coefplot: how to fix the scaling for different coefficients?!

[Lisa Wilson]

Hi

I am not getting a clear figure with coefplot. I think the problem is that the coefficients and confidence intervals of the variables are on different scales.

This is the code:

eststo clear
    

foreach s in lnsp {

    foreach v in  g mp rra ira {

        eststo `s'_`v', title(`s'_`v'): qui ivreg2 `s'_`v' inv if  event==1 ,robust bw(auto) small                
    }
        

    coefplot `s'_g, bylabel(gt) || `s'_mp, bylabel(mo) || ///
             `s'_rra, bylabel(rn) || ///
             `s'_ira  , bylabel(ir)  ///
             drop(_cons) bycoefs vertical rescale(100)    ///
              yli(0, lp(solid) lc(gray))   /// levels(99 95 90)
             title(event) name(`s', replace)       aspectratio(0, placement(top))                
}

It gives me this figure:




How can I fix the scaling here to have a better figure?

Data example is below

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(date lnsp_g lnsp_mp lnsp_rra lnsp_ira event)
12421 .4692802532253717 .14198951396707926 .10629503389465267 -.9952776840946314 0
12422 .08914587641991911 .10291307764690316 -.3318984640303393 .4418097664912821 0
12423 .03529091803674955 .04109996125757931 .22157609911148768 -.17088358851430918 0
12424 -.40765895173964317 .18539173387332397 -.2448045393640168 .31390102477364756 0
12425 -.7795970204235063 .5142133254770442 -.2705502213848616 1.060767574317742 0
12428 .11401619151197817 .26025975361029197 .46797481801336005 .2597474371801294 0
12429 -.19364316376082868 -.09974189028947933 .0030346989873786256 .057563911711255855 0
12430 .20776474761765673 -.04576813578812988 -.7547986524557153 .5274223146005728 0
12431 .5462456669565521 -.15046456920850204 .2533801960820954 -1.0102612299187559 0
12432 .21740829937344763 -.06713020348587406 .4007314312737229 -.08638537692535131 0
12435 -.0943130556477616 -.05023424709232671 -.1411726291595432 -.08000458637084051 0
12436 -.19579254714728211 .2547532420942886 .004020570122604816 .07535600427065425 0
12437 .011814342583973847 -.0384665223628744 -.004766881421617519 -.017749674613249056 0
12438 -.3952954577303407 .5130418929766485 -.039353952545760684 .03845383502026323 0
12439 .19865771351208128 -.17136365875475335 -.08457641784507519 -.12305880387026846 0
12442 -.184929500582121 -.15172187573763907 -.06425307037898165 -.12207302491143501 0
12443 -.0758445834010617 -.10248966717791674 .21184217753263201 -.28128758179382984 0
12444 .29312350646976615 -.027009516082973306 -.16455620223599166 .3226864723739702 0
12445 .3058560568832407 -.07280011286059818 -.25670945269801493 .8170486473570566 0
12446 -.11894466216942146 .11300240242352899 -.19101730208306744 .49295619750712305 0
12449 .32812852104849366 -.20352284793473885 .3377431878653461 .10539099889690995 0
12450 -.04325627853102798 -.640007449782459 .5350572696211113 -.23958322438683055 0
12451 .08746131982394889 -.04122253568984391 .17343929315802242 .21865536475324454 0
12452 .4434366135967496 -.5976933544288165 -.10846345884968134 -.04674236310928137 1
12453 -.2922161232353835 -1.5241609583330238 .33673559010865706 -.8456219972021835 1
12456 1.1090050486084282 .602494216605681 -.8097711552451052 -.448933729029525 1
12457 -.15259408785790818 -.06398275352762539 .47247001615794915 -.4183412285672805 1
12458 -.05906193891569611 .03847128328568061 -.026657655559586324 .36717136553900076 1
12459 -.016098457679547984 -.4588374177327772 -.3315882108207253 -.06077085160416651 0
12460 .13306199982793476 .12984045540265043 -.24874545326215625 .19019634225854176 0
12463 -.2866019641137818 .10877933212218949 .5059049606559034 -.3281604010069472 0
12464 .23539192752828786 .04470265843669807 -.09761899666576124 .31313018671568393 0
end
format %td date
[/CODE]

[Hemanshu Kumar]

Your data example doesn't include the variable inv, so I am unable to run your code.

[Lisa Wilson]

Thanks. Here is a larger dataset that includes variable inv

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(date lnsp_g lnsp_mp lnsp_rra lnsp_ira event inv)
12421    .4692802532253717   .14198951396707926   .10629503389465267   -.9952776840946314 0                       .
12422   .08914587641991911   .10291307764690316   -.3318984640303393    .4418097664912821 0                       .
12423   .03529091803674955   .04109996125757931   .22157609911148768  -.17088358851430918 0    .0001037300948808387
12424  -.40765895173964317   .18539173387332397   -.2448045393640168   .31390102477364756 0   .00039471382360643337
12425   -.7795970204235063    .5142133254770442   -.2705502213848616    1.060767574317742 0   -.0003026014198168511
12428   .11401619151197817   .26025975361029197   .46797481801336005    .2597474371801294 0   -.0007423219750262642
12429  -.19364316376082868  -.09974189028947933 .0030346989873786256  .057563911711255855 0    .0005207410454479737
12430   .20776474761765673  -.04576813578812988   -.7547986524557153    .5274223146005728 0   .00019629184686866953
12431    .5462456669565521  -.15046456920850204    .2533801960820954  -1.0102612299187559 0   .00035226007444310036
12432   .21740829937344763  -.06713020348587406    .4007314312737229  -.08638537692535131 0    .0004888344012117783
12435   -.0943130556477616  -.05023424709232671   -.1411726291595432  -.08000458637084051 0   .00020794923934610055
12436  -.19579254714728211    .2547532420942886  .004020570122604816   .07535600427065425 0   -.0012942947083397468
12437  .011814342583973847   -.0384665223628744 -.004766881421617519 -.017749674613249056 0     .000167796653747826
12438   -.3952954577303407    .5130418929766485 -.039353952545760684   .03845383502026323 0  .000017297604942863378
12439   .19865771351208128  -.17136365875475335  -.08457641784507519  -.12305880387026846 0    .0003565571417315326
12442    -.184929500582121  -.15172187573763907  -.06425307037898165  -.12207302491143501 0   -.0005298995317184904
12443   -.0758445834010617  -.10248966717791674   .21184217753263201  -.28128758179382984 0  -.00023676718411862037
12444   .29312350646976615 -.027009516082973306  -.16455620223599166    .3226864723739702 0   -.0002536925576144698
12445    .3058560568832407  -.07280011286059818  -.25670945269801493    .8170486473570566 0   .00010465381345717918
12446  -.11894466216942146   .11300240242352899  -.19101730208306744   .49295619750712305 0  -.00006915705026261012
12449   .32812852104849366  -.20352284793473885    .3377431878653461   .10539099889690995 0  -.00016776162403295855
12450  -.04325627853102798    -.640007449782459    .5350572696211113  -.23958322438683055 0  -.00028226394117332747
12451   .08746131982394889  -.04122253568984391   .17343929315802242   .21865536475324454 0    .0001400690480754924
12452    .4434366135967496   -.5976933544288165  -.10846345884968134  -.04674236310928137 1  -.00026531191296280064
12453   -.2922161232353835  -1.5241609583330238   .33673559010865706   -.8456219972021835 1     .000498699215280966
12456   1.1090050486084282     .602494216605681   -.8097711552451052    -.448933729029525 1   .00008855198424112721
12457  -.15259408785790818  -.06398275352762539   .47247001615794915   -.4183412285672805 1    -.011268856161307626
12458  -.05906193891569611   .03847128328568061 -.026657655559586324   .36717136553900076 1    .0001265887660513266
12459 -.016098457679547984   -.4588374177327772   -.3315882108207253  -.06077085160416651 0    -.002472796488412068
12460   .13306199982793476   .12984045540265043  -.24874545326215625   .19019634225854176 0   -.0005486098544475761
12463   -.2866019641137818   .10877933212218949    .5059049606559034   -.3281604010069472 0   -.0020734165851365475
12464   .23539192752828786   .04470265843669807  -.09761899666576124   .31313018671568393 0   -.0010484851440303972
12465  -.03073777296792546  -.00465283270456851   .08846867429443872  -.06256256111016079 0    .0005815619327903278
12466    .5029341166894419  -.03862185625244052   .01339682096022532  -1.0261379795099774 0    .0006689745885743988
12467  -.11932887493105149   -.1885650010588087   .40763215953040177   -.6712644220453114 0    -.006299996631268913
12470                    0                    0                    0                    0 0  .000038762830913151976
12471   .07776487527442022  -.26360630422050235   .19583792735356464    .7595633311172101 0  .000014154625097594052
12472   .29968817034057826   -.7970723613659239    .3501562831425593 -.052974236206832614 0    -.000284166411652988
12473   -.5946479361644026  -.46745392727038393    .5314446217103095   -.8362673659836393 0    -.001512915215064927
12474   .18796483599834524   -.1756006452857939  -.04764897965636372    .3759865751720728 0    .0010296576298001042
12477   -.5037018382052079  -.06644634848759523    .2913661295943868    .4616487563912859 0  .000033615991017381916
12478   .40555379589603735   -.9827135296605114   .47920867243676746   -.5304869388766935 0  -.00007777547905085708
12479    .1833508059578124   .05986719459956902   -.1482624025127155 -.036965618798893374 0    .0007523181266708543
12480   .08933180139311503   .06867519769667751   .09536117142116218   -.6426555158290395 0   .00046030868966851537
12481   .44052551908049864   -.9087019491324959    .0649748135515047    .7381841403550506 0  -.00016450976117771878
12484  -.08502053962312095   .11424894279906539  -.07451411023691357   .48137157393176366 0  .000052121468152683785
12485    .6057030275996218   -.2200596317386463  -.30139106208667776  -.36804614164038973 0  .000011934951373737468
12486  -.12471454467865017  -.24614183236063525   .34679511476535474      .30116800128331 0   -.0012123803963590039
12487    -.541525390023962  -.07657696751534004    .9257854477523324  -1.0332056893184358 0   .00040107548269711674
12488    .6960189851725427    .4650207097336505    -.771378721674143   .07395022949964414 0    .0001656555164710466
12491    .2250523761375378  -.24052502271182163    .2981137092765742  -.06675717108064916 0    .0001627977758392543
12492    .1083878100415383 -.053474810700432585    -.379090234129265    .2355923840574441 0    .0017194182002690619
12493 -.023440256007285498   .14824999874095823   -.3517364106707763     .699444313518903 0    .0007884412939661364
12494   .14541018587812315   .18545081541048702   .09237895154431364   -.1822646239528023 0     -.23059915577788692
12495   .43273096700831104   -.3434195944436276    .3877650437762554   -.5468967832374361 0     .000764871822498966
12498    .4651280885553777   -.3134647481262469   -.3318113740781232  -.31544823022089474 1    -.004283617021478302
12499   -.7478319443481514     .183764315008613   -.1980837350868523    .7983710245858902 1   .00036972463910815213
12500  .010309545058286407   .15157444467459413  -.07467517492243658  -.17867270168573413 1    .0006978327948990435
12501     .355268247865165   -.1865670012247378   .40933556160138096  -1.5199000598914196 1   -.0029557695847485434
12502  -.03509320354160028  -.05176015002228951   -.3462072245791603   -.3905017526153456 1    .0003179581680870628
12505    .4098021116665507    -.158110284545252   -.3516886737331113 -.050449730323952835 0      .05378878812445539
12506   -.3143846701893702   -.8087767345334171    .1489624762214543   -.7841751312477294 0    .0027485727306269083
12507   .15019141178002382    .5149907758299825   -.6434497739447655  -1.5593481982246367 0    -.002146782194386256
12508  -.22160221133086155   -.5195725351143423   .30295869821245436   .45084662935719955 0     -.06563280713485505
12509                    0                    0                    0                    0 0                       .
12512    .9206932141391182  -2.5096576848178813   1.8516418771755152  -1.8237680966331111 0     .008256040954670892
12513  -.24992517332605746    .6005766095829088   -.4354784042792037     2.19482352981398 0    .0006435656642090375
12514  -.22265449225883738   .23676673817957905  -.11817697476776345 -.017428667169271606 0   -.0000240578731332216
12515  -.10611324476221085  .031604804396177855    .1385809499008488    .5190651892891855 0   .00014123590722874265
12516     .637306216950573  -.19656447705186864   -.5054460334012916   -.8140372307413806 0   .00014199452416871882
12519   .14248430581263394   .26989618932177706  -.04392744098819357   .08824125666296467 0   .00013957966422859896
12520   -.7041770698130934  -.02844908449399952  -.00518149370552301   .28670814950580414 0  -.00008030167631392637
12521  -.07994709508394626  -.39033730740980843    .6255364016608329   -.5153248812734601 0  .000011078939423618028
12522   -.1062914601353886  -.44360608148058844    .4478626576402469   .08580366272468076 0    .0001830072684389613
12523   .40045591124569246  -.30149104341017363    -.644661948871784   .43489209322849565 0   .00010967633231909759
12526    .8750269804545153  -1.1762307849606053    .2529751148229895    -.836642520112723 0    .0001035345477667827
12527   .14159105063659655  -.22128707010509174   -.6456732213724337     .724431016279075 0    .0001303057001049105
12528   -.5718257148162407  -.17438005178001958   .20373033810177787   .37759044215827586 0   .00009411662839086711
12529    .1881967489847599    .3644562240335425   -.6425579099491117    1.549325331596135 0  -.00032866660708646576
12530   .22425221607099488   -.2374442557770351 .0005816569195697952  -.23438196657137045 0    .0000529381320131179
12533    .5401133297234969   -.0719786197298724   -.5402550009992275   1.1272541861295293 0  -.00011829353097620985
12534   -.4880136636727229   -.0468931208073915   .37186098578779436  .005617298378021829 0   .00011195330332092085
12535                    0                    0                    0                    0 0  2.6174817791865535e-06
12536    .9523915230869424   -.5030555792371878   .12309039177896963  -1.2641314748758195 0   .00001460231841089133
12537   -.7189116592423487  -.07051130345266898   1.1495473231957654  .012162013278211425 0  -.00010064172849282634
12540     .528066364384915  -.22681367030726562   .07670191434760731   .07896109252402272 0   .00006314486635967267
12541   .45809668929307035   -.3535213658094296  -.24203147789902069   .09889459714087234 0 -.000035714817780397826
12542 -.017944819476272467  -.27797703817859976  -.05310904555383367   .06545151024093787 0   -.0012082584055322521
12543  -.40186686267883026   -.3157484007688396   .17138456336937047   .42718029481309827 0   .00013060485745160872
12544    .5551156965455097  -1.6131471465275258   1.2168307454035574  -1.0074455168301304 0  -.00021068616657187272
12547    .3379184055527329   -1.000387170010665   .04051563637015576   -.5857145497675971 0   .00014796912388080616
12548   .07696010220833727    .5638231213573022  -1.1347063944648106   1.2977285998873969 0     -.01796917533374649
12549   -.7645162826626638   -.5359677198035264   1.2414292926372514   -1.018907603068036 0 -.000021756746488652105
12550  -.33364453898697777   .16372676994704172   .23869719698747538     .394590600032406 0     .003803226844388918
12551  -1.2535513511192091   .39724743034303067    .8650645779393482   .06510697650893693 0     -.19810729173092326
12554    .1121161870534062  -.02991900853991325   -.4703094442329512   .43255069329114537 1   -.0019904839009410205
12555  -.07579530055529449    .5568062190192382   -1.026868774621839   1.6394169016772517 1   -.0004689931588869173
12556   .21648836779592862    .8717372015670666  -.04962118844739116  -.10586985651902658 1  -.00014082420960416405
12557   -.1642737419575664    .9742895758375845  -.33175581185738245   .08455311003244503 1    .0003615204495665145
12558   .44377672711663685  -.44272149594275056  -.04033481302643516  -.30404241330396964 1    -.004619191223525621
end
format %td date

[Hemanshu Kumar]

Here is what I get using your data. The only thing I had to do to make the code run, was to

tsset date

At least in this graph, I cannot see what the issue is.

[Lisa Wilson]

You do not see the same figure as this is only a subset of my data. However, the issue is apparent when I use the full sample.

[Lisa Wilson]

Hi
I am also open to any other codes that can provide a better figure to visualize the results here ( but with the "vertical" option).
Thanks

[Ben Jann]

Hi, if the estimates have different orders of magnitude, then plotting them in the same graph might not be the best idea. If you want to do so, then you will need to apply some kind of standardization. One idea would be to apply standardization to the variables upfront (e.g. divide the variables by their standard deviations or some other values that make sense) and then estimate the models. However, you can also use option transform() in coefplot to rescale estimates while plotting them. Here is an example that standardizes the estimates by dividing them by their standard errors. Whether this makes sense in your application I don't know, but it solves the scaling problem (all confidence intervals will have same length with this approach).

tsset date
foreach v in  g mp rra ira {
    eststo lnsp_`v': qui ivreg2 lnsp_`v' inv, robust bw(auto) small
}
coefplot (lnsp_g lnsp_mp lnsp_rra lnsp_ira) ///
    , aseq swap eqrename(lnsp_g = gt lnsp_* = "") ///
      drop(_cons) yline(0) vertical ///
      transform(* = @/@se) ///
      ytitle(Coefficient divided by standard error)

Maybe a better option is to create a separate subgraph for each estimate, with individual scales. Example:

coefplot lnsp_g || lnsp_mp || lnsp_rra || lnsp_ira || ///
    , drop(_cons) bylabels(gt mo rn ir) vertical ///
      byopts(row(1) yrescale) yline(0)

An issue with this graph is that you probably want to align the scales such that zero is at the same place. This is not automatically possible, I believe. However, here is a manual solution using command addplot (type ssc install addplot), that can be applied after creating the graph to adjust the scales.

addplot 1:, ylabel(-10(2)10) xlabel(none)
addplot 2:, ylabel(-3(1)3) yscale(range(-3.2 3.2)) xlabel(none)
addplot 3:, ylabel(-4(1)4) yscale(range(-4.2 4.2)) xlabel(none)
addplot 4:, ylabel(-2(.5)2) xlabel(none)

ben