In Excel output trend forecast for only 12 months

[LeenaSL]

I'm using JDemetra v3.6.0

The issue is that in Excel output only 12 months of trend forecast data y_f(?) appear, seasonally adjusted forecast data appears for all 24 months in Excel.

Is this a bug?

[TanguyBarthelemy]

I don't know exactly what forecasts you're talking about.

In the GUI

Tables

You can find forecasted data in the tables (Main results ▶ Table) and the forecasted data are in italic:

D final tables

You can also find it in the D-Final tab (Decomposition (X11) ▶ D-Final-Table) inthe column which names ends with a:

Generating output

You can also generates outputs in Excel with choosing items that ending with _f and _f(?). But in my case it generates 12 forecast:

And the csv output also generates series but with 12 forecast too.

With the cruncher

You can use the package R {rjwsacruncher} to launch the cruncher on your WS and get as many forecasts as you want (but in csv files):

library("rjwsacruncher")

cruncher_and_param(
    workspace = "tmp.xml",
    cruncher_bin_directory = "C:/Users/UTZK0M/Software/jdemetra-related/jwsacruncher-3.6.0/bin/",
    v3 = TRUE,
    csv_layout = "vtable",
    policy = "complete",
    tsmatrix_series = c("y_f","y_f(24)", "y_f(36)")
)

In R with the rjdverse

The x13() function allows you to add userdefined variables (as forecasts):

library("rjd3x13")
#> 
#> Attaching package: 'rjd3x13'
#> The following object is masked from 'package:grDevices':
#> 
#>     x11

mod <- x13(AirPassengers, userdefined = c("y_f","y_f(24)", "y_f(36)"))
mod$user_defined
#> $y_f
#>           Jan      Feb      Mar      Apr      May      Jun      Jul      Aug
#> 1961 444.2958 413.5096 465.5510 497.5498 498.3543 575.2946 670.7150 654.9295
#>           Sep      Oct      Nov      Dec
#> 1961 556.7300 490.9867 422.5163 478.1584
#> 
#> $`y_f(24)`
#>           Jan      Feb      Mar      Apr      May      Jun      Jul      Aug
#> 1961 444.2958 413.5096 465.5510 497.5498 498.3543 575.2946 670.7150 654.9295
#> 1962 481.0623 452.3748 514.5929 538.7232 545.1939 635.8970 726.2182 716.4854
#>           Sep      Oct      Nov      Dec
#> 1961 556.7300 490.9867 422.5163 478.1584
#> 1962 615.3768 531.6170 462.2280 523.0999
#> 
#> $`y_f(36)`
#>           Jan      Feb      Mar      Apr      May      Jun      Jul      Aug
#> 1961 444.2958 413.5096 465.5510 497.5498 498.3543 575.2946 670.7150 654.9295
#> 1962 481.0623 452.3748 514.5929 538.7232 545.1939 635.8970 726.2182 716.4854
#> 1963 526.2767 494.8929 568.8009 583.3038 596.4359 702.8834 786.3145 791.9611
#>           Sep      Oct      Nov      Dec
#> 1961 556.7300 490.9867 422.5163 478.1584
#> 1962 615.3768 531.6170 462.2280 523.0999
#> 1963 666.3007 581.5829 510.9198 566.3876
#> 
#> attr(,"class")
#> [1] "user_defined"

^{Created on 2026-02-16 with reprex v2.1.1}

[LeenaSL]

Thank you for answering my question.

I'm only able to use the GUI and need output in Excel format. I can see forecast for series y_f(?) for 24 months in GUI but in Excel output it appears for 12 months only. In Excel output there is t_f forecast for 24 months.

[annasmyk]

That is strange...we'll retest this
In the mean time you can copy pastethe forecasts you need from GUI (Hopefully not too many)
or use the cruncher as indicated above, and turn the generated csv files into excel

[TanguyBarthelemy]

Hi @LeenaSL,
I'm able to reproduce your problem.

By default, for the items that ending with (?) (as y_f(?), ylin_f(?), cal_f(?)...), the output is 1 year (so 12 months). I didn't find any button or tools in the GUI to change this value (although it's possible in R or with the cruncher).

On the other hand, the items without (?) (t_f, i_f, sa_f...) that generates series have a length defined in the specification. In our case, it's defined in X11 ▶ Forecasts horizon:

In this example I defined 3 years of forecast (value -3).

You surely must have a forecast length of -2 (for 2 years = 24 months) for these series in your specifications.

I think the difference between these two types of elements is the seasonal adjustment phase they come from (pre-treatment part or decomposition part).

So in my opinion, it's not a bug but I raised an issue (jdemetra/jdplus-main#888) to see if it is possible to set the number of forecasts for the preprocessing part.