EC[ON]OMY

Monthly vs. yearly inflation trends in Kazakhstan 2026

At the start of 2026, inflation in Kazakhstan reflects two parallel trends. The annual rate is gradually stabilizing, while monthly price growth remains elevated. This signals strong inertia in the current inflation regime. In this environment, annual inflation increasingly reflects base effects and accumulated price growth, while monthly inflation shows the real-time speed of price increases and the persistence of underlying pressure.

An important factor remains administrative control in regulated services. The extension of utility tariff restraints through the first half of 2026 limits the risk of acceleration in service inflation and reduces second-round effects, when higher regulated costs spill over into a broad range of goods and services.

However, even with this tariff anchor in place, the structure of inflation continues to show a dominant contribution from food. This makes the overall index more sensitive to seasonal and logistics-related shocks in the food segment.

Forecast results: march 2026 and year-end trajectory

According to our estimates, inflation in February 2026 stands at 11.7% year-on-year and 1.1% month-on-month. The forecast covers not only headline inflation but also its key components, allowing a deeper assessment of the quality of disinflation.

Component breakdown for February 2026:

  • ⁠ ⁠Food products: 12.5% y/y (1.5% m/m)
  • ⁠ ⁠Non-food goods: 11.6% y/y (0.8% m/m)
  • ⁠ ⁠Paid services: 10.7% y/y (1.1% m/m)

The economic interpretation of this March profile is straightforward.

First, a plateau in annual inflation at 11.7% means that monthly price growth around 1% remains high enough to prevent rapid disinflation.

Second, the structure confirms the key risk of 2026: food prices continue to outpace other components and remain the main source of inflationary pressure. This is visible both in annual terms (12.5%) and in the monthly impulse (1.5%).

Services look more controlled in annual terms, partly due to tariff measures. However, their monthly growth of 1.1% shows that price pressure in the service sector has not disappeared.

Trajectory through end-2026

The model assumes monthly inflation will remain relatively high, in the range of 0.8–1.0% m/m. In practical terms, this means that even with gradual moderation in annual inflation, the underlying pace of price growth stays elevated, limiting the scope for a sharp decline in the headline figure.

Forecast for end-2026: 11.7% y/y, including:

  • ⁠ ⁠Food products: 12.4% y/y (1.1% m/m)
  • ⁠ ⁠Non-food goods: 10.4% y/y (0.8% m/m)
  • ⁠ ⁠Services: 11.2% y/y (0.7% m/m)

Forecast accuracy: february Check

It is important to assess model performance. Previously, the National Bureau of Economic Research projected February inflation at 11.7% y/y, which exactly matched the official figure.

However, food inflation turned out stronger than expected:

  • ⁠ ⁠Forecast: 12.2% y/y
  • ⁠ ⁠Actual: 12.7% y/y

Services were precisely on target (10.8% y/y), and non-food goods were close (forecast 11.7% vs. actual 11.6%).

In short, the headline forecast was accurate. The deviation came from structure rather than level, with inflation proving more food-driven than anticipated.

Forecast methodology

The model is built around statistical robustness and comparability across time.

  • ⁠ ⁠Data horizon: Monthly observations from January 2011 through the latest available data, ensuring a long history to identify seasonality and recurring inflation patterns.
  • ⁠ ⁠Stationary transformation: Models are estimated on stationary representations of the data. This avoids spurious relationships and unstable coefficients, which are common in non-stationary macro series.
  • ⁠ ⁠Consensus of independent econometric models:

The final projection is not based on a single specification. It is an aggregated result.

Key advantages:

o Reduced model risk – averaging smooths temporary over- or under-estimation.

o Greater resilience to structural breaks and outliers.

o Better structural consistency, especially when forecasting both headline and components.

Forecasting components is not cosmetic. It is a stress test. It shows whether disinflation is broad-based or dependent on a single anchor.

Statistical error metrics

Model credibility depends on actual performance.

  • ⁠ ⁠In 2025, the average monthly error was 0.1 percentage points in both annual and monthly terms.
  • ⁠ ⁠In February 2026, deviation was 0.0 p.p. y/y and 0.2 p.p. m/m.

These values suggest the model adequately captures current inflation dynamics and short-term inertia. That said, econometric models always face natural limits. Unexpected supply shocks, logistics disruptions, or sudden administrative decisions can cause short-term deviations.

This material is analytical in nature and does not constitute individual financial or investment advice. The actual inflation path may deviate from projections due to supply shocks, tariff policy changes, external price movements, or shifts in economic behavior.

National Bureau of Economic Research specifically for EconomyKZ.org

Scroll to Top