The zero that saved Petrobangla 460 crore taka it never had to pay

A piece of news published by The Business Standard newspaper caught my eye yesterday — Govt secures 90% LNG premium cut from US firm Gunvor, approves 117 cargoes.

This was news for me. I have read about strange pricing in this market, but a supplier cutting its premium by 90% in five days was a new one.

Let's take a step back. Let me explain briefly what LNG is and how it is priced.

LNG and Pricing

LNG is liquefied natural gas. You take regular gas, freeze it into liquid at a liquefaction facility, load it onto specialized LNG carrier vessels, and ship it to countries that do not have enough gas of their own.

The big buyers sit in two blocks. One is Japan, South Korea and China. The other is Europe. This gave the world two benchmarks: JKM (Japan Korea Marker) for Asia, and TTF for Europe. When a country imports LNG, it benchmarks against whichever block it sits closer to. Bangladesh is in the JKM zone.

The gas itself is the cheap part. At Henry Hub, the American gas benchmark, gas trades at $2.81 per MMBtu right now. Freezing it and shipping it to Asia adds maybe another six to eight dollars. JKM is trading at $21. 

Everything above the cost stack, call it eleven dollars and change, is scarcity. In a shortage, the price of LNG detaches from what it costs to make and becomes whatever a desperate buyer will pay. Keep this in mind for later.

Deals come in two broad types.

Spot: you buy a cargo from the market for prompt delivery, priced as JKM + X. The X is a premium over the benchmark, covering distance, timing, and the seller's mood. (There are also short-term contracts of one to two years in between, but let's not overload.)

Long term: contracts running 5 to 20 years, priced off an energy index. In the Middle East that is usually crude oil:

A × Brent + B

In the US it is natural gas:

M × Henry Hub + N

A and M are factors, they decide how much the LNG price moves (up or down) with the index. B and N are constants, a fixed dollar amount stacked on top.

Now the Petrobangla deal

I will be referring to this table going forward.

As per officials, Gunvor's initial offer had two legs. The front leg, cargoes for 2026 to 2028, priced off the spot benchmark at JKM + $0.10, because gas is scarce right now and Gunvor wants the market price for those years.

The tail leg, cargoes from 2029 to 2038, priced at 122% × Henry Hub + $5.35, basically cost-plus American gas for the years when a wall of new supply is expected from Qatar and the US. In LNG, scarce years ride the market, glut years ride the cost, a sensible stucture.

On July 28, the CCEA (cabinet committee on economic affairs) gave in-principle approval for 78 cargoes. Between July 28 and August 6, Petrobangla negotiated with Gunvor, and out of that the deal grew to 117 cargoes with both prices lower: the tail down to 121% × HH + $5.20, and the front, per the official story, up tenfold to JKM + $0.875.

That file reached the CCGP (the purchase committee, the one that approves the actual money) on August 7 and died there. The Energy Division withdrew its own proposal from the meeting. No decision, no reason given. 

Five days later it came back and got approved, with the front premium collapsed to JKM + $0.0875. A 90% cut, extracted in five days by the master negotiators at Petrobangla ... from a seller holding every card in this market right now.

Global commodity trading doesn't follow the Kawran bazaar fish market school of negotiation tactic. Energy traders do not quote ten times above their own opening offer, and they definitely do not fold 90% in five days for nothing.

Look at the two numbers again. 0.875 and 0.0875. This likely is not a price movement. It's more of a zero going missing after the decimal.

Now look at what else changed between the withdrawn file and the approved one. The tail formula: identical. The front structure: identical, five cargoes in 2026, six in 2027, three in 2028, in both. The delivery window: identical.

Out of everything in that file, exactly one number moved between August 7 and August 12, and its two versions differ by exactly one decimal place. Negotiations usually move packages. But typos move... digits.

So what happened here?

The straightforward reading: Gunvor negotiated once, and that was before August 7. The trade was volume for price. Take 117 cargoes instead of 78, and both legs get cheaper. The tail from 122% + $5.35 to 121% + $5.20, and the front from JKM + $0.10 to JKM + $0.0875.

Why would Gunvor cut both legs? Because the discount was pocket change compared to what the volume offered them. The price cuts are worth roughly $63 million to Bangladesh over the life of the deal. The 39 extra cargoes are worth roughly $1.2 billion of revenue to Gunvor, on a cost-plus formula. 

Keep in mind, Gunvor is a trader, not a producer. It buys American gas on the same Henry Hub index it sells to us on, so the tail margin locks in the day the contract is signed. Giving up $63 million of rate card to book $1.2 billion of guaranteed volume is an easy yes for any trading desk.

Then, somewhere in the paper pushing between the negotiation table and the cabinet file, JKM + $0.0875 became JKM + $0.875. One zero gone, the premium up tenfold, nothing else touched. The public record cannot say whose paper dropped the zero, Gunvor's offer letter or Petrobangla's summary sheet.

But the withdrawal points at Dhaka. You pull your own file from a committee when you find your own error in it.

Once they caught it, they withdrew the file, fixed the digit, and resubmitted. Fair enough, mistakes happen. But instead of quietly correcting it, or even just hiding it, somebody briefed it out as evidence of newfound negotiation prowess.

Let's price the prowess anyway. The claimed saving, $0.875 down to $0.0875 over 14 cargoes, comes to roughly 460 crore taka. The movement against Gunvor's actual opening offer is about 7 crore taka. 

The rest was saved against a number that lived for five days in a withdrawn file. And even the inflated premium was barely 4% of a cargo's delivered price. A week of drama over the smallest line on the invoice.

Two more things for context. The deal is called G2G, government to government, though the only government on the other side is Gunvor's board. It is a no-tender, directly negotiated 13-year purchase, happening because Bangladesh promised Washington around $15 billion of US LNG back in February, while Qatar sits on force majeure delivering half its cargoes.

This is good diversification; you don't want all your required energy to pass through a single conflict-prone strait.

Is this whole Petrobangla-Gunvor story provable from the outside? No. But it is cheaply checkable. Government summary sheets may write amounts in words as well as in figures, and "zero point eight seven five" when spelled out cannot lose a zero. 

One look into the August 7 summary sheet settles it. So does Gunvor's offer letter. If that letter says $0.0875, there was never a price hike and never a 90% cut. There was one negotiation, one typo, and one press release.

The devil is in the decimals.

—

Taukir Aziz is a finance professional and a trustee of Panam Institute, a Dhaka-based think tank