Government procurement accounts for roughly 12-15% of GDP across developed economies, dwarfing many sectors that receive intense attention in trade negotiations. Yet it remains the most consequential carve-out in the multilateral trading system—a vast economic territory where the foundational principles of non-discrimination, transparency, and national treatment apply only selectively, if at all.

This asymmetry is not accidental. When the GATT was drafted in 1947, procurement was explicitly excluded from Article III national treatment obligations, reflecting an early consensus that states should retain sovereign authority over how they spend public funds. Seven decades later, that exclusion has hardened into an architectural feature of global trade governance, sustained by a plurilateral rather than multilateral framework.

The institutional question is sharpening. As digital procurement platforms reduce transaction costs, as green public procurement reshapes demand signals across supply chains, and as state-owned enterprises blur the public-private boundary, the case for bringing procurement under coherent international discipline grows more urgent. Equally, so do the counterarguments—particularly from developing economies that view procurement preferences as among the few remaining instruments of legitimate industrial policy. Understanding this tension requires examining the architecture of the Agreement on Government Procurement, the substantive resistance to expanding it, and the parallel innovations emerging through regional trade agreements that may ultimately reshape the multilateral conversation.

The Agreement on Government Procurement, revised in 2012 and entering force in 2014, operates as a plurilateral instrument within the WTO framework—binding only its 22 parties (counting the EU as one) rather than the full membership. This structural choice reflects a calculated trade-off: deeper commitments among willing participants in exchange for narrower geographic coverage.

The GPA's substantive disciplines are notably ambitious. Parties commit to national treatment and non-discrimination for covered procurement, transparent tendering procedures, independent domestic review mechanisms, and electronic publication of opportunities above specified thresholds. The 2012 revision modernized these obligations to accommodate electronic procurement and clarified rules on technical specifications to prevent disguised discrimination.

Yet coverage operates through a positive-list approach. Each party schedules the central, sub-central, and other entities subject to disciplines, along with covered goods, services, and construction services, all above monetary thresholds calibrated by SDR. This à la carte structure produces dramatic asymmetries—one party may cover a federal procurement market many times the size of another's bound commitments, generating persistent reciprocity grievances.

Membership expansion has proven sluggish. China's accession negotiations, initiated in 2007, remain unresolved fifteen years later, largely because successive offers have failed to satisfy incumbent parties regarding state-owned enterprise coverage and sub-central procurement. Russia, Australia, and several others sit in similar limbo. The plurilateral model, designed to enable progress among the willing, paradoxically generates high entry barriers as incumbents demand ever-more comprehensive offers.

The institutional dilemma is structural. Lowering accession standards would dilute the agreement's substantive value to existing members; maintaining them perpetuates a club whose membership reflects 1990s economic geography rather than contemporary trade flows.

Takeaway

Plurilateral agreements solve the problem of negotiating among heterogeneous parties by creating clubs of the willing, but the same exclusivity that enables deeper commitments can ossify into a barrier against the very expansion the architecture was meant to facilitate.

The reluctance of most developing economies to join the GPA is neither accidental nor merely strategic foot-dragging. It rests on a coherent economic argument: discriminatory procurement constitutes one of the few remaining policy instruments compatible with WTO disciplines that can support domestic industrial development, particularly for sectors characterized by significant learning effects or scale economies.

The theoretical foundation traces to infant industry reasoning, refined through contemporary work on directed technical change and demand-side industrial policy. Public procurement creates anchor demand—predictable, large-volume orders that can sustain nascent domestic suppliers through the learning curve. Korea's electronics sector, Brazil's aerospace industry, and China's renewable energy manufacturers all benefited from procurement preferences during critical developmental phases.

The counterarguments are equally substantive. Empirical work by Evenett, Hoekman, and others suggests that procurement discrimination frequently fails its developmental objectives—generating rents for politically connected incumbents rather than productivity gains, raising costs for public services, and inviting retaliatory measures that harm export-oriented sectors. The fiscal cost of procurement preferences in some developing economies has been estimated at 15-25% above competitive benchmarks.

A more sophisticated middle position is emerging. Rather than binary inclusion or exclusion, scholars propose graduated GPA accession with explicit special and differential treatment provisions: extended transition periods, threshold flexibilities, reserved categories for SMEs and disadvantaged suppliers, and offset arrangements for technology transfer. The 2012 revised text already incorporates several such mechanisms in Article V.

The deeper question concerns institutional sequencing. Whether procurement liberalization should precede or follow industrial upgrading remains contested, with substantial evidence on both sides depending on sectoral characteristics and governance capacity.

Takeaway

Industrial policy debates often founder on the assumption that liberalization and developmental discretion are binary choices. The more productive frame asks which procurement disciplines are compatible with which developmental strategies under which institutional conditions.

While the multilateral procurement project stalls, regional trade agreements have become the laboratory for procurement liberalization innovation. The CPTPP, USMCA, EU-Japan EPA, and RCEP each contain procurement chapters that diverge meaningfully in scope, mechanism, and enforcement—generating a natural experiment in institutional design.

The CPTPP's Chapter 15 represents perhaps the most ambitious recent template. It extends coverage to sub-central entities for most parties, applies disciplines to a broad range of services including construction, and incorporates detailed transparency obligations including mandatory electronic notice publication. Critically, it links procurement disciplines to investor-state and state-state dispute settlement, creating enforcement teeth absent from the GPA framework, which relies on WTO dispute settlement and domestic review procedures.

USMCA took a notably different path. Its procurement chapter substantially narrowed coverage relative to NAFTA, with the United States and Canada effectively maintaining bilateral GPA commitments while Mexico's procurement obligations were significantly reduced. This represents a deliberate institutional retrenchment, reflecting domestic political concerns about Buy American policies and procurement-based industrial strategy.

RCEP's approach is more cautious still—committing only to cooperation and transparency on procurement, with no binding market access obligations. This reflects the inclusion of parties at vastly different developmental stages and the influence of ASEAN's consensus-driven negotiating culture.

These divergent approaches matter beyond their immediate parties. Regional templates often migrate to subsequent agreements and eventually influence multilateral negotiating positions. The procurement provisions in current EU and UK trade negotiations explicitly draw on CPTPP language, suggesting an emerging convergence around the high-ambition model—at least among economies prepared to accept its disciplines.

Takeaway

When multilateral progress stalls, regional agreements become institutional laboratories. The variation across their procurement chapters is not noise but signal—revealing which design choices different economies can sustain politically.

Government procurement liberalization sits at the intersection of trade economics, industrial policy, and institutional design—which is precisely why progress has been halting. The economic case for disciplines is strong; the political economy of implementation is genuinely difficult.

The most promising path forward likely runs through institutional layering rather than multilateral breakthrough. Continued GPA membership expansion, even at slow pace, combined with regional procurement innovations that generate replicable templates, may ultimately produce greater coverage than any grand bargain.

What requires sustained analytical attention is the calibration question: which disciplines, applied to which entities, above which thresholds, with what flexibilities, generate net welfare gains for parties at different developmental stages. The procurement frontier will not be settled by ideology in either direction, but by careful institutional engineering responsive to economic and political constraints alike.