Terminology

Executive summary

Government procurement is the process by which government agencies buy goods and services from the commercial sector. The procurement activities are often governed by regulations and laws that aim to increase transparency, fairness, and the efficient use of taxpayer dollars (Wikipedia, 2023). However, inefficiencies still exist in most countries, either because these regulations are not effective, or because they are not applied to all procurement processes (more here).

Inefficiency of government procurement is important. Approximately $13 trillion is spent per year on government procurement globally (Open Contracting Partnership, 2020, p. 6). We estimate that 15% of government procurement may be inefficient^[1] with considerable uncertainty (75% confidence interval: 5% – 30%, excluding corruption). We therefore estimate that globally approximately $0.65 trillion to $3.9 trillion dollars is lost due to inefficiencies in public procurement. Roughly $740 billion of this loss is in middle-income countries (more here). The experts we spoke to did not believe that any one source of inefficiency in procurement is most important because different inefficiencies are closely related. Experts did agree that administrative costs of the procurement process are unlikely to be a significant source of inefficiency (more here).

More efficient government procurement can improve health and income in several ways: by enhancing service quality, freeing up funds for better public services, boosting economic growth through increased investor confidence, and reducing citizen costs for public services (more here).

In this project, we attempt to identify promising interventions to tackle government procurement inefficiencies. We focus on passive waste (unintentional inefficiencies), rather than active waste (inefficiencies from fraud or corruption) partly because we think we are less likely to find tractable interventions to reduce active waste than passive waste.

We have done a broad search for interventions that would help reduce passive waste in procurement and found that supporting Lift (an impact accelerator program related to procurement) looks promising. Two other interventions that we found somewhat promising are pooled procurement (a collaboration between different buyers, where they jointly procure for the same product), and supporting civil society actors to monitor project execution. We started by compiling a longlist of potential interventions with examples and potential implementers, and selected two to look into in more detail. We based our selection on whether we could find implementing organizations (as a proxy for tractability), as well as any evidence that was readily available for the interventions (more here). We list the interventions and our main findings below.

• Supporting an initiative similar to the impact accelerator Lift could be worth considering for donors interested in improving public procurement. Lift is an impact accelerator program by the Open Contracting Partnership (OCP) that offers technical and financial support to teams aiming to improve public procurement. We think the application process could select motivated local teams that may be able to influence government procurement reforms with a relatively small amount of support. Our main uncertainty is that it is likely hard to predict which initiatives will succeed. (more here).
• Supporting pooled procurement of health care products seems promising to us. Pooled procurement is a fairly well-established and tested tool that has historically resulted in substantial price reductions of health products and various other benefits. Our impression is that pooled procurement is currently still underused and will likely grow in importance for governments in the future (more here).
• Supporting civil society actors to increase monitoring of project execution has proven to be effective in the past, but we think most organizations we identified lack enough government endorsement and support to be highly effective (more here).
• Encouraging the use of e-procurement (meaning that the government digitizes a specific part of the procurement process, for example by advertising tenders online and/or bid submission online) does not seem promising, but we have low confidence in this assessment. We spent limited time on this since from a brief review it did not seem neglected. Our impression is that the World Bank is spending significant amounts of money on installing e-procurement systems in low- and middle-income countries. There might be interventions that aim to complement, such as change-management programs and training to sustainably implement e-procurement software purchased by the World Bank. We did not research these interventions (more here).

Increasing government procurement efficiency can lead to improved health and income by improving government services, reducing government spending, and increasing economic growth

Government procurement is the process by which government agencies buy goods and services from the commercial sector. The procurement activities are often governed by regulations and laws that aim to increase transparency, fairness, and the efficient use of taxpayer dollars (Wikipedia, 2023). However, inefficiencies still exist in most countries, either because these regulations are not effective, or because they are not applied to all procurement processes.

Government procurement regulations typically apply to all contracts for public works, services, and supplies made by a public authority. Nevertheless, there are some exceptions for strategic or politically sensitive sectors. For example military acquisitions, public health, energy supply, or public transport may be subject to different procurement rules (Wikipedia, 2023).

We can think of four main ways through which more efficient procurement can lead to improved health and income.

1. Efficient procurement can contribute directly to better quality government services such as healthcare, education, and infrastructure. These lead to better health outcomes and higher incomes, among other positive outcomes.
2. By reducing government costs through efficient procurement processes, more funds become available for government services, which can also lead to more provision of or better quality of services and the same effects as above.

3. There are several indirect ways in which better procurement can lead to economic growth and hence increased income. A transparent and effective procurement system could boost investor confidence, which may attract foreign investments. Transparent procurement practices and reduced waste could also increase political trust and political stability, which may lead to economic growth.^[2] More efficient procurement could also reduce administrative costs for organizations, which can contribute to economic growth. Finally, effective procurement is likely to favor relatively productive firms, which could then expand and increase the average productivity of domestic firms.

4. In some cases, citizens directly pay part of the costs of products or services that are procured by the government, for example with public transport or electricity. More effective procurement could reduce prices and hence lower out-of-pocket spending of the population, increasing disposable income. Disposable income could also increase if the government reduces taxes because of lower costs as described in point 2.

We could think of two more potential routes to improved health or higher incomes, although we are uncertain about whether they would occur and believe these effects are likely too small to take into account. We will include them for the sake of completeness. The first is that improved procurement processes can lead to increased taxpayers’ trust and hence lower tax evasion leading to a higher tax revenue which could be invested towards better health care services or economic growth. Secondly, effective procurement methods could lead to more equal opportunities for local businesses, which could mean that government investments are distributed amongst more organizations. This could mean government spending goes to fewer wealthy people on average, since we expect the organizations that would counterfactually get to deliver the most goods or services are on average owned by richer people (either because these are the people with more government connections, or because they have become richer from delivering large amounts of goods and services to the government).

We estimate that roughly 5%-30% of government procurement can be classified as passive waste

In this project, we focus on passive waste (unintentional inefficiencies), rather than active waste (inefficiencies from fraud or corruption), because our client asked us to and because we think we are less likely to find tractable interventions to reduce active waste than passive waste.

Quantifying passive waste in government procurement is challenging, as there is no standard measure for (unintentional) inefficiency, and estimates inevitably depend on the benchmarks used to define efficiency. We approached this question primarily by asking experts, but we refrained from delving into a precise definition of efficiency to make the most of our interview and research time. Consequently, our use of the term ‘inefficiency’ is somewhat broad and lacks precision. Here, we use it to refer to passive waste rather than corruption.

Michael Jarvis, Executive Director of the Trust, Accountability, and Inclusion (TAI) Collaborative, described a scenario in which a lot of procurement is suboptimal, but 15% is egregious. Mihaly Fazekas, Scientific Director of the Government Transparency Institute, and assistant professor at the Central European University, suggested three academic papers that might help to determine the proportion of passive waste:

• Bandiera et al. (2009) measure active and passive waste separately in the procurement of 21 generic goods by 208 Italian public bodies between 2000 – 2005.^[3] Over this time, these goods are occasionally available to purchase from the central procurement agency, Consip, and so the authors “observe periods when the public body must purchase the goods on its own and periods when it can also buy it from Consip” (p. 1279). They hypothesize that if a public body is corrupt, they will be less likely to purchase from Consip, because centralized procurement limits opportunities for corruption. In contrast, an unintentionally inefficient public body is more likely to take advantage of a clear opportunity to purchase goods more cheaply through Consip.^[4] As a result, by observing how public bodies respond to specific goods becoming available through Consip, the authors estimate the extent to which procurement behavior is driven by active or passive waste. The findings suggest that on average passive waste accounts for at most a 10% increase in prices (p. 1280). However, the paper suggests that there are outliers paying 55% more than others, and 80% of waste is passive—which would suggest more than 40% price increases due to inefficiency.

• Borges de Oliveira et al. (2019) models unit prices achieved through public procurement in Brazil between 2014 – 2016, using independent variables that capture characteristics of the procurement process. The authors then use this model to estimate how improvements to the process could reduce prices, and suggest a 16% decrease is achievable—primarily through a combination of increasing quantities for tenders (7.2% decrease), publishing all tenders (3.3% decrease), and retiming purchases away from the end of the budgetary year (1.8% decrease).^[5]

  • Fazekas et al. (2021) appears to be an extension of this methodology to healthcare procurement in seven countries and two states in Brazil for the period 2012 – 2018 (36,600 tenders in total). The authors find substantial differences within and across countries, and suggest that there is scope for improvement of ~14%. The drivers of this are more varied than in the analysis for Brazil: 2.1% improvement is attributed to product bundling, while 1-1.5% improvement comes from a number of factors such as quantity of goods tendered for, and advertisement period length.

  • These articles do not fully exclude corruption: for example, it’s plausible that tenders were being actively not advertised or shortened.^[6] As such, they may overestimate passive waste. However, in these articles the authors estimate savings from specific process changes—not maximum savings possible from addressing all inefficiencies—such that they could underestimate both active and passive waste.^[7]

Based on the above, we would suggest 15% of government procurement may be inefficient due to passive waste with considerable uncertainty (75% confidence interval: 5%-30%). To put this into context, approximately $13 trillion is spent per year on government procurement globally (Open Contracting Partnership, 2020, p. 6), so we estimate that globally approximately $0.65 trillion-$3.9 trillion dollars is lost due to inefficiencies in public procurement. We estimate this for specific income groups in Table 1 below.

Table 1: Estimated annual passive waste, by World Bank income group

In expert interviews and in academic literature the focus appears to be on quantifying price inefficiency. We did not find any measure of inefficient public procurement due to reduced quality of goods and services purchased.

There are many closely related sources of inefficiencies in government procurement, and experts indicated that we could not point to one most important one

Figure 1 below captures our first principles thinking on the possible sources of inefficiency in public procurement.

Figure 1: First principles breakdown of procurement process and potential inefficiencies. The gray arrows at the bottom indicate in which stages the interventions we focused on (described in the next section) have the most impact.

Note. The four stages in the process are partially based on Fazekas and Blum (2021, p. 8). Inefficiencies have been added based on first principles thinking, and expert interviews.

The experts we spoke to did not believe that any one source of inefficiency was most important. Mihaly Fazekas stressed that procurement is an ecosystem, such that it’s not possible to tackle only one inefficiency at a time.

However, both Gavin Hayman, Executive Director of the Open Contracting Partnership, and Mihaly Fazekas confirmed that administrative costs are unlikely to be a significant source of inefficiency, given that the amount spent on the procurement process itself (in terms of staff or equipment) is so much less than the overall spending.

We have done a broad search for interventions that would help reduce inefficiency in government procurement and found that only pooled procurement looked promising

Approach to identifying promising interventions

Our approach to this report was informed by the client’s brief, which emphasized 1) tractability over importance and neglectedness, and 2) an initial focus on breadth rather than depth. As a result, we started by compiling a longlist of potential interventions based on academic literature, gray literature, and case studies. We categorized each intervention based on our first principles breakdown of the procurement process.^[8] We also added estimates of effectiveness from Fazekas and Blum (2021) where available, and spent ~10-20 minutes looking for implementing organizations for each case.

We then ordered the potential interventions based on our assessment of how promising they were. Following discussions with the client, we placed heavy emphasis on having identified potential implementers.

Table 2 below shows the five interventions on the longlist that we initially considered most promising, plus a sixth which we considered separately, and a very short summary of our conclusions. We share more information in the longlist, and in the sections below.

Table 2: Description of longlisted interventions from our research

Note. Our assessment of how promising interventions are remains low confidence, and is based on limited review.

In the next two sections, we share our findings for the two interventions that we researched in more detail than the other interventions. We give a description of the intervention, an overview of the potential implementers and their historical hits, a description of how we think a funder could work on this intervention, and the case for and against looking into the intervention further.

Pooled procurement

We are interested in pooled procurement given a relatively large amount of evidence on effectiveness (e.g., related to affordability, availability, and quality of products), many success stories across various contexts and health products, and given that we think there is (growing) room for more funding. Note that we reviewed pooled procurement as a market shaping tool in a previous Rethink Priorities report (Kudymowa et al., 2023, pp. 28-33). A quick introduction:

• “Essentially, pooled procurement^[10] means that buyers ‘pool’ their financial, technical or human resources to purchase products. The basic idea is that it can increase the buyers’ bargaining power by aggregating their demand, and increase efficiencies through sharing of human resources and technical capacity” (Kudymowa et al., 2023; Parmaksiz et al., 2022, p. 2).

• It can be used across many sectors and products, but it is, by far, most prominently used and discussed for health products. Thus, this is what we focus on in the following.

• Pooled procurement can be used to achieve various goals, such as to reduce prices, increase availability and secure supply stability, speed up drug access, and improve procurement efficiency and quality standards. This is particularly relevant for low- and middle-income countries (LMICs) as many countries are heavily reliant on imported health products^[11], which are often unaffordable^[12] or unavailable.^[13]

• Many different pooled procurement modalities exist with varying degrees of coordination,^[14] geographic levels (within or across countries), and types of products. The higher the degree of coordination, the higher the complexity and cost, but also the larger the potential benefits (Rao, 2006, p. 27).

While we decided to focus on the health sector in the following section, we would like to briefly discuss pooled procurement in other sectors. Our impression after a few hours of search is that focusing on other sectors is unlikely to yield much useful/actionable information. A few thoughts:

• We struggled to find information on pooled procurement outside the health sector. Our impression is that >90% of search results on pooled procurement are related to the health sector, and the few that aren’t are mostly focused on high-income countries. Moreover, much of what we found cannot be directly called pooled procurement, but seemed more vague (e.g., countries coordinating on procurement for infrastructure, but hard to know what exactly they do and what the benefits are).^[15] We don’t know precisely why pooled procurement seems less common in other sectors, but here are some thoughts:

  • We suspect that this might be partly because the health sector has a comparatively high share of standardized products that are needed in large quantities (e.g., essential medicines) and that are well-suited for pooled procurement (in comparison to, say, power plants, which are too specialized for demand aggregation).

  • This, however, does not explain why homogenous infrastructure projects (such as roads) and commonly used materials (such as concrete) appear less common in pooled procurement projects. It could have to do with fluctuating demand over time or the fact that, while healthcare is often a big part of government activity, some other sectors like construction are more typically outsourced to other providers who procure inputs on their own (e.g., concrete).

  • We would also like to note that while ‘pooled procurement’ is a commonly used term in health, it seems to be used less in other sectors. ‘Centralized procurement’ or ‘centralized demand’ seem to be more common terms in other sectors. However, we did not find much even with different search terms.

• Given that we found so little evidence/discussion of pooled procurement in other sectors, we expect it would also be much harder to find concrete funding opportunities.
• Our prior is that it is unlikely that pooled procurement is more impactful in sectors other than health. This is because the procurement of health products has the direct life-saving benefit of making medicines more available, faster, cheaper, and at a higher quality, which is not the case for other examples we’ve seen so far (e.g., procurement of buses or textbooks).
• Nonetheless, here are some examples we found of pooled procurement in other sectors:
  • E-buses in Chile (Cities, Climate, Finance Leadership Alliance, n.d.)
  • UNESCO (2023) recommends pooled procurement of textbooks in sub-Saharan Africa
  • ATEX (African Trade Exchange, n.d.) is a platform for pooled procurement of several products (e.g., fertilizer and other agricultural products, food):
    • The ATEX was launched to mitigate supply chain challenges due to the Russia-Ukraine war and making goods available more reliably and at competitive prices, to mitigate the risk of food shortages (reliefweb, 2022). We have not seen any info on their reported benefits.
  • The Digital Impact Alliance’s (DIAL, 2019) analysis suggests that pooled procurement of the aid sector’s spend on mobile communication channels might improve pricing and quality
  • Power pools (e.g., the Wikipedia, 2024):
    • Our impression is that power pools are a related concept to pooled procurement (though not exactly the same thing). Power pools create a common electricity market and pool the electricity demand and supply across countries. We haven’t investigated the concept in detail.
  • Proenca and Simard (2018) investigated ways to make public procurement of standardized goods and services in Brazil more efficient. They recommend establishing a pooled procurement mechanism by creating a central procurement agency and estimate that “carrying out the proposal will save an estimated 15-22% of total expenditure in standardized goods and services. This represents an additional 360-530 million USD in the government budget each year” (p. 2). They also point to other studies that found cost savings of ~15-20% due to pooled procurement (e.g., Dimitri et al., 2006)—which is in about the same ballpark as typical cost savings reported in the health sector (see here).

Implementing organizations and historical hits

Pooled procurement dates back to the 1970s and is now a fairly well-established tool, particularly among large global health organizations (e.g., Global Fund, Gavi, UNICEF, WHO). Many success stories (and some known failures) exist, though causal attribution and extrapolation of findings is difficult.^[16]

Several analyses point to pooled procurement being successful in many cases, particularly at reducing drug prices (by ~15-30% in many cases, in some cases up to 99%):

• Dubois et al. (2021) found price reductions of 15%, on average, when reviewing data from seven LMICs and 40 essential medicines from 16 therapeutic areas.^[17]

• A USAID (Rao et al., 2006, p. 46) report found cost savings of 7-88%^[18] (over unit price) for several large pooled procurement mechanisms such as ECDS,^[19] ACAME,^[20] GCC,^[21] and GDF^[22].

• Kim and Skordis-Worral (2017) found that the Global Fund’s voluntary pooled procurement scheme reduced the price of an ARV from 2009 to 2013 by ~16-19%.^[23]

• See also the table below for more results & successes.

There is also evidence on various other benefits, e.g. increased availability of medicines and quality of products.^[24]

On the other hand, there are also some known failures and major barriers/limitations to pooled procurement, such as:

• Failures (adapted from Kudymowa et al., 2023, pp. 32-33):
  • The Pacific Island countries have discussed pooled procurement for two decades, but eventually dismissed the idea due to concerns with national sovereignty and political tensions, high advance funding requirements for administration and operations, and insufficient aggregate demand to result in price reductions.

  • The East African Community (EAC) has discussed pooled procurement for essential medicines since 2007 and a proposal has been approved by the Council of Ministers in 2008. Two of the major reasons stated for introducing this mechanism are: (1) donor-funded supply (e.g., by PEPFAR and the Global Fund) of some essential medicines is insufficient to meet the demand, which means that there is still a major treatment gap;^[25] (2) another goal of the EAC is to promote local manufacturing of drugs.^[26] However, little progress has been made since 2018 for unclear reasons.^[27]

  • The Asthma Drug Facility was discontinued in 2014, partly due to funding difficulties and partly due to lack of demand from countries.
• Barriers / challenges / limitations (adapted from Nemzoff et al., 2019; Parmaksiz et al., 2022):
  • Regulatory and legal barriers (e.g., regulatory misalignment across countries/regions, lack of clarity around applicable laws)
  • Uncertain cash flow (e.g., inability or limited willingness to make initial capital investments)

  • The politics of pricing (e.g., debates on whether there should be a single price and/or price transparency)^[28]

  • Conflicting competition among manufacturers (e.g., tension between local manufacturing and international tendering for health products)
  • Loss of autonomy of participating members
  • Geographic considerations (e.g., in some cases, high shipping costs might be the main reason for high prices [as in small island states], which would not be alleviated by pooled procurement)
  • High level of market concentrations (i.e., low level of supplier competition) makes pooled procurement less effective at reducing prices
  • Requires specific market features, including:
    • “(i) Large enough volumes; (ii) a supplier that can supply such volumes, and (iii) a buyer that commits to purchasing those volumes. For (ii) and (iii) to work, it is critical for there to be trust between buyers and suppliers” (see Nemzoff et al., 2019)
    • Doesn’t make sense in sectors/areas with highly specialized or quickly evolving needs, or very complex products that are hard to standardize

  • (Theoretical) trade-off between short-term economic benefits and longer-term availability of products. For example, pooled procurement could reduce competition among suppliers, and small and medium-sized suppliers might be driven out, potentially leading to shortages and higher prices in the longer run; however, to our knowledge, this negative effect has not been demonstrated in practice yet.^[29]

• Central contracting^[30] in many Latin American and Caribbean countries;

• Revolving Fund focused on vaccines, Strategic Fund focused on other medicines and vector control, currently planned expansion to NCDs
• Countries access short-term loans for medicines/vaccines
• Bulk purchases at low price and other services (e.g., technical support in quality assurance, supply chain management, demand forecasting)
• Funding source: Fee charged to members as % of purchase price
• Revolving Fund has helped increase PAHO vaccination rates from 25% in 1978 to >80% in 1993 (Ahmad et al., 2022)
• Strategic Fund reduced prices of ARVs by up to 97%, and of of Amlodipine (cardiovascular medicine) in some countries by up to 99% (Lal et al., 2022)
• Quality assurance for all procured products
• Central contracting of initially only vaccines, and later other essential medicines for 14 Pacific Islands + several other (mostly sub-Saharan African) countries
• Countries receive interest-free short-term loans for medicines
• VII also plans shipments and purchases, and offers contract and logistical support
• Long-term vision is for countries to graduate out of support, but increasing demand for VII support is expected in coming years
• Funding source: Donor funds (BMGF, GAVI, UNICEF)
• Vaccines arrive several months sooner (UNICEF, n.d., pp. 71-72) in countries; prevents stock-outs
• Procurement at more affordable prices (have not found any concrete figures)
• From 2015-2019, VII facilitated (UNICEF, 2020) delivery of ~560M doses of vaccines, and ~400k packs of ARVs
• Combination of various market shaping interventions (e.g., pooled procurement, market analyses, demand forecasts, technical assistance to regulators and manufacturers) and actors (e.g., UNICEF, WHO, PATH)
• Support was implemented in 3 phases, each focusing on a slightly different goal and strategy
• Pentavalent vaccine reached fully satisfied demand
• Price of pentavalent vaccine has dropped by ~70% from 2006 to 2023
• It is estimated that Hib and HepB components alone could avert 10M deaths and 390M DALYs and generate >$250B in economic value in 73 Gavi countries in 2001-2020
• Pooled procurement from 2007 to 2016 to make pediatric ARVs more accessible in 40 LMICs
• Combined with other interventions (e.g. consolidating ARV formulations, assistance with supply chain management)
• Implemented by CHAI and Unitaid; pediatric ARVs funded by national programs, Global Fund and PEPFAR
• Highly successful, but transition away from pooled procurement was likely inadequately executed
• Unitaid estimated that ~210k deaths were averted due to this program and that each $1 spent on this program had $11 in benefits from mortality reductions) and ~$2 in cost savings
• Estimated price reduction of ARVs between ~35-81%
• Transition away from pooled procurement has been criticized for inadequate planning, uncertain long-term market sustainability
• National pooled procurement mechanism across the country’s 9 provinces
• Provinces aggregate demand into a national 3-year contract, with multiple awards per product and minimum volume pledges for vendors
• Price preferences are given for local producers
• Price of ARVs has dropped by 68% in South Africa since 2011
• Public sector is estimated to save $326M over three years
• An estimated 2M extra people could be treated (on top of 4.7M who currently receive treatment)
• CHAI report (CHAI, 2021) found that price per unit is even lower than what Global Fund or PEPFAR pay
• Recent pilot project in which 23 cancer centers in India introduced pooled procurement for 40 (mostly essential) cancer drugs
• Group contracting, i.e. joint negotiation of drug prices, and selection of suppliers and distributors, but actual purchasing is done by individual cancer centers
• Entire process took 1 year
• Process was supported by admin grant from a large philanthropic foundation in India (Tata Trust)
• Price reductions of 23-99% (median: 82%); higher with generics than with brand names and newly patented medicines [note that savings were calculated in comparison to the maximum retail prices, i.e., not prices previously paid]
• World’s largest TB drug procurer, established in 2001 with ~130 countries using their procurement; procured drugs worth >$300M in 2020
• Stop TB’s operating costs are ~$8-10M per year, (Stop TB, 2021) and the operating costs for just the GDF are likely lower than that
• [Note that we came across this initiative at a very late stage of the report and didn’t have time to investigate further details.]
• Reported $31M of savings in 2018 due to price reductions, which represents 14% of TB drug costs delivered by GDF in 2017
• If savings were used to buy more TB drugs, “the savings would be sufficient to provide 960,000 people with drug-sensitive TB medicines, 53,000 people with shorter regimens for drug-resistant TB, or up to 37,000 people with conventional regimens for drug-resistant TB” (Stop TB, 2018)
• Price reductions of MDR-TB drugs of around 27-33% within a year

Case studies and success stories on pooled procurement are fairly abundant. See Table 3 below for a variety of examples across different contexts, modalities, and health products:

Table 3: Selected examples of pooled procurement

At first glance, pooled procurement does not seem particularly neglected, as it is a fairly well-established and commonly employed tool by large global health organizations and often funded by BMGF. However, our impression is that pooled procurement is currently still underused and will likely grow in importance for governments in the future. Moreover, our impression is that the set of health products and settings in which pooled procurement could be gainfully implemented is vast, as the benefits do not seem to be restricted to certain products and seem to to generally outweigh the downsides of pooled procurement.^[31]

The introduction of pooled procurement mechanisms has been or is debated in many countries/settings, pointing to significant interest:

• There have been recent calls to introduce pooled procurement in various countries/areas, e.g.:
  • In India to “reduce costs, ensure quality, and prevent stockouts of life-saving drugs” (Drishti IAS, 2023; Neelakantan & Kulkarni, 2023). We have not found any active advocacy organization working on this in a quick search.
• A replication of the PAHO Revolving Fund for vaccines in Southeast Asia (Ahmad et al., 2022). The authors suggest obtaining seed funding from donors such as BMGF. We’re not aware of concrete ongoing efforts to introduce this.
• Several African countries have shown interest in pooled procurement, e.g.:

  • A recent report by the Government of Botswana (2021) (in collaboration with PEPFAR, USAID, and other organizations) investigates the potential for introducing pooled procurement in Botswana and highlights the country’s “interest and political will”, but also points out regulatory / legal barriers that need to be resolved first and that technical support is required. The report also highlights the actions and/or interest related to pooled procurement of other African countries, such as Rwanda,^[32] Gabon,^[33] and Tanzania.^[34]

• Pooled procurement has been recommended in the WHO guideline on country pharmaceutical pricing policies (2020, pp. 32-34)^[35] and in several instances by the CGD (Silverman et al., 2019, p. 50; Kaufman et al., 2021).

• Some initiatives faced funding constraints, pointing to room for more funding, e.g.:
  • The UNICEF Vaccine Independence Initiative (VII) (see in table above) planned an expansion to other essential medicines, but was only able to raise $35M out of the desired $100M in donor contributions (Nemzoff et a., 2019, p. 9). Much of the $35M came from BMGF.

We expect that pooled procurement will grow in relevance for LIC governments as they transition to MICs due to what a 2019 CGD report calls the “triple transition in global health procurement” (adapted from Silverman et al., 2019, p. xii):

• Transition from donor aid: As economies grow, countries are gradually transitioning out of donor-funded procurement of health products:
  • However, we would like to point out that this transition is unlikely to be very quick and sudden. For example, 9 countries may transition out of Gavi support until 2040, which represents ~20% of Gavi disbursements. For the Global Fund, 10 countries may transition until 2040, but these only represent ~1% of disbursements (Georghe & Baker, 2023). Moreover, countries transitioning out of Gavi support can still participate in UNICEF Supply Division pooled procurement, so the transition seems more gradual than we initially thought.
• Epidemiological transition: Procurement of health products will gradually need to shift from communicable to non-communicable diseases.
• Transition in health system organization: Government procurement of health products will become more important as countries transition away from out-of-pocket spending towards universal health coverage.

What we think a donor could do

Donors could consider several options:

1. Fund advocacy initiatives that aim to introduce a pooled procurement mechanism, e.g.:

   1. Partnerships for African Vaccine Manufacturing (PAVM) “has been collaborating with various stakeholders to advocate for a pooled procurement mechanism for vaccines across the continent” since 2021 (Kaseya & Tshangela, 2023).^[36] We have not come across other advocacy initiatives, and we have not found much detail on the PAVM.^[37] We contacted the PAVM program manager to learn whether they have room for more funding and are still waiting for a reply.

2. Fund existing pooled procurement initiatives of large global health organizations, e.g.:

   1. PAHO Strategic Fund: PAHO seems to still have room for more funding^[38] (at least as of 2021, we aren’t quite sure how this changed since then) and is willing to expand to more medicines (e.g., for NCDs). The idea would be to fund the functioning of the Strategic Fund rather than to fund the purchase of drugs. An advantage of this would be that the benefits (e.g., price reductions) would directly translate into government savings as the drugs themselves are country-funded.

      1. Note that we briefly tried to determine the relationship between the Strategic Fund and the Revolving Fund. Our impression is that they operate fairly independently, though they have substantial overlap in how they are set up. For example, they operate under the same funding modalities and lending conditions, but they differ somewhat in their support services provided and their member countries (see Nemzoff et al., 2019, p. 5). We couldn’t figure out if they share the same staff at PAHO.

   2. Other pooled procurement initiatives of large global health organizations, e.g. Gavi, CHAI, Unitaid etc. (see table above for examples). While these are also well-established off-the-shelf options and there seems to be room for more funding in at least some cases (e.g., here), a “downside” of many of these initiatives is that governments may not (or only partially) benefit from government savings, as medicines are often funded by donors or via co-financing arrangements.

      1. By “downside” we mean that this doesn’t quite fit under the scope of government procurement, which was the intended scope of this report. However, it might be worthwhile to consider pooled procurement from a wider lens. Our guess is that donor-funded pooled procurement could potentially be more cost-effective than procurement done directly by governments. See the following footnote for an explanation.^[39]

We aren’t sure whether it would be more worthwhile to fund advocacy or to fund the expansion of existing initiatives:

• On the one hand, it might be preferable to fund an existing mechanism and help it expand to other medicines or areas, such as the PAHO Strategic Fund or the UNICEF VII. Both have a good track record and room for more funding, and they want to expand to a wider range of health products. That might be a relatively safe, easy, and quick option.
• On the other hand, a downside of many existing mechanisms is that they create parallel systems for procurement and aren’t necessarily set up for the long-term. What the PAVM is advocating for would be essentially a more country-owned pooled procurement mechanism as a replacement of Gavi/UNICEF vaccine procurement support once countries are no longer eligible for that support. So this would be thinking a bit more long-term. However, the outcome would be more uncertain, and we find it hard to estimate how promising advocacy could be (though there definitely seems to be lots of country interest for pooled procurement).

It does not seem straightforward to determine for which health product and for which country pooled procurement makes sense, as this depends heavily on various economic, political, and legal factors. A starting point could be medicines that fit the following criteria (from a 2021 CGD/ PAHO panel discussion [Kaufman et al., 2021]): “(1) medicine types that have recently been added to the WHO Essential Medicines List^[40] (e.g., monoclonal antibodies); (2) new innovations for chronic diseases (e.g., insulin analogues for the treatment of diabetes); (3) essential medicines that are in short supply or are at risk of short supply; and (4) medicines with a concentrated supplier base to increase market competition (e.g., biosimilars)” (Kaufman et al., 2021).

Overall, do we think a donor should look into this further? Case for/against

We do not recommend further research into pooled procurement at this point, but think that keeping an eye out for emerging initiatives over the next years in the field could be worthwhile. Further desk research would unlikely be fruitful for determining plausible and good options for donor funding, but expert input might be useful.

Our current key uncertainties are:

• We haven’t been able to find off-the-shelf funding options that seem to be a good fit for donors and we aren’t totally sure whether these exist (currently or in the near future).

• We found very little information on the costs of implementing such a mechanism^[41] (only for the Stop TB GDF, see here). We expect there is likely a large variation in costs (and in cost-effectiveness).

  • Another related aspect is that we don’t know the counterfactual costs (i.e., how much governments have to pay for the procurement in the absence of a pooled procurement mechanism). This is probably difficult to figure out based on desk research, but is an important figure to estimate total cost savings of governments.

• Given that regulatory/ legal misalignment is a key barrier to pooled procurement, we are uncertain in which countries pooled procurement can be currently implemented. Moreover, the success also heavily depends on political will, and we don’t currently have a good overview of where the political will for pooled procurement is sufficiently strong.

Why would a donor want to look more into this?

• Pooled procurement is a fairly well-established and tested tool that credibly resulted in substantial price reductions of health products and various other benefits in many cases, such as increasing availability and quality of medicines in LMICs. It is also widely endorsed, including by WHO and CGD.
• Even though pooled procurement does not look particularly neglected, there is a vast set of potential countries, products, and circumstances in which it is not used yet and could be considered. We think that pooled procurement will likely grow in importance in the future.
• There seems to be room for more funding: Some previous attempts have failed and some initiatives have expanded to a smaller extent than envisioned (at least partly) due to funding constraints. Moreover, large upfront investments can be needed, which countries are sometimes unwilling or unable to pay. We expect these upfront costs are mainly for two reasons:

  • If the chosen modality is a revolving fund (as is the case with PAHO), this requires an initial capital investment by donors or members, which would then be replenished through purchasing fees.^[42] For example, the OECS/PPS^[43] established a revolving fund across countries. Participating countries had to “commit one third of their annual pharmaceutical budget to the mechanism, even before its establishment” (Parmaksiz et al., 2022, p. 9).

  • It doesn’t seem easy to set up a pooled procurement mechanism given all the coordination, research, negotiation, and administration that needs to go into this, especially across countries. It would probably require feasibility studies, assessing the regulatory/economic environment, then potentially require some regulatory harmonization (e.g., regarding quality standards, or product registration processes) across countries. Then probably lots of negotiations on how to exactly set up the mechanism (i.e., which products, what quality standards, what kind of set up).^[44]

Why should a donor not fund this?

• Many initiatives seem to have struggled due to regulatory and/or legal barriers and misalignment across countries. These barriers need to be overcome first before a national or international pooled procurement mechanism can be introduced.^[45]

• While there seems to be a lot of interest, we haven’t seen many concrete advocacy organizations that promote pooled procurement.
  • While funding large global health organizations could potentially be a good off-the-shelf funding option, in many cases, the savings resulting from price reductions would not fully benefit national governments, as medicines are donor-funded or funded via co-financing arrangements. We haven’t investigated this line of thinking further (i.e., whether funding initiatives like Gavi could be a good option), as it was out of scope for this project.

Support civil society actors to increase monitoring of project execution

We are interested in this intervention given some evidence of effectiveness, and some indication that there is scope for improvement. In particular:

• Fazekas and Blum (2021, pp. 18-19) reference two high-quality field experiments studying civil society supervision of government contract execution. One of these studies, Lagunes (2017), found that an extended audit of small-scale infrastructure works by a civil society organization (CSO), explicitly supported by the national corruption agency, led to a 51% cost reduction.^[46] The other field experiment, which focused on increasing community participation in (anonymous) feedback and review, found that this had some impact in reducing the costs of labor but the effect on the cost of materials was insignificant.^[47]

• Schwartz et al. (2020, p. 30) cite two papers that indicate that many government projects that are started are not finished: “Williams (2017), analyzing a database of 14,000 development projects in Ghana, found that one-third of projects that start are never completed, wasting on average one-fifth of local government investment. In a sample of Nigerian federal government social sector projects, Rasul and Rogger (2016) found that a quarter were not completed”.

This evidence prompted us to look at this intervention in more detail. The main ways that we would expect this intervention to improve procurement are:

• Reduce costs of execution (including due to overruns)
• Increase likelihood of completion of projects
• Improved monitoring of contract execution could have a selection-effect which reduces the number of low-integrity bidders (such as firms that intentionally underbid initially, anticipating overruns and re-negotiations later in the process).

Implementing organizations and historical hits

Table 4 lists organizations that have implemented this kind of work, and some examples of success. In general, we found that many “success stories” related to earlier stages of  procurement rather than the execution phase.^[48] This may be in part because “successful” monitoring of execution could look the same as a scenario where monitoring was not needed.

Table 4: Selected examples of civil society monitoring of project execution

We estimate that funding for the organizations we have identified is < $3m per year, which we describe further below. These efforts generally seem to focus on large infrastructure projects, which seems reasonable to us given the large public spend in this sector.

• CoST: according to annual reports from 2019 (CoST, 2019), 2020 (CoST, 2020) and 2021 (CoST, 2021), the international secretariat raised £1m – £2.6m per year, with a decreasing trend over time. The majority of CoST funding comes from the UK Foreign, Commonwealth and Development Office (FCDO). In addition, CoST programs in each country/ city can raise funds, but according to a 2019 review of CoST, in two years self-funding amounted to only £0.3m and programs were highly dependent on funding from the secretariat (Davila et al., 2019, p. 33). It’s unclear whether this has increased since 2019.
• AFIC: the procurement-related work at AFIC appears to have been funded by the William and Flora Foundation for $150,000 per year between 2018 – 2023 (Hewlett Foundation, n.d.). Their funding is not only focused on efficiency but also includes a gender equality element.
• Tracka NG: expenditure was roughly $200,000 per year in 2018 and 2019 (BudgIT, 2019, p. 40). This figure was not reported for 2020 onwards but we expect this to be slightly higher given expansion to more states. The Tracka (n.d.) lists the EU, Osiwa, British Council, Luminate, and MacArthur Foundation as supporters. The parent organization, BudgIT, is also funded by a variety of donors including BMGF.
• We could not quickly identify how much GIZ is spending on Uganda’s Contract Monitoring System.
• We have not estimated the funding for Transparency International or Planet as we do not believe they have sustained programs for this kind of work.

We are likely to have missed some relevant organizations, and we remain uncertain about the extent to which the World Bank and other donors may already fund monitoring for co-financed projects. However, we expect that there is a gap in execution monitoring for government procured projects, whether in infrastructure or other domains.

What we think a donor could do

It seems plausible that a donor could make a grant to CMS, Tracka NG, AFIC, or CoST to train and deploy more monitors in countries of interest. Additionally, donors could consider:

• funding capacity to increase follow up with government officials after monitoring (which currently seems limited), or
• funding increased engagement with (unpaid) community champions and CSOs.

Tracka reported in 2019 that it costs $500 per month to hire a project officer (BudgIT, 2019, p. 8).

Overall, do we think a donor should look into this further? Case for/against

We do not recommend that donors look into this further as a standalone topic. While the intervention is likely to have low costs, most of the organizations we identified do not appear to have enough government endorsement and support to be highly effective.

We believe that the impact of an intervention without government endorsement would be significantly lower than the 51% estimated by Lagunes (2017), but we are uncertain about by how much. We asked Viktor Nestulia, Head of Ukraine Support at OCP, whether initiatives need government endorsement to succeed. He said that he believes that initiatives that are collaborations between civil society and government organizations are most likely to have an impact. It is important that there is at least potential for political will, which can then be encouraged by implementers. He confirmed that often some parts of a government are in favor and others are against a reform, and he said that in those cases OCP looks for intermediaries who can facilitate dialogue between the parties to bring them closer together. All in all, this has confirmed our thought that some government endorsement is necessary.

We believe that a good way to identify new implementers with government endorsement could be through OCP’s Lift program (more below).

Why would a donor want to look more into this?

• There’s academic and anecdotal evidence that a significant proportion of government-funded projects are not completed, and beyond that it seems likely that a further proportion may be low quality or delayed.^[53]

• Lagunes (2017) indicates that CSO monitoring in conjunction with government endorsement can be highly effective in reducing construction costs.
• This appears to be neglected—and is likely particularly so for small-scale projects that could directly impact communities’ access to water and healthcare.

Why should a donor not fund this?

• Academic evidence is limited, and only for small-scale public works in one country
• While there are implementing organizations in countries of interest, it is difficult to find concrete examples of impact.
• In many cases, it’s unclear how much endorsement these organizations receive from governments, and this may be important because the key evidence for this intervention, Lagunes (2017), measures the impact of CSO monitoring at the local level with clear endorsement by the central government.
• It’s possible that there is already some monitoring commissioned by the World Bank or other donors on larger projects, which could lead to duplication of efforts.

We deprioritized looking into e-procurement due to concerns about neglectedness and effectiveness

E-procurement is a collection of interventions in which the government digitizes a specific part of the procurement process. Examples include publishing tender announcements and contract awards electronically on a public website, online availability of tender documents, electronic management of qualification documents of bidders, and digital submission of tenders through specialized IT systems (Fazekas & Blum, 2021, p. 9). E-procurement was one of the interventions suggested in the brief, and we took the following steps to review it early in the project:

• Considered the evidence available in Fazekas and Blum (2021),
• Read a report on e-procurement by Charity Entrepreneurship (they did not recommend e-procurement as an area for new charities. Charity Entrepreneurship gave us permission to link to this version of the report),
• Spoke to multiple experts about who is currently rolling out e-procurement and how much success they are achieving.

Our impression from this shallow review is that the World Bank is spending significant amounts of money on installing e-procurement systems in low- and middle-income countries.

Since we think the World Bank is active in this area, we deprioritized further research into e-procurement systems, but note that there might be room for the following interventions:

• Full digital transformation/ change management programs to sustainably implement e-procurement software purchased by the World Bank
• Providing incentives to government officials to encourage use of existing e-procurement systems
• Training to increase use of existing e-procurement systems.

We did not spend time on researching these interventions, and are not sure how a philanthropic donor could practically contribute to them.

A program like Open Contracting Partnership’s Lift might be worth considering for donors interested in improving procurement

It may be worthwhile supporting a program similar to Open Contracting Lift: an 18-month impact accelerator program that offers technical and financial support to teams aiming to improve public procurement. Teams receive $35,000 and 200 hours of technical assistance.

Both government departments and independent implementers can apply to the program, and the teams often include representatives from both government and civil society organizations (Viktor Nestulia). Nestulia indicated that Lift funds a broad spectrum of projects related to open contracting, as long as they solve an existing problem. The aims of the projects are varied: some focus on improving efficiency, while others focus on outcomes such as sustainability and inclusion of underrepresented groups in procurement. We understand that civil society monitoring of project execution, or improving the use of e-procurement could also fall under this program.^[54]

Lift started its first program in 2019 with five teams (OCP, 2019), did a second round in 2021 with seven teams (OCP, 2021), and in the second half of 2023 they ran a third round with ten teams (OCP, 2023).^[55] We note that in 2021, OCP indicated that they wanted to expand the 2023 Lift program to 15 teams (OCP, 2021, p. 8). They did not achieve this due to a lack of funding (Gavin Hayman, email exchange).

From scanning the list, we think the two most promising examples currently being supported are a program that aims to expand access to basic medicines in Brazil, and one that aims to strengthen Bangkok’s resilience to flooding. The reason we think these could have a high social return on investment (SROI) is that they could save government money as well as positively impacting health or saving lives in other ways.

This year, Lift received more than 150 proposals but only selected ten of these to start the program in September. We have not been able to find information online on rejected programs, but Viktor Nestulia indicated that he believes there are many promising projects that Lift is not able to fund. He indicated that the main bottleneck is OCP’s resources and capacity to support teams. Two types of projects that may be interesting to fund through this platform are projects that support monitoring by civil society organizations (more here), or projects that aim to increase the use of e-procurement systems (more here).

We think (with moderate confidence) that supporting a future selection of Lift proposals, or a similar initiative, could be an interesting opportunity for funders wanting to work on government procurement, because:

• The application process filters for teams that are both motivated and skilled. These teams are locally based, which lowers implementation barriers.
• There are Lift applications by mixed teams with part government actors and part civil society actors, which we think increases the chance of success.
• The combination of technical assistance and financial support offered by Lift can significantly impact smaller organizations striving for specific, yet potentially influential reforms.
• Given the cost of support offered, several initiatives could be supported. This means there is opportunity to learn and gain successes despite a lower chance of success per initiative.
• Lastly, we think programs currently funded by Lift may not have been chosen for their high social return on investment (SROI), for example, two of the winning initiatives take place in the United States where we think (with limited confidence) that the SROI would be lower than in LMICs. Therefore, there is potentially an opportunity to support initiatives that are not currently prioritized by Lift.

Our main uncertainty is that it is difficult to predict a priori which initiatives have a high chance of success, and that it may be difficult to get small scale initiatives to share the information needed to make this judgment.

While a lot of money is being spent on improving some areas of procurement, we think it is likely that other parts of the area are neglected

We did not research whether government procurement as a whole is neglected, because it’s such a wide-ranging topic. While we think a lot of money is being spent on improving some areas of procurement (for example e-procurement by the World Bank, as explained above), we think it is likely that other parts of the area are neglected. We therefore focused on determining neglectedness per individual intervention.

In terms of neglected geographical areas, Michael Jarvis shared that Asia receives minimal philanthropic attention, and that Latin America has had successful projects that struggle to scale due to funding limits.

Notes

1. ^{^}

    We use the term inefficiency somewhat loosely; more on that here.

2. ^{^}

    For example, at the time of writing there were large protests in Panama against a mining contract that was recently signed by the government and is perceived by the population as unlawful. This has led to direct economic losses, and a lower investment-grade rating from JP Morgan (Murray, 2023).

3. ^{^}

    Such as car rentals, office goods, phone contracts and software.

4. ^{^}

    We have not read the paper in detail, but an underlying assumption seems to be that bodies are aware of the opportunity to achieve lower prices through Consip, i.e., they are not so inefficient as to miss this fact.

5. ^{^}

    Other reductions come from: reallocating contracts from medium and large companies to small companies (1.4%), using one-off electronic auctions instead of framework contracts (1.2%), and lowering spending concentration for buyers of very high concentrations (1.1%).

6. ^{^}

    Mihaly Fazekas confirmed that the methodology could capture corruption, but believed the results should mostly capture inefficiency.

7. ^{^}

    For example, Fazekas et al. (2021, p. 31) generate their savings estimate assuming 20% of those products currently advertised for only 1-12 days are moved to longer advertisement procedures, but it’s possible that the savings estimate would be greater for a more efficient system where say 50% of these products were advertised for 12+ days.

8. ^{^}

    It’s unlikely that our longlist is exhaustive, but the interventions listed do cover all of the kinds of inefficiency that we identified using our first principles approach.

9. ^{^}

    As described in the longlist, there are several types of e-procurement. We highlight the selection of e-procurement which we think are most promising for a philanthropic donor to focus on here, and describe more types of e-procurement separately in the longlist.

10. ^{^}

     “Also referred to as joint, bulk, group, centralized, cooperative or collaborative procurement” (Parmaksiz et al., 2022, p. 2).

11. ^{^}

     For example, in sub-Saharan Africa, the proportion of imported health commodities is on average ~80% in terms of total market value. The remainder (~20%) is local manufacturing (Silverman et al., 2019, p. 18).

12. ^{^}

     In LMICs, prices for basic generic medicines (such antihypertensives and pain relievers) are in many cases 20-30x more expensive than in wealthy countries (Silverman et al., 2019, p. 5).

13. ^{^}

     “Nearly 2 billion people have no access to basic medicines” (Chan, 2017, p. 14).

14. ^{^}

     The level of coordination can range from informal agreements to share price information among participants over group contracting (i.e., joint price negotiation and selection of suppliers) to fully integrated supply chain operations (i.e., a centralized distribution organizations that combines group purchasing and supply chain-related activities) (Nemzoff et al., 2019, p. 3).

15. ^{^}

     For example, the East African Community (n.d.) coordinates infrastructure procurement (among other sectors), but it’s unclear to us what exactly they do and how (e.g., CPCS, 2016).

16. ^{^}

     This is partly because most of the evidence is based on case studies and simple before-after comparisons of outcomes (e.g., drug prices), and partly because pooled procurement is rarely implemented in isolation, but usually combined with other market shaping tools, such as demand forecasting. Moreover, there are many different modalities in which pooled procurement can be implemented, which makes comparison across mechanisms challenging.

17. ^{^}

     We think the internal validity of this study is relatively strong. The study uses a fixed effects regression model within-molecule-country variation in prices. They also use a two-stage Heckman procedure to correct for unobserved selection (as there might be unobserved selection for which drugs pooled procurement is used), and various other robustness checks based on within-molecule-country variation, using product-level fixed effects.

18. ^{^}

     Our best guess is that these figures are based on simple before-after comparisons of prices.

19. ^{^}

     Eastern Carribean Drug Service.

20. ^{^}

     Association Africaine des Centrales d’Achats de Médicaments Essentiels.

21. ^{^}

     Gulf Cooperation Council.

22. ^{^}

     Global Drug Facility.

23. ^{^}

     This analysis was done using a difference-in-difference analysis.

24. ^{^}

     See Kudymowa et al. (2023, pp. 30-31) for more detail.

25. ^{^}

     “Though most of the medicines for the treatment of HIV/AIDS, TB and malaria in the region are imported under donor funded programmes like the Global Fund, the Clinton Health Access Initiative (CHAI) or the US President’s Emergency Plan for AIDS Relief (PEPFAR), the limited budget of these programmes can only treat a small percentage of the population in these countries who need treatment” (Syam, 2014, p. 5).

26. ^{^}

     “In fact, the high price of essential medicines like ARVs charged by multinational pharmaceutical companies, rather than import tariffs, is the predominant reason for the high cost of medicines. […] Therefore, the EAC Partner States have to pursue policies to support the development of a local pharmaceutical industry which can supply affordable and quality medicines in the region. Ownership of production facilities by local nationals may offer several advantages including continuity of production and supply in the face of changing economic circumstances which can avoid disruptions in the pharmaceutical supply chain, building domestic technological capacity and skill development, securing a competitive market environment that may constrain the pricing power of multinational suppliers” (Syam, 2014, pp. 5-6).

27. ^{^}

     UNECA (2023, p. 11) cites several challenges, including: “limited awareness of the benefits of pooled procurement at the national level”, “lack of clear potential role for local manufacturers in the regional pooled procurement process”, and “limited financing and complex financing mechanisms among the member states.”

28. ^{^}

     According to Nemzoff et al. (2019, pp. 9-10), whether prices are transparent has implications for the magnitude of prices and whether pooled procurement arrangements are politically stable.

29. ^{^}

     Parmaksiz et al. (2022): “Although this explanation might be plausible, none of the included studies have demonstrated pooled procurement’s supplier competition reducing effect in practice. This lack of empirical evidence was also mentioned by Toulemon and Burns & Lee.”

30. ^{^}

     See Nemzoff et al. (2019, p. 3) for an overview of what central contracting entails and how it relates to other pooled procurement approaches.

31. ^{^}

     For example, according to the WHO (2020, pp. 25-27), “the effects on reduced price of pooled procurement due to economies of scale is likely to be generalizable”. According to a recent CGD / PAHO panel discussion: “Given that pooled procurement for a wide range of supplies can produce net gains—even if some of the products may be theoretically better suited to decentralized approaches—the benefits of collective negotiating power for a basket of purchases tend to balance out the downsides” (Kaufman et al., 2021).

32. ^{^}

     “Rwanda is open to pooled procurement agreements with a specific provision in the Rwanda Public Procurement Authority (RPPA) that foresees pooled procurement agreements in the future. However, despite this willingness to explore pooled procurement mechanisms within the region, gaps in the country’s policy and regulatory frameworks hinder the advancement of joint purchasing with other countries. One such gap is the lack of an autonomous medical regulatory authority (MRA) which has been cited as a challenge to the harmonization of medicine regulations in the EAC” (p. 48).

33. ^{^}

     “In November 2019, Gabon convened with eight other countries, including five Small Island Developing States, to commit to the establishment of a pooled procurement mechanism for vaccines. This new pooled procurement group is supported by the World Health Organization (WHO) Regional Office for Africa” (p. 47).

34. ^{^}

     “Tanzania signed an MOU in 2018 on the Provision of the SADC SPPS. Through this agreement, the country will share pricing and supplier information with other SADC member states. This agreement is expected to reduce pharmaceutical and medical supplies costs by 40 percent” (p. 54).

35. ^{^}

     WHO (2020, p. 33) found in a systematic literature review that “on balance, pooled procurement is likely to deliver more desirable than undesirable effects, as indicated by evidence of reduced prices of health products under a pooled procurement arrangement. […] The effects on reduced price of pooled procurement due to economies of scale are likely to be generalizable. […] This policy seems to be acceptable to government authorities given broad adoption of policies at different levels of administrative jurisdiction.”

36. ^{^}

     See African Union & Africa CDC (2022) for more detail on the planned procurement mechanism.

37. ^{^}

     We would like to note that during the writing of this report, Gavi pledged $1B to boost African vaccine manufacturing and has approved its African Vaccine Manufacturing Accelerator (AVMA). We aren’t sure how exactly AVMA and PAVM intersect with each other (Buguzi, 2023).

38. ^{^}

     CGD / PAHO panel discussion: “Gross-Galiano highlighted the need to increase the Fund’s capitalization, pointing out that while the Strategic Fund expanded to place purchase orders amounting to over US$235 million in 2020, it has not grown as much as the PAHO Vaccine Revolving Fund” (Kaufman et al., 2021).

39. ^{^}

     We will use Gavi as an example here: (1) We expect that any savings by Gavi are reinvested into further vaccine support, resulting in direct health benefits for populations. We are unsure how governments typically spend their savings from pooled procurement (i.e., if these are used to procure more health products). (2) We also think that supporting Gavi could be cheaper. As the procurement works in parallel to government systems, our guess is there isn’t the need for lots of resource-intensive regulatory alignment across countries—which is needed if the procurement works directly via governments.

    On the other hand, donor-funded pooled procurement initiatives aren’t intended as long-term solutions, and supporting country-led initiatives might potentially be a better long-term solution for countries transitioning away from donor support.

40. ^{^}

     See also Rethink Priorities’ recent report on the WHO Essential Medicines List (Kudymowa et al., 2023).

41. ^{^}

     We’ve seen lots of data on the costs spent on health products through pooled procurement, but not on the administrative / implementation costs of starting and running such a mechanism.

42. ^{^}

     Nemzoff et al. (2019) write: “However, donors’ and countries’ limited willingness to make the initial capital investment has hindered the expansion of these funds”.

43. ^{^}

     Organisation of the Eastern Caribbean States Pharmaceutical Procurement Service.

44. ^{^}

     This could also decrease the success probability of an initiative, reducing the SROI.

45. ^{^}

     Regulatory harmonization is a priority of the newly-founded African Medicines Agency (Wikipedia, 2023), but it hasn’t started operating yet.

46. ^{^}

     The study includes 200 small-scale infrastructure works in Peru, which the paper describes as “roads, sidewalks, exercise courts, public markets, and other miscellaneous projects” (p. 23). These were randomly assigned to control and treatment. For projects in the treatment group, the relevant district government (specifically the mayor) was informed of and reminded of the monitoring efforts; the treatment took the form of four letters across eighteen months from the Office of the Comptroller General and Proetica, a civil society organization. The author finds that contract length does not decrease.

47. ^{^}

     The relevant paper is Olken (2007). We have not read this paper, and rely on Fazekas and Blum’s description of the results. The paper’s abstract states that “increasing grassroots participation in monitoring had little average impact, reducing missing expenditures only in situations with limited free‐rider problems and limited elite capture”.

48. ^{^}

     For example, see this case study (Basel Institute on Governance, n.d.) from Pakistan, which reports savings from the tender phase, identified by comparing the final contract to the initial bid.

49. ^{^}

     This source (Transparency International, n.d.) shows 18 of the countries: Mexico, Honduras, Panama, Ecuador, Peru, Colombia, Paraguay, Argentina, Germany, Italy, Latvia, Hungary, Bulgaria, Rwanda, Zambia, Pakistan, India, South Korea, China, Indonesia (pp. 2-3). This database (Basel Institute on Governance, n.d.) includes some information on specific pacts up to 2019.

50. ^{^}

     The monitor’s reports are available Transparency International (2020). We skimmed these briefly but did not identify recommendations made by the monitor that clearly linked to savings or increased likelihood of project completion.

51. ^{^}

     CBM-I are part of Integrity Watch Afghanistan (n.d.), who also conducted monitoring for a World Bank infrastructure project in 2011 – 2013 (Integrity Watch Afghanistan, 2011 – 2013).

52. ^{^}

     The presentation was given at a Center for Effective Global Action (CEGA) conference (2016). We have not watched the presentation, but skimmed the slides.

53. ^{^}

     We do not think that all projects that are currently uncompleted could be improved by monitoring the execution. For example, uncompleted projects can also be caused by incompetent firms winning the bid, rather than lack of monitoring. We do not know how much of the problem would be solved by this intervention.

54. ^{^}

     For example, OCP recently published a case study (Dilotsotlhe, 2023) about civil society contract monitoring in Uganda, though this was not Lift-supported.

55. ^{^}

     The current program incubates projects in Brazil, Chile, Lithuania, Philippines, Senegal, Thailand, Uganda, UK, USA.