Stock selection is the primary driver of alpha in EM equities and the most fertile ground for systematic strategies. At the security level, company fundamentals matter: earnings dynamics, balance sheet quality and valuation signals have historically carried real predictive content. Country-level returns, however, are a different story. Macro forces tend to dominate: monetary policy cycles, terms of trade, external demand and financial conditions.

Country timing has traditionally been the domain of discretionary managers, whose macro narratives and on-the-ground conviction shape top-down positioning. Systematic equity managers, by contrast, have historically focused on bottom-up stock-picking, while country exposures tend to be treated as risks to be contained. A handful of managers incorporate top-down signals, but most approach the opportunity set through a fundamental, rather than a macro, lens.

Yet, the landscape is shifting. High-quality, point-in-time macroeconomic data spanning financial conditions, labour markets, trade dynamics, inflation and more now offer the raw material for a disciplined, systematic approach to country allocation. Advances in modelling techniques allow us to extract signals from dozens of macro indicators across EM, a task that was once too noisy and too sparse to tackle rigorously. The goal is not to reduce the risk budget dedicated to stock selection but help to ensure that country exposures result from a deliberate and dynamic country allocation process. This should complement bottom-up alpha by harvesting macro information that fundamental models leave on the table.

Where discretionary managers have long relied on narrative and on-the-ground judgment, systematic managers can now bridge that gap with a repeatable, data-driven process, backing macro conviction with hard data.

Opportunity set in EM

A risk decomposition of EM versus developed markets (DM) highlights a key nuance: while stock-level risk dominates both, country-level risk is significantly higher in EM (Figure 1). As a result, the opportunity set for country timing is structurally larger in EM, although stock selection models are typically not built to exploit it.

Figure 1. Country risk contribution (% of total risk): EM versus DM

Source: Man Numeric. Data as at Jan 2008 – Jun 2026.

Problems loading this infographic? - Please click here

Cross-country return dispersion over time further highlights this gap. EM countries continue to exhibit wider return spreads than their DM counterparts, reflecting greater heterogeneity in economic structure, policy regimes and external vulnerabilities (Figure 2).

Figure 2. Rolling 12-month country return dispersion: EM versus DM

Source: Man Numeric, MSCI. Data as at Jan 2008 – Jun 2026.

Problems loading this infographic? - Please click here

A simple fundamental country timing model

A natural starting point for country timing is a cross-sectional model that blends relative valuation and 12-month price momentum, standardised across EM countries, into long/short country positions. In practice, this approach has a structural weakness. As valuation and momentum are slow-moving, persistent signals, the resulting country exposures can remain static for years at a time rather than rotating with the macro cycle – as illustrated by South Korea (Figure 3). Due to its chronic valuation discount to EM peers, driven by corporate governance concerns and a concentrated, export-oriented industrial structure, the model has maintained a long Korean equities position for most of the past two decades, largely independent of the prevailing macro backdrop.

Figure 3. South Korea: Value and Momentum country position over time

Source: Man Numeric, MSCI, Bloomberg. Data as at Jan 2008 – Jun 2026.

Problems loading this infographic? - Please click here

To evaluate the model's standalone efficacy as a country-timing signal, we converted its standardised score into a country-level long/short position, sized in proportion to conviction (capped at ±5%) and applied it directly to each country's own market total net US dollar return.

Applied across all EM countries, this fundamental model generated a positive cumulative profit and loss (PnL) over the period examined (Figure 4), but its risk-adjusted profile remained unimpressive, with a gross Sharpe ratio under 0.5, alongside a significant turnover. Furthermore, over 40% of the value added since 2008 was generated during the past 12 months, as the model benefited from the dramatic rally in South Korea. On the negative side, the model would have been particularly vulnerable to the Russia drawdown in 2022 as the country became uninvestable.

Figure 4. Cross-sectional Value + Momentum model: cumulative PnL and summary performance, all EM countries

Source: Man Numeric, MSCI, Bloomberg. Data as at Jan 2008 – Jun 2026.

Problems loading this infographic? - Please click here

Problems loading this infographic? - Please click here

This is precisely why, in practice, many quantitative EM strategies do not target country exposures directly. The model draws on the same valuation and momentum information that underpins bottom-up stock selection; its country tilts also tend to be positively correlated with a manager's existing stock-picking alpha, limiting the diversification benefit on offer. In short, country tilts emerge as a by-product of bottom-up stock selection, rather than as a deliberate, risk-budgeted allocation decision.

The issue is not that fundamental models are flawed, but ignoring the country dimension can potentially leave a portfolio exposed to macro regimes it was never designed to navigate. A systematic country allocation overlay, informed by macroeconomic data rather than stock selection alpha residuals, can help mitigate this blind spot without necessarily compromising the bottom-up process.

Enhancing country allocation through a macro-based country timing model

Historically, systematic equity managers struggled to model country returns because the data was sparse, noisy and prone to restatement. That has now changed. Leveraging a broad array of macroeconomic datasets, a rigorous country timing model can combine fundamental economic logic with machine learning techniques. This enables us to view the world through two distinct lenses:

• Time series (TS): evaluating each country against its own history captures idiosyncratic dynamics. For example, is the current macro environment in Brazil improving or deteriorating relative to Brazil’s historical norm?
• Cross-sectional (XS): comparing all countries simultaneously mimics a global allocator’s relative value view, such as, is Brazil more attractive than Indonesia or South Africa right now?

These two signals form a combined view that balances country-specific dynamics with relative positioning across the universe, and accounts for the nuance of EM. For instance, it recognises that a rise in oil prices impacts a net-importer like South Korea very differently than a net-exporter like Saudi Arabia. It also explicitly models the external influence of major economic blocs, for example how US rates or Chinese industrial output spill over into specific EM economies based on their trade linkages.

Case study: navigating South Korea’s macro cycle (2021–2026)

To illustrate how the model navigates idiosyncratic and volatile macro cycles, we examine South Korea over the last five years. Korea presents a particularly rich test case. It is a market that exemplifies the interplay between global trade dynamics, domestic policy and sector-level sensitivities; it is highly integrated within global supply chains, exposed to commodity and semiconductor cycles, and subject to a domestic policy environment that periodically diverges from regional peers.

It is important to note that this single-country study is an illustration of the model's mechanics and its results should not be generalised across markets.

Figure 5. South Korea time series model score

Source: Man Numeric. Data as at Jul 2021 – Jun 2026.

Problems loading this infographic? - Please click here

The model’s TS score for South Korea significantly fluctuated throughout the 2021–2026 period (Figure 5), with the evolution of the underlying drivers revealing the model’s analytical depth:

2022: Credit risk gives way to trade stress
The model’s conviction shifted from sovereign credit concerns to external trade vulnerabilities. June’s rotation, from sovereign CDS carry to a bearish Utilities signal, was triggered by a deteriorating merchandise trade balance; by July that same trade balance had become the score’s largest detractor, reflecting the energy price volatility unleashed by the war in Ukraine. The score bottomed in August on weak trade and softening US business inventory data, with a modest energy-export pickup limiting the decline, then revisited that low in December as defensive Utilities returns and soft Energy and Materials exports pointed to a broader risk-off tone amid the global tightening cycle.

2023–2024: Domestic momentum leads, then fades
January 2023 brought the year’s sharpest positive move, a sector rotation that signalled domestic momentum was turning before the global backdrop caught up, even as weak US existing home sales and soft Materials exports held the level back. By April, defensive-sector underperformance and real GDP growth had taken over as the dominant driver, despite a Consumer Discretionary and Staples headwind from rising soft-commodity prices. The model stayed bullish throughout the rest of 2023 on a firming trade balance and softening core inflation, before a fresh sector rotation pulled the score back to its cycle low in September 2024.

Early 2026: Korea’s export engine regains its footing
The score’s most substantial improvement came in early 2026. A sudden drop in silver and gold prices against Korea’s Materials sector lifted it sharply in February; a brief rotation in March gave way to an April recovery driven by mixed terms of trade. Throughout, the level stayed anchored by a strengthening merchandise trade balance and energy-related export growth, evidence that Korea’s export engine had regained its footing on semiconductor strength and softening oil prices.

Over the 2008–2026 period, the macro timing model’s Korea position shifted repeatedly (Figure 6) as its underlying drivers rotated from sector-level dynamics to external trade and terms-of-trade signals, standing in contrast to the static, multi-year long tilt generated by the fundamental model (Figure 3).

Figure 6. South Korea: macro timing model (final) position over time

Source: Man Numeric. Data as at Jan 2008 – Jun 2026.

Problems loading this infographic? - Please click here

Evaluating the model

To evaluate the model's efficacy, we apply the same signal-to-position scaling and cap used for the fundamental model, only using data that was available at each point in time; the model's periodic retraining introduces no look-ahead into the backtest. Over the 2008–2026 backtest window used for Figure 4, this produced a cumulative PnL with a materially higher realised Sharpe ratio (0.85) than the standalone fundamental approach examined earlier, along with a 28% lower turnover. This provides evidence that a dynamic, macro-informed view of country risk can potentially deliver a more attractive risk-adjusted profile than a static cross-sectional signal (Figure 7) and has a place in a systematically managed EM portfolio.

The macro model's drawdown profile is also more contained over the same window: a maximum peak-to-trough drawdown of 4.6%, against 8.7% for the fundamental model. Because it is built on macroeconomic data rather than valuation or price information, the macro model's daily returns are only weakly related to the fundamental model examined above, with a correlation of just 0.08 over the 2008–2026 window, making it a largely uncorrelated source of alpha.

Both models rebalance daily, implying substantial annualised turnover. Translating either signal into a live, cost-efficient strategy would require addressing turnover (through position smoothing, lower rebalancing frequency, or explicit trading constraints) to seek to remain viable once realistic transaction costs are applied.

Figure 7. Macro timing model (combined TS+XS): cumulative PnL and summary performance

Source: Man Numeric, MSCI, Bloomberg. All EM Countries, data as at Jan 2008 – Jun 2026. Results are gross of transaction costs.

Problems loading this infographic? - Please click here

Problems loading this infographic? - Please click here

Conclusion

Stock selection remains the foundation of systematic EM investing, but it is typically not designed to navigate the country dimension. Country-level risk in EM is structurally elevated relative to DM, and the unintended exposures that emerge from bottom-up optimisation can quietly accumulate, becoming most consequential precisely when macro stress makes disciplined positioning matter most. A natural first instinct, blending valuation and momentum into a cross-sectional country signal, falls short in practice. Given both inputs are slow-moving, the resulting positions can stay static for years and the approach delivers an unremarkable Sharpe ratio below 0.5 despite meaningful turnover.

The country timing framework seeks to address this gap by viewing macro conditions through two complementary lenses – time series (each country against its own history) and cross-sectional (all countries against each other) – blended into a single view and continually retrained on point-in-time macroeconomic data. The case study illustrates the value of this approach. Where the static cross-sectional signal stays locked into the same long-run tilt, irrespective of the prevailing macro backdrop, the macro-driven positions all rotate across successive regime shifts. It is precisely this adaptiveness, which is structurally difficult for a valuation- or momentum-driven signal to replicate given how slowly those inputs move, that underpins the model’s attractive risk-adjusted performance over the long run.

Importantly, this is not a replacement for stock selection but a complement to it: a systematic overlay that transforms country risk from an unmanaged residual into a deliberate, data-driven source of alpha. As macro differentiation across EM continues to deepen, we believe the case for integrating this capability into institutional investment processes will only strengthen.