1. Data sources
Auto-loaded from /dataKeep these three files in the site'sdata folder. No visitor upload is required.Retail prices
data/delhi_petrol_diesel_prices.xlsxColumns: Date, Petrol Price (Rs/Litre), Diesel Price (Rs/Litre)
Loading…
USD/INR
data/USD_INR Historical Data.csvColumns: Date, Price (INR per USD)
Loading…
Indian Basket crude
data/crude_price_clean.xlsxColumns: Year, Month, Price ($/bbl)
Loading…
Input assumptions
Retail is assumed to be ₹/litre. Indian Basket crude is assumed to be US$/barrel. USD/INR means rupees per US dollar. A barrel is fixed at 158.987 litres.
The calculated “wedge” is not tax alone. It can include refining economics, freight, marketing margin, dealer commission, taxes/cess, inventory effects, and timing differences.
Headline metrics
Interpretation: compare retail with crude/FOB in the same unit, then use the change and lag views. A flat retail price while crude falls increases the retail-minus-crude wedge; that does not by itself prove a tax change.
Price levels in the same unit
Petrol vs international benchmarks
Diesel vs international benchmarks
Indexed comparison — base = 100
Best for seeing whether retail prices move proportionately with crude/product benchmarks despite different levels.
Retail minus benchmark wedge
Petrol wedge
Diesel wedge
Change pass-through
Horizon changes
Change ratio — diagnostic
Ratio = retail change ÷ benchmark-equivalent change over the chosen horizon. Very small benchmark moves are suppressed because they create unstable ratios. Use regression beta for the more robust pass-through estimate.
Lag sensitivity
Petrol: retail changes vs lagged benchmark changes
Diesel: retail changes vs lagged benchmark changes
Report-ready observations
Automatic summary
Methodology notes
Monthly mode: retail uses the effective month-end price (latest available Delhi observation on or before month-end), which safely carries unchanged administered prices across months. FX and optional daily FOB observations are averaged by calendar month; crude monthly data is averaged if multiple observations exist. Daily mode: every retail date is kept; FX and benchmarks are matched to the latest available observation on or before that date. A monthly crude benchmark used in daily mode should be interpreted as a monthly-period benchmark, not a real-time daily crude quote. Lag regression: OLS of 1-period retail price changes on benchmark-equivalent price changes, tested from lag 0 to the selected maximum. Beta measures unit-for-unit pass-through; R² measures explanatory fit. This is descriptive, not causal.