With this issue you appear to be a part of a small minority who understands and recognizes the risks.
Systemic risk: what is it? It can be defined as system-wide promises that can't be kept.
Such unkept promises are breakdowns of expected behavior.
[1] Transactional (commercial and financial transactions, deliveries, schedules)
[2] Operational (signalling, power transmission, messaging)
The effects of such breakdowns may be:
(A) Primary
(B) Secondary
(C) Immediate
(D) Delayed
In complex systems, there may be linked breakdowns
For example, at the height of the financial crisis, banks simply stopped lending to each other. Technically speaking there was no operational breakdown. Instead the system was permeated with toxic financial instruments, and every bank knew it. Thus, there was a breakdown of trust. That stopped the flow of credit, and a whole slew of secondary effects followed, causing the global economy to slide into near-collapse.
This was contagion: unchecked propagation of risky financial instruments which began to fail, causing mistrust of ability to pay, causing separate financial quarantines which stopped the global flow of money.
Part of the problem lay with insufficient capitalization requirements, Basel I, II, and now III, but that's beyond the scope of this forum. What's not beyond this forum is the fact that systemic risk - promises that couldn't be kept - jumped across all sectors globally and generated secondary risks such as debt that are barely controlled. They may yet precipitate another meltdown. Lesson: systemic risk has cross-connects.
---
As noted in your linked presentation algorithms and HFT affect finance in ways that are not understood, and cause crashes that can't be explained. As Slavin says, the only solution they're sure of is to "hit the red button". Or as stated upstream "The only certain countermeasure is to shut down and revoke trades."
Message 27182215
We've discussed algorithmic trading and HFT extensively:
Message 27498427
Message 27366843 ... and more.
---
We've seen the breakdown of relatively simple technology time and again: Three Mile Island, Chernobyl, Fukushima, jet crashes. Western media are having a field day because Chinese bullet trains collided after lightning strikes. These are fairly simple failures, caused by unexpected conditions. Relatively, algorithms in an jet airliner are child's play. If you want real complexity, look at global finance and economics where failure-prone algorithms are operating. What happens when something unexpected occurs? As Mandelbrot, Nassim Taleb, and the recent global crash taught us, Black Swans can be killers. But we still haven't learned.
Global use of algorithms and HFT is acknowledged to fail unpredictably, for reasons that can't be explained. Yet use continues, for billions of daily transactions involving trillions of dollars. Has the system been stress-tested for systemic effects? No. (Why bother when it's already breaking down every month?) Are the capital requirements - particularly with respect to $600 trillion in derivatives and HFT - sufficient? We have no idea. Can networks be brought down operationally? If brought down operationally, can transactional effects be controlled?
Should we abandon use of algorithms? Of course not. They're wonderful tools. But with respect to global finance, anyone who tells you the risks are understood and under control is lying. Not mistaken: lying.
Jim
|
