The centre for tax analysis in developing countries

Overseas Development Institue Institue for Fiscal Studies

One of the key functions of tax policy units is to develop and assess tax policy proposals. What problem do the proposals try to address? Is there evidence that the proposed reform would actually help to tackle the issue identified, and do so more effectively than other policies? Which parts of society would be particularly affected by the proposal – either positively or negatively? And how much revenue would the reform yield or cost?

Since TaxDev was established, this type of appraisal and costing of tax policy proposals has been a major part of our work with our partners at the Ministries of Finance of Ethiopia, Ghana, Rwanda and Uganda. Such analysis has fed into annual Budgets and built on broader reviews of existing tax policies and systems.

In doing this, we have strived to make sure all policy proposals are subject to proper scrutiny, with the findings of analysis clearly and consistently set out. This helps to ensure that the analysis can feed into decision-making by senior policymakers and politicians.

To this end we have drawn on the approaches for policy appraisal and costing used in organisations like the UK’s HM Treasury (i.e. Finance Ministry) to develop structured processes and templates for both policy appraisals and policy costings. These set out various steps to assess the extent to which a policy proposal would likely meet its aims and its other impacts on taxpayers and the economy (the policy appraisal) and how it would affect revenues (the policy costing).

To help both our partner governments and other governments to follow this process, we have published two manuals – one on policy appraisal, and the other on policy costing. As well as explaining the various steps involved in each process, these also discuss the issues that one might consider in the context of example reforms to value-added tax and corporate income taxes. A spreadsheet showing several worked examples of policy costings and suggested template documents for presenting findings are also available to download.

The policy costing process

Our policy costing manual divides the costing process into three stages:

  1. A static costing that assumes taxpayer behaviour is unaffected by the reform. This is the simplest type of costing, requiring no assumptions about how responsive taxpayers are to tax rates and rules. It is a useful starting point, and depending on the estimates and the policy objectives and constraints, it may be enough to ‘rule out’ a particular policy from further consideration.
  2. A behavioural costing allows for behaviour to change along at least one margin for taxpayers that would be directly affected by a reform. The behavioural effects that could be incorporated will differ for different policies but could include changes to labour supply or taxable income for a change in income tax rates, or changes in consumer demand for a change in indirect taxes. Behavioural costings are more complex and require sourcing appropriate estimates of how responsive taxpayers are to tax policy. In general, behavioural responses will not be large enough to reverse the sign on a static costing – e.g. to make a tax cut pay for itself – but sometimes they might, and otherwise they may still have notable effects on revenue costs or yields.
  3. A wider economic effects costing. For many reforms, a behavioural costing will be sufficient. However, some reforms are large enough that their impacts on the wider economy – e.g. on the level of GDP – are important enough to try to incorporate into a costing. In doing this it is important not to double-count certain responses in both stage 2 and stage 3.

In each stage it is also important to be clear about the assumptions made – e.g. on future growth in tax bases and the scale of any behavioural elasticities or fiscal multipliers – and the uncertainties around them. Policymakers need to know how confident they can be about any costing estimates in order to take informed decisions. 

The policy appraisal process

Costing a policy is just one element of appraising its potential impact, even if in some cases it is the main outcome of interest for policymakers. A full policy appraisal should take a broader view of the likely impact of a tax policy change, drawing on available data, consultation, and existing evidence from similar policy changes to understand economic, distributional and administrative implications.

The constituent steps of this evidence-gathering process can include a wide array methods and sources. Regardless of the type of evidence used, however, a systematic approach to completing and presenting a policy appraisal provides structure to the exercise and clarity in summarising conclusions. Our appraisal manual sets out a total of ten steps to be completed in a clear and concise policy appraisal document. These are:

  • Clearly defining the policy change being appraised;
  • Identifying the rationale for the proposed change;
  • Assessing whether evidence suggests that the proposed change would in fact address the policy problem that provides its rationale;
  • Estimating the revenue cost or yield of the policy for the government;
  • Setting out which groups would be particularly affected by the proposals and in what ways;
  • Examining other potential impacts, including unintended ones;
  • Discussion of legal and administration issues;
  • Considering whether alternative policy proposals may be better able to address the policy problem in hand;
  • Documenting key assumptions and uncertainties underlying the appraisal;
  • And potential plans for ex-poste monitoring and evaluation if a policy is implemented.

Concluding thoughts

The manuals do not provide detailed instructions on how to cost and appraise specific policies – indeed they could not unless they were thousands of pages long. The data, methods and issues to consider will differ for different policies, and analysts at Ministries of Finance and Revenue Authorities need to use their knowledge of policy analysis and their tax systems to select the appropriate data, evidence and assumptions on a case-by-case basis.

However, our experience with our government partners suggests that embedding this bespoke analysis in a consistent and standardised overarching approach can help to embed evidence-based tax policy analysis in the tax policymaking process.


Published on: 16th July 2021

Skip to main content