Triangular arbitrage: Difference between revisions

Revision as of 03:08, 14 July 2011

Revision as of 03:30, 14 July 2011

Line 27:

==Evidence for triangular arbitrage==

Research examining high-frequency exchange rate data has found that [[mispricing]]s do occur in the foreign exchange market such that executable triangular arbitrage opportunities appear possible.<ref name="Fenn et al. 2009">{{Cite journal | title = The Mirage of Triangular Arbitrage in the Spot Foreign Exchange Market | journal = International Journal of Theoretical and Applied Finance | volume = 12 | issue = 8 | date = 2009 | pages = 1105-1123 | author = Fenn, Daniel J. | author2 = Howison, Sam D. | author3 = McDonald, Mark | author4 = Williams, Stacy | author5 = Johnson, Neil F. | url = http://www.worldscinet.com/ijtaf/12/1208/S0219024909005609.html | accessdate = 2011-07-06}}</ref> However,In mostobservations of thesetriangular apparentarbitrage, the constituent exchange rates have exhibited strong [[correlation]].<ref name="Aiba et al. 2002" /> A study examining exchange rate data provided by [[HSBC Bank (Europe)|HSBC Bank]] for the [[Japanese yen]] (JPY) and the [[Swiss franc]] (CHF) found that although a limited number of arbitrage opportunities typicallyappeared onlyto exist for 1as ormany 2as 100 seconds, 95% of them lasted for 5 seconds or less, and potentially60% onlylasted yieldfor very1 second or less. Further, most arbitrage opportunities were found to have small profitsmagnitudes, (usuallywith less94% thanof JPY and CHF opportunities existing at a difference of 1 [[basis point]], which translates into a potential arbitrage profit of $100 USD on aper $1 million USD trade)transacted.<ref name="Fenn et al. 2009" />

There are variations in the number of triangular arbitrage opportunities that occur during different hours of the 24 hour foreign exchange trading day. Perhaps slightly counter-intuitively, more triangular arbitrage opportunities arise during hours when [[market liquidity]] is at its highest; for example, from about 8-10 AM GMT when both Asian and European foreign exchange traders are active, and from 2-4 PM GMT when both European and American traders are active. During more liquid periods, triangular arbitrage opportunities also tend to exist for shorter durations than during less liquid periods. The reason for these differences is that in liquid periods the bid-ask spread is narrower and prices move around at a higher frequency due to the large volume of trading. This results in more price misalignments and thus more potential arbitrages. The high trade frequency, however, also ensures that the mispricings are quickly traded away and thus that any arbitrage opportunities are short-lived.

Using data from [[HSBC Bank (Europe)|HSBC Bank]], researchersResearchers have shown a decrease in the incidence of triangular arbitrage opportunities from 2003 to 2005 for the [[Japanese yen]] (JPY) and [[Swiss franc]] (CHF)and whichhave canattributed bethe explaineddecrease byto broader adoption of electronic trading platforms and trading algorithms during the same period. Such electronic systems have enabled traders to trade rapidly and react hastily to changes in price. The speed gained from these technologies improved trading efficiency and the correction of mispricings, allowing for less incidence of triangular arbitrage opportunities.<ref name="Fenn et al. 2009" />

Mere existence of triangular arbitrage opportunities does not necessarily imply that a trading strategy seeking to exploit currency mispricings is necessarily profitable. Electronic trading systems allow the three constituent trades in a triangular arbitrage transaction to be submitted very rapidly. However, there exists a delay between the identification of such an opportunity, the initiation of trades, and the arrival of trades to the party quoting the mispricing. Even though such delays are only milliseconds in duration, they are deemed significant. For example, if a trader places each trade as a [[limit order]] to be filled only at the arbitrage price and a price moves due to market activity or new price is quoted by the third party, then the triangular transaction will not be completed. In such a case, the arbitrageur will face a cost to close out the position that is equal to the change in price that eliminated the arbitrage condition.<ref name="Fenn et al. 2009" />