Quarterization query

[Daniel Gaskin]

I am attempting to turn an annual time series into quarterly data.

I have attempted to use denton and cubic splines but to no avail ... could anyone shed some light on an example of how to turn annual data into quarterly data?

Thanks in advance!

Daniel

[Nick Cox]

In general, or so I assert, this only makes sense if you have some quarterly series to use to estimate the quarterly equivalent of a yearly series.

denton (SSC) -- please note the request in the FAQ Advice to explain where user-written programs you refer to come from -- is one such method.

I can't see that any kind of splines would help you here.

You have provided precisely no details of exactly what you tried and exactly what went wrong, so "follow the pattern of denton" seems the only possible advice. Again, please study the FAQ Advice for indications of how to ask more answerable questions.

[David Poensgen]

I am currently doing some work involving methods of temporal disaggregation; if you specified more clearly what the data you have looks like (especially wether you have a related high-frequency-series you want to use or you are looking for a univariate method, wether the series to be disaggregated stock or flow data etc.) it would make helping you much easier. If you have a dta-file for Stata, a description of its structure would be helpful. Moreover you sound like you are not even sure which method to use (there are quite a few variations)? Are you asking for technical assistance or "statistical advice"?

If the latter, you should tell us what you are planning with the high frequency estimation as well. (By the way, it is my personal and humble opinion so far that, at least in most cases, it is the better idea to stay away from disaggregation and just work with annual data).

Edit:
I found this article to be a good starting point, as it contains a very brief comparison of methods - and also an overview on existing software implementations:
http://journal.r-project.org/archive...ax-steiner.pdf (Sax/Steiner, Temporal Disaggregation of Time Series, R Journal 5/2 (2013), pp. 80-87).
Also, i tried opening the Excel file in the thread duplicate, to no avail.

[sotomayor]

I applied denton before with no problems. I checked what I did several months ago because I have another variable that I need to be in quarters. First I proceed with the annual data (varyear), I tsset year, yearly and save it as datayear.dta. Then I have another file with my other variable in quarters the one that I will use for interpolation. I applied first encode since my variable period was a string variable. Then I applied tsmktim to indicate that my variable is in quarters, I saved the file too as dataquarter.dta.

Then I follow this command:

denton varyear using newfile.dta, interp(varquarter) from(dataquarter.dta) gen(newvarquarter)

varquarter is the variable from my dataquarter file and newvarquarter is the new variable that will be generated by denton. I think I am following the command just fine, but I received the message "invalid syntax" . My data is not a stock is a flow (exports and gdp). Any help will be appreciated.

[Nelson S.]

Hey man I'm actually new at this command also but I think it deppends on the theorical background of your data, for example if you want to convert quarterly gdp data, you should use an indicator like the monthly indicator of economic activity as your reference indicator for the denton command because is a proxy of the behavior of the monthly gdp.

[Nelson S.]

Dear Sotomayor, I actually used your sintax to run denton myself and worked like a charm