[Nick Cox]

OK, that's probably an answerable question if we could see the data. But as far as I can see you are showing us nothing about those observations that might explain and I don't have any guesses.

We just need to be able to reproduce the problem: clearly you aren't expected to post the entire dataset. bank_id 2 would presumably be enough,

[Ciaran OFlynn]

Hi Nick,

Below is ten lines before the cutoff ofbank_id 1 and all of bank_id 2.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float mdate byte bank_id double fedfunds float(date2 fedfunds2)
680 1  .4000000059604645 680   .4
681 1  .4000000059604645 681   .4
682 1                .45 682   .4
683 1  .4099999964237213 683  .41
684 1  .5400000214576721 684  .54
685 1                 .7 685  .65
686 1  .6600000262260437 686  .66
687 1  .7900000214576721 687  .79
688 1                .95 688   .9
689 1  .9100000262260437 689  .91
690 1 1.0399999618530273 690 1.04
691 1 1.1533333333333333 691 1.15
692 1  1.159999966621399 692 1.16
693 1  1.149999976158142 693 1.15
694 1 1.2033333333333334 694 1.15
517 2               1.25   .    .
518 2 1.2999999523162842   .    .
519 2                  .   .    .
520 2 1.2466666666666666   .    .
521 2                  .   .    .
522 2                  .   .    .
523 2 1.0166666666666666   .    .
524 2                  .   .    .
525 2                  .   .    .
526 2  .9966666666666667   .    .
527 2                  .   .    .
528 2                  .   .    .
529 2 1.0033333333333334   .    .
530 2                  .   .    .
531 2                  .   .    .
532 2               1.01   .    .
533 2                  .   .    .
534 2                  .   .    .
535 2 1.4333333333333333   .    .
536 2                  .   .    .
537 2                  .   .    .
538 2               1.95   .    .
539 2                  .   .    .
540 2                  .   .    .
541 2               2.47   .    .
542 2                  .   .    .
543 2                  .   .    .
544 2 2.9433333333333334   .    .
545 2                  .   .    .
546 2                  .   .    .
547 2               3.46   .    .
548 2                  .   .    .
549 2                  .   .    .
550 2               3.98   .    .
551 2                  .   .    .
552 2                  .   .    .
553 2  4.456666666666667   .    .
554 2                  .   .    .
555 2                  .   .    .
556 2  4.906666666666666   .    .
557 2                  .   .    .
558 2                  .   .    .
559 2  5.246666666666667   .    .
560 2                  .   .    .
561 2                  .   .    .
562 2  5.246666666666667   .    .
563 2                  .   .    .
564 2                  .   .    .
565 2  5.256666666666667   .    .
566 2                  .   .    .
567 2                  .   .    .
568 2               5.25   .    .
569 2                  .   .    .
570 2                  .   .    .
571 2  5.073333333333333   .    .
572 2                  .   .    .
573 2                  .   .    .
574 2  4.496666666666667   .    .
575 2                  .   .    .
576 2                  .   .    .
577 2 3.1766666666666667   .    .
578 2                  .   .    .
579 2                  .   .    .
580 2  2.086666666666667   .    .
581 2                  .   .    .
582 2                  .   .    .
583 2               1.94   .    .
584 2                  .   .    .
585 2                  .   .    .
586 2  .5066666666666667   .    .
587 2                  .   .    .
588 2                  .   .    .
589 2 .18333333333333332   .    .
590 2                  .   .    .
591 2                  .   .    .
592 2                .18   .    .
593 2                  .   .    .
594 2                  .   .    .
595 2 .15666666666666668   .    .
596 2                  .   .    .
597 2                  .   .    .
598 2                .12   .    .
599 2                  .   .    .
600 2                  .   .    .
601 2 .13333333333333333   .    .
602 2                  .   .    .
603 2                  .   .    .
604 2 .19333333333333333   .    .
605 2                  .   .    .
606 2                  .   .    .
607 2 .18666666666666668   .    .
608 2                  .   .    .
609 2                  .   .    .
610 2 .18666666666666668   .    .
611 2                  .   .    .
612 2                  .   .    .
613 2 .15666666666666668   .    .
614 2                  .   .    .
615 2                  .   .    .
616 2 .09333333333333334   .    .
617 2                  .   .    .
618 2                  .   .    .
619 2 .08333333333333333   .    .
620 2                  .   .    .
621 2                  .   .    .
622 2 .07333333333333333   .    .
623 2                  .   .    .
624 2                  .   .    .
625 2 .10333333333333333   .    .
626 2                  .   .    .
627 2                  .   .    .
628 2 .15333333333333332   .    .
629 2                  .   .    .
630 2                  .   .    .
631 2 .14333333333333334   .    .
632 2                  .   .    .
633 2                  .   .    .
634 2                .16   .    .
635 2                  .   .    .
636 2                  .   .    .
637 2 .14333333333333334   .    .
638 2                  .   .    .
end
format %tm mdate
format %tm date2

[Nick Cox]

Thanks for the data example.

There are no observations with a non-missing value for fedfunds2 that could be used to replace a missing value of fedfunds, so it's not a surprise that nothing was changed. Positively put, all the observations with non-missing fedfunds2 are observations that also have non-missing fedfunds -- they are at the top of the listing I don't know whether the explanation lies in the original data or in what you did earlier.

I can tell you that an incorrect xtset setting has no side-effects for mipolate (SSC), which makes no use of any such settings.

[Ciaran OFlynn]

Hi Nick,

Perhaps I could rephrase. Now that I have interpolated my data, and in doing so trebled the size of my datset, how do I go about adding new variables to my dataset? If I have new complete monthly dataset I want to add to my panel, Is there a function that will allow me to fill this down, by panel identifier and date?

[Nick Cox]

That sounds like a merge or an interpolation or both.

[Ciaran OFlynn]

If all the data points are known, and they just need to be copied down in a cascade, interpolation would be unnecessary - no?

I tried merging, but stata returned the error "variables ... do not uniquely identify observations in the master data; r459". I assume this is caused by attempting to merge time-series data into a panel.