Measuring the economic impact of beneficial ownership transparency (full report)

  • Publication date: 12 May 2022
  • Authors: Oxford Insights, Lateral Economics, Open Ownership

Annex I: Literature review

Measuring the tangible economic impact of beneficial ownership transparency: a nascent field

To date, most arguments in favour of BOT have focussed on non-monetised benefits, emphasising instead the social value arguments in favour of increasing company transparency in order to prevent financial crime through case studies and impact stories.

Historically, a leading argument in favour of beneficial ownership transparency has always been that as a reform designed to reduce the use of anonymous shell companies to launder the proceeds of crime and corruption, BOT is a reform to pursue in order to reduce financial crime. In 2003, the Financial Action Task Force (FATF) was the first body to set forth global standards on BOT, with the aim of helping to “get rid of the cloak of secrecy concerning the ultimate owner of a company, foundation, association or any other legal person, and prevent their misuse for crime and terrorism.” [164]

In the years following the original FATF publication, organisations such as Open Ownership, Global Witness, Transparency International and the Open Government Partnership have continued to make the case for BOT as an important tool in the anti-corruption space without attempting to quantify the economic impact of reform. Instead, almost all of the BOT advocacy materials we encountered used case-study examples to demonstrate the positive impact of BOT registers where they have been implemented. Examples of this include Open Ownership’s impact stories, [165] and Transparency International’s work to identify how beneficial ownership data had been used to uncover cases of corruption or financial mismanagement by officials in the Czech Republic, Turkmenistan, Brazil and Denmark (amongst other jurisdictions). [166]

In addition to making the case for beneficial ownership as being in the public interest, the literature we encountered also identifies a number of more readily monetisable and quantifiable benefits of BOT.

In addition to arguments which emphasise social value, research and advocacy materials on the subject of BOT have also identified a number of quantifiable and monetisable benefits associated with BOT (as laid out in this report’s logic model).

Indeed, it is worth acknowledging that benefits associated with a reform are more often than not measurable and monetisable in principle. Crucially however, there are various degrees to which these calculations can be readily made. [167] In the UK, HM Treasury provides four definitions of benefit types which can arise from government reforms, which can be used as a cursory tool to map out the BOT benefits encountered across literature and in our logic model:

Benefit type Examples of BOT benefit identified in literature
Cash releasing benefit (CRB) – a monetisable benefit which is ‘cashable’ for a particular stakeholder group, releasing additional funding for a government, business, individual etc. Reduced due diligence costs for businesses

Reduced AML compliance costs for banks

Increased government revenue from asset seizures (likely to be limited in size – but still a CRB)
Non cash releasing benefit (Non CRB) – a monetisable but not cash-releasing benefit Further law enforcement efficiency when investigating cases of potential financial mismanagement (reduction in police time and resource)

Less public money wasted on unfit providers (better value for money in government contracts)

Reduced risk of losses for shareholders involved in ‘pump and dump’ schemes

Increasing investor protection by reducing the risk of misallocated funds, and of investors breaking bribery laws

Reduced tax evasion and tax avoidance
Quantifiable benefit – a quantifiable but not monetisable benefit Reduced perception of corruption (quantifiable via indexes such as the Corruption Perception Index*)

Increased prosecutions or convictions in cases related to corruption / financial mismanagement (reduced impunity)

Increasing international economic stability by making it easier to gauge risk in financial transactions

Increased data use by civil society and government officials
Qualitative benefit – a benefit which can be qualitatively identified, but not readily quantified More trust in public institutions and business environment

Increased insight into company ownership structures for civil society, government and businesses (normative value of transparency)

*Corruption Perception Index

Table adapted from HM Treasury’s Guide to Developing the Programme Business Case. [168]

Whilst literature identifies a number of benefits that are in principle quantifiable or monetisable, there has been little done to measure any of these benefits to date.

Generally, the sources we consulted which describe the economic benefits that one could logically expect to arise from BOT, do not look to quantify these benefits. This echoes findings from a 2021 European Commission report which looks to assess the impact of high value datasets, (including company ownership data), which states that whilst experts consulted were able to speak to case study examples demonstrating company data value, they could not refer to robust figures or means of quantification. [169]

Moreover, a number of studies explicitly refrain from attempting to conduct an economic analysis to either forecast, or assess the quantitative impact of BOT. We found two recurring explanations for reticence here:

1. Some studies refrain from quantitative impact analysis on the rationale that it is too early to assess impact.

A number of studies have taken the stance that beneficial ownership as a nascent policy field is too difficult to measure due to a lack of empirical data. For instance, the 2019 Adam Smith International report states:

“it is largely too early in the implementation of BOT to measure its impact on most of these hoped-for benefits on a systematic comprehensive basis” [170]

Similarly, a 2020 Transparency International helpdesk response paper outlined how the supporting evidence for BOT reform remained “anecdotal”, referring to “a dearth of empirical evidence” and the current impossibility of identifying the causal effects of implementation. [171]

Meanwhile, even within the UK government, which is widely regarded as a forerunner in terms of BOT reform, some implementers have expressed doubt about being able to quantifiably measure impact due to a lack of policy maturity. For instance, the BEIS Post-Implementation Review of the UK PSC register from 2019 refrains from any quantitative impact measurement, reasoning that it is “too early to evaluate its wider economic effects”. [172]

2. Other studies refrain from quantitative impact analysis on the rationale that it is too difficult to attribute potential economic benefits directly to beneficial ownership transparency.

A further challenge which surfaced throughout research is that isolating beneficial ownership transparency (and the design choices it invokes) as the sole drivers of economic benefit is extremely difficult, especially without baseline data. As one of our interviewees for this project pointed out, “beneficial ownership is just one piece of the puzzle”, and is always part of a broader package of transparency reforms. The same view is reflected in the Adam Smith report which states that “BOT is only one element of systems to address corruption or money-laundering, for example, which makes attribution more challenging”. [173]

The UK government also acknowledged the challenge of attribution. A 2014 impact assessment of its ‘Transparency and Trust’ programme of reforms, which included an assessment of the costs and benefits of creating the UK’s centralised PSC register, makes no attempt to quantify the economic benefits of the register. Attribution emerges as a key obstacle when discussing BOT’s impact upon law enforcement resourcing costs:

“We should also note that beneficial ownership reform is only one part of the Transparency and Trust package. While it is difficult to predict reliably change in the crime rate related to any one part of the package, we could consider that the overall combined effect from implementing the comprehensive package is likely to be greater than the sum of its parts.” [174]

Given the challenges associated with attribution and data availability, benefits that are technically monetisable or quantifiable have often been captured only qualitatively.

For instance, the authors of the 2019 Post Implementation Review of the PSC register made a decision to use qualitative interviews to discuss the value of UK PSC data with law enforcement, civil society organisations and financial and business institutions. This was attributed to a lack of data maturity following the registers release in 2016 and the staggered introduction of reporting measures which followed. [175] Based on these interviews, the resulting report was able to speculate qualitatively about the presence of benefits such as reduced due diligence costs, but wasn’t able to monetise this as a cash releasing benefit (as it is in principle). The conclusions made by the Post Implementation Review indicated that 4 out of the 5 objectives of the PSC regulations were being met. They were deemed successful in the areas of increasing company transparency, increasing confidence in the business environment, facilitating economic growth and facilitating investigations into economic crime. The only objective for which no evidence was found, was reducing illicit activity and improving corporate behaviour. [176]

Similarly, the 2014 BEIS impact assessment of the ‘Transparency and Trust package’ refrains from monetizing reductions in law enforcement resource time and costs (in principle, a non-cash releasing, but montetisable benefit):

“there is no reliable or systematic way of attributing reductions in law enforcement agencies’ costs or the consequences directly and exclusively to enhanced transparency of company beneficial ownership. For this reason, the benefits resulting from reduced costs to law enforcement remain non-monetised.” [177]

Despite the barriers to measurement, there remains a widespread assumption that these economic benefits of BOT are significant, and it has been implied that they outweigh the costs associated with BOT reform and the costs associated with inaction in this area.

As characterised by a report commissioned by the European Union into high value datasets, which includes a discussion of the value of BOT data, “beneficial ownership datasets are unanimously considered of high value by the literature”. [178]

Whilst there were a number of sources consulted which mentioned the costs of beneficial ownership transparency (including set up costs and maintenance for government, and compliance costs for business), it was consistently implied that these costs would be eclipsed by the wider economic benefits of beneficial ownership transparency, even if the latter could not be readily quantified.

For instance, the Adam Smith International 2019 report states that centralised beneficial ownership registries do incur some costs for businesses, including submission fees and time taken for business to report, but these are “outweighed by the benefits that business can derive from greater transparency and from the risks of resistance to BOT”. The report argues that smaller costs to businesses pale in comparison to risk costs associated with the lack of a register – such as the reputational risk of a business being involved in a corruption scandal. [179]

Elsewhere, the UK Government's 2014 Impact Assessment for a public centralised beneficial ownership registry sizes costs for government and businesses over 10 years at around £1088 million total – but frames this alongside the cost of organised crime and fraud in the UK, which is estimated at £24 billion annually according to Home Office. [180] [181]

We came across no literature during our research which explicitly made the case against beneficial ownership registries based on cost-concerns.

Existing work to track economic impact of beneficial ownership transparency

Where literature does attempt to put figures on the benefits associated with BOT reform, these are often high-level and lack methodological rationale.

Such figures can be found cited in both government materials and in advocacy settings, and are more often than not linked to the costs associated with broad areas of criminal activity, such as corruption or money laundering.

Sources consulted generally fell short of outlining exactly how, and to what extent BOT would minimise these costs: for example, the aforementioned 2014 UK government impact assessment references the £24 billion cost of organised crime estimate, but offers no insight into how a centralised BOT registry will affect the figure. [182]

Moreover, these numbers are rarely accompanied by methodological rationale, and sometimes their provenance is unclear. As an illustrative example of all of the above, a 2011 World Bank report titled The Puppet Masters, which calls for a more effective approach to beneficial ownership transparency, begins with the assertion that corruption siphons around $40 billion annually away from legitimate economies. [183] There is no indication of how this figure was calculated, or by whom, but it also been cited by the United Nations, [184] news outlets [185] and academic papers. [186] A number of figures have been similarly ‘recycled’ without any easily identifiable methodological grounding.

Nonetheless, there are a very limited number of studies that have sought to quantify and even monetise the benefits of beneficial ownership transparency, often with clear limitations.

The first attempt to monetise the benefits of BOT we encountered dates back to 2002, when the UK government published a Regulatory Impact Analysis (RIA) which seeks to place value on the benefits of BOT implementation options for law enforcement purposes. This includes the creation of a “modern database” of BOT information searchable by name (early ideation of the UK PSC register). The RIA summarises its methodology as follows:

The RIA estimated the benefits of different levels of BOT, varying by timeliness and access to data and the linkage of databases on company ownership with other databases. After consulting with agencies involved in pursuing and recovering the proceeds of financial crimes, HM Treasury/DTI concluded there were three different levels of benefits of BOT, corresponding to the different options, particularly 5%, 15%, and 30% orders of economy. These percentages were then applied to estimates of costs involved in pursuing and preventing financial crimes, as well as the value of recoveries from financial crimes, to generate economic benefit estimates. [187]

To illustrate, the RIA cited an average expenditure per head for a law enforced specialised in recovery of £30k. Taking reported estimates of 1,700 financial investigators in UK law enforcement, multiplying by £30k, and applying 5% yields the estimated £2.55 million savings for this element in the first order of economy scenario (i.e. 5%). [188] Similar calculations were undertaken for the other benefit elements – i.e. reduced prevention efforts, greater recoveries, and reduced business risk – for first (5%), second (15%), and third (30%) order of economy scenarios.

The RIA estimates appear reasonable, with cost and recovery estimates based on official data, and assumed orders of economy based on consultation with agencies, in which they were asked to give a range (e.g. 5-10%) of assumed economies in their operations or improvements in fraud recovery. In all cases, HM Treasury/DTI followed the lowest estimate provided.

The RIA is candid about the margins for error associated with this approach, noting that quantifying the impact of financial crime is “notoriously difficult”, and therefore acknowledging that it does not look to make precise benefit calculations for each option put forth. Perhaps due to the choice of following lowest estimates, the figures generated were regarded as a conservative underestimation by the Open Government Partnership. [189]

In 2015 the UK government published a further impact assessment which covers limited economic benefits of BOT reform, with a specific focus on costs and benefits for businesses responding to direct requests from the public for PSC information. [190] The report suggests that costs for businesses to familiarise themselves with guidance, and to respond to these requests are likely to amount to £137.1 million over 10 years. Based on a prescribed flat fee of £12 which businesses would charge to process PSC requests, and the average number of requests businesses would receive, benefits to businesses are estimated to be £47.3 million over 10 years. Crucially, the benefits calculated here are limited to the ‘cashable’ benefits for businesses associated with a fee-based approach; the report does not attempt to quantify the wider benefits of being able to access this information for requesters.

Outside of the UK context, we only found two examples of work which looks to quantify the monetisable economic benefits of BOT. The first, a study conducted by PwC, focused on BOT within the context of the wider value of the business information sector in Italy, and identified a constant growth in the last three years (from a value of 57 million euro in 2017 to 58 million euro in 2018 and 60 million euro in 2019). However, the report is not publicly available, only referenced in the European Commission report on high value datasets, which provides no insight into its methodological approach or the specific types of business information considered in the analysis.

Finally, the European Commission impact assessment itself conducts a macro-economic impact assessment which looks to size the value of company and company ownership data. To achieve this, the report looks back to a study conducted by Graham Vickery in 2011 [191] to estimate the market size of public sector information (PSI) in its broadest sense, and applies forecasts from the European data market monitoring tool [192] to predict a baseline scenario for growth up to 2025. From here, company and company ownership data is estimated to represent 6% of total PSI market size, giving it a representative value of 3016 million euros in 2020 across all EU member states, growing to 4132 million euros by 2025 (see Table 1 annex). [193]

Crucially, however, these estimates are limited by the fact that the report doesn’t extrapolate on how a 6% market share value for company and company ownership data was identified, only that it was determined based on “existing literature” and the study’s research – which raises some questions as to the accuracy of these large macro-economic figures. Moreover, the report refers to “company and company ownership data” throughout. Whilst beneficial ownership data is referenced during the discussion, no specific estimations are attributed to BOT data in isolation.

Of all the studies we encountered, the joint Companies House and BEIS 2019 valuation of Companies House [194] data is the best example of work which measures the monetisable benefits of beneficial ownership data specifically.

Indeed, of all the existing research reviewed, the Companies House/BEIS report was one of the most relevant to this study, for three reasons:

  1. It is the only study we came across which has attempted to isolate the benefits of beneficial ownership data transparency using a specific BOT intervention (the introduction of the UK PSC register) as one of the focuses of analysis. As emphasised in the EU/Deloitte report, the Companies House valuation is the first of its kind in this sense – as other countries like France and Denmark, which have also moved forward with the creation of PSC registers, have not conducted similar analyses. [195]
  2. The report also considers the economic implications of making company data available under a paid-access model only – and hence comes closest to measuring the impact of making a BOT register freely accessible to the public – one of the key design choices underpinning BOT interventions at the start of this project.
  3. The valuation is accompanied by a thorough methodological discussion of potential approaches for valuing the benefits of company data and their limitations and advantages, which is likely to be instrumental to informing some of the avenues explored in this paper.

The key findings of the report concerning the value of beneficial ownership data are as follows:

  • Based on willingness to pay survey responses, beneficial ownership data was estimated to account for 4% of the total value of all Companies House data – or approximately £40 million to £120 million of aggregate benefit per year.
  • Yet for “high use” users – who are characterised as “public good” users working in transparency organisations, government departments or law enforcement entities – the report acknowledges that the value of PSC data rises to 13% of the total value.
  • An illustrative policy application shows that introducing subscription fees to the PSC register would generate a net welfare loss. This is because of transactions that would no longer take place because of the fees, and represents what economists would label as lost consumer surplus. [196]

These findings were calculated through a willingness to pay (WTP) survey-based approach and a discrete choice model, whereby the authors broke down Companies House datasets and services in order to isolate their specific elements, such as PSC data. Respondents were then asked how much they would be willing to pay to use or access each element. Furthermore, for the intermediaries and ‘public good’ providers who were the most frequent and intensive users of PSC data, WTP surveys were supplemented with additional qualitative research exploring their use of data, its associated costs and revenues and the potential availability of alternative data sources. This was to ensure the robustness of findings, given that there were a limited number of respondents from the ‘public good’ user groups.

In addition to quantifying the benefits of PSC data, the report also makes a solid economic case for the open access provision of company data. An estimated demand curve is used to highlight how introducing an annual subscription charge for Companies House data would severely reduce numbers of users and transactions, leading to a net economic loss to the service despite a hypothetical new revenue stream. This was achieved by estimating demand based on the results from the WTP survey to associate a new number of users for each annual price range.

Although this element of the report refers to Companies House data more broadly, it nonetheless provides some methodological grounding for pursuing a similar approach for measuring the varying economic impact of making BOT data available under paid / free models – an important intervention design choice identified at the start of this work as a potential avenue for measurement.

The report’s wider exploration of potential approaches to benefit measurement also provides valuable insight for this research.

In addition to its findings, the report also includes a discussion paper outlining methodological options for measuring the economic value of company data, outlining the different methods available to evaluators, and the kinds of data points which would be required to execute them:

Method Data required Category
Point expansion
Observing a single point of the demand curve and applying pre-determined elasticity assumption to estimate the demand curve for a good or service. Simplified approach but pragmatic in situations where demand information is very limited.
Market data, empirical data Market value-based analysis
Residual imputation method
Estimating the economic rent or marginal
revenue of a non-priced input to production provides an output of known value. The value of the non-priced input is calculated by subtracting all other costs of production from the value of the output that is produced; hence the ‘residual’ value is assigned to the non-priced input.
Market data, production data Market value-based analysis
Direct estimation
Using econometrics to directly estimate the demand curve for a good or service when data is available. This approach is the most accurate, but is reliant on the accessibility of data.
Market data, empirical data Market value-based analysis
Production function
Quantifying relationship between an input to production and the output. Marginal product of data is calculated by holding all other inputs constant and looking at the change in output when only the measured input changes.
Market data, production data Production function
Inference from market values
Using comparable income streams or sales values to estimate value for non-priced assets. The marginal value associated with non-priced uses of CH data could be assessed by assuming that the price associated with ‘paid for’ uses of the data also represents the WTP of open access users.
Market data Inferring the value of non-priced uses from market data
Revealed preference
Observing relationship between demand for market-price goods and preferences for related non-priced goods. Analysing choices made by consumers and consequently attributing value to individual features in terms of their willingness-to-pay.
Market data, attribute data and survey data Non-market valuation methods
Stated preference
Simulating markets for trading non-priced goods through survey-based approaches. Discrete choice experiments are an attribute-based approach that break a good or service down into its characteristics or features and provide a value for changes in the provision of these attributes.
Attribute data, survey data Non-market valuation methods

Table and method summaries adapted from Valuation of Companies House Data, Report 1: Methodological Framework [197]

After weighing up the trade-offs associated with each approach, the BEIS/Companies House report is also candid about the limitations of its chosen methodology. This, in turn, raises important considerations about how the results of different methodological approaches to economic measurement can be interpreted.

Following a discussion of the advantages and trade-offs associated with each approach, the authors of the report chose to pursue a stated preference approach, using a discrete choice experiment (DCE) to estimate user willingness-to-pay (WTP). It is acknowledged that stated preference / WTP is not necessarily a first choice approach from a valuation perspective, since Companies House data is largely unpriced.

However, this decision was made for a number of reasons which are outlined in the methodological framework:

  • a stated preference approach can be applied across various data types and user groups, ensuring consistency of perspective across the valuation study;
  • non-market valuation methods are not necessarily reliant on third party data, and all data requirements could be met through survey design and Companies House product data, and;
  • a stated preference approach allowed the authors to design a survey which incorporates in-built consistency checks for the WTP estimates collected.

Whilst these concerns, particularly surrounding data availability, justify the use of a WTP approach, the discussion nonetheless serves as an important illustration of how certain methodological approaches might impact the way we can frame and understand its estimates.

Ultimately, the value identified in the Companies House / BEIS report will never be realised or released back into government budgets – it produces a monetised but non-cashable benefit estimation. This could, in theory, be a potential trade off associated with other methodological approaches explored in this study; for instance, if stakeholders were more interested in measuring cash releasing benefits, to assess the amount of budget freed up by an intervention, then alternative approaches should be pursued.

Looking to other policy areas for insight: measuring the economic impact of corruption, fiscal transparency and open contracting

The BEIS valuation of Companies House data serves as a particularly useful starting point when looking at potential methods for valuing beneficial ownership data. However, the fact that the study is the first of its kind, and that PSC data only constitutes a fraction of the report still means that the field of literature surrounding this piece of work is limited.

Given the lack of literature which actually quantifies the economic impact of BOT, we have looked to three other jurisdictions for insight into potential approaches for measurement, to explore how economic impact has been quantified in literature on anti-corruption, fiscal transparency and open contracting.

Our findings from each of the jurisdiction have been summarised below.

1. Measuring the economic impact of corruption

In contrast to beneficial ownership transparency, there is a wealth of literature seeking to quantify impacts of corruption. Given the logical connections between corruption and BOT, some of this work could reasonably be applied to the methodologies explored in our research.

There is a clear link between corruption and certain company structures. [198] As outlined in this study’s logic model, beneficial ownership transparency provides law enforcement, journalists, and NGOs with the tools and data to explore these structures and identify cases of potential corruption and other illicit financial flows, whilst acting as a deterrent for the same. Consequently, as the 2019 Adam Smith International report neatly summarises, “the potential impact of effective BOT can be viewed as a potential reduction in the scale of corruption facilitated by the opacity of company structures”. [199]

There is a wide body of literature looking to quantify, and often monetise, the impact of corruption on a range of economic indicators – many of which are also present in our logic model. Cross-country panel studies draw upon corruption indices like the ICRG (International Country Risk Guide) risk of corruption measure and the FFC (Freedom from Corruption) index, combined with longitudinal techniques, [200] random effect models [201] and data envelopment analyses [202] to put an economic value on the economic effects of corruption.

There is some underexplored potential for the methodologies in this research to draw upon some of these methods to quantify the relationship between BOT and economic impact, if it is assumed that a reduction in corruption would act as an intermediary logical step here.

Despite more progress in the field of quantifying the effect of corruption, analyses remain subject to critiques which could also be applied to work seeking to measure the economic impacts of BOT.

A 2015 UK Department for International Development (DfID) report exploring the causes and effects of corruption identifies a number of difficulties in terms of measuring the economic impacts of corruption. These include: a lack of clear definitions to set the parameters for measurement, aggregation problems when attempting to demonstrate direct causality and, probably the most significant in the context of the report, “the limited range of methodologies that can adequately demonstrate causal relationships between corruption and [economic] growth”. [203]

Elsewhere, economists at the World Bank have drawn attention to issues concerning data availability, quality and coverage as key challenges for econometric analyses of the impacts of corruption. Echoing the compromise taken the Companies House / BEIS report, the World Bank economists here, who were looking to estimate the effects of corruption in Bangladesh, chose to set aside the ‘preferred’ approach in favour of a method which better suited the data available, carrying out a cross-country empirical regression, as opposed to a time series econometric analysis. [204]

Whilst the vast majority of research measuring the economic impact of corruption has employed cross-sectional methodologies – like the one selected by the World Bank economists working on Bangladesh – these approaches have also been critiqued for presenting only a “snapshot in time”, which does not fully represent corruption and economic growth over a longer period. [205] The aforementioned report suggests longitudinal approaches as a possible solution, but as identified by the World Bank, data availability can be a barrier to models which assess impact across time. [206]

Others have recognised that analyses sometimes neglect to account for other institutional factors, echoing the fears around attribution surfaced in discussions earlier in the literature review. For instance, Al-Sadig finds that when the quality of institutions is controlled for, the negative effects of corruption on FDI identified by Wei, Habib and Zurawicki, and Voyer and Beamish “disappears”. [207]

Another final challenge concerns the universality of analyses – for instance, considerable differences have been found in terms of corruption’s impact on FDI flows both within and outside the OECD. [208]

2. Measuring the economic impact of fiscal transparency [209]

Despite a fairly small body of work into the economic impacts of fiscal transparency, the research we encountered in this area raised important questions regarding the trade-offs associated with causal analysis, which have been factored into our discussion of potential methodologies for measuring BOT benefits.

Fiscal transparency and beneficial ownership transparency can be viewed as neighbouring fields, which share some policy objectives in terms of controlling public sector corruption and boosting citizen trust in government, and look to achieve them through principles of open, reliable, and timely data publication. Given these shared objectives and mechanisms for leading to impact, we conducted a review of econometric studies in the fiscal transparency field to determine how researchers have tracked impact, and assess how this might be relevant to our work.

Overall, the literature covering the economic impacts of fiscal transparency is much more modest than that on corruption. The studies we encountered in this field tended to focus on identifying causal associations between fiscal transparency and a range of positive outcomes. For instance, a number of papers including Hameed, [210] Glenersster and Shin, [211] Alt and Lassen, [212] and De Simone et al. [213] find that even after controlling for several variables, fiscal transparency is positively associated with higher credit ratings, lower public debt, and control of corruption.

Crucially, however, as causal models, none of the work here attempts to size or monetise the benefits identified. Instead, studies seek to prove the existence of qualitative benefits with confidence, and as such, address the ‘attribution’ problem identified in literature on BOT and corruption.

Whilst they might be described as an academic gold standard in terms of linking benefits to a certain variable, accurate casual models can be extremely time-consuming and expensive to produce and often result in findings that are narrow in scope. As such it is unsurprising that they have not yet been produced for BOT – a relatively nascent policy field compared to fiscal transparency.

3. Measuring the economic impact of open contracting

In the analyses we encountered which measure economic impact in the open contracting space, case studies were often used to make simple and persuasive pre- and post-implementation comparisons. Whilst little work has been done to put an economic value on contracting data itself, research here emphasises the need for baseline, counterfactual and costing data in order to track impact.

The final jurisdiction we analysed in the literature review was open contracting. The connection between procurement transparency and beneficial ownership information is a crucial one; when open contracting and beneficial ownership information are combined, it is possible to reach conclusions on where contracts are being concentrated and whether the government is getting value for money. Without beneficial ownership information, it is much more difficult to determine any possible conflicts of interest using open contracting data, and to ensure questionable or unreliable suppliers are not re-engaged under another name. [214]

Given the interoperability of data in each field, and some of the shared objectives of open contracting and BOT in terms of improving public procurement outcomes and reducing illicit activity, it seemed natural to look to literature measuring the economic impact of open contracting for insight.

Perhaps due to the relative maturity of the open contracting advocacy space, we found more examples of impact quantification linked to e-procurement systems than BOT registers. However, we didn’t encounter any research which attempted to size the economic effects of open contracting as a policy across countries, in the same way that the Deloitte / European Commission report and the BEIS/Companies House 2019 paper sought to put a value on company ownership data.

Instead, quantitative impact tracking in the open contracting space has been largely tied to specific case study examples. A number of examples we identified here relate to open contracting reforms in Ukraine, where the introduction of a new e-procurement platform, Prozorro, led to a range of observed benefits, including average savings of 15% for healthcare organisation publishing contracts on the platform. [215]

Open Contracting Partnership’s monitoring evaluation and learning work in this area has been crucial to generating these descriptive statistics. By establishing a number of key indicators at the start of the ProZorro project in 2015, OCP and the Prozorro team have been able to establish a number of baselines off which progress can be measured. Initial findings pointed towards a number of positive impacts, including a 9.7% increase in savings overall when comparing estimated contract value with actual contract value.

Despite this progress, a 2019 study in which Kovalchuk et al. used a regression discontinuity approach to show that the implementation of ProZorro was linked to an increase in the number of bidders, cost savings, and reduced contracting times, pointed to a lack of data preceding the implementation of ProZorro as a key barrier to robust implementation. According to one of the author’s of the report, C. Kenny, the lack of data for small “below threshold” contracts before implementation meant that results were “still informative, but not as conclusive as we had wanted”. [217]

In addition to emphasising the importance of collecting baseline data both at the start of an intervention, and before it is implemented, literature in the field of open contracting also emphasises the importance of work to estimate the costs of an intervention, so that these figures can be used in future cost-benefit analyses. In another paper focussing on ProZorro, Vissapragada uses the Open Government Costing Framework and Methods developed by Results for Development to estimate that the initiative cost over 4.6 million euros to set up and put into operation, arguing that these figures provide “a first step towards conducting a cost-benefit analysis of open government reforms”. [218]

The emphasis on cost is an important one, and perhaps provides some optimism around the potential for further cost-benefit analyses in the beneficial ownership space using costing data from the UK government, some of which has been documented in the 2014 impact assessments and 2019 Post Implementation Review of the PSC register. [219] [220]


[164] Financial Action Task Force. (2019). Best Practices on Beneficial Ownership for Legal Persons.

[165] Open Ownership. Impact stories. Last updated April 2021.

[166] Transparency International. (2021). Out in the open: How public beneficial ownership registers advance anti-corruption.

[167] HM Treasury. (2018). Guide to Developing the Programme Business Case.

[168] Ibid. p. 20.

[169] European Commission. (2020). Impact Assessment study on the list of High Value Datasets to be made available by the Member States under the Open Data Directive.

[170] Davila, J., et. al. (2019). Towards a Global Norm of Beneficial Ownership Transparency. Adam Smith International., p. 34.

[171] Van der Merwe. T. (2020). U4 Helpdesk Answer, Beneficial ownership registers: progress to date. Transparency International., p. 16.

[172] Department for Business, Energy and Industrial Strategy (BEIS). (2019). Post-Implementation Review of the People with Significant Control Register., p. 40

[173] Davila, J., et. al. (2019). Towards a Global Norm of Beneficial Ownership Transparency. Adam Smith International. p. 34.

[174] Department for Business, Innovation and Skills (BIS). (2014). Final Stage Impact Assessments to Part A of the Transparency and Trust Proposals (Companies Transparency)., p. 25.

[175] Department for Business, Energy and Industrial Strategy (BEIS). (2019) Post-Implementation Review of the People with Significant Control Register., p. 12.

[176] Ibid. p. 39-40.

[177] Department for Business, Innovation and Skills (BIS). (2014). Final Stage Impact Assessments to Part A of the Transparency and Trust Proposals (Companies Transparency)., p.25.

[178] European Commission. (2020). Impact Assessment study on the list of High Value Datasets to be made available by the Member States under the Open Data Directive., p. 25.

[179] Davila, J., et. al. (2019). Towards a Global Norm of Beneficial Ownership Transparency. Adam Smith International., p. 43.

[180] Department for Business, Innovation and Skills (BIS). (2014). Final Stage Impact Assessments to Part A of the Transparency and Trust Proposals (Companies Transparency)., p. 25.

[181] Home Office. (2013). Understanding organised crimes: estimating the scale and the social and economic costs.

[182] Department for Business, Innovation and Skills (BIS). (2014). Final Stage Impact Assessments to Part A of the Transparency and Trust Proposals (Companies Transparency).

[183] Van der Does de Willebois, E., et. al. (2011). The Puppet Masters: How the corrupt use legal structures to hide stolen assets and what to do about it. World Bank., p. ix.

[184] United Nations Office on Drugs and Crime (UNODC). Asset Recovery. Accessed February 2022.

[185] The India Economic Times. (2010). Corruption robs developing nations of $40 billion annually: World Bank.

[186] Vlasic, M. V. and Noell, J. N. (2010). Fighting Corruption to Improve Global Security: An Analysis of International Asset Recovery Systems. Yale Journal of International Affairs.

[187] HM Treasury / Department of Trade and Industry (DTI). (2002). Regulatory Impact Analysis: Disclosure of Beneficial Ownership of Unlisted Companies., p. 6-7.

[188] Ibid. p. 58.

[189] Open Government Partnership. (2019). Global Report: Beneficial Ownership.., p. 5.

[190] This is because the PSC information on the register is only confirmed to be accurate in confirmation statements every 12 months. Therefore, sometimes onlookers wish to request more up to date information directly from companies. In other instances, onlookers also wish to request the date of birth of a PSC which is required for certain banks’ due diligence processes, but isn’t available on the public register, to protect against identity fraud.

[191] Vickery, G. (2011). Review of Recent Studies on PSI Re-Use and Related Market Developments. European Commission.

[192] Data landscape. Data Landscape: The European Data Market Monitoring Tool. Accessed February 2022.

[193] European Commission. (2020). Impact Assessment study on the list of High Value Datasets to be made available by the Member States under the Open Data Directive., p. 390.

[194] Department for Business, Energy and Industrial Strategy (BEIS), Companies House. (2019). Valuing the User Benefits of Companies House Data: Policy Summary.

[195] European Commission. (2020). Impact Assessment study on the list of High Value Datasets to be made available by the Member States under the Open Data Directive. p. 137.

[196] Department for Business, Energy and Industrial Strategy (BEIS), Companies House (2019). Valuing the User Benefits of Companies House Data: Policy Summary.

[197] Department for Business, Energy and Industrial Strategy (BEIS), Companies House (2019). Valuing the User Benefits of Companies House Data, Report 1: Methodological Framework. p. 16-21.

[198] Department for Business, Energy and Industrial Strategy (BEIS). (2017). Register of Overseas Entities.

[199] Davila, J., et. al. (2019). Towards a Global Norm of Beneficial Ownership Transparency. Adam Smith International., p. 34.

[200] Rahman, A., Kisunko, G. and Kapoor, K. (2000). Estimating the Effects of Corruption: Implications for Bangladesh. The World Bank.

[201] Alemu, A. M. (2012). Effects of Corruption on FDI Inflow in Asian Economies. Seoul Journal of Economics. Volume 25, no. 4.

[202] Faruq, H., Webb, M. and Yi, D. (2013). Corruption, Bureaucracy and Firm Productivity in Africa. Review of Development Economics. Volume 17, no. 1.

[203] Department for International Development (DFID). (2015). Why corruption matters: understanding causes, effects and how to address them. p. 38

[204] Rahman, A., et. al. (2015). Estimating the Effects of Corruption: Implications for Bangladesh. The World Bank.

[205] Department for International Development. (DFID) (2015). Why corruption matters: understanding causes, effects and how to address them. p. 38.

[206] Ibid.

[207] Al-Sadig, A. (2009). The effects of corruption on FDI flows. Cato Journal. Volume 29, no. 2. p. 283.

[208] Department for International Development (DFID). (2015) Why corruption matters: understanding causes, effects and how to address them. p. 46.

[209] Fiscal transparency can be defined as “the comprehensiveness, clarity, reliability, timeliness, and relevance of public reporting on the past, present, and future state of public finances”. Source: IMF. Why Fiscal Transparency Matters. Accessed February 2022.

[210] Hameed, F. (2005). Fiscal Transparency and Economic Outcomes. International Monetary Fund (IMF). p. 17.

[211] Glennerster, R. and Shin, Y. (2008) Does Transparency Pay? International Monetary Fund (IMF). p. 206.

[212] Alt. J. E. and Lassen, D. (2006). Fiscal transparency, political parties, and debt in OECD countries. European Economic Review. Volume 50, no. 6.

[213] De Simone, E., et. al. (2017). The Impact of Fiscal Transparency on Corruption: An Empirical Analysis Based on Longitudinal Data. The B.E. Journal of Economic Analysis & Policy. Volume 17, no. 4.

[214] Open Ownership, Open Contracting and Oxford Insights, (2021). Integrity in IMF Covid-19 financing: did countries deliver on their procurement & beneficial ownership transparency commitments?.

[215] Amin, L. (2017). Making the case for open contracting in healthcare procurement. Transparency International.

[216] Kovalchuk, A., Kenny, C. and Snyder, M. (2019). Examining the Impact of E-Procurement in Ukraine. Center for Global Development.

[217] Kenney, C. (2019). Measuring the impact of open contracting. Centre for Global Development.

[218] Vissapragada, P. (2017). Open Government Case Study: Costing the ProZorro e-Procurement system. Results for Development.

[219] Department for Business, Innovation and Skills. (2014). Final Stage Impact Assessments to Part A of the Transparency and Trust Proposals (Companies Transparency).

[220] Department for Business, Energy and Industrial Strategy (BEIS). (2019). Post-Implementation Review of the People with Significant Control Register.

Next page: Annex II: Bibliography