A glimpse into banking climate disclosures

Taskforce on Climate-related Financial Disclosures (TCFD)

To their credit all of the leading listed banks and insurers in Australia have committed to implementing the 2017 TCFD recommendations.  Having given a lot of thought to the challenges of quantifying the physical risks ourselves, we were really excited to see what the banks came up with.

To summarise our findings – this is just the first round of disclosures in response to the TCFD and we are impressed by how quickly companies have responded.  But are we ready to say that the proportion of high risk properties only increases by 1-2% over the next 40 years and do the users of the information understand what this means?

A snapshot of recent disclosures by Australian banks

Let’s start with a snapshot of the progress each company has made with assessing physical risk. One easy call-out before we get into some of the detail:  it’s a great start.  The extent of the 2018 disclosures is indicative of a huge amount of work having been undertaken in a relatively short period of time.

NAB and ANZ have been participating in the UNEP FI TCFD Pilot[1].  So it’s maybe not surprising that they have not yet provided results of their modelling – a global initiative involving 14 other banks will inevitably be slower moving.

There are also good reasons for deferring the provision of results until they are known with more certainty.  We particularly like ANZ’s plan to trial a specific peril (namely flood – planned for 2019) in a specific area and to build the framework out from there for other perils and regions.  Given the uncertainties involved, this approach has the benefit of being able to adapt the framework as the results emerge.

Whilst NAB and ANZ appear to have focussed on the framework, Westpac and CommBank have managed to get to the finishing line, or near to it.  This is a remarkable achievement in the short time available considering the complexity in the underlying modelling.

The physical risk is assessed as minor for the residential mortgage portfolios

Both Westpac and CommBank show limited variation in the physical risk under a 4 degree warming scenario (a quite severe scenario).

Westpac indicate that only 1.7% of the portfolio would be exposed to higher physical risk by 2050. Read here. Commbank measure it slightly differently and show only 1% of properties as being high risk by 2060.  Read here.

We understand that the proportions are intended to indicate the impact of climate change on the bank’s credit risk and not to the risk of damage to the property. The latter will increase by much higher levels. In fact CommBank estimate the increase in property damage will be 27% by 2060 – a plausible figure.

The credit risk is lower than the risk of damage to the property for a number of reasons. Some of the key ones are:

• The homeowner will often have insurance in place covering most of the damage to the property as well as equity in the home.
• Ultimately a key issue will be whether the increase in natural hazard risk will have an adverse effect on property values. Potentially property values will reduce for a small proportion of properties – for example those at low elevations and near to the coast. As long as the proportion of properties affected is small, the 1-2% may be plausible.

Source:  Westpac 2018 Sustainability Performance Report

Were the disclosures helpful?

It appears that the 4 degree warming scenario is being adopted by each bank as a severe scenario to provide a conservative view of the risk (to use Westpac’s own wording). In this respect at least the basis being used is consistent across companies, which should assist comparability. However, we have concerns that this may be where the comparability ends, given the differences in the way the numbers are reported and the complexity of the underlying analysis. Does the industry need to agree on a common reporting basis and alignment of scenarios?

We also have some concerns about the extent of the Westpac and CommBank disclosures at this early stage in the context of the massive uncertainty about the nature of the future risk.  In fact the uncertainty of the estimates is not mentioned at all as far as we could tell.  Even ignoring that we can’t reliably predict how extreme events will change, there are questions about levels of availability and affordability of insurance in future, and impacts to property values, health and wellbeing.

It’s not clear that users of the disclosures will understand why the quantified risk is so low (ie 1-2%), especially under a severe 4 degree scenario.  In the absence of a common reporting basis, maybe some extra explanation of the reasons for the low proportions may have helped.

This is probably not the right question to ask.  It is inevitable that the estimates will be wrong, but the results can still be useful in understanding the nature of the risk if they reflect the best evidence at a point in time.  Maybe a better question would be whether the results are sufficiently reliable at this early stage.

In summary, we are moving in a good direction, and quite quickly.  But there is a lot still to be done.   Further standardisation will assist to make the disclosures useful.

References

CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.