A country growth forecast is a build, not an opinion, and it has exactly four inputs. Assemble them in this order.
Labour comes first, because it is the one term the projections have largely already settled. Start from the working-age population growth the projections imply, add or subtract a participation drift you are willing to defend, and convert headcount to hours if the working week is trending. India over the coming decade gets working-age growth of about 0.85% a year plus a modest participation gain, so roughly 1.00%. The United States gets roughly 0.35%.
How do you forecast capital growth?
Capital comes second, and you never forecast gK directly. Forecast the investment rate, take the capital-output ratio from your own perpetual-inventory series, and let gK = s/(K/Y) − δ produce the growth rate. India at s = 34.23% and K/Y = 2.78 with δ = 5% gives 7.31%; the United States at s = 21.89% and K/Y = 2.98 gives 2.35%. The indirect route is better precisely because it forces the rising denominator into view, and a rising capital-output ratio is the mechanism that ends every investment-led growth run.
Third comes the residual, anchored on the realised decade with any deviation argued explicitly. Extrapolating a good decade is the commonest error in country forecasting, and the base rate for a middle-income economy's TFP accelerating is low.
Inflation comes last and only if you need a nominal number, taken from the central bank's target rather than the trailing average, since the terminal state assumes the target is met.
Advertisement
How do you test and bound the forecast?
Then state a band and three falsifiers. The band comes from running the residual low, base and high, not from a confidence interval you invented. India at 0.80%, 1.30% and 2.00% gives real growth of 4.33%, 4.83% and 5.53%; the United States at 0.80%, 1.20% and 1.60% gives 1.85%, 2.25% and 2.65%.
And then do the arithmetic that most forecasts skip. A widely quoted real forecast of 6.50% for India requires a residual of 2.975%, which is 2.01 times the 1.480% actually realised over the decade to FY2024. That is not a refutation of the forecast; TFP can accelerate, and reallocation out of agriculture plus formalisation are the named mechanisms by which it might. It is a demand that whoever quotes 6.50% says out loud that they are assuming the residual doubles, because until the decomposition is on the page nobody notices they have.
The three falsifiers for the India build, stated so they could actually fire: the investment rate falls below 30% of output for three consecutive years; the working-age participation gain reverses; or measured TFP growth over any rolling five-year window falls below 0.8%. Write yours the same way, as observable events with thresholds and windows, or they are not falsifiers.