The Software the State Imagines

In quantum mechanics, the instrument decides which reality can be recorded. A detector calibrated to measure one property is blind to all the others: the system responds only in the basis the apparatus interrogates, and what happens outside that basis keeps existing — just without leaving a trace in the record. Meanwhile, what nobody measures remains in superposition: all its possible versions coexist, left to probability, until a measurement forces the collapse into one concrete outcome.

That rule governs more than electrons. A public policy also measures with the instruments it designs: its legal categories are the measurement basis, and any economic reality that falls outside those categories will keep happening without leaving a trace in the tax record. And observing has a cost — whoever measures becomes part of the system. Dominican technology policy making has just produced a perfect object of study.

The Ley 30-26, enacted on June 18, redrew the tax treatment of software transactions. The new regime sorts the software universe into three categories: acquisition with transfer of intellectual property, free of withholding; licensing and leasing —subscriptions, SaaS and equivalent arrangements— subject to a 15% withholding on the gross value when the provider is abroad; and associated services, taxed with ITBIS when invoiced separately, under the current regime for intangibles. The design looks orderly on paper. The problem is that it legislates for a market that exists less and less and taxes the one that exists more and more.

The rewarded category —buying software together with its property rights— is today an almost mythological figure in technology commerce. No global provider sells its intellectual property: Microsoft, SAP, Oracle, OpenAI and the whole catalog that sustains Dominican companies' operations is rented, subscribed to, metered by consumption. The zero-withholding tax benefit thus applies to a transaction that occurs only in exceptional cases —custom development, asset purchases— while the 15% withholding on gross value falls on practically all corporate technology spending. For an economy that imports nearly all its software, that withholding works as a silent tariff on digital inputs: the foreign provider rarely absorbs it, the contract is adjusted, and the extra cost ends up on the local buyer's balance sheet. The reduction from 27% is real relief and brings the country closer to regional standards; the incentive architecture, however, still points in the opposite direction from the transformation the State itself preaches.

The second confusion is subtler and costlier. The regime leaves licensing —local and cross-border— outside ITBIS because it involves intangible goods, and keeps the tax on separately invoiced services. Translated into market structure: the proprietary license gets tax relief, while implementation, integration, migration and support —the natural territory of local talent and of the open-source ecosystem, which lives off services because its license is free— bear the 18%. Whoever sells licenses wins twice: their product is exempt from ITBIS, and so are their ancillary services, invoiced alongside the license. Whoever sells knowledge with no license attached pays in full. The architecture, without intending to, builds an inverted industrial policy.

The figures in the T al Cubo® report, Perspectivas de la Digitalización República Dominicana 2025, show on whom it rains. Of public ICT spending of 2025 —RD$2,784 million across 179 providers,— the top 12 companies concentrate 73.8%, and within that group two firms account for almost a third of the total, both leveraged on proprietary licensing by exclusivity. The tax treatment consolidates exactly that engine: it makes the incumbent who invoices licenses cheaper and the competitor who invoices services more expensive. The concentration the data describes finds in that architecture an unintended ally.

Here the Nash equilibrium operates with textbook precision. Given these rules, each player's optimal response is predictable: the provider will package every service as an accessory to the license to inherit the exemption; the buyer will keep renting from abroad with the 15% built into the price; nobody will migrate to the rewarded category because the global market simply lacks sellers willing to transfer their intellectual property. Each participant maximizes their gain given what the others do, none improves by moving alone, and the resulting equilibrium delivers less revenue than projected, less competitiveness for local services and more depth in concentration. A mechanism designed without market observers is a system that chooses its effects by probability: the collapse arrives with enforcement, and it arrives against the clock of competitiveness.

The good news is that the regulations for the law are still being written. That intermediate state is the window physics grants to policy: the system still admits measurement before the collapse. Here it is worth distinguishing two ways of looking. Oversight will observe the rules once published, draw up records and issue reports when the effects are already running. Parrhesia —the frank speech that Foucault recovered from the Greeks— measures now, from inside the ecosystem, accepting the cost of pointing out that the tax categories describe an imaginary market. Technology chambers, local developers, service firms and business associations have consultation spaces ahead of them that are worth more than ten years of later observations. Mechanisms reviewed after they have entered into force get corrected by shoving.

The country deserves a software tax regime designed around the market that exists: one where the digital input is treated as productive investment, where the local service competes on equal terms with the imported license, and where the legal definition follows economic reality instead of waiting for it. The Ley 30-26 opened that door by modernizing the rates; its regulations will decide whether the framework completes the modernization or leaves it halfway. Because in technology policy, as in physics, what nobody observes is left to probability — and the poorly calibrated instrument never registers the reality it claimed to seek.

The State design pursuing the Goal 2036 taxes real software to reward an imaginary software production, and will end up collecting from the economy that the very goal kept from growing —a bill that will be paid by those left out of the game.