Twelve steps from opening the model to downloading a finished scenario. Most people are running their first one inside five minutes — the rest of this page is here for when you want to do something specific. Click any screenshot to enlarge it.
Almost every question we get about the model comes down to one of these. Two minutes here saves a lot of hunting later.
Choosing fewer topics, tables or years changes only what is on your screen. The model always solves every variable for every year from 2005 to 2065. You can never leave something out by accident.
Base is our forecast and never moves. Revised is your scenario. Difference is one minus the other. If your changes seem to have done nothing, you are almost certainly looking at Base.
The other 68 are calculated from them. You cannot type over next year’s fifteen-year-olds, because they are already determined by births fifteen years ago. That is a feature: it is what keeps a scenario internally consistent.
Start by picking what you want to model. You can select a single country, or tick several and have the program add them together: select every Western European country and the tables come back as one combined Western Europe — one population, one workforce, one economy.
Next, decide what appears on screen. This is presentation only: whatever you choose to display, the model still solves every variable for every year behind the scenes. You are picking a window, not restricting the calculation.
You must have at least one topic showing, and you can have as many as you like. The last one works for a single country only — not on an aggregate.
Show one or two topics at a time. Everything can be on screen at once if you want it, but you will spend the session scrolling. Close a topic and reopen it later — nothing is lost.
For every topic you can show any combination of three tables. The choice applies to all topics at once, and you can change it whenever you like.
| Table | What it shows |
|---|---|
| Base case | Our forecast, unchanged. Your fixed reference point. |
| Revised case | Your scenario. Also the table that carries the tooltips explaining each variable, and the one you need on screen to view trend charts. |
| Difference | Revised minus Base — the size and direction of the change you have made. |
Best to use Revised only, and select the others only when you need to compare results — this minimises scrolling.
The model covers 2005 to 2065 — actual data to 2024, forecast from 2025 onward. That is around sixty columns, far more than fits on a screen, so you choose which years to show. Again, this is display only: the model solves every year regardless.
Rather than tick years one at a time, set a step size and the program shows every nth year relative to your base year. With 2025 as base and a step size of 5 you get 2010, 2015, 2020, 2025, 2030 and so on — a readable, evenly spaced view of a sixty-year run.
This cannot be changed. It is the latest actual year — you cannot change that year, or any earlier year’s data.
Step size also controls the table you get when you go to change a variable (step 7), which is where it really earns its keep. A step size of 5 means entering five or six numbers instead of forty.
Nothing loads until you tell it to. Click the red button at the top right of the menu screen to pull in the data for your chosen countries and display it to your chosen specification. The button then changes its own label depending on what it is about to do.
| If it reads | Clicking it will |
|---|---|
| Applying Changes | Load the data for the countries you have selected, and display it. |
| Update Data Display | Redraw the screen only — because all you changed was a display setting such as the year range. Your loaded data and any scenario stay exactly as they are. |
Most of the 82 variables are outcomes, not inputs. You will know them because the button beside them in the Revised table is labelled as non-changeable.
Take the 15–24 age group. Next year’s figure is this year’s, plus the fourteen-year-olds moving up, minus the twenty-four-year-olds moving out, adjusted for the age profile of migrants and for deaths. There is no sensible way to type a different number over it — and if you did, the rest of the model would no longer add up. The program does that arithmetic for you.
The non-changeable variables are the ones a politician could not arbitrarily change. The changeable ones are the levers that policy, investment and social change genuinely pull.
They are not dead ends, though. Click the button beside any of them and you get a chart of that variable across the full run, so you can see the shape of what your scenario has produced.
The rest — fourteen of them — carry a change button. These are the levers:
You can only change future years — currently 2026 onward. History is history.
Clicking the change button opens a chart of that variable from 2005 to 2065. If you have already made changes, both the base case and your revised case are drawn together, so you can see at a glance how far you have moved it.
Click Change at the top left of the chart and a small table opens, showing only the years you chose to display — so your ‘step size’ setting from step 3 decides how much typing you are about to do. It arrives pre-filled with the existing forecast.
That last step is what makes this quick. The program fills every intervening year by drawing a straight line between the values you did enter. Set 2030 and 2050, and the twenty years between them are interpolated for you.
After clicking Save Change, you are returned to the chart for the variable, which now displays both the old/original data and the new values you just entered for it. If the line looks wrong, it probably is — fix it here rather than after running the scenario. Your other options are Cancel or Clear, as described below.
Three options sit at the foot of the table:
| Option | Result |
|---|---|
| Cancel | Nothing is changed. You go back to the main screen. |
| Clear | Empties the forecast values so you can enter a different set. |
| Save change | Writes your new values into the underlying database ready for modelling. Nothing is recalculated yet — that is step 9. |
Click Run New Scenario at the top right of the tables. Every table recalculates using your new values. It is normally very quick.
Check that Revised is ticked in the data control box. The Base table does not change — that is the whole point of it — so if Base is the only table on screen, a successful scenario looks identical to no scenario at all.
You do not have to do everything at once. Change one variable, run it, look at the result, then change a second variable and run again — the first change is still in place, so you are watching the combined effect build up. That works across sessions too.
To isolate the effect of a single variable, run it, download the result (step 11), then reset to base (step 10) before testing the next one. To build a compound story instead, just keep going.
Click Reset Data to Base Case at the top of the tables and everything — both the Base and the Revised tables — goes back to our original forecast.
Resetting discards the scenario you have built. If it is worth keeping, download it first.
At any point you can download what you are looking at as an .xlsx file — before running a
scenario, after running one, or several times through a session.
Downloading is also how you keep more than one scenario. The model holds a single revised case at a time, so to compare three futures side by side, run each, download it, and assemble them in Excel.
The last two topics — Distribution of Households by Income Segment, and Household Expenditure by Income Segment — divide households into five income groups. You decide where the lines fall. Select either topic in the data control area and a slider appears.
Set your breakpoints at levels that mean something for your business — the income at which households start buying what you sell — and the model will tell you how many households cross that line each year between now and 2065.
| Variable | Topic | What changing it does |
|---|---|---|
| Birth rate (per 000 women 15–49) | Population | Drives total births, and through them the whole age structure, workforce and household count from that year on. |
| Migration rate (% of population) | Population | Adds or removes people at a given age profile — so it moves births, workers and households, not just headcount. |
| Male propensity 15–64 | Labour Force | Share of prime-age men in work. Typically 70–90%. |
| Male propensity 65–69 | Labour Force | Typically a low percentage (15%–40%). |
| Male propensity 70–74 | Labour Force | Always low, around 10% at best. |
| Female propensity 15–64 | Labour Force | Share of prime-age women in work — usually the largest untapped labour reserve. |
| Female propensity 65–69 | Labour Force | Typically a low percentage (15%–40%). |
| Female propensity 70–74 | Labour Force | Always low, around 10% at best. |
| Share of previous year’s GDP invested in Fixed Capital | Resources per worker | Gives a measure of fixed resources behind each worker. |
| Growth rate of GDP per worker | Productivity | Your expected trend in productivity per worker. Default is the latest actual year, halving in value over 40 years. |
| Growth rate of Total GDP | Productivity | Your expected growth rate per annum for the total economy. Note that it retroactively changes GDP per worker to fit. |
| Wage ratio | Household Income | The share of output per worker that reaches the worker — the link between economy and household. |
| Household expenditure as % of income | Household Income | The savings-versus-spending decision; drives total consumer expenditure and PCE. |
| Median as % of average income | Households by Income | Income disparity. Lower means a longer tail and fewer households near the average. |
| Income segment breakpoints | Households by Income | The four cut-offs defining five income groups, and the number and percent of households in each. |
Full definitions for all 82 variables are on the data and methodology page.
The Base table never changes — that is what makes it useful as a reference. Tick Revised in the data control box, or better, tick Difference, which shows only what moved.
Yes. Tick as many as you like and the model aggregates them into a single set of tables. The two income-segment topics are the exception — those work for one country at a time.
Because it is already determined: it is the births of twenty years ago, adjusted for migration and mortality. Change the birth rate or the migration rate instead, and the model works it through for you.
No. Enter values for the years you care about, click Fill Forecast Data, and the program interpolates a straight line between them. Setting your step size first (step 3) makes this much faster.
The model holds one revised case at a time, so run the first, download it, reset to base, then run the second. Compare the two spreadsheets side by side.
No — only years from 2024 onward. Data to 2023 is actual, sourced from the United Nations and the World Bank.
Real 2023 US dollars, converted at 2023 exchange rates, so countries and years are directly comparable without any adjustment.
Email info@globaldemographics.com and tell us the country and the variable you were working with — it is usually a quick answer.
If you have not yet seen what the model covers — 109 countries, plus China national and national urban, 82 variables to 2065 — start there.