residual 10593.7341 30 353.124469 R-squared = 0.7526 Adj R-squared = 0.7113 Total 42815.6867 35 1223.30533 Root MSE = 18.792 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2296071 .0406696 5.646 0.000 .1465486 .3126655 cheat .03401 .0153663 2.213 0.035 .0026279 .0653921 ccool .0188358 .0315956 0.596 0.556 -.045691 .0833626 ccom1 -14.53645 11.14994 -1.304 0.202 -37.30766 8.23476 ccom -34.4339 12.22146 -2.817 0.008 -59.39346 -9.47434 cons 5.829398 3.350063 1.740 0.092 -1.012344 12.67114 reg dco2 dntek cheat ccool ccom1 ccom (dntek cheat ccool ccom1 cnetsum cnumel > e cprpcoal) if year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 16.90 Model 32088.9239 5 6417.78478 Prob > F = 0.0000 Residual 10726.7628 30 357.558761 R-squared = 0.7495 Adj R-squared = 0.7077 Total 42815.6867 35 1223.30533 Root MSE = 18.909 dco2 Coef. Std. Err. t P>Itl [95% Conf. Interval] ccom -41.93512 27.99561 -1.498 0.145 -99.10979 15.23955 dntek .2190724 .0540585 4.053 0.000 .1086702 .3294745 cheat .0336549 .0155082 2.170 0.038 .0019828 .0653269 ccool .0163987 .0328265 0.500 0.621 -.0506421 .0834395 ccoml -11.2841 15.64569 -0.721 0.476 -43.23685 20.66866 cons 5.396151 3.670671 1.470 0.152 -2.10036 12.89266 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) list year co2 pco2 if year == 1998 year co2 pco2 39. 1998 1485.4 1477.243 15 Tech1.log 1/7/00 log close 16 Tech2.log 1/7/00 / > *** Detrend using series as a function of time and time squared > / sort year gen t = year - 1959 gen t2 = t^2 reg nonhitek t t2 if year > 1959 & year < 1998 Source I SS df MS Number of obs = 38 + F( 2, 35) = 3838.06 Model I 78121131.4 2 39060565.7 Prob > F = 0.0000 Residual I 356201.041 35 10177.1726 R-squared = 0.9955 + Adj R-squared = 0.9952 Total I 78477332.4 37 2121008.98 Root MSE = 100.88 nonhitek Coef. Std. Err. t P>|t| [95% Conf. Interval] + t 89.57803 6.124895 14.625 0.000 77.14383 102.0122 t2 1.045451 .1523153 6.864 0.000 .7362346 1.354668 cons 2119.505 51.79796 40.919 0.000 2014.35 2224.661 pred tntek reg co2 t t2 if year > 1959 & year < 1998 Source | SS df MS Number of obs = 38 + F( 2, 35) = 141.08 Model 1173540.35 2 586770.175 Prob > F = 0.0000 Residual 145569.29 35 4159.12256 R-squared = 0.8896 + Adj R-squared = 0.8833 Total 1319109.64 37 35651.6119 Root MSE = 64.491 . co2 1 Coef. Std. Err. t P>|t| [95% Conf. Interval] t 33.95098 3.915488 8.671 0.000 26.00211 41.89984 t2 -.4775276 .0973713 -4.904 0.000 -.6752018 - .2798534 cons 776.0454 33.11311 23.436 0.000 708.8222 843.2686 pred tco2 gen dntek = nonhitek - tntek gen dco2 = co2 - tco2 1 Tech2.log 1/7/00 gen cheat = htdgyear [_n] - htdgyear[_n-1] (1 missing value generated) gen ccool = cldgyear [n] - cldgyear[_n-1] (1 missing value generated) gen cfossl = ffppcomr [ n-1] - ffppcomr[_n-2] (2 missing values generated) gen celec = eleprrer [_n] - eleprrer[_n-1] (1 missing value generated) gen celecl = eleprrer [n-1] - eleprrer [_n-2] (2 missing values generated) gen cnumref = numrefin [n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [n] - netsumca [n-1] (1 missing value generated) gen cextra = extralos [n] - extralos [_n-1] (1 missing value generated) gen cpdctu = cpvtyund[ - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur [n] - cpvtysur[_n-1] (2 missing values generated) gen cnumele = numeleut [n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal [_n] - propcoal [_n-1] (1 missing value generated) reg dco2 dntek cheat ccool celec1 celec (dntek cheat ccool celec1 cnetsum cnu > mele cprpcoal) if year > 1961 & year < 1998 (2SLS) Source SS df MS Number of obs = 36 F( 5, 30) = 2.40 Model 45401.9393 5 9080.38787 Prob > F = 0.0605 Residual 97227.5073 30 3240.91691 R-squared = 0.3183 Adj R-squared = 0.2047 Total 142629.447 35 4075.12705 Root MSE = 56.929 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] celec I 26.42162 61.93791 0.427 0.673 -100.0725 152.9157 2 Tech2.log 1/7/00 dntek .4161244 .1316183 3.162 0.004 .1473241 .6849248 cheat .0308702 .0483983 0.638 0.528 -.0679724 .1297128 ccool .0277248 .0961688 0.288 0.775 -.1686782 .2241278 celecl 25.65706 39.58241 0.648 0.522 -55.18101 106.4951 cons I 5.249081 10.69211 0.491 0.627 -16.58713 27.08529 predict pdco2 (2 missing values generated) gen res = dco2 - pdco2 (2 missing values generated) sort year gen resl = res[_n-1] (3 missing values generated) reg res res1, noconstant Source I SS df MS Number of obs = 36 F( 1, 35) II 125.39 Model 86714.4116 1 86714.4116 Prob > F = 0.0000 Residual 24203.6713 35 691.533465 R-squared = 0.7818 Adj R-squared = 0.7756 Total 110918.083 36 3081.05786 Root MSE II 26.297 res Coef. Std. Err. t P>|t| [95% Conf. Interval] resl 9443893 .0843358 11.198 0.000 7731785 1.1156 predict reshat (3 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) gen co2hat = tco2 + pdco2 + reshat (3 missing values generated) sort year list year co2 pco2 co2hat if year > 1959 year co2 pco2 co2hat 1. 1960 787.9 . . 2. 1961 793.7 . . 3. 1962 826.6 849.9268 . 4. 1963 858.6 868.607 846.5774 5. 1964 895.2 906.7777 897.3272 6. 1965 934.5 966.4081 955.4742 7. 1966 985.7 1017.399 987.2653 3 Tech2.log 1/7/00 8. 1967 1011.5 1029.756 999.8203 9. 1968 1069.2 1086.026 1068.785 10. 1969 1115.6 1092.537 1076.646 11. 1970 1149 1068.521 1090.301 12. 1971 1162.5 1100.846 1176.85 13. 1972 1223.7 1166.109 1224.335 14. 1973 1277.9 1205.359 1259.747 15. 1974 1233.1 1206.354 1274.86 16. 1975 1197 1181.993 1207.252 17. 1976 1269.5 1203.174 1217.347 18. 1977 1307.4 1242.638 1305.275 19. 1978 1320.9 1301.18 1362.341 20. 1979 1337.7 1293.344 1311.967 21. 1980 1289.4 1287.324 1329.213 22. 1981 1256 1289.778 1291.739 23. 1982 1189.1 1210.842 1178.943 24. 1983 1182.5 1232.261 1211.728 25. 1984 1245.2 1294.459 1247.465 26. 1985 1244.6 1328.91 1282.39 27. 1986 1245.9 1340.387 1260.766 28. 1987 1286.6 1368.801 1279.569 29. 1988 1348.6 1415.759 1338.129 30. 1989 1360.8 1430.417 1366.993 31. 1990 1346.1 1395.804 1330.059 32. 1991 1330.8 1357.438 1310.499 33. 1992 1352.1 1369.327 1344.17 34. 1993 1379.8 1374.29 1358.021 35. 1994 1398.4 1380.811 1386.014 36. 1995 1411.7 1370.931 1387.542 37. 1996 1460.5 1371.36 1409.862 38. 1997 1480 1377.845 1462.027 39. 1998 1485.4 1366.091 1462.565 / * Get the standard errors correct * / drop pco2 pdco2 reg dco2 dntek cheat ccool celecl cnetsum cnumele cprpcoal if year > 1961 & y > ear < 1998 Source I SS df MS Number of obs = 36 + F( 7, 28) = 2.46 Model 54375.6598 7 7767.9514 Prob > F = 0.0420 Residual 88253.7868 28 3151.92096 R-squared II 0.3812 Adj R-squared II 0.2265 Total 142629.447 35 4075.12705 Root MSE = 56.142 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + dntek .3373357 .1369748 2.463 0.020 .0567555 . 6179158 cheat .0256565 .0548182 0.468 0.643 -.0866335 .1379466 ccool .0037869 .1078407 0.035 0.972 -.2171149 .2246886 celecl 26.00033 34.67416 0.750 0.460 -45.02647 97.02713 cnetsum 1.977037 1.088829 1.816 0.080 -.2533281 4.207403 cnumele .1434818 .3187264 0.450 0.656 -.5093996 .7963631 4 Tech2.log 1/7/00 cprpcoal 113.3318 717.8583 0.158 0.876 -1357.134 1583.798 cons -35.69672 22.4991 -1.587 0.124 -81.78403 10.39059 predict jdco2 (2 missing values generated) reg celec dntek cheat ccool celec1 cnetsum cnumele cprpcoal if year > 1961 & > year < 1998 Source SS df MS Number of obs = 36 F( 7, 28) = 3.75 Model 2.00184997 7 .285978568 Prob > F = 0.0055 Residual 2.13815167 28 .07636256 R-squared = 0.4835 Adj R-squared = 0.3544 Total 4.14000164 35 . 118285761 Root MSE = .27634 celec Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .0000804 .0006742 0.119 0.906 -.0013006 .0014615 cheat .000283 .0002698 1.049 0.303 -.0002697 .0008357 ccool .0005762 .0005308 1.085 0.287 -.0005112 .0016635 celec1 .2366476 .1706705 1.387 0.177 -.1129551 .5862504 cnetsum .0130937 .0053594 2.443 0.021 .0021155 .0240718 cnumele -.0046928 .0015688 -2.991 0.006 -.0079063 -.0014792 cprpcoal 4.05057 3.533388 1.146 0.261 -3.187247 11.28839 cons .0279816 .1107434 0.253 0.802 -.1988659 .2548291 predict jcelec (2 missing values generated) newey jdco2 dntek cheat ccool celecl jcelec if year > 1961 & year < 1998, t (y > ear) lag (1) Regression with Newey-West standard errors Number of obs = 36 maximum lag : 1 F( 5, 30) = 28.14 Prob > F = 0.0000 Newey-West jdco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .4161244 .0605748 6.870 0.000 .2924141 .5398348 cheat .0308702 .0181335 1.702 0.099 -.0061633 .0679038 ccool .0277248 .029021 0.955 0.347 -.0315439 .0869935 celecl 25.65706 15.10975 1.698 0.100 -5.201163 56.51528 jcelec 26.42162 25.88745 1.021 0.316 -26.4476 79.29084 cons 5.249081 5.649554 0.929 0.360 -6.288848 16.78701 log close 5 Tech3.log 1/7/00 /******** > *** Detrend the series > * / sort year gen dntek = nonhitek[ : - 1.03148*nonhitek[_n - 1] (1 missing value generated) gen tntek = 1.03148*nonhitek[_n-1] (1 missing value generated) gen cheat = htdgyear[_n] - htdgyear[_n-1] (1 missing value generated) gen ccool = cldgyear[_n] - cldgyear [_n-1] (1 missing value generated) gen cfoss = ffppcomr [_n] - ffppcomr [_n-1] (1 missing value generated) gen celec = eleprrer [_n] - eleprrer[_n-1] (1 missing value generated) gen cnumref = numrefin[ n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap[_n] - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [n] - netsumca[_n-1] (1 missing value generated) gen cextra = extralos [n] - extralos[_n-1] (1 missing value generated) gen cpdctu = cpvtyund[_n] - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur [ n] - cpvtysur[_n-1] (2 missing values generated) gen cnumele = numeleut [_n] - numeleut [_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal[_n] - propcoal [_n-1] (1 missing value generated) gen celecl = celec[_n-1] (2 missing values generated) gen cfoss1 = cfoss[_n-1] 1 Tech3.log 1/7/00 (2 missing values generated) / * 1 * / gen tco2 = co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 20.24 Model 30102.967 4 7525.74175 Prob > F = 0.0000 Residual 11527.5507 31 371.856475 R-squared = 0.7231 Adj R-squared = 0.6874 Total I 41630.5177 35 1189.44336 Root MSE = 19.284 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2757985 .0373419 7.386 0.000 .1996392 .3519579 cheat .0347627 .0157057 2.213 0.034 .0027306 .0667947 ccool .0306602 .0321682 0.953 0.348 -.0349473 .0962677 celecl -34.9154 9.646783 -3.619 0.001 -54.59015 -15.24066 cons 16.49371 3.354586 4.917 0.000 9.651983 23.33543 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1473.885 -6.11469 1480 / * 1.0025 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.0025*co2[ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 2 Tech3.log 1/7/00 Source SS df MS Number of obs = 36 F( 4, 31) = 20.61 Model 30523.176 4 7630.79401 Prob > F = 0.0000 Residual 11476.7904 31 370.219047 R-squared = 0.7267 Adj R-squared = 0.6915 Total 41999.9665 35 1199.99904 Root MSE = 19.241 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2775046 .0372596 7.448 0.000 .2015131 .3534961 cheat .0349749 .0156711 2.232 0.033 .0030134 .0669363 ccool .0310143 .0320973 0.966 0.341 .0344486 .0964772 celec1 -35.26631 9.625521 -3.664 0.001 -54.89769 -15.63493 cons 13.45798 3.347192 4.021 0.000 6.631338 20.28463 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1474.48 -9.219649 1483.7 / * 1.005 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.005*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 20.97 Model 30946.4647 4 7736.61618 Prob > F il 0.0000 Residual 11436.3686 31 368.915115 R-squared = 0.7302 Adj R-squared = 0.6953 Total 42382.8333 35 1210.93809 Root MSE = 19.207 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2792108 .037194 7.507 0.000 .2033532 .3550684 cheat .0351872 .0156435 2.249 0.032 .0032821 . 0670923 ccool .0313683 .0320407 0.979 0.335 -.0339792 .0967158 3 Tech3.log 1/7/00 celec1 -35.61719 9.608555 -3.707 0.001 -55.21397 -16.02042 cons 10.42226 3.341293 3.119 0.004 3.607646 17.23687 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1475.075 -12.32473 1487.4 / * 1.0075 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.0075 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 21.32 Model 31372.7804 4 7843.1951 Prob > F H 0.0000 Residual 11406.2653 31 367.944041 R-squared II 0.7334 Adj R-squared = 0.6990 Total 42779.0457 35 1222.25845 Root MSE = 19.182 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2809169 .037145 7.563 0.000 .2051593 .3566746 cheat .0353994 .0156229 2.266 0.031 .0035363 .0672625 ccool .0317224 .0319985 0.991 0.329 -.0335391 .0969838 celecl -35.96808 9.5959 -3.748 0.001 -55.53904 -16.39711 cons 7.386531 3.336892 2.214 0.034 .5808952 14.19217 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 4 Tech3.log 1/7/00 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1475.67 -15.42975 1491.1 / * 1.01 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.01*co2 [_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 21.65 Model 31802.1815 4 7950.54537 Prob > F = 0.0000 Residual 11386.5162 31 367.306973 R-squared = 0.7364 Adj R-squared II 0.7023 Total 43188.6977 35 1233.96279 Root MSE = 19.165 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2826233 .0371128 7.615 0.000 .2069312 .3583154 cheat .0356116 .0156093 2.281 0.030 .0037762 .0674471 ccool .0320764 .0319708 1.003 0.323 -.0331285 .0972814 celec1 -36.31896 9.58759 -3.788 0.001 -55.87297 -16.76494 cons 4.350814 3.334002 1.305 0.202 -2.448928 11.15056 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1476.265 -18.5348 1494.8 / * 1.015 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.015*co2[ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) 5 Tech3.log 1/7/00 reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 22.25 Model 32670.0036 4 8167.5009 Prob > F = 0.0000 Residual 11378.0128 31 367.032672 R-squared = 0.7417 Adj R-squared = 0.7084 Total 44048.0164 35 1258.51475 Root MSE = 19.158 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval ] dntek .2860354 .0370989 7.710 0.000 .2103716 .3616992 cheat .0360361 .0156035 2.309 0.028 .0042125 .0678596 ccool .0327845 .0319589 1.026 0.313 -.0323961 .0979651 celecl -37.02067 9.584009 -3.863 0.001 -56.56738 -17.47395 cons -1.720619 3.332757 -0.516 0.609 -8.517821 5.076584 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1477.455 -24.7448 1502.2 / * 1.02 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.02*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 22.79 Model 33550.0692 4 8387.51731 Prob > F = 0.0000 Residual 11410.8466 31 368.091827 R-squared = 0.7462 Adj R-squared = 0.7135 Total 44960.9159 35 1284.5976 Root MSE = 19.186 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2894477 .0371524 7.791 0.000 .2136748 .3652206 cheat .0364605 .015626 2.333 0.026 .004591 .0683299 6 Tech3.log 1/7/00 ccool .0334926 .032005 1.046 0.303 -.031782 .0987671 celecl -37.72243 9.597827 -3.930 0.000 -57.29733 -18.14754 cons -7.792066 3.337562 -2.335 0.026 -14.59907 - -.9850633 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1478.645 -30.95484 1509.6 / * 1.025 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.025*co2 [_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 23.24 Model 34442.3379 4 8610.58447 Prob > F = 0.0000 Residual 11485.0183 31 370.484463 R-squared = 0.7499 Adj R-squared = 0.7177 Total 45927.3562 35 1312.21018 Root MSE = 19.248 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2928599 .037273 7.857 0.000 .2168411 .3688787 cheat .036885 .0156767 2.353 0.025 .0049121 .0688579 ccool .0342007 .0321088 1.065 0.295 -.0312856 .0996871 celecl -38.42425 9.62897 -3.990 0.000 -58.06266 -18.78584 cons -13.86353 3.348392 -4.140 0.000 -20.69262 -7.034435 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 7 Tech3.log 1/7/00 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1479.835 -37.16489 1517 / * 1.03 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.03*co2[_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 + F( 4, 31) = 23.61 Model 35346.8238 4 8836.70594 Prob > F II 0.0000 Residual 11600.5284 31 374.210595 R-squared = 0.7529 + Adj R-squared = 0.7210 Total 46947.3522 35 1341.35292 Root MSE = 19.345 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2962724 .03746 7.909 0.000 .2198723 .3726725 cheat .0373094 .0157554 2.368 0.024 .0051762 .0694427 ccool .0349089 .0322699 1.082 0.288 -.0309059 .1007237 celecl -39.12601 9.677271 -4.043 0.000 -58.86293 -19.38909 cons -19.93496 3.365188 -5.924 0.000 -26.79831 -13.07162 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1481.025 -43.3749 1524.4 / * 1.035 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.035*co2 [n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) 8 Tech3.log 1/7/00 reg dco2 dntek cheat ccool celec1 if year < 1998 Source I SS df MS Number of obs = 36 F( 4, 31) = 23.90 Model 36263.4323 4 9065.85807 Prob > F = 0.0000 Residual 11757.3869 31 379.270547 R-squared = 0.7552 Adj R-squared = 0.7236 Total 48020.8192 35 1372.02341 Root MSE = 19.475 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .2996845 .0377124 7.947 0.000 .2227697 .3765994 cheat .0377339 .0158615 2.379 0.024 .0053842 .0700837 ccool .0356171 .0324873 1.096 0.281 -.0306412 .1018754 celecl -39.82774 9.742477 -4.088 0.000 -59.69766 -19.95783 cons -26.00641 3.387863 -7.676 0.000 -32.916 -19.09682 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1482.215 -49.58496 1531.8 /* 1.04 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.04*co2[_n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.11 Model 37192.2798 4 9298.06994 Prob > F = 0.0000 Residual 11955.5803 31 385.663879 R-squared = 0.7567 Adj R-squared = 0.7254 Total 49147.86 35 1404.22457 Root MSE = 19.638 dco2 I Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3030968 .0380289 7.970 0.000 .2255364 .3806573 9 Tech3.log 1/7/00 cheat .0381584 .0159946 2.386 0.023 .0055371 .0707797 ccool .0363252 .03276 1.109 0.276 - -.0304892 .1031396 celecl -40.52952 9.824248 -4.125 0.000 -60.56621 -20.49284 cons -32.07787 3.416298 -9.390 0.000 -39.04545 -25.11028 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1483.405 -55.79503 1539.2 / * 1.045 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.045*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.23 Model 38133.2977 4 9533.32444 Prob > F = 0.0000 Residual 12195.119 31 393.390935 R-squared = 0.7577 Adj R-squared = 0.7264 Total 50328.4167 35 1437.95476 Root MSE = 19.834 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3065091 .038408 7.980 0.000 .2281755 .3848427 cheat .0385828 .0161541 2.388 0.023 .0056364 .0715293 ccool .0370333 .0330865 1.119 0.272 -.0304471 .1045138 celecl -41.23128 9.922178 -4.155 0.000 -61.46769 -20.99486 cons -38.1493 3.450352 -11.057 0.000 -45.18634 -31.11226 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year 10 Tech3.log 1/7/00 list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1484.595 -62.00504 1546.6 / * 1.05 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.05*co2[_n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source 1 SS df MS Number of obs = 36 + F( 4, 31) = 24.28 Model 39086.5589 4 9771.63973 Prob > F = 0.0000 Residual 12475.989 31 402.451258 R-squared = 0.7580 Adj R-squared II 0.7268 Total 51562.5479 35 1473.21565 Root MSE = 20.061 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3099216 .0388477 7.978 0.000 .2306911 .3891521 cheat .0390073 .0163391 2.387 0.023 .0056836 .0723311 ccool .0377415 .0334654 1.128 0.268 -.0305116 .1059946 celecl -41.93306 10.03579 -4.178 0.000 -62.40118 -21.46493 cons -44.22075 3.489859 -12.671 0.000 -51.33837 -37.10314 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1485.785 -68.2151 1554 /* 1.055 - * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.055*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) 11 Tech3.log 1/7/00 reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.25 Model 40051.9539 4 10012.9885 Prob > F = 0.0000 Residual 12798.1994 31 412.845142 R-squared = 0.7578 Adj R-squared = 0.7266 Total 52850.1533 35 1510.00438 Root MSE = 20.319 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3133338 .0393462 7.964 0.000 .2330867 .3935809 cheat .0394318 .0165487 2.383 0.023 .0056805 .0731831 ccool .0384495 .0338948 1.134 0.265 - -.0306794 .1075784 celecl -42.63482 10.16456 -4.194 0.000 -63.36557 -21.90407 cons -50.29219 3.534637 -14.228 0.000 -57.50113 -43.08325 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1486.975 -74.42515 1561.4 / * 1.06 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.06*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.16 Model 41029.533 4 10257.3833 Prob > F = 0.0000 Residual 13161.76 31 424.572905 R-squared = 0.7571 Adj R-squared = 0.7258 Total 54191.2931 35 1548.32266 Root MSE = 20.605 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] + 12 Tech3.log 1/7/00 dntek .316746 .0399011 7.938 0.000 .2353671 .3981249 cheat .0398562 .0167821 2.375 0.024 .0056289 .0740835 ccool .0391577 .0343728 1.139 0.263 -.0309462 .1092615 celec1 -43.33655 10.30792 -4.204 0.000 -64.35969 -22.31341 cons -56.36363 3.58449 -15.724 0.000 -63.67425 -49.05302 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1488.165 -80.63515 1568.8 / * 1.065 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.065*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 24.00 Model 42019.3269 4 10504.8317 Prob > F II 0.0000 Residual 13566.641 31 437.633579 R-squared = 0.7559 Adj R-squared = 0.7244 Total 55585.9679 35 1588.17051 Root MSE = 20.92 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3201582 .0405102 7.903 0.000 .2375371 .4027793 cheat .0402808 .0170383 2.364 0.025 .005531 .0750305 ccool .0398658 .0348975 1.142 0.262 -.0313081 .1110398 celecl -44.03834 10.46526 -4.208 0.000 -65.38239 -22.6943 cons -62.43509 3.639206 -17.156 0.000 -69.8573 -55.01289 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year 13 Tech3.log 1/7/00 list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1489.355 -86.84523 1576.2 / * 1.07 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.07*co2[_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celec1 if year < 1998 Source SS df MS Number of obs = 36 + F( 4, 31) = 23.79 Model 43021.3606 4 10755.3401 Prob > F = 0.0000 Residual 14012.8785 31 452.028338 R-squared = 0.7543 + Adj R-squared = 0.7226 Total 57034.239 35 1629.54969 Root MSE = 21.261 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3235707 .0411711 7.859 0.000 .2396018 .4075397 cheat .0407052 .0173162 2.351 0.025 .0053885 .0760218 ccool .0405739 .0354668 1.144 0.261 -.0317611 .112909 celecl -44.74012 10.63598 -4.206 0.000 -66.43235 -23.04788 cons -68.50653 3.698572 -18.522 0.000 -76.04982 -60.96325 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1490.545 -93.05525 1583.6 / * 1.075 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.075*co2[ n-1] (1 missing value generated) gen dco2 = co2 [_n] - tco2 14 Tech3.log 1/7/00 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 23.54 Model 44035.5037 4 11008.8759 Prob > F = 0.0000 Residual 14500.4754 31 467.75727 R-squared = 0.7523 + Adj R-squared = 0.7203 Total 58535.9791 35 1672.45655 Root MSE = 21.628 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3269829 .0418812 7.807 0.000 .2415655 . 4124002 cheat .0411296 .0176149 2.335 0.026 .0052038 .0770555 ccool .041282 .0360786 1.144 0.261 -.0323007 .1148648 celecl -45.44186 10.81945 -4.200 0.000 -67.50827 -23.37545 cons -74.57798 3.76237 -19.822 0.000 -82.25139 -66.90458 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1491.735 -99.26529 1591 / * 1.08 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.08*co2 [ n-1] (1 missing value generated) gen dco2 = co2 n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 23.24 Model 45061.8723 4 11265.4681 Prob > F = 0.0000 Residual 15029.3888 31 484.818994 R-squared = 0.7499 Adj R-squared = 0.7176 Total 60091.2611 35 1716.89317 Root MSE = 22.019 dco2 I Coef. Std. Err. t P>|t| [95% Conf. Interval] 15 Tech3.log 1/7/00 dntek .3303951 .0426382 7.749 0.000 .2434339 .4173563 cheat .0415541 .0179333 2.317 0.027 .0049789 .0781293 ccool .0419901 .0367307 1.143 0.262 -.0329226 .1169028 celecl -46.14363 11.015 -4.189 0.000 -68.60887 -23.67838 cons -80.64942 3.830373 -21.055 0.000 -88.46152 -72.83732 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1492.925 -105.4753 1598.4 / * 1.085 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.085*co2 [ n-1] (1 missing value generated) gen dco2 = co2 [n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source SS df MS Number of obs = 36 F( 4, 31) = 22.90 Model 46100.4464 4 11525.1116 Prob > F = 0.0000 Residual 15599.6443 31 503.214332 R-squared = 0.7472 Adj R-squared = 0.7145 Total 61700.0907 35 1762.85973 Root MSE = 22.432 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3338074 .0434396 7.684 0.000 .2452118 .4224031 cheat .0419786 .0182703 2.298 0.028 .0047159 .0792412 ccool .0426983 .037421 1.141 0.263 -.0336224 .119019 celecl -46.8454 11.22203 -4.174 0.000 -69.73288 -23.95792 cons -86.72086 3.902364 -22.223 0.000 -94.67978 -78.76194 predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) 16 Tech3.log 1/7/00 sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1494.115 -111.6853 1605.8 / * 1.09 * / drop tco2 dco2 pdco2 pco2 gen tco2 = 1.09*co2[_n-1] (1 missing value generated) gen dco2 = co2 [ n] - tco2 (1 missing value generated) reg dco2 dntek cheat ccool celecl if year < 1998 Source | SS df MS Number of obs = 36 F( 4, 31) = 22.54 Model 47151.2263 4 11787.8066 Prob > F = 0.0000 Residual 16211.2396 31 522.943213 R-squared = 0.7442 Adj R-squared = 0.7111 Total 63362.4659 35 1810.35617 Root MSE = 22.868 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] dntek .3372197 .0442829 7.615 0.000 .2469041 4275354 cheat .0424031 .0186251 2.277 0.030 .004417 .0803891 ccool .0434064 .0381475 1.138 0.264 -.034396 .1212088 celec1 -47.5472 11.4399 -4.156 0.000 -70.87902 -24.21537 cons -92.79232 3.978126 -23.326 0.000 -100.9058 -84.67888 - . predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) sort year list year co2 pco2 pdco2 tco2 if year == 1998 year co2 pco2 pdco2 tco2 39. 1998 1485.4 1495.304 -117.8954 1613.2 log close 17 Tech4.log 1/7/00 / > *** Detrend the series > / sort year gen dntek = nonhitek [ n] - 1.03148*nonhitek[_n - 1] (1 missing value generated) gen dco2 = co2[_n] - 1.007727*co2[ r - 1] (1 missing value generated) gen tntek = 1.03148*nonhitek[_n-1] ( (1 missing value generated) gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) gen cheat = htdgyear[_n] - htdgyear[_n-1] (1 missing value generated) gen ccool = cldgyear - cldgyear[_n-1] (1 missing value generated) gen cfoss = ffppcomr [n] - ffpcomr[_n-1] (1 missing value generated) gen cres = eleprrer - eleprrer [_n-1] (1 missing value generated) gen ccom = eleprcor[ - eleprcor[_n-1] (1 missing value generated) gen ccoal = ffppcoar[_n] - ffppcoar[_n-1] (1 missing value generated) . gen cgas = ffppngr[_n] - ffppngr[_n-1] (1 missing value generated) . gen coil = ffppoilr - ffppoilr[_n-1] (1 missing value generated) gen cnumref = numrefin[ n] - numrefin[_n-1] (1 missing value generated) gen crefcap = refincap[_n] - refincap[_n-1] (1 missing value generated) gen cnetsum = netsumca [ n] - netsumca[_n-1] (1 missing value generated) gen cextra = extralos [_n] - extralos[_n-1] (1 missing value generated) 1 Tech4.log 1/7/00 gen cpdctu = cpvtyund [_n] - cpvtyund[_n-1] (2 missing values generated) gen cpdcts = cpvtysur[_n] - cpvtysur[_n-1] (2 missing values generated) gen cnumele = numeleut [_n] - numeleut[_n-1] (1 missing value generated) gen propcoal = hrscoal/hrstotal gen cprpcoal = propcoal [n] - propcoal[ [_n-1] (1 missing value generated) gen cresl = cres [_n-1] (2 missing values generated) gen ccoml = ccom [_n-1] (2 missing values generated) gen cfoss1 = cfoss [_n-1] (2 missing values generated) gen ccoall = ccoal [_n-1] (2 missing values generated) gen cgas1 = cgas [_n-1] (2 missing values generated) gen coill = coil [ n-1] (2 missing values generated) / > *** Residential Electricity > * / . reg dco2 dntek cheat ccool cresl cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1998 & year > 1960 (2SLS) Source I SS df MS Number of obs = 36 F( 5, 30) = 19.01 Model 32698.8207 5 6539.76415 Prob > F = 0.0000 Residual 10116.866 30 337.228866 R-squared = 0.7637 Adj R-squared = 0.7243 Total 42815.6867 35 1223.30533 Root MSE = 18.364 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -36.72072 26.67595 -1.377 0.179 -91.20027 17.75883 dntek .2185198 .0577016 3.787 0.001 .1006774 .3363623 cheat .0305465 .0153697 1.987 0.056 -.0008426 .0619356 ccool .019039 .0319962 0.595 0.556 -.0463061 .084384 cresl -20.77218 14.37943 -1.445 0.159 -50.1389 8.594546 2 Tech4.log 1/7/00 cons I 4.730238 3.632671 1.302 0.203 -2.688665 12.14914 - predict pdco2 (2 missing values generated) gen pco2 = tco2 + pdco2 (2 missing values generated) test cres cresl ( 1) cres = 0.0 ( 2) cresl = 0.0 F( 2, 30) = 8.63 Prob > F = 0.0011 gen res = abs (co2 - pco2) (2 missing values generated) gen pres = abs (co2 - pco2) / co2 (2 missing values generated) sort year egen sres = sum (res) if year > 1961 & year < 1998 (3 missing values generated) egen spres = sum (pres) if year > 1961 & year < 1998 (3 missing values generated) replace sres = sres/36 (36 real changes made) replace spres = spres/36 (36 real changes made) list sres spres if year == 1997 sres spres 14. 12.24878 .0100937 sort year list year co2 pco2 year co2 pco2 1. 1960 787.9 2. 1961 793.7 3. 1962 826.6 832.7742 4. 1963 858.6 867.9325 5. 1964 895.2 894.7097 3 Tech4.log 1/7/00 6. 1965 934.5 951.6642 7. 1966 985.7 1009.519 8. 1967 1011.5 1010.425 9. 1968 1069.2 1056.519 10. 1969 1115.6 1118.855 11. 1970 1149 1136.014 12. 1971 1162.5 1166.568 13. 1972 1223.7 1195.436 14. 1973 1277.9 1250.613 15. 1974 1233.1 1230.401 16. 1975 1197 1194.785 17. 1976 1269.5 1228.558 18. 1977 1307.4 1285.546 19. 1978 1320.9 1355.125 20. 1979 1337.7 1337.681 21. 1980 1289.4 1299.796 22. 1981 1256 1259.905 23. 1982 1189.1 1198.699 24. 1983 1182.5 1206.558 25. 1984 1245.2 1243.121 26. 1985 1244.6 1280.168 27. 1986 1245.9 1261.914 28. 1987 1286.6 1280.426 29. 1988 1348.6 1336.241 30. 1989 1360.8 1378.077 31. 1990 1346.1 1349.238 32. 1991 1330.8 1330.994 33. 1992 1352.1 1349.149 34. 1993 1379.8 1369.035 35. 1994 1398.4 1400.692 36. 1995 1411.7 1407.763 37. 1996 1460.5 1430.094 38. 1997 1480 1478.708 39. 1998 1485.4 1477.432 / > *** Summary Stats > ******* / su dco2 dntek cheat ccool cfoss cres ccom ccoal cgas coil cnumref crefcap cne > tsum cextra cpdctu cpdcts cnumele cprpcoal if year < 1998 Variable Obs Mean Std. Dev. Min Max dco2 37 9.498726 34.52621 -76.60513 63.25078 dntek 37 -3.355278 87.84338 -231.6407 183.346 cheat 37 -4.918919 216.0386 -710 353 ccool 37 -1.351351 104.9924 -291 256 cfoss 37 .0108108 . 3211125 -1.15 .8200002 cres 37 -.1 .3391166 -.7000008 1 ccom 37 -.0945946 .3299377 -.6000004 1 ccoal 37 -.0018919 .1437946 -.13 .74 cgas 37 .0362162 .2346079 -.71 .49 coil 37 .0143243 .816056 -2.61 2.22 cnumref 37 -3.918919 10.48644 -43 14 crefcap 37 .1516216 4648567 -1.029999 1.16 4 Tech4.log 1/7/00 cnetsum 37 14.72432 9.749998 2 43.79999 cextra 37 .0113514 .0297361 -.04 .05 cpdctu 37 -.472973 3.286539 -19.91 .26 cpdcts 37 - -.3635135 3.328039 -19.95 .8099995 cnumele 37 63.97297 42.49607 -54 137 cprpcoal 37 .0010541 .0144497 - -.0297088 .0360969 sort year gen ggdp = (gdp[_n] - gdp[_n-1])/gdp[_n-1] (1 missing value generated) gen gco2 = (co2[_n] - co2[_n-1])/co2[_n-1] (1 missing value generated) reg gco2 ggdp, noconstant Source SS df MS Number of obs = 38 F( 1, 37) = 61.42 Model .02635254 1 .02635254 Prob > F = 0.0000 Residual .015876083 37 .000429083 R-squared = 0.6240 Adj R-squared = 0.6139 Total I .042228623 38 .00111128 Root MSE = .02071 gco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ggdp I .6500644 .0829499 7.837 0.000 .4819919 .8181369 log close 5 Seas Factr %chng NSA MonthNSA % SA Avg SA % Jan 97.7 1.29 1.32 Feb 96.4 -1.3 1.38 7.0 1.43 8.4 Mar 96.8 0.4 1.52 10.1 1.57 9.7 Apr 98.7 2.0 1.46 -3.9 1.48 -5.8 May 102.1 3.4 1.47 0.7 1.44 -2.7 Last Week May 1.53 4.8 1.50 1.3 Seas Factr %chng NSA MonthNSA % SA Avg SA % Jan 97.7 1.29 1.320368 Feb 96.4 -1.3 1.38 6.976744 1.431535 8.419377 Mar 96.8 0.4 1.52 10.14493 1.570248 9.689783 Apr 98.7 2.0 1.46 -3.947368 1.47923 -5.796406 May 102.1 3.4 1.47 0.684932 1.439765 -2.667946 Last Week May 1.53 4.794521 1.498531 1.304791 RFG Premiums 25 PADD 2 20 PADD 5 15 PADD 1 Cents per gallon 1Anticipated Premium 5 0 1/3/2000 1/10/2000 1/17/2000 1/24/2000 1/31/2000 2/7/2000 2/14/2000 2/21/2000 2/28/2000 3/6/2000 3/13/2000 3/20/2000 3/27/2000 4/3/2000 4/10/2000 4/17/2000 4/24/2000 5/1/2000 5/8/2000 5/15/2000 5/22/2000 -5 1441 1998 co2 = 1485.4 1999 CO2 Est + 1999 CO2 = 1496,9 Acves = - .3 = - - (-35) = 10.5 ACVCS1 = -. 1 . (-21.8) = 2.18 in 12.68 Theat = 226 . (.0285) = 6.44 A cool = - 133 . (.0171) 52.27 D Ntck =- 132. . (.214) = - 28.25 --- - -11.4 5.3 =6.1 [email protected] @ Tech2v4 Fri May 26 15:07:49 2000 Page 1 tm / / / / / / / / / / / / Statistics/Data Analysis User: Leigh L. Linden Project: CO2 Concavity 1. /******** > *** Detrend the series > ********/ 2. sort year 3. gen dntek = nonhitek [_n] - 1.03148*nonhitek[_n - 1] (1 missing value generated) 4. gen dco2 = co2[_n] - 1.007727*co2[_n - 1] (2 missing values generated) 5. gen tntek = 1.03148*nonhitek[_n-1] (1 missing value generated) 6. gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) 7. gen cheat = htdgyear[_n] - htdgyear[_n-1] (1 missing value generated) 8. gen ccool = cldgyear| - cldgyear[_n-1] (1 missing value generated) 9. gen cfoss = ffppcomr [_n] - ffppcomr[_n-1] (2 missing values generated) 10. gen cres = eleprrer[_n] - eleprrer[_n-1] (1 missing value generated) 11. gen ccom = eleprcor[_n] - eleprcor[_n-1] (2 missing values generated) 12. gen ccoal = ffppcoar[_n] - ffppcoar[_n-1] (2 missing values generated) 13. gen cgas = ffppngr[_n] - ffppngr[_n-1] (2 missing values generated) 14. gen coil = ffppoilr[_n] - ffppoilr[_n-1] (2 missing values generated) 15. gen cnumref = numrefin[ n] - numrefin[_n-1] (2 missing values generated) 16. gen crefcap = refincap [_n] - refincap[_n-1] (2 missing values generated) Tech2v4 Fri May 26 15:07:49 2000 Page 2 17. gen cnetsum = netsumca [n] - netsumca[_n-1] (2 missing values generated) 18. gen cextra = extralos[ n] - extralos[_n-1] (2 missing values generated) 19. gen cpdctu = cpvtyund[_n] - cpvtyund[_n-1] (3 missing values generated) 20. gen cpdcts = cpvtysur[_n] - cpvtysur[_n-1] (3 missing values generated) 21. gen cnumele = numeleut[_n] - numeleut[_n-1] (2 missing values generated) 22. gen propcoal = hrscoal/hrstotal (1 missing value generated) 23. gen cprpcoal = propcoal[_n] - propcoal[_n-1] (2 missing values generated) 24. gen cresl = cres[_n-1] (2 missing values generated) 25. gen ccom1 = ccom[_n-1] (2 missing values generated) 26. gen cfoss1 = cfoss[_n-1] (2 missing values generated) 27. gen ccoall = ccoal[_n-1] (2 missing values generated) 28. gen cgas1 = cgas [_n-1] (2 missing values generated) 29. gen coill = coil[_n-1] (2 missing values generated) 30. /******* > *** Residential Electricity > *******/ 31. reg dco2 dntek cheat ccool cresl cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1999 & year > 1960 Instrumental variables (2SLS) regression Source I SS df MS Number of obs = 37 F( 5, 31) = 20.65 Model I 33308.5971 5 6661.71941 Prob > F = 0.0000 Residual I 9750.18308 31 314.522035 R-squared = 0.7736 . Tech2v4 Fri May 26 15:07:49 2000 Page 3 Adj R-squared = 0.7370 Total 43058.7801 36 1196.07723 Root MSE = 17.735 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -35.02527 30.24463 -1.158 0.256 -96.70961 26.65906 cheat .0285269 .0141858 2.011 0.053 -.0004053 .0574591 ccool .017093 .0301959 0.566 0.575 -.0444919 .0786779 cresl -21.78913 15.42529 -1.413 0.168 -53.24922 9.670965 dntek .2136202 .0592512 3.605 0.001 .0927765 .3344639 cons 5.340061 3.694117 1.446 0.158 -2.19414 12.87426 32. predict ncpdco2 (option xb assumed; fitted values) (2 missing values generated) 33. gen pco2 = tco2 + ncpdco2 (2 missing values generated) 34. gen res = abs (co2 - pco2) (3 missing values generated) 35. gen pres = abs (co2 - pco2) /co2 (3 missing values generated) 36. sort year 37. egen sres = sum (res) if year > 1961 & year < 1998 (4 missing values generated) 38. egen spres = sum (pres) if year > 1961 & year < 1998 (4 missing values generated) 39. replace sres = sres/36 (36 real changes made) 40. replace spres = spres/36 (36 real changes made) 41. list sres spres if year == 1997 sres spres 3. 12.39082 .01018 . 42. sort year 43. list year co2 pco2 year co2 pco2 Tech2v4 Fri May 26 15:07:49 2000 Page 4 1. 1960 787.9 2. 1961 793.7 3. 1962 826.6 832.6566 4. 1963 858.6 867.6783 5. 1964 895.2 895.8261 6. 1965 934.5 951.2713 7. 1966 985.7 1009.09 8. 1967 1011.5 1011.834 9. 1968 1069.2 1056.323 10. 1969 1115.6 1117.83 11. 1970 1149 1137.076 12. 1971 1162.5 1169.457 13. 1972 1223.7 1196.567 14. 1973 1277.9 1252.147 15. 1974 1233.1 1230.643 16. 1975 1197 1194.24 17. 1976 1269.5 1231.07 18. 1977 1307.4 1287.222 19. 1978 1320.9 1352.597 20. 1979 1337.7 1336.859 21. 1980 1289.4 1299.277 22. 1981 1256 1260.767 23. 1982 1189.1 1198.896 24. 1983 1182.5 1207.682 25. 1984 1245.2 1242.691 26. 1985 1244.6 1280.264 27. 1986 1245.9 1262.526 28. 1987 1286.6 1279.201 29. 1988 1348.6 1335.492 30. 1989 1360.8 1379.206 31. 1990 1346.1 1351.47 32. 1991 1330.8 1328.634 33. 1992 1352.1 1346.341 34. 1993 1379.8 1372.792 35. 1994 1398.4 1401.747 36. 1995 1411.7 1407.625 37. 1996 1460.5 1429.432 38. 1997 1480 1475.549 39. 1998 1485.4 1479.12 40. 1999 1490.802 44. drop pco2 res pres sres spres 45. reg dco2 dntek cheat ccool cres1 cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1999 & year > 1960, noconstant Instrumental variables (2SLS) regression Source I SS df MS Number of obs = 37 F( 5, 32) = Model 29039.2378 5 5807.84757 Prob > F = Residual 17249.5958 32 539.04987 R-squared = Adj R-squared = Tech2v4 Fri May 26 15:07:49 2000 Page 5 Total 46288.8337 37 1251.04956 Root MSE = 23.217 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -81.1811 34.27072 -2.369 0.024 -150.9883 -11.37392 cheat .0228864 .0183833 1.245 0.222 -.014559 .0603319 ccool .0029232 .0391893 0.075 0.941 -.0769027 .0827491 cresl -4.371342 19.70403 -0.222 0.826 -44.50714 35.76445 dntek .1359288 .0683671 1.988 0.055 -.0033304 .2751879 46. predict pdco2 (option xb assumed; fitted values) (2 missing values generated) 47. gen pco2 = tco2 + pdco2 (2 missing values generated) 48. test cres cres1 ( 1) cres = 0.0 ( 2) cres1 = 0.0 F( 2, 32) = 10.30 Prob > F = 0.0004 49. gen res = abs (co2 - pco2) (3 missing values generated) 50. gen pres = abs (co2 - pco2) /co2 (3 missing values generated) 51. sort year 52. egen sres = sum (res) if year > 1961 & year < 1998 (4 missing values generated) 53. egen spres = sum (pres) if year > 1961 & year < 1998 (4 missing values generated) 54. replace sres = sres/36 (36 real changes made) 55. replace spres = spres/36 (36 real changes made) 56. list sres spres if year == 1997 sres spres 23. 17.03412 .0141688 Tech2v4 Fri May 26 15:07:49 2000 Page 6 57. sort year 58. list year co2 pco2 year co2 pco2 1. 1960 787.9 2. 1961 793.7 3. 1962 826.6 828.7922 4. 1963 858.6 879.2025 5. 1964 895.2 887.4487 6. 1965 934.5 963.8498 7. 1966 985.7 1017.475 8. 1967 1011.5 1006.969 9. 1968 1069.2 1060.961 10. 1969 1115.6 1137.046 11. 1970 1149 1145.16 12. 1971 1162.5 1157.799 13. 1972 1223.7 1185.255 14. 1973 1277.9 1245.12 15. 1974 1233.1 1189.606 16. 1975 1197 1206.03 17. 1976 1269.5 1222.884 18. 1977 1307.4 1260.669 19. 1978 1320.9 1353.203 20. 1979 1337.7 1335.421 21. 1980 1289.4 1269.496 22. 1981 1256 1254.397 23. 1982 1189.1 1201.122 24. 1983 1182.5 1203.716 25. 1984 1245.2 1238.225 26. 1985 1244.6 1271.419 27. 1986 1245.9 1264.855 28. 1987 1286.6 1282.988 29. 1988 1348.6 1328.258 30. 1989 1360.8 1379.703 31. 1990 1346.1 1360.981 32. 1991 1330.8 1339.668 33. 1992 1352.1 1343.537 34. 1993 1379.8 1371.368 35. 1994 1398.4 1398.142 36. 1995 1411.7 1415.278 37. 1996 1460.5 1429.303 38. 1997 1480 1480.993 39. 1998 1485.4 1479.374 40. 1999 1508.46 59. /**** > *** Summary Stats > / 60. su dco2 dntek cheat ccool cfoss cres ccom ccoal cgas coil cnumref crefcap cne > tsum cextra cpdctu cpdcts cnumele cprpcoal if year < 1998 Tech2v4 Fri May 26 15:07:49 2000 Page 7 Variable Obs Mean Std. Dev. Min Max dco2 37 9.498726 34.52621 -76.60513 63.25078 dntek 37 -5.090339 91.44317 -239.8102 183.579 cheat 37 -4.918919 216.0386 -710 353 ccool 37 -1.351351 104.9924 -291 256 cfoss 37 .0108108 .3211125 -1.15 .8200002 cres 37 -.1 .3391166 -.7000008 1 ccom 37 -.0945946 .3299377 -.6000004 1 ccoal 37 -.0018919 .1437946 -.13 .74 cgas 37 .0362162 .2346079 -.71 .49 coil 37 .0143243 .816056 -2.61 2.22 cnumref 37 -3.918919 10.48644 -43 14 crefcap 37 .1516216 .4648567 -1.029999 1.16 cnetsum 37 14.72432 9.749998 2 43.79999 cextra 37 .0113514 .0297361 -.04 .05 cpdctu 37 -.472973 3.286539 -19.91 .26 cpdcts 37 -.3635135 3.328039 -19.95 .8099995 cnumele 37 63.97297 42.49607 -54 137 cprpcoal 37 .0010541 .0144497 -.0297088 .0360969 61. sort year 62. gen ggdp = (gdp[_n] - gdp[_n-1])/gdp[_n-1] (1 missing value generated) 63. gen gco2 = (co2[_n] - co2[_n-1])/co2[_n-1] (2 missing values generated) 64. reg gco2 ggdp, noconstant Source SS df MS Number of obs = 38 F( 1, 37) = 65.27 Model . 026951415 1 . 026951415 Prob > F = 0.0000 Residual . 015277208 37 .000412898 R-squared = 0.6382 Adj R-squared = 0.6284 Total .042228623 38 . 00111128 Root MSE = .02032 gco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] ggdp .6561952 .0812201 8.079 0.000 . 4916277 .8207627 65. log close Stata Corporation 702 University Drive East College Station, Texas 77840 409-696-4600, fax 409-696-4601 http://www.ela.doe.gov/pub/energy.overview/monthly.energy/mert9-9 Table 9.9 Retail Prices of Electricity Sold by Electric Utilities (Cents per Kilowatthour) Residential Commercial Industrial Other Total 1973 Average 2.5 2.4 1.3 2.1 2.0 1974 Average 3.1 3.0 1.7 2.8 2.5 1975 Average 3.5 3.5 2.1 3.1 2.9 1976 Average 3.7 3.7 2.2 3.3 3.1 1977 Average 4.1 4.1 2.5 3.5 3.4 1978 Average 4.3 4.4 2.8 3.6 3.7 1979 Average 4.6 4.7 3.1 4.0 4.0 1980 Average 5.4 5.5 3.7 4.8 4.7 1981 Average 6.2 6.3 4.3 5.3 5.5 1982 Average 6.9 6.9 5.0 5.9 6.1 1983 Average 7.2 7.0 5.0 6.4 6.3 1984 Average 7.15 7.13 4.83 5.90 6.25 1985 Average 7.39 7.27 4.97 6.09 6.44 1986 Average 7.42 7.20 4.93 6.11 6.44 1987 Average 7.45 7.08 4.77 6.21 6.37 1988 Average 7.48 7.04 4.70 6.20 6.35 1989 Average 7.65 7.20 4.72 6.25 6.45 1990 Average 7.83 7.34 4.74 6.40 6.57 1991 Average 8.04 7.53 4.83 6.51 6.75 1992 Average 8.21 7.66 4.83 6.74 6.82 1993 Average 8.32 7.74 4.85 6.88 6.93 1994 Average 8.38 7.73 4.77 6.84 6.91 1995 Average 8.40 7.69 4.66 6.88 6.89 1996 Average 8.36 7.64 4.60 6.91 6.86 1997 Average 8.43 7.59 4.53 6.91 6.85 1998 January 7.87 7.22 4.36 6.37 6.57 February 7.97 7.29 4.31 6.63 6.52 March 8.01 7.28 4.33 6.72 6.53 April 8.23 7.31 4.30 6.69 6.51 May 8.49 7.45 4.41 6.69 6.67 June 8.53 7.61 4.65 6.83 6.97 July 8.58 7.69 4.85 6.84 7.21 August 8.57 7.67 4.78 6.69 7.14 September 8.43 7.55 4.62 6.56 6.95 October 8.25 7.44 4.42 6.76 6.69 November 8.04 7.11 4.32 6.11 6.39 December 7.92 7.11 4.30 6.69 6.46 Average 8.26 7.41 4.48 6.63 6.74 1999 January R 7.55 R 6.92 R 4.24 R 6.51 R 6.37 February R 7.90 R 7.12 R 4.29 R 6.39 R 6.44 March R 7.87 R 7.08 R 4.16 R 6.54 R 6.36 April R 8.07 R 7.01 R 4.21 R 6.53 R 6.34 May R 8.24 R 7.13 R 4.28 R 6.60 R 6.44 June R 8.40 R 7.33 R 4.50 R 6.63 R 6.76 July R 8.46 R 7.47 R 4.76 R 6.66 R 7.04 August R 8.39 R 7.40 R 4.84 R 6.63 R 7.02 September R 8.33 R 7.36 R 4.53 R 6.61 R 6.80 October R 8.34 R 7.33 R 4.43 R 6.66 R 6.64 November R 8.07 R 7.06 R 4.24 R 6.32 R 6.35 December R 7.91 R 6.81 R 4.17 R 6.47 R 6.34 Average R 8.14 R 7.18 R 4.40 R 6.55 R 6.60 2000 January 7.61 6.82 4.15 5.98 6.29 R=Revised. 1 of 2 5/24/2000 10:42 AM http://www.era.doe.gov/pup/energy.overview/monthly.energy/mer9-9 Notes: Prices are calculated by dividing revenue by sales. Revenue may not correspond to sales for a particular month because of electric utility billing and accounting procedures. That lack of correspondence could result in uncharacteristic increases or decreases in the monthly prices. See Note 7 at end of section. Geographic coverage is the 50 States and the District of Columbia. Sources: See end of section. 2 of 2 5/24/2000 10:42 AM Table A2. Annual U.S. Macroeconomic and Weather Indicators Year /1987 1988 1989 $1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Macroeconomic Real Gross Domestic Product (billion chained 1996 dollars 6093 6349 6569 6684 6669 6891 7054 7338 7537 7813 8165 8516 8867 9219 9514 GDP Implicit Price Deflator (Index 1996=1.000) 0.779 0.805 0.835 0.868 0.897 0.917 0.942 0.961 0.982 1.000 1.017 1.029 1.043 1.061 1.077 Real Disposable Personal Income (billion chained 1996 Dollars) 4563 4766 4885 4991 5026 5200 5254 5388 5533 15678 5885 6125 6367 6620 6880 Manufacturing Production 1.418 1.504 (Index, 1992=1.000) 0.928 0.971 0.990 0.985 0.962 1.000 1.037 1.100 1:159 1.213 1.298 1.361 1.464 Real Fixed Investment (billion.chained 996,dollars) 856 887 911) 895 833 887 958 1046 1109 1213 1316 1472 1590 1702 1780 Real Exchange Rate 1.152 1.153 1.159 1.102 (Index, 990=1.000) NA NA NA 0.999.41.007 1.013 1.057 1.034 0.961 1.017 1.104 Business Inventory Change (bilion(chained 1996 dollars) 8.4 16.9 14.2 8.9 - -6.8 1-4.7 3.6 121 14.1 10.1 22.1 25.1 0.9 9.9 12.5 Producer Price Index ((index. 1982=1:000) 1.028 1.069 1.122 163 1.165 .172 1.189 1.205 1.248 1.277 1.276 1.244 1.255 1.303 1.308 Consumer Price Index (index, 1982 1984=1'000) 1.137 1.184 1.240 1.308 1.363 1.404 1.446 1.483 1.525 1.570 *1.606 1.631 1.667 1.714 1,746 Petroleum Product PricerIndex (Index 1982=1:000) 0.568 0.539 0.612 0.748 0.671 0.647 *0.620 0.5911 0.608 0.701 0.680 0.513 0.610 0.782 0.716 Non-Farm Employment (millions) 102.0 105:2 107.9 109.4 108.3 108.6 110.7 114.14 117.2 119.6 122.7 125.8 128.6 131.2 133.0 Commercial Employment ((millions) 65.2 67.8 70.0 -70.8 71.2 73.2 76.1 78.8 81.1 83.9 86.6 89.5 91.8 94.1 Total:Ind ustrial Production (index, 1992=1/000) 0.932 0.974 0.991 0.989 0.970 1.000 1.034 1.091 1.144 1.195 1.270 1.324 1.370 415 1.451 Housing Stock (millions) 99.8 101.6% 102.9 103.5 104.5 105.5 106.8 108.2 109.6 111.0 1125 114.3 115.8 117.0 118.3 Weather a Heating Degree-Days U.S 4334 4653 4726 4016 4200 4441 4700 4483 4531 4713 4542 3951 4177 4251 4464 New England 6546 6715 6887 5848 5960 6844 6728 6672 6559 6679 6662 5680 6007 6354 6478 Middle Atlantic 5699 6088 6134 4998 5177 5964 5948 5934 5831 5986 5809 4812 5334 5555 5712 U.S Gas Weighted 1391 4804 4856 4139 1337 4458 24754 4659 4707 4980 4802 4185 4409 4429 4703 Cooling Degree Days(U.S.). 269 283 1156 1260 1331 1040 1218 1220 1293. 1180 1156 1410 1277 1240 1234 Population-weighted degree-days. A degree-day indicates the temperature variation from 65 degrees Fahrenheit (calculated as the simple average of the daily minimum and maximum temperatures) weighted by 1990 population. Notes: Historical data are printed in bold; forecasts are in italics. Sources: Historical data: latest data available from: U.S. Department of Commerce, Bureau of Economic Analysis; U.S. Department of Commerce, National Oceanic and Atmospheric Administration; Federal Reserve System, Statistical Release G.17(419); U.S. Department of Transportation; American Iron and Steel Institute. Macroeconomic projections are based on DRI/McGraw-Hill Forecast CONTROL0300. Energy Information Administration/Short-Term Energy Outlook -- May 2000 43 Thu Jun 1 13:07:02 2000 Page 1 EIAl.do tm / / / / / / / / / / / / Statistics/Data Analysis User: Leigh L. Linden Project: CO2 Concavity 1. /******** > *** Detrend the series > ********/ 2. sort year 3. gen dntek = nonhitek[_n] - 1.03148*nonhitek[_n - 1] (1 missing value generated) 4. gen dco2 = co2[_n] - 1.007727*co2[_n - 1] (1 missing value generated) 5. gen tntek = 1.03148*nonhitek[_n-1] (1 missing value generated) 6. gen tco2 = 1.007727*co2[_n-1] (1 missing value generated) 7. gen cheat = htdgyear[_n] - htdgyear[_n-1] (1 missing value generated) 8. gen ccool = cldgyear[_n] - cldgyear[_n-1] (1 missing value generated) 9. gen cfoss = ffppcomr[_n] - ffppcomr[_n-1] (2 missing values generated) 10. gen cres = eleprrer[_n] - eleprrer[_n-1] (1 missing value generated) 11. gen ccom = eleprcor[_n] - eleprcor[_n-1] (2 missing values generated) 12. gen ccoal = ffppcoar[_n] - ffppcoar[_n-1] (2 missing values generated) 13. gen cgas = ffppngr[_n] - ffppngr[_n-1] (2 missing values generated) 14. gen coil = ffppoilr[_n] - ffppoilr[_n-1] (2 missing values generated) 15. gen cnumref = numrefin[_n] - numrefin[_n-1] (2 missing values generated) 16. gen crefcap = refincap[_n] - refincap[_n-1] (2 missing values generated) Tech2v4 Thu Jun 1 13:07:02 2000 Page 2 EIAl.do 17. gen cnetsum = netsumca[n] - netsumca[_n-1] (2 missing values generated) 18. gen cextra = extralos[_n] - extralos[_n-1] (2 missing values generated) 19. gen cpdctu = cpvtyund[_n] - cpvtyund[_n-1] (3 missing values generated) 20. gen cpdcts = cpvtysur[_n] - cpvtysur[_n-1] (3 missing values generated) 21. gen cnumele = numeleut - numeleut[_n-1] (2 missing values generated) 22. gen propcoal = hrscoal/hrstotal (1 missing value generated) 23. gen cprpcoal = propcoal[_n] - propcoal[_n-1] (2 missing values generated) 24. gen cresl = cres[_n-1] (2 missing values generated) 25. gen ccom1 = ccom[_n-1] (2 missing values generated) 26. gen cfoss1 = cfoss[_n-1] (2 missing values generated) 27. gen ccoall = ccoal[_n-1] (2 missing values generated) 28. gen cgas1 = cgas[_n-1] (2 missing values generated) 29. gen coill = coil[_n-1] (2 missing values generated) 30. /******* > *** Residential Electricity > *******/ 31. reg dco2 dntek cheat ccool cres1 cres (dntek cheat ccool cresl cnetsum cnumel > e cprpcoal) if year < 1999 & year > 1960 Instrumental variables (2SLS) regression Source SS df MS Number of obs = 37 + F( 5, 31) = 20.59 Model 33281.8043 5 6656.36086 Prob > F = 0.0000 Residual 1 9733.19261 31 313.973955 R-squared = 0.7737 Tegh2v Thu Jun 1 13:07:02 2000 Page 3 EIAil.d Adj R-squared = 0.7372 Total 43014.9969 36 1194.86103 Root MSE = 17.719 dco2 Coef. Std. Err. t P>|t| [95% Conf. Interval] cres -33.81025 30.21827 -1.119 0.272 -95.44082 27.82031 cheat .027059 .0141735 1.909 0.066 -.001848 .055966 ccool .0117836 .0301695 0.391 0.699 -.0497475 .0733148 cres1 -22.31244 15.41185 -1.448 0.158 -53.74511 9.12023 dntek .2150862 .0591996 3.633 0.001 .0943479 .3358245 cons 5.675934 3.690897 1.538 0.134 -1.8517 13.20357 32. predict ncpdco2 (option xb assumed; fitted values) (2 missing values generated) 33. gen pco2 = tco2 + ncpdco2 (2 missing values generated) 34. gen res = abs (co2 - pco2) (2 missing values generated) 35. gen pres = abs (co2 - pco2) /co2 (2 missing values generated) 36. sort year 37. egen sres = sum (res) if year > 1961 & year < 1998 (4 missing values generated) 38. egen spres = sum(pres) if year > 1961 & year < 1998 (4 missing values generated) 39. replace sres = sres/36 (36 real changes made) 40. replace spres = spres/36 (36 real changes made) 41. list sres spres if year == 1997 sres spres 3. 12.41092 .0101946 42. sort year 43. list year co2 pco2 year co2 pco2 Techav4 Thu Jun 1 13:07:02 2000 Page 4 EIAI.do 1. 1960 787.9 2. 1961 793.7 3. 1962 826.6 832.6114 4. 1963 858.6 867.3605 5. 1964 895.2 896.6985 6. 1965 934.5 951.2264 7. 1966 985.7 1008.833 8. 1967 1011.5 1012.767 9. 1968 1069.2 1055.934 10. 1969 1115.6 1117.148 11. 1970 1149 1136.98 12. 1971 1162.5 1170.386 13. 1972 1223.7 1197.09 14. 1973 1277.9 1252.568 15. 1974 1233.1 1232.562 16. 1975 1197 1193.717 17. 1976 1269.5 1231.819 18. 1977 1307.4 1286.824 19. 1978 1320.9 1352.466 20. 1979 1337.7 1338.035 21. 1980 1289.4 1299.336 22. 1981 1256 1262.01 23. 1982 1189.1 1199.387 24. 1983 1182.5 1207.21 25. 1984 1245.2 1243.342 26. 1985 1244.6 1280.6 27. 1986 1245.9 1262.907 28. 1987 1286.6 1279.128 29. 1988 1348.6 1335.273 30. 1989 1360.8 1379.995 31. 1990 1349 1352.034 32. 1991 1335 1330.902 33. 1992 1359 1352.124 34. 1993 1385 1378.572 35. 1994 1410 1407.606 36. 1995 1421 1418.873 37. 1996 1471 1439.389 38. 1997 1493 1486.616 39. 1998 1495 1491.97 40. 1999 1507 1500.68 44. /**** > *** Summary Stats > *******/ 45. su dco2 dntek cheat ccool cfoss cres ccom ccoal cgas coil cnumref crefcap cne > tsum cextra cpdctu cpdcts cnumele cprpcoal if year < 1998 Variable Obs Mean Std. Dev. Min Max dco2 37 9.83951 34.45789 -76.60513 63.25078 dntek 37 -5.090339 91.44317 -239.8102 183.579 cheat 37 -4.918919 216.0386 -710 353 August 24, 1999 TO: Joe Aldy FROM: Howard Gruenspecht SUBJECT: 1998 Energy Use Per your request, I have attached some information regarding energy use during 1998 and projections of energy use for 1999. Based on some back-of-the envelope calculations, it appears that weather and production declines in energy-intensive industries might have depressed energy use by 1 to 1.3 quads during 1998. Taken together, a reduction in energy use of this magnitude would translate into a reduction of roughly 20 to 25 MMT in energy-related carbon emissions given the likely fuel composition of the reduction in energy use. The latest short-term energy outlook (STEO, August 1998) produced by the Energy Information Administration projects that energy use in 1999 will be 1.8 percent higher than the 1998 level, and that carbon emissions from energy will increase by 31 MMT over the 1998 level. This is subject to some corrections for changes in data reporting as well as the resolution of key "wildcards" over the rest of the year. Let me know if you need more. Çã HKG Draft: 8-19-99 Energy Use and Carbon Emissions for 1999: Projections from the Short Term Energy Outlook The Short Run Energy Outlook (STEO), prepared monthly by the Energy Information Administration, reflects year-to-date energy and economic data as well as the latest projections of economic and energy market conditions over the coming months. STEO is probably the single best source for an examination of how 1999 is shaping up from an energy perspective. The latest STEO (August 1999) projects that 1999 U.S. energy consumption will be 96.4 quads, a 1.8 percent increase over the 1998 level. By way of comparison, energy consumption grew by only 0.5 percent between 1997 and 1998. Carbon emissions from energy are projected to increase by 31 MMT between 1998 and 1999. Note that when EIA checked into an apparent inconsistency between energy #s and an earlier set of carbon #s they had provided, they indicated that there were potentially some number problems associated with the recent sale of power plants by utilities (who report to EIA) to independent power producers (who have taraditionally not reported, but who are now starting to report.) In the context of electricity restructuring. They indicated that the Sept 1999 STEO could change by up to 0.5 percent in emissions as they try to reconcile reporting. The big "wildcards" in short term energy forecasting are weather, oil prices, and the level of economic activity during the remainder of the year. Consider, for example, the STEO projection for 1998 energy use that was issued one year ago in August, 1998. In late 1998, oil prices dropped dramatically towards the end of the year and the economy grew faster than projected, both of which tend to increase energy consumption. However, mild winter weather at the end of 1998 cut in the other direction by sharply reducing heating demand. As a result of these offsetting factors, actual 1998 energy use exactly equalled the August 1998 forecast of 94.7 quadrillion Btus, but its composition was different - less natural gas due to a mild heating season and more oil due to a robust economy and lower oil prices. Of course, it is just happenstance that the combination energy-increasing and energy-decreasing developments in the final 4 months 1998 led to overall 1998 energy usage identical to that projected in the August 1998 STEO. If all of the big wildcards line up in one direction between now and the end of the year, actual energy consumption for 1999 could turn out to differ substantially from the August 1999 STEO projection. Like Yogi Berra said, "it ain't over until it's over." However, from information available today it appears more likely than not that energy use and emissions growth in 1999 will exceed last year's exceptionally low levels. HG/PP Draft 8-19-99 Industrial Energy Use: Explaining Recent Data Energy consumption in the industrial sector, which accounts for more than 1/3 of the nation's total energy use, decreased slightly between 1997 and 1998, a period over which GDP grew by 3.9 percent and overall industrial production grew nearly as fast. The question before us is why this happened, and whether we can count on a continued "decoupling" of industrial output growth and industrial energy use. Manufacturing account for about 55 percent of total industrial energy use. Agriculture, construction, forestry and fisheries, and mining account for the remainder. While all manufacturing uses some energy, about two-thirds of manufacturing energy use, or over one-third of total industrial energy use occurs in the following industries: chemicals, paper products, steel, and petroleum refining. Therefore, it makes sense to focus on these sectors in examining drivers of recent industrial energy use trends. Between 1997 and 1998 there was an absolute output decline in steel and paper, while chemicals production grew at a slow rate. The change in energy consumption in a particular industry depends on two main factors: (1) the change in production level, and (2) the change in energy used per unit of production. Both of these can be influenced by both short and long-run factors. The larger the role played by short-run factors, the less likely is it that the single-year trend will continue into the future. Certainly, long-run factors are important. Output growth for energy-intensive industries has been slower than the growth of overall industrial production. There has also been a steady trend of energy-efficiency improvements in most energy-intensive industries. Because of these long-run factors, which we would expect to continue, industrial energy consumption has been growing at a much slower rate than industrial production over the last decade. These long run trends could be strengthened by DOE and EPA programs designed to promote energy efficiency improvement in industry. However, our review of available information suggests that short-run factors, notably a flat or declining level of output in the steel, paper, and energy-intensive chemicals sectors between 1997 and 1998, also played a very significant role in driving down 1998 industrial energy use. Influences such as the recent economic upheavals in Russia and Asia, which significantly impacted trade patterns in energy-intensive goods, might have played a major effect. For example, if production in steel, paper, and chemicals between 1997 and 1998 had continued to grow at the average rates experienced during the 1992 to 1997 period, total industrial energy consumption would have risen by .3 percent instead of falling by .7 percent between 1997 and 1998. Assuming that this additional energy use had the same composition as average industrial energy use, GHG emissions would have been roughly 5 MMT higher in 1998. Finally, it is important to recognize that year-over-year comparisons in energy use reflect the change in production rather than its absolute level. Data for the first half of 1999 suggest no evidence of a "bounceback" in production of steel, energy-intensive chemicals, or paper. However, even if output were to remain fixed at its 1998 level in 1999, comparisons of 1999 to 1998 energy data would not show the same "kick" from these production levels as do comparisons of 1998 to 1997 data. To the extent that we cannot expect continued absolute declines in these sectors, the year-over-year decline in industrial energy use during a period of strong economic growth is probably best seen as an anomaly rather than the new expectation, at least until more data is in. Growth rates (%) 1992 1993 1994 1995 1996 1997 Avg. 1998 (year over year) '92-'97 Iron and Steel 4.2 7.1 6.1 3.5 1.3 4.2 4.4 -2.6 Paper Products 0.4 2.4 -0.7 1.9 0.1 4.0 1.4 -1.8 Chemicals 5.1 2.8 2.4 4.4 4.1 4.5 3.9 0.9 Petroleum Products 0.9 2.9 -0.2 1.8 2.2 2.8 1.7 2.0 All Industrial Output 3.1 3.5 5.4 4.9 4.5 6.0 4.6 3.7 Real GDP 2.7 2.3 3.5 2.3 3.4 3.9 3.0 3.9 Industrial energy use 3.4 0.6 2.7 0.9 2.9 0.4 1.8 -0.7 Çã HKG Draft: 8-19-99 The Impact of Weather on 1998 Energy Use Although most energy uses is are independent of weather conditions, heating and cooling - lumped together under the term space conditioning - are important exceptions. Energy used to heat and cool residential and commercial buildings accounted for nearly 15 percent of total expenditures by end-users on energy in recent years. There is significant year-to-year variation in weather, which for space conditioning purposes is most usefully measured in terms of heating and cooling degree-days. For example, over the past 50 years, annual U.S. population-weighted heating-degree days ranged from 4016 to 4958, while population weighted cooling degree-days ranged from 1029 to 1348. More "degree-days" of either type implies more energy use for space conditioning. Winter weather is more important than summer weather in driving energy use. This reflects the fact that in terms of heating and cooling expenditures, the United States is a "cold" country. In 1990, one of the warmest years on record, Americans used roughly 4 times as much primary energy and spent roughly 2.3 times as much to heat their homes and businesses as to cool them. The expenditure ratio is less than the primary energy use ratio because cooling energy, virtually 100 percent électricity, has a significantly higher cost to the customer per unit of energy content than average heating energy (a combination of natural gas, fuel oil, and electricity). Because swings in weather can have quite a large impact on energy use levels, it is important to take account of their role when interpreting annual variation in energy statistics. For example, in 1998, the U.S. economy grew by 3.9 percent, yet energy use and greenhouse gas emissions increased only slightly over 1997 levels. Some have tried to interpret this outcome as a clear indication existing Administration programs and new technologies in the energy market have successfully "decoupled" economic growth and energy use trends. While "decoupling" may or may not be taking place, it is clear that weather variation also played in determining the change in energy use between 1997 and 1998. Heating degree days were 13 percent lower in 1998 compared to 1997, while cooling degree days were 15 percent higher. Because of the dominance of heating, the overall net effect of climate changes between 1997 and 1998 was a significant reduction in space conditioning energy demand. A rough calculation suggests that if weather had remained unchanged between 1997 and 1998, primary energy consumption in 1998 would have been more than 1 percent above its actual value and carbon dioxide emissions from 1998 energy use would have been 15 to 20 MMT higher than their actual level. The Department's programs and new technologies may also be having an impact on energy consumption, but that effect is likely to be a gradually increasing one that should not show up so dramatically over a one-year time period. The risk in overinterpreting weather related swings in energy consumption as an indicator that we are successfully "decoupling" energy use and the economy is that weather will continue to fluctuate. If we try to take credit for the weather now, those who seek to question the value of our programs and technologies will use it against us when the thermometer moves in an unfavorable direction. Heating degree days during the first four months of 1999 were up 7 percent from their level in the comparable 1998 period. Data on cooling degree days for this summer are not yet available. Even though anecdotal evidence suggests a hot summer, last summer was also very hot, so the year-to-year change in cooling degree days should be relatively small The heating data that is already in hand, together with the summer data on cooling degree days and heating degree days in the first part of the upcoming winter, will together determine how 1999 ultimately shapes up relative to both 1998 and "normal" conditions. Based on what we know so far, we shouldn't count on getting the same type of "energy dividend" from weather changes that we experienced in 1998. Çã