618:
348:
819:
is the
Bartlett kernel and can be thought of as a weight that decreases with increasing separation between samples. Disturbances that are farther apart from each other are given lower weight, while those with equal subscripts are given a weight of 1. This ensures that second term converges (in some
69:
data. The abbreviation "HAC," sometimes used for the estimator, stands for "heteroskedasticity and autocorrelation consistent." There are a number of HAC estimators described in, and HAC estimator does not refer uniquely to Newey–West. One version of Newey–West
Bartlett requires the user to specify
613:{\displaystyle X^{\operatorname {T} }\Sigma X={\frac {1}{T}}\sum _{t=1}^{T}e_{t}^{2}x_{t}x_{t}^{\operatorname {T} }+{\frac {1}{T}}\sum _{\ell =1}^{L}\sum _{t=\ell +1}^{T}w_{\ell }e_{t}e_{t-\ell }(x_{t}x_{t-\ell }^{\operatorname {T} }+x_{t-\ell }x_{t}^{\operatorname {T} })}
675:
332:
115:
887:, the CovarianceMatrices.jl package supports several types of heteroskedasticity and autocorrelation consistent covariance matrix estimation including Newey–West, White, and Arellano.
790:
736:
817:
873:
155:
199:
763:
709:
259:
232:
1571:
966:
299:
279:
175:
135:
81:
1250:
334:. This means that as the time between error terms increases, the correlation between the error terms decreases. The estimator thus can be used to improve the
1520:
1494:
1470:
1446:
1285:
205:
62:
829:
986:
928:
884:
1576:
626:
304:
87:
820:
appropriate sense) to a finite matrix. This weighting scheme also ensures that the resulting covariance matrix is
821:
71:
1439:
Topics in
Advanced Econometrics: Estimation, Testing, and Specification of Cross-section and Time Series Models
891:
1508:
335:
1147:
1420:"Usage Note 40098: Newey–West correction of standard errors for heteroscedasticity and autocorrelation"
1063:
1019:"A Simple, Positive Semi-definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix"
178:
768:
714:
339:
42:
1566:
1128:
1086:
1041:
58:
53:
in 1987, although there are a number of later variants. The estimator is used to try to overcome
795:
1516:
1490:
1484:
1466:
1460:
1456:
1442:
1291:
1281:
836:
specifies the "maximum lag considered for the control of autocorrelation. A common choice for
76:
Regression models estimated with time series data often exhibit autocorrelation; that is, the
38:
34:
843:
140:
1543:
1162:
1120:
1078:
1033:
184:
46:
957:, the Newey–West corrected standard errors can be obtained in PROC AUTOREG and PROC MODEL
741:
687:
237:
210:
1398:
54:
50:
1504:
1205:
1059:
954:
284:
264:
160:
120:
1560:
1480:
202:
1309:
30:
990:
913:
produces Newey–West standard errors for coefficients estimated by OLS regression.
1064:"Heteroskedasticity and autocorrelation consistent covariance matrix estimation"
66:
1362:
950:) in the context of a time-series dataset produces Newey–West standard errors.
1341:
1185:
1166:
924:
in the
Econometrics toolbox produces the Newey–West estimator (among others).
77:
26:
1295:
1380:
1222:
1548:
1531:
1148:"Automatic positive semidefinite HAC covariance matrix and GMM estimation"
1363:"Heteroscedasticity and autocorrelation consistent covariance estimators"
1323:
1419:
234:
are "point-wise" consistent estimators of their population counterparts
1132:
1105:
1090:
1045:
1018:
157:
is the covariance matrix of the residuals. The least squares estimator
935:
module includes functions for the covariance matrix using Newey–West.
917:
1532:"Econometric Computing with HC and HAC Covariance Matrix Estimators"
1124:
1082:
1037:
1515:(Third international ed.). Harlow: Pearson. pp. 637–642.
939:
906:
1223:"time series – Bartlett Kernel (Newey West Covariance Matrix)"
342:
when the residuals are heteroskedastic and/or autocorrelated.
1489:. Princeton: Princeton University Press. pp. 408–410.
1441:. New York: Cambridge University Press. pp. 195–198.
1106:"Automatic lag selection in covariance matrix estimation"
987:"Newey West estimator – Quantitative Finance Collector"
846:
798:
771:
744:
717:
690:
629:
351:
307:
287:
267:
240:
213:
187:
163:
143:
123:
90:
137:
is the design matrix for the regression problem and
70:
the bandwidth and usage of the
Bartlett kernel from
828: = 0 reduces the Newey–West estimator to
84:of the error covariance is constructed from a term
1204:
867:
811:
784:
757:
730:
703:
669:
612:
326:
293:
273:
253:
226:
193:
169:
149:
129:
109:
902:include a function for the Newey–West estimator.
1465:. Princeton University Press. pp. 279–285.
1253:[Generalized Least Squares estimation].
670:{\displaystyle w_{\ell }=1-{\frac {\ell }{L+1}}}
65:in the models, often for regressions applied to
1186:"sandwich: Robust Covariance Matrix Estimators"
1012:
1010:
1008:
327:{\displaystyle X^{\operatorname {T} }\Sigma X}
110:{\displaystyle X^{\operatorname {T} }\Sigma X}
1251:"Verallgemeinerte Kleinst-Quadrate-Schätzung"
967:Heteroskedasticity-consistent standard errors
261:. The general approach, then, will be to use
8:
1342:"Regression with Newey–West standard errors"
1104:Newey, Whitney K.; West, Kenneth D. (1994).
1017:Newey, Whitney K; West, Kenneth D (1987).
1547:
855:
851:
845:
803:
797:
776:
770:
749:
743:
722:
716:
695:
689:
649:
634:
628:
601:
596:
580:
567:
556:
546:
527:
517:
507:
497:
480:
470:
459:
445:
436:
431:
421:
411:
406:
396:
385:
371:
356:
350:
312:
306:
286:
266:
245:
239:
218:
212:
186:
162:
142:
122:
95:
89:
946:to several estimation commands (such as
41:model where the standard assumptions of
978:
57:(also called serial correlation), and
1572:Regression with time series structure
1399:"Robust covariance matrix estimation"
7:
1245:
1243:
1180:
1178:
1176:
82:heteroscedastic consistent estimator
1324:"plm: Linear Models for Panel Data"
602:
568:
437:
362:
357:
318:
313:
144:
101:
96:
14:
1280:(7th ed.). Boston: Pearson.
45:do not apply. It was devised by
1536:Journal of Statistical Software
1310:"CovarianceMatrices.jl package"
1406:Gretl User's Guide, chapter 22
792:row of the design matrix, and
607:
539:
80:are correlated over time. The
33:to provide an estimate of the
1:
785:{\displaystyle t^{\text{th}}}
731:{\displaystyle t^{\text{th}}}
1513:Introduction to Econometrics
1437:Bierens, Herman J. (1994).
1276:Greene, William H. (2012).
1203:Greene, William H. (1997).
1593:
1146:Smith, Richard J. (2005).
1113:Review of Economic Studies
830:Huber–White standard error
301:to devise an estimator of
1381:"statsmodels: Statistics"
1167:10.1017/S0266466605050103
812:{\displaystyle w_{\ell }}
72:Kernel density estimation
879:Software implementations
201:. This implies that the
868:{\displaystyle T^{1/4}}
150:{\displaystyle \Sigma }
37:of the parameters of a
869:
822:positive semi-definite
813:
786:
759:
732:
705:
671:
614:
502:
475:
401:
336:ordinary least squares
328:
295:
275:
255:
228:
195:
194:{\displaystyle \beta }
171:
151:
131:
111:
1549:10.18637/jss.v011.i10
1060:Andrews, Donald W. K.
870:
814:
787:
760:
758:{\displaystyle x_{t}}
733:
706:
704:{\displaystyle e_{t}}
672:
615:
476:
455:
381:
329:
296:
276:
256:
254:{\displaystyle E_{i}}
229:
227:{\displaystyle e_{i}}
196:
172:
152:
132:
112:
1530:Zeileis, A. (2004).
1462:Time Series Analysis
1367:Econometrics Toolbox
1278:Econometric analysis
1207:Econometric Analysis
844:
796:
769:
742:
715:
688:
684:is the sample size,
627:
349:
305:
285:
265:
238:
211:
185:
179:consistent estimator
161:
141:
121:
88:
23:Newey–West estimator
606:
572:
441:
416:
43:regression analysis
1577:Estimation methods
1457:Hamilton, James D.
1155:Econometric Theory
865:
809:
782:
755:
728:
701:
667:
610:
592:
552:
427:
402:
324:
291:
271:
251:
224:
191:
167:
147:
127:
107:
59:heteroskedasticity
1522:978-1-4082-6433-1
1496:978-0-691-01018-2
1472:978-0-691-04289-3
1448:978-0-521-41900-0
1287:978-0-273-75356-8
1255:www.uni-kassel.de
779:
725:
665:
453:
379:
294:{\displaystyle e}
274:{\displaystyle X}
170:{\displaystyle b}
130:{\displaystyle X}
35:covariance matrix
1584:
1553:
1551:
1526:
1500:
1476:
1452:
1424:
1423:
1416:
1410:
1409:
1403:
1395:
1389:
1388:
1377:
1371:
1370:
1359:
1353:
1352:
1346:
1338:
1332:
1331:
1320:
1314:
1313:
1306:
1300:
1299:
1273:
1267:
1266:
1264:
1262:
1247:
1238:
1237:
1235:
1233:
1219:
1213:
1212:
1210:
1200:
1194:
1193:
1182:
1171:
1170:
1152:
1143:
1137:
1136:
1110:
1101:
1095:
1094:
1068:
1056:
1050:
1049:
1023:
1014:
1003:
1002:
1000:
998:
989:. Archived from
983:
949:
945:
934:
923:
912:
901:
897:
874:
872:
871:
866:
864:
863:
859:
818:
816:
815:
810:
808:
807:
791:
789:
788:
783:
781:
780:
777:
764:
762:
761:
756:
754:
753:
737:
735:
734:
729:
727:
726:
723:
710:
708:
707:
702:
700:
699:
676:
674:
673:
668:
666:
664:
650:
639:
638:
619:
617:
616:
611:
605:
600:
591:
590:
571:
566:
551:
550:
538:
537:
522:
521:
512:
511:
501:
496:
474:
469:
454:
446:
440:
435:
426:
425:
415:
410:
400:
395:
380:
372:
361:
360:
333:
331:
330:
325:
317:
316:
300:
298:
297:
292:
280:
278:
277:
272:
260:
258:
257:
252:
250:
249:
233:
231:
230:
225:
223:
222:
200:
198:
197:
192:
176:
174:
173:
168:
156:
154:
153:
148:
136:
134:
133:
128:
116:
114:
113:
108:
100:
99:
47:Whitney K. Newey
16:Statistical tool
1592:
1591:
1587:
1586:
1585:
1583:
1582:
1581:
1557:
1556:
1529:
1523:
1509:Watson, Mark M.
1505:Stock, James H.
1503:
1497:
1479:
1473:
1455:
1449:
1436:
1433:
1431:Further reading
1428:
1427:
1418:
1417:
1413:
1401:
1397:
1396:
1392:
1379:
1378:
1374:
1361:
1360:
1356:
1344:
1340:
1339:
1335:
1322:
1321:
1317:
1308:
1307:
1303:
1288:
1275:
1274:
1270:
1260:
1258:
1249:
1248:
1241:
1231:
1229:
1227:Cross Validated
1221:
1220:
1216:
1211:(3rd ed.).
1202:
1201:
1197:
1184:
1183:
1174:
1150:
1145:
1144:
1140:
1125:10.2307/2297912
1108:
1103:
1102:
1098:
1083:10.2307/2938229
1066:
1058:
1057:
1053:
1038:10.2307/1913610
1021:
1016:
1015:
1006:
996:
994:
993:on 24 June 2018
985:
984:
980:
975:
963:
947:
943:
932:
921:
910:
899:
895:
894:, the packages
881:
847:
842:
841:
799:
794:
793:
772:
767:
766:
745:
740:
739:
718:
713:
712:
691:
686:
685:
654:
630:
625:
624:
576:
542:
523:
513:
503:
417:
352:
347:
346:
308:
303:
302:
283:
282:
263:
262:
241:
236:
235:
214:
209:
208:
183:
182:
159:
158:
139:
138:
119:
118:
91:
86:
85:
55:autocorrelation
51:Kenneth D. West
39:regression-type
17:
12:
11:
5:
1590:
1588:
1580:
1579:
1574:
1569:
1559:
1558:
1555:
1554:
1527:
1521:
1501:
1495:
1481:Hayashi, Fumio
1477:
1471:
1453:
1447:
1432:
1429:
1426:
1425:
1411:
1390:
1372:
1354:
1333:
1315:
1301:
1286:
1268:
1239:
1214:
1195:
1172:
1161:(1): 158–170.
1138:
1119:(4): 631–654.
1096:
1077:(3): 817–858.
1051:
1032:(3): 703–708.
1004:
977:
976:
974:
971:
970:
969:
962:
959:
920:, the command
909:, the command
880:
877:
862:
858:
854:
850:
806:
802:
775:
752:
748:
721:
698:
694:
678:
677:
663:
660:
657:
653:
648:
645:
642:
637:
633:
621:
620:
609:
604:
599:
595:
589:
586:
583:
579:
575:
570:
565:
562:
559:
555:
549:
545:
541:
536:
533:
530:
526:
520:
516:
510:
506:
500:
495:
492:
489:
486:
483:
479:
473:
468:
465:
462:
458:
452:
449:
444:
439:
434:
430:
424:
420:
414:
409:
405:
399:
394:
391:
388:
384:
378:
375:
370:
367:
364:
359:
355:
323:
320:
315:
311:
290:
270:
248:
244:
221:
217:
190:
166:
146:
126:
106:
103:
98:
94:
15:
13:
10:
9:
6:
4:
3:
2:
1589:
1578:
1575:
1573:
1570:
1568:
1565:
1564:
1562:
1550:
1545:
1541:
1537:
1533:
1528:
1524:
1518:
1514:
1510:
1506:
1502:
1498:
1492:
1488:
1487:
1482:
1478:
1474:
1468:
1464:
1463:
1458:
1454:
1450:
1444:
1440:
1435:
1434:
1430:
1421:
1415:
1412:
1407:
1400:
1394:
1391:
1386:
1382:
1376:
1373:
1368:
1364:
1358:
1355:
1350:
1343:
1337:
1334:
1329:
1325:
1319:
1316:
1311:
1305:
1302:
1297:
1293:
1289:
1283:
1279:
1272:
1269:
1256:
1252:
1246:
1244:
1240:
1228:
1224:
1218:
1215:
1209:
1208:
1199:
1196:
1191:
1187:
1181:
1179:
1177:
1173:
1168:
1164:
1160:
1156:
1149:
1142:
1139:
1134:
1130:
1126:
1122:
1118:
1114:
1107:
1100:
1097:
1092:
1088:
1084:
1080:
1076:
1072:
1065:
1061:
1055:
1052:
1047:
1043:
1039:
1035:
1031:
1027:
1020:
1013:
1011:
1009:
1005:
992:
988:
982:
979:
972:
968:
965:
964:
960:
958:
956:
951:
942:, the option
941:
936:
930:
925:
919:
914:
908:
903:
893:
888:
886:
878:
876:
860:
856:
852:
848:
839:
835:
831:
827:
823:
804:
800:
773:
750:
746:
738:residual and
719:
696:
692:
683:
661:
658:
655:
651:
646:
643:
640:
635:
631:
623:
622:
597:
593:
587:
584:
581:
577:
573:
563:
560:
557:
553:
547:
543:
534:
531:
528:
524:
518:
514:
508:
504:
498:
493:
490:
487:
484:
481:
477:
471:
466:
463:
460:
456:
450:
447:
442:
432:
428:
422:
418:
412:
407:
403:
397:
392:
389:
386:
382:
376:
373:
368:
365:
353:
345:
344:
343:
341:
337:
321:
309:
288:
268:
246:
242:
219:
215:
207:
204:
203:least squares
188:
180:
164:
124:
104:
92:
83:
79:
74:
73:
68:
64:
60:
56:
52:
48:
44:
40:
36:
32:
28:
24:
19:
1542:(10): 1–17.
1539:
1535:
1512:
1486:Econometrics
1485:
1461:
1438:
1414:
1405:
1393:
1384:
1375:
1366:
1357:
1349:Stata Manual
1348:
1336:
1327:
1318:
1304:
1277:
1271:
1261:21 September
1259:. Retrieved
1257:. Uni-Kassel
1254:
1232:15 September
1230:. Retrieved
1226:
1217:
1206:
1198:
1189:
1158:
1154:
1141:
1116:
1112:
1099:
1074:
1071:Econometrica
1070:
1054:
1029:
1026:Econometrica
1025:
995:. Retrieved
991:the original
981:
952:
937:
926:
915:
904:
889:
882:
837:
833:
825:
681:
679:
75:
31:econometrics
22:
20:
18:
1385:statsmodels
933:statsmodels
78:error terms
67:time series
63:error terms
25:is used in
1561:Categories
973:References
340:regression
27:statistics
1567:Estimator
1296:726074601
805:ℓ
652:ℓ
647:−
636:ℓ
588:ℓ
585:−
564:ℓ
561:−
535:ℓ
532:−
509:ℓ
488:ℓ
478:∑
461:ℓ
457:∑
383:∑
363:Σ
319:Σ
206:residuals
189:β
145:Σ
102:Σ
1511:(2012).
1483:(2000).
1459:(1994).
1062:(1991).
961:See also
944:--robust
896:sandwich
117:, where
1133:2297912
1091:2938229
1046:1913610
765:is the
711:is the
61:in the
1519:
1493:
1469:
1445:
1294:
1284:
1131:
1089:
1044:
997:18 May
931:, the
929:Python
918:MATLAB
680:where
338:(OLS)
1402:(PDF)
1345:(PDF)
1151:(PDF)
1129:JSTOR
1109:(PDF)
1087:JSTOR
1067:(PDF)
1042:JSTOR
1022:(PDF)
940:Gretl
911:newey
907:Stata
885:Julia
840:" is
177:is a
1517:ISBN
1491:ISBN
1467:ISBN
1443:ISBN
1328:CRAN
1292:OCLC
1282:ISBN
1263:2023
1234:2022
1190:CRAN
999:2009
898:and
281:and
49:and
29:and
1544:doi
1163:doi
1121:doi
1079:doi
1034:doi
955:SAS
953:In
948:ols
938:In
927:In
922:hac
916:In
905:In
900:plm
890:In
883:In
181:of
1563::
1540:11
1538:.
1534:.
1507:;
1404:.
1383:.
1365:.
1347:.
1326:.
1290:.
1242:^
1225:.
1188:.
1175:^
1159:21
1157:.
1153:.
1127:.
1117:61
1115:.
1111:.
1085:.
1075:59
1073:.
1069:.
1040:.
1030:55
1028:.
1024:.
1007:^
875:.
832:.
824:.
778:th
724:th
21:A
1552:.
1546::
1525:.
1499:.
1475:.
1451:.
1422:.
1408:.
1387:.
1369:.
1351:.
1330:.
1312:.
1298:.
1265:.
1236:.
1192:.
1169:.
1165::
1135:.
1123::
1093:.
1081::
1048:.
1036::
1001:.
892:R
861:4
857:/
853:1
849:T
838:L
834:L
826:L
801:w
774:t
751:t
747:x
720:t
697:t
693:e
682:T
662:1
659:+
656:L
644:1
641:=
632:w
608:)
603:T
598:t
594:x
582:t
578:x
574:+
569:T
558:t
554:x
548:t
544:x
540:(
529:t
525:e
519:t
515:e
505:w
499:T
494:1
491:+
485:=
482:t
472:L
467:1
464:=
451:T
448:1
443:+
438:T
433:t
429:x
423:t
419:x
413:2
408:t
404:e
398:T
393:1
390:=
387:t
377:T
374:1
369:=
366:X
358:T
354:X
322:X
314:T
310:X
289:e
269:X
247:i
243:E
220:i
216:e
165:b
125:X
105:X
97:T
93:X
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.