The problem
For a central bank, climate change is not an environmental brief; it is a threat to price stability and financial soundness. Mongolia loses an average of USD 0.43 billion a year to natural and biological hazards, equivalent to 3.3% of GDP, and its periodic dzud winters translate directly into banking-sector stress. The 2009/10 dzud killed 8.5 million head of livestock, 20% to 25% of the national herd, and cost around USD 287 million, close to 4% of that year's GDP. When a herder cannot repay, the shock lands on the loan book, and from there on the credit channel the central bank relies upon.
Who this is for
This is written for central bankers, financial regulators and prudential supervisors in commodity- and agriculture-exposed economies, and for the DFIs supporting them. It assumes familiarity with monetary policy transmission and Basel-style prudential supervision.
Why this matters now
The 2023/24 dzud made the transmission concrete. The Bank of Mongolia allowed payment postponement on a MNT 1.2 trillion (about USD 354 million) herder loan portfolio, because without forbearance 51.2% of that portfolio would have been downgraded to Special Mention, with the associated provisioning hitting bank capital. A physical climate event became, within a single season, a question of loan classification and capital adequacy. That is the mechanism every exposed central bank now has to model before it is tested by its own version of a dzud.
Absent forbearance, 51.2% of the herder loan portfolio would have been reclassified to Special Mention after the 2023/24 dzud, converting a climate event into a capital-adequacy problem. Source: Bank of Mongolia (2025).
How climate risk transmits into monetary policy
The reports identify four transmission channels a central bank must track. Physical shocks such as dzuds and droughts constrict food supply and push up food inflation, feeding inflation expectations and pressure on the policy rate. Loan defaults concentrated in agriculture and mining stress the banking sector and weaken the pass-through of that policy rate through the credit channel. Chronic physical and transition impacts shift the natural rate of interest by altering productivity, labour supply and risk premiums. And Mongolia's external dependence, with mining at 92% of exports and around 65% of GDP, couples commodity-price and trade shocks to the exchange rate. These are not tail scenarios; they are the operating environment.
The transition side: a concentrated, high-carbon loan book
Transition risk is equally material because the economy is carbon-heavy and the banks are concentrated in it. More than 20% of Mongolian bank loans sit in mining, quarrying, manufacturing, construction and transport, sectors that carried an 18.1% non-performing loan rate in 2023. A disorderly transition, whether through trading-partner decarbonisation, coal-price shocks or stranded assets, would fall on exactly the exposures that are already the weakest, which is why the Bank of Mongolia treats sectoral concentration as a systemic-risk question rather than a per-loan one.
What the Bank of Mongolia is actually doing
The practical response is supervisory and analytical, not aspirational. The Bank approved an ESG Risk Management System regulation in 2023 by Governor's decree, embedded "greening the financial sector" in its 2022 Monetary Policy Guidelines, and set a 10% green-loan target for banks. Most importantly, in the 2024/25 cycle it ran a bottom-up climate scenario-analysis pilot across seven banks, including five systemically important ones, using NGFS scenarios adapted to Mongolian conditions, and it is extending its long-standing Forecasting and Policy Analysis System to carry climate variables. The supervisory design applies proportionality: enhanced assessment, including site visits, is triggered for loans above MNT 1 billion at systemic banks and MNT 500 million at others, drawing on international frameworks including the NGFS scenarios, PACTA, PCAF and IFRS S1 and S2.
Risks, limitations and what a robust approach requires
- Inconsistent bank-level methodologies undermine comparability. Divergent ESG assessment approaches across banks mean supervisory data cannot yet be aggregated with confidence; harmonised reporting templates are a precondition, not a refinement.
- Sectoral concentration makes the risk non-linear. With mining at 65% of GDP, a transition shock does not scale smoothly; it threatens a cascade, which conventional per-loan provisioning is not built to absorb.
- Compounding events erode recovery capacity. Successive dzuds, combined with a herd that has grown 44% while grassland shrank 20%, mean each shock lands on a system with less slack than the last.
- Monetary policy room is limited. With imports near 66% of GDP and a heavily dollarised system, external shocks dominate domestic levers, so climate-adjusted policy has to work alongside prudential tools, not in place of them.
Conclusion
Mongolia shows the sequence a climate-exposed central bank should follow: quantify the physical and transition exposure, prove it in the loan book (as the 51.2% herder-portfolio figure did), then build the analytical capacity, scenario analysis embedded in the forecasting system and harmonised ESG supervision, before calibrating prudential or monetary levers. The lesson for other commodity- and agriculture-dependent economies is that the diagnostic work, not the choice of instrument, is what determines whether the next climate shock is absorbed or amplified.