Liquidity, Regionalization & Asset Allocation

For global investors, 2025 was one of the most undercurrent-filled years of the 21st century. Unlike the bursting of the dot-com bubble in 2001 or the global financial crisis in 2008, markets in 2025 did not experience a prolonged, large-scale liquidation cycle or a “storm-like” sequence of relentless crashes.

Yet it is clear that, amid geopolitical uncertainty, uncertainty over US fiscal and monetary policy, uncertainty across multiple countries’ economic fundamentals, and the ebbing of globalisation in favour of regionalisation, equities, bonds, commodities and crypto have all been pricing in a future that is more cautious and more defensive.

Against that backdrop, liquidity allocation has become less concentrated in equities and bonds than it once was. Commodities, FX and rates attracted greater attention in 2025. At the same time, investors have been steadily reducing leverage and trimming exposure to higher-risk assets—one of the direct reasons the crypto bull market ended in Q4 2025.

So, where do markets go in 2026? As in 2025, implied expectations embedded in derivatives-market data have already offered an answer.

Liquidity: Not Abundant

At the start of 2025, one major “bullish” factor in investors’ minds was Donald Trump’s formal inauguration. The prevailing view was that Trump would trigger more rate cuts, inject more liquidity into markets, and drive asset prices higher.

Indeed, between September and December 2025, amid “concerns about a weakening labour market”, the Federal Reserve delivered three “defensive” rate cuts and, in December, announced the end of quantitative tightening. But this did not produce the liquidity flood investors had hoped for.

From October 2025 onwards, the Effective Federal Funds Rate (EFFR) gradually moved towards the midpoint of the “rate corridor”. In the following months, EFFR crossed that midpoint and drifted towards the upper bound of the corridor—hardly a sign of easy liquidity.

EFFR is the core short-term market rate in the US. It reflects funding liquidity conditions in the banking system and how the Fed’s policy stance (hikes or cuts) is transmitted in practice. In relatively loose-liquidity regimes, EFFR tends to sit closer to the lower end of the corridor, as banks have less need for frequent overnight borrowing.

In the final months of 2025, however, banks clearly faced liquidity tightness—a key driver of the rise in EFFR.

The SOFR–IORB spread further highlights the degree of stress. If EFFR primarily reflects cash-market conditions, SOFR, secured funding collateralised by US Treasury securities, captures a broader liquidity shortage. Since October 2025, SOFR has remained above the Interest Rate on Reserve Balances (IORB), indicating that banks have been willing to pay a higher rate premium to “bid” for liquidity.

Notably, even after the Fed stopped shrinking its balance sheet, the SOFR–IORB spread did not fall sharply in January. One plausible explanation is that, during 2025, banks deployed a significant share of their liquidity buffers into financial investments rather than extending credit to the commercial, industrial, and real estate sectors.

Over the past year, commercial and industrial lending contracted meaningfully versus 2024, and consumer credit showed similar weakness. By contrast, VettaFi data suggest that margin debt rose 36.3% over the past year, reaching an all-time high of $1.23T in December 2025, while investors’ net debit balances also expanded to $ -814.1 billion—broadly matching the pace of margin debt growth.

As liquidity requirements grow to push markets higher, the banking system is showing signs of strain, and demand for short-term funding has increased. The fix is straightforward: either reduce margin lending and pull liquidity back, or obtain liquidity support from the Fed and the repo market.

For the economy as a whole, the first option is preferable—lower system-wide leverage and strengthen resilience in banks and the financial system—but it would also imply lower valuations and a sharp equity sell-off. Given the midterm-election backdrop, the White House is unlikely to accept that path.

As a result, in 2025 alone, the repo market expanded from roughly $6T to more than $12.6T—over three times its size during the 2021 bull market. In 2026, repo may need to expand further to support equity-market performance.

Repo transactions typically use US Treasuries—“high-quality assets”—as collateral. Historically, Treasury notes (T-notes) have been the most important form of collateral. But since mid-2023, that has changed, in part because the issuance and outstanding stock of Treasury bills (T-bills) has increased in an “exponential” fashion.

This is not benign: a rising share of T-bills in total government debt often signals deteriorating sovereign credit perceptions. As investors begin to doubt a government’s repayment capacity, they may become less willing to buy long-dated bonds at relatively low yields.

To reduce debt-servicing pressure, the government leans more heavily on T-bill financing—raising the T-bill share further and reinforcing investor doubts in a vicious cycle.

A higher T-bill share has another consequence: liquidity dynamics become less stable. Since a large portion of the liquidity supporting equities is channelled via repo, a greater reliance on T-bills implies more frequent rollovers and a shorter average liquidity “life”.

With overall leverage and margin debt already pushing beyond historical peaks, more frequent and more violent liquidity swings weaken the market’s shock-absorption capacity—setting the stage for potential cascading liquidations and large price moves.

In short: the quality of USD liquidity deteriorated markedly in 2025, with no clear sign of improvement so far.

So, in this macro context, how have investors’ expectations and portfolios changed?

Risk Premia and “Strict Diversification”

One cost of poorer-quality USD liquidity is that USD-based long-term funding costs remain elevated. This is intuitive: as USD asset markets become more fragile, US Treasury debt expands sharply (reaching USD 38.5 trillion by December 2025), and US fiscal, monetary and foreign policy turn more uncertain and less predictable, the perceived probability of systemic risk rises over time—prompting long-term Treasury investors to demand greater compensation.

Since long-term financing rates are typically anchored to the 10-year Treasury yield, it is telling that the 10-year yield fell only 31 bps over the past year—far less than the 75 bps decline in policy rates. This implies long-term funding costs stayed above 4%.

High funding costs constrain positioning. When a risk asset’s implied forward return falls below Treasury yields, holding that risk asset long-term becomes unattractive. Crypto is a textbook example: as implied forward returns declined, investors progressively reduced exposure, and the market moved into a bearish phase.

Compared with expensive long-term liquidity, short-term liquidity funded via T-bills is materially cheaper. But T-bill funding is also short-duration, creating an environment naturally favourable to speculation: investors can borrow short, apply high leverage, push prices up quickly and exit. Markets may look buoyant in the short run, but speculative froth makes rallies difficult to sustain—something clearly visible in the liquidity-sensitive crypto market.

Meanwhile, after decades, “strict diversification” made a comeback in 2025. Unlike the traditional 60/40 approach, liquidity has been spread across a broader set of instruments rather than confined to USD assets.

In fact, throughout 2025, investors steadily reduced the share of USD and USD-pegged assets in portfolios. Although persistent net outflows did not visibly hit US equities, incremental liquidity was allocated more heavily to non-US markets.

Assets tightly pegged to USD or USD-denominated leverage (crypto, WTI oil, the dollar itself) underperformed, while assets less tied to the dollar (such as precious metals) delivered far stronger performance than other major asset classes.

Notably, simply holding euros or Swiss francs performed no worse than holding the S&P 500. This suggests a profound shift in investor logic—one that goes beyond a single business cycle.

The New Order

What most deserves reassessment in 2026 is not a linear question like “will growth be stronger?”, but rather the fact that markets are adopting a new pricing grammar. Over the past two decades, returns often rested on two implicit assumptions: first, supply chains were organised around maximum efficiency, suppressing costs and stabilising inflation; second, central banks provided powerful backstops during crises, systematically compressing risk premia.

Both assumptions are now weakening. Supply chains increasingly prioritise control and redundancy; fiscal and industrial policy appears more frequently in profit models; and geopolitics has shifted from tail risk to constant noise. “Regionalisation” is less a slogan than a change in the constraint set facing the global economic system.

In this framework, the key is not to bet on a single direction, but to realign exposures to three more reliable “hard variables”: supply constraints, capital expenditure, and policy-driven order flow.

Together, they point towards a set of assets: commodity-linked equities, the AI infrastructure chain, defence and security themes, and select non-US markets that improve portfolio correlation structures. At the same time, the core question in rates and government bonds is no longer “how much tailwind will rate cuts bring?”, but how the new term structure reshapes the distribution of returns.

Regionalisation: Not “Decoupling”, but a New Cost Function

Equating “regionalisation” with “full decoupling” tends to understate its true impact. A more accurate description is that globalisation’s objective function has shifted from “efficiency at all costs” to “efficiency under security constraints”.

Once security becomes a binding constraint, many variables that previously sat outside valuation models—supply-chain redundancy, energy security, access to critical minerals, export controls on key technologies, and the rigidity of defence budgets—begin to enter discount rates and earnings expectations in various forms.

This produces two direct consequences for asset pricing. First, risk premia become less likely to revert to structurally low levels: political and policy uncertainty becomes an everyday variable, and markets require greater compensation. After all, nobody wants to bear “Cuban equity risk”, and today, even in US equities, that “Cuban equity risk” is no longer zero.

Second, global beta explains less, while regional alpha matters more: under different blocs and policy functions, the same growth and the same inflation can produce very different valuations and capital flows. For allocators, diversification in the age of regionalisation looks less like splitting assets evenly by country and more like diversifying across supply-chain position and policy elasticity.

Equities: From “Buying Growth” to “Buying Location”

If 2010–2021 equity allocation was largely about “buying growth and falling discount rates”, 2026 is more about “buying location”. “Location” refers to where a market sits on three maps: the resource map, the compute map and the security map. As the world emphasises supply-chain autonomy and critical infrastructure security, markets positioned at key nodes are more likely to earn a structural premium, even if their domestic macro picture is imperfect.

In an era where security is the top priority, increasing inventories of gold, silver, copper and other non-ferrous metals can be rational even if they are not immediately needed. Supply chains can be disrupted without warning (as last year’s trade tensions showed), sharply raising costs and forcing major countries to hold larger mineral reserves against potential shocks.

Structurally rising demand for critical minerals, combined with long-cycle supply constraints, makes commodities behave more like “supply-side assets” than mere mirrors of the traditional business cycle. Options-market implied expectations reflect this: although investors see signs of overheating in some non-ferrous metals markets (particularly silver), traders still anticipate further upside potential for gold over the longer run.

This logic also provides a clearer allocation case for equities in resource-rich countries. Copper-linked equities—Chile is a prime example—partly reflect foundational shifts in electrification and in demand for industrial infrastructure.

Precious-metals resource equities—South Africa is a typical case—combine commodity upside with the double-edged nature of risk premia: when commodities rise, profits and the currency may reinforce each other; when risk rises, politics and external financing conditions can amplify volatility. For portfolio construction, resource-country equities are better understood as a “supply-constraint factor” than simply emerging-market beta.

Another central theme is AI. AI discussions are easily pulled towards application-layer narratives, but allocators should focus on balance-sheet realities: compute, energy, data centres, networks, and cooling. These links share two traits: higher capex visibility and often benefit from joint support from policy and industry.

Rather than treating AI as another software-valuation game, it may be more robust to view it as a new wave of infrastructure build-out. Higher compute density ultimately translates into greater power and engineering demand, shifting more of the return distribution upstream and into midstream “real-economy” segments.

Under regionalisation, computing infrastructure is also spreading geographically. Higher security redundancy and localisation requirements increase the strategic value of key hardware and intermediate goods.

Markets such as Korea, positioned at the industrial interface of global compute infrastructure via semiconductors and critical electronics, are often seen as more direct equity expressions of the AI capex cycle. For portfolios, the value of this exposure is not only “faster growth”, but “more observable capex and more stable policy support”.

In addition, “defence and security” has returned to investors’ agendas for the first time since the end of the Cold War. Influenced by Trump’s “Donroeism” and the Russia–Ukraine war, both the US and Europe are placing defence higher on the priority list.

The distinctive feature of defence assets is that demand does not come from marginal household consumption; it is closer to a fiscal function constrained by national security. Once budgets step up, the political resistance to reversing them is greater, so order visibility is typically stronger. This gives defence-related equities a more defensive allocation role in a regionalised world: when conflict and sanctions risk rise, they can add resilience at the portfolio level.

That said, defence-sector price sensitivity often runs ahead of fundamentals: event-driven repricing followed by mean reversion is common. A more robust framing is to treat it as a portfolio “tail insurance” or risk-hedging factor, rather than a linear-growth core holding. Its value lies in reducing drawdowns, not in guaranteeing outperformance every quarter.

Hong Kong equities and mainland China assets are another area worth considering. Labelling them simply as “cheap” is insufficient; their allocation value stems from two factors. First, pricing often bakes in pessimistic expectations early, leaving room for rebalancing.

Second, their policy function and sector composition differ from those of US and European assets, potentially improving portfolio correlation structure. In the age of regionalisation, correlations do not automatically fall; they can rise during risk events. Structurally different assets can therefore provide more meaningful hedging.

Rates and Treasuries: Keep the Curve Steepening

The core tension in 2026 rates markets can be summarised in one line: the front end is more a function of the policy path, while the long end is more a container for term premia.

Rate-cut expectations do help front-end yields decline, but whether the long end follows depends on whether inflation tail risks, fiscal supply pressure and political uncertainty allow term premia to keep compressing. In other words, long-end “stubbornness” may not mean markets have mispriced the number of cuts; it may mean markets are repricing long-run risk.

Supply dynamics amplify this structural difference. Changes in US fiscal funding composition directly affect supply–demand across maturities: the front end is easier to absorb when money markets have capacity. In contrast, the long end is more prone to pulse-like volatility driven by risk budgets and term premia.

The portfolio implication is clear: duration exposure should be managed in layers, avoiding a one-path bet on “inflation fully disappearing and term premia returning to ultra-low levels”. Curve-structure trades (for instance, steepening strategies) persist not merely because of superior trading skill, but also because they align with the different pricing mechanisms of the front and long ends.

Crypto: Separate Accounting for “Digital Commodities” and Secondary Risk Assets

In 2026, the key for crypto is not simply “will it rise?”, but sharper internal differentiation. Bitcoin is more readily understood as a non-sovereign, rules-based supply asset that is portable across borders—a “digital commodity”. Under a regionalisation narrative, it is more likely to absorb demand for alternative payment systems and hedges.

By contrast, a subset of tokens that behave more like equity-style risk assets are priced more on growth stories, ecosystem expansion and risk appetite. When risk-free yields remain attractive, regulation becomes clearer, and traditional capital markets offer more mature funding and exit channels, equity-like tokens must offer higher risk compensation to justify allocation.

As a result, crypto allocation is better approached via “separate books” rather than a single basket: place bitcoin in a commodity/alternative-asset framework, using small weights to obtain portfolio-level convexity; treat equity-like tokens as high-volatility risk assets with stricter return hurdles and clearer risk budgets. The core of the regionalisation era is not to embrace every new asset, but to identify which assets remain more explainable under the new constraints.

Use Hard-Constraint Assets as the Core, Use Structural Divergence as the Return Engine

Putting the above together, a 2026 portfolio looks more like managing a set of “hard constraints”: supply constraints restore the strategic role of commodities and resource equities; capex supports earnings visibility across the AI infrastructure chain; policy-driven orders enhance the resilience of defence and security; the return of term premia reshapes the distribution of duration returns; and select non-US assets provide reflexive hedging through valuation structure and policy functions.

This does not require perfect prediction of every event. On the contrary, the rarest skill in the age of regionalisation is to place the portfolio in a position that relies less on flawless forecasting: let hard assets and infrastructure absorb structural demand; let curve structures absorb structural divergence; and let hedging factors absorb structural noise.

Trading in 2026 is no longer about “guessing the answer”, but about “acknowledging constraints”—and rewriting asset-allocation priorities accordingly.

Disclaimer: The information provided herein does not constitute investment advice, financial advice, trading advice, or any other sort of advice, and should not be treated as such. All content set out below is for informational purposes only.