What the copper curve taught me about patience

Six weeks ago I did not know what a warrant was. I won't pretend I fully do now. But I can tell you what the copper curve has been patiently explaining to me since June, which turns out to be less about metal than about time.

A metal has to be somewhere

A financial asset can be an entry in a ledger and nothing else. A tonne of copper concentrate has to physically be somewhere — in a warehouse, on a ship, in a smelter's queue — and somebody is paying, every day, for it to be there.

That single fact produces the curve. When the forward price sits above spot — contango — the market is effectively paying you to wait, covering the cost of holding the metal. When it sits below — backwardation — the market wants the metal now and will pay a premium to skip the queue.

So the curve isn't a forecast, exactly. It's the market's opinion about time, priced. Reading it as a prediction is the beginner's mistake, and I made it for about a fortnight.

What operations actually is

I sit between the traders and the books that keep them honest. Daily mark-to-market and hedging reports, reconciling physical positions against financial ones. Long/short maintenance on client accounts. Research on junior mining projects that might one day feed the concentrates pipeline.

The romantic version of a desk is a decision. The real version, from where I sit, is a reconciliation — and I've become slightly evangelical about this, because reconciliation is where you find out what the book is really made of.

The two trades

A few weeks in, an ad-hoc reconciliation surfaced two duplicate executions — 800 tonnes that existed twice in a place they should have existed once. Fixing it was a five-minute conversation.

The interesting part isn't that it was found. It's that nothing had gone wrong yet. Nobody had lost anything. It was simply a position that would have been wrong later, quietly, on a day when it would have been much more expensive to be surprised.

Most of the errors worth catching haven't cost anything yet.

I keep turning that over. Almost all the checking work I've been given has that shape: it's not detecting damage, it's removing the conditions for damage. Which is deeply unglamorous and, I now think, most of what a well-run book is.

On patience

Here's the thing four years of school prepared me badly for. The market opens and my instinct is to have a view. Immediately. Preferably an interesting one.

The desk's rhythm is nothing like that. A concentrate cargo takes weeks to move. An assay takes as long as it takes. Twice now I've watched a price action confirm something a trader said a week earlier, and what struck me both times wasn't the call — it was that they'd been willing to hold the view for a week without needing the market to agree with them in the meantime.

I'm not there. I still want the confirmation on the same day. But I've started noticing the wanting, which people tell me is the first part.

Six weeks, four things

• Ask what a number is for before asking whether it's right. Half the reports I touch have a decision behind them and the decision determines the tolerance.

• Physical markets have friction, and the friction is information. Freight, warehousing, assay timing — none of it is noise around the price. Some of it is the price.

• The person who reconciles the book is the person who understands the book.

• Say “I don't know” quickly. On a desk it costs nothing at 9am and a great deal at 4pm.

Ask me in six months and I expect most of this will read as naive, which is roughly the point of writing it down now. The curve, meanwhile, will still be there, still pricing the cost of waiting, entirely unbothered by anyone's need to be right this week. It's the least impatient thing in the building.