- Public financial management is a relatively placid field of international development. Other development specialists have fierce arguments over the faults and merits of, say, Millennium Villages, or randomized controlled trials. For the most part, PFM specialists acknowledge that while there is a strong technical core to public finances and budgeting, there is a lot we don’t yet know about the political and institutional side of establishing and reforming systems in developing countries. Arguments are rarely had in the open.One of the more profound disagreements in the field is over the proper sequencing of PFM reforms.Given that many governments face challenges with budget formulation and budget execution; as well as external oversight and internal control, where should they start? Is it necessary for bureaucracies to first establish control over inputs before they can link allocations to results? Must we establish credible annual budgets before we worry about medium-term budgeting? Are there budgetary basics as well as advanced budgetary systems? If so, what are the prerequisites for advanced systems to work?This post could have been titled “Why Allen Schick is always right”, because the major exponents of the sequencing debate implicitly base their claims on Schick’s work. They are also mostly wrong. Schick’s seminal 1998 article explains “Why most developing countries should not try New Zealand’s reforms”. In a presentation given the same year that was unfortunately never published, he makes the more general argument and warns aspiring reformers to “look before you leapfrog”.
Schick established the notion that there is a sequence to budget development, and that budgetary basics must come first. This means that control over inputs precedes control over outputs; cash accounting before cost accounting; rules before flexibility; as well as integrated and centralized departments before autonomous agencies.
The PFM field has since taken sequencing in several directions. First of all, concerns over sequencing are still often ignored in practice. In a survey of PFM reforms in 31 African countries, Matt Andrewsfinds that the overwhelming majority of governments have tried to introduce “advanced” reforms. This includes medium-term expenditure frameworks, accrual accounts and performance budgets. Many of them failed. The most obvious conclusion from Schick’s argument is that we should not try advanced budgetary reforms in developing countries. Perhaps unsurprisingly, this argument has not caught on.
Recently, the most influential take on sequencing has been to advocate rapid progress through “platforms”. Instead of shying away from reforms that go beyond basic budget systems altogether, or trying to immediately implement more advanced budgetary systems, a quick step-by-step approach is advocated.
The most commonly cited formulation of the platform approach suggests countries should first establish credible budgets; then move to establish internal accountability; then link policy and planning; and finally establish performance management. This is all meant to take place over an eight year period(!).
By the letter, the platform approach is very faithful to Schick’s “Basics First” message. The wild optimism of this approach, however, has since come in for an entirely uncharacteristic (for the PFM community) amount of criticism and ridicule. For instance, Richard Allen has rightly argued that today’s rich countries took centuries to move through these stages.
But this debate misses the most important point about Schick’s original paper. Yes, the elements of budget reform do somehow build upon one another. But, more importantly, a country’s PFM system – and its bureaucracy more broadly – is inexorably linked to wider societal norms and private sector practices.
The establishment of Weberian bureaucracies requires governments to develop administrative formality and controls over inputs. This process happens either in parallel with similar advances in the private sector or because a more advanced private sector demands that the lagging public sector cleans up its act. Further reforms towards a New Zealand style contractualism – where managers are given discretion over the use of funds and held accountable for outputs or outcomes – rely on societal norms of formality and transparency to be already firmly in place. Schick is quite explicit that public sector norms may catch up with the private sector, or both may advance in lockstep, but only very rarely does the public sector move ahead in isolation of the private sector.
Budget systems are a product of the broader political, economic and social context in which they came about. They are not very portable. The broad societal changes required to get from pervasive informality to acceptance of formal contractual arrangements may not take centuries. But for those sitting in a Ministry of Finance, they probably take long enough to be well beyond their planning horizons.
Virtually all comprehensive budget reform strategies implicitly assume that reforms within central finance agencies can spearhead similar changes in the wider public sector. They also assume that public sector reforms will somehow take society along for the ride. To the extent that we ponder sequencing issues, we focus on the technical requirements of PFM systems (i.e. we need the budget to be predictable over 12 months before we can enforce ceilings for 3 years). We hardly ever focus on the external preconditions for reforms to make sense in the first place. Matt Andrews and others have long called for the “other stuff” of PFM reform to be given primary importance. To date, the internal technical logic of a PEFA assessment will still win out over any other consideration. But the unfortunate fact is that a discussion about sequencing makes little sense if it does not include non-technical sequencing as well. Unless we prioritise getting a better understanding of how PFM interacts with the wider political environment, we’re unlikely to move beyond what we knew 15 years ago.
Why most publications about pfm sequencing are missing the point.