House of Commons - Office for Budget Responsibility

Office for Budget Responsibility - Treasury Contents

2 The tasks of the permanent OBR

35. The resources the OBR will need and how it should be structured will depend on the tasks it is to carry out. The broad parameters of the OBR's remit have already been set:

The Office for Budget Responsibility (OBR) will make independent assessments of the public finances and the economy. It will have direct control over the forecast and make all the key judgments that drive the official projections. It will have full access to the necessary data and analysis produced by the Treasury.
The OBR will also present a range of outcomes around its forecasts to demonstrate the degree of uncertainty. Based on these range of outcomes, in each Budget and Pre-Budget Report the OBR will confirm whether the Government's policy is consistent with a better than 50 per cent chance of achieving the forward looking fiscal mandate set by the Chancellor.
The OBR will also have a role in making an independent assessment of the public sector balance sheet, including analysing the costs of ageing, public service pensions and Private Finance Initiative contracts.[26]

These tasks can be distilled into forecasting, assessing the likelihood that the fiscal mandate will be achieved, and commentary about the sustainability of finances. However, that still leaves considerable detail unsettled.

Forecasting

WHAT WOULD SUCCESSFUL FORECASTING LOOK LIKE?

36. However sophisticated the modelling, and however accurate the data used, forecasting will always require a considerable amount of judgement.[27] In future those judgements will be made by the OBR rather than the Chancellor. Our witnesses warned that giving the OBR responsibility for producing forecasts will not necessarily produce more accurate forecasts. As Professor Besley said, "this organisation is bound to fail if that is the way in which we judge it, because we know that the success of short-term economic forecasting is extremely limited", and Professor Wren-Lewis agreed that "forecasts are always wrong".[28]

37. One of the reasons for establishing the OBR was to remove both the possibility that a Government might distort the forecasts (not necessarily intentionally), and the fear of such bias. Our witnesses also pointed to the fact that the existence of the OBR could improve public and political understanding of forecasting. Forecasts are inherently fallible but the OBR should aim to avoid a bias in either direction in each forecast of the economy and of public finances. If in the event there is a bias over time, the judgement of it would depend on the transparency and coherence of the underlying judgements, as well as the performance of other forecasters. Mr Chote suggested that it would be "very desirable" to set out "a reasonably user-friendly way how you approach the task of thinking about economic forecasting and economic analysis".[29]

38. One of the ways in which we will judge whether the OBR is a success is whether there is greater public understanding of the purpose and limitations of the forecasting process, and realistic expectations of what it can deliver.

39. There should, and will certainly, be analysis of the accuracy of OBR forecasts. Their quality and authority can be measured over time, relative to other forecasts. Absolute accuracy is not a useful criterion. Persistent pessimism or optimism will become apparent over time, justifiable on the basis of published explanations and methodology.

40. We can envisage a situation in which a government argues that an erroneous forecast had led it to pursue an ill-fated policy, thereby attempting to escape responsibility for its own decisions.[30] The permanent OBR might therefore be considered to have an incentive to produce cautious forecasts. By contrast there is some evidence that Treasury forecasts have tended to be optimistic. This may reflect government's incentive to "talk up" the country's economic performance. This matters because official economic forecasts can in themselves influence economic performance, affecting business confidence, currency, stock and government debt markets.

41. The Chancellor has reserved the right to use Treasury forecasts rather than OBR forecasts if he considers it necessary. Indeed, Professor Wren-Lewis considered that if "the Treasury decides that the OBR model is wrong in some sense, I think basically then it is up to the Treasury to decide whether it wants to move to an alternative model or an alternative way of doing things whereby it produces its own forecast and does not rely on the OBR."[31] In such cases, the Government would have to explain and defend its decision. The need for such an explanation suggests the Government would not take such a decision lightly, without evidence, or without considering the consequences for the public perception of its commitment to independent forecasting. The legislation establishing the OBR should not require future governments to use OBR forecasts. It is possible that the power of the Government to use its own forecasts will counterbalance any incentive the OBR might have to be overly pessimistic. The OBR's reputation would suffer if it were shown that its forecasts were so significantly biased that the Government no longer considered them a reasonable base for policy-making. However, it would be a major step for Government to do this. Once the decision had been made, the OBR's credibility would be severely, and possibly terminally, damaged.

42. Simon Hayes of Barclays Capital was concerned that the existence of both MPC and OBR forecasts as 'official' forecasts could make the communication of economic policy difficult:

If the two forecasts are materially different, it will be difficult for the authorities to argue that monetary and fiscal policy are appropriately co-ordinated. If they are engineered so as not to be materially different, it raises the question of why effort is being duplicated across two public institutions.[32]

We do not consider this a problem. The MPC is charged with setting interest rates on the basis of its own economic judgement; it is appropriate for it to retain its own forecasts. Part of the rationale for the OBR is to raise public understanding of forecasting; the uncertainties and judgements in the process are demonstrated by two forecasts.

CONTROL OF MODELS AND DATA

43. The forecasts produced by the interim OBR used the Treasury's models and sub models. Simon Hayes considered that the OBR should "own" macro-economic forecasts:

There is a set-up cost associated with this but it need not be an onerous one. Diminishing returns quickly set in with macroeconomic forecasting—the accuracy gains from increased sophistication are questionable beyond a basic level—and the Treasury model should provide a suitable baseline.[33]

Sir Alan Budd agreed, but considered that "ownership" was not necessarily straight forward:

The OBR must be able to use the models that are required for forecasting processes. Ownership is a slightly difficult issue, I am not even quite sure how it is defined, but certainly if a model is being used to produce OBR forecasts then in that sense it must be owned by the OBR. There is not just one model, there is a suite of models, but any model which the OBR is using it must own in the sense that it knows what it contains and is responsible for its outputs.[34]

Professor Wren-Lewis agreed that "the OBR has to be in charge of the model it uses."[35]

44. Several witnesses drew attention to the way in which models could be tested and reviewed by outside experts, and this is something that the permanent OBR may wish to consider.[36] There was general agreement that the OBR should be open about the forecasting models it used. The OBR should have discretion in the models it uses in drawing up its forecasts. It is a matter for the organisation itself as to whether it is content to use the Treasury models, or wishes to make changes. Whatever course the OBR takes, there would be benefits in it being as transparent as possible about the models it uses. The OBR should also be cautious about attempting to increase the sophistication of the model in search of dubious increases in accuracy. As many witnesses pointed out, a sophisticated model cannot remove the need for forecasters to exercise their judgement and incremental benefits to an already highly complex model may be nugatory or perverse.

45. The OBR will also need to have confidence in the data underpinning the forecasts. Many witnesses considered the OBR should be given "statutory access" to data, other than the personal data used by HMRC. As well as using existing data, the OBR may need to suggest changes to the data collected. As Mr Chote said:

I think it should be deciding what data it needs to do the job it is tasked with doing. There is an iterative process there about looking at what you have through the normal channels and then saying, "Well, actually, this could be done better. We need more, different information," or "Is this really being approached in the right way?" That sort of relationship you have seen in the past with the Bank of England interacting with the Office for National Statistics, raising any concerns they may have there, and I think the OBR would presumably raise similar sorts of concerns if it found, for example, that it was not getting the information it thought it needed to be able to validate the social security spending forecast, to take one example. The OBR should not be at all afraid of saying, "We want more or different information," if it is available.[37]

In written evidence, Professor Besley noted:

Some kind of link to ONS would be useful—the Bank as you know is asked every year to comment on ONS's performance and OBR could be given the same status both in relation to ONS statistics relevant to the conduct of fiscal policy.[38]

The OBR should have the power to check the quality of fiscal data itself, and to request that the ONS does so. It should also be free to use any existing data it thinks fit in constructing the forecast, and to recommend changes in data collection, if it considers that this would improve the forecast or its ability to assess the likelihood of achieving the fiscal mandate. Where a recommendation may increase the overall cost of the OBR's work, it should be required to seek an external view on the benefits and costs of the change, and report this to the Treasury and to this Committee.

FREQUENCY OF FORECASTS

46. There was some disagreement in the evidence we received about how often forecasts should be produced. The Industry Act 1975 requires the Government to produce a minimum of two forecasts a year. Forecasts will have to be produced with each fiscal event. In general opinion favoured two forecasts,[39] but it has been suggested to us that four forecasts a year would be worthwhile.[40] It has also been suggested that the OBR should monitor the fiscal environment between forecasts to establish how the reality is deviating from the forecast.[41] In normal circumstance, we see no case for more than two forecasts a year: even if short-term forecasting were more precise than it actually is, changing tax and spending plans within year is intrinsically undesirable. While we agree with the NIESR that "expert commentary on state of the public finances should not be left to twice yearly forecasting rounds";[42] we do not consider that this is a task for the OBR; such commentary is already widely available. While there may be cases when sudden shocks to the economy mean it is desirable to have a further forecast, we expect that to be rare. While forecasting will be a key task of the OBR, there is no point in devoting resources, either to increase detail or frequency, to forecasts which will have no practical application or benefits. In normal circumstances, the OBR should produce two sets of forecasts a year. An extra forecast would be desirable if there were significant monetary or fiscal policy changes or significant external shocks.

Sustainability analysis

47. The interim OBR was given two tasks in relation to fiscal sustainability: beginning work on an independent assessment of the public sector balance sheet and fiscal sustainability and providing an initial discussion of public sector liabilities and their implications for the public finances.[43]

48. In its advice to the Chancellor, the interim OBR proposed an annual report on the longer-term sustainability of the public finances. This would include long-term fiscal projections.[44] It would be particularly important that the organisation examine factors such as the Private Finance Initiative and pension liabilities, which some have argued may currently not be properly presented. It should consult with other interested parties before deciding how these should be represented.

49. The OBR's engagement in this will not be without controversy. It is notable that although each year the Congressional Budget Office prepares a baseline budget, showing projections for the next 10 years, its independence is safeguarded by legal rules about how that budget should be constructed:

The baseline is constructed according to rules set forth in law, which generally instruct CBO to assume that current spending and revenue laws continue without change. Thus, the baseline is not a prediction of future budget outcomes. Rather, it reflects CBO's best judgment about how the economy and other factors will affect federal revenues and spending under existing laws. Each summer, CBO updates its baseline projections, incorporating a new economic forecast and the effects of laws that have been enacted to date in that session of Congress.[45]

We also note that "in accordance with the CBO's mandate to provide objective and impartial analysis, CBO's reports contain no policy recommendations."[46]

50. We support the interim OBR's recommendation that the permanent OBR should produce an annual report on the long-term sustainability of the public finances. This report should contain no policy recommendations.

Commentary

51. The terms of reference for the interim OBR required it to make a judgement on whether the Government's policy was consistent with a better than fifty per cent chance of achieving its fiscal mandate.[47]

52. Several of our witnesses argued that the new OBR should have a wider commentary and research function on fiscal sustainability, arising out of the functions outlined above. For example, Professor Besley argued for an OBR with a broad remit along the line of "safeguarding fiscal sustainability in the UK" and be able to comment on issues falling within that remit:

Taking an example from the past, had the OBR existed since, say, 2000, I would hope it would have produced commentary on some of the risks around fiscal policy with respect to the taxation that was being raised from the financial sector and that it would have felt perfectly justified to raise such an issue independently because it felt it was an important part of the outlook. To constrain artificially in any way the ability of this body to comment on what is germane to that broad objective I think would compromise its independence and compromise its institutional integrity.[48]

53. The staff of the International Monetary Fund (IMF) went further than Professor Besley. They suggest that the OBR could have a role in "scoring" individual government policy and reform suggestions from non-governmental sources, including the Opposition.[49] Professor Wren-Lewis also advocated the costing of Opposition policies by the OBR; although Professor Besley and Robert Chote both expressed concern about the resource implications of such a move, which would be considerable.[50]

54. There are other risks associated with an OBR with a commentary function that is too broad. The Institute for Fiscal Studies stated in its February 2010 'Green Budget', and Robert Chote reiterated in his evidence to us, that the OBR should comment on whether changes to the fiscal rules were consistent with long-term sustainability. However, the Green Budget continued:

An OBR should not be tempted to offer advice on broader economic policy issues, as some of its overseas counterparts do (for example, in Sweden). The cost of commenting on broader economic policy could be getting drawn unnecessarily into conflict with other parts of the policymaking process and thereby reducing the credibility of the OBR on its key function—the production of high-quality independent fiscal forecasts.[51]

55. Dr George Kopits, the Chairman of the Hungarian Fiscal Council, emphasised the importance of the OBR avoiding "normative judgements or formulating policy recommendations, as this may be seen as a departure from nonpartisanship."[52]

56. Simon Hayes, of Barclays Capital, drew a distinction between the OBR's becoming involved in the setting of the mandate—which he considered could impinge on its independence—and commenting on the likelihood of the Government achieving its mandate—which he considered to be desirable.[53] The interim OBR's advice to the Chancellor examined this issue, but did not make a recommendation.[54] He further suggested that the OBR could promote best practice in fiscal policy.[55]

57. Dr Frank Eich, of The Pensions Corporation, said that he saw an important role for the OBR in raising the quality of public debate on longer-term economic and public finance issues. Dr Eich thought that the OBR should produce work that is accessible to non-specialists and actively engage with Parliament, media and the public to improve understanding.[56]

58. The OBR's contribution to public understanding should not be confused with self promotion. This commentary function should be one of informing public debate through disseminating better understanding of fiscal policy and long-term economic trends, identifying possible risks in the structure of the economy and provision of data. Beyond any duties set out on the face of the legislation, the statute should give the OBR absolute discretion over the work it undertakes. The legislation should leave the OBR able to conduct work on the fiscal policies of political parties along the lines proposed by Robert Chote in evidence on 16 September 2010. There may be also particular subjects which the Treasury, or this Committee, consider should be examined by the OBR. We would expect the OBR to consider such proposals carefully and, where appropriate, to explain its decisions.

59. As an additional defence of OBR impartiality, we recommend that the OBR's core tasks should be set out on the face of the legislation.

60. A particular source of controversy is whether the OBR should comment on the Government's fiscal mandate. There were a variety of views among our witnesses and the interim OBR did not reach a conclusion.

61. Some argued that the OBR should restrict its commentary to whether the mandate was achievable and government policy consistent with it. They suggested that the Monetary Policy Committee was insulated to some extent from political controversy by the fact it was not involved in setting the inflation target and argued that the OBR could benefit in a similar way.[57] Simon Hayes was concerned that the OBR could become too closely associated with a mandate that had been adopted on its advice. This could undermine its perceived objectivity when assessing the likelihood of the mandate being met.[58]

62. At the other extreme, NIESR argued that sustainability was integral to the mandate and the OBR would be best placed to assess the current fiscal mandate and manage change to the fiscal mandate itself. However they accepted this would probably involve separating the OBR from the official forecasts.[59]

63. Robert Chote argued that the OBR should be able to comment on whether the mandate was consistent with short-term sustainability. However, it should stay clear of issues such as the pace of consolidation.[60]

64. We agree with Mr Chote that it would be inappropriate for the OBR to have a role in setting the fiscal mandate. This should be a political decision. Once the mandate is set, any OBR commentary should be based on aggregate fiscal figures, not on individual measures.

26 http://www.hm-treasury.gov.uk/data_obr_work.htm Back

27 Qq131-132 Back

28 Q 82 Back

29 Q 133 Back

30 Q 111 [Professor Wren-Lewis] Back

31 Q 129 Back

32 Ev w25 Back

33 Ev w61 Back

34 Q 28 Back

35 Q 128 Back

36 Q 123 Back

37 Q 124 Back