Pre-election policy costings

Last week Robert Chote, Chair of the Office for Budget Responsibility, told the Treasury Select Committee that allowing the OBR to cost party policies before an election would ‘offer the prospect of improving the quality of policy development for parties… and potentially improve the quality of public debate’. This question was raised in October last year when Ed Balls called for the OBR to ‘provide independent scrutiny and certification’ of Labour’s tax and spend commitments before the 2015 election. At present, the OBR is explicitly prohibited from considering ‘the effect of any alternative (non-government) policies’, and it looks unlikely that this ban will be lifted in time for the next general election. As part of our work on ‘Year Five’ of the Coalition, the Institute for Government has today published two case studies exploring pre-election policy costing mechanisms in Ireland and Australia. A third study, on the Netherlands, will follow shortly. We found that these mechanisms can improve the quality of policies and deliver more informed public debate by encouraging parties to be open about the costs of their proposals. Here we set out some of the lessons from international practice about how such systems work. One key lesson is that, while the ultimate goal of a costing mechanism should be to encourage parties to be more transparent about costs, it is often necessary to ensure confidentiality of the costing process itself. This may seem counter-intuitive, but experience in Ireland and Australia shows that parties must be confident that policies and costings will not be shared with their political opponents or the media until the party is ready to release that information. A second lesson is to minimise the scope for ministerial interference. Ministers should not be aware of officials’ involvement in the process. It is also important to limit ministerial involvement in granting access to costings. In Ireland, where all parties can submit policies to be costed by Department of Finance officials, the costing process is overseen by the Secretary General (permanent secretary equivalent) of the Department of Finance. The government of the day has no role in deciding when parties can submit their policies. This contrasts with UK convention on pre-election contact with the Opposition, which must be agreed by the Prime Minister. A third important feature, common to the Irish, Australian and Dutch costing mechanisms, is that they allow for dialogue between parties and the costing body. This allows the team carrying out costings to ask parties for additional information, and enables the parties to ask questions related to their costing. While this may increase the resource demands of the process, the benefit is that policies can be refined and improved in an iterative process before the party commits and publishes its plans. A fourth lesson is the timing of when parties can request policy costings. In Australia parties can have policies costed at any time during the electoral cycle, allowing parties to develop their policies over a longer period. Alternatively, a costing facility of this sort could function solely in the few months before the election, as is the case in Ireland. Either way, any expansion to the OBR’s remit would of course require additional resources. A separate question is whether any such facility would be open right up until polling day, or until the start of the campaign. A fifth point is the need for clear terms of access to government data. Independent costing bodies are often reliant on information and data from government departments. It is important to put in place clear and enforceable agreements around what access these bodies have to government data, and how long departments should take to respond to requests for information. A sixth lesson for political parties is the importance of co-ordination. The costing body may receive a high number of requests which require a fairly swift turnaround. It will be important for parties to co-ordinate their requests, perhaps through a single point of contact, to ensure overall coherence of policy requests and to prioritise urgent work at peak times. Having a single point of contact on the party side can also allow for a freer exchange of information: in Ireland, we found that a good relationship between the individuals involved in costing could sometimes result in a freer discussion of the pros and cons of a policy. A question to consider is whether the OBR should assess party manifestos as a whole, rather than costing individual policies. In Ireland and Australia parties do not submit their complete manifestos, but in the Netherlands the Central Planning Bureau assesses the cumulative fiscal and economic impact of all policies in a manifesto, and publishes analysis of all manifestos a few weeks prior to the election. This helps to inform public debate and provides a disciplining effect on parties who must be sure that their sums add up and that their overall policy mix is coherent and complementary. One final point to take into account is that costing processes can only take you so far. While it is of course important to ensure that party manifestos are based on fiscal realities, not even the best and most accurate policy costing model can tell you whether a policy will be implementable or whether there are better alternatives out there that the party should consider.