Kazakhstan Economic Risk
Low Risk (20–40)
Composite index computed from five macro indicators with fixed, published weights. Independent model estimate — not an official assessment. Data as of 19 August 2026.
Contribution by factor (weight × sub-score)
- Inflation risk 25% · 10.2% CPIscore 42 → +10.6 pts
- Currency risk 25% · -14.2% KZT/USD (12m)score 0 → +0.0 pts
- GDP growth risk 20% · 4.1% real growthscore 21 → +4.2 pts
- Unemployment risk 10% · 4.5% unemployedscore 18 → +1.8 pts
- Interest-rate risk 20% · 16.75% base ratescore 77 → +15.4 pts
Key indicators — select a card to inspect
Inflation (CPI, year-on-year)
Source: Trading Economics / Bureau of National Statistics · Frequency: Monthly · Unit: % · Last updated: August 2026 (July 2026 print)
Actual · July 2026
10.2 %
Previous · June 2026
10.3 %
AI Forecast — independent model estimate
Horizon: 5 months (Aug – Dec 2026)
| Period | Forecast | Range | vs current |
|---|---|---|---|
| Aug 26 | 10.1 % | 9.7 – 10.5 | -0.1 |
| Sep 26 | 10.0 % | 9.6 – 10.4 | -0.2 |
| Oct 26 | 9.8 % | 9.3 – 10.3 | -0.4 |
| Nov 26 | 9.6 % | 9.1 – 10.2 | -0.6 |
| Dec 26 | 9.5 % | 9.0 – 10.2 | -0.7 |
What drives it: The forecast assumes continued gradual disinflation, a relatively tight monetary stance (real base rate around +6pp), stable food prices and no major external inflation shock. Momentum in the last three prints has been roughly -0.15pp per month, and the model extrapolates a slightly slowing version of that path. Regulated utility tariffs and services inflation are the main reasons the descent is modelled as slow rather than sharp.
External / official forecast (for comparison)
Trading Economics consensus
10.5% next print, ~10.3% end-2026, ~9.5% in 2027
Role in the risk index
CPI above 10% is more than triple the 5% target band midpoint, so inflation remains the largest single contributor to the risk index.
Kazakhstan Economic Risk — 2015–2026
Same weighted methodology applied to annual actual data (CPI, real GDP growth, KZT/USD change, unemployment, base rate).
Current (2026)
32.0
2026 · Low Risk
1 year ago
35.6
2025 · Low Risk
5 years ago
25.5
2021 · Low Risk
10 years ago
32.3
2016 · Low Risk
Risk changed by -36.3 points over the selected period (2015 → 2026).
Kazakhstan Economic Risk Forecast
AI Forecast — independent model estimate. The same weighting is applied to the forecast values of all five indicators.
Current risk score
32.0
Low Risk
3 months (Nov 2026)
28.4
-3.6 pts · Low Risk
6 months (Feb 2027)
27.5
-4.5 pts · Low Risk
12 months (Aug 2027)
32.7
+0.7 pts · Low Risk
Risk is expected to decline moderately over the next 6 months (-4.5 points), mainly because inflation is projected to ease from 10.2% toward ~9.1% and monetary policy may gradually loosen toward a 15.75% base rate. Working against that improvement: the model expects mild tenge depreciation, which raises the currency-risk component and partly offsets the inflation and interest-rate gains.
Actual vs AI Forecast — model performance
Every published forecast is stored and compared with the value that is later released. Errors are in the indicator's own units (percentage points, or tenge for KZT/USD). Pending periods have no actual release yet and are marked unavailable.
Inflation
MAE 0.13 pp · RMSE 0.13 pp · n=4
| Period | AI forecast | Actual | Error |
|---|---|---|---|
| Apr 26 | 11.2 | 11.1 | +0.1 pp |
| May 26 | 10.9 | 10.7 | +0.2 pp |
| Jun 26 | 10.4 | 10.3 | +0.1 pp |
| Jul 26 | 10.1 | 10.2 | -0.1 pp |
| Aug 26 | 10.1 | unavailable | — |
GDP Growth
MAE 0.40 pp · RMSE 0.42 pp · n=3
| Period | AI forecast | Actual | Error |
|---|---|---|---|
| 2024 | 4.6 | 4.8 | -0.2 pp |
| 2025 | 5.5 | 6.0 | -0.5 pp |
| H1 26 | 4.6 | 4.1 | +0.5 pp |
| 2026 | 4.9 | unavailable | — |
KZT/USD
MAE 4.75 ₸ · RMSE 5.02 ₸ · n=4
| Period | AI forecast | Actual | Error |
|---|---|---|---|
| May 26 | 480 | 474 | +6 ₸ |
| Jun 26 | 476 | 471 | +5 ₸ |
| Jul 26 | 472 | 470 | +2 ₸ |
| Aug 26 | 468 | 462 | +6 ₸ |
| Sep 26 | 465 | unavailable | — |
Unemployment
MAE 0.02 pp · RMSE 0.05 pp · n=4
| Period | AI forecast | Actual | Error |
|---|---|---|---|
| Q3 25 | 4.6 | 4.6 | 0.0 pp |
| Q4 25 | 4.6 | 4.5 | +0.1 pp |
| Q1 26 | 4.5 | 4.5 | 0.0 pp |
| Q2 26 | 4.5 | 4.5 | 0.0 pp |
| Sep 26 | 4.5 | unavailable | — |
Base Rate
MAE 0.17 pp · RMSE 0.20 pp · n=3
| Period | AI forecast | Actual | Error |
|---|---|---|---|
| Apr 26 | 17.75 | 17.50 | +0.25 pp |
| Jun 26 | 17.25 | 17.00 | +0.25 pp |
| Aug 26 | 16.75 | 16.75 | 0.00 pp |
| Oct 26 | 16.50 | unavailable | — |
Methodology
The risk index is fully transparent and reproducible. Each indicator is converted to a 0–100 sub-score by linear normalisation between a "benign" and a "stressed" reference level, then combined with fixed weights. Sub-scores are clipped to the 0–100 interval, so extreme readings cannot dominate the index without limit.
sub_score = clip( (value − benign) / (stressed − benign) × 100 , 0 , 100 ) risk_index = 0.25·inflation + 0.25·currency + 0.20·gdp + 0.10·unemployment + 0.20·base_rate
| Component | Weight | Normalisation | Current input | Sub-score | Contribution |
|---|---|---|---|---|---|
| Inflation risk | 25% | 3% CPI = 0 risk, 20% CPI = 100 risk | 10.2% CPI | 42.4 | +10.6 |
| Currency risk | 25% | -5% (appreciation) = 0 risk, +30% depreciation = 100 risk | -14.2% KZT/USD (12m) | 0.0 | +0.0 |
| GDP growth risk | 20% | +6% real growth = 0 risk, -3% recession = 100 risk | 4.1% real growth | 21.1 | +4.2 |
| Unemployment risk | 10% | 3.5% = 0 risk, 9% = 100 risk | 4.5% unemployed | 18.2 | +1.8 |
| Interest-rate risk | 20% | 6% policy rate = 0 risk, 20% = 100 risk | 16.75% base rate | 76.8 | +15.4 |
How each component affects the score
- Inflation risk: High consumer-price inflation erodes real incomes and forces tighter policy.
- Currency risk: Measured as 12-month change in KZT/USD. Rapid depreciation imports inflation and stresses FX debt.
- GDP growth risk: Weak or negative real growth signals falling activity, revenue and employment.
- Unemployment risk: Labour-market slack; small weight because the Kazakh rate is stable and slow-moving.
- Interest-rate risk: A high policy rate restricts credit and investment and raises debt-service cost.
Forecasting method
- Current implementation: a transparent statistical baseline — momentum extrapolation of the recent trend, damped toward a long-run anchor, with cross-indicator rules (inflation → base rate → FX) and judgement-based scenario bounds for the published ranges.
- Planned models: Linear Regression, Random Forest, XGBoost and ARIMA/SARIMAX estimated on the full historical series, with rolling-origin backtesting; the model with the lowest out-of-sample MAE/RMSE per indicator becomes the published forecast.
- Evaluation: every forecast vintage is stored, then scored against the released value using MAE and RMSE (see the model performance section), so improvement over time is measurable rather than claimed.
- Separation of sources: actual data, external/official forecasts and this project's independent forecast are stored and displayed as three distinct series — external consensus is never relabelled as an AI forecast.
Data sources
- Inflation — Trading Economics / Bureau of National Statistics · Monthly · updated August 2026 (July 2026 print)
- GDP Growth — Trading Economics / Bureau of National Statistics · Quarterly · updated August 2026 (Q2 2026 estimate)
- KZT/USD — Trading Economics market data / National Bank of Kazakhstan · Daily · updated 19 August 2026
- Unemployment — Trading Economics / Bureau of National Statistics · Quarterly · updated August 2026 (Q2 2026)
- Base Rate — Trading Economics / National Bank of Kazakhstan · Per meeting · updated August 2026
Limitations
- The index is a weighted composite, not a probability of crisis; weights are a modelling choice and are shown so that they can be challenged or re-estimated.
- Fiscal balance, current account, oil price, external debt and bank-sector indicators are not yet included, so commodity and external-financing risk is only captured indirectly.
- Exchange-rate forecasts have the widest uncertainty of all five indicators; point values should be read together with their ranges.
- Annual history uses period-average/period-end conventions per indicator; where a value is not published it is marked unavailable and never interpolated.
- The headline unemployment rate excludes informal employment and underemployment, which understates true labour-market slack in Kazakhstan.
About the project
This project analyzes Kazakhstan's macroeconomic conditions and estimates how economic risk changes over time using historical economic indicators and independent forecasting models.
It tracks five indicators — inflation, real GDP growth, the KZT/USD exchange rate, unemployment and the National Bank base rate — combines them into a transparent 0–100 Economic Risk Index, reconstructs that index back to 2015, projects it forward, and then scores its own forecasts against the data that is later released.
- Scope
- 5 indicators · annual risk history 2015–2026 · 3/6/12-month forward risk path · rolling forecast-error log
- Data
- Trading Economics country pages, National Bank of Kazakhstan, Bureau of National Statistics
- Status
- Student research / data-science project. Forecasts are independent model estimates, not official government or central-bank projections.