Data, incentives, and relationship building for improved tax enforcement

Tax authorities in developing countries often lack the information and enforcement capacity necessary to collect adequate tax revenue. Both innovative and practical research, performed in collaboration with tax authorities, can address these constraints.

The International Growth Centre (IGC) works closely with its partner governments to promote effective tax policies and systems and increase revenue generation. Historically, research and analysis on these important policy issues has been surprisingly limited due in part to the sensitivity and low availability of data.

An IGC regional conference on tax, which took place on 11 July 2018, co-hosted by the Zambia Ministry of Finance, showcased international research in tax design and enforcement, presenting a series of research projects completed in the past two years to an audience of African revenue authorities, policymakers, researchers, and civil society.

Here are some of the key research findings presented at the conference:

1. Identifying bogus firms to unlock revenue potential in India

High-performing tax administration is key for successful revenue generation. One critical area highlighted during the conference focused on improving the effectiveness of tax administration, specifically by leveraging the data available from revenue authorities to identify more efficient methods for tax collection.

i. The study

Ofir Reich (University of California, Berkeley) presented research conducted in India, which looked to identify bogus firms that only exist on paper and make money by falsely reporting transactions with genuine firms. This leads to the government receiving less tax due to the lower estimate of value added calculated. These firms are difficult to detect, requiring significant man-power for what are often poorly targeted inspections. The estimated revenue losses associated with bogus firms is around $300 million annually in India.

ii. The data

The research, conducted by Reich and co-authors Aprajit Mahajan (University of California, Berkeley) and Shekhar Mittal (University of California, Los Angeles), made use of anonymised data provided by the National Government Territory of Delhi, India, It provided information on the Value Added Tax (VAT) returns of all registered private firms.

iii. Methodology

Using a machine learning approach, they were able to analyse the data to better understand the behaviour of firms who were known to be bogus, before looking for similar suspicious behaviour in the data available for existing firms. They then targeted the firms that exhibited similar suspicious characteristics. The approach is continually refined as more bogus firms are identified and this data is fed back into the model, improving future predictions.

iv. The findings

The analysis identified different indicators of suspicious behaviour, with one example being that bogus firms tend to have low profit margins, and correspondingly, trading partners with low profit margins. The approach proved to be highly accurate; of the top 400 suspicious firms, the expectation is that at least 30% will be bogus. The research has the potential to identify significant revenue gains and demonstrates the value of utilising data held by revenue authorities in creative and innovative ways.

The researchers are now exploring options for extending the same approach to identify bogus firms in other states in India and working with revenue authorities in other countries. There is also significant potential to extend the methodological approach to other areas of tax administration.

2. Improving tax compliance with effective consumer and firm incentives 

Another major theme connecting many of the projects presented is the important role of incentives in improving the effectiveness of tax systems, including at the consumer and firm level.

A. Rwandan receipt lottery

i. The study:

Anders Jensen (Harvard University) presented a VAT project in Rwanda, funded by the IGC, examining the role of consumers in encouraging firms to comply with VAT. Researchers sought to understand what keeps consumers from participating in government schemes that incentivise them to enforce VAT (i.e., ask for a receipt when buying something at a retail store) through a survey of 550 Rwandans across major cities.

ii. The findings

One such scheme in Rwanda is a ‘receipt lottery’, where Electronic Billing Machine (EBM) receipts become lottery tickets for cash prizes. The study found that:

• Consumer participation in the lottery scheme is constrained by the sign-up criteria: 31% of those surveyed were ineligible because they don’t have access to/don’t use the internet.
• Consumer participation is also constrained by a lack of awareness of the scheme: 40% did not know about the lottery.
• Consumers face potential price increases: 55% report having to pay a ‘price penalty’ when asking for a receipt.
• Consumers are willing to participate in well-advertised schemes with a sufficiently high reward.

B. Ugandan tax evasion 

i. The study:

Another IGC project on VAT compliance in Uganda, presented by Justine Knebelmann (Paris School of Economics), is looking at the effectiveness of the VAT paper trail in dis-incentivising tax evasion.

ii. Methodology

As part of the study, the Uganda Revenue Authority (URA) sent monitoring letters to firms found to be misreporting their VAT.

iii. Preliminary results

• After receiving the letter, 20% of these firms amend their VAT reporting.
• Firms are less likely to misreport their VAT and if they do misreport, do so by a lesser amount.

Overall, VAT liability increases for 62% of buyer and 21% of seller firms that are part of the study.

C. South African small business tax compliance

i. The study

In South Africa, a project presented by Sharon Smulders (University of South Africa) examined the role of costs and incentives for small businesses to be tax compliant.

ii. The findings

• The costs to comply with taxes are high and regressive, with laws and operational procedures being very complex.
• Furthermore, these businesses are not aware of the incentives provided by the revenue authority. For example, a scheme that identifies high performing small businesses and provides incentives such as tax benefits and the opportunity to bid for government contracts.

iii. Recommendations:

• Improving education and outreach to small businesses
• Making tax filing easier
• Simplifying legislation

3. Relationship building for targeting high net worth individuals in Uganda 

While evidence shows technologies and incentive schemes can improve tax enforcement and compliance, good old fashioned relationship building can play a role as well. This is certainly the case in Uganda, where the URA has made this their primary strategy in targeting high net worth individuals (HNWIs).

i. The study

Milly Nalukwago and Ronald Waiswa (URA) shared the results of a collaboration with IGC researchers in which the URA identified potential HNWIs in Uganda, and assessed the URA’s approach to taxing these individuals.

ii. The findings

They found that establishing a special URA unit dedicated to HNWIs has been very successful in recovering taxes. The unit, which has five staff members, maintains a register of VIPs to target. It encourages voluntary tax compliance by interacting one-on-one with HNWIs.

Staff come to these meetings armed with tax information about these individuals, use soft skills, and keep interactions confidential to build trust. The URA has found that discretion and a personal touch are especially important as many HNWIs wield significant political influence or are policymakers themselves. Using these lessons, the URA has advised Kenyan revenue authorities on the topic of taxing HNWIs.

An accompanying blog outlining the conference’s key policy messages is available here.