![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
An Absence of Consequence
Feb 2013
Handbook of Exchange Rates - ed. Jessica James, Ian W. Marsh and Lucio Sarno - Wiley, 2012
My friend Jess pulled a Lewis Carroll on me a few months ago. I asked about her writing and she said that she had just finished another book and would I like a copy? I excitedly said yes, expecting another of her SF or fantasy stories. Instead, I got this.
Well, I've read it cover to cover, just to show her. I am even going to have a go at reviewing it, despite the fact that it is aimed at professionals in the field and not at financial ignoramuses like me. For Foreign Exchange (FX) is an important part of a modern economy and understanding how it works is vital in avoiding currency crises like the devaluations that were common in my youth, and, arguably, the current eurozone problems. I am hoping that at least a few take-home messages will emerge.
It is structured as an academic review, with chapters submitted by many authors loosely grouped by the editors into four areas: Overview, Models and Methods, Markets and Products, and Policy. The overview is particularly valuable in explaining the major players and some of the technical terms such as carry trade, currency swaps and algorithmic trading, though the chapter covering the historical development of FX markets would have been better placed at the start to give context to what follows.
It is true, however, that FX is a relatively new market in historical terms, only really coming into existence with the collapse of the Bretton Woods system in the early 1970s when most of the world's major currencies became free-floating. It has also developed at an incredible rate, both in terms of its volume (the average daily turnover is estimated as $2 trillion, some 16 times the GDP of the world's 35 largest economies) and its technical innovation (algorithmic trading, in which trades are performed by largely automated systems in microseconds, is now responsible more than 50% of total trades by volume).
Given the constantly changing nature of the beast, it is perhaps unsurprising that modelling of FX markets has tended to lag behind. Most of the empirically tested macroeconomic models are based on the wholly unrealistic assumption of equilibrium stock conditions and tend to break down when faced with empirical data other than that used to calibrate them. This has led to the rise of microeconomic models based on the dynamics of trades, particularly order flow, and the effects of new information on markets (assuming that all dealers see all information, but not at the same time). By far the most interesting new class of models, however, is based on modelling traders as agents with cognitive limitations who don't understand the whole market. Instead they use simple heuristic extrapolative or mean-reverting rules to take positions and adapt them based on their utility. This approach proves to be very adept at modelling the wild fluctuations and disconnects between underlying fundamentals and market rates that are empirically observed. What's more, the probability distribution of the simulated returns deviates from a normal distribution in the same way that actual returns do. Whether these models can be used to make money, only the authors probably know - they look inherently chaotic to me, which means that they are probably useless for making specific predictions. But I find it fascinating (and worrying) that introducing human ignorance into the model produces the most realistic results.
The section on methodology covered specific subjects that went largely over my head. The main impression I got was that the algorithms used by analysts generally work better than a random walk but not significantly so. There were a couple of chapters on whether particular managers post returns that are signicantly greater than those expected naturally (so-called "alpha"). The result? Most managers have zero or slightly negative alpha (due to transaction costs), but about one quarter have a significant positive effect. At least over the five year period examined.
The products section made my head spin with its talk of futures, forward swaps and derivatives, and the chapter on algorithmic trading was much too short and frankly alarming in its implications. Vital trades are made in microseconds, far faster than any human can keep track. The market is almost totally unregulated and (presumably) impossible to model, which implies that no-one can tell if it is operating safely or whether dangerous instabilities could develop and spiral out of control before anyone could stop them. If we do have a technological singularity, this will be the place where it starts.
Tellingly, the section on the consequences of FX market mechanisms on government policy - and hence on us - is by far the shortest. There are a few interesting take-home messages - government manipulation of exchange rates can be effective in reducing market volatility and improving competitiveness if done right, the Chinese renminbi was not kept artificially low in the 2000s as some commentators have claimed, and pegging exchange rates rather than letting them float freely can be the right choice for some countries in some circumstances - but in general the impression given is that the social consequences of FX trading are not of great interest to the people who study it. As the academic feel of this book makes clear, they see FX as a fascinating system that can be studied and analysed in much the same amoral way as scientists study an ecosystem or the human body. But FX is a human creation rather than a natural one, so the analogy with scientific knowledge is false and there is a moral dimension to the rules by which it operates which this book does not touch. We could, if we wanted, regulate FX out of existence by imposing controls, and for the vast majority of human history that is precisely what was done. Obviously, FX trading and the accompanying speculation have now become too deeply ingrained in western market and government operations to be safely decommissioned, but its almost total lack of regulation and the casual introduction of new instruments such as algorithmic trading are areas that regulators - and analysts - should be looking at with a very beady eye.
Handbook of Exchange Rates - ed. Jessica James, Ian W. Marsh and Lucio Sarno - Wiley, 2012
My friend Jess pulled a Lewis Carroll on me a few months ago. I asked about her writing and she said that she had just finished another book and would I like a copy? I excitedly said yes, expecting another of her SF or fantasy stories. Instead, I got this.
Well, I've read it cover to cover, just to show her. I am even going to have a go at reviewing it, despite the fact that it is aimed at professionals in the field and not at financial ignoramuses like me. For Foreign Exchange (FX) is an important part of a modern economy and understanding how it works is vital in avoiding currency crises like the devaluations that were common in my youth, and, arguably, the current eurozone problems. I am hoping that at least a few take-home messages will emerge.
It is structured as an academic review, with chapters submitted by many authors loosely grouped by the editors into four areas: Overview, Models and Methods, Markets and Products, and Policy. The overview is particularly valuable in explaining the major players and some of the technical terms such as carry trade, currency swaps and algorithmic trading, though the chapter covering the historical development of FX markets would have been better placed at the start to give context to what follows.
It is true, however, that FX is a relatively new market in historical terms, only really coming into existence with the collapse of the Bretton Woods system in the early 1970s when most of the world's major currencies became free-floating. It has also developed at an incredible rate, both in terms of its volume (the average daily turnover is estimated as $2 trillion, some 16 times the GDP of the world's 35 largest economies) and its technical innovation (algorithmic trading, in which trades are performed by largely automated systems in microseconds, is now responsible more than 50% of total trades by volume).
Given the constantly changing nature of the beast, it is perhaps unsurprising that modelling of FX markets has tended to lag behind. Most of the empirically tested macroeconomic models are based on the wholly unrealistic assumption of equilibrium stock conditions and tend to break down when faced with empirical data other than that used to calibrate them. This has led to the rise of microeconomic models based on the dynamics of trades, particularly order flow, and the effects of new information on markets (assuming that all dealers see all information, but not at the same time). By far the most interesting new class of models, however, is based on modelling traders as agents with cognitive limitations who don't understand the whole market. Instead they use simple heuristic extrapolative or mean-reverting rules to take positions and adapt them based on their utility. This approach proves to be very adept at modelling the wild fluctuations and disconnects between underlying fundamentals and market rates that are empirically observed. What's more, the probability distribution of the simulated returns deviates from a normal distribution in the same way that actual returns do. Whether these models can be used to make money, only the authors probably know - they look inherently chaotic to me, which means that they are probably useless for making specific predictions. But I find it fascinating (and worrying) that introducing human ignorance into the model produces the most realistic results.
The section on methodology covered specific subjects that went largely over my head. The main impression I got was that the algorithms used by analysts generally work better than a random walk but not significantly so. There were a couple of chapters on whether particular managers post returns that are signicantly greater than those expected naturally (so-called "alpha"). The result? Most managers have zero or slightly negative alpha (due to transaction costs), but about one quarter have a significant positive effect. At least over the five year period examined.
The products section made my head spin with its talk of futures, forward swaps and derivatives, and the chapter on algorithmic trading was much too short and frankly alarming in its implications. Vital trades are made in microseconds, far faster than any human can keep track. The market is almost totally unregulated and (presumably) impossible to model, which implies that no-one can tell if it is operating safely or whether dangerous instabilities could develop and spiral out of control before anyone could stop them. If we do have a technological singularity, this will be the place where it starts.
Tellingly, the section on the consequences of FX market mechanisms on government policy - and hence on us - is by far the shortest. There are a few interesting take-home messages - government manipulation of exchange rates can be effective in reducing market volatility and improving competitiveness if done right, the Chinese renminbi was not kept artificially low in the 2000s as some commentators have claimed, and pegging exchange rates rather than letting them float freely can be the right choice for some countries in some circumstances - but in general the impression given is that the social consequences of FX trading are not of great interest to the people who study it. As the academic feel of this book makes clear, they see FX as a fascinating system that can be studied and analysed in much the same amoral way as scientists study an ecosystem or the human body. But FX is a human creation rather than a natural one, so the analogy with scientific knowledge is false and there is a moral dimension to the rules by which it operates which this book does not touch. We could, if we wanted, regulate FX out of existence by imposing controls, and for the vast majority of human history that is precisely what was done. Obviously, FX trading and the accompanying speculation have now become too deeply ingrained in western market and government operations to be safely decommissioned, but its almost total lack of regulation and the casual introduction of new instruments such as algorithmic trading are areas that regulators - and analysts - should be looking at with a very beady eye.