How it works
From official statistics to a forecast you can use
CiO Costs combines official construction price statistics with the things that move construction costs week to week, and turns them into a chance, an expected change and a likely range for each country.
Official construction price indices where they exist (Uganda, Kenya and South Africa in full; latest readings for Ghana, Zimbabwe, Senegal and Mauritius), world oil prices, exchange rates, rainfall, and published tax changes such as fuel and cement duties.
Where a country publishes a detailed index, each material is forecast from the drivers that move it — oil for diesel, the currency for imported steel, rainfall for sand — and added back up the way the statistics office does. Elsewhere, an import-cost indicator tracks the share of costs that moves with oil and the exchange rate.
Every method is tested on years of past data, as if it had been run at the time, and kept only if it beats simple rules of thumb. In Uganda the forecasts have been about 11–14% more accurate than a simple trend, and the likely ranges are set so that 8 in 10 past outcomes fell inside.
Oil and exchange rates are refreshed every few hours and the forecasts with them. Official data is added within a day of publication. A written note for each country comes out every Monday.
Track record
Uganda, overall construction input index, tested month by month since 2021 using only the data available at each point:
| Horizon | Errors vs a simple trend | Outcomes inside the likely range |
|---|---|---|
| 1 month ahead | 13% smaller | 89% raw → about 80% after calibration |
| 3 months ahead | 11% smaller | 76% raw → about 80% after calibration |
| 6 months ahead | 11% smaller | 69% raw → about 80% after calibration |
| 12 months ahead | 13% smaller | 67% raw → about 80% after calibration |
Each month we will also publish our forecast next to the official figure.
Live track record
Every forecast the site publishes is saved. When the official figure comes out, it is checked here automatically: was it inside the likely range, and how far off was it compared with assuming no change?
| Country | Forecasts checked | Inside the range (aim 80%) | Average miss, ours / no change |
|---|---|---|---|
| No saved forecast has reached its official release yet. The first checks appear as the next figures come out (benchmarks monthly, UBOS and Stats SA monthly, KNBS quarterly). | |||
Reading the numbers
- Chance of a rise: how likely prices are to be higher by the month you pick.
- Expected change: the most likely move from today's prices to the date you choose. Where official figures are months old, today's level is itself a forecast, so the range also covers the months not yet published (the chart shows the path from the latest official figure).
- Likely range: where the change should land 8 times in 10.
Limits
- Forecasts are estimates, not certainties. Always use the range.
- Import-cost indicators see oil and currency effects only, not local materials or labour.
- Where official statistics are published late, the forecast bridges the gap with the latest market data.
- Nothing here is financial or contractual advice.
Take part
Anyone can use CiO Costs and add prices — no account. Tell us what you paid for cement, steel, sand or diesel whenever you buy, even months later, and follow the free weekly update.