Calculating Stock beta

[Tim Nivard]

Hi all,
I just started using STATA, so I don't have much knowledge about it.
My problem is the following:
I have downloaded monthly stock return data from CRSP for 5 years (2012-2016) and have to calculate an individual beta for each stock, as a control variable in estimating the ex-ante cost of equity. This is estimated as the slope coefficient obtained by regressing the returns on the market return (S&P return) for the 36 months prior to the measurement of the ex-ante cost of equity.
So for a particular stock and a particular month, I want to regress the returns from the last 3 years on the market returns.
For example stock Apple: beta in end December 2014 would be a regression of Apple's returns from January 2012 till December 2014 on the market returns during this period.
This way, i want to have a beta for December 2014 and 2015 for every firm, so i am able to merge them to another stata file.
I have to include another restriction: firms with missing data have to be excluded, so if firms do not have the 36 months stock returns needed, i need to drop them.
I hope someone can answer these tons of questions, I would be very grateful!
Thanks in advance,
Tim

[Clyde Schechter]

It would have been easier to respond to this had you shown an example of your data. The code depends on how your data is laid out and what your variable names are, and other details. So what I'll do is describe an imaginary data set and show you the code that would work with that.

Variables:

mdate, a Stata-internal-format numeric monthly variable. (If you don't know what this means, read -help datetime-, which has things you will need to know to make good use of Stata going forward.)

stock_id, an identifying variable for the stock. It might be string or some numeric code--doesn't matter for present purposes.

stock_return, a numeric variable indicating the return of the given stock in the given month

market_return, a numeric variable indicating the return of the S&P in the given month.

rangestat (reg) stock_return market_return, by(stock) interval(mdate -35 -1)

will do these regressions.

Now, I'm not sure what you mean when you say "if firms do not have the 36 months stock returns needed, i need to drop them." Do you actually want to drop them from the data set entirely? Or do you just want to not use the regression results for those stocks? I'm going to assume the latter.

gen byte full_data = !missing(stock_return, market_return)
rangestat (sum) full_data, by(firm) interval(mdate -35 -1)
foreach v of varlist b_market_return se_market_return b_cons se_cons {
    replace `v' = . if full_data_sum < 36
}

Added: I forgot to mention that -rangestat- is not part of official Stata, but it is one of the most useful user-written programs around. Robert Picard, Nick Cox, and Roberto Ferrer wrote it. You can get it by running -ssc install rangestat-, and you can read full instructions for its many uses with -help rangestat-.

[Tim Nivard]

Hi Clyde,

Thank you for your quick response and help! Unfortunately, my stata does not seem to support rangestat. When I enter the command ssc install rangestat, stata gives me the following message:
. ssc install rangestat
connection timed out -- see help r(2) for troubleshooting
http://fmwww.bc.edu/repec/bocode/r/ either
1) is not a valid URL, or
2) could not be contacted, or
3) is not a Stata download site (has no stata.toc file).

When I open the troubleshooting file, it says it may have to do with my internet connection, but I have tried to install it on several locations, so it seems unlikely that that is the problem. Is there a way to solve this, or is there a way to do this without using rangestat?

As for your other questions, you are right about everything actually. Those are the variables I use, plus an extra identifier to eventually merge the database with Compustat. I do not want to drop the firms from the data set entirely, because it does not matter if it cannot calculate the 36 months prior to 2012 for example, i care about the 36 months prior to 2015 and 2016. I've attached a screenshot of my database, so you have an idea what it looks like. Is there a better way to show you the database?
I really hope you, or anyone else, can help me with this issue

[Nick Cox]

help netio

for help on internet connection problems. One guess is that your access to the internet involves a proxy server: if so, you must tell Stata about it. The settings are probably accessible via your default browser.

[Tim Nivard]

Thank you for your help Nick, unfortunately I'm not able to see the name of the proxy and therefore I cannot tell Stata to use it. Is there another way to solve my problem without having to install rangestat?

[Clyde Schechter]

-rangestat- is a really useful command, and this problem is by no means the only time you are likely to want it. If I were you, I would get your firm's IT on the case so that you're able to get this installed. Believe me, if you do things like this problem on any regular basis, you will deeply regret not having -rangestat-. (Especially if you use large data sets: not only is -rangestat- easier, it speeds up the execution by one or more orders of magnitude!)

But, yes, we were able to do these things before -rangestat- was invented. Here's how we used to do it:

xtset stock mdate
tempfile betas
rolling beta = _b[market_return], window(36) reject(e(N) < 36) saving(`betas'): ///
    regress  stock_return market_return
merge 1:1 stock mdate using `betas'

Note: not tested, beware of typos.

[Nick Cox]

Unless you are behind some massive firewall you should be able to work out how to download from SSC. Ask around people at your workplace to find out what proxy server is being used.

[Dev Vikas]

-rangestat- is a really useful command, and this problem is by no means the only time you are likely to want it. If I were you, I would get your firm's IT on the case so that you're able to get this installed. Believe me, if you do things like this problem on any regular basis, you will deeply regret not having -rangestat-. (Especially if you use large data sets: not only is -rangestat- easier, it speeds up the execution by one or more orders of magnitude!)

But, yes, we were able to do these things before -rangestat- was invented. Here's how we used to do it:

xtset stock mdate
tempfile betas
rolling beta = _b[market_return], window(36) reject(e(N) < 36) saving(`betas'): ///
regress stock_return market_return
merge 1:1 stock mdate using `betas'

Note: not tested, beware of typos.

Hi Clyde,
I have a panel data of 3000 firms with daily stock return observation from 1991 to 2012. For each firm for each year-end date, I need to calculate a beta with 400 previous days observation, based on market model. R=α + β *Rm + ε. (one beta for each firm for each year).
My data looks as below in stata. Can you suggest me a way-out?
id Date return(R) market_ret(Rm)
4 02jan1991 -.018088 -.010877
4 03jan1991 -.007895 -.013472
4 04jan1991 -.02122 -.002926
4 07jan1991 -.03794 -.017533
4 08jan1991 .025352 -.001694
4 09jan1991 .002747 -.011094
4 10jan1991 .019178 .01058
4 11jan1991 -.005376 .001159.....

Thank you.
Vikas

[Clyde Schechter]

Just change the -rangestat- command shown in #2 by substituting your variable names for the ones there, and set the window to 400 days instead of 36 months. If your IT set up doesn't permit you to get -rangestat-, then do that for the code you show in your own quotation of my earlier post.

By the way, please read the FAQ, especially #12. There you will find excellent advice on the best way to post example data, which is by using the -dataex- command. You should use -dataex- whenever you post example data, as it is the only easily usable way to provide those who want to help you with an exact and complete replica of your Stata example. Be sure to always use it in the future. Showing list-output or, what appears in your case to be copy-paste from a spreadsheet (not even a Stata data set!) is not helpful.

[Dev Vikas]

Thank you Clyde.
I will definitely follow now onward as suggested.
Vikas

[Gabriel Davidson]

Hi all,
I am working with a CRSP dataset containing daily stock data for every NYSE listed company from 2010 to 2019. I would need to calculate a daily rolling beta with an observation period of 365 days for each company.
I have tried the above but unfortunately I am getting an error message that the variable "firm" has not been found. If I substitute "cusip" for "firm" I am getting very weird results.
I have attatched a screenshot of what my data looks like for reference. Would be great If someone could help me out ! Any support is appreciated.

[Clyde Schechter]

At least on my setup, there is no screenshot visible. In any case, all new Forum members are asked to read the Forum FAQ. They give excellent advice about the most effective ways to share information here. You will learn there, among many other helpful things, that screenshots are usually not helpful even when they do show up. The useful way to show example data is with the -dataex- command. If you are running version 17, 16 or a fully updated version 15.1 or 14.2, -dataex- is already part of your official Stata installation. If not, run -ssc install dataex- to get it. Either way, run -help dataex- to read the simple instructions for using it. -dataex- will save you time; it is easier and quicker than typing out tables. It includes complete information about aspects of the data that are often critical to answering your question but cannot be seen from tabular displays or screenshots. It also makes it possible for those who want to help you to create a faithful representation of your example to try out their code, which in turn makes it more likely that their answer will actually work in your data.

In addition to including example data (with -dataex-) when you post back, you should show what you mean by "very weird results." Do that by copying them from your Results window or log file to your computer clipboard, and then pasting them here between code delimiters so they will appear in a readable and well-aligned layout. In addition to showing those results, also explain what you find "weird" about them, unless it is blatantly obvious even to people who have no idea what stock data is.

If you do those things, you are likely to get a timely and helpful response. Without them, there just isn't enough information given in your post for anyone to figure out what the problem is.

[Gabriel Davidson]

Dear Mr Schechter, thank You for Your timely response.
Unfortunately I do not have my prior results handy at the moment.
By weird results I meant that my calculated rolling Beta, which is usually supposed to be around one was over 200 for most observations.
I have inserted my data using dataex as advised for reference.
I am trying to calculate a daily rolling beta with an observation period of 365 days seperately for each of the companies in my dataset.
Again, any support is highly appreciated.
Gabriel

* Example generated by -dataex-. For more info, type help dataex
clear
input str8 cusip double mdate str5 ticker double stock_return float market_return
"23305187" 21626 "ASHR"   .023835  .003706
"24610T10" 19620 "VFL"    .009386 -.001843
"13078810" 18805 "CWT"    .004934  .009184
"46428720" 18330 "IVV"    .001134  .001713
"03761U50" 18847 "AINV"  -.074939 -.066634
"68162K10" 19719 "ZEUS"   .008699 -.000337
"74347B11" 21910 "SPXU"   .000504  .000034
"92912J10" 18410 "IAE"    .006365 -.012375
"33939L77" 20964 "MBSD"   .002679  .004442
"46428840" 20263 "JKI"   -.007059 -.007353
"26846110" 18469 "EEG"    .000832    .0112
"46090E10" 18266 "QQQQ"   .014645  .016043
"12326C10" 21703 "BFST"    .00938 -.002765
"42774610" 21901 "HRTX"  -.007115 -.000432
"56086P20" 18547 "MMUS"   .040323  .003819
"91890520" 18637 "VHI"   -.001391 -.001376
"46428943" 18588 "IWY"          0 -.001575
"37954Y53" 21732 "CHIK"   .011235  .002928
"90214J10" 20342 "TWOU"   .000277  .004487
"91825510" 18722 "VIST"  -.005848  -.00018
"69047Q10" 18421 "ECA"    .009144  .010976
"57722310" 18569 "MTSN"   .020243  .003678
"90346E10" 19751 "SLCA"   .033153  .006141
"69313910" 18939 "PNG"   -.001666  .011736
"43851610" 19099 "HON"   -.001205 -.000504
"90210410" 20402 "IIVI"  -.003245  .001511
"53030730" 20886 "LBRDK" -.004591 -.002284
"15087010" 18802 "CE"    -.012126 -.011726
"18449610" 18266 "CLH"    .017782  .016043
"86793170" 21551 "SCON"   .056589  .001269
"25460E86" 21237 "SPDN"  -.000662  .000974
"24237020" 19565 "DF"     .002812  .000828
"46431W85" 20356 "COMT"   .000536 -.000466
"02083G10" 21166 "ALPN"  -.000894 -.000473
"74347G73" 18941 "QID"    .002222  .008616
"74347X62" 18266 "URE"   -.005806  .016043
"27785810" 18266 "ESIC"   .004128  .016043
"80688210" 18282 "SCHN"  -.014774 -.010598
"07588710" 19024 "BDX"    .013901    .0089
"11403910" 18576 "BFSB"   .027559  .004376
"92647N82" 20501 "CDC"    .011188   .01648
"03071H10" 21210 "AMSF"    -.0112  .011841
"05377410" 21640 "CAR"    .025817  .011569
"26815740" 18266 "DYNT"  -.014771  .016043
"12590610" 18266 "CSS"    .060185  .016043
"90278810" 19156 "UMBF"   .025714  .011651
"46924R10" 19471 "JXSB"   .002625  .010419
"92206C77" 18266 "VMBS"   .003027  .016043
"95825410" 21350 "WES"   -.038544 -.001017
"46429B20" 19234 "ECNS"   .013068  .000844
"55345K10" 20615 "MRC"   -.049933 -.009175
"90292510" 20362 "USAK"   .031341  .001974
"55273810" 20726 "MFM"    .005405 -.009321
"33737J13" 20944 "FBZ"     .01823  .004089
"47804J30" 21340 "JHMC"   .005803  .000703
"84476710" 18266 "OKSB"   .023055  .016043
"46564T10" 19369 "ITI"   -.005848 -.000048
"05589T10" 18267 "DMF"    .008216  .003116
"12640210" 20627 "CSWI"   .007546 -.001652
"42026110" 18266 "HWKN"   .010536  .016043
"00174410" 19459 "AHS"    .015928  .003552
"46636710" 18556 "JACK"  -.005124  .001774
"90187B40" 19481 "TWO"   -.032663  .010535
"95790J10" 20381 "TLI"   -.015253 -.001421
"90214Q76" 21852 "AFIF"   .003113  .003253
"25401T10" 20699 "NSAM"   .019339  .004201
"91359V10" 20416 "UVE"   -.061015  .001222
"74734R10" 18879 "QRE"   -.033841 -.026705
"47580P10" 20850 "JELD"   .039155  -.00089
"89678F10" 19633 "TSC"   -.006163 -.000667
"10482B10" 21640 "BHR"    .020475  .011569
"91382130" 18722 "UUU"      -.025  -.00018
"20779710" 21489 "CTWS"   .000723  .010558
"45167R10" 18266 "IEX"    .016051  .016043
"05464C10" 18268 "TASR"   .242553  .000546
"91329010" 18415 "UNTY"   .039252  .025843
"08468010" 19992 "BHLB"   .003791  .008576
"59511210" 18266 "MU"     .027462  .016043
"72919P20" 18430 "PLUG"   .046512  .001283
"61945C10" 21854 "MOS"    .039235  .009662
"92240M10" 18807 "VGR"    .002262  .008283
"31679P10" 20437 "FSAM"   .011111  .010619
"74347X79" 18570 "URTY"   .074374  .019283
"23305166" 21777 "JPN"    .010802  .014426
"03763610" 21174 "AFT"   -.003056  .001986
"46428865" 19922 "TLH"   -.002384  .010265
"33738R76" 21620 "RNLC"   .008537  .002953
"30224P20" 19956 "STAY"   .010643   .00295
"46428783" 19975 "IYM"    -.00939 -.006545
"26747510" 18296 "DY"    -.012195 -.005474
"62846410" 18273 "MYE"   -.009667  .001747
"03833610" 19415 "ATR"     .00691  .006109
"65410610" 18276 "NKE"   -.006269  .002427
"52186N10" 19341 "LF"     .009259 -.004135
"46138E56" 20765 "LDRI"   .000798  .022224
"25459W56" 18266 "TYD"   -.002593  .016043
"26875P10" 19669 "EOG"   -.028646 -.013183
"16389320" 20150 "CHMT"   -.01378 -.004539
"98920710" 21061 "ZBRA"   .003839  .004615
"69350610" 21623 "PPG"    .003191  .004985
end
format %tdDD/NN/CCYY mdate

[Clyde Schechter]

OK, thanks for the -dataex-. I can't run a rolling regression on the example data because you have only a single observation for each cusip (or ticker--they are in 1:1 correspondence). I can't see any good reason why the code shown in would not work correctly for you provided you changed the value of the -window()- option to 365 and replace firm by cusip or ticker.

So my hypothesis is either that your expectations about what the values of beta should be are incorrect, or, more likely, something is wrong with your data. Here's what I recommend you do to investigate further. Review the results you have gotten so far and identify a few of the cusip's for which the betas strike you as implausible. Then browse the data for just those cusips and see if you can spot some values that look bad. Seeing the bad data may give you a clue about why it's bad and how you might fix it.

Added: Wait, I do see a problem. It isn't even clear to me how you got -rolling- to run at all. To use -rolling- you need to -tsset- or -xtset- your data first. But you have no numeric panel variable to do that with. Both cusip and ticker are string variables, and neither -xtset- nor -tsset- will accept them. So you must have done something funky with -xtset-/-tsset- because you did get -rolling- to run nonetheless. So please post back showing the exact code you ran, starting from your -xtset- or -tsset- command, and any code you wrote to create a variable that -xtset-/-tsset- would accept as a panel variable, through the -rolling- command.

[Gabriel Davidson]

Thanks again for the quick response!
I used the code below using the -rangestat- method , assuming this will get me the daily rolling stock beta as described above in #2.

rangestat (reg) stock_return market_return, by(cusip) interval(mdate -365 -1)

I cleaned and reloaded my dataset as suggested and I believe I have sensible results now (not sure how to share the results since there are so many companies and one beta by itself for one company is not really meaningful in the dataex output).
I wanted to be certain that the formula I used provides the results I am looking for which is daily rolling stock betas calculated as (covariance(stock_return,market_return)/Var(market_return)) for each company by cusip with an observation period of one year. Would be great if someone familiar with the -rangestat- package could confirm.

Thanks alot!
Gabriel

Again, my data in case that helps (including my calculated StockBetas):

* Example generated by -dataex-. For more info, type help dataex
clear
input str8 cusip double mdate str5 ticker double stock_return float market_return double StockBeta
"23305187" 21626 "ASHR"   .023835  .003706     .892705925196281
"24610T10" 19620 "VFL"    .009386 -.001843   .19800738384552294
"13078810" 18805 "CWT"    .004934  .009184    .6732097615024479
"46428720" 18330 "IVV"    .001134  .001713    .9919297917883115
"03761U50" 18847 "AINV"  -.074939 -.066634    1.465681603271195
"68162K10" 19719 "ZEUS"   .008699 -.000337     1.64030843080252
"74347B11" 21910 "SPXU"   .000504  .000034  -2.9836481392980425
"92912J10" 18410 "IAE"    .006365 -.012375    1.376470343228398
"33939L77" 20964 "MBSD"   .002679  .004442 -.024109124641230554
"46428840" 20263 "JKI"   -.007059 -.007353     .896455511037068
"26846110" 18469 "EEG"    .000832    .0112    1.163869350977153
"46090E10" 18266 "QQQQ"   .014645  .016043                    .
"12326C10" 21703 "BFST"    .00938 -.002765    .5178547299074445
"42774610" 21901 "HRTX"  -.007115 -.000432   1.0964329703894817
"56086P20" 18547 "MMUS"   .040323  .003819    .5668207045111178
"91890520" 18637 "VHI"   -.001391 -.001376   1.5464852402988218
"46428943" 18588 "IWY"          0 -.001575    .8584965305239649
"37954Y53" 21732 "CHIK"   .011235  .002928   1.1661167005243738
"90214J10" 20342 "TWOU"   .000277  .004487    .6179837784877977
"91825510" 18722 "VIST"  -.005848  -.00018    .5614593562330275
"69047Q10" 18421 "ECA"    .009144  .010976    1.221338933176252
"57722310" 18569 "MTSN"   .020243  .003678    2.074004776235268
"90346E10" 19751 "SLCA"   .033153  .006141   1.4970851171149162
"69313910" 18939 "PNG"   -.001666  .011736    .4082542749657267
"43851610" 19099 "HON"   -.001205 -.000504   1.2813770827927304
"90210410" 20402 "IIVI"  -.003245  .001511    .7017199710083349
"53030730" 20886 "LBRDK" -.004591 -.002284    .9688106538793758
"15087010" 18802 "CE"    -.012126 -.011726   1.8553516625438589
"18449610" 18266 "CLH"    .017782  .016043                    .
"86793170" 21551 "SCON"   .056589  .001269   .38332822005779177
"25460E86" 21237 "SPDN"  -.000662  .000974   -1.001745298567352
"24237020" 19565 "DF"     .002812  .000828    .9494859254938147
"46431W85" 20356 "COMT"   .000536 -.000466    .7412149236056794
"02083G10" 21166 "ALPN"  -.000894 -.000473    .7087160838276706
"74347G73" 18941 "QID"    .002222  .008616  -1.9105879005933917
"74347X62" 18266 "URE"   -.005806  .016043                    .
"27785810" 18266 "ESIC"   .004128  .016043                    .
"80688210" 18282 "SCHN"  -.014774 -.010598   1.8796012712990073
"07588710" 19024 "BDX"    .013901    .0089    .7037437086034988
"11403910" 18576 "BFSB"   .027559  .004376    .7499288881018805
"92647N82" 20501 "CDC"    .011188   .01648    .7704799434638845
"03071H10" 21210 "AMSF"    -.0112  .011841   1.0955674531643278
"05377410" 21640 "CAR"    .025817  .011569   1.7542400876737385
"26815740" 18266 "DYNT"  -.014771  .016043                    .
"12590610" 18266 "CSS"    .060185  .016043                    .
"90278810" 19156 "UMBF"   .025714  .011651   1.5619297281678637
"46924R10" 19471 "JXSB"   .002625  .010419  -.07271651262799071
"92206C77" 18266 "VMBS"   .003027  .016043                    .
"95825410" 21350 "WES"   -.038544 -.001017    .8729295789606891
"46429B20" 19234 "ECNS"   .013068  .000844     1.07416502353452
"55345K10" 20615 "MRC"   -.049933 -.009175   1.4731644890109459
"90292510" 20362 "USAK"   .031341  .001974   1.0303423845499822
"55273810" 20726 "MFM"    .005405 -.009321  .002391261912530691
"33737J13" 20944 "FBZ"     .01823  .004089   1.6261182627022106
"47804J30" 21340 "JHMC"   .005803  .000703    .6955163206650468
"84476710" 18266 "OKSB"   .023055  .016043                    .
"46564T10" 19369 "ITI"   -.005848 -.000048    .3347822749104675
"05589T10" 18267 "DMF"    .008216  .003116                    .
"12640210" 20627 "CSWI"   .007546 -.001652   .07309732579743403
"42026110" 18266 "HWKN"   .010536  .016043                    .
"00174410" 19459 "AHS"    .015928  .003552   1.3347955532810023
"46636710" 18556 "JACK"  -.005124  .001774    .8241547508398113
"90187B40" 19481 "TWO"   -.032663  .010535    .6314490494115711
"95790J10" 20381 "TLI"   -.015253 -.001421   .17191713178648496
"90214Q76" 21852 "AFIF"   .003113  .003253  -.07705500506845772
"25401T10" 20699 "NSAM"   .019339  .004201   1.4625090414361421
"91359V10" 20416 "UVE"   -.061015  .001222   1.3553256872817092
"74734R10" 18879 "QRE"   -.033841 -.026705     .719396953726553
"47580P10" 20850 "JELD"   .039155  -.00089                    .
"89678F10" 19633 "TSC"   -.006163 -.000667   .18432959238891222
"10482B10" 21640 "BHR"    .020475  .011569    .5830313996736869
"91382130" 18722 "UUU"      -.025  -.00018    .4817436626408654
"20779710" 21489 "CTWS"   .000723  .010558   .19905917882565807
"45167R10" 18266 "IEX"    .016051  .016043                    .
"05464C10" 18268 "TASR"   .242553  .000546                    .
"91329010" 18415 "UNTY"   .039252  .025843 -.027568692368822853
"08468010" 19992 "BHLB"   .003791  .008576     .964469892616046
"59511210" 18266 "MU"     .027462  .016043                    .
"72919P20" 18430 "PLUG"   .046512  .001283   .35080912038408024
"61945C10" 21854 "MOS"    .039235  .009662   1.4439139396188418
"92240M10" 18807 "VGR"    .002262  .008283    .4984547804359235
"31679P10" 20437 "FSAM"   .011111  .010619    .6735840283628172
"74347X79" 18570 "URTY"   .074374  .019283   3.9002795151126715
"23305166" 21777 "JPN"    .010802  .014426    .7174346556580157
"03763610" 21174 "AFT"   -.003056  .001986   .30064125422595644
"46428865" 19922 "TLH"   -.002384  .010265  -.20933351308343542
"33738R76" 21620 "RNLC"   .008537  .002953    .7601910399966961
"30224P20" 19956 "STAY"   .010643   .00295   1.0840799196726572
"46428783" 19975 "IYM"    -.00939 -.006545   1.0248731430398064
"26747510" 18296 "DY"    -.012195 -.005474   1.6582223723265257
"62846410" 18273 "MYE"   -.009667  .001747   1.9603893422671688
"03833610" 19415 "ATR"     .00691  .006109    .9390253289491751
"65410610" 18276 "NKE"   -.006269  .002427   -.0899985286620045
"52186N10" 19341 "LF"     .009259 -.004135   1.2887470618558607
"46138E56" 20765 "LDRI"   .000798  .022224  -.17143878077044525
"25459W56" 18266 "TYD"   -.002593  .016043                    .
"26875P10" 19669 "EOG"   -.028646 -.013183   1.5182498365746002
"16389320" 20150 "CHMT"   -.01378 -.004539   1.2270635850033134
"98920710" 21061 "ZBRA"   .003839  .004615   1.7577883933814706
"69350610" 21623 "PPG"    .003191  .004985    .7607361246949624
end
format %tdDD/NN/CCYY mdate