Forecast
Forward projections from the prediction engine — baseline plus an optimistic/pessimistic band and a confidence interval, per macro indicator. These are PROJECTIONS, not observations: generated from observed history, never a guarantee. A country we hold no projections for is shown honestly.
Forecast
Need it as JSON? GET /api/v1/forecast/country/{iso3}.
CMQ: forward projections across 7 macro indicators, horizons 2026, 2028, 2030, 2035. These are projections, not observations — each carries an optimistic/pessimistic band and a confidence interval, anchored against the latest observed value where one is held.
Nowcast
Each indicator nowcast from its strongest leading series — inferable from observations already in hand.
For each indicator, CMQ's strongest leading series is fit by ordinary least squares over their shared observed history, then the leader's already-observed recent values project the target — so the next year or two is inferable from data already in hand. These are nowcasts from a leading correlation: correlation, not causation, and not a guaranteed forecast.
Led by Current account / GDP by 1 year (it moves withthe target; r = 0.797 over 26 obs).
| Year | Nowcast | Implied by Current account / GDP in |
|---|---|---|
| 2026 | 55.4 | 2025 (1) |
Latest observed-0.3(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 1.28 | 4.57 | -2.01 | 0.59 – 1.97 |
| 2028 | 1.84 | 7.53 | -3.86 | 0.58 – 3.09 |
| 2030 | 2.39 | 9.75 | -4.96 | 0.7 – 4.09 |
| 2035 | 3.79 | 14.19 | -6.62 | 1.06 – 6.51 |
Blended 1-yr-ahead-0.82· skill-weighted across models
Track record±1.58typical miss · 1 realized · 526% MAPE
Method: linear_trend; lookback 8y; horizon 1y; confidence 80%
Latest observed595.1bn(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 569.5bn | 623.4bn | 522.8bn | 551.8bn – 587.1bn |
| 2028 | 636.2bn | 729.5bn | 555.3bn | 589.1bn – 683.2bn |
| 2030 | 702.9bn | 823.4bn | 598.4bn | 623.1bn – 782.7bn |
| 2035 | 869.6bn | 1T | 721.9bn | 689.2bn – 1T |
Blended 1-yr-ahead629bn· skill-weighted across models
Track record±25.6bntypical miss · 1 realized · 4% MAPE
Method: weighted_composite; lookback 8y; horizon 1y; confidence 80%
Latest observed4.2(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 3.92 | 12 | -10 | 0.84 – 6.99 |
| 2028 | 4.36 | 12 | -10 | -1.06 – 9.79 |
| 2030 | 4.81 | 12 | -10 | -2.31 – 11.92 |
| 2035 | 5.91 | 12 | -10 | -4.55 – 12 |
Blended 1-yr-ahead3.56· skill-weighted across models
Track record±0.28typical miss · 1 realized · 7% MAPE
Method: weighted_composite; lookback 6y; horizon 1y; confidence 80%
Latest observed9.2k(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 8.9k | 9.8k | 8.1k | 8.6k – 9.2k |
| 2028 | 9.8k | 11.3k | 8.5k | 9.1k – 10.5k |
| 2030 | 10.7k | 12.7k | 9k | 9.5k – 11.9k |
| 2035 | 13k | 15.8k | 10.6k | 10.3k – 15.7k |
Blended 1-yr-ahead9.7k· skill-weighted across models
Track record±346.04typical miss · 1 realized · 4% MAPE
Method: weighted_composite; lookback 8y; horizon 1y; confidence 80%
Latest observed51.2(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 52.42 | 60.66 | 42.92 | 50.31 – 54.53 |
| 2028 | 55.06 | 69.33 | 38.6 | 50.22 – 59.9 |
| 2030 | 57.7 | 76.12 | 36.45 | 50.25 – 65.15 |
| 2035 | 64.3 | 90.35 | 34.25 | 50 – 78.6 |
Blended 1-yr-ahead51.44· skill-weighted across models
Track record±1.22typical miss · 1 realized · 2% MAPE
Method: linear_trend; lookback 10y; horizon 1y; confidence 80%
Latest observed3(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 1.88 | 6.62 | -3.21 | 0.79 – 2.96 |
| 2028 | 1.32 | 9.53 | -5 | -0.57 – 3.2 |
| 2030 | 0.76 | 11.36 | -5 | -1.68 – 3.19 |
| 2035 | -0.64 | 14.35 | -5 | -4.13 – 2.84 |
Blended 1-yr-ahead3.59· skill-weighted across models
Track record±1.12typical miss · 1 realized · 37% MAPE
Method: exponential_smoothing; lookback 6y; horizon 1y; confidence 80%
Latest observed64.4m(2026)· IMF DataMapper API
| Year | Baseline | Optimistic | Pessimistic | Conf. band |
|---|---|---|---|---|
| 2026 | 63.4m | 66.4m | 60.6m | 61.4m – 65.3m |
| 2028 | 64.1m | 67.2m | 61.3m | 59.4m – 68.9m |
| 2030 | 64.9m | 68m | 62.1m | 57.5m – 72.3m |
| 2035 | 66.8m | 70m | 63.9m | 53m – 80.7m |
Blended 1-yr-ahead65m· skill-weighted across models
Track record±1mtypical miss · 1 realized · 2% MAPE
Method: exponential_smoothing; lookback 10y; horizon 1y; confidence 80%
Reliability badges show each indicator's best model skill— how much it beats a naive last-value forecast, measured out-of-sample on that series' own observed history (backtested one-step-ahead). A track record, not a guarantee of future accuracy.
Generated by the prediction engine (model prediction-engine-v1) on 2026-10-31 from observed history. Projections, not observations — never a guarantee of future values.
Machine-readable: GET /api/v1/forecast/country/CMQ · Country page → · Full brief → · API docs
Fit: Govt debt / GDP ≈ 52.6536 + 2.7489·Current account / GDP. Correlation, not causation.
Led by Govt debt / GDP by 1 year (it moves withthe target; r = 0.684 over 25 obs).
| Year | Nowcast | Implied by Govt debt / GDP in |
|---|---|---|
| 2026 | -1.5 | 2025 (51.1) |
Fit: Current account / GDP ≈ -12.5449 + 0.2161·Govt debt / GDP. Correlation, not causation.
Led by Govt debt / GDP by 2 years (it moves withthe target; r = 0.681 over 24 obs).
| Year | Nowcast | Implied by Govt debt / GDP in |
|---|---|---|
| 2026 | 399.8bn | 2024 (51.4) |
| 2027 | 396.3bn | 2025 (51.1) |
Fit: GDP (current US$) ≈ -201960406728.1222 + 11706872611.1532·Govt debt / GDP. Correlation, not causation.
Led by GDP (current US$) by 1 year (it moves inversely tothe target; r = -0.536 over 45 obs).
| Year | Nowcast | Implied by GDP (current US$) in |
|---|---|---|
| 2026 | -10.21 | 2025 (543.1bn) |
Fit: CPI inflation ≈ 23.7644 + 0·GDP (current US$). Correlation, not causation.
Machine-readable: GET /api/v1/nowcast/country/CMQ?indicator= · API docs
