The Smith Machine v2.3 | RSA Economic Nowcasting & Firestore Live Edge

Smith Headline Nowcast (YoY) LIVE JEVONS

5.82% +142 bps vs Stats SA

MoM Impulse: +1.03% | Index: 101.17

Stats SA Print / Target Band 3.0%–6.0%

4.40% Midpoint: 4.50%

Kalman-Calibrated Bias: -0.03% (State-Space Adjusted)

SARB Policy Repo & Prime MPC BASELINE

7.25% Prime: 10.75%

Taylor Target: 8.75% | Output Gap: -0.45%

USD/ZAR & Brent Spot 95 ULP: R30.25/L

R17.40 / $88.50

ZAR-Brent Barrel: R1,540 | Real GDP YoY: +1.42%

31-Day High-Frequency Smith Nowcast vs. SARB Target Band (3.0% – 6.0%)

Daily Jevons-Laspeyres trajectory capturing advance retail food (17.14%) & fuel (14.35%) pass-through ahead of monthly Stats SA prints

Smith Nowcast Phillips Implied Official Stats SA (4.40%) 4.5% SARB Anchor

Jevons Elementary Sub-Indices (Base = 100.0)

Unweighted geometric mean of daily price relatives across Checkers Sixty60, Pick n Pay ASAP!, and Shoprite

COICOP 01: Food & NAB Jevons Index 102.33

Calibrated Weight: 17.14% | 30d Drift: +2.33%

COICOP 07: Liquid Fuels & Transport Index 103.61

Calibrated Weight: 14.35% | 95 ULP: R30.25/L · 50ppm Diesel: R33.29/L

Residual Sticky Core & Administered Index 100.37

Calibrated Weight: 68.51% | Core CPI Anchor: 4.50%

High-frequency staples directly track 31.49% of the Stats SA CPI basket daily, front-running official prints by 18–25 days.

Stats SA CPI COICOP Division Weighting Mapping

Young / Laspeyres expenditure weights (Sum = 100.00%)

HF Coverage: 31.49%

Macro Shock Pass-Through & SARB Taylor Rule Simulator

Log-linear ERPT & supply shock engine coupled with the SARB Taylor reaction function (Repo Baseline: 7.25% | Prime: 10.75%)

Exogenous Shock Parameters

Auto-Saved Offline

1. USD/ZAR Exchange Rate R17.40 (0.0%)

R14.00 Base: R17.40 R24.00

2. Brent Crude Oil Price ($/bbl) $88.50 (0.0%)

$50/bbl ZAR-Brent: R1,540/bbl $140/bbl

3. Agricultural / Grain Supply Shock 0.0%

-20% (Bumper Harvest) COICOP 01 Weight: 17.14% +20% (El Niño Drought)

4. Administered Eskom / Tariff Shock +12.5%

0% Default MYPD Baseline: 12.5% +25% (NERSA Clawback)

Calibrated Pass-Through Specification:

Δπ_t = 0.12·Δln(ZAR/USD) + 0.08·Δln(Brent) + 0.1714·AgriShock + 0.04·TariffShock

i_target = r* (2.5%) + π_proj + 0.5·(π_proj - 4.5%) + 0.5·(y_t - y*)

Projected Headline CPI

6.32% +0.50%

Unshocked Nowcast: 5.82%

Taylor Target Repo (i_target)

9.50% Prime: 13.00%

Current SARB Repo: 7.25%

Projected MPC Action

+225 bps hike expected

Output Gap (GDP Engine): -0.45%

SARB MPC Decision Probability Distribution

Hawkish Tightening Bias

Probability of HIKE 84%

Probability of HOLD 14%

Probability of CUT 2%

12-Month Forward CPI & Repo Path (Baseline vs. Shocked Scenario)

Dynamic mean-reverting impulse response over 12 months (M+1 to M+12)

Baseline CPI Path Shocked CPI Path

DYNAMIC FACTOR MODEL

RSA High-Frequency Real GDP Nowcasting Engine

Real-time factor extraction across Manufacturing PMI, Eskom Generation (GWh), Transnet Port Container Volumes, and Real Retail Trade Sales

1. ABSA / BER Manufacturing PMI (Index) 51.4 pts (Weight: 28%)

38.0 (Contraction) Neutral: 50.0 62.0 (Expansion)

2. Eskom Dispatchable Generation (GWh / Week) 4,420 GWh (Weight: 32%)

3,600 GWh (Stage 6 Load-Shedding) Norm: 4,500 GWh 5,200 GWh

3. Port Container Throughput (Durban/CT TEUs/wk) 81.5k TEU (Weight: 20%)

55k TEU (Congestion) Target: 84k TEU 110k TEU

4. Real Retail Trade Sales Growth (YoY %) +2.1% (Weight: 20%)

-6.0% YoY Trend: +1.8% +8.0% YoY

GDP QoQ (SAAR) +1.64% Seasonally Adj. Annualised

Real GDP (YoY) +1.42% SARB Potential y*: 1.87%

Output Gap (y_t - y*) -0.45% Feeds Taylor & Phillips

Dynamic Factor Proxy Contribution to QoQ Growth

Note: Adjusting the GDP proxies above dynamically updates the Output Gap (y_t - y*) used in Tab 1 and Tab 3's SARB Taylor Rule reaction function.

STATE-SPACE KALMAN FILTER

Live Stats SA Release Calibration & Historical Error Backtest

Kalman-gain weight updating: ŵ_{k,t} = w_{k,t-1} + K_t · (Actual_StatsSA - Nowcast_Smith)

12M Backtest MAE: 0.118% (11.8 bps)

12M Backtest RMSE: 0.146% (14.6 bps)

Ingest New Official Stats SA Monthly Print

Enter an incoming official Stats SA CPI print to execute a state-space Kalman filter step and re-anchor COICOP division weights.

Official Stats SA CPI (%)

Kalman Gain (K_t)

Innovation Error (Actual - Nowcast): -0.170%

Updated Food Weight (ŵ_01): 17.140%

Updated Fuel Weight (ŵ_07): 14.350%

Updated Core Weight (ŵ_core): 68.510%

12-Month Historical Backtest: Smith Nowcast vs. Official Stats SA CPI

Monthly pre-release Smith Nowcast compared against subsequent Stats SA actual releases

STATS SA EXPENDITURE DECILES

Income-Decile Inflation Divergence & Living Cost Inequality

Structural divergence between staple-heavy low-income deciles (Deciles 1–3) and service/credit-weighted high-income deciles (Deciles 8–10)

Plutocratic Gap (D1–3 vs Headline): +4.13% (+413 bps)

Deciles 1–3 Low-Income

Low-Income Living Cost Index

8.53% YoY

Food & NAB 48.0%

Minibus / Commuter 18.0%

Prepaid Elec / Water 14.0%

Other / Residual 20.0%

Deciles 1–3

Deciles 4–7 Middle-Income

Urban Middle Living Cost Index

7.92% YoY

Food & NAB 24.0%

Fuel & Commuter 16.5%

Housing & Utilities 25.5%

Credit, Health & Other 34.0%

Deciles 4–7

Deciles 8–10 High-Income

High-Income Living Cost Index

7.27% YoY

Food & Groceries 8.5%

Private Fuel & Vehicles 12.0%

Housing & Bond Service 24.0%

Medical (16%) + Tuition/Ins (39.5%) 55.5%

Deciles 8–10

Effective Inflation by Expenditure Decile vs. Official CPI

Side-by-side comparison against Stats SA Headline (4.40%) and Smith Nowcast (5.82%)

Key Insight: Because maize, bread, oil, and prepaid electricity carry a 62% combined weight in Deciles 1–3, low-income households experience +413 bps higher inflation than official Stats SA Headline CPI.

PERSONAL WALLET ENGINE

Personalized Household Inflation & Repo Exposure Calculator

Adjust your monthly expenditure shares (automatically normalized to 100%) to compute π_personal = ∑ w_j · π_j and home bond interest rate sensitivity

Monthly Budget (ZAR):

1. Food & Groceries (π = 9.4%) 22.0%

R7,700 / mo Normalized Share

2. Fuel & Transport (π = 7.2%) 14.0%

R4,900 / mo CEF Pump & Taxi

3. Rent / Home Bond (π = 6.1%) 28.0%

R9,800 / mo Repo Exposure > 25%

4. Electricity & Utilities (π = 12.7%) 11.0%

R3,850 / mo Eskom + Municipal

5. Medical Aid & Ins. (π = 9.6%) 15.0%

R5,250 / mo CMS Tariff Escalation

6. Discretionary / Leisure (π = 4.2%) 10.0%

R3,500 / mo Core Services & Retail

Personal Effective Inflation Rate (π_personal) Normalized Σw_j = 100%

8.04% R2,814 / mo inflation cost

Your personal inflation is +3.64% higher than Stats SA Headline (4.40%)

Formula: π_personal = ∑ (w_j · π_j) vs Smith Nowcast: +2.22%

HIGH REPO EXPOSURE (>25%)

SARB Repo Rate Debt-Service Sensitivity

Prime: 10.75%

Your Rent/Home Bond allocation (28.0% · R9,800/mo) exceeds the 25% credit threshold. Immediate monthly instalment impact of a ±50 bps SARB MPC shift on your implied 20-year Prime-linked facility:

+50 bps SARB Hike (7.75%) +R318 / month

-50 bps SARB Cut (6.75%) -R312 / month

FMCG PACKAGE & BRINE AUDITOR

Retail Shrinkflation & True Unit-Price Spread Ledger

Detects hidden unit-cost inflation across 6 core South African staples by normalizing package weight/volume downsizing and IQF poultry brine injection

Avg Nominal Sticker YoY: +4.85%

Avg Real Unit Inflation: +13.42%

RSA Consumer Staple	Nominal Shelf Price	True Unit Price	6M Volume / Yield Delta	Sticker Price YoY	Real Unit Inflation	Shrinkflation Diagnostics

THE SMITH MACHINE