Globalization is not ending

It becomes conditional

A common way to describe the present moment is to say that globalization is over (see here or here). While that might be tempting, it is probably wrong. The more interesting development is that while globalization is not quite disappearing, it is becoming conditional.

Markets still adapt and remain open. Trade remains vast, and supply chains, data flows, financial networks and technological ecosystems continue to bind states and firms together. But the assumptions that once made this openness feel stable are weakening. It no longer seems obvious that rules will remain predictable, that disputes will be handled procedurally rather than arbitrarily, that economic infrastructure will stay politically neutral, or that governments will restrain their own discretion.

I recently wrote a short essay (in Czech, available here) for the Prague-based Institute of International Relations where I describe this emerging condition as “managed openness.” I think this condition characterizes well a world in which economic liberalism survives, but under thicker layers of security logic, technological rivalry and political discretion.

This matters because the world economy is not going to return to autarky or closed blocs. Recent trade data complicate the deglobalization story. Global trade reached record levels in 2024, while at the same time the WTO reported that import restrictions in force covered 11.8 percent of world imports by mid-October 2024. The point here is not that trade is collapsing. The point is, rather, that the rules under which trade takes place are being rewritten.

The deeper issue is therefore normative. By “norms” I do not mean only legal rules or formal treaties. This is because norms also include shared expectations about what counts as normal behavior. For decades, the liberal order depended not only on legal institutions, but also on a background expectation that states would not constantly treat economic interdependence as a weapon. That expectation is now much harder to sustain. Sanctions, export controls, investment screening, industrial subsidies, tariffs and emergency measures are no longer rare exceptions. They are becoming ordinary instruments of statecraft.

Farrell and Newman’s concept of weaponized interdependence captures one part of this shift especially well. Global networks, they contend, are not flat spaces of mutual benefit. They contain hubs and chokepoints. States that control those nodes (say, in finance, data, technology or payments infrastructure) can use them to surveil, exclude or coerce others. The SWIFT system, dollar clearing, semiconductor controls and digital platforms all show that interdependence can be a source of leverage as much as cooperation.

But the current transformation goes well beyond arbitrary coercion. What matters most is that the expectation of neutrality is fading. Once business firms and governments begin to assume that access to markets, technologies or supply chains may be selectively revoked, behavior changes even before any formal breakdown occurs. Investment horizons shorten, risk premiums rise, trust becomes more regional. The growing costs of transacting then enter prices, contracts and political strategy.

This is why the current crisis of liberal order is not merely a crisis of American power, even if American power does remain central. Why? Because the postwar liberal order was never purely neutral. It was embedded in the so-called Pax Americana (sea-lane security, dollar centrality, institutional leadership and the willingness of the United States to underwrite a relatively open system) in which institutions (understood as rules, norms, and enforcement mechanisms) weren’t merely decorative. In this sense, Keohane’s seminal point in After Hegemony still remains useful. Institutions can facilitate cooperation among self-interested actors by reducing uncertainty and transaction costs (North, after all, claimed as much), even when hegemonic dominance erodes (North, Weingast and Wallis, in fact, later clarified that the point is not hegemony but “dominant coalitions”). Ruggie’s idea of “embedded liberalism” added another layer to these insights because it made clear that the postwar order worked because openness was politically embedded in domestic compromises and shared expectations, not because markets floated above politics.

The difficulty today is that the old embedding is fraying from several directions at once. Security has returned to economics. Domestic politics rewards visible intervention. Repeated crises (financial, pandemic, energy, inflationary and geopolitical) make emergency governance feel permanent. At the same time, despite the recent Hungarian election results, political liberalism is weakening more visibly than economic liberalism. V-Dem, for instance, reports that autocracies outnumbered democracies in 2024 for the first time in more than two decades. And Freedom House records a nineteenth consecutive year of declining global freedom.

The result, it seems to me, is not the end of liberalism but its fragmentation. In artificial intelligence, trade, data and industrial policy, we are already seeing overlapping regulatory nodes rather than one universal regime (consider, for instance, the EU-style legalism, OECD-style standards, UN soft law, security clubs, national sovereignty claims and private supply-chain governance). In my IIR essay, I suggest that the future will be shaped by competition among these normative nodes (see here for more on how I think about the meaning of competition), not by a simple choice between globalization and deglobalization.

For Europe, this creates a particularly difficult problem. “Economic security” is real. Small open economies cannot pretend that technology, finance and supply chains are politically neutral. But if security becomes a blank cheque for permanent subsidies, arbitrary regulation and defensive industrial policy, Europe may sacrifice the very institutional qualities that make it attractive (by these I mean legal predictability, independent courts, openness to talent and capital, and the trust needed for long-term investment).

So the question is not whether state governments should or should not intervene. They will. The question is, rather, whether intervention remains rule-bound, proportionate and reviewable or whether arbitrary exceptions become the norm.

Globalization is not ending. But the assumption that liberal openness can be treated as a stable background condition is evaporating in front of our eyes. What comes next will depend on whether “managed openness” can preserve enough predictability to avoid becoming merely another name for politicized uncertainty.