Arbitrage is the oldest trick in markets, and the least mysterious: the same thing, priced two different ways at the same time, and a trade that pockets the difference. It sounds like free money. It almost never is, and the reasons why teach you more about how markets actually work than almost anything else I know.
The core idea, in round numbers
Imagine the same shares at $100 on one exchange and $101 on another, at the same moment, in the same currency. Buy 1,000 shares where they cost $100 and sell 1,000 where they fetch $101, and you have $1,000 that appeared out of a disagreement between two prices. No view on the company. No waiting for earnings. Just a gap, and a trade that closes it.
That simplified picture is the whole idea. Everything else here is the fine print: why the gap exists, why it is usually gone before you can see it, and the shapes it takes when a normal investor runs into it, which most of us do every time we buy a US stock with Canadian dollars.
Why the gap only lasts seconds
Markets are not one place. The same company can trade on several exchanges, in two countries, in two currencies, and each venue prints its own price. Those prices should agree, and mostly they do, because an industry exists to make them agree.
Arbitrage firms park their servers in the same buildings as the exchanges' matching engines, a practice called colocation, and watch every venue at once. The moment a real gap opens, their machines buy the cheap one and sell the dear one until the gap is gone. The whole event can take less time than a blink. By the time a human sees a price difference on a screen, it is usually a stale quote, two prices in different currencies that only look different, or a gap smaller than the cost of grabbing it.
That last point is the retail reality. A one dollar gap on a small trade is not a dollar. The spread takes a piece, the commission takes a piece, and borrowing shares to sell the expensive side takes a piece. The one-day moves on my Trends page are what is left after this machinery has ironed the prices flat. Run the receipt yourself in the fee drag calculator on my Tools page: a $500 trade at a $9.95 commission needs a move of almost 4 percent, there and back, before you make a single cent.
The version Canadians meet every day: two listings, one company
Dozens of Canadian companies are interlisted: the same shares trade on the TSX in Canadian dollars and in New York in US dollars. Shopify is the famous example, and the big banks and miners do it too. One company, one claim on the same business, two prices in two currencies.
Divide the New York price by the exchange rate and you land on the Toronto price, roughly. When the two drift apart, it is usually the currency that moved, not a bargain appearing. Arbitrage desks spend all day nudging the two back into line, and they are very good at it.
The lesson for a Canadian investor is less glamorous and worth more money: the gap you actually pay is not between the two listings. It is the 1.5 to 2.5 percent your broker quietly takes when it converts your dollars on the way in, and again on the way out. That spread is bigger than most arbitrage gaps, and it is the closest thing to an arbitrage most of us will ever fund. With our own money, from the wrong side.
Norbert's gambit: the retail cousin
There is a well known Canadian workaround, named after Norbert Schlenker, the British Columbia financial planner who popularized it. You buy an interlisted stock in the currency you have, have your broker journal the shares over to the other exchange's listing, and sell them in the currency you want. Instead of the broker's 1.5 to 2.5 percent conversion spread, you pay two commissions and the bid and ask spread. On a few hundred dollars that trade is not worth it. On $10,000, the difference between a 2 percent spread and roughly $20 of commissions is about $180, and the arithmetic only improves as the amount grows.
It is arbitrage logic used defensively: not profiting from a gap, just refusing to pay someone else's. My FX cost tool puts the three prices side by side: the bank spread, a cheaper conversion service, and the gambit at a flat cost.
The honest fine print: the gambit takes a few days at most brokers, the stock's price can move while you wait, the US dollar amount you end up with is not locked in until the second sale settles, and not every broker or account type allows it. It is a known technique that I am describing, not a recommendation to use it. Whether it fits your account is a question for your broker.
The invisible version inside every ETF
If you own an index ETF, arbitrage is working for you every day without sending you a bill. An ETF is a basket of stocks wrapped in a single ticker, with one special feature: a small club of large firms, called authorized participants, can create new ETF units by delivering the underlying basket to the fund, and redeem units by handing them back.
So the moment an ETF trades above the value of what it holds, the participants buy the holdings, swap them for fresh units, and sell the units into the premium. The moment it trades below, they run the film backwards. That is why an ETF's price hugs the value of its holdings all day, in a way no single stock is forced to. The exceptions prove the rule: when markets are stressed, or the underlying stocks trade in time zones asleep while the ETF is awake, the gap can widen for a while, pricing real uncertainty rather than offering a gift.
Merger arbitrage: getting paid to wait, and to worry
When a takeover is announced at a fixed cash price, say $50 a share, the target rarely jumps all the way to $50. It might sit at $48.50 for months while regulators review it, lenders fund it, and shareholders vote. That $1.50 difference is called the spread, and it is not a pricing mistake. It is the market's price tag on the risk that the deal breaks.
Merger arbitrage funds make a living collecting those spreads, deal after deal. Most deals close, and each one pays a little. But when a deal does break, the stock usually falls back toward its pre announcement level, and one broken deal can hand back the profits of many small wins. The spread was never free money. It was wages for carrying a risk somebody had to carry.
The institutional version, in one breath
The biggest version of the trade is the one between an index and its futures. S&P 500 futures trade nearly around the clock; the index itself is 500 stocks that trade during market hours. The difference between the futures price and the cash index is called the basis, and most of it is just interest and expected dividends over the contract's life. Desks trade that gap constantly, in enormous size, which is a large part of why futures and the index already agree with each other at the morning open. Same idea as the two listings, same idea as the ETF, bigger numbers and faster machines.
So is it free money?
No. A gap is usually small for a reason: part fees, part the cost of borrowing shares to sell the expensive side, part timing, since the two legs do not always fill at the prices you saw. And part of it, in the merger case, is honest pay for a risk that occasionally collects.
What arbitrage actually pays is a wage. A wage for speed, paid to firms spending fortunes on servers and cables to be a microsecond faster. A wage for capital, paid to whoever keeps money parked and ready to move the second a gap opens. And a wage for risk, paid to whoever holds the bag while a deal or a market makes up its mind. Nobody earns it by noticing something clever on a screen at home.
Here is what it teaches a regular investor, and it is the part I keep coming back to. Prices disagree far less than people think, because an entire industry is paid to make them agree. In my Friday Close polls the crowd votes on where prices land; the arbitrage desks vote with billions, every second, and their votes are why the Toronto and New York prices of the same company almost always match to the penny once the currency is counted. For the rest of us, the useful move is not chasing gaps we cannot reach. It is controlling the costs we can: the currency spread, the commissions, the fees that compound quietly. Patience and low fees beat chasing gaps, every week of the year.