Summary of findings
1. Most governments made progress on publishing details of their emergency procurement, meeting the broad spirit of the IMF commitments.
With varying degrees of quality, granularity and regularity, these governments have published information on their pandemic-related contracting included as structured, machine-readable open contracting data:
- Afghanistan, where information is available through the new AETS portal with a COVID-19-specific tag; 
- Bolivia, where the procurement authority does not publish any open data, but does publish basic contracting information which can be searched manually on the SICOES website;
- Ecuador, where a specific open data portal for emergency contracting, Datos Abiertos, was created at the beginning of the pandemic. Data is available in the Open Data Contracting Standard(OCDS): a crucial goal for governments seeking to allow in-depth monitoring of their procurement;
- Honduras, where contracting data is published in a number of different locations, including an Institute for Access to Public Information portal, Honduras’ central Open Contracting Portal;
- Malawi, where an Excel file containing COVID-19 contract awards is available for download;
- Moldova, where contracting information is published in line with OCDS on M.tender and healthcare procurement information is published on tender.health, a joint government-civil society initiative;
- Nigeria, where the Bureau of Public Procurement have a dedicated COVID-19 contracts web page;
- the Kyrgyz Republic, where almost all government contracts are available through the Zakupi portal in OCDS format, though there is no COVID-19-specific tag;
- South Africa, where government have created a dashboard displaying information on coronavirus-related contracts;
- Uganda, where government-wide contract awards are available on the Government Procurement Portal, though there is no COVID-19-specific tag for emergency contracts.
Broadly, all these countries can be said to have met the spirit of the IMF commitments on emergency procurement. However we could not find the quarterly reports including information on pandemic-related procurement that the Nepali government had committed to publish.
Some procurement departments also circulated new regulations to emphasise procuring entities’ transparency responsibilities at the start of the pandemic. Based on our interviews, these did not seem to be driven by the IMF requirements but for domestic accountability reasons.
For example, in Malawi, the Public Procurement & Disposal of Assets Authority mandated that government entities publish contract award details on the PPDA website within five days of the contract being signed.  In Nigeria we saw that the Bureau for Public Procurement also requests that entities pass on details of COVID-19 contracts within five days of contract award.  Both these requirements seem to set out more specific and stringent publication requirements than stated in the countries’ procurement laws. 
2. However, many of our research partners said that the publication of emergency contracts was not a direct consequence of the IMF commitments. It often followed from sustained advocacy by local and international civil society, existing technical support to civil servants (including from OCP) and/or where a culture of procurement data disclosure existed before the pandemic.
Our research partners cited factors other than the IMF commitments as stronger influences on procurement agencies and finance ministries working quickly to disclose pandemic-related contracts.
In some cases this was because of clear direction-setting from political leadership and on-going reforms to improve procurement transparency. For instance, in Afghanistan, a presidential instruction called for cross-government transparency. There was also an active memorandum of understanding with Open Contracting Partnership and Integrity Watch Afghanistan to publish open data on procurement.10 Here, we also heard that the IMF commitment had not been extensively communicated throughout the Afghan government, yet the National Procurement Authority still created one of the more progressive pandemic-specific portals considered in this report based on its existing collaborations in this area.
Similarly, in Nigeria, we heard that the Bureau of Public Procurement, the implementing body, was not consulted on the IMF commitment. Their decision to create a specific page for pandemic-related contracts was a consequence of discussions with civil society, while the circular mandating transparency during the emergency was strongly influenced by the World Bank. In both cases, it appears that commitment holders (finance ministries) did not follow up with the agencies best suited to implement the commitments.
In Ecuador, progress on publishing emergency procurement information can be attributed to the procurement authority’s (SERCOP) commitment to transparency, rather than to the IMF commitments. A decision was taken early on in the pandemic by the head of SERCOP to commit to publishing all contracts in an open format.  The result of this was the publication of COVID-19 contracts in the Open Contracting Data Standard (OCDS) on the Datos Abiertos portal.  However, the creation of Datos Abiertos was well underway when the commitments were signed. SERCOP was an active participant in OCP’s Lift impact accelerator program where a lot of the technical work was done, pivoting to the pandemic as it struck the country. 
The fact that some countries committed to already-existing plans is not necessarily something to be concerned about. By adding an IMF commitment to existing plans, a government is providing itself with an additional test to which it will be held accountable. Whether or not a commitment was implemented is more important than whether it was a new policy or not. But where governments are already strong performers on open contracting, the IMF may be missing a chance to set out more ambitious pathways towards deepening and sustaining existing reforms.
We also saw that in some cases civil society organisations (CSOs) were not aware of the commitments. In Afghanistan, for example, a CSO which actively monitors government expenditure did not know that the Ministry of Finance had actually met their IMF commitment. This was because the report disclosing data was not linked on any of the usual data disclosure portals, including those on the Finance Ministry’s website.
It is important to emphasise here how helpful civil society can be when it comes to ensuring a government meets transparency goals. Across most of the countries in which we conducted research, we found that the governments who did best on procurement transparency were those who had strong track records of contracting data publication before the pandemic. In these places (Nigeria, Afghanistan, the Kyrgyz Republic, Ecuador and Moldova, for example) existing cultures of transparency are strongly supported by collaborative relationships between procurement agencies and CSOs.
In Nigeria and Afghanistan, interviewees told us that discussions with and encouragement from CSOs were among the most important factors in their decision to publish information on pandemic-related contracts. Moldova also provides an example of how civil society has been able to push for transparency during the COVID-19 pandemic: here, medical procurement information was made available on a new e-procurement system after pressure from civil society. 
As we recommend below, the IMF should encourage governments to embrace civil society involvement to help them meet their commitments.
3. The existence of the IMF commitments provided government officials with additional incentives and leverage to improve emergency procurement transparency.
While the IMF commitments may not have been the main drivers of change in all cases, they still had an influence on emergency procurement practices and policies during the pandemic.
In Afghanistan, we heard that procurement agencies began publishing emergency contracts without knowing that they were meeting IMF commitments as a consequence; a similar situation arose in Nigeria. However, interviewees still said that the IMF contributed to the imperative to publish, in a broader sense. At the highest levels of government, we heard that having a clearly written commitment to an international donor helps to drive the political will.
Similarly, in Ecuador, the publication of COVID-19 contracts was not directly influenced by the IMF commitments process. SERCOP had already worked to create the Datos Abiertos portal before the commitments were made. Nonetheless, one research partner did emphasise that the IMF commitments ‘gave further political force’ to existing reforms. 
4. The IMF commitments process was more influential in the sphere of beneficial ownership than open contracting: in four countries, it led to improvements on beneficial ownership, demonstrating the agenda-setting power of the IMF commitments in new areas.
The IMF largely smoothed the path for existing policies and reforms on open contracting, but in some cases it had a starker impact on policies and reforms related to beneficial ownership. The IMF’s impact here was due to the fact that they explicitly made beneficial ownership a priority in their negotiations with governments, though the contexts differed in each country. This demonstrates the agenda- and priority-setting power of IMF commitments in policy areas that are newer or under-developed for both government and civil society.
In the Kyrgyz Republic, we heard that the IMF treated beneficial ownership disclosure as a very high priority during discussions on the emergency commitments. As a result of this, an amendment to the Kyrgyz procurement law was made in December 2020. This mandated that buyers publish information on suppliers’ beneficial owners on the public procurement portal.  The new law is now being implemented.
In Ecuador, the commitment to publish beneficial ownership data on the SERCOP website has not yet been met. However, the IMF still appears to have had some influence here, as SERCOP issued a new resolution in September 2020 that allows the procurement agency to collect beneficial ownership information directly from suppliers at the offer stage.  Stakeholders emphasised that whilst the procurement agency had already been planning these reforms, the IMF commitment helped to push forward the change.  One research partner also emphasised that the IMF commitments helped to put beneficial ownership ‘on the radar’ of civil society in Ecuador. 
In Afghanistan, we found that the National Procurement Authority issued a circular in September 2020 mandating that bidders submit information about their beneficial owners.  This met a Prior Action required for disbursement of funds under Afghanistan’s application to the IMF under the Extended Credit Facility.  Interviewees suggested that the IMF saw this as an urgent issue: they were concerned that a planned regulation on beneficial ownership within the Ministry of Industry and Commerce was taking too long to implement, and that it was a priority to get some information published as soon as possible. 
Finally, in Moldova, interviewees highlighted new legislation mandating companies to declare beneficial ownership data to the procurement authorities, directly attributing this progress to the IMF commitments. 
However, there are still issues with the quality of beneficial ownership data in Moldova, which are explored further below.
5. That said, where countries published data on the beneficial owners of emergency suppliers, this was low in quality and quantity.
Despite these examples of progress on beneficial ownership, few governments in the scope of this report have met the wording of their emergency commitments to publish data on the beneficial ownership of suppliers in a sense that clearly meets an internationally-accepted definition of the term.  Where there is a beneficial ownership data field, the quality and detail of the data disclosed is often limited.
In Afghanistan we found that while the requirement to publish beneficial ownership information is very new and adjustments will take time, buying agencies currently upload PDF scans of paper forms. None of those we have examined use the downloadable form the government offers, which sets out a solid definition of beneficial owner.  Similarly, in Malawi and South Africa, data on COVID-19 contract contains a field for a ‘beneficial owner’ but it appears that only the names of single shareholders or company owners/directors are captured here. 
6. The commitments themselves should have been more specific and ambitious: governments are capable of more than what the wording captured, including. Emphasising, for example, rapid publication and standardised open data would have helped the pandemic response.
The language of many governments’ commitments is broad and does not hold them to particularly ambitious programmes of reform. This is especially the case for procurement transparency commitments. We are, of course, aware that commitments had to be negotiated very rapidly, and we acknowledge that governments would have followed guidelines set out by IMF teams surrounding the wording of commitments. Our comments here are made in the spirit of going even further in future to encourage more comprehensive reforms and improvements.
Bolivia’s commitments serve as a good illustration of this trend. The government committed to adhering ‘to best practices in procuring and awarding contracts related to the pandemic, including by publishing regularly documentation on procurement contracts on the government’s website.’  The lack of specificity in this commitment meant that the government could continue to publish contract information on SICOES as before – without publishing downloadable open data which is of use to civil society – and technically still meet its obligations to the IMF.
Afghanistan, too, only commits to ‘publishing quarterly reports on pandemic-related spending, including procurement contracts and the beneficial ownership of companies awarded those contracts.’  A more ambitious emergency commitment should have emphasised AGEOPS as the preferred portal for publishing COVID-19-related contracts data and committed to disclosing data in OCDS. It could also have encouraged outreach to CSOs and journalists to highlight the commitments and to set up spaces for feedback on data quality and utility. 
Most countries in this research broadly followed this trend. As outlined in our recommendations below, introducing standard language that emphasises, for example, data quality and usability, or civil society consultation, would allow governments to set a higher bar for implementation, and give procurement authorities more clarity surrounding the changes they are expected to make to meet the commitments.
It was also not always clear what particular terms referred to in a commitment’s wording. A case in point is ‘ex-post validation,’ a phrase used in many commitments. South Africa and Uganda, for example, stated that they would conduct audits of COVID-19 spending, which would include ‘ex-post validation of delivery.’  The Kyrgyz Republic, however, stated that they would ‘publish ex-post validation of delivery along with the name of awarded companies and their beneficial owner(s) for all public procurement contracts.’  From our interviews it seemed that the ‘ex-post’ element was understood here as referring to releasing data on the implementation phase of the contracting process. A glossary of terms and standard language would therefore be helpful for all parties. 
With this said, we also found that buyers or procurement authorities may not have sufficient capacity to respond to additional transparency requirements. We are aware that publishing additional information is not straightforward and numerous interviewees reported concerns around their capacity to implement commitments.
For example, we believe that a lot of COVID-19 contracting is happening at the provincial level in Afghanistan, where authorities have very little experience proactively disclosing data. Moreover, across the country, there are problems with expensive internet and electricity shortages. In the Kyrgyz Republic, the Public Procurement Department have had their mandate and responsibilities increased as a result of the government implementing the IMF commitment, but have no additional budget with which to hire new staff or web developers. In Uganda, one research partner said limited resources, including the lack of an e-government procurement platform, have left the government unable to fully implement their commitments. 
Taking into account all of these reports, there is clearly a widespread need for governments to allocate additional resources to support improved data publication, supported with IMF funds if possible. We therefore recommend that a commitment regarding allocating resources to relevant government departments is included as standard in any commitments on open contracting and beneficial ownership.
 See PPDA circular: www.ppda.mw/wp-content/uploads/2020/04/PPDA-CIRCULAR-APRIL-2020-COVID-190001.pdf, p.2.
 See BPP guidelines: www.bpp.gov.ng/wp-content/uploads/2020/05/BPP-Guideline-on-COVID-19-Procurements.pdf, p.4.
 Interview with research partners in Nigeria, February 2021. See: https://goprs.unodc.org/documents/Public_Procurement_Act_2007.pdf and www.ppda.mw/wp-content/uploads/2019/11/public_procurement_and_asset_disposal_act.pdf.
 Interviews with research partners in Afghanistan, January-February 2021.
 Datos Abiertos portal: https://datosabiertos.compraspublicas.gob.ec/.
 See here for a summary of SERCOP’s work with the OCP Lift programme: www.open-contracting.org/2020/12/20/how-covid-19-and-collective-intelligence-transformed-procurement-risks-into-opportunities-for-transparency-in-ecuador/.
 Medical procurement data in Moldova is now available on tender.health.
 Interview with research partners in Ecuador, February 2021.
 http://cbd.minjust.gov.kg/act/view/ru-ru/111125?cl=ru-ru, Article 51.
 See SERCOP’s September 2020 resolution here: https://portal.compraspublicas.gob.ec/sercop/wp-content/uploads/2020/09/RE_SERCOP_2020_0110.pdf.
 Interview with research partners in Ecuador, March 2021.
 See https://twitter.com/npa_aop/status/1311240352469315584?s=20. The circular is available in Dari here: https://ageops.net/da/documents/circulars/71 and in English here: https://cms.npa.gov.af/files/Documents/NPACirculars/1602054686-73.pdf.
 We understand that this regulation has now been approved by the Cabinet and is in the process of being implemented.
 3 Interview with research partners in Moldova, January and February 2021. For Moldova’s updated beneficial ownership legislation see this decree form July 2020: www.legis.md/cautare/getResults?doc_id=122147&lang=ro. See also this decree from November 2020 which introduces a definition of beneficial ownership on the Single European Procurement Document: www.legis.md/cautare/getResults?doc_id=124193&lang=ro.
 See, for example: www.openownership.org/uploads/definitions-briefing.pdf.
 This form can be found by downloading the Excel file here: https://ageops.net/da/documents/circulars/71.
 See www.ppda.mw/covid-19-reports/and http://ocpo.treasury.gov.za/COVID19/Pages/Reporting-Dashboard.aspx; interview with research partners in South Africa, February 2021; similar findings were discovered in Moldova and Bolivia.
 See Bolivia’s Letter of Intent to the IMF here: www.imf.org/-/media/Files/Publications/CR/2020/English/1BOLEA2020001.ashx
 See Afghanistan’s Letter of Intent to the IMF here: www.imf.org/en/Publications/CR/Issues/2020/04/30/Islamic-Republic-of-Afghanistan-Request-for-Disbursement-Under-the-Rapid-Credit-Facility-49386.
 See: https://tenders.ageops.net/en/procurement-process for the AGEOPs portal.
 See Uganda’s Letter of Intent to the IMF here: www.imf.org/en/Publications/CR/Issues/2020/05/14/Uganda-Request-for-Disbursement-under-the-Rapid-Credit-Facility-Press-Release-Staff-Report-49427. See South Africa’s Letter of Intent to the IMF here: www.imf.org/en/Publications/CR/Issues/2020/07/28/South-Africa-Request-for-PurchaseUnder-the-Rapid-Financing-Instrument-Press-Release-Staff-49612.
 See Kyrgyz Republic’s Letter of Intent to the IMF here: www.imf.org/en/Publications/CR/Issues/2020/05/12/Kyrgyz-Republic-Request-for-Purchase-Under-the-Rapid-Financing-Instrument-and-Disbursement-49408.
 For an example glossary, see: https://standard.open-contracting.org/latest/en/getting_started/contracting_process/.
 Interviews with research partners, February 2021.