Fuel Impact Forecast — Methodology

Companion to: fuel-impact-2026-05-16.html Generated: 2026-05-16 by build_report.py (fuel-impact-report skill, MC-3483) Purpose: step-by-step record of every input value used — verify the math from this doc alone.

1. Data Sources

#	Source	File	Last read
1	IATA Jet Fuel Price Monitor	`~/Dropbox/Safair/Reporting Data/Data Fuel.xlsx`	2026-05-16
2	B4i Fuel Feedback (Bernd Feucht)	`~/PKA/SecondBrain/wiki/concepts/Fuel-Prices-B4i.md`	embedded as base64 in HTML
3	Radixx route surcharge	`~/PKA/SecondBrain/wiki/concepts/fuel-surcharge-data/surcharge_history.json`	embedded as pivot JSON in HTML
4	FinDash kpi_data	`~/PKA/Projects/findash/data/kpi_data.json`	2026-05-16
5	May MTD dashboard	`~/Dropbox/Safair/Reporting Data/dashboard/dashboard_data.json`	2026-05-16

2. Schedule Constants (only non-actuals in CONSTANTS)

Parameter	Value	Basis	As-of
MAY_FLIGHTS	5,031 flights	May schedule	2026-05-16
MAY_SEATS	950,475 seats	May schedule	2026-05-16

MC-3517 note: March + April estimate constants removed. All cost fields now read directly from kpi_data.json actuals (fuel_r, fuel_uplift, avg_fuel_per_liter_uplifted, total_costs, pbt_r). NON_FUEL_PER_FLIGHT derived at runtime from April actuals: (total_costs − fuel_r) / flights.

3. Mar Actuals (kpi_data.json, year 2026, months[2])

Metric	Value	kpi_data key
Flights	5,640	`flights`
Pax	952,408	`pax`
Seats	1,062,072	`seats`
Load Factor	89.7%	`load_factor`
Yield	R1,329	`yield`
Revenue	R1,233.5M	`revenue_r`
Fuel uplift	28.613M L	`fuel_uplift` — ACTUAL
Fuel R/L delivered	R14.62	`avg_fuel_per_liter_uplifted` — ACTUAL
Fuel cost	R418.4M	`fuel_r` — ACTUAL
Non-fuel cost	R717.8M = total_costs − fuel_r	derived
Total costs	R1,136.3M	`total_costs` — ACTUAL
PBT	R127.4M	`pbt_r` — ACTUAL
Reconciliation check	revenue − fuel − non_fuel = R97.2M vs pbt_r R127.4M	gap R30.22M

4. Apr Actuals (kpi_data.json, year 2026, months[3])

Metric	Value	kpi_data key
Flights	5,307	`flights`
Pax	864,907	`pax`
Seats	999,135	`seats`
Load Factor	86.6%	`load_factor`
Yield	R1,628	`yield`
Revenue	R1,427.2M	`revenue_r` — ACTUAL
Fuel uplift	26.637M L	`fuel_uplift` — ACTUAL
Fuel R/L delivered	R26.52	`avg_fuel_per_liter_uplifted` — ACTUAL
Fuel cost	R706.5M	`fuel_r` — ACTUAL
Non-fuel cost	R683.5M = total_costs − fuel_r	derived
Total costs	R1,390.1M	`total_costs` — ACTUAL
PBT	R43.7M	`pbt_r` — ACTUAL
Reconciliation check	revenue − fuel − non_fuel = R37.2M vs pbt_r R43.7M	gap R6.56M
Non-fuel/flight	R128,797	(apr non_fuel_total) / apr flights — used for May
L/flight	5,019	apr fuel_uplift / apr flights — used for May

5. Recovery Analysis

Lens	Result
Per-pax yield recovery	R299 ÷ R378 = 79%
Per-flight rev recovery	R50K ÷ R59K = 85%
Network revenue ÷ fuel hike	R194M ÷ R288M = 67%
Total absorption (rev + nonfuel saving) ÷ fuel hike	R228M ÷ R288M = 79%

6. May Projection

6a. May Fuel R/L Derivation Chain (MC-3517)

Step 1 — April actual delivered R/L anchor:

Apr fuel_r = R706.54M  (kpi_data fuel_r)
Apr fuel_uplift = 26.637M L  (kpi_data fuel_uplift)
Apr actual R/L = 706.54M ÷ 26.637M = R26.5200/L

Step 2 — May fuel R/L derivation (M2 priority chain):

Priority 1 (preferred): May MTD actual avg_fuel_per_liter_uplifted from kpi_data.json
  → Used when present: NOT AVAILABLE — fell through to IATA estimate
Priority 2: Apr actual R/L × (IATA×FX latest) / (IATA×FX Apr)
Priority 3: Apr actual R/L × IATA ratio only (FX unavailable)
Priority 4: Apr actual × 0.97 (IATA unavailable; B4i-corroborated 3% ease)

Actual source used: Apr actual R26.52 × IATA ratio (162.89/167.96=0.9698). USD/ZAR not available — excluded from scaling. CAVEAT: monthly contracts lag IATA weekly spot.

April IATA $/bbl reference = 167.96
Latest IATA $/bbl = $162.89  (May 08, Data Fuel.xlsx)
May base R/L result = R25.7200

CAVEAT (M2): IATA-scaled estimates assume supplier contracts track IATA weekly spot. Monthly contracts lag IATA weekly; April delivered R26.52 rose while IATA fell in that period. A Rand move can flip the sign of the implied R/L change. This estimate is directional. Superseded by May MTD actual delivered R/L (priority 1) when kpi_data.json is updated.

Step 3 — B4i corroboration (direction check): May B4i tariff data shows broad decreases: Astron −332cpl, Shell −215cpl, PetroSA −171cpl, Engen −222cpl. This corroborates the downward IATA direction. Base R/L of R25.72 is anchored to actual April R26.52 minus documented easing.

Step 4 — Scenario band: Opt = base − R1.50 = R24.22 Base = R25.72 Pess = base + R2.00 = R27.72

6b. May MTD Signals

Metric	Value	Source
Period	May 2026 MTD (14 days, through 14 May)	dashboard_data.json
Flights	2222	dashboard_data.json
Pax	336517	dashboard_data.json
Revenue	R531.8M	dashboard_data.json
Yield	R1,580	dashboard_data.json
LF	80.2%	dashboard_data.json

6b-ii. Base Full-Month LF Projection Chain (MC-3517 follow-on)

May MTD LF is a partial-month observation (~14 days). Bookings continue to mature. The full-month projection uses April actual full-month LF as the maturity proxy.

May MTD LF        = 80.2%  (from dashboard_data.json kpis_mtd — partial, ~14 days)
April actual LF   = 86.6%  (from kpi_data.json months[3] load_factor — ACTUAL full month close)
Projection method = MTD 80.2% @14d → full-month 86.6% using April actual full-month close
                    as maturity proxy; assumes April-like booking maturation
Flag              = if May booking pace differs materially from April, this proxy
                    will over- or under-state full-month LF

Base — MTD actual vs full-month projected:

Variant	LF	Revenue	Fuel	Non-fuel	PBT	Margin
[live MTD]	80.2%	R1,204M	R649M	R648M	R-93M	-7.7%
[full-month proj]	86.6%	R1,300M	R649M	R648M	R3M	0.2%

6c. Schedule + Cost Inputs

Parameter	Value	Basis
Flights	5,031	May schedule
Seats	950,475	May schedule
L/flight	5,019	Apr actual fuel_uplift ÷ Apr flights
Non-fuel/flight	R128,797	Apr actual (total_costs−fuel_r)÷flights
Non-fuel total	R648M	5,031 × R128,797

6d. Three Scenarios (+ Base dual-line)

Scenario	Fuel R/L	Yield	LF	Revenue	Fuel	Non-fuel	PBT	Margin
Optimistic	R24.22	R1,628	82.2%	R1,272M	R612M	R648M	R12M	1.0%
Base case [live MTD]	R25.72	R1,580	80.2%	R1,204M	R649M	R648M	R-93M	-7.7%
Pessimistic	R27.72	R1,533	78.2%	R1,139M	R700M	R648M	R-209M	-18.3%
Base [full-month proj]	R25.72	R1,580	86.6%	R1,300M	R649M	R648M	R3M	0.2%

7. Reconciliation Assertions (A4 dual-gate)

Assertion: (revenue − total_costs) ≈ pbt_r. Gap = non-operating IS items (finance costs, tax, other income) not captured in total_costs. Warn gate: max(R15M, 1% revenue) — amber NOTE in report. Fail gate: R50M — red WARNING block in report.

Month	rev − costs	pbt_r	Gap	other_is	Warn gate	Status
Mar	R97.2M	R127.4M	R30.2M	R30.2M	R15.0M	WARN
Apr	R37.2M	R43.7M	R6.6M	R6.6M	R15.0M	PASS

7b. IS-Bridge Waterfall (A4 — sums exactly to fuel hike)

Identity: revenue_Δ − fuel_hike + nonfuel_saving + other_IS_movement = −pbt_drop

Revenue increase:     +R193.7M
Fuel hike:            −R288.1M
Non-fuel saving:      +R34.3M
Other IS movement:    R-23.7M  (apr_other_is=R6.6M, mar_other_is=R30.2M)
PBT drop:             −R83.7M
Bridge check:         R288.1M ≈ R288.1M

8. Caveats

March + April costs are ACTUALS from findash (fuel_r, fuel_uplift, total_costs, pbt_r). No estimates used.
Non-fuel/flight (R128,797) derived from April actuals. Held flat for May.
May fuel R/L anchored to April actual R26.52 × IATA easing ratio. B4i corroborates direction.
May L/flight (5,019) from April actual uplift ÷ flights. Stage-length mix ±2-3%.
Hedging not modelled — assumes spot pass-through.
Source of truth for constants: CONSTANTS dict in build_report.py.

Generated 2026-05-16 · build_report.py · MC-3483/MC-3517