Holt-Winters ETSNBS CPI Data 2021–20243/6/12-Month ForecastsConfidence Intervals

Nigerian Inflation Forecasting Model

Triple exponential smoothing (Holt-Winters additive) trained on NBS headline CPI data, producing 3, 6, and 12-month forecasts with configurable confidence intervals.

3-Month Forecast
37.8%
90% CI: 35.4% – 40.1%
16.4pp above CBN target
6-Month Forecast
39.6%
90% CI: 36.2% – 42.9%
18.2pp above CBN target
12-Month Forecast
41.0%
90% CI: 36.3% – 45.8%
19.6pp above CBN target
CBN 2025 Target
21.4%
Single-digit long-term aspiration
Last reading: 34.80% (Dec 2024)

Inflation Trajectory & Forecast

Last 24 months of actuals + 12-month forecast

Model Parameters

α — Level Smoothing0.40
Responsiveness to recent level changes. High α → more reactive.
β — Trend Smoothing0.10
Dampens trend component. Low β → smoother trend.
γ — Seasonal Smoothing0.30
Updates seasonal factors. High γ → seasonal pattern adapts faster.
Confidence Interval

Model Information

Method
Holt-Winters Additive
Season Length
12 months
Training Data
48 months
Residual σ
0.83pp

Key Inflation Drivers

FX pass-through
↑High

Naira weakness feeds import costs and PMS prices

PMS subsidy removal
↑High

Full downstream deregulation persists into 2025

Food supply shocks
↑Moderate

Flooding and insecurity in food-producing states

CBN tightening (MPR 27.5%)
↓Moderate

CRR hike and hawkish stance dampening demand

Base effects
↓Moderate

High 2024 base suggests softer YoY prints in H2 2025

Fiscal deficit monetisation
↑Low

CBN direct financing of fiscal gap reduced

Methodology

The Holt-Winters additive model (triple exponential smoothing) decomposes the NBS headline CPI series into level (L), trend (T), and seasonal (S) components updated each period by α, β, and γ respectively. Forecast: F(t+h) = (L + h·T) + S(t+h−m) where m = 12 (annual seasonality). Confidence intervals assume normally distributed residuals scaled by σ√h. Training data: 48 monthly NBS observations (Jan 2021 – Dec 2024).