From Proposal to Decision: How to Validate and Advance a Simpler Tax System

A proposed decision process—not an invitation to an existing pilot. None of these research stages postpones current legal duties or practical e-invoicing preparation. Each reform must independently meet its own legal, economic and operational conditions.

The final test of a reform is not whether its diagram is elegant or its supporters sound confident. It is whether the proposal survives legal scrutiny, credible modelling, operational testing and an honest comparison with less disruptive alternatives. A simpler tax system deserves political effort, but effort should follow evidence rather than compensate for its absence.

The ideas discussed here form a programme for investigation: a usable public baseline for electronic invoicing, proportionate commercial services, clearer responsibility, wider reverse charge where justified and legally possible, and a differentiated receipts-based alternative to corporate profit tax. They are not a package in which every element must succeed for any element to be adopted. Separating them is essential to responsible decision-making.

Give each proposal its own burden of proof

E-invoice delivery, reporting to the tax authority and the calculation of liability are separate functions. A business might benefit from standardised delivery even if a proposed tax-base replacement fails. A public baseline channel could be worthwhile even if a universal reverse charge proves unavailable. Bundling everything into an indivisible political promise would make useful improvements hostage to the most uncertain component.

For each proposal, publish the problem, the intended mechanism, affected groups, alternatives and conditions for rejection. The Commission’s Better Regulation toolbox explicitly identifies evidence-informed policymaking and legal basis, subsidiarity and proportionality as elements of its guidance.11 The practical programme below is my proposed application of those disciplines, not a claim that the Commission has endorsed this tax package.

The evidence standard should match the claim. A prototype can show that a correction message reaches a recipient. It cannot establish that a national tax reform raises revenue. A static tax calculation can show different liabilities on recorded activity. It cannot demonstrate how businesses, consumers or organised fraud would respond after implementation.

Stage one: establish the legal perimeter

Before a live tax pilot, commission a public legal memorandum with a clear distinction between conclusions, assumptions and unresolved questions. For VAT, it should identify the relevant existing options, the proposed scope beyond them and the necessary European procedure. The Commission describes supplier liability as the basic rule, defined reverse-charge exceptions and a separate route for authorised derogations.9 10

Do not present Article 395 as an automatic permission slip. Nor should a voluntary pilot be assumed lawful merely because participating companies consent. Consent does not itself change statutory liability. The Council’s historical treatment of generalised reverse charge as a conditional, time-limited measure is a warning against turning past proposals into current national authority.1

A receipts-tax replacement needs a different legal workstream: the definition and character of the levy, differentiated treatment, cross-border allocation and interaction with international obligations. The EU minimum-tax directive covers specified multinational and large domestic groups.7 The proposed replacement cannot assume that every profit-related calculation disappears for those groups simply because the domestic tax base changes.

Stage two: specify a decision-ready design

Replace slogans with worked rules. Identify the taxable person, tax base, timing, classification, deductions or credits, minimum liability, appeal route and transition provisions. For the digital component, identify who must send what to whom, which event proves delivery, what happens during an outage and who pays for correction.

Consider a hypothetical small wholesaler changing its invoicing software while also assessing the receipts-tax proposal. It needs to know how returned goods alter the new base, whether an advance counts as revenue, how a customer-status error is corrected and whether historical documents remain accessible after changing providers. A diagram showing “business to platform to state” does not answer those questions.

The design should also separate an optional value-added service from the means of meeting the public obligation. My preference is a direct public baseline, with commercial automation and support available by choice. That does not rule out government procurement of technical infrastructure. It requires the government to explain why a taxpayer must enter a separate commercial relationship if the public duty could be met otherwise.

Stage three: conduct shadow testing

A shadow exercise applies proposed rules to authorised historical or controlled test data without changing actual tax liabilities. It can expose missing classifications, inconsistent transitions and unexpected distributional effects. The input dataset, code version and assumptions should be preserved so that an independent reviewer can reproduce the result within the lawful access arrangements.

For the wholesaler, test a sale, a return, an unpaid invoice, a mixed-activity transaction and a correction spanning the proposed start date. For an exporter, trace refund timing and supplier payments separately. For a low-margin firm, show the charge during a difficult year. Include failures deliberately; a test set containing only straightforward invoices is a demonstration, not a readiness assessment.

Use observed data for empirical results and clearly labelled invented cases for testing logic. Synthetic cases are valuable because their intended outcomes are known. They are not evidence of national revenue or fraud prevalence. A report should make that boundary visible rather than allowing a successful software test to become a claim of economic validation.

Stage four: evaluate before asking for commitment

Publish fiscal, cash-flow, compliance and distributional results separately. Slovakia’s Financial Administration already distinguishes its declared corporate-tax datasets from cumulative cash-revenue reporting.12 13 A reform assessment must preserve such distinctions rather than build a single impressive total from incompatible measures.

For invoice access, compare my preferred public baseline with an improved provider-led model on equal scope and time. Publish total public and business costs, small-business access and administrative work, replaceability, security and operational evidence, recourse and competition. Require tested migration and operator takeover, not an exit clause alone. Publish sensitivity analysis as well: test higher implementation and operating costs, lower take-up and smaller or delayed benefits, separately and together. Show whether the preferred option still outperforms alternatives under those worse assumptions and where the decision would change. Independent reviewers should challenge assumptions, missing data and failure modes. Where data are protected, provide controlled verification rather than an unverifiable claim of success.

Business consultation should not be limited to the largest likely beneficiaries. Include small firms, loss-making firms, accountants, software providers including smaller providers, resellers and integrators, employees and consumers through appropriate representation. Participation grants neither a veto nor automatic protection of existing revenues. Ask the same concrete questions of each group. A disagreement about classification or payment timing is information about the proposal, not evidence that the critic profits from complexity.

Stage five: a lawful live test, if justified

Only after legal authority, operational readiness and an evaluation plan are established should a live pilot be considered. Define who participates, what the comparison is, which obligations change and how adverse effects are handled. Selection matters: a pilot made up only of enthusiastic, technologically mature businesses may say little about nationwide implementation.

Set stopping rules before launch. Examples include unacceptable correction backlogs, unresolved legal defects, material fiscal exposure or a burden on vulnerable firms beyond the agreed tolerance. The precise thresholds should be chosen by accountable decision-makers with technical advice, not invented after the results are known. Stopping a technical function does not reverse tax liabilities already incurred. Before launch, require a costed and rehearsed stop-and-takeover plan covering accrued obligations, corrections, records, contracts, continued support and the legally permitted next regime. Specify who executes and funds parallel operation, rerouting and evidence retention; a return to the old process is available only where the law permits it.

The strongest objection is that this sequence could become endless analysis. That risk is real. The answer is deadlines and bounded questions, not removing the tests. Each stage should have a named owner, a defined output and a decision date. For each reform, distinguish the policy proposer, the body legally empowered to change the rules, the operational guarantor and the evaluator. The tax administration is not automatically all four. Providers need decision dates and controlled changes to the specification, not an open-ended promise of reversibility. Uncertainty should be reduced where possible and explicitly accepted where unavoidable; it should not become either a permanent veto or a blank cheque.

Build support without claiming opponents are corrupt

Tax reform changes burdens, responsibilities and commercial opportunities. Opposition may reflect legitimate concerns about investment, low margins, legal uncertainty or transition costs. An argument does not become wrong because an adviser makes it, and a proposal does not become right because a large exporter supports it.

Public communication should state what is being replaced, what remains, who may pay more and what has not been proven. I accept a proportionate minimum annual tax in principle, not merely the idea of studying one. That is not endorsement of a particular amount: payment capacity, clear limits, reliefs and its relationship to the main tax must be assessed. So does the fact that higher revenue from a receipts-based replacement remains an unverified hypothesis, not money already available for spending.

Decide, publish and retain the ability to correct

The final decision should identify which findings justify adoption, which protections are necessary and which alternatives were rejected. Publish an implementation plan with accountable owners, support arrangements and post-implementation review. Do not declare the project complete when legislation passes or the first invoice is transmitted.

I continue to support a public baseline and optional commercial services, and to advocate testing the tax alternatives. But if the public design fails reasonable-cost replacement, competitive access or security and operational tests, it should not be adopted on the strength of that preference. An improved provider-led model may win; a narrower tax reform may be better. Publish that result and its reasons. Reform earns authority by correcting proposals in response to evidence, not by concealing the author’s starting position or protecting it from defeat.

What a business can verify now: its own scope of obligations, software support, invoice reception, export and the contact for failed processing. What remains a proposal: the public baseline and the tax reforms described here. Support for them is not a reason to postpone preparation or change today’s tax treatment.