Questions on r(198)

[Guest]

Dear all,

I am trying to use the Newey-West Standard Errors to carry out my regression, however, the following popped up:

. newey  csad c.abs_r_mt##c.abs_r_mt, lag(1)
date is regularly spaced, but does not have intervals of 1
r(198);

Below is my -dataex-:

* Example generated by -dataex-. To install: ssc install dataex
clear
input int date float(csad r_mt year abs_r_mt)
15347         .          . 2002          .
15354  4.150988  -8.779794 2002   8.779794
15361   8.47286 -11.722405 2002  11.722405
15368  2.371604   .7411363 2002   .7411363
15375  4.922579  12.524622 2002  12.524622
15382  1.898078 -.10467237 2002  .10467237
15389         0          0 2002          0
15396  1.229306  1.9515274 2002  1.9515274
15403  2.176993 -.24670455 2002  .24670455
15410  5.491458  11.407248 2002  11.407248
15417  3.275471 -2.9843094 2002  2.9843094
15424  2.618778   2.722045 2002   2.722045
15431  3.072469  -3.339813 2002   3.339813
15438    2.6599  1.5941483 2002  1.5941483
15445    2.7624   .7426223 2002   .7426223
15452  2.954676 -1.6689825 2002  1.6689825
15459  3.800143   2.567357 2002   2.567357
15466 1.5627458   .2256331 2002   .2256331
15473 2.2163022 -2.0939267 2002  2.0939267
15480 2.4541366  -4.799683 2002   4.799683
15487 3.1900074   1.408269 2002   1.408269
15494 3.8380544  -4.929328 2002   4.929328
15501 2.5736775  2.7708454 2002  2.7708454
15508 1.8126394  -1.549638 2002   1.549638
15515  4.617593   14.62258 2002   14.62258
15522 3.7301376 -2.0095348 2002  2.0095348
15529 2.0474093  1.3527323 2002  1.3527323
15536 2.1388013  -2.633168 2002   2.633168
15543 1.6265107  -.1318021 2002   .1318021
15550  1.585621 -1.1615623 2002  1.1615623
15557 2.0878482  -.6251392 2002   .6251392
15564 2.0168262  -1.122699 2002   1.122699
15571 1.4484754   .5845769 2002   .5845769
15578 1.7805477  1.5606302 2002  1.5606302
15585  1.525958  .02931488 2002  .02931488
15592 2.3180652  -2.799415 2002   2.799415
15599 1.6707025 -1.6376716 2002  1.6376716
15606 2.5378354 .026403705 2002 .026403705
15613 1.6017348 -1.8123453 2002  1.8123453
15620         0          0 2002          0
15627 2.2233071  -3.020985 2002   3.020985
15634  2.758704  -.4821299 2002   .4821299
15641 2.2093017 -1.8539804 2002  1.8539804
15648 2.1681485  2.1972647 2002  2.1972647
15655 1.6749827  -2.788559 2002   2.788559
15662  3.588643  -4.780155 2002   4.780155
15669  3.346243  -4.567672 2002   4.567672
15676 1.8651558  .03562433 2002  .03562433
15683 2.2610478   .4898008 2002   .4898008
15690 1.6952885  1.0478189 2002  1.0478189
15697  1.901793  1.8837024 2002  1.8837024
15704  3.395599  -4.671672 2002   4.671672
15711  2.772347 -1.2277025 2003  1.2277025
15718 2.1912358   3.483433 2003   3.483433
15725  3.687262   5.716919 2003   5.716919
15732 2.4519944   .3525168 2003   .3525168
15739  .9451538  .14199395 2003  .14199395
15746 1.2781053  -.9503823 2003   .9503823
15753 1.6940317   1.431836 2003   1.431836
15760 2.2230365 -1.1358991 2003  1.1358991
15767 1.9164674  2.3592293 2003  2.3592293
15774 2.4346666  -3.939136 2003   3.939136
15781 2.6690204  .04388403 2003  .04388403
15788  1.442735 .012047686 2003 .012047686
15795 1.8391362   1.584176 2003   1.584176
15802 2.3038082  -.3085499 2003   .3085499
15809  4.650578    4.98953 2003    4.98953
15816  2.918732  -1.697974 2003   1.697974
15823 4.6705866   -2.76868 2003    2.76868
15830  3.494127 -1.6048017 2003  1.6048017
15837  2.977946 -.01338571 2003  .01338571
15844 3.3964925   2.671273 2003   2.671273
15851  3.643645    1.46425 2003    1.46425
15858  4.041341   1.270017 2003   1.270017
15865  2.878992  -4.592948 2003   4.592948
15872 2.8937554  1.9754747 2003  1.9754747
15879 1.9556812 -3.3694615 2003  3.3694615
15886 3.5537255 -2.2777672 2003  2.2777672
15893  2.549982   .6448685 2003   .6448685
15900 2.0797458   1.192847 2003   1.192847
15907  3.549946 -2.6754115 2003  2.6754115
15914 2.1864173 -1.5634494 2003  1.5634494
15921 1.9790592   2.047979 2003   2.047979
15928  1.656871  -.9543859 2003   .9543859
15935  1.760976 -1.5156193 2003  1.5156193
15942  2.576479  -1.520167 2003   1.520167
15949 2.3478339  1.9608448 2003  1.9608448
15956 2.1813831  -1.797851 2003   1.797851
15963 2.5223765  -2.324635 2003   2.324635
15970 2.0894804 -.58425045 2003  .58425045
15977 2.3127463  -2.599389 2003   2.599389
15984  .8075221   .8873354 2003   .8873354
15991  1.284165    2.47579 2003    2.47579
15998 1.6533177   -3.03647 2003    3.03647
16005  3.750155  -.2284442 2003   .2284442
16012  4.875941 -1.5789025 2003  1.5789025
16019  4.833535  -2.918827 2003   2.918827
16026 2.1364899  -.8974869 2003   .8974869
16033 3.2972984     4.7895 2003     4.7895
16040  2.786816   1.797363 2003   1.797363
end
format %td date

I then tried to do it with other lags specification, unfortunately, the same problem persists:

. newey  csad c.abs_r_mt##c.abs_r_mt, lag(2)
date is regularly spaced, but does not have intervals of 1
r(198);

.

I have tried searching the error phrases both on Google and Statalist, I have come across NOTHING at all, does it mean that my case is extremely rare? Most importantly, what actually went wrong?

[Andrew Musau]

If the trading days are consecutive and there are no holes, you may consider the first date as time1, the second as time2, and so on. Therefore

sort date
gen time=_n
tsset time
newey  csad c.abs_r_mt##c.abs_r_mt, lag(1)

If these are random returns taken at different points in time, then you should think a little bit more about what you are trying to do.

[William Lisowski]

You have daily dates, but just a single weekly observation for the Monday in each week. Your tsset for this data was not specified appropriately.

Notice the difference in what tsset reports with the command you probably used, and with the correct version for your sample data.

. tsset date
        time variable:  date, 07jan2002 to 01dec2003, but with gaps
                delta:  1 day

. tsset date, delta(7 days)
        time variable:  date, 07jan2002 to 01dec2003
                delta:  7 days

In general, gaps in time series data are Not a Good Thing.

[Guest]

Dear William and Musau,

Thanks for your inputs. I have, as suggested by William, changed my tsset specification. Another question arises, and you will see it below:

* Example generated by -dataex-. To install: ssc install dataex
clear
input int date float(csad r_mt abs_r_mt)
15341         .          .         .
15372  8.436224 -12.511042 12.511042
15400 3.3225675  1.9648436 1.9648436
15431  7.411424   9.583619  9.583619
15461  5.605823  3.4607785 3.4607785
15492  5.201907  -10.41467  10.41467
15522  6.498603   13.83425  13.83425
15553 4.1513767 -3.1880274 3.1880274
15584   2.72628  1.0409108 1.0409108
15614 4.0416083  -6.223028  6.223028
15645  5.389867 -4.6561255 4.6561255
15675  6.631547 -10.604467 10.604467
15706 4.3471274 -2.0230613 2.0230613
15737  6.163964   9.239872  9.239872
15765  3.007635   1.704784  1.704784
15796  3.893607  -1.587202  1.587202
15826  12.98457 -2.1023018 2.1023018
15857  5.347927   5.392154  5.392154
15887  4.101147  -8.716791  8.716791
15918  5.460991 -1.1276236 1.1276236
15949  4.401002   -.802781   .802781
15979  4.695461   -6.41879   6.41879
16010  8.431628 -2.3680263 2.3680263
16040  4.248314    2.77055   2.77055
16071  7.948152 -.08662705 .08662705
16102  5.976416   8.824282  8.824282
16131  7.156741  8.5455265 8.5455265
16162  6.598279   2.931958  2.931958
16192  6.662524 -12.454858 12.454858
16223  4.943614 -.13449205 .13449205
16253  5.744555 -11.035892 11.035892
16284  7.770133  -4.800909  4.800909
16315  5.599969  -5.720576  5.720576
16345  6.091397   7.662899  7.662899
16376  8.607979   -6.84691   6.84691
16406  6.134023    2.50735   2.50735
16437  7.697844     -7.467     7.467
16468  8.073707   -7.12877   7.12877
16496 4.2788343  10.215192 10.215192
16527  6.864724  -6.697318  6.697318
16557 10.891788  -8.308406  8.308406
16588  8.302193 -10.174953 10.174953
16618  6.594253  -1.306629  1.306629
16649  8.303922  1.3506366 1.3506366
16680   11.2914  11.181755 11.181755
16710  6.459751 -1.1536884 1.1536884
16741  7.689352   -5.73104   5.73104
16771  6.338261   .6175821  .6175821
16802 4.3295684   4.632648  4.632648
16833 9.1836815   8.995499  8.995499
16861  6.077281   3.366835  3.366835
16892  9.024069  3.6101246 3.6101246
16922 13.108974   7.246772  7.246772
16953 12.284817   19.91872  19.91872
16983  8.959021   2.630195  2.630195
17014  9.179778  -9.174119  9.174119
17045  7.507405   3.316923  3.316923
17075  5.662868   5.209397  5.209397
17106  6.947441  2.1305964 2.1305964
17136 10.264326  10.832777 10.832777
17167  9.773865   9.396377  9.396377
17198  9.617068  18.595917 18.595917
17226  11.13708   8.833343  8.833343
17257  9.992418  16.287004 16.287004
17287  12.05036   23.37832  23.37832
17318 15.995577   5.864699  5.864699
17348  19.03857 -4.5951257 4.5951257
17379  10.58591   13.97101  13.97101
17410 12.910087    18.6199   18.6199
17440  9.278922   2.939992  2.939992
17471 11.715993  -7.272669  7.272669
17501  10.56437   -13.4348   13.4348
17532  5.599703   15.03479  15.03479
17563  9.197095 -12.709573 12.709573
17592  6.502603  10.337528 10.337528
17623  11.54764 -32.929096 32.929096
17653  9.982803   7.562576  7.562576
17684  7.970956  -5.452303  5.452303
17714 11.655103  -29.46325  29.46325
17745  9.313706   7.531732  7.531732
17776 10.240597  -27.24382  27.24382
17806  9.038309  -3.427655  3.427655
17837 13.412267  -27.49006  27.49006
17867  8.933171  17.450539 17.450539
17898  7.807279  -.3152318  .3152318
17929  7.190331  11.987215 11.987215
17957  8.212899   7.716379  7.716379
17988   8.45233  17.186045 17.186045
18018  7.430564   3.718705  3.718705
18049  6.803858    8.14052   8.14052
18079   8.82491   8.066118  8.066118
18110  9.652463   15.51348  15.51348
18141  7.239003 -23.743225 23.743225
18171   8.52014   5.252904  5.252904
18202  6.789934  13.509326 13.509326
18232 12.144785  11.025484 11.025484
18263  5.653536   -.949526   .949526
18294  7.025649  -8.584202  8.584202
18322  5.034777   7.229745  7.229745
18353  8.442362   3.608491  3.608491
18383  9.804841   -9.97802   9.97802
18414  7.413955  -9.025985  9.025985
18444  6.932698  -8.916621  8.916621
18475  5.629972  16.368925 16.368925
18506  6.990912   4.834938  4.834938
18536  6.541029   1.402888  1.402888
18567  10.63143  13.805274 13.805274
18597  8.259201   -2.79766   2.79766
18628  5.767099 -1.0636408 1.0636408
end
format %td date

I have a monthly data now, and I have specified it as

tsset date, delta(28 days)
        time variable:  date, 01jan2002 to 01jan2011, but with gaps
                delta:  28 days

,

I do not see where did the gaps come from. If Stata was referring to the first row that had dashes in it, it is due to the nature of the regression specification- we subtract stock prices at time t-1 from stock prices at time t, i.e. today's price minus yesterday's price, to find the price difference, hence, we would end up having 1 less observation. Why, to Stata at least, is that classified as a gap, which William deem highly unfavourable?

[Clyde Schechter]

You don't have regular observations at exactly 28 day intervals here. In fact only 7 of your observations are precisely 28 days after the preceding one. To see this:

sort date
gen gap = date - date[_n-1]
tab gap

If your data are supposed to be at regular 28 day intervals, then something has gone wrong in the creation of this data set, and that needs to be revisited and fixed. If your expectation that it would have regular 28 day intervals is wrong, then you may need to rethink your plans for this data. It will be suitable for some purposes, but not for any that require equally spaced observations.

[Nick Cox]

A few seconds' work looking at your dates shows that they are monthly dates given as the first day of each month. That explains simply why they aren't a fixed number of days apart.

clear
input int date 
15341  
15372  
15400 
15431 
15461 
15492 
15522 
end 

format date %td

list, sep(0)

     +-----------+
     |      date |
     |-----------|
  1. | 01jan2002 |
  2. | 01feb2002 |
  3. | 01mar2002 |
  4. | 01apr2002 |
  5. | 01may2002 |
  6. | 01jun2002 |
  7. | 01jul2002 |
     +-----------+

The remedy is to work with equivalent monthly dates with something like

gen mdate = mofd(date) 
format mdate %tm
tsset mdate

It's important to be familiar with your data in detail before you get into any attempt at modelling.

[William Lisowski]

Let me add that the problem posed in post #4 is essentially the same as the problem posed earlier at

https://www.statalist.org/forums/for...-in-newey-west

which apparently you did not see my answer to. I will repeat the advice I gave there.

Stata's "datetime" variables are complicated and there is a lot to learn. As you learned in this topic, there is a difference between "monthly" dates and "daily dates chosen one per month." There is much else for you to learn.

As you apparently have not already read the very detailed Chapter 24 (Working with dates and times) of the Stata User's Guide PDF, do so now. Everyone should do so before attempting to work with Stata dates and times.

After that, the help datetime documentation will usually be enough to point the way. You can't remember everything; even the most experienced users end up referring to the help datetime documentation or back to the manual for details. But at least you will get a good understanding of the basics and the underlying principles. An investment of time that will be amply repaid.

All Stata manuals are included as PDFs in the Stata installation (since version 11) and are accessible from within Stata - for example, through the PDF Documentation section of Stata's Help menu.

[Guest]

Dear Clyde, Nick and William,

Thanks for your (very) valuable inputs, I have dealt with it successfully.