Non-technical summary
- Lenders face various risks from climate change. To build on industry efforts to better understand these risks, as part of our stress testing programme in 2022 we worked with the largest banks in Aotearoa New Zealand to assess climate-related risks in two sets of sensitivity analyses, focussing on:
- risks from drought and emissions pricing for their agricultural exposures, the focus of this article, and
- flood risk for their residential mortgage exposures. We published the flood risk results in March 2023.
- The exercise was successful in its main purpose of supporting banks to build their capability to identify and assess climate-related risks, find solutions to the significant data and modelling challenges, and apply the learnings to their management of these risks.
- We designed the exercises to be potentially stressful, in line with our regular stress testing approach.
- The exercise consisted of sensitivity analysis, not scenario analysis. Sensitivity analysis focusses on illustrating the effect of changing one variable at a time, or a small set of closely related variables, while holding all else constant.
- The work that banks did for these exercises fed into the 2023 Climate Stress Test, which is underway at the time of writing. This is a full-fledged scenario stress test, with multiple climate and non-climate related variables changing together over time.
- Across this exercise, analysis was limited to banks’ dairy and sheep & beef portfolios.
- To assess drought risk, banks estimated defaulted exposures for two different severities of drought: a one-year drought, and two years of drought back-to-back. We prescribed common assumptions for milk and meat prices and other economic variables, and regional drought effects on farms’ production levels, working expenses, and land prices. Results are conditional on these background assumptions. The one-year drought resulted in 7% of banks’ sheep & beef exposures and 8% of dairy exposures defaulting, and the two-year drought resulted in double those amounts, compared with around 3% in a baseline with no drought.
- For emissions pricing, we prescribed four different emissions price levels and for each of these banks estimated the proportion of their agricultural borrowers that would be unprofitable, conditional on the same background assumptions. An emissions price of NZ$15 per tonne of CO2 equivalent — the lowest level of emissions pricing considered in the exercise and the most relevant for assessing risk through 2030 — resulted in 8% of dairy exposures and 22% of sheep & beef borrowers being unprofitable, compared with 6% and 15% in a baseline with no emissions pricing. As expected, this increased as the emissions price increased.
- There is notable variation in individual bank results driven in part by differences in methodologies and assumptions, highlighting the exploratory nature of the exercise.
- We collected qualitative information on actions banks may consider taking to support customers in adapting to and mitigating risks from drought and emissions pricing. The quantitative results are absent any mitigating actions by banks or their customers.
