Arbitrage Lives in the Friction

The Arbitrage Nobody Sees on Their Screen

Stop Watching the Spread. Start Mapping the Friction

Everyone trades the obvious pairs and earns almost nothing. The real arbitrage sits in multi-leg routes through capital-controlled markets, where most people never bother to look.

Most people who chase arbitrage are staring at the wrong screen.

They open two exchanges, find a price gap, and assume the gap is the opportunity. It is not. The gap is the symptom. The opportunity is the reason the gap exists in the first place, and that reason is almost always friction.

Arbitrage exists because two markets price the same thing differently, and because something stops those prices from snapping back into line. Remove that something and the gap closes in milliseconds. Gold trades at a different price in London and New York for about as long as it takes an algorithm to notice, and then equilibrium is restored. Same with efficient crypto pairs. Same with liquid FX.

So the gap is not the prize. The cost of closing it is.

What the cost actually is

That cost wears different clothes depending on the market. Sometimes it is physical, the price of moving a thing from one place to another. A laptop that costs 500 pounds in London is more expensive in Singapore because someone has to ship it, clear it, and carry the risk in between. Sometimes the cost is energy, meaning time, effort, and the sheer complexity of pulling off the execution. Sometimes it is market friction, thin liquidity and wide spreads. And very often, in the markets worth caring about, it is legislative: capital controls, compliance walls, banking restrictions.

If none of those existed, arbitrage would not exist either. The whole game reduces to three questions. What is the arbitrage. How do you find it. How do you execute it and scale it.

The example that exposes the trap

Here is the example I use to break people of the screen-watching habit.

Ethereum is trading at 2,000 dollars on Coinbase in the US. On an Indian exchange the equivalent price is 2,085. You buy at 2,000, you sell at 2,085, you pocket an 85 dollar spread, roughly 4.2 percent. On paper you are done.

Except you are not. You sold in India, so you are now holding rupees. The trade only counts when you can bring that value back to dollars and run the cycle again. That return leg is where the energy sits. That is the actual problem, and it is the reason the spread existed at all. Anyone can find the gap. Closing the loop is the work.

Capital and cycling, the part people skip

Before you touch a single trade, you have to answer two operational questions: how much capital you are deploying, and how you are cycling it.

Say you have 100,000 dollars. You could run it as one cycle of the full amount per day, or you could run 10,000 dollars ten times over. Those are completely different operations with different risk profiles, even though the headline number is identical.

Now add time. A realistic cycle, once you factor in banking hours, liquidity windows, and waiting on counterparties, often takes one to two hours. If a single cycle takes two hours, ten cycles is twenty hours of work, which is not a business, it is a breakdown. So you optimize. You decide capital per cycle, cycles per day, and time per cycle, and you build around what is actually feasible. This is operational arbitrage. It has very little to do with the clean theory people quote.

Why the obvious pairs pay nothing

Most people look at simple pairs. Rupees to USDT. Dirham to USDT. These markets are efficient, which means the arbitrage that exists is roughly equal to the cost of the friction. You do the work and you keep almost none of it.

The opportunity shows up in multi-leg routes. Dirham to USDT to rupees. Rand to USDT to dollars to euros. Central African francs to USDT to euros to dollars. Every leg you add introduces a fresh inefficiency, and those inefficiencies stack. That stack is where the meaningful money lives. It is also invisible on any single exchange screen, which is exactly why it survives.

Where to actually look

You are not looking for countries. You are looking for corridors. Trade corridors, remittance corridors, capital-restricted economies, markets where liquidity sits unevenly on the two sides.

A few things are almost always true. USDT is somewhere in the chain. The US dollar is somewhere in the chain. And China is somewhere in the background, even when you cannot see it directly.

Then you stop thinking about single markets and start thinking about systems. Bolivia touches Argentina and Brazil. Brazil opens onto dollar markets, Argentina onto euro markets, and Bolivia has its own financial threads running toward Switzerland. Suddenly you are not looking at one country, you are looking at a node in a network. If you can move value through Bolivia into Argentina into Brazil into Switzerland and back to dollars, you may have built something compounded that no screen will ever show you. But you only see it if you map the flows yourself.

Finding the flows

None of this is sitting in a tidy report waiting for you. You assemble it.

You check Binance across spot, P2P, and express. You compare the XE headline rate against the rate people actually transact at, because those two numbers are rarely the same. You hunt down local exchanges. You ask local traders the only question that matters: where do you buy, and where do you sell. Sometimes the answer rearranges your entire route. Someone tells you, we do not use Bitcoin here, we use XRP, or only USDT, and the path you planned no longer applies.

Then you read around it. Forums, Reddit, Bloomberg, local-language platforms where the real conversation happens. AI tools can help you process the pile once you have it, but the signal has to exist somewhere first, and finding it is a human job.

I also recently wrote an article on Binance’s P2P Trading and the Debanking Risk it carries.

White, gray, and black

Every route has to be classified honestly. Fully compliant is white. Semi-compliant is gray. Non-compliant is black. A lot of real arbitrage runs through gray. That is just the truth of these markets, and pretending otherwise gets people hurt.

Here is the way I think about the risk. Picture your capital running through a row of washing machines. As long as every machine keeps spinning, value keeps moving and the cycle pays. The moment one machine stops, the money inside it is locked. If 10,000 dollars freezes mid-cycle, you cannot redeploy it, and the whole system slows down behind it.

The machines most likely to seize are the transition points, where gray flow meets a regulated interface, where money hits a bank, where someone asks you to prove your source of funds. That is where most failures happen. Not in the spread. In the plumbing.

Treat due diligence as a science

Once you have an opportunity, take it apart completely. Down to the granular level.

Know the limits: per transaction, per day, per week, per month, crypto versus fiat, resident versus non-resident. Know the infrastructure: which exchanges, which banks, which on-ramps and off-ramps. Know the regulatory friction: travel rule delays that can run 48 to 72 hours, wallet restrictions, the triggers that wake up a compliance desk. Know the cost structure: trading fees, spread, conversion cost, deposit and withdrawal fees.

Do not draw a straight line across a map and call it a route. Real execution is step by step, with exact movement instructions and defined checkpoints. Break every cycle into the smallest pieces it will tolerate, and for each piece write down the cost, the time, the risk, and what it depends on. Nothing stays abstract.

Then build the risk table. Probability of a bank freeze. Delay scenarios. Source-of-funds challenges. Counterparty risk. If funds get frozen, you might wait fifteen days, produce documentation, and still not know the outcome. If your source of funds is weak, the whole structure collapses. The more honestly you map this, the more stable and profitable the operation becomes. The mapping is not paperwork. It is the edge.

How to actually execute

Do not concentrate. A sane structure might be 5,000 dollars of exposure per market, across five markets at once, with total exposure capped at 25,000. If one market fails, the system survives. That is the entire point of spreading it.

Do not overuse a route, either. Trade Tanzania today, then step away for a few days and rotate into Guatemala, Chile, Mozambique, and come back later. Markets degrade when you lean on them too hard. People notice patterns. Liquidity tightens. So you rotate.

And you do not do this alone. The partner model is not optional. You need someone local who handles the in-country flow and someone external who brings capital and infrastructure. Usually the local partner runs the ground game, you run the capital and the routing, and you split the profit. Sometimes they bring the capital and you execute, and the split adjusts to match. Trust gets built slowly, on small cycles, and then you scale. Walk into Mozambique with no partner and you will fail. Try to copy a route everyone already knows and you will be standing in line with thousands of people doing the same thing.

The economics, without the fantasy

The math is simple once the operation is real. Two percent per cycle on 100,000 dollars of daily turnover is 2,000 dollars a day. That is 20,000 in ten days and 60,000 in a month. Those numbers are reachable, but only with discipline, real diversification, and constant route discovery. The day you stop discovering new routes is the day the returns start dying.

A tip worth more than the rest

Go look at the P2P market on Binance. Read the feedback. Look at who the buyers and sellers actually are and what payment methods they used. Google a payment method and it will tell you which country it belongs to.

Do that for a while and a pattern jumps out. Certain players are operating in more than one market at once. Someone in Bangladesh and Pakistan. Someone in India and Nepal. Someone in the UAE and Jordan. Someone in Turkey and Azerbaijan. You will keep seeing two or three countries clustered around the same trader, and the right reaction is to ask why.

The answer is that they have found a liquidity pool that trades in both directions, and they are working it. That is the whole secret, sitting in public, in the feedback section. Use it.

Research makes the difference. The five Ps still apply: prior planning prevents poor performance.

This is a working framework, not a theory. If you want to go deeper into a specific corridor or build a structured model around one of these routes, that is the next conversation. For now, stop watching the spread and start mapping the friction. The money was never in the gap. It was always in the cost of closing it.

About Faisal Khan

Faisal Khan has spent over 30 years working in cross-border payments, banking access, remittances, stablecoins, money transmitter licensing, and MSB-friendly banking. He runs Faisal Khan LLC, a consultancy that helps companies navigate the hard parts of regulated financial services: getting banked, getting licensed, structuring compliant payment flows, and finding the right partners to make it all work.

He is the founder of DealHarbor, a marketplace for regulated financial services deal flow, and The Money Wiki, a knowledge platform covering banking, payments, licensing, and financial systems globally.

If something in this article is relevant to what you are working on, book a free 15-minute call. No pitch, no obligation. Just a straight conversation about whether there is something worth exploring together.

Book a free 15-minute discovery call

faisalkhan.com