Flood initiative shows cross-border risk

Listen to this article


As financial policymakers from the Group of 20 countries gathered in Scotland, there was a debate about policy reform – and the lessons that financiers should learn from past banking disasters.

However, for another – more offbeat – clue about the policy challenges that await the G20, financiers might do well to take a glance at a meteorological initiative that is under way in Reading, England, to measure European flood risk.

On the face of it, the task of predicting river and sea levels might seem unconnected to the world of finance. Water risk, after all, is a tangible problem linked to the natural environment. The credit disasters, however, were triggered by human fear and greed, often in a seemingly ethereal realm of cyber space.

Yet, the way that scientists in Reading are trying to measure flood risk has fascinating parallels with a topic now confronting G20 leaders – namely, macroprudential regulation.

In recent years, European meteorologists have embarked on an intriguing experiment. Flood warnings in Europe were once issued by national bodies, and based on a binary system of alerts, meaning that they either predicted a flood – or not. But national bodies now co-operate with each other, to produce pan-European, cross-border forecasts.

Moreover, meteorologists have embraced a more complex method of analysis known as “ensemble” forecasting. This essentially tries to measure flood risks by using dozens of computers that all run simulations of how a river (or sea) might behave, based on slightly different initial inputs.

The resulting collection of forecasts is used to create a composite vision of the future, instead of a simple binary result. At Reading, for example, over four dozen computers are now used to create this “ensemble” forecast, producing a highly nuanced picture of European flood risk.

Now, in theory, these reforms look sensible. After all, water levels can be affected by a host of cross-border factors – and a composite assessment of risks is usually a better reflection of real life than crude, binary models. Yet, there is a catch. In the past, meteorologists have found it difficult to get all the different European bureaucrats to co-ordinate, or even use data in the most rational way, owing to their diverging domestic pressures. As a result, as an academic study notes, “results suggest that flood forecasters may not instinctively use ensemble predictions in the way that promoters [of this] think they should”.

The collective forecasts have left meteorologists exposed to political risk. As forecasts become more nuanced, there is a growing onus on the meteorologists to use their judgments about when to ring alarm bells – or not. Imagine, for example, a scenario where one out of those 50-odd computers in Reading predicts that Canary Wharf is about to flood. “Should we call for an evacuation of Canary Wharf, or not?” laments one academic. Evacuations are costly. But if it turned out that an evacuation had not been carried out if one of those 50 computers flashed red, there would be recriminations.

That is where the parallels for financial regulation crop up. In recent years, financial policymaking has also been dominated by national regulators, working with rules and goals that often seemed clear and binary. Think of all those inflation targets, or capital adequacy standards.

But now financial policymakers, like meteorologists, are trying to co-operate across borders and adopt a more holistic vision of risk. Hence the new dialogue between G20 ministers and the focus on “macroprudential supervision”.

In many ways, that shift looks as sensible as adopting “ensemble” forecasting for flood risk. Yet, the more that financial bureaucrats move into a world of cross-border macroprudential supervision, the more that regulation will become riddled with complex political judgments.

What will central banks or regulators do in the future if consumer prices seem benign, but property prices are not? Or if one country is booming, but another is not?

There is, of course, no easy answer. The Bank of England has made one attempt to recognise the complexities of forecasts by using a “fan chart” to illustrate inflation probabilities. That approach might conceivably be extended to financial risk too.

But perhaps the real lesson is that while the type of reforms that the G20 are calling for are eminently sensible, they will not be easy to implement. Perhaps it is apt that the headquarters of the FSA, the British regulatory body, is in Canary Wharf: if nothing else, it is a stark reminder that predictive, proactive policymaking is an art, not science – for banks, just as much as rivers.

Copyright The Financial Times Limited 2017. All rights reserved. You may share using our article tools. Please don't copy articles from FT.com and redistribute by email or post to the web.