Financial downturns, plane crashes, and train wrecks

Image “Panic of 1873 bank run”, Illus. in: Frank Leslie’s Illustrated Newspaper, 1873 Oct. 4, p. 67. – see Wikipedia for further details

Tim Harford, wrote a great blog post today, “Time to bring in the crash investigators”, which opened as follows:

The NTSB is capable of providing a clear and authoritative narrative, explanations and conclusions about the crisis

After a financial train wreck, it’s time to begin to learn lessons from the disaster and prevent its recurrence.

I found that the blog post led me to a number of further thoughts around this topic, which I have set out below.

In his blog post, Tim Harford referred to views that it is perhaps not the usual regulators (“Securities and Exchange Commission, the Financial Services Authority, or the G20’s Financial Stability Board“) that might be best placed, or most appropriately structured, to act as an investigator, but rather a board styled around something like the National Transportation Safety Board (NTSB). The reasons given were due to the unbiased way in which the NTSB can criticize its own rules, and the way in which the NTSB focuses on reviewing the facts available.

I think there is a lot of validity to this, and left a comment on the article, with further support for this mindset:

Another reason supporting “crash investigator” style thinking is their broad understanding that most crashes are caused not by a single failure, but rather by a series of multiple failures in the system – so, when a plane crashes, it’s not due to pilot error (the onboard computers would issue warnings), rather it might be due to pilot error (eg, a tired pilot), flying during exceptionally stormy weather, with a simultaneous breakdown on the plane’s onboard computer systems.
Financial investigators need to think the same way, and not stop work after discovering the first ‘issue’ that they come across, but rather understand the complex integration of different systems, and multiple simultaneous failure points.
In both systems (financial markets, and air/rail networks), smaller failures have typically smaller consequences, but when large, multiple points of failure occur the consequences can be dramatic (a plane crash in the middle of an ocean, or a financial market going into melt-down). Therefore, despite the lack of frequency of such major cases, the implementation of a robust team of investigators (with the appropriate mindset) is a worthwhile investment.

Even in everyday life, things generally don’t turn into bigger problems on their own, but rather when multiple different issues collide simultaneously.

This discussion leads me to three questions:

1) How much time, energy and cost should one invest to understanding historical ‘crashes’, to prevent future crashes?

  • To be clear, I am not advocating not spending any time, or spending too little time – I do believe that “lessons learned” can help reduce future risk, and minimize pain and suffering.
  • That said, there must be a threshold by which the investigation/learning leads to diminishing returns – where is that point, and how do the investigators know when they have reached it?
  • Interestingly, in the case of many public situations (eg, the financial downturn experienced in the last five years), there is extensive private interest in forming points of view, with supporting critical analysis, for the reasons for the downturn, driven partly by intrinsic interest (eg, a desire to know the ‘answer’), extrinsic interest (eg, selling books), and driven by a desire for fame/notoriety.

2) Shouldn’t crash investigators look forward, and try to predict future crashes, and their respective causes?

  • Crash investigators might minimize risk of repetition of historical crashes, but that doesn’t mean that there won’t likely be further crashes (Wikipedia has page which sets out a long list of economic crises). In financial markets the history of crashes and downturns, to varying degrees, is widely acknowledged to be caused by different factors.  Of course, there would likely have been even more economic crises if we hadn’t learned some past lessons.
  • This question also largely comes down to a consideration as to whether the skill-sets and approach taken by these crash investigators is the most effective, and efficient way to determine, and avoid/manage risk? There are plenty of other market participants (not just in financial markets, but also, for example, in the aerospace sector, with regard to avoiding plane crashes) who believe they can (and maybe also believe that they are best placed to) address that risk.
  • Crash investigators might use their combined knowledge, and unique skill-sets and approach to develop a unique, holistic, unbiased point of view as to future risk levels. Such a board might then report to the state (political leaders), regulators, and the market players themselves. They might not even need regulatory ‘power’ – if their accuracy was high, market forces would eventually ensure that they would be listed to. The real challenge might be to determine who would finance such a board, with sufficient funding to give them ability to do their jobs properly (perhaps a co-operative funding model, by incumbent market players?)

3) How much redundancy should one build into a system, to prepare oneself for when these ‘rare’ multiple, simultaneous errors occur?

  • If crashes result from systematic failures at many parts of a system, it’s already highlighting that the redundancy built into the system has failed to do it’s job (if crashes result from only one failure, it highlights a lack of any redundancy). And even with redundancy, if crashes still occur, doesn’t one need to not only build in redundancy at each stage, but also further redundancy, either at each stage, or on a macro level, if all stages fail together.
  • Such ‘redundancy’ measures typically increases costs, and is often seen as inefficiency, until the problem actually occurs, which of course it is significantly less like to do so if the redundancy is implemented (a paradox!).
  • One way to approach this is try to ‘predict’ the risk – one might apply gut feeling (perhaps sufficient, in the case of many everyday life matters) or otherwise develop simple, or complex models to simulate different possible outcomes.
  • Interestingly, stakeholders in the respective market (eg, airline passengers, or the general public, in the case of the financial economy) would be very interested to know how the market players had judged the necessary investment, and level of acceptable risk, were such models to be employed.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: