Why do these traders get billions to play with, unchecked?
Because they're supposed to be hedging every bet
One of the most misunderstood concepts in all of finance is that of “arbitrage”. It looks very much like speculation from afar but it isn't, it's very much the opposite. There are also a lot of people who describe what they do as arbitrage and they're damned liars: they're speculators. Ivan Boesky used to claim to be an arbitrageur but he wasn't even a speculator, he was an insider trader.
Back to basics
Arbitrage, pure arbitrage, is not a bet or a speculation. It's nailing down a sure thing. Forget finance for a moment – and think bookies and the gee-gees. A bookie will always have a balanced book: whichever horse comes in, the house wins (same with casinos). They do this by changing the odds dependent upon where the weight of money is going.
A bookie who runs his book because of his own views on which horse will win, rather than simply considering the maths of how much is at stake and where, is running an unbalanced book and will likely come a cropper. He is speculating on the outcome of the race, not conducting arbitrage among other people's views and bets on such.
In finance, arbitrage is buying and selling in different markets at the same time to lock in those small profits that uneven price movements can offer. An example: BP shares are traded in both New York and London. Over there in dollars, over here in pounds. So, maybe someone buys lots of BP shares in London: the London price moves up. The New York price hasn't yet moved so if you're quick, you can buy in New York, sell in London and make a penny or two per share.
Similarly, perhaps the dollar/pound rate changes so that what were the same prices for BP in the two places are now not. Buy and sell accordingly. But note that these things happen at the same time: this is the definitional point about arbitrage. There's a pricing mismatch, so buy and sell to lock in that gain and book that tiny percentage profit.
We can take this a stage further and it's still arbitrage. Many financial instruments are close substitutes for each other. There's a predictable relationship between the price of gold mine shares, gold futures, gold options, and the underlying price of physical gold. If the latter changes then you (well, no, not you or me, but these Masters of the Universe) should be able to predict where all the other prices will go. While they're not there yet, buy and sell again to lock in those infinitesimal profits.
A slight detour: Exchange Traded Funds. These are a relatively new development. Think of them perhaps as unit trusts on steroids. A unit trust holds a great big sack of investments: they're a way for us unsophisticated investors to be able to hold such a great big bag chosen for us by experts.
Unit trusts might specialise: in developing economies, in UK stocks, in large UK companies, UK SMEs, etc. ETFs, generally speaking, are similar but on drugs. You might have an ETF that purely tracks the price of gold. It might be structured to be a “short”, ie, the value of the ETF goes up when the gold price falls. As far as we're concerned they're a black box that grants us, by simply buying or selling them on an exchange, access to the complex trading strategies of the big boys on currencies, commodities, rates and stock indices – that great panoply of exotica that the markets provide for our delectation.
And ETFs are complex, should have a pre-determined relationship with an underlying price and, joyously, they publish what they've got within that black box. So, if you're an arbitrageur, you've got a lovely set of opportunities here.
Ninja-like speed needed in buying and selling
ETFs tend to be fairly lightly traded: not by design, just because there are thousands of the damn things and they're just going to be less traded than those main indices or markets that they're based upon. Which is great for arbitrage because lightly traded means prices move slowly. So, if the gold price moves you might have a minute or two to arbitrage between that price and the prices of a gold ETF or two.
Say gold moves up. You buy the gold ETF before that price moves at the same time selling gold itself. You've now locked in your profit, which is the difference between the price that has moved and the one that hasn't yet. Of course, your actions have just moved that ETF price to or closer to where it ought to be which is the service that arbitrage does to markets as a whole. It keeps prices in line across markets and among substitutes by trading when they get out of line.
This is how Ricardo's Iron Law of One Price works: people buy and sell stuff, move it around, so that prices are the same everywhere for the same item.
Billions and billions to play with
This arbitrage is risk-free: you're buying and selling at the same time so you're “hedging”. It also has tiny profit margins attached to it. Fractions of a penny on moving tens or even hundreds of pounds worth of bits and bobs.
So, because it's risk-free, those on arbitrage desks get given lots and lots of money to play with. It wouldn't be out of order at all for a senior trader to have access to $10bn, say, to play with. They need these sorts of sums to do their work. You're making maybe a few hundred quid every time you move a million across one of these price disparities. It takes a few days, possibly even a week or two for the money to come back after everyone's settled up and paid.
It all only works if you've got tonnes of money at your fingertips and if you stick rigidly to doing that risk-free arbitrage.
Which brings us to the great temptation of every arbitrageur. With access to all that cash, with all those fabulous trading tools and computers and screens and methods of gambling, well, surely I'm smarter than all those other people?
I'm going to start speculating: I'll not put the hedge on every time. So, for example, instead of playing the price differences between Nikkei (the Japanese equivalent of Dow Jones 30 or the FTSE 100) futures in Singapore and the Nikkei itself in Japan, I'm going to start making bets on the direction of the Nikkei itself. And I'm going to do this with the huge budgets that I'm given.
That's what the rogue trader Nick Leeson  did and he broke Barings when he did so.
Now of course there are systems in place to stop people doing this: but if you came to trading from the back office, where those systems exist and are devised, you may well know how to beat those systems. As Leeson did. The reason those systems exist is because every banker, however grasping and venal, knows that speculation is different from arbitrage.
With speculation we know that people are taking risk, that's the whole point of it. So we limit how much they are allowed to risk: the nightmare of this sort of investment banking is the thought that the bloke you thought was doing arbitrage, with the budget allowed for arbitrage, was actually speculating, taking the risk that you most certainly do not want being taken with that amount of money.
That line between arbitrage and speculation is really quite important. ®