I thought Eurokopek actually had a pretty good explanation. There is a limit to that basis since at some point it gets big enough to arb away by borrowing in one currency and lending in the other.
You can't arb this, as there is no hedge for the liquidity/banking system balance sheet factor... You can speculate and take advantage of the basis to obtain a certain return, but this will involve taking risk and thus will not be an arb. I can elaborate further, if you wish.
Well if it was possible to borrow USD at rate X in $ and lend GBP at rate Y in pounds you could borrow the $ at X, convert it at spot to GBP and lend that amount in pounds at Y and at the same time buy/sell the futures (or enter into a forward contract) that is off from what those lending rates implied. When the futures expired your exchange risk would be negated by the opposite positions you'd taken and you would have made more on the GBP loan (or the opposite if it was off in the opposite direction) than you paid on your USD borrow. The only reason that wouldn't work is if you for some reason couldn't borrow at X or lend at Y (with the appropriate tenor to match the future). Is that what you're saying is the case? In which case I wonder if you can't borrow at X why you would call that the interest rate? What am I missing here?
Gonna be a lengthy one... I hope I am able to convey my thinking with some clarity. Imagine that cash is precious and that, apart from all the "normal" risks you face when lending it to someone (e.g. counterparty risk), there is also the possibility you wouldn't have enough on hand to deal with any of your own issues that may arise. This fundamental principle, IMHO, is the underlying reason for the existence of all the different bases out there (tenor basis, x-ccy basis, etc). Since nobody can model or quantify liquidity, there is simply no good way in finance to deal with these effects, which is why they always come out as "fudge factors" that make all the traditional "no arbitrage" logic work. X-ccy basis is a specific manifestation of this effect, where there is a perception that USD cash is the most precious cash of all. Lending out USD cash makes you feel the most vulnerable and exposed, which means that you'd need to be paid extra to do it. Obviously, this, IMHO, is sort of the underlying basic principle. In reality, many things may affect the way you feel about lending out your precious dollars. For instance, if you were of the opinion that the supply of USD relative to all the other stuff will decrease in the future, you might feel more reluctant to lend your USD. Likewise, if you felt that you (and all your fellow mkt participants) are likely to have more demand for USD in the future (maybe due to increased regulation or fear of a genuine crisis), you'd also be less willing to part with your USD cash. Pre 2008 all such effects, while always present in theory, were happily ignored, because in a world of infinite banking system liquidity, the extra idiosyncratic value of USD cash was 0. So all these "arbs" were never really "arbs", but we were happy to pretend they were. Not any more.
If I can borrow USD at X% interest I can borrow USD at X% interest, right? If that's the case, what about the purely mechanical arbitrage I described doesn't work. If USD are more precious than that is reflected by a higher interest rate and it's not longer X, it's X+X1. But if we agree that the interest rate is X% and I can borrow at X%, and then lend at Y% in GBP and the forward exchange rate doesn't reflect that reality, why can't I enter into the transaction I described for risk free arbitrage profit? It's purely mechanical.
Obviously M will have to speak more to this, but that level of arb would, in my eyes, require a bot to quick to spot it and move it. I was under the impression that the mkt would equilibrate to quickly and close the gap. i.d. if you have noticed this, surely others have as well, which like any system, obs. change the system; thus, the equilibrium is reformed. I know arb exists. I am just saying that this sounds arcane than necessary.
You can enter into such a transaction, sure... But, as a trivial example, what if I told you that such a trade is going to be subject to an arbitrary tax assessed every day? Would you still consider it to be "risk free arbitrage"?
We agree, the market will almost instantly arb out any imbalance. Which is why I question this idea that there could be a persistent "basis" in the market that is being alleged.