Working paper · Optimal petroleum product reserves All research →
The southern haulage corridor at dusk A stylised wide steppe landscape: a low dusk sky over a flat horizon, a single straight road running to a vanishing point, and a line of coal-haulage trucks queued along it.
Research pitch

The reserve you need is a pricing rule, not a stockpile

The idea

Mongolia caps fuel at a fixed price. That single regulation — not the railway, not the border — decides how much fuel the state must stockpile. Index the cap and the requirement falls 61%.

Drag the rule

Move one regulation. Watch the reserve.

Reserve required against price-cap pass-through A line chart. The required reserve holds flat at about 35 days of national consumption while pass-through is below 0.5, falls sharply to about 23 days, and declines to 13.6 days at full pass-through. A shaded band marks the 11 to 18 days Mongolia actually holds. Held today — about 11–18 days Mongolia’s cap today Pass-through ρ — share of a cost shock reaching the pump Days of consumption

Reserve required
— days
Export-days
—
Stock
— kt
Private imports
— kt/mo
Computed at the observed August 2026 calibration — a 2,840 ₮/L retail cap, 3,599 ₮/USD, 800 USD/t landed, 5% duty, 10% VAT, excise suspended. Twenty-one solved equilibria; the line is the model’s output, not a fit. Export-days count days of coal haulage the stock sustains — the thing that actually stops when the fuel does.
0

Pass-through in the cap today. It is a fixed price, not an indexed one — so Mongolia sits on the flat part of the curve, where nothing it does to the reserve changes what it needs.

−61%

Fall in the required stock between a rigid cap and a fully indexed one. No new tank, no new railway, no new supplier.

32

Landlocked developing countries sized against a 90-day rule written for states that can call a tanker. Most of them cap fuel prices too.