The recent NSW experience was, by contrast, highly positive. If either party won government, there would have been no post-election budget surprises or "black holes"; both parties had an accurate and independent estimate of the effect of their policies on the budget to present to voters. Policies were costed rapidly and reliably. The aggregate impact of the parties' policies on the forward estimates was calculated and published in budget impact statements for both parties on the Monday before the election weekend.

While published costings were accurate and reliable, another important indicator of success was the large number of costings not published. Of 476 policies submitted by the parties for costing (through their parliamentary leaders), only 179 were announced. Overwhelmingly, this was due to different ideas and options being tested in development. Both parties used the process to refine and improve policies in response to costing information. The availability of independent and reliable information on costs helped the respective NSW parliamentary leaders determine and advocate publicly the most cost-effective policies – the ones that would deliver the best outcomes for the lowest costs – from their respective political perspectives.

NSW was the first Australian jurisdiction to establish a PBO, with the Parliamentary Budget Officer Act 2010. An acting PBO (former NSW auditor-general Tony Harris) was appointed in early 2011, shortly before that year's election, and was only asked to cost government policies. The 2014-15 PBO had a full term (from September 2014 to the end of June 2015); this was the first time the NSW PBO had operated under legislation as amended in 2013.

The NSW arrangements differ in many respects from those in the Commonwealth. Importantly, the NSW PBO is not permanent, it is a pre- and briefly post-election appointment, and can only cost policies proposed by the premier and opposition leader. Neither minor parties nor individual MPs are allowed to submit policies for costing. The Commonwealth PBO has a more detailed and prescriptive remit, including post-election costings, and a much bigger budget.

The NSW Labor Party has taken the view that the PBO should be an ongoing position. There are arguments for this: it would provide better support to the Parliament on policy costings in all years, not only before the election, improving Parliament's policy capacity and enhancing transparency in government. There are, on the other hand, arguments for a shorter term: given the pre-election period is when costings are most needed and relevant, a PBO's costing work in other years may have little if any relevance to decision-making, and there are budget savings from having a term appointment rather than a permanent PBO.

Were there to be an ongoing role for a NSW PBO, it would need to be broader than simply election policy costings – otherwise, it would be hard to justify on cost-effectiveness grounds. This could include: providing advice to parliamentary committees to help them test different options being considered in inquiries; publishing materials to broaden understanding of the budget choices facing NSW; and providing advice on budget and costing matters to minor parties and individual MPs. These may be less attractive for government, given it receives such advice from its own departments; few governments go out of their way to give their opponents free, high-quality advice. What happens in the future will, however, be determined by the wishes of the NSW Parliament, and a ministerial review of the act is due by November this year.

Not every aspect of the NSW experience was ideal, and the ministerial review will address these. The PBO's post-election report outlines ways in which processes could be improved, including a firmer start point for the forward estimates, allowing parliamentary leaders to withdraw policies (in 2014-15, this was worked around by leaders not announcing policies, which had the same effect but was clumsy), shorter deadlines for turnaround of information close to the election date, and various other technical changes. Even so, the overall and unanimous conclusion from the political and public service participants in the 2014-15 process was that it was highly successful.

Important contributors to the success of the NSW PBO in 2014-15 included confidentiality, quality of analysis (as a result of having a strong mix of excellent staff with complementary skills), responsiveness, flexibility and good relationships with other NSW agencies.

Confidentiality is a fundamental prerequisite for an effective PBO. No political party will submit policies if it thinks details will be leaked. In 2014-15, there were no leaks from either the PBO or – just as importantly – the many NSW government agencies that were contacted to provide information required for costing.

Speed was also important. The average time taken to complete a costing was eight days, with a range between one and (in the worst case) 48 days. This range reflects the broad variety of costings, some of which were simple and others which were complex and required information from agencies to complete. Notably, though, delays were experienced mainly in relation to very difficult costing requests received in the year before the election, and timetables became progressively faster closer to the election.

Liaison with other agencies was key to providing useful costings. Of the 476 costing requests received, 217 (45 per cent) required information from a government agency for the PBO to complete the costing. Under section 16 of the PBO Act, agency heads must respond to information requests within 10 business days "or such other period as is agreed between the head of the agency and the parliamentary budget officer". Most agencies were committed to quick turnarounds and the average time to respond was less than 10 days; however, 51 information requests took more than 10 days to be returned, and nine more than 15 days. Although early performance was patchy, with some laggard agencies, turnaround times reduced considerably from February onwards. This likely reflects not only improved understanding of PBO processes and streamlined procedures in agencies but also the obvious difficulty any agency (in any jurisdiction) experiences in pinning down information when key staff are on holidays in December-January.

All of the costed policies that were announced were also released, together with their costing, in budget information statements for each party before the election. Not every policy could be costed – for some, there was inherent uncertainty or unavailability of information. In these cases, the NSW PBO took the innovative approach of providing what information and analysis it could in the form of briefing papers. This material is still available online at the NSW Parliament website. Importantly, though, policies that were not released, together with their associated costing material, remain confidential.

A PBO helps develop and maintain public trust in political processes. Other states are now considering establishing their own PBOs to achieve similar levels of transparency and better fiscal information as now applies in NSW. Arguably the best-performing and oldest such body worldwide is the United States' Congressional Budget Office; many other PBO equivalents have been established in recent years around the world. Not all have worked perfectly – one African PBO has been prosecuted for corruption; the first Canadian PBO had a combative relationship with government that saw it cut back his funding significantly – but all have had the potential to improve the operations of their respective governments and most have delivered.

Stephen Bartos was NSW parliamentary budget officer in 2014-15.