Handling time in row-based input tables when rolling forecast

In HR Library based models, many of the input tables used are what we term “row-based” input tables. Each row in the table represents a particular rule used in the simulation. This type of table has several advantages. First of all, input can be made very concise. In the example below the columns ‘Job cluster’ and ‘Location’ are empty in the first row. This means that for the year 2018 and onwards, all job clusters and locations will use the same hiring rate. On the second row, a specific rule for Norway has been created. Starting from 2019, a different hiring rate will be used for Norway only. The generic rule on row number 1 continues to be used for all other locations and job clusters.

Row
Time
Job cluster
Location
Hiring rate
1
2018
100%/year
2
2019
Norway
80%/year

Used in this manner, row-based tables allow us to input sophisticated rules using minimal space.

What happens to a rule when we use ‘rolling forecast’ to adjust the time horizon of the model? The table below shows a typical setup of a model’s simulation horizon. The period between ‘start time’ and ‘start future’ is termed history and is represented in the model by the inputs in the records part of the HR Library. The period between ‘start future’ and ‘stop time’ is termed future. In this part of the time horizon, values from history combined with rules from the drivers section of the model are used to calculate future projections of the workforce.

Rolling forecast
Rolling forecast

When we roll the forecast forward by one year, Smia is designed to retain the values within history and future and discard the others. Hence, the values for 2016 are discarded in the records section and for 2018 in the drivers section.

history
future
start time
start future
stop time
Before rolling forecast
2016
2017
2018
2019
2020
After rolling forecast
2017
2018
2019
2020
2021

Our row-based table now looks like this:

Row
Time
Job cluster
Location
Hiring rate
1
2019
Norway
80%/year

The rule for 2019 is retained, while the rule for 2018, which is no longer part of future, is removed. This can in some cases be annoying, because we often use the ‘start future’ point in time to input generic rules that are valid for the whole future. Hence, it is often desirable to retain rules corresponding to the value of ‘start future’ as it was before rolling the forecast. Our desired result after rolling forecast is the following:

Row
Time
Job cluster
Location
Hiring rate
1
2019
100%/year
2
2019
Norway
80%/year

How can we achieve this in HR Library based models?

The answer is to refer directly to the built-in Smia object ‘start future’ instead of the year 2018. We set the original input table up like this:

Row
Time
Job cluster
Location
Hiring rate
1
‘start future’
100%/year
2
2019
Norway
80%/year

We now have a specific rule for 2019, but also a generic rule valid from the start of future. When we roll the forecast forward, both rows are retained.

To allow for inputting ‘start future’ and retaining the value, a change must be made in the formula of the ‘time’ column. To make the change, enter design mode by clicking on the blue triangular ruler symbol in the toolbar: Design mode icon Then click on the column header in the table to edit the formula.

Original formula:

data {r=row| } as future default empty

Updated formula:

{r=row| } as future default empty

Confirm with the Enter key. You can now write ‘start future’ instead of the year. ‘start future’ refers to the start of future and the row-based input table shows this value directly.

Design mode button
Design mode button

If you do not have design mode access or would like assistance in making the adjustments in your model, please contact us.

Comments

Let us know what you think

You have to log in before you can comment on this article.

Please enter your account information to log in.

Notifications

Subscribe You must log in before you can subscribe to email notifications on this article.