The words 'public procurement' do not raise strong feelings in the minds of most readers. But they should.

The way government spends taxpayers’ money on everything from pens to highways has great bearing on how efficiently it functions. Wasteful procurement means that the government’s budget is able to provide less of the services that really matter to taxpayers: education, health, public transport and so on.

On the other hand, a lean government that keeps unnecessary expenses at a minimum can serve its citizens better and maintain their trust.

Public procurement is a substantial part of the national income and government budget; estimates suggest that it is about 19.8 per cent of the gross domestic product in Pakistan.

Our research has revealed a simple, low-cost policy that can make government leaner.

Many policies have been tried out around the world, but they often contradict each other.

For example, international organisations (the Organisation for Economic Co-operation and Development and World Bank have often espoused this approach, for example) tend to advocate a policy of writing strict rules to govern procurement procedures. Such policies entail strong monitoring and oversight institutions to ensure compliance with the rules. While this may seem safer, it has a downside: procurement officers have to deal with red tape, filling out extensive forms and providing complete documentation.

Other governments and businesses have pushed in the other direction, to give greater power to procurement officers, trusting them to use their discretion to seek out value for money. This approach calls for lighter monitoring during procurement, and strict enforcement of rules only in cases of wrongdoing. This philosophy underlay the reforms to the public procurement system in the United States under Bill Clinton in 1993 that made an appeal for government to make an “investment in smart buying”. It also underlies Netflix’s five-word policy for expensing: “Act in Netflix’s best interest”.

This second type of policy is often accompanied by pay-for-performance schemes that reward procurement officers when they perform well: spend more efficiently and receive a bonus.

We had the chance to test these two policies against each other, and our results can guide policy.

Over the past several years, in a research project with Oriana Bandiera (London School of Economics) and Andrea Prat (Columbia University), we have engaged with the Government of Punjab to deploy a methodology for measuring value for money of generic goods and developed a web portal named Punjab Online Procurement System (POPS). The project was conducted jointly with the Public Procurement Regulatory Authority and Punjab Information Technology Board.

First of all, there was a big difference in prices paid by different public bodies for exactly the same good – as shown in the figures. Each point in the figures is a purchase. The horizontal position of the point represents the price the public body actually paid for the item. We can see that the prices paid for the same item vary dramatically. Using the detailed data in POPS, the vertical axis represents what the item is actually worth (according to a methodology we developed, adjusting for differences in the precise attributes of the item and the location in which it is purchased).

We highlight in orange items that are worth roughly the same amount. We find that high performers pay as little as Rs3.50 for pens worth Rs25, while poor performers pay up to Rs115. Prices paid for toner worth Rs3,500 ranges from Rs1,550 to Rs11,000. High performers pay Rs60 for registers worth Rs150, while poor performers pay Rs700: public bodies are paying as little as Rs0.70 and as much as Rs1.80 for paper worth Rs1.25 per sheet.

These numbers raise the question of why offices are paying such different amounts for such simple, off-the-shelf items. These items may seem trivial as each purchase is relatively small. However, keep in mind that they are bought across the entire government of Pakistan. In the offices we worked with in the health, higher education, agriculture and communication and works departments, such costs accounted for over half of non-salary expenditures.

They also pose a challenge to the government: Is it possible to find ways to get poor performers to stop overpaying and behave more like the high performers?

We worked with the government to design and experimentally test two policy initiatives to improve performance of frontline procurement agents.

Our first initiative reduced the red tape faced by procurement officers when making purchases, simplifying or bypassing much of the pre-audit monitoring of their procurement of generic goods. We wondered if this increased autonomy would allow frontline procurement agents to improve value for money by striving to lower prices, or if it would make them more wasteful since they knew they wouldn’t get caught.

Our second initiative introduced performance pay for procurement officers, rewarding them financially for achieving greater value for money.

After running the experiment for two years, we learned that both giving procurement officers autonomy and performance pay can improve the value for money achieved in public procurement in Pakistan.

The figure below shows how much these initiatives were able to save in the around 600 small public offices we worked with, and how they compare to the cost of 150 additional hospital beds or operating 10 schools. The autonomy initiative saved the government Rs13 million, or around 8pc of the spending by the offices we worked with, and was virtually costless to the government. The pay for performance initiative saved Rs4 million, and cost Rs2.65 million, and so provided a solid 45pc return on investment for the government. Offices where both initiatives were implemented together displayed a 26pc rate of return (RoR).

The results of these initiatives are encouraging, suggesting that real savings are possible without cutting back on the items being purchased to provide services with, and guiding us to how to achieve them. This was a small sample, the potential for savings across government is substantial.

Perhaps more importantly, we learned more about when each type of policy is likely to be successful. Monitoring compliance with rules is costly. It takes time for civil servants to document that they are following the rules, it takes time and effort for monitors to verify these documents.

The costs of this red tape have to be balanced against the benefits of using rules to improve value for money and the quality of services delivered to citizens. When monitoring is very costly, either because monitors are themselves corrupt/inefficient, or because the time cost and inefficiency associated with red tape are too onerous, monitoring is ineffective. In this situation, performance incentives for procurement officers don’t work, they are too tangled in red tape to do anything to improve performance. By contrast, giving more power and discretion to bureaucrats improves performance. In the longer term, jobs that offer the autonomy to effect positive change may attract individuals with stronger motivation to work in the public interest into the civil service, amplifying the benefits.

The bottom line is that we need to check our instinct to react to the inefficiency and corruption we often see in government procurement by imposing additional monitoring on the bureaucrats. In many places, the elaborate system of pre-audit monitoring they face is probably doing more harm than good, and should be replaced with greater autonomy and rigorous post-audit of purchases.

More broadly, creating a civil service that empowers bureaucrats who want to make a difference and serve their country, while reigning in the excesses of bad apples who exploit the system will serve Pakistan better than a rule-bound behemoth. For Pakistan to develop as a nation, it is imperative for the design of that civil service and all public policies to be based on evidence of what works, what doesn’t, when and why. This is as true of procuring pens and paper as for any other policy.

The Analytical Angle is a monthly column where top researchers bring rigorous evidence to policy debates in Pakistan. The series is a collaboration between the Centre for Economic Research in Pakistan, Evidence for Policy Design at Harvard Kennedy School, and Dawn.com. The views expressed are the authors’ alone.