Long-term multi-year forecasts for impairment and capital are often based on the premise that tomorrow will largely resemble yesterday. Shock events, such as the 1997 Russian debt default, the 2008 Financial Crisis and the recent cost of living crisis, provide useful inputs for developing forecasting models. These events remain relevant because of the extended period of geopolitical stability that ran from the late 1990s until the early 2020s.
This meant that when shocks occurred, they occurred within a relatively unchanging global environment but didn’t fundamentally change that environment. The same system boundaries and operating assumptions continued to apply, thus allowing good predictive models to be built.
However, the world has changed. The old rules no longer apply, and we are experiencing what are arguably the greatest shifts in the geopolitical landscape since the end of the Second World War. Isolationism, state-based armed conflicts, trade wars, and weakened international institutions are all systemic threats to the UK’s financial stability. Consequently, new approaches are required to ensure that our economic forecasts remain accurate and relevant in this new environment.
This blog covers
|
The thorny problem of structural breaks in macroeconomic data
Economic forecasting presents serious technical challenges for financial modellers. Traditional macroeconomic tools assume that, after a severe downturn, key indicators revert to the BAU historical position, not a potential reset to a completely different base case. To put it another way, our models extend the “in context” patterns of the past, not the uncertainties of an unknown future.
Things such as conflict risks are also inherently hard to capture, and it can be difficult to know how to incorporate such events during the model build phase and at model application. These are not "black swan" events, but like black swans, they do not correlate well with traditional economic predictors of risk.
Furthermore, recent BoE research demonstrates that as the size of a geopolitical shock increases, severe non-linearities emerge in the economic response that traditional models completely fail to capture.
Geopolitical risk: The regulatory position
UK financial regulators no longer view geopolitical risk as a peripheral, qualitative concern. The regulatory position has shifted from passive observation to active engagement, as evidenced by BoE requirements that were explicitly stated in the 2025 Bank Capital Stress Test (BCST).
The BCST required banks to model a severe "tail risk" scenario driven entirely by geopolitical shocks far outside of recent experience. The severity of the scenario illustrates the scale of the forecasting problem that needs solving. The scenario includes a rapid fragmentation of world trade resulting in a 20% reduction in global trade volumes, a 300% increase in natural gas prices, a 5% fall in GDP and a 28% reduction in house prices.
This type of scenario is difficult to model statistically.
Not surprisingly, the Post Model Adjustment (PMA) has become the risk managers’ best friend for capturing risks of this type;
“We’ll just slap on a little something extra for conflict and other geopolitical risks that our models don’t account for.”
Yet, as the PRA highlights in SS1/23, PMAs are meant to be temporary fixes based on rigorous root cause analysis, not rough and ready permanent add-ons. PMA adjustments will always be needed but we should not be over reliant upon them. Where practically possible, modelled outcomes should take priority.
Learning from climate risk?
The geopolitical upheavals we are now experiencing represent a difficult problem for forecasters, yet we have solved a similar problem before. We should learn from this experience.
Over the last decade, the UK financial sector successfully transformed Climate Risk from a vague concern into well-defined quantifiable risk metrics. This was done by splitting climate risk into two distinct, actionable categories, and we can also apply this to conflict risk:
- Physical event risk. A conflict can be viewed as analogous to an extreme weather event that destroys value. For example, wars and hurricanes both result in direct physical destruction of collateral and reduce customers’ financial capacity to service their debt obligations.
- Transition risk. As we move from the old world order to the new, this is analogous to the move from a high carbon emission economy to a net zero one. It captures the structural shifts in international relations, trade and military power. This transition risk covers things like reshoring of operational activities that increases firms’; costs, and writing-off profitable ventures and revenue streams in jurisdictions that turn hostile.
Framing Geopolitical Risk in this way, we can apply tried and tested alternative approaches to deal with it, that we know work for climate risk.
This includes methods such as:
Non-linear statistical approaches. For example, Markov Switching models that apply different modelling assumptions across different time periods, allowing for varying periods of stability and volatility. One can therefore potentially apply a mixture of peaceful, stable global activity with more volatile ones within a single scenario.
Forward looking, narrative-based scenario creation. These are derived from a blended data driven / expert view of what “the new normal” might look like in the future, and hence, provide a baseline for long term forecasts to trend to.
Bottom up, index-led adjustments. These are based on customers’ individual sensitivity to change in different regions, districts, industries and event types. In principle, similar to flood risk indices used in climate risk modelling that provide an overlay to the predictions from the core forecasting model.
The role of the PMA will no doubt remain, but widening the set of forecasting tools that we have and applying these to geopolitical risks can move us forward.
Final thoughts
The increase in conflicts and other geopolitical events present an existential challenge to traditional macroeconomic forecasting and risk management. The era of relying on models based on decades-long periods of relative stability is over.
The Bank of England now expects firms to possess the analytical capabilities necessary to predict the impact of severe, geopolitical-driven macro-financial disruptions. To meet regulatory expectations and protect their balance sheets over the long term, UK financial institutions must acknowledge the limitations of their current frameworks and develop new approaches to geopolitical risk, just as they have adapted to climate risk in the past.
How Jaywing can help
At Jaywing, we work with banks and lenders to strengthen credit risk forecasting frameworks in volatile environments. We can help with improving model resilience, reducing over-reliance on PMAs, and embedding forward-looking risk into capital and impairment forecasts.
We can also support firms to:
- Identify structural breaks and shifts in macroeconomic data
- Apply non-linear and state-dependent modelling approaches
- Build narrative-driven stress scenarios aligned to regulatory expectations
- Develop granular, index-led overlays at sector, regional and portfolio level
- Strengthen model governance and documentation to withstand supervisory scrutiny
Our teams combine deep technical modelling expertise with hands-on experience of BCST, IFRS 9 and ICAAP frameworks.
If geopolitical instability is stretching the limits of your current forecasting framework, we would welcome a conversation.
Get in touch to discuss how your models would perform in response to a severe geopolitical event and where targeted enhancements could strengthen resilience.