00:00 OK,
00:01 I think let's get going.
00:03 Well,
00:03 welcome everybody.
00:04 Uh,
00:05 good morning,
00:05 good afternoon,
00:06 good evening,
00:07 depending on where you are.
00:08 Um,
00:09 my name is Deon Filmer,
00:10 the director of the World Bank's Development Research Group.
00:12 It's my pleasure to welcome you today to
00:15 this November edition of our policy research talks.
00:19 Um,
00:20 as some of you may know,
00:21 these talks provide us the opportunity to present
00:23 work coming out of the bank's research department.
00:26 With the goal of sharing,
00:27 uh,
00:28 the findings with colleagues inside and outside the department,
00:31 along with others outside the World Bank.
00:33 So with that,
00:34 I'd like to welcome our online audience,
00:36 both on Webex,
00:37 um,
00:38 as well as on YouTube.
00:40 Uh,
00:41 our presenter today is my colleague,
00:42 uh,
00:43 Oyebola Okunnobe,
00:45 who's an economist in the Human Development
00:47 and Public Services team in the department.
00:49 Um,
00:50 she
00:51 joined the department shortly after completing her
00:53 PhD at Harvard University in 2016,
00:56 and she's a research associate at the Institute for Fiscal Studies UK
01:00 and an advisory Committee member of the Nigerian Tax Research Network.
01:06 So today Oyebola will present her research on the potential
01:09 of technology to transform 3 core tax administration functions.
01:14 Um,
01:14 identifying the tax base,
01:16 reducing compliance costs,
01:18 and monitoring compliance.
01:20 She's also going to discuss factors that limit the effectiveness of
01:24 tax digitalization and increasing tax revenues,
01:27 such as a lack of complementary enforcement capacity.
01:32 So we're very grateful today to have uh Kiara Bronke as a discussant.
01:36 Uh,
01:36 Kiara is the Practice Manager for the Fiscal
01:39 Policy and Sustainable Growth Unit in the Macroeconomics,
01:42 Trade and Investment Global Practice of the World Bank.
01:45 She leads a group of macro fiscal economists and tax
01:48 experts who work on fiscal policy and sustainable growth.
01:52 She's over 25 years of professional
01:53 experience in managing multi-stakeholder and multi-disciplinary programs
01:58 at various international organizations with positions in the UK government,
02:02 the OECD,
02:03 the IMF,
02:04 and the ADB,
02:05 among others,
02:06 prior to the World Bank.
02:08 She's published on taxation and public ser spending,
02:11 and recently,
02:12 uh,
02:12 she co-authored an approach paper for the conclave of ministers on Finance,
02:16 of Finance on financing human capital development.
02:20 So today,
02:21 I'll ask Ayebola to talk for approximately,
02:23 to present for approximately 45 minutes,
02:25 after which we'll hear from Kiara for about 10 to 15 minutes.
02:29 Uh,
02:30 we'll conclude the session with questions and answers from the audience.
02:33 If you have a question,
02:34 please use the either the raised hand function on Webex
02:38 or signal to me in the chat,
02:39 um,
02:40 and that you have a question,
02:41 and I'll call on you.
02:43 To pose your question out loud.
02:45 If you're following on YouTube,
02:46 uh,
02:47 please submit your question in the chat there and it will get related
02:51 to me.
02:52 Um,
02:53 just a reminder,
02:54 we're recording the session today.
02:56 So with that over to you,
02:58 Oyeboa.
03:01 Thank you,
03:01 Dion,
03:03 for that introduction and thank you very much to
03:05 every single one of you for joining this conversation
03:08 on the use of technology in tax mobilization.
03:12 So,
03:13 given that this was a self-selected audience,
03:15 maybe this first slide is not essential,
03:18 but I think that it's important to establish this case
03:21 to start with.
03:22 So,
03:22 in a place like the World Bank,
03:23 where we have colleagues working on
03:26 Very important things like health,
03:29 education,
03:29 social protection,
03:30 climate.
03:31 Why do we still care about taxes?
03:34 And the simple answer
03:36 is that
03:37 taxes are what fund all of these things that I just mentioned,
03:40 these core
03:42 functions of government,
03:43 from investing in public goods to providing social protection,
03:47 governments needs taxes to function.
03:49 And this is highlighted more than ever before
03:52 in the response to the current pandemic.
03:56 Now,
03:57 besides taxes,
03:58 the other sources of revenue that governments typically would
04:01 rely on are foreign aid and natural resource revenues,
04:04 and taxes have the advantage of being a lot less volatile and more sustainable.
04:09 And one final thing to highlight in the importance of taxation is that it's
04:13 a core part of the social contracts that governments have with their citizens.
04:17 So,
04:17 any government that's seeking to tax its subjects
04:20 needs to bargain with them over policy
04:23 and provide the types of public goods that they desire.
04:25 And so,
04:26 we find that
04:27 taxes play this role,
04:28 this accountability role as well.
04:32 Now,
04:33 what we do also observe is that low-income countries,
04:36 they tend to collect less tax.
04:38 So,
04:38 this graph is showing the relationship between tax revenues
04:42 as a percentage of GDP on the Y axis
04:45 and
04:47 GDP per capita on the X axis.
04:50 And what we see here is that a country,
04:52 a low-income country like the Central African Republic
04:55 here is at less than 10% of GDP that they collect in taxes,
04:59 compared to a country like Norway,
05:00 that's closer to 40%.
05:02 And what we also need to keep in mind is that these rich countries today,
05:06 France,
05:06 Norway,
05:07 Luxembourg,
05:08 100 years ago,
05:09 so in 1920,
05:10 they were exactly here
05:11 at about 10%
05:13 of GDP.
05:14 And so this capacity and ability to tax
05:16 is something that countries develop over time.
05:20 And why is it that we observe such
05:22 low rates of tax collection in low-income countries?
05:27 There are a few reasons.
05:29 The first one that I'll talk about is
05:31 that
05:32 countries have limited information on their tax base.
05:35 So one of the key findings in the recent tax literature
05:38 is that information is central to tax collection.
05:41 In cases um where there's high levels of informality,
05:45 the government has less opportunity
05:48 to have information on what are taxable transactions and assets.
05:51 And this is a
05:53 crucial function that the government needs to have the
05:56 capacity to understand what the tax base is.
05:59 Because from
06:00 Places all across the board at different income levels.
06:03 What we find is that taxes that are subject to things like third party
06:07 information,
06:08 where another entity,
06:09 for instance,
06:10 an employer is reporting one's wages to the government,
06:14 levels of compliance are much higher in those settings.
06:16 And so,
06:17 when large majority of the population
06:19 is not having any
06:21 um formal links through which there's other parties reporting their income,
06:26 um
06:26 this could lead to lower tax collection.
06:30 The second one I will discuss is administrative capacity.
06:34 And so,
06:35 for those of us who spent time in government offices in
06:38 developing countries,
06:39 this scene would not be too unfamiliar.
06:42 Many times you have a lot of manual processes
06:45 that leads to inefficiencies as well as high compliance costs for the taxpayers.
06:51 The 3rd
06:53 dimension of capacity that I would like to emphasize is enforcement.
06:57 So,
06:57 at the end of the day,
06:58 tax officials are typically the ones who go in the field to enforce tax
07:03 laws.
07:03 And if they are compromised,
07:05 this leads to leakages in revenue.
07:07 In addition,
07:08 even thinking outside the tax
07:10 authority,
07:11 if the
07:12 legal systems of the courts are not functioning properly,
07:15 and there are no credible threats of either financial or criminal
07:19 consequences for evasion,
07:21 then you find that collection levels will continue to remain low.
07:26 Now,
07:26 if we switch hats for a second,
07:28 at this same time,
07:30 we know that there's
07:31 ways in which technology
07:34 is reshaping different functions of the government.
07:36 So,
07:37 we have examples from things like procurement,
07:39 education,
07:39 voting,
07:40 social protection,
07:41 and in fact,
07:42 our 2016 World Development report was exactly on this topic.
07:48 So the big question that we're asking today is,
07:50 can technology
07:51 improve tax collection?
07:53 And this is in light of the fact that there's been a rapid adoption
07:57 of technology in taxes over the last decade.
07:59 So,
07:59 even from the time that I started working on this in about 2014 until now,
08:04 just the rates at which different countries have moved
08:06 forward in their journey has been quite impressive.
08:09 And we know that this would only continue to increase.
08:12 We are now um
08:13 moving to a world where social distancing,
08:16 remote engagement
08:18 are becoming more and more of the norm.
08:19 Look at all of us on this floor.
08:21 Uh,
08:22 and this we think,
08:23 I think will continue also in taxation.
08:25 And so,
08:26 we can only expect the use of technology to increase.
08:29 There is also a lot of enthusiasm around the potential of technology.
08:33 So,
08:34 there's several,
08:35 um
08:35 both from the research side but most important from
08:38 the policy side about how technology might change things.
08:40 And it's important to have a critical look at
08:42 the extent to which this potential is being realized
08:45 and how we might be able to maximize the impact.
08:49 So,
08:49 what we're going to do today is that I'm going to walk you through 3 papers.
08:53 The first one is an overview of the literature
08:55 to understand um
08:57 what we've learned so far on the promise and limitations of
09:00 IT for tax mobilization.
09:03 And then I'm going to do deep dives into two
09:06 case studies from low,
09:08 very low-income context,
09:09 where we can look at what was the technological intervention,
09:13 what was the
09:15 impact that it had on tax collection,
09:17 or the challenges,
09:18 what are the lessons learned?
09:21 And so there's a paper like this and one in Tajikistan.
09:24 And so to give you the punchline first,
09:27 the fact is that yes,
09:29 technology can help the tax authority and
09:31 in
09:31 three dimensions.
09:33 One is helping the tax authority to define compliance,
09:36 meaning to identify what needs to be taxed.
09:38 It could be a major source of information on what the tax base is.
09:42 Two is to facilitate compliance by making it easier for people to pay.
09:45 And 3 is to monitor compliance,
09:48 to detect evasion
09:50 when it occurs because it almost always does occur.
09:54 Very quickly,
09:55 I need to break the news that technology is by no means a silver bullet.
09:59 Uh,
09:59 its impact is going to depend on other dimensions of capacity in the government,
10:04 one of which is the enforcement capacity.
10:08 At the end of the day,
10:09 it's going to require,
10:11 um,
10:11 it could provide a lot of
10:13 resources,
10:13 but enforcing the tax laws,
10:16 having that capacity remains crucial.
10:18 The other point that I would emphasize today
10:20 is that technology creates winners and losers.
10:23 Uh,
10:24 people who benefit
10:25 may
10:26 easily comply,
10:27 but the losers may act strategically to undermine the impact
10:30 of the technology that has been introduced.
10:35 So let's dive into the first speaker.
10:37 This is joint work with Fabricio Santoro,
10:40 who is at the
10:41 ICTD,
10:42 the International Center for Tech and Development,
10:44 leading,
10:45 uh,
10:45 a
10:45 great project they have going on on the on the digitalization of taxation.
10:56 So as I mentioned,
10:58 we will be looking at the role of technology in transforming 3 functions,
11:02 3 of the core functions of the tax administration.
11:05 The first one being to define compliance,
11:07 to facilitating it,
11:08 and 3 monitoring it.
11:13 So let's start with how technology might
11:15 help governments to understand what the tax base
11:17 is.
11:18 This could happen in many ways,
11:19 for instance,
11:20 through creating completely new
11:22 databases,
11:23 such as
11:24 taxpayer
11:25 registration drive or property registration drive.
11:27 Essentially,
11:28 there's this process of making
11:30 The tax base more legible to the state,
11:32 helping the state to understand
11:34 who are my taxable
11:35 individuals,
11:36 what are taxable um assets that they are,
11:38 and
11:39 having these electronic databases makes
11:41 the landscape a lot more clear to the taxing
11:44 authority.
11:46 A second way this might happen is
11:48 to capture transactions right as they're occurring.
11:51 So,
11:51 for instance,
11:52 the use of electronic billing machines.
11:54 As soon as the
11:55 customer is
11:56 checking out in the store,
11:58 that information on the transaction,
11:59 what was spent
12:00 is immediately transmitted to the tax authority.
12:04 Third is using existing databases to create
12:09 to as sources of third-party information to understand
12:12 more fully what this tax liability of an individual is.
12:16 And to illustrate this,
12:17 we can think of
12:18 how the tax authority might build the profile of the taxpayer.
12:21 So,
12:21 for instance,
12:23 there could be um reports of wages that come from an employer,
12:27 similar to the W-2 system we have in the US
12:30 um for a business,
12:31 it could be this electronic sales register that I just described.
12:34 In fact,
12:35 you could also have credit card companies
12:37 put in all the sales that the company has.
12:40 Um,
12:40 you could have their clients supporting how much they paid to them.
12:42 You could have the electricity company
12:45 saying this is how much power that they use for a factory,
12:47 which gives a proximity of what they're producing.
12:49 Um,
12:49 on the VHC you could have the people that buy from them
12:53 report how much VAT they paid to them,
12:54 and it serves as a cross-check on what the firm itself reports.
12:57 You can think of
12:58 on and on using all these pieces of information,
13:01 the government can create a
13:03 more comprehensive profile
13:05 of the taxpayer and what all their liabilities can be.
13:09 OK,
13:09 thank you.
13:10 Thanks for alerting me to it.
13:12 Um,
13:14 OK,
13:14 so I was talking about
13:17 The role of
13:19 Um,
13:20 technology in defining compliance and thinking about how it could
13:22 be used to link different databases on an individual.
13:26 And so these tiles were showing different sources
13:29 of information that the tax authority could have
13:32 on the full profile of a taxpayer.
13:35 So,
13:35 for example,
13:36 we could have employers that are reporting the wages as in a W-2 in the US.
13:40 We could have the electronic sales register
13:44 reporting um purchases as I just described on the previous slide.
13:48 We could have an electricity company
13:50 report how much the
13:52 Um,
13:53 company's bill is as a proxy for what the factory is producing.
13:56 We could have VAT,
13:58 um,
13:58 purchases report how much they paid.
14:01 And so the idea is that
14:03 with
14:04 different sources of third-party information,
14:06 a tax authority can have
14:08 rich information on what um
14:10 the tax level of an individual or firm should be.
14:14 So now moving to the next
14:17 um
14:19 area in which technology may help to
14:22 improve taxation
14:23 is in making compliance easier,
14:25 so facilitating compliance
14:27 and
14:29 There are many ways that this could happen,
14:30 and I'll start with the most basic,
14:31 which is having a good,
14:33 nice,
14:34 clear website
14:35 where all the tax rules are clearly stated,
14:37 how to guide,
14:38 what exactly needs to be paid,
14:40 when and how.
14:41 Um,
14:42 in many developing countries,
14:43 you find out that individuals feel the need to have an
14:47 insider,
14:47 like a contact in the tax office to guide them through the process.
14:51 And a lot of this needless um interactions between
14:54 taxpayers and tax officials could be solved with clear
14:58 information,
14:58 even
14:59 disseminated through a website.
15:01 Another major thing is how the tax authority communicates.
15:04 So,
15:04 the modern methods use an email,
15:06 SMS,
15:07 voice calls,
15:08 reminders,
15:08 things that could um
15:10 alert the
15:11 individuals
15:12 of information from the tax authority.
15:16 Next,
15:16 we move on to things like e-filing and e-payments,
15:18 which we'll discuss quite a bit here.
15:20 And um we could also have
15:22 things like pre-fuel returns.
15:24 This would be great,
15:24 even in the US where the government essentially sends
15:28 the taxpayer
15:30 information that they've gathered from the different sources.
15:32 So like I was showing from the previous tile,
15:34 um,
15:35 from the previous slide,
15:36 where all the taxpayer then needs to do.
15:38 is to sign and submit.
15:40 And so the goal in all of these is to have an easy transparent process
15:45 that would not require any personalized
15:47 interactions between taxpayers and tax officials,
15:50 and you find that these tend to have a dual effect.
15:53 So not only does it
15:54 make compliance costs lower,
15:56 but then it also reduces instances of corruption and harassment that happen.
16:03 The third big bucket of ways in which technology can improve tax collection is
16:09 through monitoring compliance,
16:11 detecting evasion when it's happening.
16:14 And so we could think of the tax authority using IT to
16:18 automate
16:19 regular crosschecks
16:20 of self-reported tax liability.
16:22 So
16:23 not only do they,
16:24 does it help to collect this information,
16:26 but how How do you
16:27 actually then use it to detect
16:30 where,
16:30 oh,
16:31 why did your um suppliers say that you paid X,
16:33 but instead you're claiming to have paid Y?
16:36 And so that's a very promising way of
16:38 um
16:39 monitoring compliance.
16:40 And one interesting application also is using
16:44 all this information to build a risk profile
16:46 for taxpayers.
16:47 So they're interested in um
16:50 Uh,
16:50 exciting machine learning methods where using the
16:52 vast information that the tax authority is collecting
16:56 on individual taxpayers,
16:58 they can build a
17:00 way of predicting
17:01 who is likely to be evading
17:03 based on past behavior.
17:06 These are some promising avenues.
17:08 And so very quickly now,
17:10 I want to then point us in the
17:14 other direction,
17:15 which is
17:16 thinking about the limitations.
17:17 Why is it that sometimes
17:19 tax um
17:20 IT projects and taxes don't yield the desired
17:23 results?
17:26 So the first point here is thinking about infrastructure and connectivity.
17:29 This sounds really basic,
17:30 I know,
17:31 but it's surprising how often it happens where
17:35 the tax authority might purchase software
17:38 and they realize that,
17:39 oh,
17:39 we actually don't have all that we need to
17:42 run this and we don't have the right computers
17:45 that we need to make this work.
17:46 And then by the time you get the computers,
17:48 the license has expired.
17:49 And so
17:50 thinking very carefully about what the The structure and connectivity needs are.
17:55 And we can also think about this on the side of taxpayers
17:58 to think about
17:59 what is it going to be,
18:01 what type of technology would be easy for people to adopt.
18:03 And one nice example from this we see is from Rwanda where
18:07 they had an
18:08 initial filing,
18:09 e-filing system,
18:10 but then they created another version
18:13 that could be used on mobile phones with it,
18:15 it's called the N Declaration,
18:17 such that um
18:18 to address equity concerns.
18:23 The second point under the limitations of technology,
18:26 which I think is
18:28 perhaps
18:30 The
18:34 Perhaps the,
18:35 uh,
18:36 one of the most important is to think about the human factor.
18:39 So,
18:40 technology is a tool at the end of the day,
18:41 it's only as good as its users.
18:43 And we might think about taxpayers who
18:45 may need training and subsidies to actually use
18:48 the technology that has been made available.
18:50 And very importantly,
18:51 we need to think about tax officials
18:53 who need the right skills
18:55 and training
18:56 to be able to use technology.
18:58 as well as the need to have aligned incentives.
19:02 Um,
19:02 there's a very interesting paper by our colleagues in the World Bank,
19:05 um,
19:06 in Madagascar,
19:08 where
19:09 there were different IT interventions introduced in customs,
19:13 and it was literally trying to
19:15 be one step ahead of the tax officials
19:17 as it kept finding ways to
19:20 um circumvent the controls that have been put in with the technology system.
19:24 And so at the end of the day,
19:26 it's um
19:27 IT that IT is not going to enforce the tax laws.
19:30 We need to
19:32 put very careful attention
19:33 in the incentives and the training that tax officials have.
19:38 Other things to think about are
19:40 um
19:41 the overall institutional strategy for adopting technology.
19:44 One mistake is to think that,
19:46 oh,
19:47 this is the IT department,
19:49 and they are going to deal with it.
19:51 No,
19:51 technology has to be thought up through in the entire
19:54 um
19:55 Work that the institution does.
19:57 So one example
19:58 is imagine you start e-filing,
20:00 and now all of a sudden,
20:01 all these officials whose previous job it was
20:04 to collect tax revenues are now redundant,
20:06 we must think about how do they
20:08 find new roles in the tax authority.
20:11 And the final point here is thinking about the regulatory environment,
20:15 um,
20:16 timely prosecution of evaders,
20:17 that's what really sends a strong deterrent message
20:20 on
20:22 tax compliance,
20:23 as well as the need for modern laws to support
20:27 this,
20:28 um,
20:29 to be able to actually use these data points like
20:31 I was describing with the tiles across different agencies,
20:33 they need to be the right environment,
20:35 the right laws that would
20:37 support that type of sharing.
20:40 So,
20:41 this was a very quick overview.
20:43 The paper just came out.
20:44 I saw that Ryan
20:45 posted it in the chat earlier.
20:46 Thank you.
20:47 So it's a fresh hub
20:49 of the press working paper that gives a lot of those examples on
20:53 the um high-level,
20:54 um
20:56 Limitations and promise of technology.
20:59 Um,
20:59 but now what I'm going to do is to move into specific examples from my research,
21:05 um,
21:05 in thinking about
21:06 how this plays out in practice.
21:08 And so,
21:08 the first paper that I'll be discussing
21:10 is on real estate taxes in Liberia.
21:13 And this is a close partnership that I had with the
21:17 Liberian Revenue Authority.
21:18 Over several years,
21:19 they're like my second family,
21:21 and um just thinking through
21:23 different ways in which technology
21:27 could be used in that context.
21:29 And so,
21:30 what
21:31 this paper would be doing would be looking at that first bullet
21:34 under identifying the tax base in the potential of technology,
21:38 and it would also illustrate
21:40 one of the limitations,
21:41 which is thinking about the broader regulatory environment.
21:46 So,
21:47 the context in which this study occurs is that there's extremely high
21:51 rates of noncompliance with the real estate tax
21:55 in Liberia.
21:56 One of the major issues is that there is no um comprehensive cavester,
22:00 no information on
22:01 who owns what property,
22:03 where things are located.
22:05 What are the values of these properties.
22:07 And so,
22:08 um,
22:09 what ends up happening is that many properties
22:11 remain completely out of the tax register.
22:14 Some analysis I did for them
22:16 many years ago was showing that about
22:18 5% of property owners
22:21 in the residential property were actually estimated to be on the tax roll.
22:24 So,
22:24 huge.
22:25 I mean,
22:25 it's
22:27 huge room for increasing compliance.
22:30 What we also noticed from analyzing the data that
22:32 existed was that among those on the tax roll,
22:34 there are high delinquencies in subsequent years.
22:37 So among those that paid in one year,
22:38 only about 40% actually paid in the next year.
22:43 indicating that there are probably um major issues on the enforcement side.
22:47 And the tax rates are comparable to other settings,
22:49 it's about 0.25% for residential property
22:52 and 1.5% for commercial property.
22:56 So,
22:56 in this case,
22:57 the technology investment
22:59 was what I'll call a very low-tech investment.
23:02 It was um training youth to use open-source software on tablets
23:07 to go around and enumerate properties in a specific um pilot
23:12 community.
23:13 So,
23:13 they were supposed to get the ownership.
23:14 They had a quick survey with the residents.
23:17 So up is someone sitting with the residents to discuss
23:20 who owns this property,
23:21 getting that um contact information.
23:23 And
23:24 at the end of this exercise,
23:25 what happened was they were able to create a new electronic property database
23:30 that previously did not exist.
23:31 And so this pilot we did was in with about
23:35 close to 600 properties that were enumerated
23:37 and in subsequent phases,
23:38 the government went on to expand this to about 40,000 properties.
23:45 Then we use this information to conduct a randomized
23:48 intervention where we wanted to see what the impact was
23:52 of
23:53 alerting people that their lack of compliance had been detected by the state.
23:58 So two questions here was,
24:01 what happens when you alert them that the state knows
24:03 that they own this property and they've not paid,
24:05 and 2,
24:06 was there any
24:08 additional impacts from informing them about
24:10 what the penalties of noncompliance are.
24:14 This is the um
24:17 way we run this
24:18 experiment.
24:19 Our control group
24:21 here um
24:22 get a notice
24:23 that only gives them information on the tax requirements.
24:27 It says,
24:27 Dear property owner,
24:29 these are the requirements for property tax,
24:31 please comply.
24:33 In our first treatment group,
24:34 we have that basic information,
24:37 but then it's personalized
24:39 with the name of the property owner as well as a photograph of their property,
24:43 which we derived from this new electronic database collected in the
24:47 pilot project.
24:49 The second treatment group
24:52 has
24:52 a,
24:53 has some information
24:54 on what are the legal penalties for noncompliance,
24:57 it cites relevant portions of the revenue code.
25:00 And then in the final group we have both of these elements included.
25:07 What do we find?
25:08 So,
25:10 The first outcome is people come into the tax office as instructed in the letter
25:15 to register their properties,
25:17 and what we find is about 4.4% of people do this in our control group.
25:22 This goes up a bit in our among the people who
25:25 get the picture of their property and their name on it,
25:29 but not statistically significant.
25:31 But when we have the combination of this
25:33 detection information as well as the penalty information,
25:36 there's about a tripling
25:38 of the registration rate.
25:40 Um,
25:41 in terms of the payments,
25:42 we find that uh this goes up from about 2.2%
25:46 to almost 10% in our combined group again.
25:49 So,
25:50 pretty large impacts on people choosing to um
25:54 register and also following through and actually making payments.
25:58 In terms of the amounts paid,
26:00 we find an
26:01 increase,
26:02 uh,
26:02 in our combined group again,
26:03 but given the large
26:05 amounts of zeros,
26:06 people that just do not pay anything at all,
26:08 we don't detect a statistically diff significant impact.
26:12 When we use other methods like doing an inverse hyperbolic sign,
26:15 then we do see an increase.
26:19 And to think about what these numbers mean,
26:22 what it actually,
26:23 it's quite cost effective because in the first year,
26:28 the tax revenues that are collected,
26:29 they more than cover the cost of
26:31 this entire enumeration process and the delivery
26:34 of these notices to the property owners
26:37 and We look to see,
26:38 OK,
26:38 what happens after the first year,
26:39 and we see that there's a sustained impact.
26:41 So
26:42 4 years later we're still seeing that people who
26:45 received the
26:46 detection and penalty notice
26:48 in 2017
26:49 are still more likely to make a payment,
26:52 and we don't observe these for the other two individual treatments.
26:56 Now,
26:56 one might look at
26:58 these results and think,
26:59 wow,
26:59 that's great.
27:00 We're able to quadruple the payment rates,
27:03 but one could also look at it and think that
27:05 there are people who,
27:06 even after knowing that the government is telling you that Ishani,
27:10 I know that you own this house.
27:12 I know that you've not paid your tax,
27:14 uh,
27:15 and,
27:15 and these are the penalties if you do not pay,
27:17 people just
27:18 don't respond.
27:19 And this is not unique to our paper.
27:21 We find this in other studies where large shares of the population
27:24 who are faced with this type of information where there's perfect detection,
27:28 the government is telling you that we know what you've done.
27:31 They just do not comply.
27:33 And in our setting,
27:34 we are quite confident that this is not because
27:36 they don't understand the notice or didn't get it.
27:39 So the hypothesis is that maybe they just have a very low
27:42 expectation that any penalties are going to be enforced on them.
27:46 Fine,
27:46 the government might know,
27:47 but
27:47 I've never seen anybody go to jail for
27:50 not paying their taxes.
27:52 And so why do I think they would care about little me?
27:55 And so we do a second experiment,
27:56 uh,
27:56 we conduct this experiment to test this hypothesis.
28:00 So what we have here is that in our control group,
28:03 we have it as
28:05 similar as possible to
28:07 the combined treatment group where there is the information
28:10 on what the penalties are and also the detection information
28:13 that tells people that shows the photo.
28:16 And then in the treatment group,
28:18 there's an additional message that lets them know
28:20 that they've been selected for intensive follow-up.
28:23 If there's no response,
28:24 the LRA will proceed to the next stage of enforcement actions.
28:28 And this message is deliberately quite vague because
28:32 um just given the.
28:35 The legal and the
28:37 administrative context,
28:38 the LRA at that point could not commit
28:40 to what exactly these enforcement actions would be.
28:43 Typically,
28:44 um,
28:44 the end,
28:45 what they,
28:46 what they do is repeated visits to the person,
28:49 half cajoling,
28:50 half threatening,
28:51 but essentially just trying to use their words and their visits to
28:55 get them to eventually pay.
28:57 And
28:58 In fact,
28:59 there have been different parts of the law
29:01 that have been adopted in different years.
29:03 So,
29:03 before they could lock up buildings,
29:05 but then they're no longer allowed to do that
29:07 because of the,
29:08 a government rule,
29:09 a,
29:09 a legal ruling.
29:10 And so,
29:11 in fact,
29:11 one time,
29:12 one of the LRA
29:15 officials told me that they feel like a toothless bulldog,
29:17 because in terms of their actual
29:19 ability to enforce,
29:20 they have to use very mild means to cajole people.
29:24 And so,
29:26 what we do find here is that this message actually still does have quite a
29:30 bit of a significant impact where there's a 7%
29:33 point increase in people responding to this notice.
29:36 And in terms of their payments,
29:37 we also see a similar
29:39 increase.
29:40 In terms of the amount paid,
29:41 this goes up significantly by about um
29:44 close to $60
29:45 on average,
29:46 meaning that for each
29:49 person that paid,
29:50 I think it was an average of about $600
29:52 paid by each person that paid.
29:55 And so,
29:56 one of the lessons from this um Liberia study.
29:59 So,
29:59 the first thing is that,
30:00 yes,
30:01 technology can provide crucial information.
30:04 This was a,
30:05 this
30:06 was a data source that
30:08 did not previously exist,
30:10 but from collecting this data,
30:12 the government was able to look and know where the um
30:16 who exactly had not complied.
30:18 And
30:18 as I mentioned,
30:19 it was a technology that was very
30:22 um that was selected.
30:24 Taking into account very much so the conditions of the country.
30:27 There are more fancy ways one could have done this.
30:29 You could have drones,
30:30 you know,
30:31 using satellite imagery,
30:32 but they also had an incentive of providing employment for young people.
30:36 And so this low-tech way still delivered significant outcomes.
30:40 Um,
30:40 and this is not limited to this setting.
30:42 So,
30:43 this same
30:44 idea we could think of in
30:46 information,
30:46 other sources of information,
30:48 um,
30:48 third party information that could give
30:50 the government information on the tax base.
30:53 But I think one of the main lessons that's coming out of
30:55 this study is that information on tax base is not sufficient.
30:59 It's not going to,
31:00 we might know exactly
31:02 who owes what,
31:03 but unless there is credible
31:05 enforcement
31:06 that can be carried out,
31:08 we find that this is not going to be sustained
31:11 in the long run.
31:12 So,
31:13 for example,
31:13 in this case,
31:14 by the 2nd year,
31:15 those um impacts are completely gone.
31:19 So we saw this first example looking at technology
31:22 use in
31:24 understanding the tech space.
31:25 We're going to move on to a second example that's
31:29 going to be looking at
31:32 the role of technology and facilitating.
31:34 So this is joint work with Victor Poloquin,
31:38 um,
31:39 And it was done in partnership with the Tajikistan Tax Committee.
31:44 Again,
31:44 we're very grateful for their support and interest in learning together
31:49 with us and,
31:50 and,
31:50 um,
31:51 experimenting in trying to understand exactly the best
31:54 ways to roll out this e-filing program.
31:56 And uh there's a very um
31:59 strong World Bank team that was involved in,
32:02 in setting this up.
32:05 So,
32:05 like I said,
32:06 this
32:07 paper will be looking at facilitating compliance,
32:09 and in terms of the limitation,
32:11 one of the key things we'll highlight is the human factor.
32:18 The question that we're asking in Tajikistan is,
32:20 how does the introduction of e-filing,
32:23 in this case,
32:23 it's replacing in-person submissions to a tax official,
32:27 how does it affect firms' compliance costs,
32:29 their tax payments,
32:31 and their bribe payments?
32:33 We're working with small and medium firms in
32:37 uh Duchambe.
32:38 And what we do is the randomized encouragement of e-filing.
32:42 So e-filing is available to every,
32:43 was available to everyone,
32:45 um,
32:45 but we,
32:46 um,
32:47 try to encourage a randomly selected group of people to adopt at a higher rate.
32:51 So we're able to achieve
32:53 a 93% adoption in this
32:55 treatment group
32:56 compared to 60% in our control group.
32:58 And it's this difference in the adoption rates that we use to
33:02 identify the impact of e-filing.
33:06 Uh,
33:07 the context in the setting,
33:09 again,
33:09 it's a very low-income setting,
33:10 and here,
33:11 firms face very high compliance costs.
33:13 They have to go every month,
33:15 leave whatever you're doing in your business,
33:17 go and line up at the tax office,
33:19 submit different forms,
33:20 get different signatures before you're considered to have complied.
33:24 Uh,
33:24 and corruption was also a major concern.
33:26 So almost a third of firms see that they expect to give gifts
33:30 when they meet with tax officials.
33:33 The tax officials in this setting are very,
33:35 very key part of the story because they have strong incentives
33:39 as well as discretionary powers.
33:40 So in terms of the incentives,
33:42 each official monitoring taxpayers
33:44 is given a target
33:46 of how much tax they're expected to
33:48 collect from the people that they're monitoring,
33:49 so they have an incentive to collect as much as possible.
33:53 And they also have discretion in that until they sign the declaration,
33:57 the tax,
33:59 the
33:59 firm is not considered to have complied.
34:01 And so
34:02 there's a holdup opportunity where if I tell you I'm like,
34:05 Dion,
34:05 no,
34:06 this is too low,
34:07 really,
34:07 you didn't make any money last month.
34:09 You need to bump this up.
34:11 And until he complies with what I want,
34:14 I do not accept that declaration.
34:16 So they're able to influence the declaration upwards.
34:19 Or um from the different,
34:21 um.
34:22 Interviews and focus groups we did,
34:24 there's also the opportunity to
34:26 help firms with advice on how they might minimize their liabilities.
34:30 And in this setting,
34:31 the government was particularly interested in
34:33 using e-filing to one,
34:35 improve their service deliveries to cut this compliance costs down,
34:38 but also to reduce corruption.
34:42 The design of this experiment is that we have a control group who we invite to the tax
34:49 authority seminar to get a general training on taxation.
34:54 Then there's the first treatment group that
34:56 gets information and training on e-filing,
34:58 so they get a walkthrough of the entire system.
35:00 This is how you use it,
35:01 this is all the way,
35:03 this is exactly
35:04 what needs to happen.
35:05 And our third group
35:07 gets in addition to that,
35:09 logistical help for registration.
35:11 Now,
35:11 in
35:12 Tajikistan at this time,
35:13 it was quite a hassle to actually start e-filing.
35:17 You had to submit several documents
35:19 um
35:20 to the tax office,
35:21 and they would review.
35:23 Over about 2 weeks,
35:24 and once it's approved,
35:25 they would come to your
35:27 premises and help to install it on your computer
35:30 and give you an e-token,
35:32 and that was the point at which you could start e-filing.
35:34 So what this intervention did was to create
35:37 um
35:38 was to have a firm,
35:39 a private firm,
35:40 facilitate this entire process for firms so that
35:43 they could actually start e-filing quite easily.
35:46 Um,
35:47 in thinking about our.
35:50 Um,
35:51 results,
35:51 we
35:53 expect that e-filing,
35:55 the impact will vary by the prior compliance behavior of firms.
35:59 And in terms of thinking about firms' prior compliance behavior,
36:02 we're using the risk profile score as a proxy for firms
36:05 that are more likely to collude or evade.
36:08 So,
36:09 We think about two types of firms.
36:11 One is those that are in active collusion with tax authority,
36:14 with tax officials,
36:16 where they like they like paper filing because the tax official helps them.
36:20 And the second set of firms are those that are not in this um cohesive equilibrium.
36:26 And we think that for both types of firms,
36:28 if they start to e-file,
36:29 their compliance costs reduce.
36:31 We think that the unofficial payments would also reduce,
36:34 although,
36:35 in one case,
36:36 it's the payments for collusion,
36:38 and in the other case,
36:39 it's coercive payments because we've heard a
36:41 lot of that anecdotally of coercive payments
36:43 um
36:44 obtained by tax officials.
36:46 When it comes to the tax payments,
36:48 we think that this could actually go
36:49 in different directions.
36:51 So for our
36:52 collusive firms,
36:53 if they no longer have the opportunity to collude,
36:56 then
36:57 their tax payments might increase.
36:59 On the other hand,
37:00 for the non-collusive firm for whom the tax official is really almost like
37:05 an enforcement agent.
37:06 And forcing enforcing them to pay more taxes
37:09 than maybe they actually wanted to.
37:11 So,
37:12 this is the example I had again earlier about saying,
37:14 you're not allowed to
37:16 pay this low level,
37:18 I want you to increase it.
37:20 For them,
37:20 we think that if you remove this tax official at the point of payment,
37:24 their um payments might actually go down.
37:28 I just wanted to say a little bit about our use of um
37:32 risk profile scores um
37:35 to proxy for baseline tax evasion.
37:38 This is a method that's used in many countries to target audits.
37:42 Essentially,
37:42 the tax authority will have a proprietary algorithm
37:45 based on
37:46 observed farm characteristics and prior audits that have been conducted
37:49 that will predict which types of firms
37:51 are more likely to be evading.
37:53 And so,
37:53 based on this prediction,
37:55 firms get different scores.
37:56 And in our setting,
37:57 And we do have evidence that this risk score
38:00 is predictive of tax compliance because we see,
38:02 for example,
38:03 that
38:03 even after controlling for um
38:05 observables,
38:07 like the number of employees,
38:08 the turnover,
38:09 um,
38:10 the firms that were that had higher risk scores
38:13 in the previous year,
38:14 in the coming year,
38:15 they actually pay
38:16 lower taxes.
38:18 So,
38:19 what do we find?
38:19 The first result is on the
38:22 e-filing adoption rates.
38:24 We find that the training
38:26 does not actually have a significant impact relative to the control group.
38:30 What
38:31 matters quite a bit is the logistical help,
38:33 helping firms actually start to e-file.
38:35 And this again goes to some points I had
38:38 earlier about the adoption costs on the side of um
38:41 taxpayers.
38:43 In terms of the impacts of e-filing,
38:45 we find that it reduces
38:47 the amount of time that firms spend
38:50 by about 5 hours each month.
38:52 So this is about 40% of the total time they spend on taxes.
38:55 And
38:56 if you want to quantify this,
38:58 just the savings from the compliance costs,
39:00 they compensate for the entire program after 7 months.
39:03 So if you take the wage of the accountant,
39:05 multiply that over 7 months,
39:07 the savings compensate for.
39:10 This,
39:11 uh,
39:11 Eylin.
39:12 What we also find is that bribes do go down,
39:15 but this is only among the risk,
39:17 low-risk score firms.
39:19 We don't observe this change among our high-risk of firms,
39:22 but as I'll show you in the next slide,
39:23 one thing we actually find,
39:25 actually,
39:25 not yet,
39:26 is that among the higher-risk score firms,
39:29 um,
39:29 many of them actually revert to e-filing.
39:31 They revert from e-filing to paper filing.
39:34 And so,
39:35 um,
39:35 we think that might be one of the reasons
39:36 why we don't see an impact on their compliance,
39:39 on their bribe payments.
39:43 In terms of the revenues,
39:44 we find that e-filing is actually helping to close this gap
39:50 that exists in the control group,
39:51 for instance,
39:52 between our low-risk firms that pay more taxes than the higher-risk ones.
39:55 Once people start e-filing,
39:57 this gap ends up being closed.
39:59 Because the lower,
40:00 the high-risk score firms,
40:02 their tax payments increase quite substantially.
40:05 For the lower-risk firms,
40:06 we don't see a significant
40:08 reduction,
40:09 but we do see that
40:10 in terms of how likely they have to pay zero tax,
40:13 uh,
40:13 we do find impacts on that.
40:17 Some of that evidence on thinking about high risk of firms is that one,
40:22 they're less likely to adopt e-filing,
40:24 suggesting that they do like
40:25 the paper filing and the interaction with tax officials,
40:28 and then they are also significantly more likely to drop out
40:32 over the
40:34 6,
40:34 after 6 months,
40:35 and
40:36 all the increase that we see in the tax revenues,
40:38 they dissipate once they
40:40 move back to paper filing.
40:43 And so what do we learn from this um case study?
40:48 One is that yes,
40:50 technology could really be great and not only in reducing compliance costs,
40:54 but also for increasing tax payments and reducing corruption.
40:58 And this type of
40:59 intervention is relevant to other technologies that
41:02 reduce the amount of in-person interactions with tax officials.
41:06 Another key lesson from the study is that when adoption is voluntary,
41:11 users may respond quite strategically,
41:13 and
41:14 the people for whom technology might
41:16 impose the most constraints
41:18 on their behavior,
41:19 they might actually be the least likely to adopt.
41:22 And so,
41:22 it's important for us to think about the heterogeneity
41:25 in users' capacity to adopt and how they might be
41:28 responding to uh minimize the impact of technology on them.
41:32 In the case of Tajikistan,
41:33 one thing they did was that
41:35 after
41:37 People,
41:37 once you start to e-file,
41:39 we've already known that
41:40 you have the capacity to do so,
41:42 then you're not allowed to go back to paper filing.
41:47 So,
41:47 now I'm going to um start wrapping up and to
41:52 Try to bring some of those lessons to think about
41:55 how we um do our policy engagements,
41:58 um,
41:58 with countries
41:59 and
42:00 some key considerations that we
42:03 could have in mind when we are thinking about tax digitalization projects.
42:07 And the first one is to really think
42:09 hard about whether technology is the binding constraint
42:14 in this given situation,
42:16 um,
42:17 or are there other parts,
42:18 other things like enforcement or the dimensions of capacity
42:22 that are maybe more pressing.
42:23 And this is a bit
42:25 difficult because I've been in many conversations with uh tax
42:30 leaders of tax agencies,
42:32 and there's a lot of excitement about oh if we just had this new
42:36 way of
42:37 capturing this information,
42:38 if we had a
42:39 perfect cadaster,
42:40 how great things would be.
42:44 But often,
42:45 one week,
42:46 one thing to look at is what has been
42:48 done with the information that's already collected to see
42:51 where the weakest link in the whole tax collection process might be.
42:55 And so,
42:56 it might be less exciting and more difficult to actually
43:00 work on like an HR reform or thinking about how to To change the
43:04 incentives of tax officials
43:06 than it is about just
43:07 having a big procurement and getting this new technology,
43:10 um intervention.
43:11 And so,
43:12 it's really um important to do what
43:15 one could think about like a diagnostic,
43:16 and I wanted to give a shout out to
43:19 some work
43:20 done by the World Bank MTI.
43:23 Uh,
43:23 group on innovations and tax compliance because
43:26 there,
43:26 they have created diagnostics for really thinking through
43:30 different aspects of tax compliance.
43:32 So,
43:32 what are the gaps in enforcement?
43:33 What are the gaps in facilitation?
43:35 Um,
43:36 what are the gaps in terms of tax morale
43:38 and the trust that people have in the government?
43:40 That I think is,
43:41 is core before we can
43:42 um really
43:44 be confident that any investments we're making are going to change the situation.
43:49 Second main point is to think about the appropriate technology
43:52 for the current level of development for the country.
43:54 It's
43:55 really not useful to get the most state of the art technology if you don't have
44:00 the right skills and um
44:03 uh capacity to use it effectively.
44:05 And so thinking about what makes the most sense
44:08 for the current context.
44:10 And one,
44:12 Thought to keep in mind is that,
44:13 you know,
44:14 um,
44:15 low-income,
44:15 like
44:16 countries in East Africa,
44:17 for example,
44:18 if you think about like mobile
44:20 payments,
44:21 um,
44:21 a lot of it is done on
44:23 not smartphones,
44:24 so USSD payments,
44:25 but
44:26 is extremely effective in,
44:28 in,
44:28 um,
44:29 transferring
44:30 um money
44:31 amongst people.
44:33 The third point here is to
44:34 think very carefully about the impact of technology on different
44:38 agents.
44:39 Uh,
44:39 like I said earlier,
44:40 they're winners and losers,
44:41 and to think about what type of strategic behavior we might anticipate.
44:44 Um,
44:45 it's important to be one step ahead,
44:47 to think,
44:48 OK,
44:49 who is going to probably going to try to sabotage this?
44:51 How might we create safeguards to prevent that?
44:58 And thinking more broadly,
44:59 um,
45:00 the,
45:01 with the use of IT systems,
45:03 the vast amounts of data that have been made available now to tax authorities.
45:06 And so one thing that I find very exciting is this potential that now is available for
45:12 um
45:13 our policy partners to
45:15 work closely with them in answering research and policy questions
45:19 that may then bring new insights that actually
45:21 help to also increase tax revenues in return.
45:24 And uh this could be you doing um
45:27 randomized experiments like the projects that I described,
45:29 but that doesn't have to be the case.
45:31 Many times,
45:32 data that already exists,
45:34 um,
45:34 could be used to analyze the impacts of previous projects
45:38 or previous policy interventions to understand
45:40 the impact.
45:42 And of course,
45:43 to be able to do all of this,
45:44 to use this type of data,
45:45 um,
45:46 while protecting taxpayers,
45:48 there needs to be a very strong regulatory framework,
45:51 um,
45:51 to,
45:51 on data sharing.
45:52 And this,
45:53 as we all know,
45:54 has been the focus of our most recent WDR
45:57 on,
45:57 um,
45:58 social,
45:59 and they have the idea of having a social contract,
46:01 um,
46:01 surrounding data use.
46:04 So,
46:05 um,
46:06 pulling this all together,
46:07 uh,
46:08 main takeaways are that,
46:10 yes,
46:10 governments can mobilize tax revenue and we can using technology and
46:14 we can think of this along the lines of using it to
46:17 define compliance,
46:18 that's to understand the tax base,
46:19 to facilitate,
46:20 uh,
46:21 compliance as well as to monitor it.
46:24 Uh,
46:26 technology is not a silver bullet.
46:27 We need
46:28 complementary investments in people and institutions
46:31 to reap the benefits,
46:32 and
46:33 any program must always anticipate how different
46:36 subgroups may be affected and how they would
46:38 likely respond.
46:40 And finally,
46:41 that the large amount of data that we're having now
46:43 creates new opportunities to use,
46:46 um,
46:48 This,
46:48 to have partnerships with tax authorities that will create
46:51 insights that would help policymaking and revenue generation.
46:55 So,
46:56 thank you all again for your attention and I will hand it over now.
47:03 Thanks Ayubola for that really clear,
47:05 uh,
47:06 what's that?
47:07 Techno
47:08 technology plant,
47:09 um,
47:10 technology flower.
47:11 um,
47:12 thanks,
47:12 that's really clear and,
47:13 and for somebody who doesn't work in this area,
47:15 really,
47:16 very clear and and really informative,
47:18 so thank you very much.
47:19 Um,
47:20 Kiara,
47:20 over to you if you're ready,
47:22 there you are.
47:23 OK,
47:23 great,
47:23 over to you,
47:24 Kiara.
47:28 for inviting us to,
47:30 to discuss today.
47:31 I must say I enjoyed it and I think Ayebola said very interesting things.
47:36 I'll get to them
47:37 in a little while.
47:39 I wanted to emphasize um
47:43 That,
47:43 uh,
47:44 coming from operations,
47:45 um,
47:46 first of all,
47:47 this type of uh applied research and looking at
47:50 the impact and evaluating the impact of certain,
47:53 um,
47:54 policy decisions and implementation of such policies.
47:57 I'm talking in general,
47:58 not just about the technology.
48:00 Um,
48:01 it's extremely valuable in terms of also being able to change the course
48:07 if we maintain a certain flexible
48:09 attitude
48:10 while we implement also operations,
48:12 and I'll
48:14 perhaps give you
48:15 a couple of examples if I get there,
48:16 at least one example.
48:19 Um,
48:21 I also say that we uh certainly value um the collaboration with DAC.
48:28 Uh,
48:28 also as a global practice where we work more analytical work.
48:32 Now,
48:32 AEBel has made reference to the
48:35 report that is going to be launched in January
48:39 on um innovation in tax compliance,
48:42 and that's a joint effort with the global
48:44 governance practice,
48:46 um,
48:47 and that tries to look at the problem of compliance from uh in different dimensions.
48:54 But I think having a collaboration with the research
48:57 group of the bank gives us the strength of also
49:01 both retroactively like Aebola was doing in a way,
49:04 testing a hypothesis that for those who are already in operations
49:08 may already guess,
49:09 but bringing the evidence and in a robust.
49:13 Uh,
49:13 way and rigorous way,
49:15 it's certainly helpful to think,
49:17 uh,
49:17 to push this thing forward
49:18 in operation.
49:19 Now,
49:20 coming to the potential of technology,
49:23 clearly,
49:24 and I don't know how many of you are really,
49:26 are aware of the
49:27 DRM approach that we,
49:29 the new DRM approach that was presented in June to the board,
49:33 to,
49:34 to the World Bank board.
49:36 Um,
49:36 the,
49:37 the objective is certainly to promote more revenue collection.
49:41 But in a much fairer
49:43 way,
49:44 so that would entail fighting tax evasion,
49:49 avoidance,
49:50 in other words,
49:51 broadening the base and
49:53 while trying to minimize the distortions in the system and to make the tax system
50:00 equitable.
50:01 So in that sense,
50:02 one could think that also
50:04 the digital transformation and technology can become an enabler
50:08 of this vision of helping to mobilize more resources
50:12 and in a fairer way.
50:14 And I think implicitly what uh Ayebola was testing was trying to
50:19 create a level playing field when you looked,
50:21 say,
50:22 at taxing property real estates,
50:26 um,
50:26 and how actually
50:27 um
50:29 Which is often also an indication of proxy,
50:32 a proxy of income
50:34 and how
50:36 Um,
50:37 sorry,
50:37 I don't know if Ryan is talking to me,
50:39 but anyway,
50:39 I hope you can hear me well.
50:41 Uh,
50:42 if you can't,
50:42 please help me.
50:44 What I'm trying to say is that,
50:45 uh,
50:46 by the,
50:46 the,
50:47 she's a tested hypothesis in the Liberia case
50:50 where we are trying to,
50:52 um,
50:53 in the spirit of broadening the base and trying to find ways to tax the wealth.
50:59 Uh,
51:00 more than
51:01 labor income or
51:03 more than those who do not own wealth.
51:06 Um.
51:08 Property,
51:09 real estate become an issue.
51:11 Certainly an interesting area to look at.
51:14 Um,
51:14 and certainly technology,
51:16 that's the hypothesis,
51:17 could help us to,
51:19 um,
51:20 assess and to improve the databases.
51:24 But then what she showed is that,
51:26 and I like that because one could consider what you presented today.
51:30 It's almost like a framework.
51:32 If I need to go in and do an operation on revenue mobilization using technology in a
51:37 low-income country,
51:39 your framework of looking at all the dimensions,
51:41 what are the databases,
51:43 your,
51:43 the tiles that you presented,
51:45 what is working,
51:46 what is not?
51:47 Do we have it?
51:48 Well,
51:48 the reality is that in most countries,
51:50 including Tajikistan,
51:51 Uh,
51:51 and Liberia,
51:53 those tiles are often missing,
51:56 or databases are incomplete
51:58 or,
51:59 um,
51:59 need to be built.
52:00 They're just simply non-existent.
52:03 So
52:04 this is the first caveat that I think that we also take away.
52:08 It's important to go in with a diagnostic basically.
52:12 And so what this research is helping us to
52:16 really understand what are the pieces of the puzzle
52:19 that we need to assess
52:21 before we think that the digital transformation will bring.
52:26 What I was trying to,
52:28 the objective,
52:29 which is mobilizing resources and in a fair way,
52:32 and also trying to tax those who can afford it more.
52:36 Um,
52:38 and I think what uh
52:39 you didn't say so explicitly,
52:42 but it was that actually only with your control groups you were able to
52:48 push people to be compliant.
52:50 Um,
52:51 but even within the control group,
52:53 a lot were still non-compliant.
52:55 There was a good percentage.
52:57 You could see an improvement,
52:58 but we were not reaching even 100% within
53:01 the
53:02 treatment group.
53:03 So clearly indicating that there are other pieces that are missing,
53:07 and I think it's exactly
53:09 the soft part,
53:10 which is not the
53:11 digital transformation,
53:12 but having to look at what is the governance environment in the country.
53:17 What is the political commitment and champions of,
53:20 um,
53:21 of,
53:21 of the government to really want to
53:24 To pay taxes,
53:25 and I think that's
53:27 all this resonates,
53:28 I think especially in colleagues that work on public sector reforms and taxation
53:34 as part of that and institutional strengthening,
53:37 it's exactly these pieces,
53:39 but I value deeply this research because,
53:44 as I was saying,
53:45 why we know it intuitively by working on the country and in the country context.
53:50 The research of having this analytical approach,
53:54 um,
53:55 I think first of all,
53:56 strengthen our ability to,
53:59 to have an analytical approach which is deeper
54:03 and broader.
54:04 And also um
54:06 in terms of er the policy dialogue.
54:12 The experience that I had,
54:14 uh,
54:14 many years ago in Tanzania was looking at,
54:16 uh,
54:18 service delivery,
54:18 and this is in a way what you're doing.
54:20 Ayebola is looking at,
54:22 she's assessing in Tajikistan and bringing the service closer to the taxpayers,
54:26 and that's the e-filing in a way.
54:28 So closing the gap between the tax collector
54:31 and the taxpayer and making it more efficient.
54:35 Um,
54:36 my experience was that even before reaching to that stage,
54:39 and I was looking down at the service delivery more in
54:43 education and health and how the government could actually assess,
54:47 uh,
54:48 even if the service was delivered,
54:50 not even the quality of the service,
54:51 but whether the service was reaching.
54:54 The beneficiaries.
54:56 Um,
54:56 indeed,
54:57 you need,
54:58 technology was certainly a mean,
55:00 uh,
55:01 to close certain gaps and distances and also geographical spaces,
55:05 and that's e-filing would do the same,
55:08 but,
55:08 uh,
55:08 the,
55:09 the incentive to comply
55:11 was very limited and the more we were digging.
55:15 Um,
55:16 basically,
55:16 you bring transparency,
55:18 and I think that's what you were also
55:20 bringing through the research,
55:21 certainly on Liberia.
55:23 Transparency means that certain groups that have strong interests,
55:29 are,
55:29 are becoming evident.
55:30 And so,
55:31 and so this is the other dimension when I talk about the governance environment.
55:36 Um,
55:37 so we need to,
55:38 so technology helps,
55:40 but it's,
55:41 um,
55:43 But it needs to be applied uh with a,
55:46 with a framework.
55:47 And so
55:48 I like this approach because I think
55:50 that you're looking at different dimensions and
55:53 you also try to bring in in your evaluations also
55:57 the behavioral side
55:59 of and responses of the heterogeneity of the groups.
56:03 I think moving forward,
56:05 I think in the case of Tangiskison,
56:07 there are two things.
56:08 One is that clearly,
56:11 In a way,
56:11 implicitly you highlighted that it was not a strong political commitment.
56:15 It's only
56:16 at the end that they decided if people opt to do the e-filing,
56:20 they're not allowed to go back to paper filing.
56:24 But strong groups of risk lovers.
56:28 Basically have tested the system and found that they can still get away.
56:33 So how do we close that part?
56:36 The other one that the technology is now capturing,
56:39 and I don't think is your example,
56:41 is not,
56:41 is your research,
56:42 not at all,
56:43 is the fact that you were able to do the research
56:45 with small and medium sized enterprises that were in the system.
56:49 So,
56:49 we were not really capturing the formality through
56:52 this because we were looking at corporate income taxation
56:55 and
56:56 companies that are formally registered at above a certain size.
57:01 But the reality is that in low-income country,
57:03 actually informality is really large and so
57:06 what we need,
57:07 what technology could help us if we could fill in some
57:10 of the databases that you were showing in your earlier chart.
57:14 It's precisely to enable also with a lot of training
57:17 as well and assuming that there is political commitment.
57:21 Um,
57:22 those databases that indirectly can help us to
57:27 Understand who's the beneficiary owner of the small businesses,
57:30 where are the small businesses operating.
57:33 And why they are so small that they prefer to stay implicitly,
57:37 sorry,
57:38 they prefer to stay in the informality and
57:43 Perceived as invisible,
57:45 but actually visible to a certain
57:47 um
57:49 And I'm thinking of Tajikistan because I spent 3 years there
57:53 as a country manager.
57:56 What it will be interesting to analyze,
57:58 what is it that is preventing them from moving to from informality
58:04 where they are actually harassed.
58:08 But in a different way,
58:10 and moving to
58:11 the formality
58:13 and uh sorry,
58:14 this is an important question because
58:17 what's happening in the,
58:18 in Tajikistan is that you have,
58:20 we have a lot of small
58:22 Businesses,
58:23 traders.
58:25 They tend to be small,
58:26 but when they go to the market or to the places where they need to sell their goods.
58:32 They are
58:33 asked to pay fees.
58:36 Those are taxes in many ways one could consider them.
58:38 It's just that they're often they're not going into their treasury accounts,
58:43 um,
58:43 but they're pocketed by small groups or by local authorities,
58:48 but without any transparency whatsoever on the use of these funds.
58:53 And
58:54 those fees have to be paid even if that day
58:57 there was no trade
58:59 in every single day at the market,
59:01 so it's very different from a tax system
59:03 where we would only pay when there is
59:06 income
59:07 to declare
59:08 or a tax base to declare.
59:11 And yet
59:12 they're not moving from informality to formality,
59:15 and I think it would be very interesting to understand
59:18 over time,
59:19 assuming that this e-filing were you know,
59:23 become part of the
59:26 tax administration,
59:27 rather than there would be an incentive.
59:29 For the informal operators in the informal sector moving to the formal sector,
59:34 I hope
59:35 I'm not sure I'm clear.
59:36 I made it clear myself,
59:37 but
59:38 in this heterogeneity we're missing one part,
59:41 and,
59:41 and just by coincidence Ibola done this exercise
59:44 in Tajikistan that I knew very well,
59:47 and I,
59:48 I was wondering what is it that would make them go and
59:51 declare and use that system that would allow them to pay only.
59:55 On the income
59:57 earned,
59:58 not always paying this,
1:00:00 uh,
1:00:00 you know,
1:00:01 fees that are actually completely delinked from their income.
1:00:06 Anyway,
1:00:06 I think
1:00:07 this is just an illustration of potential
1:00:10 uh research that we can continue to carry out also jointly,
1:00:15 um,
1:00:16 and that can also help us to not only to address the issues of revenue,
1:00:20 but also I think more issues
1:00:22 of redistribution and transparency and accountability,
1:00:26 but also to the benefits of the taxpayer,
1:00:28 not only to the tax administration.
1:00:31 And with this,
1:00:32 uh,
1:00:32 Deon,
1:00:33 I want to thank you again and look forward to our collaboration
1:00:37 and also we are working together on aA
1:00:41 conference on a yearly basis,
1:00:43 so I look forward to the next one next year.
1:00:45 Over to you.
1:00:46 Thank you.
1:00:48 Thanks,
1:00:48 Kiara.
1:00:49 Yeah,
1:00:49 thank you for,
1:00:49 we,
1:00:50 we had a very successful tax conference this past year
1:00:52 and look forward to next year's one as well.
1:00:54 Um,
1:00:55 Oyebola,
1:00:56 before opening it up,
1:00:57 I,
1:00:57 I want to give you a chance to maybe respond to some of the issues that,
1:01:00 that Kiara raised,
1:01:01 if you just want to respond or even elaborate on them.
1:01:07 Thank you so much,
1:01:08 uh,
1:01:08 Kiera for these great comments.
1:01:10 Uh,
1:01:11 I'm glad to know that this resonates with you,
1:01:13 and yes,
1:01:14 it was very interesting to learn that you did spend time in Tajikistan
1:01:18 as country manager.
1:01:19 So firsthand,
1:01:20 um,
1:01:21 knowledge of the context.
1:01:23 The few of the points,
1:01:24 a few points you made that I just wanted to touch on,
1:01:27 which I,
1:01:27 I mean,
1:01:27 they're all
1:01:29 very spot on.
1:01:30 And the first one,
1:01:30 I was thinking about,
1:01:32 what are the other things.
1:01:33 So I,
1:01:34 beyond technology,
1:01:35 beyond enforcement,
1:01:37 um,
1:01:38 what else,
1:01:38 what is
1:01:39 keeping us at only about 30% and keeping us from going all the way up to like 100%
1:01:44 compliance?
1:01:45 And I feel like this is
1:01:47 such an important question,
1:01:48 um,
1:01:49 in developing countries and
1:01:52 A few of the things,
1:01:52 one you already mentioned was about political um
1:01:56 resistance,
1:01:57 uh,
1:01:57 but some other things are around things of just like.
1:02:01 The norm that is built.
1:02:03 So if
1:02:04 it's just,
1:02:04 I know that very few people around me are
1:02:07 paying,
1:02:07 then I'm less likely to actually want to pay and just the overall sense
1:02:12 that I was indicating in one of the earlier slides about
1:02:15 this idea of a social contract between government and citizens,
1:02:18 and
1:02:18 there's only so far that
1:02:20 one can go with um
1:02:23 enforcement in and of itself,
1:02:26 until it's like an accepted
1:02:29 part of
1:02:30 the interaction that the citizens accept that this is what
1:02:33 I pay as being a part of the society that is providing
1:02:37 these different types of public goods and services,
1:02:40 um,
1:02:41 we will by no means reach higher levels of,
1:02:44 um,
1:02:46 Tax compliance.
1:02:47 That's,
1:02:47 that's my take on it.
1:02:48 And
1:02:49 there's,
1:02:49 um,
1:02:50 that's why one of the things that I like a lot about the
1:02:53 enforcement,
1:02:54 facilization,
1:02:54 and trust framework that the innovations and tax compliance work is doing because
1:02:58 they pay a lot of attention to the sense of
1:03:01 what does it take for
1:03:03 taxpayers to actually have trust in the system and believe that
1:03:07 them contributing to it is a meaningful thing to do.
1:03:10 Um,
1:03:12 The other
1:03:13 point I wanted to respond to was thinking about
1:03:15 how technology might help to bridge this gap between people
1:03:19 being in the informal sector and what it might take for them to formalize.
1:03:23 And I think that there are many factors that affect this decision of firms to
1:03:28 go from the
1:03:30 Cross from the informal sector to the formal,
1:03:32 but at the core of it,
1:03:33 one could think of it as them doing a
1:03:35 cost-benefit analysis.
1:03:37 What are the costs of
1:03:38 me remaining in the
1:03:40 informal sector?
1:03:41 And some of them you listed quite well,
1:03:43 like this harassment and there's multiple fees that they need to pay.
1:03:47 And
1:03:48 how does that change if they move to the
1:03:51 um formal sector?
1:03:52 I think that there are also
1:03:54 other things that there may be other costs that firms perceive that keep them
1:03:59 from moving.
1:04:00 So,
1:04:00 it could be that
1:04:01 it's not only just the,
1:04:03 um,
1:04:04 corporate income
1:04:06 taxes that maybe they might be thinking about,
1:04:08 but then there's Social Security contributions.
1:04:11 There are many other
1:04:12 types of um responsibilities
1:04:15 and expectations that come
1:04:17 with moving into the formal sector.
1:04:19 And so I imagine that for most firms,
1:04:21 they are
1:04:21 um
1:04:23 conducting this analysis and seeing which
1:04:25 um regime they prefer to be in.
1:04:27 And the other point is that often,
1:04:29 even for firms that may be
1:04:31 indifferent,
1:04:32 the costs of formalizing could also be quite substantial.
1:04:35 And so,
1:04:36 there's um
1:04:38 Work and there's different research that's been done on
1:04:40 how the firms respond when you help them to try to
1:04:43 overcome these um
1:04:45 formalization costs.
1:04:46 So,
1:04:47 that's one thing that potentially,
1:04:49 if these costs,
1:04:50 if it's a more automated
1:04:52 uh
1:04:52 process,
1:04:52 many countries now are doing things that they call a one-stop shop,
1:04:55 so that I,
1:04:56 I no longer have to go to like 6 different offices to formalize my business.
1:04:59 Um,
1:04:59 that's one of the
1:05:01 um opportunities I think that technology might help in this
1:05:03 formalization
1:05:05 journey.
1:05:05 And I thought it was also important to
1:05:07 Uh,
1:05:08 put out there that often when we're talking about formalization,
1:05:11 people
1:05:12 are thinking about,
1:05:13 you know,
1:05:14 the markets woman that's selling,
1:05:16 selling tomatoes.
1:05:17 Um,
1:05:18 that's,
1:05:19 that's really not where your revenues are going to come from.
1:05:22 One,
1:05:24 When I think of this informal
1:05:26 sector,
1:05:27 um,
1:05:27 one of the more promising,
1:05:29 um,
1:05:30 dimensions of these uh professionals.
1:05:32 So there's a nice work by Daiso Gimbo that looks at
1:05:36 um taxing people like accountants,
1:05:37 doctors,
1:05:38 officers,
1:05:38 dentists.
1:05:39 These are people who
1:05:41 many times will fly under the tax radar in developing countries,
1:05:44 but actually could be
1:05:46 important sources of revenue if they are
1:05:49 made to
1:05:50 join the tax base.
1:05:51 So I will stop it there,
1:05:53 but thanks again,
1:05:53 those are really great comments,
1:05:55 um,
1:05:55 and
1:05:56 very,
1:05:56 um,
1:05:57 just spot on insights from your time in the field.
1:06:02 Yeah,
1:06:03 thanks both.
1:06:03 Um,
1:06:04 and reminder if you,
1:06:05 if you have any questions,
1:06:06 please raise your hand or just,
1:06:07 uh,
1:06:07 drop a line in the chat.
1:06:08 Um,
1:06:09 since there aren't any questions right there,
1:06:10 I'm,
1:06:11 I'm gonna take the opportunity to pose some
1:06:13 questions that occurred to me during the talk.
1:06:14 So,
1:06:15 um,
1:06:16 One is,
1:06:17 um,
1:06:17 it was interesting
1:06:18 how you described the Liberia case where they
1:06:22 sort of combined,
1:06:23 um,
1:06:24 you know,
1:06:24 the youth employment
1:06:25 with the tax where they use the young people to sort of plot out the,
1:06:30 the,
1:06:30 the real estate.
1:06:31 Um,
1:06:32 and you mentioned in passing that then they had taken that to scale.
1:06:37 It seemed to me that that might open up some real opportunities for,
1:06:40 for collusion or something,
1:06:41 you know,
1:06:41 if you go to scale,
1:06:43 uh,
1:06:43 and you have a lot of young people,
1:06:45 um,
1:06:46 Uh,
1:06:48 Throughout there,
1:06:49 there's potential collusion there.
1:06:51 So that,
1:06:51 that just triggered something in me.
1:06:53 I,
1:06:53 I wonder if it's difficult,
1:06:55 you know,
1:06:56 at that level,
1:06:56 you might want drones or you might want
1:06:58 satellite imagery because you need that distance.
1:07:01 Anyway,
1:07:01 I see there's a couple of questions,
1:07:02 so
1:07:03 let me turn it over to people who have their hand up.
1:07:06 I see,
1:07:07 um,
1:07:07 Rajiv,
1:07:08 you've raised your hand,
1:07:10 and then
1:07:11 I see
1:07:13 Bob has raised his hand.
1:07:16 And
1:07:19 Um,
1:07:21 Uh,
1:07:21 Javi,
1:07:21 so let's go in that order.
1:07:23 So,
1:07:23 Rajiv,
1:07:24 Javi,
1:07:24 and,
1:07:25 oh sorry,
1:07:25 Rajiv,
1:07:26 yeah,
1:07:26 Javi and Bob,
1:07:27 sorry,
1:07:27 that's the order that hands were raised.
1:07:29 So,
1:07:30 so Rajiv over 2.
1:07:34 Thank you,
1:07:34 John,
1:07:34 and uh
1:07:35 thank you,
1:07:36 Ebola for an excellent presentation.
1:07:39 I think you have raised,
1:07:41 you know,
1:07:42 issues that are very important
1:07:43 for the tax administrations of uh
1:07:46 low income countries and middle income countries both
1:07:49 because technology is one factor,
1:07:51 you know,
1:07:51 which can really
1:07:54 remove the barriers as far as the voluntary compliance of the tax base is concerned.
1:07:58 So it's,
1:07:58 it's a very
1:08:00 important administrative tool that I think with
1:08:02 the experience of the countries who have
1:08:04 adopted electronic filing of returns and electronic processing of returns.
1:08:09 They have realized that it has led to an enormous growth in the
1:08:13 taxpayers and the number of taxpayers who file the returns.
1:08:16 Uh,
1:08:16 but just a couple of points that I wanted to make were that
1:08:20 while technology is extremely important,
1:08:22 As far as the administrative issues are concerned,
1:08:25 the role of technology and the political issues as to how the
1:08:30 evasive behavior of the taxpayers can be changed,
1:08:33 how the information that flows into the
1:08:37 tax administration through technological means,
1:08:39 that can be the power of that information can be harnessed by the
1:08:42 by the policymakers.
1:08:44 That issue is something that,
1:08:45 you know.
1:08:46 Needs to be addressed because
1:08:49 what happens is that technology,
1:08:50 which is an important means,
1:08:52 becomes an end in itself and that should not happen because ultimately,
1:08:55 as far as the tax administration is concerned,
1:08:58 the important goal is to raise revenue and mobilize
1:09:01 revenue.
1:09:02 So if If the
1:09:04 governance mechanisms
1:09:05 are only
1:09:07 helping
1:09:08 some taxpayers but not the taxpayers
1:09:11 who probably would with some small nudge would be able to come into the tax base,
1:09:15 then there is there are concerns,
1:09:18 so.
1:09:20 Uh,
1:09:20 so technology is,
1:09:22 should be seen as a means and not an end.
1:09:24 And I also want to raise an important question of
1:09:28 whether,
1:09:29 you know,
1:09:30 technology is an answer to all the questions because,
1:09:33 you know,
1:09:33 sometimes the human touch is an important,
1:09:36 is an important element because many times what happens
1:09:38 is that the tax administration of the countries,
1:09:41 uh
1:09:42 by,
1:09:43 you know,
1:09:44 engaging with
1:09:45 the taxpayers and looking at the tax records and
1:09:49 Uh,
1:09:49 doing audits and
1:09:51 so they gain a lot of experience of
1:09:54 how the tax is being evaded
1:09:56 and sometimes
1:09:58 when we,
1:09:59 uh,
1:10:00 you know,
1:10:00 look at technology only as a,
1:10:02 as an end,
1:10:03 these skills,
1:10:04 these human skills,
1:10:05 and that
1:10:06 that human touch starts,
1:10:07 you know,
1:10:07 vanishing,
1:10:08 which is,
1:10:09 which is important to realize that we should not
1:10:12 let that human intelligence and human touch in
1:10:15 these this entire administrative processes to vanish.
1:10:19 So,
1:10:19 those are two important points.
1:10:20 Thank you.
1:10:22 Thanks,
1:10:22 Rajiv.
1:10:22 Um,
1:10:23 I see Ishani's also raised her hand,
1:10:24 so let's just go with Bob this round,
1:10:26 and then I'll give Ayebola a chance to respond,
1:10:29 and then we'll go to uh,
1:10:30 uh,
1:10:30 sorry,
1:10:31 let's go to Shavi
1:10:33 and then give Ayebola a chance to respond and then go to Bob and Ishani afterwards.
1:10:37 That's OK.
1:10:38 So Shabi,
1:10:38 over to you.
1:10:40 All right,
1:10:40 um,
1:10:41 so,
1:10:42 uh,
1:10:42 Olibola,
1:10:43 thanks a lot for the,
1:10:43 uh,
1:10:44 the presentation.
1:10:44 This was very,
1:10:45 very,
1:10:46 very nice.
1:10:46 So I,
1:10:47 I also had a question on,
1:10:48 on Liberia.
1:10:49 And so
1:10:50 I mean,
1:10:51 you know,
1:10:51 if I'm a,
1:10:52 you know,
1:10:52 a resident,
1:10:53 and I,
1:10:54 you know,
1:10:54 so I only see that down,
1:10:56 which is,
1:10:57 you know,
1:10:57 I,
1:10:57 I,
1:10:57 you know,
1:10:58 I,
1:10:58 I
1:10:59 basically,
1:10:59 I gotta fork out money from,
1:11:00 from my pocket.
1:11:02 So is there any way that,
1:11:03 that these taxes are then linked to
1:11:06 Uh,
1:11:06 improvements in,
1:11:07 in,
1:11:07 you know,
1:11:08 street lighting or garbage collection or,
1:11:11 or what have you.
1:11:12 So,
1:11:12 so in other words,
1:11:12 I mean,
1:11:13 do they see the benefits of,
1:11:15 of their taxes,
1:11:16 uh,
1:11:16 at work,
1:11:17 and,
1:11:17 and is,
1:11:18 you know,
1:11:18 is there a way that,
1:11:19 that,
1:11:19 you know,
1:11:20 this could be,
1:11:20 uh,
1:11:21 emphasized,
1:11:21 uh,
1:11:22 so,
1:11:22 you know,
1:11:23 to improve,
1:11:23 to improve,
1:11:24 uh,
1:11:24 enforcement of,
1:11:25 of,
1:11:25 of,
1:11:25 of tax collection?
1:11:27 Um,
1:11:27 that was it.
1:11:28 Thank you.
1:11:29 Thanks.
1:11:30 Ebola back to you.
1:11:32 OK.
1:11:33 Thank you for these um great set of questions.
1:11:35 So,
1:11:35 Dianne,
1:11:36 I'll start with yours about uh these young people and whether
1:11:39 they could end up being like compromised agents as well.
1:11:42 And I think that's um it's very possible that some of that is going on.
1:11:46 So,
1:11:47 um,
1:11:47 one of the
1:11:49 reasons that they went the route of using young people
1:11:52 was for some of these um political economic considerations that we
1:11:56 were discussing where
1:11:57 they actually try to find,
1:12:00 um,
1:12:01 People
1:12:02 from different communities.
1:12:04 And so the whole idea of having a government job,
1:12:07 uh being given to someone in your,
1:12:09 in your,
1:12:10 from your neighborhood was
1:12:11 a very,
1:12:12 uh,
1:12:12 they thought that was part of the benefits that people are getting.
1:12:15 And they also believed that these people would know the communities
1:12:19 a lot better
1:12:20 and so
1:12:21 they would be able to win the trust of people to get the information that they needed,
1:12:25 the correct information for registration.
1:12:28 And so
1:12:29 even if you use drones,
1:12:31 typically that would only get you started and
1:12:33 then you still have to do some ground truthing
1:12:35 to complement that information.
1:12:37 So this was just cutting out that one step and they did use,
1:12:40 you know,
1:12:41 some satellite imagery.
1:12:42 There was a very
1:12:44 resourceful person who was very good with GIS in the office that would Pretty much
1:12:49 make out maps for people and give it to them.
1:12:52 They will take them to the field and say,
1:12:53 OK,
1:12:54 these are all the properties in this block that you've been assigned,
1:12:57 and you need to come back with information
1:12:59 on them.
1:12:59 So,
1:12:59 it was a bit of a mix and match between the
1:13:02 um
1:13:03 kind of the more um satellite imagery method,
1:13:06 as well as the boots on the ground
1:13:08 method.
1:13:10 And what I did just from going on the field sometimes with these people,
1:13:12 what I did see was
1:13:14 At first I was worried that maybe people would be hostile to them,
1:13:18 like,
1:13:18 oh,
1:13:19 you are on the side of the government
1:13:21 trying to come to tax me,
1:13:22 but many of the interactions that interactions that I witnessed were
1:13:26 quite positive and pleasant because
1:13:28 they saw these people as being
1:13:30 from the community
1:13:31 and coming to
1:13:33 to do this work.
1:13:34 So
1:13:34 it's,
1:13:35 I mean,
1:13:35 am I saying that no one,
1:13:36 like,
1:13:37 oh,
1:13:37 I'm just going to skip over your house because I know you.
1:13:40 Um,
1:13:40 from church,
1:13:41 you know,
1:13:41 that probably happened quite a bit,
1:13:43 but the
1:13:44 ideal would be to have checks and balances in place such that
1:13:47 any information is verified through,
1:13:50 um,
1:13:50 um,
1:13:51 other sources.
1:13:53 The second question
1:13:55 from Rajiv,
1:13:56 this is a very um great comments about technology.
1:13:58 I love that phrase,
1:13:59 not being an end in itself.
1:14:01 And I think that
1:14:02 sometimes the excitement could almost make it seem that um
1:14:06 as long as we can get this new um
1:14:08 ITAS system or this new software,
1:14:11 all the problems would be solved,
1:14:13 but no,
1:14:15 in fact,
1:14:15 I,
1:14:16 your point about
1:14:17 being careful not to lose the human touch
1:14:20 and the human intelligence,
1:14:22 I feel like that's,
1:14:23 that's really valid and it's a bit of we see a little bit of that in the
1:14:28 Tajikistan paper,
1:14:29 because like I was
1:14:31 saying,
1:14:31 we do find
1:14:32 that among these
1:14:34 um
1:14:35 lower-risk groups for whom the tax officer
1:14:37 maybe previously was serving as a deterrent,
1:14:40 they actually we find them just start
1:14:43 reporting lower taxes.
1:14:44 Um,
1:14:45 for instance,
1:14:45 the likelihood that they report having zero
1:14:50 amount of tax liability in a month,
1:14:52 that increases.
1:14:53 And so,
1:14:53 there is that sense that people can
1:14:57 Almost like hide behind technology,
1:14:59 they no longer have a human interaction and
1:15:02 again going on the field there
1:15:04 properties that I maybe we visited like that I go into,
1:15:08 now I'm going to Liberia.
1:15:09 And I have people telling me like,
1:15:11 oh,
1:15:11 yes,
1:15:11 we went to this property in 2018,
1:15:13 and they remember it,
1:15:14 and there's just ways that people
1:15:16 are able to remember things on the field,
1:15:18 if they're from that context that
1:15:20 are quite,
1:15:21 uh,
1:15:22 you know,
1:15:22 are quite remarkable,
1:15:23 and the ideal would be to find a way
1:15:26 to
1:15:26 marry these two,
1:15:27 to use these human knowledge
1:15:30 to the extent possible while
1:15:33 controlling for the negative parts that could come from the human interactions.
1:15:37 Um,
1:15:37 Javi,
1:15:38 on your question about
1:15:40 how these taxes are used,
1:15:42 so that's one thing that's unique about Liberia.
1:15:44 Most other places,
1:15:47 property taxes are local.
1:15:48 So real estate tax is used for roads,
1:15:50 for lighting,
1:15:51 for
1:15:52 uh
1:15:54 local public goods,
1:15:55 but in Liberia,
1:15:55 it all goes to the treasury.
1:15:56 And so,
1:15:58 there is no
1:15:59 direct um
1:16:01 link between the taxes that I pay in my community
1:16:04 and the,
1:16:05 um,
1:16:06 and the types of infrastructure of public goods that we get.
1:16:10 That said,
1:16:11 on the broader level,
1:16:12 thinking about taxes in general,
1:16:14 the government has been doing a lot
1:16:16 in the,
1:16:16 in recent years to try to emphasize this link between
1:16:20 tax
1:16:20 revenues and property and um
1:16:23 and public goods.
1:16:24 So,
1:16:24 for instance,
1:16:24 there's certain roads where you see a big sign board that says your taxes at work,
1:16:28 LRA and they,
1:16:30 they're trying to message
1:16:31 to show things that
1:16:33 taxes are doing.
1:16:34 But
1:16:34 I think that
1:16:36 On the whole,
1:16:37 this does not seem to me to be very,
1:16:39 um,
1:16:41 very
1:16:42 core motivation.
1:16:43 It's not
1:16:44 something that people are very conscious of,
1:16:46 especially when it now comes to the
1:16:48 um real estate taxes that are only centrally collected.
1:16:56 Thanks,
1:16:56 Ayebola.
1:16:57 Um,
1:16:57 we have 5 minutes left,
1:16:58 so let me turn to the two people who've raised their hands,
1:17:01 Bob and then Ishani,
1:17:03 and then I propose we,
1:17:04 we give Ayebola a chance and maybe Kira,
1:17:06 if you want to come in at the very end.
1:17:09 As comment,
1:17:09 but then we'll wrap up after that.
1:17:10 Thanks,
1:17:11 Bob.
1:17:12 Hi.
1:17:12 Yes,
1:17:13 thanks for this very interesting talk.
1:17:14 One thing that struck me about like your research projects is that they
1:17:17 seem to
1:17:18 require uh
1:17:19 quite a long gestation period and like quite heavy ex ante investment
1:17:24 in terms of
1:17:25 gaining the,
1:17:25 the trust and respect of
1:17:28 Tax collection,
1:17:29 uh,
1:17:30 agencies.
1:17:31 So I was wondering whether you
1:17:33 could sort of elaborate a bit on sort of what that means
1:17:36 for the bank's operational engagement.
1:17:38 So what are the do's and don'ts that follow sort of from
1:17:42 your work?
1:17:42 And,
1:17:43 you know,
1:17:43 is it reasonable
1:17:44 to sort of,
1:17:45 uh,
1:17:45 expect to be able to replicate
1:17:47 sort of your,
1:17:48 your research
1:17:50 very broadly,
1:17:51 or is this sort of a boutique niche?
1:17:53 type of intervention that requires a lot of
1:17:56 uh
1:17:56 dedicated expertise and
1:17:58 resources.
1:18:00 Thanks.
1:18:05 I Obola,
1:18:05 thanks for the
1:18:06 insightful presentation.
1:18:07 Um,
1:18:07 so I was just wondering if you could talk a little bit about,
1:18:11 um,
1:18:12 sort of the,
1:18:12 the
1:18:13 estimation of,
1:18:14 of costs and benefits,
1:18:15 right?
1:18:15 So you,
1:18:16 you touched upon that a little bit.
1:18:17 But,
1:18:18 so for example,
1:18:19 related to Chavi's point
1:18:20 about e-filing,
1:18:22 for instance,
1:18:22 there might be
1:18:24 benefits that
1:18:25 accrue over a longer time horizon.
1:18:28 And so it sort of struck me that,
1:18:29 you know,
1:18:29 the estimate that you provided was sort of on the lower end of things,
1:18:32 and I wonder if there are these sort of bigger picture
1:18:35 um benefits that might accrue
1:18:37 over time that we don't capture and then we sort of
1:18:38 walk away thinking this is a relatively small intervention with,
1:18:41 you know,
1:18:41 modest costs and modest benefits,
1:18:43 but really,
1:18:44 you know,
1:18:44 it has the potential to be
1:18:46 a much bigger thing.
1:18:50 Thank you.
1:18:51 Um,
1:18:51 so I'll start with Bob.
1:18:53 Thanks,
1:18:53 Bob,
1:18:54 this is,
1:18:54 you're exactly right.
1:18:55 These things take a very long time to,
1:18:58 to germinate,
1:18:59 so
1:19:00 I think
1:19:01 the first,
1:19:02 like,
1:19:02 one of the um,
1:19:05 Sources I ref I referred to
1:19:07 was an analysis that I've done in Liberia in 2013.
1:19:10 So that was the first engagement with them,
1:19:12 and it was essentially just looking at all
1:19:14 the tax data that they had on property taxes
1:19:17 and working through a process of thinking,
1:19:19 where are the gaps,
1:19:20 you know,
1:19:20 what's
1:19:21 What is the coverage on the
1:19:23 properties in the tax base?
1:19:24 What's the
1:19:25 rate at which people are paying
1:19:27 uh from year to year?
1:19:28 How,
1:19:29 you know,
1:19:29 just trying to like map out
1:19:31 this whole um
1:19:33 process of revenues and
1:19:35 where those gaps are and
1:19:37 uh
1:19:38 created,
1:19:38 this is work sponsored by IGC I was still a PhD student at the time,
1:19:42 and
1:19:43 created this whole document that was essentially like a guide to
1:19:47 the um
1:19:48 collection of real estate taxes in Liberia.
1:19:50 And
1:19:52 During from the time I've been there,
1:19:53 there've been maybe like
1:19:55 4 or 5 different
1:19:57 managers of the real estate department,
1:19:59 and what they've told me is that they tend to go to this work that I did for them way back
1:20:04 to give them like once they join,
1:20:05 it's a great orientation into what are the issues,
1:20:08 what are the leakages,
1:20:09 what are the concerns
1:20:11 in this.
1:20:11 And so
1:20:12 I think that
1:20:13 Providing
1:20:14 things that are valuable to the tax authority early on in the relationship
1:20:18 is really crucial
1:20:20 in
1:20:20 building these um
1:20:22 long-term relationships.
1:20:23 And so I found that
1:20:26 at different intervals,
1:20:27 it's not until now in 2021,
1:20:29 where a research paper
1:20:30 is coming out that they didn't have to wait
1:20:33 many years to
1:20:34 get benefits,
1:20:35 um,
1:20:36 all along the process.
1:20:37 So from
1:20:38 Training people in doing this field,
1:20:40 um,
1:20:41 enumeration,
1:20:42 and then working with them also when they
1:20:44 expanded it,
1:20:45 they've been very practical ways and so I feel that
1:20:48 trying to be for me,
1:20:49 I always think
1:20:50 about trying to be as useful as possible
1:20:53 to the um tax authorities,
1:20:55 especially early on
1:20:56 and I know that we do this a lot,
1:20:58 for instance,
1:20:58 through like the ASC and analytical work.
1:21:00 Now thinking about
1:21:02 the research
1:21:04 there I think the
1:21:05 um.
1:21:06 The,
1:21:07 uh,
1:21:08 these things take a lot of time,
1:21:09 um,
1:21:10 doing different surveys,
1:21:11 different fieldwork,
1:21:12 and in fact,
1:21:13 um,
1:21:14 I,
1:21:16 In the in the Tajikistan project,
1:21:19 I was just chuckling to myself that probably all the World Bank
1:21:22 team that worked on it have forgotten they ever did this project
1:21:25 because it's so many years later.
1:21:27 And so one thing I think that could help in these interactions is that
1:21:31 when we do have World Banks.
1:21:33 that are very time-bound,
1:21:35 we could have
1:21:37 partnerships with researchers who have the ability to spend
1:21:40 much longer on a project such that once the,
1:21:43 um,
1:21:44 immediate deliverables
1:21:45 are
1:21:46 done,
1:21:47 these,
1:21:47 um,
1:21:47 research learnings could continue
1:21:49 over,
1:21:50 over time.
1:21:51 So that's one way which I think we could,
1:21:54 um,
1:21:54 partner together to
1:21:56 be able to get
1:21:58 research,
1:21:58 um,
1:21:59 findings that do take a longer time,
1:22:01 given the fact that many bank projects
1:22:03 are time-bound.
1:22:04 Um,
1:22:05 have much shorter time horizons.
1:22:08 And to Ishani's point,
1:22:10 no,
1:22:10 you're exactly right that
1:22:11 it was really the lower bound that I give that if we only thought about the
1:22:15 savings,
1:22:15 the compliance costs,
1:22:16 but
1:22:17 there are many things that we are unable to quantify directly in this context about
1:22:22 the overall efficiency of the tax system.
1:22:24 So
1:22:25 before they used to have people type up so people submit in person,
1:22:28 then you have to type up all these returns.
1:22:30 There could be mistakes,
1:22:31 and then it's months or weeks until this data could actually be useful for
1:22:35 any type of analysis and um policy decisions.
1:22:38 So,
1:22:39 um,
1:22:40 thinking more broadly about the overall
1:22:43 improvement and efficiency
1:22:44 of the tax system
1:22:46 is,
1:22:46 uh,
1:22:47 I think that that's really where the big gains are.
1:22:49 And
1:22:50 the facts,
1:22:50 even like Kierra is saying that
1:22:53 having a more efficient
1:22:55 process might then be more attractive to people outside the tax net to come in.
1:22:58 So one could think of so many dimensions in which
1:23:01 e-filing might actually be improving overall
1:23:04 efficiency of the system,
1:23:06 but I'm not able to put numbers on those from the specific
1:23:10 project that we did.
1:23:12 And,
1:23:13 um,
1:23:13 Yan,
1:23:14 did you want me to just go ahead and
1:23:16 conclude,
1:23:16 or what,
1:23:17 what did you want?
1:23:18 I think we're a little over time,
1:23:19 so I think we maybe we should stop there,
1:23:21 um.
1:23:22 Uh,
1:23:23 I'm gonna just thank all of those of you who asked questions.
1:23:26 I thought it was really interesting conversation.
1:23:28 Thank you,
1:23:29 Kiara,
1:23:29 for joining us today for your,
1:23:30 for your really your useful insights and it just so happens that
1:23:33 you had some on the boots on the ground experience to,
1:23:36 to mirror back to,
1:23:37 to Oyebola and then
1:23:39 especially thank you to Oyebola for the
1:23:40 really interesting presentation.
1:23:42 I think we all learned a lot from this and,
1:23:44 and,
1:23:44 and
1:23:45 given us a lot of food for thought and as you've mentioned
1:23:47 and so some of this exchange just like a fruitful research,
1:23:51 um,
1:23:52 down the line,
1:23:52 so looking forward to that as well.
1:23:54 So thanks everyone and have a good rest of your day.
1:23:57 Thank you all.
1:23:59 Thank you.
1:23:59 Thank you.
1:24:00 Aye,
1:24:01 yeah,
1:24:01 thanks.
- add-style
- lp-body-content