When markets test conviction, trust the process
In highly volatile markets the pressure to act can be overwhelming. A disciplined investment process helps resist costly reactions.
Autheurs
Uncertainty is the greatest test of investment managers’ mettle. Frazzled reaction, or stunned inaction, could determine the fate of their hard-earned gains. As fundamental investors, we believe that having a robust investment process should position you for a wide range of outcomes. Whether you are a value equity investor, an emerging debt savant, or managing portfolios with capital preservation objectives, the principle is the same: a well-defined process provides the discipline to navigate environments where visibility is low and emotions run high.
Yet it is precisely in these moments – when uncertainty is most acute – that the temptation to deviate from that process is greatest. Geopolitical shocks, sharp market moves, and a constant flow of news can create a sense that action is necessary. The challenge for investors is to distinguish between situations that genuinely warrant a reassessment of underlying assumptions, and those that simply test conviction.
But what happens when these underlying assumptions get tested day after day?
Periods of sustained disruption tend to trigger a familiar set of expectations in financial markets. As uncertainty rises, investors typically look to US Treasuries to provide a hedge; commodities (particularly energy) to rally on supply concerns; and the US dollar to strengthen as a safe haven. More broadly, there is an implicit belief that diversification should “work”: that different asset classes will behave in offsetting ways, cushioning portfolios against the shock.
These expectations are grounded in historical experience, and in the roles that different assets are intended to play within a portfolio. Government bonds have often provided protection during risk-off episodes, commodities have responded to supply disruptions, and the dollar has benefited from its reserve currency status.
However, these relationships are often treated as more stable – and predictable – than they really are. The assumption that markets will respond according to a familiar script can be comforting, but it risks oversimplifying a more complex reality.
Using correlations as an example…
Correlations are often key inputs when managers determine the strategic asset allocation of a portfolio, or when deciding whether to put on a new position. These inputs often underpin the rationale for diversification and arise from the assumption that relationships are stable enough to guide decision–making.
Correlations are not static. Relationships between equities and other asset classes shift meaningfully as crises unfold, often moving well ahead of the peak in market volatility. Patterns that may have held in more stable environments can weaken, strengthen, or even reverse over relatively short periods. Figure 1 below shows how correlations change going into and during events.
Average correlations with Global Equities
Source: Schroders, Refinitiv Datastream, 25 March 2026. Correlations are 12-weekly. Pre-Event = latest correlation entering the conflict. Average correlations across seven geopolitical events: 9/11 and Iraq invasion (2001), Madrid and Moscow bombings (2004), Crimea & ISIS (2014), North Korea-US tensions (2017), Russia-Ukraine war (2022), Israel-Hamas conflict (2023), Israel-Iran conflict (2025).
This means investing is often viewed as more art than science, as no single asset behaves consistently across all periods. Assets typically viewed as defensive do not always respond in a uniform way. As I alluded to in a recently published paper, cross-asset interaction will vary depending on current market context, and whether markets interpret the shock as primarily a growth risk, an inflation risk, or both. In the current Middle East context, investors, central banks, and policymakers are weighing the balance between growth risks and lingering inflationary pressures.
This means that diversification does not disappear during periods of stress. It becomes more dynamic, and less predictable, than many investors assume.
But wait, even the shift in direction is not guaranteed…
Correlations do not shift in a consistent direction. Even when looking at the same pair of assets, the way relationships evolve can differ markedly from one geopolitical episode to another.
2026 Iran War: The sharp decline in gold prices in mid-March 2026 is a clear example of shifting return drivers. Rising inflation concerns began to dominate its geopolitical hedging role, with gold once again moving in line with real yields after that relationship had broken down in recent years.
For example, assets often assumed to provide diversification, like commodities or gold, do not exhibit a uniform relationship with equities across events. In some cases, correlations become more negative, offering a degree of protection. In others, they move in the opposite direction, with assets becoming more positively correlated just as markets come under pressure.
The same is true for traditional safe havens, where the strength and timing of the diversification benefit can vary depending on the nature of the shock. Looking at two recent geopolitical conflicts involving military conflict, Figure 2 below shows how correlation shifts can often go in different directions.
Change in correlations vs global equities
Source: Schroders, Refinitiv Datastream, 25 March 2026. Correlations are 12-weekly. Change in correlations calculated as difference between pre-event and first 12 weeks of the event.
This inconsistency reflects the fact that not all crises are the same. Some are primarily growth shocks (for example, the Euro sovereign crisis of 2011); others are driven by supply disruptions (for example, the supply chain disruptions post Covid) or inflation concerns (the expected inflationary surge derived by the imposition of US tariffs), and many involve a combination of both. Therefore, it is only logical that there is no single pattern that investors can reliably depend on. There is no single geopolitical “playbook” that holds across episodes.
The dispersion between events is also pretty wide
Looking across events, the most striking feature is not just that correlations change, but how widely outcomes can vary. Even traditionally defensive assets like US treasuries exhibit a wide dispersion of correlations. Figure 3 below shows how average correlations between equities and other assets range between +0.5 to -0.5. This means that at times, correlations between equities and defensive assets are meaningfully negative, providing clear diversification benefits. At others, that relationship is far weaker, offering less protection.
This dispersion highlights an important point: averages can be misleading. While it is tempting to rely on historical norms, the experience during individual episodes can differ significantly from those averages. What matters for investors is not just the central tendency, but the range of possible outcomes, especially during periods of stress.
Correlation
Source: Schroders, Refinitiv Datastream, 25 March 2026. Correlations are 12-weekly. During Event = 12-week window upon start of event. Average correlations across seven geopolitical events: 9/11 and Iraq invasion (2001), Madrid and Moscow bombings (2004), Crimea & ISIS (2014), North Korea-US tensions (2017), Russia-Ukraine war (2022), Israel-Hamas conflict (2023), Israel-Iran conflict (2025). Range shows min-to-max range of 7 event-window correlations.
More fundamentally, this variability suggests that uncertainty in cross-asset relationships is not an anomaly, but a feature of markets. At the risk of sounding like a broken record, the behaviour of different asset classes during geopolitical shocks is shaped by a combination of factors – growth expectations, inflation dynamics, policy responses and market positioning – all of which are pretty much guaranteed to differ from one episode to the next.
In that sense, uncertainty is structural, not accidental. The fact that correlations behave differently across events is not a failure of diversification, but a reflection of the complex and changing nature of the risks that investors are trying to manage.
Which brings me back to my main point…
Taken together, these findings point to an uncomfortable conclusion: while diversification remains essential, its behaviour in the short term is far less predictable than many investors assume. Correlations shift, sometimes sharply, and often in ways that depend on the nature of the shock rather than the asset class itself.
This doesn’t mean that diversification is broken. The variability of correlations does not undermine the case for diversification but paradoxically reinforces it. If cross-asset relationships were stable and predictable, investors could simply position for a single outcome. The uncertainty underscores the importance of maintaining a diversified portfolio designed to perform across a range of scenarios.
This is precisely why a disciplined investment process matters. We have already established that attempting to position portfolios around geopolitical events is inherently challenging. Investors cannot reliably predict which hedge will work, know when correlations will shift, or reposition portfolios in real time with any degree of consistency.
In practice, attempts to do so often prove counterproductive. Decisions made in the midst of heightened uncertainty tend to be driven as much by narrative and recent price action as by fundamentals. This can lead to chasing protection after it has already repriced, or reducing exposure to risk assets after the bulk of the adjustment has occurred - effectively locking in losses rather than managing them.
A well-constructed investment process, by contrast, is designed with these uncertainties in mind. It knows that diversification will not always behave consistently over short horizons, and that different assets will respond differently depending on the nature of the shock.
This does not mean ignoring new information or failing to reassess the outlook. It means distinguishing between reviewing assumptions and reacting to noise. In most cases, geopolitical shocks warrant careful evaluation (did someone say emergency investment committee meeting?) but not a wholesale reset of portfolio strategy.
Trust the (investment) process
Investors face the temptation to act in the face of uncertainty, but history shows that market behaviour in these periods is neither uniform nor predictable. Correlations shift, often quickly and inconsistently, and diversification does not always look the way investors expect in the moment.
In this inherently uncertain world of investment, the moments that most tempt us to abandon process are often the moments when process matters most.
Subscribe to our Insights
Visit our preference center, where you can choose which Schroders Insights you would like to receive.
Autheurs
Topics