-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsur_practice_with_data.jl
735 lines (659 loc) · 25.9 KB
/
sur_practice_with_data.jl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
using LinearAlgebra, Distributions
# #Set up
# #need to calculate VCV matrix per Hamilton:
# #omegahat_0 = (1/T)SUM[y_t - X'_t*betahat_0][y_t - X'_t*betahat_0]'
# #or... can we use the residuals from OLS regressions?
# corr_true = -.45
# Sigma_true = [.5 corr_true*sqrt(.5)*sqrt(.3); corr_true*sqrt(.5)*sqrt(.3) .3]
# H_true = LinearAlgebra.cholesky(Sigma_true).U
#Generate data
function embed(x,p)
n = length(x)
pp = p+1
m = zeros(n-pp+1,pp)
for i in 1:pp
m[:,i] = x[(pp-i+1):(n-i+1)]
end
return m
end
using CSV, DataFrames, Distributions, LinearAlgebra
#Read Time Series Data
df1 = DataFrame(CSV.read("JuliaTSvars.csv"; header=["DATE", "logdiffEPRu", "logdiffUER", "logdiffRMW", "logdiffRAHE"], datarow=14))
EPR = df1[:,2]
UER = df1[:,3]
MW = df1[:,4]
AHE = df1[:,5]
#generate lags of variables
EPRlags1 = embed(EPR, 13)
UERlags1 = embed(UER,13)
MWlags1 = embed(MW,13)
AHElags1 = embed(AHE,13)
#drop contemporaneous value
#if 4 eqns, we may need 4 X matrices
EPRlags = convert(Matrix, EPRlags1[:,2:14])
UERlags = convert(Matrix, UERlags1[:,2:14])
MWlags = convert(Matrix, MWlags1[:,2:14])
AHElags = convert(Matrix, AHElags1[:,2:14])
# Xepr = [EPRlags1[:,2:14] UERlags1 AHElags1 MWlags1]
# Xuer = [EPRlags1 UERlags1[:,2:14] AHElags1 MWlags1]
# Xahe = [EPRlags1 UERlags1 AHElags1[:,2:14] MWlags1]
# Xmw = [EPRlags1 UERlags1 AHElags1 MWlags1[:,2:14]]
XSUR = [EPRlags UERlags AHElags MWlags]
yepr = EPRlags1[:,1]
yuer = UERlags1[:,1]
yahe = AHElags1[:,1]
ymw = MWlags1[:,1]
y1 = yepr
y2 = yuer
y3 = yahe
y4 = ymw
ySUR = [EPRlags1[:,1];UERlags1[:,1];AHElags1[:,1];MWlags1[:,1]]
nobs = size(y1)[1]
# #Data Generating stuff
# for i in 1:nobs
# error_vec = H_true'*(randn(2,1))
# x1[i] = rand(Normal(1,1))
# x2[i] = rand(Normal(1,1))
# y1[i] = beta11 + beta12*x1[i] + error_vec[1]
# y2[i] = beta21 + beta22*x2[i] + error_vec[2]
#
# return x1, x2, y1, y2
# end
#Define some terms that will be useful later when running the Gibbs
#sampler.
#bigy = ySUR #not used
bigX1 = [ones(size(XSUR)[1]) XSUR]
# bigX2 = bigX1 #since all X's are the same
# bigX3 = bigX1
# bigX4 = bigX1
X1X1 = bigX1'*bigX1
#no need to repeat these since all X's are the same
# X1X2 = bigX1'*bigX2
# X1X3 = bigX1'*bigX3
# X1X4 =
# X2X1 = bigX2'*bigX1
# X2X2 = bigX2'*bigX2
#recall, that later on, it may call for X2Y1 (for example) and all X's are the same
X1Y1 = bigX1'*y1
X1Y2 = bigX1'*y2
X1Y3 = bigX1'*y3
X1Y4 = bigX1'*y4
ytilde = zeros(4,nobs)
xtilde = zeros(4,212,nobs)
#ytilde: #eqns & nobs, xtilde:#eqns, total variables, nobs
for i in 1:nobs
ytilde[:,i] = [y1[i,1] y2[i,1] y3[i,1] y4[i,1]]'
xtilde[:,:,i] = [1 XSUR[i,:]' zeros(159)';
zeros(53)' 1 XSUR[i,:]' zeros(106)';
zeros(106)' 1 XSUR[i,:]' zeros(53)';
zeros(159)' 1 XSUR[i,:]']
# return ytilde, xtilde
end
#Define the prior hyperparameters;
β0 = zeros(212,1)
#change B0 and β0 according to the number of (total) variables
B0 = zeros(212,212)+I*1000
Omega = zeros(4,4)+I #R0 in Greenberg
v0 = 0 #what are appropriate values for this?
v1 = v0 + 53 #v0 + number of parameters
#Gibbs sampler
iter = 1000
burn = 200
sig1_final = zeros(iter-burn,1)
sig2_final = zeros(iter-burn,1)
sig3_final = zeros(iter-burn,1)
sig4_final = zeros(iter-burn,1)
corr12_final = zeros(iter-burn,1)
corr13_final = zeros(iter-burn,1)
corr14_final = zeros(iter-burn,1)
corr23_final = zeros(iter-burn,1)
corr24_final = zeros(iter-burn,1)
corr34_final = zeros(iter-burn,1)
betas_final = zeros(iter-burn,212)
B1_final = zeros(212,212)
#be sure to change betas_final to reflect no. of total variables
Sig1 = 1
Sig2 = 1
Sig3 = 1
Sig4 = 1
Sig12 = 0
Sig13 = 0
Sig14 = 0
Sig23 = 0
Sig24 = 0
Sig34 = 0
function sur_gibbs_ian(Sig1, Sig2, Sig3, Sig4, Sig12, Sig13, Sig14, Sig23, Sig24, Sig34, B0, β0, Omega, v0)
for i in 1:iter
#for i in 1:5 #test loop
#Sample from Conditional for Beta
#note X1X1 stands in for X1X2, X1X3, etc since all X's are the same
matrix1 = [Sig1*X1X1 Sig12*X1X1 Sig13*X1X1 Sig14*X1X1;
Sig12*X1X1 Sig2*X1X1 Sig23*X1X1 Sig24*X1X1;
Sig13*X1X1 Sig23*X1X1 Sig3*X1X1 Sig34*X1X1;
Sig14*X1X1 Sig24*X1X1 Sig34*X1X1 Sig4*X1X1]
matrix2 = [ Sig1*X1Y1 + Sig12*X1Y2 + Sig13*X1Y3 + Sig14*X1Y4;
Sig12*X1Y1 + Sig2*X1Y2 + Sig23*X1Y3 + Sig24*X1Y4;
Sig13*X1Y1 + Sig23*X1Y2 + Sig3*X1Y3 + Sig34*X1Y4;
Sig14*X1Y1 + Sig24*X1Y2 + Sig34*X1Y3 + Sig4*X1Y4]
B1 = inv(matrix1 + inv(B0))
β1 = matrix2 + inv(B0)*β0
#carryover from code I found... what does this do?
H_beta = LinearAlgebra.cholesky(Symmetric(B1)).U
betas = B1*β1 + H_beta'*randn(212,1)
#randn(212,1) = total number of variables aka, one row of Xj
#Sample from Wishart conditional for Sigmainv
tempp = zeros(4,4)
for j in 1:nobs
temp_term = (ytilde[:,j] - xtilde[:,:,j]*betas)*(ytilde[:,j] - xtilde[:,:,j]*betas)'
tempp = temp_term + tempp
end
#inv(tempp + inv(Omega)) is R1 in Greenberg
R1 = inv(tempp + inv(Omega))
R1Sym = LinearAlgebra.cholesky(Hermitian(R1)).U'*LinearAlgebra.cholesky(Hermitian(R1)).U
Sigmainv = rand(Wishart(nobs+v1, R1Sym))
Sigma = inv(Sigmainv)
Sig1 = Sigmainv[1,1]
Sig2 = Sigmainv[2,2]
Sig3 = Sigmainv[3,3]
Sig4 = Sigmainv[4,4]
Sig12 = Sigmainv[1,2]
Sig13 = Sigmainv[1,3]
Sig14 = Sigmainv[1,4]
Sig23 = Sigmainv[2,3]
Sig24 = Sigmainv[2,4]
Sig34 = Sigmainv[3,4]
s1 = Sigma[1,1]
s2 = Sigma[2,2]
s3 = Sigma[3,3]
s4 = Sigma[4,4]
s12 = Sigma[1,2]
s13 = Sigma[1,3]
s14 = Sigma[1,4]
s23 = Sigma[2,3]
s24 = Sigma[2,4]
s34 = Sigma[3,4]
#if i < burn #for testing
if i > burn
sig1_final[i-burn,1] = s1
sig2_final[i-burn,1] = s2
sig3_final[i-burn,1] = s3
sig4_final[i-burn,1] = s4
corr12_final[i-burn,1] = s12/(sqrt(s1)*sqrt(s2))
corr13_final[i-burn,1] = s13/(sqrt(s1)*sqrt(s3))
corr14_final[i-burn,1] = s14/(sqrt(s1)*sqrt(s4))
corr23_final[i-burn,1] = s23/(sqrt(s2)*sqrt(s3))
corr24_final[i-burn,1] = s24/(sqrt(s2)*sqrt(s4))
corr34_final[i-burn,1] = s34/(sqrt(s3)*sqrt(s4))
betas_final[i-burn,:] = betas'
B1_final[:,:] = B1
end
end
return sig1_final, sig2_final, sig3_final, sig4_final, corr12_final, corr13_final, corr14_final, corr23_final, corr24_final, corr34_final, betas_final, B1_final
end
output = sur_gibbs_ian(Sig1, Sig2, Sig3, Sig4,
Sig12, Sig13, Sig14, Sig23, Sig24, Sig34,
B0, β0, Omega, v0)
TSsig1 = mean(output[1])
TSsig2 = mean(output[2])
TSsig3 = mean(output[3])
TSsig4 = mean(output[4])
TSsig12 = mean(output[5])
TSsig13 = mean(output[6])
TSsig14 = mean(output[7])
TSsig23 = mean(output[8])
TSsig24 = mean(output[9])
TSsig34 = mean(output[10])
TSbeta = output[11]
TSB1 = output[12]
TSβ1 = zeros(size(TSbeta)[2],1)
for i in 1:length(TSβ1)
TSβ1[i] = mean(TSbeta[:,i])
return TSβ1
end
TSR1 = [TSsig1 TSsig12 TSsig13 TSsig14;
TSsig12 TSsig2 TSsig23 TSsig24;
TSsig13 TSsig23 TSsig3 TSsig34;
TSsig14 TSsig24 TSsig34 TSsig4]
####PANEL DATA ESTIMATION
df2 = DataFrame(CSV.read("FinalPanelLong.csv"; header=["Date", "EPRAL", "EPRAK", "EPRAZ", "EPRAR", "EPRCA", "EPRCO", "EPRCT", "EPRDE", "EPRDC", "EPRFL", "EPRGA", "EPRHI", "EPRID", "EPRIL", "EPRIN", "EPRIA",
"EPRKS", "EPRKY", "EPRLA", "EPRME", "EPRMD", "EPRMA", "EPRMI", "EPRMN", "EPRMS", "EPRMO", "EPRMT", "EPRNE", "EPRNV", "EPRNH", "EPRNJ", "EPRNM", "EPRNY", "EPRNC", "EPRND", "EPROH",
"EPROK", "EPROR", "EPRPA", "EPRRI", "EPRSC", "EPRSD", "EPRTN", "EPRTX", "EPRUT", "EPRVT", "EPRVA", "EPRWA", "EPRWV", "EPRWI", "EPRWY", "AHEAL", "AHEAK", "AHEAZ", "AHEAR", "AHECA",
"AHECO", "AHECT", "AHEDE", "AHEDC", "AHEFL", "AHEGA", "AHEHI", "AHEID", "AHEIL", "AHEIN", "AHEIA", "AHEKS", "AHEKY", "AHELA", "AHEME", "AHEMD", "AHEMA", "AHEMI", "AHEMN", "AHEMS",
"AHEMO", "AHEMT", "AHENE", "AHENV", "AHENH", "AHENJ", "AHENM", "AHENY", "AHENC", "AHEND", "AHEOH", "AHEOK", "AHEOR", "AHEPA", "AHERI", "AHESC", "AHESD", "AHETN", "AHETX", "AHEUT",
"AHEVT", "AHEVA", "AHEWA", "AHEWV", "AHEWI", "AHEWY", "EffMinWageAL", "EffMinWageAK", "EffMinWageAZ", "EffMinWageAR", "EffMinWageCA", "EffMinWageCO", "EffMinWageCT", "EffMinWageDE", "EffMinWageDC", "EffMinWageFL", "EffMinWageGA",
"EffMinWageHI", "EffMinWageID", "EffMinWageIL", "EffMinWageIN", "EffMinWageIA", "EffMinWageKS", "EffMinWageKY", "EffMinWageLA", "EffMinWageME", "EffMinWageMD", "EffMinWageMA", "EffMinWageMI", "EffMinWageMN", "EffMinWageMS", "EffMinWageMO", "EffMinWageMT",
"EffMinWageNE", "EffMinWageNV", "EffMinWageNH", "EffMinWageNJ", "EffMinWageNM", "EffMinWageNY", "EffMinWageNC", "EffMinWageND", "EffMinWageOH", "EffMinWageOK", "EffMinWageOR", "EffMinWagePA", "EffMinWageRI", "EffMinWageSC", "EffMinWageSD", "EffMinWageTN",
"EffMinWageTX", "EffMinWageUT", "EffMinWageVT", "EffMinWageVA", "EffMinWageWA", "EffMinWageWV", "EffMinWageWI", "EffMinWageWY", "UERAL", "UERAK", "UERAZ", "UERAR", "UERCA", "UERCO", "UERCT", "UERDE", "UERDC", "UERFL",
"UERGA", "UERHI", "UERID", "UERIL", "UERIN", "UERIA", "UERKS", "UERKY", "UERLA", "UERME", "UERMD", "UERMA", "UERMI", "UERMN", "UERMS", "UERMO", "UERMT", "UERNE", "UERNV", "UERNH",
"UERNJ", "UERNM", "UERNY", "UERNC", "UERND", "UEROH", "UEROK", "UEROR", "UERPA", "UERRI", "UERSC", "UERSD", "UERTN", "UERTX", "UERUT", "UERVT", "UERVA", "UERWA", "UERWV", "UERWI",
"UERWY"], datarow=2))
#create lags of each state's variables
EPRmat = convert(Matrix, df2[:,2:52])
eprALlags = embed(EPRmat[:,1],13)
eprAKlags = embed(EPRmat[:,2],13)
eprAZlags = embed(EPRmat[:,3],13)
eprARlags = embed(EPRmat[:,4],13)
eprCAlags = embed(EPRmat[:,5],13)
eprCOlags = embed(EPRmat[:,6],13)
eprCTlags = embed(EPRmat[:,7],13)
eprDElags = embed(EPRmat[:,8],13)
eprDClags = embed(EPRmat[:,9],13)
eprFLlags = embed(EPRmat[:,10],13)
eprGAlags = embed(EPRmat[:,11],13)
eprHIlags = embed(EPRmat[:,12],13)
eprIDlags = embed(EPRmat[:,13],13)
eprILlags = embed(EPRmat[:,14],13)
eprINlags = embed(EPRmat[:,15],13)
eprIAlags = embed(EPRmat[:,16],13)
eprKSlags = embed(EPRmat[:,17],13)
eprKYlags = embed(EPRmat[:,18],13)
eprLAlags = embed(EPRmat[:,19],13)
eprMElags = embed(EPRmat[:,20],13)
eprMDlags = embed(EPRmat[:,21],13)
eprMAlags = embed(EPRmat[:,22],13)
eprMIlags = embed(EPRmat[:,23],13)
eprMNlags = embed(EPRmat[:,24],13)
eprMSlags = embed(EPRmat[:,25],13)
eprMOlags = embed(EPRmat[:,26],13)
eprMTlags = embed(EPRmat[:,27],13)
eprNElags = embed(EPRmat[:,28],13)
eprNVlags = embed(EPRmat[:,29],13)
eprNHlags = embed(EPRmat[:,30],13)
eprNJlags = embed(EPRmat[:,31],13)
eprNMlags = embed(EPRmat[:,32],13)
eprNYlags = embed(EPRmat[:,33],13)
eprNClags = embed(EPRmat[:,34],13)
eprNDlags = embed(EPRmat[:,35],13)
eprOHlags = embed(EPRmat[:,36],13)
eprOKlags = embed(EPRmat[:,37],13)
eprORlags = embed(EPRmat[:,38],13)
eprPAlags = embed(EPRmat[:,39],13)
eprRIlags = embed(EPRmat[:,40],13)
eprSClags = embed(EPRmat[:,41],13)
eprSDlags = embed(EPRmat[:,42],13)
eprTNlags = embed(EPRmat[:,43],13)
eprTXlags = embed(EPRmat[:,44],13)
eprUTlags = embed(EPRmat[:,45],13)
eprVTlags = embed(EPRmat[:,46],13)
eprVAlags = embed(EPRmat[:,47],13)
eprWAlags = embed(EPRmat[:,48],13)
eprWVlags = embed(EPRmat[:,49],13)
eprWIlags = embed(EPRmat[:,50],13)
eprWYlags = embed(EPRmat[:,51],13)
EPRlaglong = [eprALlags; eprAKlags; eprAZlags; eprARlags; eprCAlags;
eprCOlags; eprCTlags; eprDElags; eprDClags; eprFLlags; eprGAlags;
eprHIlags; eprIDlags; eprILlags; eprINlags; eprIAlags; eprKSlags;
eprKYlags; eprLAlags; eprMElags; eprMDlags; eprMAlags; eprMIlags;
eprMNlags; eprMSlags; eprMOlags; eprMTlags; eprNElags; eprNVlags;
eprNHlags; eprNJlags; eprNMlags; eprNYlags; eprNClags; eprNDlags;
eprOHlags; eprOKlags; eprORlags; eprPAlags; eprRIlags; eprSClags;
eprSDlags; eprTNlags; eprTXlags; eprUTlags; eprVTlags; eprVAlags;
eprWAlags; eprWVlags; eprWIlags; eprWYlags]
yEPRpanel = EPRlaglong[:,1]
xEPRpanel = EPRlaglong[:,2:end]
AHEmat = convert(Matrix, df2[:,53:103])
aheALlags = embed(AHEmat[:,1],13)
aheAKlags = embed(AHEmat[:,2],13)
aheAZlags = embed(AHEmat[:,3],13)
aheARlags = embed(AHEmat[:,4],13)
aheCAlags = embed(AHEmat[:,5],13)
aheCOlags = embed(AHEmat[:,6],13)
aheCTlags = embed(AHEmat[:,7],13)
aheDElags = embed(AHEmat[:,8],13)
aheDClags = embed(AHEmat[:,9],13)
aheFLlags = embed(AHEmat[:,10],13)
aheGAlags = embed(AHEmat[:,11],13)
aheHIlags = embed(AHEmat[:,12],13)
aheIDlags = embed(AHEmat[:,13],13)
aheILlags = embed(AHEmat[:,14],13)
aheINlags = embed(AHEmat[:,15],13)
aheIAlags = embed(AHEmat[:,16],13)
aheKSlags = embed(AHEmat[:,17],13)
aheKYlags = embed(AHEmat[:,18],13)
aheLAlags = embed(AHEmat[:,19],13)
aheMElags = embed(AHEmat[:,20],13)
aheMDlags = embed(AHEmat[:,21],13)
aheMAlags = embed(AHEmat[:,22],13)
aheMIlags = embed(AHEmat[:,23],13)
aheMNlags = embed(AHEmat[:,24],13)
aheMSlags = embed(AHEmat[:,25],13)
aheMOlags = embed(AHEmat[:,26],13)
aheMTlags = embed(AHEmat[:,27],13)
aheNElags = embed(AHEmat[:,28],13)
aheNVlags = embed(AHEmat[:,29],13)
aheNHlags = embed(AHEmat[:,30],13)
aheNJlags = embed(AHEmat[:,31],13)
aheNMlags = embed(AHEmat[:,32],13)
aheNYlags = embed(AHEmat[:,33],13)
aheNClags = embed(AHEmat[:,34],13)
aheNDlags = embed(AHEmat[:,35],13)
aheOHlags = embed(AHEmat[:,36],13)
aheOKlags = embed(AHEmat[:,37],13)
aheORlags = embed(AHEmat[:,38],13)
ahePAlags = embed(AHEmat[:,39],13)
aheRIlags = embed(AHEmat[:,40],13)
aheSClags = embed(AHEmat[:,41],13)
aheSDlags = embed(AHEmat[:,42],13)
aheTNlags = embed(AHEmat[:,43],13)
aheTXlags = embed(AHEmat[:,44],13)
aheUTlags = embed(AHEmat[:,45],13)
aheVTlags = embed(AHEmat[:,46],13)
aheVAlags = embed(AHEmat[:,47],13)
aheWAlags = embed(AHEmat[:,48],13)
aheWVlags = embed(AHEmat[:,49],13)
aheWIlags = embed(AHEmat[:,50],13)
aheWYlags = embed(AHEmat[:,51],13)
AHElaglong = [aheALlags; aheAKlags; aheAZlags; aheARlags; aheCAlags;
aheCOlags; aheCTlags; aheDElags; aheDClags; aheFLlags; aheGAlags;
aheHIlags; aheIDlags; aheILlags; aheINlags; aheIAlags; aheKSlags;
aheKYlags; aheLAlags; aheMElags; aheMDlags; aheMAlags; aheMIlags;
aheMNlags; aheMSlags; aheMOlags; aheMTlags; aheNElags; aheNVlags;
aheNHlags; aheNJlags; aheNMlags; aheNYlags; aheNClags; aheNDlags;
aheOHlags; aheOKlags; aheORlags; ahePAlags; aheRIlags; aheSClags;
aheSDlags; aheTNlags; aheTXlags; aheUTlags; aheVTlags; aheVAlags;
aheWAlags; aheWVlags; aheWIlags; aheWYlags]
yAHEpanel = AHElaglong[:,1]
xAHEpanel = AHElaglong[:,2:end]
MWmat = convert(Matrix, df2[:,104:154])
mwALlags = embed(MWmat[:,1],13)
mwAKlags = embed(MWmat[:,2],13)
mwAZlags = embed(MWmat[:,3],13)
mwARlags = embed(MWmat[:,4],13)
mwCAlags = embed(MWmat[:,5],13)
mwCOlags = embed(MWmat[:,6],13)
mwCTlags = embed(MWmat[:,7],13)
mwDElags = embed(MWmat[:,8],13)
mwDClags = embed(MWmat[:,9],13)
mwFLlags = embed(MWmat[:,10],13)
mwGAlags = embed(MWmat[:,11],13)
mwHIlags = embed(MWmat[:,12],13)
mwIDlags = embed(MWmat[:,13],13)
mwILlags = embed(MWmat[:,14],13)
mwINlags = embed(MWmat[:,15],13)
mwIAlags = embed(MWmat[:,16],13)
mwKSlags = embed(MWmat[:,17],13)
mwKYlags = embed(MWmat[:,18],13)
mwLAlags = embed(MWmat[:,19],13)
mwMElags = embed(MWmat[:,20],13)
mwMDlags = embed(MWmat[:,21],13)
mwMAlags = embed(MWmat[:,22],13)
mwMIlags = embed(MWmat[:,23],13)
mwMNlags = embed(MWmat[:,24],13)
mwMSlags = embed(MWmat[:,25],13)
mwMOlags = embed(MWmat[:,26],13)
mwMTlags = embed(MWmat[:,27],13)
mwNElags = embed(MWmat[:,28],13)
mwNVlags = embed(MWmat[:,29],13)
mwNHlags = embed(MWmat[:,30],13)
mwNJlags = embed(MWmat[:,31],13)
mwNMlags = embed(MWmat[:,32],13)
mwNYlags = embed(MWmat[:,33],13)
mwNClags = embed(MWmat[:,34],13)
mwNDlags = embed(MWmat[:,35],13)
mwOHlags = embed(MWmat[:,36],13)
mwOKlags = embed(MWmat[:,37],13)
mwORlags = embed(MWmat[:,38],13)
mwPAlags = embed(MWmat[:,39],13)
mwRIlags = embed(MWmat[:,40],13)
mwSClags = embed(MWmat[:,41],13)
mwSDlags = embed(MWmat[:,42],13)
mwTNlags = embed(MWmat[:,43],13)
mwTXlags = embed(MWmat[:,44],13)
mwUTlags = embed(MWmat[:,45],13)
mwVTlags = embed(MWmat[:,46],13)
mwVAlags = embed(MWmat[:,47],13)
mwWAlags = embed(MWmat[:,48],13)
mwWVlags = embed(MWmat[:,49],13)
mwWIlags = embed(MWmat[:,50],13)
mwWYlags = embed(MWmat[:,51],13)
MWlaglong =[mwALlags; mwAKlags; mwAZlags; mwARlags; mwCAlags;
mwCOlags; mwCTlags; mwDElags; mwDClags; mwFLlags; mwGAlags;
mwHIlags; mwIDlags; mwILlags; mwINlags; mwIAlags; mwKSlags;
mwKYlags; mwLAlags; mwMElags; mwMDlags; mwMAlags; mwMIlags;
mwMNlags; mwMSlags; mwMOlags; mwMTlags; mwNElags; mwNVlags;
mwNHlags; mwNJlags; mwNMlags; mwNYlags; mwNClags; mwNDlags;
mwOHlags; mwOKlags; mwORlags; mwPAlags; mwRIlags; mwSClags;
mwSDlags; mwTNlags; mwTXlags; mwUTlags; mwVTlags; mwVAlags;
mwWAlags; mwWVlags; mwWIlags; mwWYlags]
yMWpanel = MWlaglong[:,1]
xMWpanel = MWlaglong[:,2:end]
UERmat = convert(Matrix, df2[:,155:205])
uerALlags = embed(UERmat[:,1],13)
uerAKlags = embed(UERmat[:,2],13)
uerAZlags = embed(UERmat[:,3],13)
uerARlags = embed(UERmat[:,4],13)
uerCAlags = embed(UERmat[:,5],13)
uerCOlags = embed(UERmat[:,6],13)
uerCTlags = embed(UERmat[:,7],13)
uerDElags = embed(UERmat[:,8],13)
uerDClags = embed(UERmat[:,9],13)
uerFLlags = embed(UERmat[:,10],13)
uerGAlags = embed(UERmat[:,11],13)
uerHIlags = embed(UERmat[:,12],13)
uerIDlags = embed(UERmat[:,13],13)
uerILlags = embed(UERmat[:,14],13)
uerINlags = embed(UERmat[:,15],13)
uerIAlags = embed(UERmat[:,16],13)
uerKSlags = embed(UERmat[:,17],13)
uerKYlags = embed(UERmat[:,18],13)
uerLAlags = embed(UERmat[:,19],13)
uerMElags = embed(UERmat[:,20],13)
uerMDlags = embed(UERmat[:,21],13)
uerMAlags = embed(UERmat[:,22],13)
uerMIlags = embed(UERmat[:,23],13)
uerMNlags = embed(UERmat[:,24],13)
uerMSlags = embed(UERmat[:,25],13)
uerMOlags = embed(UERmat[:,26],13)
uerMTlags = embed(UERmat[:,27],13)
uerNElags = embed(UERmat[:,28],13)
uerNVlags = embed(UERmat[:,29],13)
uerNHlags = embed(UERmat[:,30],13)
uerNJlags = embed(UERmat[:,31],13)
uerNMlags = embed(UERmat[:,32],13)
uerNYlags = embed(UERmat[:,33],13)
uerNClags = embed(UERmat[:,34],13)
uerNDlags = embed(UERmat[:,35],13)
uerOHlags = embed(UERmat[:,36],13)
uerOKlags = embed(UERmat[:,37],13)
uerORlags = embed(UERmat[:,38],13)
uerPAlags = embed(UERmat[:,39],13)
uerRIlags = embed(UERmat[:,40],13)
uerSClags = embed(UERmat[:,41],13)
uerSDlags = embed(UERmat[:,42],13)
uerTNlags = embed(UERmat[:,43],13)
uerTXlags = embed(UERmat[:,44],13)
uerUTlags = embed(UERmat[:,45],13)
uerVTlags = embed(UERmat[:,46],13)
uerVAlags = embed(UERmat[:,47],13)
uerWAlags = embed(UERmat[:,48],13)
uerWVlags = embed(UERmat[:,49],13)
uerWIlags = embed(UERmat[:,50],13)
uerWYlags = embed(UERmat[:,51],13)
UERlaglong = [uerALlags; uerAKlags; uerAZlags; uerARlags; uerCAlags;
uerCOlags; uerCTlags; uerDElags; uerDClags; uerFLlags; uerGAlags;
uerHIlags; uerIDlags; uerILlags; uerINlags; uerIAlags; uerKSlags;
uerKYlags; uerLAlags; uerMElags; uerMDlags; uerMAlags; uerMIlags;
uerMNlags; uerMSlags; uerMOlags; uerMTlags; uerNElags; uerNVlags;
uerNHlags; uerNJlags; uerNMlags; uerNYlags; uerNClags; uerNDlags;
uerOHlags; uerOKlags; uerORlags; uerPAlags; uerRIlags; uerSClags;
uerSDlags; uerTNlags; uerTXlags; uerUTlags; uerVTlags; uerVAlags;
uerWAlags; uerWVlags; uerWIlags; uerWYlags]
yUERpanel = UERlaglong[:,1]
xUERpanel = UERlaglong[:,2:end]
py1 = yEPRpanel
py2 = yUERpanel
py3 = yAHEpanel
py4 = yMWpanel
panY = [yEPRpanel;yUERpanel;yAHEpanel;yMWpanel]
panX1 = [xEPRpanel xUERpanel xAHEpanel xMWpanel]
#number of variables (aka K) will be different if we include indicators
#indicator variables
dummymat = zeros(51,51)+I
dummyvec = ones(161) #length of one "wave"
stateindicators = kron(dummymat, dummyvec)
panX = [panX1 stateindicators]
# testpanelmodel = [EPRlaglong AHElaglong MWlaglong UERlaglong stateindicators]
# panY = testpanelmodel[:,1]
# panX = testpanelmodel[:,2:107]
#not working yet
# #export for use in other programs
# statapany = DataFrame(panY)
# CSV.write("statapanY.csv", panY)
# CSV.write("statapanX.csv", panX)
pnobs = 8211
#Define some terms that will be useful later when running the Gibbs
#sampler.
pbigX1 = [ones(size(panX)[1]) panX]
pX1X1 = pbigX1'*pbigX1
#no need to repeat since all X's are the same
#recall, that later on, it may call for X2Y1 (for example) and all X's are the same
pX1Y1 = pbigX1'*py1
pX1Y2 = pbigX1'*py2
pX1Y3 = pbigX1'*py3
pX1Y4 = pbigX1'*py4
pytilde = zeros(4,pnobs)
pxtilde = zeros(4,416,pnobs)
#ytilde: #eqns & nobs, xtilde:#eqns, total variables, nobs
for i in 1:pnobs
pytilde[:,i] = [py1[i,1] py2[i,1] py3[i,1] py4[i,1]]'
pxtilde[:,:,i] = [1 panX[i,:]' zeros(312)';
zeros(104)' 1 panX[i,:]' zeros(208)';
zeros(208)' 1 panX[i,:]' zeros(104)';
zeros(312)' 1 panX[i,:]']
return pytilde, pxtilde
end
#Give the panel data TS priors
pβ0 = [TSβ1; zeros(204)]
#change B0 and β0 according to the number of (total) variables
pB0dg1 = diag(TSB1)
pB0dg2 = ones(204)*100
pB0dg3 = [pB0dg1; pB0dg2]
pB0 = Diagonal(pB0dg3) + zeros(416,416)
pR0 = TSR1
pv0 = 0 #what are appropriate values for this?
pv1 = v0 + 104 #v0 + number of parameters
#Gibbs sampler
piter = 1000
pburn = 200
psig1_final = zeros(iter-burn,1)
psig2_final = zeros(iter-burn,1)
psig3_final = zeros(iter-burn,1)
psig4_final = zeros(iter-burn,1)
pcorr12_final = zeros(iter-burn,1)
pcorr13_final = zeros(iter-burn,1)
pcorr14_final = zeros(iter-burn,1)
pcorr23_final = zeros(iter-burn,1)
pcorr24_final = zeros(iter-burn,1)
pcorr34_final = zeros(iter-burn,1)
pbetas_final = zeros(iter-burn,416)
pB1_final = zeros(416,416)
#be sure to change betas_final to reflect no. of total variables
pSig1 = TSsig1
pSig2 = TSsig2
pSig3 = TSsig3
pSig4 = TSsig4
pSig12 = TSsig12
pSig13 = TSsig13
pSig14 = TSsig14
pSig23 = TSsig23
pSig24 = TSsig24
pSig34 = TSsig34
function psur_gibbs_ian(Sig1, Sig2, Sig3, Sig4, Sig12, Sig13, Sig14, Sig23, Sig24, Sig34, B0, β0, R0, v0)
for i in 1:iter
#for i in 1:5 #test loop
#Sample from Conditional for Beta
#note X1X1 stands in for X1X2, X1X3, etc since all X's are the same
matrix1 = [Sig1*pX1X1 Sig12*pX1X1 Sig13*pX1X1 Sig14*pX1X1;
Sig12*pX1X1 Sig2*pX1X1 Sig23*pX1X1 Sig24*pX1X1;
Sig13*pX1X1 Sig23*pX1X1 Sig3*pX1X1 Sig34*pX1X1;
Sig14*pX1X1 Sig24*pX1X1 Sig34*pX1X1 Sig4*pX1X1]
matrix2 = [ Sig1*pX1Y1 + Sig12*pX1Y2 + Sig13*pX1Y3 + Sig14*pX1Y4;
Sig12*pX1Y1 + Sig2*pX1Y2 + Sig23*pX1Y3 + Sig24*pX1Y4;
Sig13*pX1Y1 + Sig23*pX1Y2 + Sig3*pX1Y3 + Sig34*pX1Y4;
Sig14*pX1Y1 + Sig24*pX1Y2 + Sig34*pX1Y3 + Sig4*pX1Y4]
B1 = inv(matrix1 + inv(B0))
β1 = matrix2 + inv(B0)*β0
B1herm = Symmetric(Hermitian(B1))+I*100 #adding to diagonal to make it sufficiently large for PSD
#take the proceeding B1 and B1herm out and uncomment out the preceding B1herm
#B1 = Matrix(Diagonal(B1)) #we did this to look at the diagonal elements when offdiags to zero
#B1herm = Symmetric(Hermitian(B1))
B1Sym = LinearAlgebra.cholesky(B1herm).U'*LinearAlgebra.cholesky(B1herm).U
#carryover from code I found... what does this do?
H_beta = LinearAlgebra.cholesky(Symmetric(B1Sym)).U
betas = B1*β1 + H_beta'*randn(416,1)
#randn(416,1) = total number of variables aka, one row of Xj
#Sample from Wishart conditional for Sigmainv
tempp = zeros(4,4)
for j in 1:pnobs
temp_term = (pytilde[:,j] - pxtilde[:,:,j]*betas)*(pytilde[:,j] - pxtilde[:,:,j]*betas)'
tempp = temp_term + tempp
end
#inv(tempp + inv(R0)) is R1 in Greenberg
R1 = inv(tempp + inv(R0))
R1Sym = LinearAlgebra.cholesky(Hermitian(R1)).U'*LinearAlgebra.cholesky(Hermitian(R1)).U
Sigmainv = rand(Wishart(nobs+v1, R1Sym))
Sigma = inv(Sigmainv)
Sig1 = Sigmainv[1,1]
Sig2 = Sigmainv[2,2]
Sig3 = Sigmainv[3,3]
Sig4 = Sigmainv[4,4]
Sig12 = Sigmainv[1,2]
Sig13 = Sigmainv[1,3]
Sig14 = Sigmainv[1,4]
Sig23 = Sigmainv[2,3]
Sig24 = Sigmainv[2,4]
Sig34 = Sigmainv[3,4]
s1 = Sigma[1,1]
s2 = Sigma[2,2]
s3 = Sigma[3,3]
s4 = Sigma[4,4]
s12 = Sigma[1,2]
s13 = Sigma[1,3]
s14 = Sigma[1,4]
s23 = Sigma[2,3]
s24 = Sigma[2,4]
s34 = Sigma[3,4]
#if i < burn #for testing
if i > burn
psig1_final[i-burn,1] = s1
psig2_final[i-burn,1] = s2
psig3_final[i-burn,1] = s3
psig4_final[i-burn,1] = s4
pcorr12_final[i-burn,1] = s12/(sqrt(s1)*sqrt(s2))
pcorr13_final[i-burn,1] = s13/(sqrt(s1)*sqrt(s3))
pcorr14_final[i-burn,1] = s14/(sqrt(s1)*sqrt(s4))
pcorr23_final[i-burn,1] = s23/(sqrt(s2)*sqrt(s3))
pcorr24_final[i-burn,1] = s24/(sqrt(s2)*sqrt(s4))
pcorr34_final[i-burn,1] = s34/(sqrt(s3)*sqrt(s4))
pbetas_final[i-burn,:] = betas'
pB1_final[:,:] = B1
end
end
return psig1_final, psig2_final, psig3_final, psig4_final, pcorr12_final, pcorr13_final, pcorr14_final, pcorr23_final, pcorr24_final, pcorr34_final, pbetas_final, pB1_final
end
poutput = psur_gibbs_ian(pSig1, pSig2, pSig3, pSig4,
pSig12, pSig13, pSig14, pSig23, pSig24, pSig34,
pB0, pβ0, pR0, pv0)
psig1 = mean(poutput[1])
psig2 = mean(poutput[2])
psig3 = mean(poutput[3])
psig4 = mean(poutput[4])
psig12 = mean(poutput[5])
psig13 = mean(poutput[6])
psig14 = mean(poutput[7])
psig23 = mean(poutput[8])
psig24 = mean(poutput[9])
psig34 = mean(poutput[10])
pβ1 = poutput[11]
pB1 = poutput[12]
TSβ1 = zeros(size(TSbeta)[2],1)
for i in 1:length(TSβ1)
TSβ1[i] = mean(TSbeta[:,i])
return TSβ1
end
TSR1 = [TSsig1 TSsig12 TSsig13 TSsig14;
TSsig12 TSsig2 TSsig23 TSsig24;
TSsig13 TSsig23 TSsig3 TSsig34;
TSsig14 TSsig24 TSsig34 TSsig4]
#what do we do about PSD-ness? off-diagonal problem?
#