Attachment F: A Challenge to the Purchaser / Provider Model?

Based on Contract Budgeting, (by Marc Robinson)
  - QUT School of Economics and Finance -

Key issues are:

• contract budgeting is traditional program budgeting, with a purchaser / provider split
• program budgeting experienced intrinsic problems over many years in measuring / defining outputs due to the too numerous / too complex functions of government.
• 'efficient' prices (as compared with costs) can only be known if competition exists
• where outputs can not be measured / defined, long term relationships with a supplier (rather than cost competition) is essential - and 'efficient' prices can thus not be known.
• without an equity owner, there is little real pressure to improve performance
• cost pressures (rather than purchaser pressure) seem to have more influence on agencies, who respond in ways with short term gains and poor long term outcomes
• thus contract budgeting is not implementable

Contract budgeting involves transforming relationship between government and its line agencies from one of funding to a purchase relationship. Central decision makers determine outputs and then contract with agencies to purchase them. Contract budgeting was initiated in New Zealand in 1989, and has spread internationally. Contract budgeting combines reconstruction of government on 'market' lines with program budgeting. Program budgeting has been practiced since the 1960s, based on monitoring of outputs. It was a reaction against input budgeting. Quasi-market models have been applied where no consumer purchase is involved (eg by Thatcher Government for health etc which were influenced by US models) by a formal purchaser / provider split; which allows competitive tendering. Contract budgeting attempts to extend this to all central government function. This is far more than has been done elsewhere.

Output budgeting: Without contracting, contract budgeting is like program budgeting, though the goals are productive efficiency rather than allocative efficiency (because of increased financial constraints since the 1970s). However there is no recognition of how similar contract budgeting is to program budgeting, and thus little learning from history. Program budgeting always fell short of expectations due to: highly aggregated programs which could not fully specify output quantities; difficulties in estimating inputs; and arbitrary allocation of cost overheads. These are fundamental problems in concept due to:

• the number of government outputs being too large to manage. In NZ number of categories initially rose by an order of magnitude, and then was rationalised.
• Inability to measure some highly variable outputs (a frequently occurring problem)

The latter can only be budgeted on an input basis. Problems in allocating some costs to outputs are serious, a difficulty which is recognised also in the private sector. Further problem is that, for many government functions, decision is not what level of output to provide, but what level of capacity to maintain (eg for emergency services).

Concept of contract: Notion of contract involves arms length relationship where risk is taken by supplier. This can not be done where purchaser is also the owner. Program budgets were just plans which agencies did best to meet. With contract budgeting there are meant to be stronger sanctions. Where did this model come from? Organisational economics favours contracting, but envisages a far more complex set of possible relationships, than classical contracting. The contract model comes from economics (with its focus on competition), but this is influenced by the need for cost control and so focused only on price competition. Second influence is managerialist philosophy which favours clear objectives; freedom to manage. Assigning a task to someone and letting them do it, is similar to role of prices in allowing others to ignore input costs in goods and services they buy. Governments are seen to need to be demanding buyers of services.

Contractual budgeting: Inability to measure or specify outputs makes it impossible to assess whether agencies met goals. However ignoring this, the problem of what happens if agency fails to perform can be considered. If fail to meet cost targets, there is no equity owner to motivate improved performance. Even if agencies are given capital to manage, this only delays the problem. Ultimately government (as owner) has to cover any losses.

Pricing problem: Without competition there is a fundamental problem in knowing the efficient cost of production as compared with the actual cost of production. However there are many areas of government which are intrinsically unsuited to competition (see below).

Agency strategies to avoid contract failure: Failure can be avoided (despite accrual accounting) by: diverting capital funds to current costs (eg in areas like human capital development, R&D capital costs are treated for accounting purposes as current costs, and can be avoided in short term, despite adverse long term effect); or by reducing unit costs through more intensive / less well remunerated labour intensive work. The latter may be reasonable if excessively generous arrangements existed, but this may not be the case.

Contract budgeting with output measurement difficulty: Where such difficulties exist agencies can respond to cost pressures with reduced output quantity / quality. Where cost cuts are imposed, agencies respond this way, and call it 'productivity'. In NZ, a study found little evidence of purchaser pressure on agencies, and mainly budget pressure - with loss of emphasis on cost effective production. Others found evidence of declining strategic expenditure as above. Financial stress yields rapid short term productivity gains, but slows long term gains. This arises whether or not contracts are used.

Competition and owner / purchaser distinction: reformers have tried to make agencies, not government, bear the cost of contract failure. This arises from notion that competition forces 'efficient' prices. However there are flaws in hard contracting model in government (through attempts to separate role as purchaser and provider). Imposing capital charge on agencies in each time period is quite different (and less flexible) than requirement to produce average rate of return over time. Government can not know efficient costs without competition, yet evidence suggests that many functions are not amenable to competition. Where output is hard to define / measure, purchaser will strive to keep provider who is believed to be doing a good job (thus competitive price tendering is risky). Long term relationships are vital to contracting for complex services - making competition difficult. Localized natural monopoly can also be a factor in this. the competition in quasi markets has thus often been less than envisaged when established. A bottom line financial measure like that in private sector is not generally possible in government.

Conclusion: Contract budgeting tries to remodel the public sector by analogy with a stylized view of the private sector. But the model does not fit, and it is not implementable.