FT: Technology can help stabilise currency markets
You may not be aware that when foreign currency traders make most of their deals during the day, they are essentially doing so with “free” money. That is to say, they buy or sell currencies many times during the day, but only pay or receive interest on their closing overnight balance. The UK’s Financial Times suggests that if, instead of interest being charged daily on closing balances, interest is charged real-time, e.g. by the minute or second, then currency traders would trade less often and speculate far less, thereby reducing currency fluctuations.
Historically, real-time interest was not feasible because of the technology challenges (number of transactions and an inability to calculate interest, charge it, and pay it quickly enough). That no longer applies.
It’s an interesting twist - the technology that enabled the wildfire that is modern FX trading (where speculative transaction volumes are many many times greater than “real trade” transactions) could now be used to damp it down. But it would take governments and central banks to drive such a change. I can’t see trading banks doing it - using the “free” float is a core component of the current banking paradigm, i.e. how they make money.
Trackback uri
November 27th, 2007 at 8:55 pm
I did some University study based research approx 18 years ago that concluded that while speculative FX volumes greatly out-number ‘real trade’ volumes this does not have a medium or long term impact on the exchange rate … although it almost certainly does have an impact on the degree of volatility in the short term
although my research is now out of date, and precedes much of the automated trading that goes on now (enabled by the technology you refer to), I would be surprised if the conclusion was any different if the research were repeated now
so, surely, the public policy reason for the change you seem to support, could only be to reduce short term volatility
November 27th, 2007 at 9:32 pm
Thanks Gavin. Re your remarks that medium/long term fundamentals will eventually win out, I agree. However when a minor currency like the NZ$ becomes one of the world’s most traded currencies, with a daily turnover of $10b vs, annual real trade of $40b, that means short-term volatility, which adds costs for business (whether directly or through hedging).
I wasn’t supporting the FT’s idea (I’m not an expert in such matters) merely passing on what a publication with some credibility in financial markets has suggested. Together with most of my peers, I do worry about FX rate volatility and related costs, so ideas which might address it are obviously of interest.
November 28th, 2007 at 8:38 am
Jim, I agree that Currency fluctuation is a major issue, I would go so far as to say heading towards becoming the biggest single economic issue we face. If the FT’s suggestion can restrain the volatility of global currency markets, that is to my mind a good thing, a movement in currency is ok, fundamentals change, but for the NZD to move 12% odd within one month ( July-August 2007 ) with no change in fundamentals, that’s a huge shift, one that can’t be good for our economy if it continues on and indeed as appears inevitable really gets worse.
November 28th, 2007 at 9:07 am
Making interest real time instead of end of day wouldn’t make any difference. All it would do would add a little to the spread. The main issue in speculation is the ability to leverage capital.
Change that and you change the market.
What would work though is a Financial Transactions Tax which would add a charge to every financial transaction. Those who trade more, pay more. This has been around for a while and has resurfaced in the last decade via the Tobin Tax proposal.
Some more background here
http://www.newrules.org/finance/tobin.html
November 28th, 2007 at 1:44 pm
Gavin said…
much of the automated trading that goes on now…
Yeah Gavin, it is called Algorithmic Trading and it is quite popular these days. They (financial analysts), that by 2015, brokers would be out of a job, since algorithmic trading will be the main trading platform (see here).
I am developing something similar ie, financial market real-time analysis, however the algorithmic trading capability would be next the step in implementation, since that’s require some good architecture, which is not my depth although the software library (API) functionality for that has been written. All that is remained is application development, which I have just started a few weeks now.
I have talked to various fund managers, banks such as ASB Securities to find out what they’re using. From the info I got from feedback, is that not alot of analysis is done, mostly Excel spreadsheet and stuff like that. I am so surprised that some of our top fund managing companies don’t use analysis at all. I asked them why, they sais, they mainly look that the management team in the make up of some companies in their portfolios. They examine their past history, of where they had been involved and what their achievements had been like. I won’t say names here, because some of those are quite high profile since it may damage their public perceptions, which lead to investors discouragement in investing with them.
November 30th, 2007 at 7:15 pm
Here is another article about algorithmic trading. As a quote from the article from Andrew Lo, director of the Massachusetts Institute of Technology’s Laboratory for Financial Engineering. “Now it’s an arms race“. “Everyone is building more sophisticated algorithms, and the more competition exists, the smaller the profits.“. I think Andrew Lo is correct here. I don’t invent algorithm myself, but I just implement algorithms that are coming out of the regular research journals in Finance/Economics/Physics/Signal Processing/Statistics/Feedback Control/Machine Learning/Pattern Recognitions/Data-mining. There are lots of algorithms that are now being applied in financial analysis, which were originally published in non-economic displines, eg - Physics/Signal Processing/Feedback Control/Machine Learning. I think that I would be disadvantage since I don’t invent my own new algorithms, but solely rely on what other researchers are publishing, and so as everyone else (I say the majority). I have had informal talks with researchers at the AUT Math Department , my former lecturers and former researchers at the Physics Department at Auckland and they are keen to collaborate in algorithmic development in the future. But before all that happens, I think that I need to convince some Venture Capitalists out there that what I am doing has great potential, by showing them a proto-type first.
Artificial intelligence applied heavily to picking stocks.
January 3rd, 2008 at 2:05 pm
Hi - just wanted to say good design and blog - cu Ronny R.