00:05 Well,
00:06 welcome everybody.
00:07 Thanks for joining us today
00:09 at this um,
00:11 April edition of our policy research talks.
00:14 As,
00:14 um,
00:15 many of you know,
00:16 these talks give us an opportunity to present our work,
00:20 uh,
00:20 coming out of the World Bank's research department.
00:23 With the goal of sharing the findings
00:25 with colleagues inside and outside the department,
00:27 uh,
00:28 as well as,
00:28 uh,
00:29 outside of the World Bank.
00:30 So a warm welcome to everybody,
00:32 uh,
00:32 in person and
00:35 online through the Webex or the live stream.
00:38 Uh,
00:38 today's topic is revenue mobilization
00:42 is
00:43 especially critical for low and middle income countries.
00:46 Uh,
00:46 where tax systems often operate under significant constraints,
00:50 such as informality,
00:51 low tax capacity,
00:52 political hurdles,
00:54 and limited data analytics,
00:56 which frequently lead to ad hoc policy adjustments
00:59 rather than comprehensive reform.
01:01 However,
01:02 as we'll hear today,
01:03 advances in data analytics are opening new possibilities for evidence-based
01:07 real-time
01:08 policymaking.
01:10 To guide us through these new developments,
01:11 we're joined by Dario Totarolo,
01:13 an economist from the macroeconomics and growth
01:16 team here at the Development Research Group.
01:19 His research focuses on public economics in developing countries
01:21 and seeks to understand how public policies affect workers,
01:25 businesses,
01:25 and consumers.
01:27 Also the distributional impacts and the unintended effects
01:30 due to design issues related to tax policy.
01:34 And how governments can address tax avoidance and evasion,
01:37 especially at the top of the income distribution,
01:39 and with
01:40 April 15th coming soon.
01:43 Um
01:45 Prior to joining the bank,
01:46 he was an assistant professor at the University of Nottingham
01:48 and a postdoctoral fellow at the Institute for Fiscal Studies.
01:52 He holds a PhD in Economics from UC Berkeley and a
01:55 BSC and MSC in Economics from
01:58 Uh,
01:59 Universidad Nacional de la Plata.
02:01 He received the National Tax Association's
02:03 Outstanding doctoral dissertation prize in 2020.
02:08 I'm also uh delighted to welcome our discussion today,
02:11 Gabriele Naust.
02:13 She's a practice manager in the global unit of the
02:15 poverty and equity Global department at the World Bank.
02:18 Previously,
02:19 she served in various positions as,
02:21 as lead poverty economist in the
02:23 West and Central Africa region of the World Bank,
02:26 Latin America
02:27 and Caribbean.
02:29 And Eastern Europe and Central Asian regions.
02:33 She has led technical work
02:34 and policy dialogue on the distributional impact of fiscal policies,
02:38 ex ante analysis of the distributional impacts
02:41 of climate change and climate action,
02:43 and on methods and tools to decompose changes in poverty and inequality.
02:48 She's
02:49 regional analysis on poverty and inequality
02:51 in Africa,
02:52 global analysis on fiscal policy in developing countries,
02:55 and the barriers to women's labor force participation and agency.
02:59 So,
03:00 uh,
03:00 the
03:02 The
03:03 rules of the game today,
03:04 Dario will present for approximately 45 minutes,
03:08 after which we'll hear from Gabriela for about 10 minutes or so.
03:11 Then we'll open the floor to Q&A,
03:14 uh,
03:14 from the audience.
03:15 Um,
03:16 if you have a question,
03:17 please use the raise hand option on the Webex.
03:21 Um,
03:21 and if you're in the room,
03:22 obviously raise your hand and go to the mic.
03:25 Um,
03:26 just a reminder,
03:27 the session is being recorded,
03:29 uh,
03:29 so please ensure that your microphones are,
03:31 um,
03:32 muted when you're not speaking.
03:34 With that over to you,
03:35 Dario.
03:36 Thank you,
03:36 Leo.
03:37 Uh,
03:37 how do I change to the
03:48 That one
03:49 going through that.
03:53 But,
03:53 uh,
03:58 It's still not 2.
04:01 So I guess,
04:02 is this,
04:03 we have to do it in Webex?
04:07 Because I see here and slide show.
04:14 Bro,
04:14 yeah.
04:22 There you go.
04:23 Yeah.
04:24 OK.
04:26 full screen.
04:28 OK.
04:31 Let's begin.
04:32 Thank you,
04:33 um,
04:33 those joining online.
04:36 And those who made it in person despite the weather.
04:40 Um,
04:41 so now I realize by looking at the,
04:43 the cover of the slide that you know it could be,
04:46 well,
04:47 the,
04:47 the cover of a TED talk or even like a Quentin Tarantino
04:51 movie.
04:52 But don't get your hopes up,
04:54 still about taxes,
04:56 so.
04:57 So let,
04:58 let me,
04:58 let me start by highlighting two quotes,
05:01 um.
05:02 The first one
05:04 says that taxes are a necessary evil.
05:07 But,
05:08 um,
05:09 and the fewer we have of them,
05:10 the better.
05:12 What I want to argue here is that
05:15 taxes from an efficiency perspective are fundamentally bad
05:18 because they often
05:20 discourage something
05:21 and make the pie smaller.
05:22 So if you're familiar with the supply and demand chart diagram on the left
05:27 with is a tax here
05:28 and we see that quantities are decreasing.
05:31 But countries
05:32 need tax revenue to achieve
05:34 development goals,
05:35 so this is a green area.
05:38 And in the process
05:39 they typically create
05:41 uh losers and winners so the incidence of attacks or who
05:44 is benefited and affected is shared by consumers and producers or
05:48 whoever is affected by it.
05:51 Selecting
05:53 the
05:53 tax mix and levels is usually at the forefront of policies policies agendas,
06:00 and here we have a figure that I like a lot but we see
06:02 at the top of the figure that um tax levels relative to the GDP
06:07 and the figure is splitting this into um different income levels so high income,
06:12 upper middle,
06:13 lower middle and low income countries.
06:15 And also by the type of tax,
06:16 so direct taxes on income or profits and indirect taxes on mostly um consumption.
06:23 And then we have other sources of revenue.
06:26 So what's really striking in this figure is that you know if
06:28 you focus on the bars in the middle for each country group,
06:32 we can see that
06:33 they are very similar.
06:34 So in terms of indirect taxes are quite similar across the spectrum.
06:39 And we see big differences in,
06:41 in,
06:41 in,
06:42 in sorry in direct
06:43 taxes such as those on wages and income and so on,
06:47 so this is really where
06:49 um high income economies make a difference.
06:53 But,
06:54 um,
06:55 typically,
06:56 as you can see they do it in different ways and colors
06:59 and
07:01 What we see here is that there are some vanilla principles,
07:03 what I put at the top of the slide neutrality,
07:06 equity,
07:07 um,
07:07 and simplicity.
07:09 And we see
07:11 That,
07:12 um,
07:13 these are useful guides,
07:14 but in practice what we often see
07:17 is that setting up efficiency,
07:18 efficient and equitable tax systems is,
07:21 is tough,
07:21 it's a tall order
07:23 and here's the second quote that I want to highlight
07:25 which says that in developing countries tax policies is often the
07:28 art of the possible rather than the pursuit of the optimal.
07:32 I want to highlight 4 main barriers.
07:35 The first one is informal economy,
07:36 so typically employment and informal consumption,
07:40 and there is a,
07:41 a good policy talk by Pier
07:43 on this topic.
07:44 The second one is weak capacity and high compliance costs for taxpayers.
07:49 There is a very good talk by our colleague Oyola.
07:52 There are also political roadblocks
07:54 typically captured by elites,
07:56 um,
07:57 in the society and also resistance from
08:00 lower income people.
08:02 And then the 4th factor which is the one I'm going
08:04 to focus most in this talk is limited data analytics,
08:08 typically because
08:09 uh
08:10 you know tax or governments don't have time.
08:13 Or don't,
08:13 don't have the capacity to do it or they see no value in this.
08:17 But
08:18 we are lagging behind,
08:19 and this is one of the points that I want to make in,
08:21 in the talk.
08:24 Taken together
08:26 these four barriers,
08:27 um could rationalize why we have
08:30 inefficient tax and inequitable tax structures
08:33 that perpetuate in time,
08:35 and these are typically characterized by service taxes.
08:40 And that typically are production efficient
08:43 and also
08:44 marginal reforms or changes
08:48 rather than more holistic reforms,
08:49 comprehensive reforms.
08:53 So let me give you a timely example of what I mean by service
08:58 tax.
08:58 So taxes are typically taxes that are much easier to collect.
09:02 So in developing countries you know you create revenue quite fast.
09:06 But at the cost of
09:09 production inefficiencies.
09:11 And this is a key question to keep in mind when we think about these taxes.
09:14 The revenue efficiency gains offset production efficiency losses
09:18 for that tax
09:19 and also comparing to other tools.
09:22 So the fact that I want to focus on
09:24 for this example is the tax on mobile money.
09:27 Mole money adoption and usage has increased substantially in the last,
09:31 I would say 5 or 10 years,
09:33 5 years especially.
09:35 But so has the tax on mobile money transactions,
09:38 especially in African countries.
09:40 So this is
09:42 Possesses a complex dilemma.
09:44 So here we have 4 factors,
09:45 2 in favor,
09:46 2 against.
09:47 The first one is that this tax helps mobilize a fair amount of revenue,
09:52 so you can see in this figure
09:54 that I'm borrowing from a paper,
09:56 um,
09:57 from the ICTD
09:58 group,
09:59 and also some from the,
10:01 the Zimbabwe Tax Administration.
10:04 Especially in Zimbabwe,
10:05 right?
10:06 You can see that in 2024 it was like 5% of all the
10:09 revenue they collect.
10:12 The second positive aspect is that
10:15 with this tax,
10:17 countries typically can reach the informal sector.
10:19 So if there are some informal businesses that
10:21 are making some transactions with their phones,
10:24 then they are going to pay a tax.
10:26 And then the negative aspects are that of course,
10:29 as you know,
10:29 they discourage more money usage
10:31 and
10:33 Some people argue that they could be regressive.
10:35 We still don't know the answer to this,
10:37 so we need more evidence.
10:40 So let me illustrate this
10:42 from 4 countries.
10:45 I'm uh going to,
10:47 to borrow some figures from a report from the ICTD.
10:51 And I'm going to show you some evidence for Ghana,
10:53 Uganda,
10:54 Tanzania,
10:54 and Zimbabwe.
10:57 Starting with Ghana,
10:58 in this figure we can see
11:00 the blue bars are the total number of transactions done with mobile,
11:04 with their mobile phones.
11:06 And then the orange line is the value of those transactions.
11:09 And then the,
11:10 the yellow bar here
11:12 shows the introduction of the tax.
11:14 So what we can see very quickly,
11:15 this is more,
11:16 you know,
11:16 macro facts.
11:17 We can see that
11:19 despite the introduction of the tax,
11:20 we see that
11:22 the number of
11:23 er transactions keep increasing
11:25 and we also see that the orange line is,
11:27 is increasing as well.
11:29 When we go to Uganda,
11:31 we see the same
11:32 pattern,
11:32 right?
11:33 The transactions keep going up,
11:35 the green line keeps going up.
11:37 And similarly,
11:39 If we look at Tanzania,
11:40 so everything seems to be going up.
11:43 But there is a visible economic drag in this figure,
11:46 and as you can see if we project a line
11:49 through the,
11:51 the yellow,
11:52 um,
11:53 time series over here,
11:54 you can see that there is like a missing
11:57 mass
11:58 that we can calculate.
11:58 This is connected to the first figure that I showed you today,
12:02 the supply and demand,
12:03 when there is,
12:03 you know,
12:04 we are losing some welfare
12:05 when we introduce a tax.
12:08 The story in Zimbabwe is a little bit different,
12:11 um.
12:13 Zimbabwe introduced its tax in October 2018
12:16 and you can see it now just in,
12:17 in two months.
12:19 Three months,
12:20 they collected a significant amount of revenue,
12:22 but then in 2019 they collected 10% of total revenue from this tax.
12:26 So this is
12:27 No,
12:28 it's a lot of,
12:29 it's,
12:29 it's,
12:29 it's very,
12:30 it's a very important source for,
12:31 for these countries.
12:33 And this is a revenue that can be used for
12:35 development goals,
12:36 you know,
12:36 to fix roads and
12:38 social protection and so on.
12:41 As a benchmark because now it's uh
12:43 popular these days,
12:44 you can see that even you know it's collecting more
12:47 revenue than other taxes such as custom duties.
12:51 Now,
12:51 what's the issue
12:53 The key question is what's happening with money usage and
12:56 in particular also incidents.
12:59 So when we look at macroevolution of um number of the volume of transactions
13:04 er done with mobile phones.
13:06 So you can see that
13:07 in,
13:08 in Zimbabwe in particular,
13:09 there was a,
13:10 a big increase when there was a cash cri crisis at the beginning of the 2018.
13:17 When they introduce the tax,
13:18 we see that it decreases a little bit,
13:20 but then it keeps increasing
13:21 and the things get really bad,
13:23 so there is like a 60% decrease
13:25 when there were some more macro
13:27 restrictions at force.
13:28 So these are uh restrictions,
13:30 some bannings on some bans on mobile money transactions.
13:34 So here's where you see that
13:36 this is um
13:37 really affecting the economy.
13:38 So the point of this graph is that
13:41 it could be more than taxes and we need
13:43 More,
13:44 uh,
13:45 research to disentangle the effect of these
13:47 taxes related to
13:49 other confounding factors.
13:51 So what are some early reflections that I want to make out of this is
13:56 The tax can help mobilize
13:57 valuable revenue,
13:59 and it can help tax the informal sector.
14:01 There are visible welfare losses.
14:04 It doesn't seem to be breaking the economy.
14:05 It's an open question.
14:07 And
14:08 the third point is that the design
14:09 varies a lot across countries.
14:11 So it goes for example,
14:12 flat 0.5% in Uganda to
14:15 22 bands in Tanzania where there is like an increased
14:19 fee that they have to pay depending on the value of the transaction.
14:23 And then the second point I want to make
14:24 is that eyeballing trends like from this figure.
14:28 Eyeballing trends over here,
14:30 it's informative,
14:31 but it's not enough.
14:33 So we need more forensic microanalysis
14:36 to set a precedent
14:38 and in particular to
14:39 estimate two unknowns.
14:40 The first one is the size of the distortion.
14:43 And also the marginal value of public funds.
14:45 Marginal value of public funds is essentially
14:48 uh the welfare loss for each additional dollar collected in revenue.
14:52 And the second one is the incidence,
14:54 who is benefiting,
14:55 who is being affected by this,
14:56 by this tax and by any tax.
14:59 So more generally,
15:00 the point that I want to make in this talk is that countries have reached a tax
15:04 and data crossroads.
15:06 Um,
15:08 and
15:09 It's key to recognize that there are,
15:11 that these 4 barriers that I mentioned at the beginning of the talk.
15:15 Are not laws of nature and these are choices.
15:19 So for example,
15:20 whether to give a
15:21 tax relief via exemptions
15:24 or reduce rates on capital income or a particular sector
15:28 and the decision to invest in
15:30 tax capacity,
15:32 collecting information,
15:33 communications
15:35 to get,
15:35 you know,
15:35 kind of alleviate the political economy of reforms.
15:39 And investing in data analytics,
15:40 then,
15:40 you know,
15:41 seeing the value of data analytics for policy choices,
15:43 for policy decisions.
15:46 It's key to recognize that there is a tax credibility revolution,
15:50 um,
15:51 and
15:52 the key is that there is a landscape where we have like electronic government,
15:56 e-wallets,
15:57 electronic filing,
15:58 electronic invoices.
16:00 Artificial intelligence,
16:01 high frequency data,
16:03 global coordination,
16:04 and so on.
16:05 Believe me,
16:06 countries are getting so much data,
16:08 so much information that they don't know what to do with it.
16:10 So it's sitting in many cases it's sitting in the tax administration,
16:14 and there is value in this.
16:15 What can they do with this?
16:16 They can.
16:18 Improve the use of second best taxes such as value added tax,
16:22 personal income tax.
16:24 And they can better understand the best tools
16:26 like it's happening in Africa with mobile mining.
16:28 We need to understand
16:30 what is the size of the elasticity of that.
16:32 So for this I'm,
16:34 uh,
16:34 I have two points here.
16:35 The first one is we can use data
16:37 for ex-ante design so we can understand the anatomy of a tax,
16:41 a tax that is currently,
16:42 currently in place.
16:44 And the second point is an ex post analysis where we actually evaluate
16:48 past reforms and ongoing policies,
16:51 and we help governments contract counterfactuals
16:54 of policies or counterfactual scenario.
16:56 What would happen if we do this or that.
17:00 So the outline of the talk.
17:02 Um,
17:03 I try to be comprehensive.
17:04 I know where to skip if I run out of time,
17:07 but
17:07 I wanna talk about national taxes and subnational taxes,
17:10 in particular,
17:12 the value of the tax,
17:13 then the personal income tax,
17:15 and lastly subnational taxes because I,
17:17 I feel it's a very important topic that has received much less attention.
17:25 OK.
17:26 So let me start with the most important
17:28 uh source of revenue in,
17:30 in low and middle-income countries.
17:32 Um,
17:34 this tax has suffered from
17:36 efficiency and equity,
17:38 equity shortcomings,
17:39 but it's still
17:41 preferable to other taxes,
17:43 and the point that I want to make here is that this is a tax that can be improved.
17:48 It's a tool.
17:49 What caught my attention to,
17:50 to get into this topic and
17:53 write papers and so on is that
17:54 this tool is often
17:56 used to affect the economy and here I put P and Q.
17:59 So prices.
18:00 And quantities or activities,
18:03 but this is a very,
18:04 it's a costly way to pursue equity.
18:07 Or
18:08 stimulate the economy
18:10 and often has unintended consequences.
18:13 And the key,
18:14 again,
18:14 I'm going to repeat this many times today,
18:15 but the impact is usually not measured.
18:18 And the incidence is not well understood.
18:21 And this is especially problematic or
18:23 um there is
18:25 less evidence in low-income countries and low and middle-income countries.
18:30 So I'm going to provide some lessons um
18:32 for Argentina and Mexico.
18:34 This is in blue because it's connected to,
18:35 to prices,
18:36 so I try to play with this.
18:38 And then for Peru,
18:39 I'm going to show you some evidence on uh VAT that was
18:42 meant to stimulate a particular sector.
18:46 So let me begin with that.
18:50 This is a,
18:51 why,
18:51 why should we,
18:52 why should we care about
18:53 VAT cuts and prices because many countries are cutting the grocery tax rates
18:59 on food on a scale not seen before,
19:01 with the ultimate goal of helping the most
19:03 vulnerable cope with the soaring cost of living.
19:05 So essentially they cut the VAT
19:07 with the hope that prices will decrease
19:09 and this will increase purchasing power of people in difficult times.
19:13 So here you have a list,
19:14 non-exhaustive list from um a paper I'm working on
19:18 uh where we have on the left countries that have
19:20 abolished the VAT
19:23 or sales tax on food.
19:25 Then in the middle some countries that have
19:27 done partial VAT cuts and then some countries that were considering at some point.
19:31 This is
19:32 mostly for
19:33 after 2020.
19:35 So I'm highlighting in,
19:36 in yellow two countries,
19:38 um,
19:39 Bangladesh and Argentina.
19:40 I'm going to focus on this.
19:42 Why did I choose,
19:43 uh,
19:43 these two countries?
19:45 Well,
19:45 first of all,
19:46 because,
19:46 uh,
19:47 Bangladesh has provided unconditional support to Argentina in the last,
19:51 uh,
19:52 World Cup.
19:52 So,
19:53 I think it's fair to,
19:55 to give a space to this.
19:56 But second,
19:57 so I was in Bangladesh in February,
19:59 uh,
20:00 for a week and I was struck
20:01 by the
20:03 Number of VAT cuts that have
20:04 implemented since 2023.
20:07 Uh,
20:07 in particular,
20:08 there was,
20:09 when I was there,
20:10 it was a training on tax expenditures
20:12 and there was this VAT card I implemented at the end of uh 24
20:16 uh on
20:17 cooking oils.
20:19 Uh and the idea was to ease prices and ensure that there was enough supply.
20:24 This tax
20:25 expired recently.
20:26 There was big pressures to extend it to June 2025 and this is a very common thing,
20:31 you know,
20:31 there is a temporary thing and then people try to make it permanent.
20:35 But
20:36 there is no analysis of this tax and uh I,
20:39 I,
20:39 I,
20:39 what I want to argue is that it's possible to do it.
20:43 The second example is um
20:46 Argentina.
20:46 Why Argentina?
20:47 Because I have a paper on this.
20:49 It's the
20:49 best paper ever.
20:51 And
20:52 Argentina cut the VAT,
20:54 the value added tax,
20:55 so tax on consumption,
20:57 uh,
20:57 from 21% to 0%,
20:59 so pretty dramatic large change on basic foodstuff.
21:03 It was in place for 4 months and a half,
21:05 and the goal was to increase the purchasing power of,
21:08 of people in,
21:08 in difficult times.
21:10 So this is a photo of a supermarket
21:12 in Argentina.
21:13 You can see that there is a banner saying IVA,
21:16 that's VAT in Spanish.
21:19 And,
21:19 and you can see that they are saying that there are over
21:22 1900 products with 0% VAT and then they list the different categories.
21:28 So in a recent paper,
21:30 we analyzed the effectiveness of these VAT cuts
21:33 um
21:34 to lower prices and reach the target population.
21:38 What do we do?
21:39 We,
21:40 the products that you see in the supermarkets,
21:43 we have data
21:44 from these places.
21:46 So for every product in the supermarket,
21:48 every product has a barcode
21:50 attached to it and it's unique
21:52 and you can track that barcode over time.
21:54 So we have
21:55 all this data
21:57 for
21:57 all the supermarkets in Argentina where we observe the price and
22:00 the number of units that they sold weekly or monthly.
22:05 What do we do with this data?
22:06 We split
22:07 all the products that you see here on the shelves into
22:09 those that were affected by the VAT cut and those that were not.
22:12 So for example,
22:13 tea,
22:13 milk,
22:14 eggs,
22:14 and pasta
22:15 were subject to 0% VAT but not coffee,
22:18 cereal,
22:19 and crackers.
22:20 So we split these goods into these groups and then we track
22:23 their prices and consumption
22:25 over time.
22:27 What do we find?
22:28 We find that
22:29 um
22:30 what do we do with this?
22:31 So we compare prices of these two groups over time.
22:35 What we find is in this figure we can see time on
22:37 the horizontal axis and then the price change on the vertical axis.
22:41 And here what you can see is
22:44 That before the reform,
22:45 which is the first vertical
22:47 line,
22:48 you see that prices were evolving similarly between these two groups,
22:51 so that's why it oscillates around 0.
22:53 And then
22:55 The VAT quickly lowered prices in these chain supermarkets.
22:59 So these chain supermarkets are like the most formal and biggest ones,
23:02 uh,
23:03 like Walmart,
23:04 right?
23:04 So you know,
23:05 within one week you see that there is almost full pass through.
23:07 So it's a
23:08 large drop in prices.
23:09 It's relatively stable over time.
23:13 But it's not full pass through.
23:14 So full pass through would be if it reaches the blue line.
23:18 When the VAT cut
23:20 ended,
23:21 prices shot up even higher than the pre-reform level.
23:24 So this is one of the risks of temporary
23:27 tax changes.
23:28 And there is an asymmetry that has been documented in our in our papers.
23:32 So this means that prices were,
23:34 if the policy had not taken place,
23:36 probably prices will be lower.
23:39 But the twist,
23:40 the twist of the,
23:41 the twist of the paper is that
23:43 novelty of the paper is that
23:45 the government perhaps preempting that there could be an asymmetry.
23:49 They imposed
23:50 some caps.
23:52 On how much prices could increase,
23:53 but they did this for some of the treated goods.
23:56 So for example,
23:56 let's take rice.
23:59 Within rice,
23:59 you always,
24:00 you typically have basmati,
24:02 brown rice.
24:04 Organic rice and so on.
24:05 So there were caps on like the fanciest
24:09 rices like
24:10 uh organic rice but not brown rice,
24:12 the most standard one or white rice.
24:16 When we look at
24:17 those that had a cap,
24:18 we actually see that prices increase.
24:21 Less,
24:22 so it was binding
24:23 and
24:24 but they overdid it,
24:25 right?
24:25 So they were,
24:26 they reached uh
24:27 uh lower level
24:29 relative to the pre-reformed one.
24:32 Now a key question is,
24:33 this is chain supermarkets.
24:34 What happens in more
24:36 independent grocery stores which many people shop in those places?
24:41 In independent grocery stores,
24:43 so now
24:44 I'm going to move from this figure with weekly data to monthly data,
24:47 and I'm going to split this effect,
24:49 price effect into chain supermarkets
24:52 like Walmart and independent grocery stores.
24:54 We see the red line,
24:56 independent grocery stores,
24:57 we see that prices fell
24:58 less.
25:00 After the VAT cut.
25:01 So we see that there is a pass-through,
25:02 it's like a drop of 6%,
25:04 it's like 35% pass-through rate.
25:07 So prices fell less after the VAT cut
25:10 and they overshot more
25:12 after the VAT was reinstated.
25:15 So a key question for welfare analysis is where do people shop?
25:18 Because you know you see that
25:20 in independent stores it's much more limited than prices are increasing more
25:24 and the gaps are really binding in Walmart type of supermarkets.
25:29 Well
25:31 It turns out that low income people are substantially more likely to shop in the,
25:35 in these places,
25:37 independent grocery stores.
25:39 So this means that they benefited less
25:41 from the VAT cut
25:42 and they were negatively affected by the VAT increase.
25:46 OK,
25:46 so this is what in the paper we have a model where we
25:49 estimate
25:50 numerically the consequences of the welfare effect of this,
25:53 this reform.
25:56 One digression that I want to make
25:58 is
25:59 the data that we are using.
26:01 So this is the underlying what,
26:02 what I,
26:02 what do I mean by scanner data.
26:04 This is a screen,
26:05 screenshot from our data.
26:07 It's hard to see,
26:07 but let me tell you what we have here.
26:10 In this column,
26:11 the one,
26:12 people online cannot see what I'm point,
26:14 but
26:15 there is a,
26:16 a column here where we have a unique barcode.
26:18 This is the same
26:20 internationally,
26:21 so you can track the same goods,
26:22 so this is Coca-Cola.
26:24 And then we have the month of the year and then we see
26:27 in which store that is like a fake ID doesn't mean anything,
26:30 but there is a code attached to
26:33 that store,
26:34 right?
26:34 So this is the type of data that we are using.
26:37 We can georeference this data.
26:40 Um,
26:41 this is for a,
26:41 for a city in Argentina.
26:43 This is the type of data we are using in this paper.
26:46 So preempting a question that many of you might have
26:50 is this data only available in high
26:53 or upper middle income countries?
26:56 The answer is no.
26:57 This is not just not just in,
26:58 in those countries,
26:59 but it's actually prevalent in many countries.
27:02 We have an ongoing agenda with um
27:05 global tax program funded by the global tax program,
27:08 and thank you,
27:09 and the IFS and,
27:11 and CEFIP.
27:13 Where we have market data for over 20 countries.
27:16 And with this data,
27:17 so you see the countries here on the,
27:19 the chart on the right.
27:21 The one we
27:22 have money to,
27:23 to purchase are those in uh black.
27:26 So it's like a fee that you pay and you have a license to use it.
27:29 But then you can see that there is,
27:30 there are many countries
27:32 at different levels of development.
27:34 So it's a very interesting source of data
27:36 uh where you can track prices and consumption over time.
27:40 What are we doing with this?
27:41 We are analyzing,
27:42 for example,
27:43 the tax on feminine hygiene products.
27:45 These are reforms that are taking place in many countries.
27:48 Uh,
27:48 and also price and consumption,
27:50 uh,
27:50 habits.
27:53 Something that I want to,
27:54 to emphasize,
27:55 er,
27:55 let me see,
27:56 yeah.
27:57 Is
27:58 why do we care about,
27:59 why do we need this type of data?
28:01 Because we want to,
28:02 we want to know what the answer to this question is.
28:04 How is this?
28:05 How are these reforms,
28:06 VAT?
28:07 Are they
28:09 effective?
28:09 Are they or are they reaching the goal that they aim to?
28:12 And
28:13 this is just preliminary evidence from Mexico.
28:15 In Mexico,
28:16 this is using public data that can be downloaded online actually.
28:20 But in Mexico they abolished the VAT on women's sanitary products in 2022.
28:26 We check very quickly what happened with prices.
28:28 It turns out
28:29 the prices are dropping by
28:30 only 50% of the VAT cut,
28:32 right?
28:32 So it's like 50% is being
28:34 pocketed by the supply chain,
28:36 supermarket manufacturers.
28:39 So it's very important to kind of try to have a sense of like global evidence on this.
28:44 There is some global evidence,
28:46 not global,
28:47 but 55 developed economies,
28:48 and they find that path through is full in those countries.
28:52 So we see,
28:52 we,
28:53 we just have one example here but it seems that
28:56 at least in Mexico it's half of it.
28:59 And then the last point that I want to make about scanner data is
29:02 um
29:03 there are many of these market companies
29:05 that provide this type of data and
29:07 every country has one,
29:08 and they have this,
29:09 you go to a supermarket,
29:10 you see when they scan
29:12 that goes to a database.
29:14 The way I got access to this data was uh uh an anecdote
29:18 on LinkedIn.
29:19 So I contacted someone there
29:21 that I knew work at a company that have this data,
29:24 and then I set up a meeting
29:26 and then they gave me a quote and then I purchased the data.
29:29 This is,
29:31 so it requires some entrepreneurship,
29:33 but these data are available in many,
29:35 many,
29:35 uh,
29:35 countries.
29:37 It's expensive for researchers,
29:38 it's not expensive for countries,
29:39 for governments,
29:40 so.
29:48 1 2nd.
29:55 The last example that I want to um talk about,
29:58 how much time do I have uh Dion?
30:02 But
30:05 20,
30:06 yeah,
30:06 OK.
30:07 So I think
30:07 I,
30:08 I'm gonna go fast on this one because I'm presenting
30:10 this paper at the end of April in a,
30:13 a conference here at the bank,
30:14 but
30:14 the point that I want to make,
30:15 we have an ongoing collaboration with the tax
30:18 administration and the Ministry of Finance in Peru.
30:20 There was a recent in 2022 they got the VAT rate
30:23 for hotels and restaurants with the goal of stimulating the sector.
30:28 And we have this collaboration where
30:31 Peru has very strict data privacy,
30:33 um,
30:34 uh,
30:35 regulations.
30:36 So the way to access,
30:37 to work with them was
30:38 they assign a data scientist and then we collaborate with them,
30:41 we send codes,
30:42 they
30:43 send output back
30:45 and,
30:45 and this is a collaboration that it's time consuming but it,
30:48 you know,
30:48 it's an alternative way in which we can make things uh work.
30:52 So in this case in Peru,
30:54 it seems that the,
30:55 the VAT card is not working as intended.
30:57 So in particular there is partial take up which is shown in this figure.
31:01 Um,
31:03 The
31:04 tax is decreasing,
31:05 so this is a chart showing that tax,
31:08 taxes are decreasing by 50%.
31:10 There is a drop in revenue from the government.
31:13 Uh,
31:13 I apologize and I'm going fast here,
31:15 but I want to cover everything.
31:17 There doesn't seem to be any effect on employment,
31:20 so zero effect.
31:22 And then there is some
31:24 effect on sales and a decreasing purchases.
31:26 So the way we interpret this is that this is a sector that is known for a high evasion
31:31 and what we argue is that when you lower
31:33 the output tax,
31:35 then
31:36 firms start to accumulate a lot of credit.
31:39 So if they were engaged in fraudulent
31:42 activities by purchasing fake invoices and so on,
31:46 the incentive after this,
31:47 this reform is to start doing so.
31:50 So we are
31:51 trying to um
31:53 Dig into the mechanisms,
31:55 but uh this is still preliminary and I think it's uh important to,
31:58 to get more evidence.
32:01 So the takeaways,
32:02 VAT cuts have,
32:03 uh,
32:04 often have effects that policymakers didn't anticipate.
32:07 There is value of data analytics for tax policy decisions,
32:11 and there are some contentious areas around permanent VAT
32:15 changes that,
32:15 uh,
32:16 I didn't touch,
32:17 but I think are very important.
32:18 One of them is whether we should be having,
32:21 if we want to help the poor people,
32:23 do,
32:23 should we reduce tax rates or,
32:25 you know,
32:25 more targeted tax rebates like,
32:27 uh,
32:27 many people are.
32:29 Uh,
32:29 analyzing,
32:30 um,
32:31 today.
32:31 And then the political economy sometimes it's not feasible to remove exemptions.
32:35 It has proven to be very difficult in countries like
32:39 Bangladesh
32:40 and Kenya more recently.
32:44 Moving on to the second,
32:46 uh,
32:46 part of the talk,
32:47 personal income tax.
32:49 It's a tax we aspire to rely on but often fall
32:53 short of.
32:54 This is the most progressive tax
32:56 that we have,
32:57 but it's not very redistributive.
33:00 The reason is that
33:01 the tax base is porous.
33:04 In particular because there is informality
33:06 and there are preferential treatments such as special deductions or
33:10 even um
33:12 reduced rates.
33:13 So,
33:13 but typically it reaches the top
33:16 5%,
33:17 even
33:17 lower than that of the adult population in,
33:20 in many low and middle income countries.
33:23 I want to focus on two key design features.
33:26 The first one is the exemption floor.
33:28 So the exemption floor is the floor
33:30 that if you make more money than that,
33:32 you,
33:32 you become subject to the tax.
33:35 There is a beautiful paper by Anders Jensen
33:37 where he shows that as countries develop.
33:40 The employee share
33:42 Increases,
33:43 which means that they
33:44 uh transition from self-employment to employee employment.
33:48 And as that happens,
33:50 the threshold.
33:52 Decreases
33:54 across the income distribution,
33:55 which means that more people become subject to it.
33:59 But
33:59 in practice and in the short run,
34:01 it has proven very hard politically challenging and
34:04 costly to expand the tax base and,
34:06 and I want to argue
34:08 that by focusing on one example from,
34:10 from Argentina.
34:11 Argentina experienced with
34:13 in the last 22 years with base,
34:16 base expansion via
34:18 under indexing.
34:19 Underindexing means that they
34:21 Nominal incomes increase,
34:23 but tax thresholds are fixed and therefore there is bracket creep,
34:26 so more people,
34:27 you know,
34:27 your tax liability increases and more people start paying taxes.
34:31 This figure here shows the evolution of the share of
34:34 former workers
34:36 that
34:36 pay the tax,
34:37 and you can see that there is a lot of variability.
34:40 The reason for this is that there are different threats,
34:42 so there was a period of formalization until 2008,
34:46 but then there was inflation,
34:47 which means that more people became subject to the tax,
34:50 and it has been politically costly for
34:52 different governments,
34:53 right wing,
34:53 left wing,
34:54 and so on.
34:55 The government has been adjusting.
34:57 So you know,
34:58 we see that the tax base is expanding but all of a
35:00 sudden they have to cut it because it's politically very costly.
35:05 And
35:05 it seems
35:06 to me at least that they have overdid it in the last period of time.
35:11 Um,
35:12 you know,
35:12 again,
35:12 this is a good tax we have been discussing
35:15 in,
35:15 in many countries they often discussing like ways to decrease it rather than
35:20 increase it,
35:20 but again it's it's like a political economy
35:22 issue that it's very difficult to sort out
35:25 and we need data for this.
35:26 So this is just a screenshot of uh
35:28 the metro station in Buenos Aires.
35:30 Where they are saying damn tax on labor and there is like a national strike to,
35:34 you know,
35:35 kind of
35:36 decrease it.
35:38 The second feature that I want to focus on is the scope.
35:42 There is a long standing debate on whether
35:45 income taxes should be comprehensive,
35:47 meaning they tax
35:48 all the income sources,
35:50 or whether we should have like a dual.
35:53 Schedule our tax system in which you have like progressive rates for labor income
35:58 and then
35:59 different flat rates for for
36:01 for capital income.
36:03 there is a debate.
36:05 Um,
36:06 what I want to argue here is that regardless of this debate,
36:09 it's key to have
36:11 a better diagnostic
36:12 and better impact measures of um
36:15 what's going on in low and middle-income countries,
36:18 in particular,
36:19 to
36:21 Shed some evidence on the,
36:22 on this debate on comprehensive versus dual.
36:24 It's key to understand that capital
36:26 and labor income boundaries are often
36:29 blurry
36:30 and unclear.
36:32 And the second one is that
36:33 there is some,
36:33 some,
36:34 there is a paper by Piketty and coauthors that show that the tax differential,
36:38 so whether you apply a lower tax rate to capital income,
36:40 that differential
36:42 declines with the income shifting in artisticity.
36:44 So
36:45 income shifting ethicity is the ability for let's say a capital rentier
36:49 or an
36:50 executive of a firm.
36:52 To declare income as labor income or to declare it as dividend,
36:56 OK?
36:57 So it's very important to measure the size of this elasticity.
36:59 We know zero or very little about this.
37:03 So for this second point I want to highlight some
37:07 great evidence from Honduras.
37:08 This is work by Tiago Scott Pierre who is or was sitting here
37:13 and
37:14 Um,
37:15 and it's,
37:15 it's a great type of collaboration that they managed to develop
37:19 with the tax administration in Honduras.
37:22 What do they do in this work?
37:24 They assess the effective tax rate of top earners.
37:27 The effective tax rate is the percentage of the income that they pay in taxes,
37:31 right?
37:32 The key of this work is that they have
37:34 a comprehensive view
37:36 of
37:36 all the income sources of Honduras
37:39 by linking
37:40 personal income tax records to corporate
37:43 tax records.
37:44 And with this
37:46 Analysis with these tools,
37:47 what they do is they assign and distributed profits,
37:51 so retained earnings from a company,
37:53 they assign that to the shareholders of those companies.
37:57 What do they find?
37:57 They find that at least in Honduras,
37:59 it seems that
38:00 effective tax rates increase,
38:02 so the progressivity increases all the way to the top.
38:05 This is for the richest,
38:06 so something I want to emphasize is
38:09 personal income tax in Honduras applies to the top,
38:12 um,
38:12 75,000 people,
38:14 like top 1.5%.
38:16 OK,
38:16 so it's really targeted at the top.
38:19 We see that it increases with,
38:20 with,
38:20 uh,
38:21 income.
38:23 But the key is that when you decompose this into different sources,
38:26 we see that.
38:28 It's mostly undistributed profits at the top,
38:31 so it's the blue shaded area,
38:33 what matters and
38:34 Uh,
38:35 it's what dominates at the top,
38:36 and this is what it,
38:37 what is.
38:38 Uh,
38:39 working as a backstop
38:40 for the personal income tax in the sense,
38:42 you know,
38:43 the reason we get progressivity all the way to the top,
38:45 we have a tax rate of
38:47 2 25%
38:48 for both personal income tax and corporate income tax.
38:52 But this is a very,
38:54 I would say
38:55 peculiar situation.
38:56 In most countries,
38:57 the personal income tax becomes regressive at the top.
39:01 So we have this evidence.
39:03 From Argentina.
39:04 I'm not lucky like uh Tatiana,
39:06 Tiago and,
39:07 and,
39:07 and coauth to have a collaboration with Argentina,
39:09 but
39:10 something that is really fascinating for Argentina is that they publish.
39:14 Tax tabulations
39:15 since 1990 or so,
39:18 very detailed tax stabulations online.
39:20 These,
39:20 these are public and this is an intermediate
39:22 step that I think many countries should be
39:24 uh doing.
39:26 So with these tax stabulations online,
39:27 what I did was to
39:29 sort people based on their
39:30 total income,
39:32 and here I show that for the richest,
39:34 so you know for the top.
39:35 0.1%,
39:37 so the richest 4 1000 people in Argentina,
39:39 the tax system becomes regressive at the top,
39:42 especially for the richest
39:43 2000 people.
39:45 What's going on?
39:46 It's that capital income share.
39:48 Increases at the top
39:50 and capital income is taxed at preferential rates lower than the other,
39:55 the,
39:55 the regular personal income tax.
39:59 So what's missing from this evidence is
40:01 um offshore income and wealth,
40:03 for example,
40:04 it's not here and consumption through the firm.
40:06 The second point is very hard to
40:08 think empirical to address.
40:11 There's a beautiful paper by Le where they do this in Portugal with some
40:14 data from I think an app or mobile phones.
40:16 I don't remember well.
40:19 The key for offshore and income wealth,
40:21 key question is how much
40:23 is out there.
40:24 I have a paper on this.
40:26 In Argentina,
40:27 at least in Argentina it seems that there is a lot.
40:29 There was a tax amnesty in 2016.
40:32 That reveal assets worth 2 21% of GDP
40:36 when they disclose this income,
40:38 this wealth.
40:39 When you disclose a stock,
40:41 you have to start paying taxes on the flow.
40:43 So that's why there was a large fiscal externality,
40:46 positive externality and the the capital income tax base more than doubled,
40:50 and this was
40:52 a permanent change,
40:53 uh,
40:53 enduring change.
40:55 So this is,
40:56 there is a key role for automatic exchange of information
40:59 and the common reporting standard.
41:01 And
41:02 again I want to emphasize countries,
41:03 and I know this because I spoke with tax administrations,
41:06 they are receiving.
41:07 Millions of records.
41:08 They don't know,
41:09 they don't know what to do with it,
41:10 how to clean it,
41:11 how to harmonize it,
41:12 and how to link it.
41:13 So again,
41:14 I think there is space here for data analytics to
41:17 um
41:18 to help.
41:20 The last
41:21 topic that I want to,
41:23 to talk about are subnational taxes.
41:26 These are widespread
41:27 with accessible data
41:29 and yet they are understudied.
41:32 There is a key question which is how local governments finance
41:35 development and at what economic and political cost.
41:39 This is a cri a critical question that received much less attention.
41:43 Um,
41:46 But you know,
41:46 data analytics again it's much needed here uh to,
41:49 to understand the anatomy of these taxes
41:52 and what is the impact of different policies that they have implemented over time.
41:56 I'm going to focus on two taxes which are the most important at the local level,
42:01 local business tax
42:03 and then property tax,
42:04 if I have time.
42:05 Otherwise I will skip it.
42:08 So local business tax um
42:11 It's like the,
42:12 I would say it's like the comfort food
42:15 of
42:16 public finances.
42:18 Uh,
42:18 because it's like,
42:19 um,
42:20 you know,
42:20 you get a lot of it,
42:21 but it's,
42:22 it's unhealthy.
42:24 So it's a fiscal crush with economic cost.
42:28 Why?
42:28 er this is due to lower political cost compared to a property tax which is typically
42:33 uh
42:34 unpopular
42:35 and
42:36 politically costly.
42:38 It's easier to collect
42:40 because there's
42:40 the revenue is typically concentrated at the very top,
42:43 and I'm going to show you a figure on this.
42:46 And then the incidence of the tax,
42:47 so who
42:48 is really affected by tax,
42:50 it's,
42:50 it's fuzzier
42:51 compared to the property tax.
42:55 The property tax,
42:57 is it a sleeping giant or is it a tax that is doomed to fail?
43:00 There is increasing consensus that the tax is underutilized,
43:04 but it's getting a lot of um academic attention
43:06 and there are many studies uh on this tax,
43:09 and I think it's great to,
43:10 to have this.
43:12 So why
43:13 should we
43:15 care about subnational taxation because
43:17 there are also interesting macro dynamics
43:19 um
43:20 in different countries.
43:22 It's hard to get surprisingly.
43:25 Uh,
43:25 good data,
43:26 macro data on
43:28 um
43:29 Collection levels for different countries,
43:32 um,
43:33 to me this is surprising because
43:36 it's surprising also that there are not more academics
43:38 working on this because if you think about it.
43:42 In Argentina and Mexico there are over 2000 municipalities.
43:45 In Brazil,
43:46 5000 municipalities,
43:47 in Nigeria,
43:48 800.
43:50 In the Philippines,
43:51 1 1500 municipalities.
43:53 So this,
43:54 the opportunity for
43:56 PhD students and for academics and so on to
43:59 shed light on this tax is endless,
44:01 you know,
44:01 it's a matter of reaching out and see who's interested in collaborating.
44:05 Um,
44:06 going back to the,
44:07 to the macrodynamics,
44:08 this figure is very concerning.
44:09 So we have here the composition of the tax revenue for the 24 provinces of Argentina.
44:15 We have the turnover tax,
44:16 property tax and other taxes.
44:19 Turnover tax
44:20 is a tax,
44:21 it's like the it's VAT but worse.
44:23 It's like a tax on gross sales,
44:25 so you can deduct
44:27 anything.
44:28 So it's
44:29 perhaps one of the
44:30 worst taxes that exist in theory.
44:33 Um,
44:34 it has advantages and disadvantages,
44:36 but in principle for efficiency it's,
44:38 it's,
44:38 it's very bad,
44:39 and we see that you know,
44:40 over time,
44:41 at the beginning in 1980,
44:42 roughly 50%
44:45 the,
44:45 it was 50% turnover tax and property tax.
44:49 Today they
44:50 essentially depend on,
44:51 on turnover taxes.
44:52 84% of subnational revenue,
44:54 own revenue is coming from this.
44:57 A similar pattern
44:58 emerges in municipalities,
45:00 so the,
45:00 the lowest level that we have in Argentina,
45:03 we see that they also impose their own.
45:06 Turnover tax at
45:08 low rates,
45:08 but still,
45:09 you know,
45:09 it's,
45:10 they are becoming,
45:11 they are depending more and more
45:13 on a tax that
45:14 in principle in an upper middle in an upper middle income country,
45:18 I,
45:19 I think this is a little bit concerning,
45:20 but again we need to balance,
45:22 we need better diagnostic,
45:23 we need to understand
45:24 what is the value of raising this extra revenue
45:27 versus not taxing at all or using a different instrument.
45:31 So that's why I put in the title Good Politics,
45:33 uh,
45:33 Bad Policy as a question mark.
45:36 So what I want to highlight here is some er ongoing
45:39 work that I have with uh Santiago Garria and Isidio Guarducci
45:44 where we are using.
45:46 Micro
45:47 tax data from uh all the provinces of Argentina.
45:51 And
45:52 I want to highlight 5 styled facts.
45:55 So again,
45:55 turnover tax is a tax applies very low rates
45:58 to to gross sales.
46:00 It's very easy to collect and administer.
46:02 Uh,
46:02 it provides a stable source of revenue,
46:05 but the main issue is that
46:07 it creates like a cascading where you tax the same transaction several times,
46:12 so this compounds into the price.
46:14 Um,
46:15 so it's,
46:15 it's,
46:16 you know,
46:16 in theory really bad.
46:17 So let me show you 5 stylized facts and what,
46:20 uh,
46:20 where are we going,
46:21 where are we heading with this.
46:23 The first one is that when you split
46:25 the,
46:26 the sample of firms into different bins,
46:28 so here the bottom 90%.
46:31 Based on their total sales,
46:33 the next 9%,
46:35 the next
46:35 0.9% and the
46:39 The largest group of firms,
46:40 the top 0.1%,
46:42 so these are essentially
46:44 almost 400 firms.
46:45 So you can see that
46:46 all the activity
46:48 at least in Argentina is concentrated here,
46:50 so like 57%
46:52 is explained by this group.
46:55 The second fact that I want to show is that these firms have multiple activities.
46:59 So we have data at the firm level,
47:02 firm month activity level.
47:04 So you can see that
47:05 in red,
47:06 we have the top 1%,
47:07 the last triangle is the top 1%.
47:10 The number of activities that they have is,
47:12 you know,
47:12 like 9.
47:13 So they have
47:14 some services,
47:15 transportation,
47:16 and then they sell something and so on.
47:20 The 3rdtu fact is that they operate in multiple jurisdictions,
47:23 so we know where these firms are selling their goods.
47:27 Right?
47:27 And we see that
47:29 this increases with firm size,
47:30 so we have percentiles of annual sales.
47:33 We see that
47:34 the top,
47:35 the,
47:35 the largest 400 firms.
47:38 Um,
47:38 they have,
47:39 they operate in about on average 20 jurisdictions.
47:43 There are 24 in Argentina,
47:45 so they sell almost to
47:47 to the entire country.
47:50 The fourth fact is that there are multiple rates
47:53 um that increase with firm size,
47:55 so they are relative,
47:56 so there are multiple rates.
47:58 So here we see for different sectors,
47:59 primary sector,
48:00 manufacturing,
48:01 retail,
48:01 financial,
48:03 and,
48:03 and services.
48:04 At the very top
48:05 were,
48:06 were.
48:07 That explains most of the macroeconomy,
48:09 we saw that they increase because there are some progressive rates
48:12 that they have introduced over time.
48:13 So that explains why the,
48:15 the revenue has increased
48:17 er
48:18 for this particular tool.
48:20 And the last
48:22 fact that I want to show is that dispersion,
48:24 tax rate dispersion.
48:26 Also increase it with firm size.
48:29 And this is concerning.
48:31 Because
48:32 the dispersion of tax rate is one of the sources of misallocation in an economy,
48:36 right?
48:37 So here what we are
48:39 doing in this work is we are applying the framework of CALO
48:43 to measure what is
48:45 the
48:45 size of these misallocations
48:47 and what would happen
48:49 in a particular country,
48:50 in this case Argentina,
48:51 if they if they um
48:54 eventually remove this,
48:55 this expression.
48:56 So if they
48:56 apply uniform rates
48:58 somehow
48:59 and again this is,
49:00 this is a framework that could be used
49:02 um in other settings with similar
49:05 tax databases
49:06 and,
49:06 and so on.
49:11 5 minutes.
49:14 Last topic that I want to mention is property tax.
49:17 Property tax
49:19 As I mentioned before,
49:20 it's underutilized.
49:22 It can help raise revenue equitably.
49:24 Why?
49:24 Because real estate is a major repository of wealth,
49:27 uh,
49:27 in,
49:28 in many societies.
49:31 Historically,
49:33 Municipalities have applied
49:35 flat.
49:37 Proportional tax schedules,
49:38 even like fees,
49:40 like a fixed fee,
49:41 an amount of
49:42 money.
49:43 The tax base also varies a lot across
49:45 municipalities.
49:47 So it's a tax that I think it's archaic in design
49:50 and
49:51 governments haven't really
49:52 thought about
49:53 improving their design because it's like they inherited
49:56 this and then it's very hard to change.
49:58 And what I,
49:59 what I think is surprising is that it constant,
50:01 constant
50:03 contrast sharply with
50:05 modern
50:06 tax tools such as personal income tax or
50:08 wealth taxes that are progressive in nature.
50:11 So that to me is uh uh surprising.
50:14 So what do we do in a recent paper we study a progressive reform
50:17 in uh Trese Febrero.
50:20 Uh,
50:20 it's a large urban,
50:22 urban municipality in,
50:23 in Buenos Aires,
50:24 Argentina.
50:25 It's characterized by law enforcement
50:29 and low compliance.
50:31 So here we have a few
50:32 of the
50:34 share of taxpayers.
50:35 We have the,
50:36 the,
50:36 so it's it's a tax that they pay monthly.
50:39 So that's why on the horizontal axis we have the 12 monthly installments.
50:43 And you can see that
50:44 roughly like 50% of the people pay the tax,
50:47 and it's very polarized.
50:48 So they either pay the 12 installments or they don't pay it.
50:52 And this is,
50:53 I think it's a style fact that would be common in,
50:55 in many settings.
50:57 So what do we do in this paper,
50:59 we study a progressive reform.
51:01 This is a reform that
51:02 for political reasons and for equity concerns,
51:05 the government decided to provide a,
51:08 um,
51:08 so the,
51:09 the schedules which is the blue one,
51:11 was historically regressive at the beginning and
51:14 then becoming progressive at the top.
51:16 They decided to give a
51:17 tax cut
51:19 for low-value properties.
51:20 We are going to call these low-value properties poor.
51:23 And it was based on a threshold
51:25 based on the evaluation of the property,
51:27 so
51:27 properties valued at 750,000
51:30 pesos,
51:31 less than that,
51:32 they experienced a VAT cut which is a red line.
51:35 Those in the middle,
51:36 no changes,
51:37 and then at the top there was a
51:39 VAT,
51:40 um,
51:40 a property tax increase.
51:44 So with this,
51:45 what do we do?
51:46 We analyze the effect of making the tax progressive
51:49 on compliance
51:51 and revenue,
51:52 and what we do is we focus on the thresholds,
51:54 the threshold
51:56 where there was,
51:57 there is a
51:58 tax cut for the poor and the threshold where there is a tax increase for the rich.
52:02 We analyze the direct effects on compliance
52:05 and then we also have an information an information experiment where we
52:09 tell the people what's going on with the other people,
52:12 and that will become clearer now.
52:15 So what is the causal effect of making a tax progressive?
52:18 What we find is for,
52:19 for poor,
52:20 so here is a chart where we put the
52:23 property value on the horizontal axis and the
52:25 change in tax compliance on the vertical axis.
52:28 We see that
52:30 those to the left of the threshold,
52:31 we call it poor,
52:32 those to the right are the middle.
52:34 We see that tax compliance is increasing.
52:37 For those who received a tax cut.
52:41 For the rich,
52:41 we see the opposite,
52:42 different threshold,
52:44 those to the right are the rich,
52:45 we see the tax compliance decreases.
52:48 We have some elasticities in the paper,
52:50 but
52:51 I want to spend time on that.
52:53 And then for the cross rate effects,
52:55 what do I mean by cross rate effects?
52:58 We have an experiment where we
53:00 tell
53:01 randomly
53:02 some groups in the,
53:05 some poor people we tell
53:07 that the rich are not paying more taxes.
53:09 And then for rich people,
53:11 we tell,
53:12 for some of them randomly we tell
53:14 them that
53:15 the poor are paying less taxes.
53:17 When we do that,
53:18 we follow,
53:19 um,
53:20 we measure the,
53:21 the changing,
53:22 so for the poor,
53:23 what we do is we measure the effect
53:25 of receiving a letter
53:27 on tax compliance.
53:29 What we compare is a letter,
53:31 so it's
53:32 within the poor group,
53:33 some are receiving a letter just saying your taxes are lower,
53:36 and the other group is your taxes are lower and the rich are paying more.
53:40 When we do that.
53:42 Those that also learned that the tax became more progressive,
53:46 there is like an extra effect on tax compliance over
53:49 and above the direct effect that I showed before.
53:53 And then for the middle and the and the rich,
53:56 it's,
53:56 it's,
53:56 it's less clear but we see that there is a little bit of uh a decrease for,
54:00 for the rich.
54:00 So when,
54:01 when the rich learn that the tax became more
54:04 And that the tax
54:05 decrease for the poor,
54:06 there is an extra negative effect on tax compliance,
54:09 uh,
54:09 for that group.
54:11 So with this um
54:13 I'm gonna stop.
54:14 So this uh food,
54:15 I think I,
54:16 I believe I provided food for thought um based on the several illustrations that I
54:21 provided today.
54:23 I wanna argue or what I argue in the
54:25 talk is that domestic resource mobilization is at the policy
54:29 data crossroads
54:30 and the data analytics can help pave the way for better tax policy and
54:36 administrations.
54:37 Um,
54:38 modest aside,
54:38 I believe we are very good at this here at the bank and,
54:41 and please reach out to us at the and Data,
54:45 um.
54:46 To,
54:47 to,
54:47 to discuss this more.
54:49 And the last thing that I want to mention is that
54:51 what I mentioned in this talk is just really scratching the surface.
54:54 It's the tip of the iceberg.
54:56 There is a flagship report we are producing on
54:59 domestic resource mobilization with Pierre Mavis,
55:02 Oyebola and many collaborators and,
55:05 uh,
55:05 Tiago and so on,
55:06 and I,
55:07 I,
55:08 uh,
55:08 want you to stay tuned because this is gonna come out soon.
55:11 So thank you so much and looking forward to the discussion.
55:26 Yeah it's
55:43 Great.
55:44 So thanks,
55:45 thanks very much for inviting me.
55:47 Um,
55:48 this is,
55:48 uh,
55:49 really a,
55:49 a pleasure.
55:50 I think that we're all super happy to have this kind of evidence,
55:54 uh,
55:54 provided,
55:55 um,
55:55 as we're all interested in.
55:58 How,
55:58 um,
55:59 how,
56:00 how to leverage some of these powerful digital tools and data
56:04 to help governments not only administer taxes better
56:08 and more effectively but also to see to what extent that,
56:12 uh,
56:12 that new data and and analytics can
56:15 improve the,
56:16 the,
56:17 uh,
56:17 range of implementable
56:19 policies in a way that
56:21 both efficiency and equity are improved.
56:24 Um,
56:24 so,
56:24 uh,
56:24 what I'm going to argue is that res despite this,
56:27 uh,
56:27 recent growth in,
56:29 in evidence,
56:30 we,
56:30 we are very far away from
56:32 what we really need.
56:33 Um,
56:34 there's a huge agenda,
56:35 I think,
56:36 on the,
56:36 on the horizon on this,
56:38 and I,
56:38 I,
56:38 I think I would,
56:39 um,
56:40 Uh,
56:41 argue that a lot of the questions that policymakers really wanna know
56:46 have to do with how economic agencies agents will respond to policy changes.
56:52 So
56:53 how will consumers respond?
56:56 How will suppliers respond when it's,
56:58 uh,
56:58 about VAT,
57:00 um,
57:01 and,
57:01 and to some extent that behavioral response partly depends on how much they know.
57:06 But it also partly depends on the capacity of uh government
57:10 government
57:11 to inform
57:12 and also
57:13 to what extent the government is able to,
57:16 you know,
57:16 administer in a way
57:18 that um takes advantage of all the data analy analytics that that exist
57:23 so there's a I think a lot of questions around
57:26 those behavioral responses on both the supplier and,
57:29 and,
57:29 uh,
57:30 if it's if it's a personal income tax,
57:32 it'll be also on.
57:33 Uh,
57:34 or corporate income tax on,
57:35 on,
57:35 on the economic agents sort of feeling this pain.
57:39 So,
57:39 um,
57:40 the work that.
57:42 Uh,
57:42 Dario just showed us,
57:44 uh,
57:44 um,
57:44 with a kind of an example on,
57:47 uh,
57:47 Argentina for VAT is really interesting.
57:50 Um,
57:51 it's similar to,
57:52 to what
57:53 I mean,
57:53 other,
57:54 other places have found,
57:55 but more importantly,
57:57 what is interesting is the,
57:59 the unintended consequences,
58:00 right?
58:01 So in the case of,
58:01 uh,
58:02 Argentina and Mexico there was very little
58:05 reduction in prices,
58:06 no reduction in prices in the case of Peru.
58:09 So a lot of,
58:09 a lot of the
58:10 reduction in VAT was pocketed by the suppliers.
58:14 Um,
58:14 there
58:15 is some evidence that in fact,
58:17 uh,
58:17 inequality increased as opposed to
58:20 you know,
58:21 declining,
58:21 which was part of the intent of the policy,
58:24 um,
58:25 and this again has to do with the behavior of,
58:27 of firms profiteering,
58:30 uh,
58:30 particularly where low,
58:31 low income households shop.
58:33 Um,
58:34 there's no discernible income on sales or employment in the case of Peru,
58:38 and there's some evidence of reducing compliance in the case of,
58:42 um.
58:43 Uh,
58:44 uh,
58:44 again,
58:44 of Peru,
58:45 um,
58:46 so my question with this and,
58:47 and particularly in the context of very tight
58:50 budgets in most of our developing countries is
58:53 could we expect the opposite if increases in VAT were on the table,
58:59 um,
58:59 and I don't think,
59:00 I don't think that that's necessarily the case
59:02 and so this kind of asymmetry of VAT going up or down is,
59:05 is really a testable
59:08 question.
59:08 Um,
59:09 there's,
59:09 uh,
59:09 some
59:10 really,
59:10 uh,
59:11 nice evidence from Mexico from the 2014 reform which had
59:15 an increase in VAT,
59:17 uh,
59:18 where,
59:18 uh,
59:19 Pierre Bachas found,
59:20 um,
59:21 that,
59:21 uh,
59:22 the pass through is,
59:23 uh,
59:24 in formal firms is actually
59:26 about 77%.
59:28 Uh,
59:28 obviously no pass through in more informal settings,
59:33 um,
59:33 but this is sort of one country,
59:36 one data point,
59:37 not clear if
59:38 this is the case for lower and low,
59:41 low middle income countries.
59:43 It's also unclear that,
59:44 um,
59:46 what,
59:46 what it means for revenue and equity and efficiency,
59:50 particularly when,
59:51 um,
59:52 taking into account the.
59:54 The fact that firms
59:56 or or consumers can switch very easily from shopping in a,
1:00:00 in a modern modern formal store to an informal store,
1:00:03 that elasticity of moving
1:00:06 your choice of where you shop
1:00:08 is especially important in
1:00:10 very low income settings where
1:00:12 informality is the norm.
1:00:13 Um,
1:00:14 it's not,
1:00:15 not uncommon to see high income people
1:00:18 buying food in,
1:00:20 let's say,
1:00:20 um,
1:00:21 street markets and,
1:00:22 and things in places like that,
1:00:24 um,
1:00:24 in developing countries,
1:00:26 so that elasticity
1:00:27 is,
1:00:28 is important to figure out.
1:00:29 Second,
1:00:30 to what extent
1:00:32 would,
1:00:32 uh,
1:00:33 VATCIT integration help?
1:00:36 Some tax administration offices are now able to
1:00:40 link.
1:00:41 Up
1:00:41 sort of VAT receipts with their corporate income tax receipts
1:00:45 and they may have an incentive to become more formal.
1:00:48 So the question is how much
1:00:50 of the reduction in compliance in sort of traditional stores
1:00:54 that are formal but potentially have a large informal,
1:00:57 um,
1:00:58 sort of
1:00:59 impetus.
1:01:00 Uh,
1:01:00 I always think about in my own country
1:01:02 in Bolivia you can.
1:01:04 You,
1:01:05 you go to a formal store,
1:01:06 you'll get asked,
1:01:07 do you wanna pay,
1:01:07 do you want a receipt or no?
1:01:09 And that determines the price,
1:01:10 right?
1:01:11 You,
1:01:11 you,
1:01:11 you,
1:01:11 and you can negotiate
1:01:13 how much of the
1:01:15 VAT savings
1:01:17 goes to the,
1:01:18 to the supplier and,
1:01:19 and how much goes to the consumer who,
1:01:21 who stop,
1:01:22 who doesn't pay,
1:01:23 pay it.
1:01:23 So that kind of negotiation is also interesting.
1:01:26 Um.
1:01:30 Um,
1:01:30 and then another question is how,
1:01:32 what would,
1:01:33 what can we expect from eliminating VAT exemptions?
1:01:36 So most countries are not necessarily considering increasing VAT,
1:01:40 but rather getting rid of
1:01:41 exemptions.
1:01:42 This,
1:01:43 uh,
1:01:43 figure comes from,
1:01:44 from some work done across multiple countries around
1:01:48 the world using our fiscal microsimulation modeling.
1:01:51 Um,
1:01:52 which essentially shows that VAT exemptions are largely
1:01:55 going to the top of the distribution,
1:01:57 mostly because,
1:01:58 of course,
1:01:58 the rich consume more.
1:02:00 So it's not
1:02:01 surprising
1:02:02 that those benefits go to the top more than the bottom.
1:02:05 But in addition to,
1:02:07 to that simple fact,
1:02:08 you also have,
1:02:09 there is also some evidence that larger firms are much more likely to claim,
1:02:14 claim VAT exemptions or reduced rates.
1:02:16 Um,
1:02:17 then,
1:02:18 then,
1:02:18 uh,
1:02:19 say smaller or more
1:02:21 informal firms,
1:02:22 more not informal,
1:02:23 but you know,
1:02:24 smaller,
1:02:24 smaller stores
1:02:25 where,
1:02:26 which is where a lot of the,
1:02:27 uh,
1:02:28 poorer population potentially is,
1:02:30 uh,
1:02:30 consuming.
1:02:31 So there's good reason to believe that eliminating
1:02:34 these uh VATT exemptions will improve efficiency.
1:02:37 It'll raise revenue,
1:02:38 etc.
1:02:39 It's very politically difficult to do as,
1:02:41 uh,
1:02:41 Dario was mentioning
1:02:42 it,
1:02:43 but,
1:02:43 um,
1:02:44 but.
1:02:45 The,
1:02:45 the,
1:02:46 you know,
1:02:47 one could do some simulations.
1:02:48 We did some simulations showing that,
1:02:50 oh,
1:02:50 if you
1:02:51 eliminate these exemptions and then you give households,
1:02:55 um,
1:02:56 you know,
1:02:56 safety nets or
1:02:58 some other form of mitigating mechanism,
1:03:01 you'd,
1:03:01 you'd be better off.
1:03:02 You'd reduce
1:03:03 inequality,
1:03:04 you'd reduce poverty,
1:03:05 etc.
1:03:06 We,
1:03:06 We often do these things to justify
1:03:09 some of the the the policies,
1:03:10 but actually impacts are are uncertain because we don't know
1:03:15 how
1:03:15 these kinds of reforms will affect compliance.
1:03:18 They're hard to to estimate,
1:03:21 um,
1:03:22 will suppliers become more,
1:03:24 you know,
1:03:25 be willing to do,
1:03:26 uh,
1:03:27 uh.
1:03:28 Capture some of that additional or uh revenue to what extent will,
1:03:33 uh,
1:03:33 consumers decide if you eliminate exemptions,
1:03:36 will they decide to go to the informal markets instead?
1:03:39 How much can you really expect in terms of revenue collections?
1:03:42 How much can you really expect
1:03:44 from,
1:03:45 um,
1:03:45 those changes?
1:03:47 And then it's unclear how,
1:03:48 how the enforcement
1:03:50 will actually affect equity and efficiency,
1:03:53 because the enforcement on its own,
1:03:55 even if you didn't no change in tax policy,
1:03:57 also has a distributional impact.
1:04:00 And
1:04:00 that again is unclear.
1:04:02 We don't know,
1:04:02 so many questions there.
1:04:04 Um,
1:04:04 PIT,
1:04:05 um,
1:04:06 Dario showed a very nice example from Honduras.
1:04:11 Um,
1:04:11 I think there are several examples around in the literature
1:04:14 that really focus on the fact that,
1:04:16 uh,
1:04:17 high income earners can play around.
1:04:20 And,
1:04:20 uh,
1:04:21 with their either with their undistributed profits
1:04:23 or the capital income and wealth,
1:04:25 uh,
1:04:25 so that they can uh
1:04:27 sort of declare it in whatever is most
1:04:30 convenient to them.
1:04:31 So tax a lots of tax avoidance happening at the top.
1:04:35 And,
1:04:36 um,
1:04:36 I guess a big question for all of us is to what extent,
1:04:39 um,
1:04:40 do international information sharing and coordination.
1:04:43 Policies help particularly in low income countries that have limited capacities,
1:04:48 um,
1:04:48 so,
1:04:49 and how,
1:04:49 how,
1:04:50 what is the best way?
1:04:51 What is the lowest hanging fruit
1:04:53 when you're dealing with a low capacity
1:04:55 country to actually make use of some of these,
1:04:58 um,
1:05:00 you know,
1:05:00 innovations in international coordination.
1:05:04 And then finally,
1:05:05 um,
1:05:05 with respect to the work that he presented on,
1:05:08 on turnover taxes,
1:05:10 even though he was presenting mostly for subnational,
1:05:13 I would argue that in a lot of countries,
1:05:15 2/3 of countries in Africa have some form of simplified tax regimes for SMEs.
1:05:20 These are turnover taxes
1:05:22 for small medium enterprises.
1:05:25 Um,
1:05:25 and here there are a lot of questions
1:05:28 that would,
1:05:29 you know,
1:05:30 that have to do more
1:05:31 with how these,
1:05:32 um.
1:05:34 These turnover taxes are actually implemented and the and the potential
1:05:38 efficiency and equity
1:05:40 uh considerations that come out of it.
1:05:42 So for instance,
1:05:43 a lot of,
1:05:44 um,
1:05:44 these,
1:05:45 uh,
1:05:45 uh,
1:05:46 turnover taxes don't have a tax-free threshold.
1:05:48 So you have
1:05:49 even very subsistence level firms
1:05:53 being liable for these turnover taxes,
1:05:56 um,
1:05:56 even though they're,
1:05:57 you know.
1:05:58 They really should be exempt,
1:06:00 um,
1:06:01 and in countries where they do have a little bit of progressivity,
1:06:04 um,
1:06:05 at the end of the day,
1:06:06 the,
1:06:07 the,
1:06:07 the,
1:06:08 the effective rates are not progressive partly because
1:06:11 the lowest income people don't necessarily have the information about
1:06:14 what's what the tax policy really is and how to,
1:06:18 uh,
1:06:19 how to,
1:06:19 you know,
1:06:20 benefit from,
1:06:21 from,
1:06:21 from that.
1:06:23 The sort of some that progressivity
1:06:25 and then finally,
1:06:26 uh,
1:06:26 on the,
1:06:27 on the property taxes,
1:06:28 I think
1:06:29 it's really interesting that you're showing the
1:06:31 differences in compliance between poor and non-poor
1:06:35 and I think more generally there's a lot of room here to do
1:06:38 kind of a combination of information.
1:06:40 Between what's available from administrative sources,
1:06:43 but also,
1:06:44 uh,
1:06:44 surveys on,
1:06:46 on how people are
1:06:47 perceiving,
1:06:48 um,
1:06:49 changes in policy
1:06:50 and therefore how their,
1:06:52 uh,
1:06:52 willingness to pay taxes changes over time.
1:06:56 With that,
1:06:56 let me end.
1:06:57 Thank you.
1:07:04 Thanks,
1:07:04 Gabriela,
1:07:05 and thanks,
1:07:05 Dario for the
1:07:07 presentation.
1:07:07 Um,
1:07:09 Um,
1:07:10 let's open it up for Q&A again.
1:07:12 If you have a question online,
1:07:13 please raise your hand,
1:07:14 and I can call on you.
1:07:16 Um,
1:07:17 if you're in the room,
1:07:18 please,
1:07:18 uh,
1:07:18 could you go to the mic?
1:07:20 Um,
1:07:21 just you wanna line up behind the microphone.
1:07:24 Hi,
1:07:24 uh,
1:07:25 very nice presentation,
1:07:26 Dario.
1:07:26 I have a question for you.
1:07:28 So if I understood correctly,
1:07:29 there's a lot of evidence within Argentina and across
1:07:32 countries that this pass through of the VAT can be
1:07:37 almost full,
1:07:37 can be 50%.
1:07:39 So do we know what drives these differences and connected to what uh the discussion,
1:07:45 uh,
1:07:46 uh,
1:07:46 it was being discussed,
1:07:47 the role of information,
1:07:48 no?
1:07:48 So is it like in.
1:07:51 Big retailers is easier to,
1:07:52 you know,
1:07:53 like accompany these policies with information campaigns of OK,
1:07:56 we're removing this tax so you should expect this
1:07:59 uh pass through or this cut in this percent
1:08:02 or informational campaigns of
1:08:05 incentivizing shopping around,
1:08:06 you know,
1:08:07 you should be expecting this to,
1:08:08 you know,
1:08:08 cost this much so look out for,
1:08:10 you know,
1:08:11 so what do we know about uh
1:08:14 how to,
1:08:14 you know,
1:08:15 implement these,
1:08:15 uh,
1:08:16 policies along with other campaigns?
1:08:19 So let's take a couple,
1:08:21 um.
1:08:23 Hey Dario,
1:08:24 uh,
1:08:24 indeed great presentation.
1:08:26 I definitely learned a lot,
1:08:27 um,
1:08:28 I was wondering,
1:08:29 uh,
1:08:29 in terms of like taxes.
1:08:30 I mean there's also some taxes that try to.
1:08:34 Overcome some inherent market failures that are there thinking of,
1:08:38 you know,
1:08:39 VAT taxes on
1:08:41 uh products with high sugar content or carbon taxation or congestion taxes
1:08:46 so
1:08:46 given
1:08:47 yeah
1:08:48 that they try to,
1:08:49 um,
1:08:49 you know,
1:08:50 overcome some market failures,
1:08:51 is there some work or should we try to focus more on taxes that are
1:08:55 less distorting by itself and maybe go more towards
1:08:59 an efficient market equilibrium?
1:09:00 Thanks.
1:09:03 And then
1:09:06 Um,
1:09:08 well,
1:09:08 again,
1:09:08 thank you again for,
1:09:09 for a very interesting,
1:09:10 it's a,
1:09:10 it's really a nice summary of all this work
1:09:13 that that that you and the colleagues have done.
1:09:15 I just have a comment regarding the subnational exactly property taxes,
1:09:20 and you mentioned how
1:09:22 you're surprised why not as many because data is there
1:09:25 fortunately I have to disagree there is really not there.
1:09:28 I mean not enough of it,
1:09:30 uh,
1:09:30 in,
1:09:31 in most countries unfortunately.
1:09:33 It's very difficult to get um micro level data on property taxation
1:09:38 uh because um.
1:09:41 There is no
1:09:43 consolidated database where you can get it or
1:09:46 then you have to travel around the country,
1:09:48 you know,
1:09:49 search by municipality,
1:09:51 uh.
1:09:52 Unless you focus on one,
1:09:53 but the,
1:09:53 the reason why is it's important to actually have a country level
1:09:57 kind of scope of data for,
1:09:59 for property tax is that
1:10:01 it's not really only,
1:10:03 I mean,
1:10:03 the,
1:10:03 um,
1:10:04 we claim that it's actually equitable tax but like all these over the last decades
1:10:09 studies have shown that it's,
1:10:10 it's regressive.
1:10:12 And there are different reasons why that is.
1:10:14 One of them is that,
1:10:15 um,
1:10:15 theoretically it's an,
1:10:16 it's a progressive tax,
1:10:17 of course it depends on how you see it.
1:10:19 Is it an excise?
1:10:20 Is it,
1:10:20 uh,
1:10:21 it's a tax on income,
1:10:22 but also,
1:10:23 um,
1:10:24 depending on the structure
1:10:26 you may end up with actually a regressive tax,
1:10:28 um,
1:10:28 and you show the example of Argentina where they had a tax break on the lower value
1:10:35 properties
1:10:35 and I would be very curious to see if you could
1:10:37 do the similar study in about 5 to maybe 10 years.
1:10:41 And see what happens because what theory suggests is that
1:10:46 such a policy is gonna
1:10:48 allocate
1:10:50 income toward the lower income housing
1:10:52 where you're gonna have now
1:10:53 higher income individuals owning lower income housing
1:10:57 where it's gonna increase their value and lower income individuals are
1:11:00 not gonna be able anymore to afford actually low income housing.
1:11:04 So,
1:11:04 um,
1:11:04 uh,
1:11:05 this is the,
1:11:06 the,
1:11:06 the limitation in,
1:11:07 I mean,
1:11:08 besides of all the issues that the policy design actually
1:11:11 can change so for property tax is really important.
1:11:13 What is only progressivity also progressivity in the medium and long term.
1:11:17 And another thing is for property taxes if no other taxes and,
1:11:21 uh,
1:11:21 implementation is here crucial because in addition to just
1:11:25 tax design you're gonna have
1:11:27 what is gonna impact progressivity is also simply.
1:11:31 Administration like you're gonna have tax assessment in
1:11:34 some countries but I'm sorry pro property value assessment
1:11:37 it's gonna be more favorable for high income house.
1:11:39 In fact,
1:11:40 uh,
1:11:40 there's some a study I from Italy I can
1:11:43 share with you that shows that high income actually
1:11:45 properties are under
1:11:47 assessed
1:11:48 relative to lower income.
1:11:50 So all these different,
1:11:51 I mean certainly it's a very,
1:11:53 um,
1:11:53 it's very useful but
1:11:55 just ideas if you have a chance to do it would be great to have such evidence.
1:11:59 Thank you.
1:12:00 OK,
1:12:00 let's take one more question and then I'll
1:12:02 get back to you.
1:12:03 Thanks.
1:12:05 Great presentation Dario.
1:12:06 Um,
1:12:07 I was wondering about how to think about the
1:12:10 political economy which came out throughout your presentation.
1:12:13 It seemed like there was some lack of clarity,
1:12:16 I mean,
1:12:16 uh,
1:12:17 not in your presentation,
1:12:18 but
1:12:19 in the political economy literature there's no clear explanation for
1:12:24 what,
1:12:24 how to think about even sort of
1:12:27 tax revolts or recent movements we've seen in Africa and
1:12:30 for example,
1:12:31 in Kenya when a
1:12:33 tax reform was announced,
1:12:34 people were on the streets were they on the streets because it's populist,
1:12:38 they don't understand.
1:12:39 Uh,
1:12:40 that taxes are important,
1:12:42 or is it that something else is it lack of
1:12:44 trust that the government will use the taxes well?
1:12:47 So with that sort of general background,
1:12:49 I wanted to ask you two specific questions one.
1:12:52 Uh,
1:12:53 what is the role of the,
1:12:55 the data revolution
1:12:57 in addressing some of these political economy constraints?
1:13:00 So for example,
1:13:01 when you present the evidence on
1:13:04 the lack of pass through of a
1:13:07 reduction in the VAT,
1:13:09 does that change?
1:13:10 Is it surprising to the tax administrators?
1:13:13 And does it lead them to think about changing the policy
1:13:17 or communicating
1:13:18 to if the demand came was a populist demand from
1:13:22 poor lower income people,
1:13:24 does that help them communicate that look we tried using this
1:13:27 and now we understand from all this nice big data,
1:13:31 uh,
1:13:31 that it didn't work as intended
1:13:34 and second on the other side of the politics of elite capture,
1:13:38 maybe this is really a naive idealistic question,
1:13:41 but.
1:13:42 Is there any talk about moving towards something that's more like a lump sum tax?
1:13:47 I mean,
1:13:48 some of the top brackets you're presenting data on
1:13:51 are as few as 500 firms,
1:13:54 or,
1:13:54 you know,
1:13:55 80,000 people.
1:13:56 Um,
1:13:57 and some of the really high
1:13:59 brackets are the ones who are offshoring presumably we kind of know
1:14:04 who are these big entities.
1:14:06 Uh,
1:14:06 has there been any discussion of,
1:14:08 you know,
1:14:09 just more lump sum taxation per head,
1:14:12 uh,
1:14:13 and what are the,
1:14:14 what is the political economy of elite capture on the top?
1:14:17 Like,
1:14:18 presumably there's a lot of
1:14:20 resistance to that kind of a movement,
1:14:23 anything written on that.
1:14:24 Thanks so much.
1:14:26 Thanks.
1:14:26 A lot of wide ranging questions.
1:14:29 Go ahead,
1:14:29 Darryl.
1:14:33 I'm gonna start.
1:14:35 Hi,
1:14:35 yeah,
1:14:36 I will start from the,
1:14:37 the end
1:14:38 all the way.
1:14:39 I think it's easier.
1:14:41 I,
1:14:42 yeah,
1:14:42 I think Mavi is sitting behind you would be
1:14:46 better equipped to,
1:14:47 to,
1:14:48 but I think those are very,
1:14:49 very good points,
1:14:50 um,
1:14:51 I think you know what I know is a lump sum tax
1:14:53 like a poll tax that Margaret Thatcher implemented in the UK.
1:14:58 It was
1:14:59 complicated,
1:15:00 so I think I lump some I wouldn't see a space for such a policy,
1:15:05 um.
1:15:07 On the
1:15:08 VAT cards communication,
1:15:10 I think it's key.
1:15:11 I think there is a disconnect.
1:15:13 Uh,
1:15:14 between research and and
1:15:16 policy action,
1:15:18 and there should be more communication in the sense like policymakers
1:15:21 should try to communicate better like the fears that er.
1:15:24 Gabriela was showing with the,
1:15:26 the,
1:15:26 the incidence of eliminating tax expenditure,
1:15:29 that's key
1:15:30 to communicate the policy reform.
1:15:32 I don't really know how to do it.
1:15:34 Argentina has done it successfully in some cases like the,
1:15:37 the tax amnesty,
1:15:38 they were,
1:15:39 they were great communicating that like to,
1:15:41 you know,
1:15:42 entice participation.
1:15:43 I feel like that that was one of the,
1:15:46 one of a good example of what countries could do.
1:15:49 That's something I,
1:15:51 I discussed in the paper like this
1:15:52 part of the political economy,
1:15:54 but
1:15:54 I'm not an expert on that topic.
1:15:56 I think it's very relevant,
1:15:57 but
1:15:59 I wouldn't do enough justice
1:16:01 to that.
1:16:02 Um,
1:16:02 Violetta,
1:16:04 thank you for your comment.
1:16:05 That's great.
1:16:07 I think I,
1:16:07 I disagree on the part that
1:16:10 There is a lot of data,
1:16:12 and it requires a lot of entrepreneur,
1:16:14 entrepreneurship,
1:16:15 right?
1:16:15 And let me give you an example of um Juan Luis Caboni who works with us.
1:16:19 He used to work at a in a municipality in Argentina,
1:16:23 that didn't,
1:16:24 they didn't have any data analytics.
1:16:26 There was,
1:16:28 you know,
1:16:28 a computer,
1:16:30 data was sitting there,
1:16:31 so he took the computer
1:16:33 that they needed capacity.
1:16:34 He arranged all the files,
1:16:36 monthly files together,
1:16:37 and they build the data and now they are using it.
1:16:39 So I feel like every municipality has this type of data it's just a matter of
1:16:43 going knocking the door,
1:16:45 trying to be proactive and try to
1:16:47 make progress.
1:16:48 That's that's what I would like to do and
1:16:51 I agree with you that you know these sometimes are regressive,
1:16:54 we don't really know,
1:16:55 we need,
1:16:55 you know,
1:16:56 we have data from 5 municipalities from Brazil,
1:16:58 Colombia,
1:16:59 Argentina.
1:17:00 We put them together,
1:17:01 we can get like an overview of what's going on in Latin America,
1:17:04 and I think that's feasible.
1:17:08 On the market failures and I think it's it's a very good point.
1:17:12 I think the talk was more focused on,
1:17:14 I agree that externalities.
1:17:17 Can be,
1:17:18 can be justified.
1:17:19 Lower rates on goods that create externalities either positive or negative.
1:17:24 Uh,
1:17:25 or higher rates on those
1:17:26 with negative externalities that those are valid points and I
1:17:30 I didn't focus too much on the talk on that.
1:17:33 Um
1:17:36 Claudia on the.
1:17:39 Marketing in large stores that,
1:17:40 that's very interesting.
1:17:44 The way I think about this is
1:17:46 Walmart is a company that
1:17:49 they comply with taxes,
1:17:51 so you know when you go to a supermarket you always get a ticket.
1:17:54 So these type of places they typically use at least in Argentina they use it
1:17:59 they use these policies as a marketing device,
1:18:01 so they all of a sudden they put no VAT
1:18:04 because there is like an extra effect
1:18:06 on the month.
1:18:07 So if you,
1:18:08 if you don't put the sign,
1:18:09 the amount would be lower,
1:18:10 in other words.
1:18:12 I think supermarkets are fully aware of that.
1:18:15 It's different in
1:18:16 independent grocery stores where tax evasion might be 50%
1:18:21 and,
1:18:21 and it's like Gabriela was saying,
1:18:23 you know,
1:18:23 you go there
1:18:24 and you invite the,
1:18:26 the,
1:18:26 the owner to evade,
1:18:27 you say,
1:18:28 is it the same price if I pay in cash,
1:18:30 you're saying can we share the incidence of the of the evaded tax.
1:18:35 I think that there's a little bit different,
1:18:37 the marketing,
1:18:38 so I don't,
1:18:38 I don't know how much they want to
1:18:40 communicate this to consumers because to begin with,
1:18:43 they are not paying.
1:18:45 Um,
1:18:45 or they are evading 50% of the tax.
1:18:48 Um
1:18:50 This could be one of the reasons that explains
1:18:52 partial path through,
1:18:53 but there are different explanations.
1:18:55 One of them is imperfect competition,
1:18:58 and
1:18:58 I believe that the mechanisms,
1:19:00 we don't understand
1:19:01 them very well,
1:19:02 and there is space to do more research on this.
1:19:05 And then going
1:19:06 to Ari,
1:19:07 thank you very much for taking the time to
1:19:10 to read the papers and,
1:19:11 and
1:19:12 yeah,
1:19:12 and think about these problems.
1:19:14 I think it's a
1:19:17 It's very interesting,
1:19:18 your observation on the inequality.
1:19:20 I don't think we,
1:19:21 we talk much about that in the,
1:19:22 in the paper.
1:19:24 Uh,
1:19:24 in,
1:19:25 in Argentina,
1:19:26 um,
1:19:27 I also think it's very important to
1:19:28 measure the elasticity of switching across stores.
1:19:31 We don't have any evidence of that,
1:19:33 it's hard to get,
1:19:34 but it's,
1:19:34 it's an important topic
1:19:36 that people should be
1:19:38 thinking more carefully about.
1:19:40 Um,
1:19:43 Eliminating VAT exemptions,
1:19:45 yeah,
1:19:45 totally,
1:19:46 that,
1:19:46 that's so important,
1:19:47 it seems that there is some professional consensus that
1:19:51 many countries that's one of kind of a low hanging fruit,
1:19:55 politically very costly,
1:19:56 but to
1:19:58 improve a second best tax
1:19:59 such as the VAT.
1:20:01 And it has proven politically very costly like Bangladesh,
1:20:05 January 9th,
1:20:06 they
1:20:07 increased the,
1:20:08 they,
1:20:08 they
1:20:09 motionized the tax rates on 100 items,
1:20:12 1,
1:20:12 10 days later they retracted the decision
1:20:15 because they were starting to be protests.
1:20:17 I think that that part is,
1:20:18 I think it's very exciting,
1:20:19 it's something that I.
1:20:22 Maybe communication can can help like this type of figures that you show,
1:20:25 but another way to compensate with an extra,
1:20:29 an additional uh an alternative supplementary program.
1:20:33 Um,
1:20:35 yeah.
1:20:35 So,
1:20:36 yeah,
1:20:36 that's
1:20:37 one reflection.
1:20:38 And then for the turnover tax,
1:20:40 yeah,
1:20:40 I think it's um.
1:20:42 It's more prevalent than we think,
1:20:44 so local business taxes are prevalent in developed,
1:20:47 even in the US there are like 7 states that they have a gross receipt tax,
1:20:52 so that's that's the turnover of the US.
1:20:55 Tax rates are very low,
1:20:56 but
1:20:57 um.
1:20:58 It's very silent and we don't have enough evidence on this,
1:21:01 and yeah,
1:21:02 I would like to see more if possible.
1:21:05 So yes,
1:21:06 with this,
1:21:06 I think I'm gonna stop.
1:21:09 Great,
1:21:09 thank you.
1:21:09 Are there,
1:21:10 we have time for maybe one or two questions,
1:21:13 very,
1:21:13 very brief ones.
1:21:18 Nice presentation Dario.
1:21:19 Uh,
1:21:20 a couple of questions on,
1:21:21 on the local taxes.
1:21:24 Uh,
1:21:24 you show the distribution of different,
1:21:26 uh,
1:21:26 um,
1:21:27 sources of the taxation at the local level,
1:21:30 but a big chunk of the,
1:21:32 the,
1:21:32 uh,
1:21:32 tax revenue comes from transfer from the federal government
1:21:36 that they go and you don't display it,
1:21:38 uh,
1:21:38 so I don't know
1:21:39 how much of the,
1:21:40 uh,
1:21:41 tax revenue you're capturing with those graphs.
1:21:44 And the other,
1:21:44 uh,
1:21:45 point is you mentioned the thing about mobile money.
1:21:48 And the taxation of mobile money.
1:21:50 I was wondering when people pay,
1:21:52 uh,
1:21:52 with mobile devices,
1:21:54 are they evading,
1:21:55 for example,
1:21:56 the,
1:21:56 uh,
1:21:57 some of the taxes,
1:21:58 the VAT tax,
1:21:59 for example,
1:22:00 and if they can evade,
1:22:02 can you get to this point that Gabriela was mentioning
1:22:05 whether you can switch between across stores using the mobile payments.
1:22:12 Um
1:22:14 On the,
1:22:15 on the mobile money,
1:22:17 they,
1:22:18 if there is a VAT or like for instance Uganda or Tanzania,
1:22:22 I don't remember,
1:22:22 I think it's Uganda,
1:22:23 they have excise tax,
1:22:25 they have the VAT and then there is like a fee,
1:22:28 like,
1:22:28 so 3 taxes,
1:22:30 they do pay those taxes,
1:22:32 the one that I think is more relevant in terms of revenue is the fee
1:22:36 that they pay.
1:22:37 But
1:22:37 otherwise there would be some,
1:22:39 some,
1:22:39 in some cases there will be some taxes.
1:22:41 Sometimes these services,
1:22:43 these
1:22:44 transactions are not taxed,
1:22:46 and
1:22:47 I
1:22:47 think those are part of the
1:22:49 OECD pillar one,
1:22:51 if I remember
1:22:53 correctly.
1:22:53 And then on the local,
1:22:55 local
1:22:57 government,
1:22:58 it's true that from the figures I'm omitting grants
1:23:01 and transfers that come from the central government.
1:23:04 For provinces,
1:23:05 at least in Argentina,
1:23:06 that's a big chunk,
1:23:07 um,
1:23:09 so all revenues could be,
1:23:11 but,
1:23:11 but still,
1:23:11 you know,
1:23:12 subnational taxes can be,
1:23:15 can represent of the total
1:23:17 adding subnational and national taxes,
1:23:19 subnational taxes are
1:23:21 20%.
1:23:22 Of all the resources that Argentina,
1:23:25 so it's a significant amount of the GDP,
1:23:27 and it has remained constant over time.
1:23:29 So it's not that what is changing is the composition of national taxes,
1:23:33 but then the
1:23:34 share that they collect their own revenue is relatively constant over time.
1:23:38 So it's really the composition,
1:23:39 the way they are taxing
1:23:41 the different tax bases.
1:23:43 And for municipalities it is even more important like municipalities,
1:23:46 40% of the revenue that they get is own revenue,
1:23:49 on average,
1:23:50 40% is what they get and the rest is what they,
1:23:53 what the,
1:23:53 the.
1:23:55 The national government transfer to them.
1:23:57 So it's even more important for cities.
1:24:00 And
1:24:00 so I think that's why
1:24:03 even from a macro perspective it's
1:24:05 really important to understand the consequences.
1:24:08 OK,
1:24:09 we have one question online uh from Alistair.
1:24:11 Do you wanna unmute yourself and ask your question?
1:24:14 Uh,
1:24:14 but
1:24:14 please make it brief because we are running out of time.
1:24:17 Alistair.
1:24:22 Hi,
1:24:22 can you hear me?
1:24:23 Yes.
1:24:25 Uh fantastic.
1:24:26 Thanks,
1:24:26 excellent,
1:24:27 excellent presentation Dario.
1:24:29 I just,
1:24:29 we,
1:24:29 there was a bit of discussion here about the um
1:24:33 about the political economy of the VAT of VAT uh
1:24:36 and base broadening VAT reform and getting rid of the,
1:24:39 obviously very,
1:24:40 very badly targeted uh reduced rates.
1:24:42 Um,
1:24:43 one issue,
1:24:45 It's got quite a bit of attention recently,
1:24:48 so for example it's been
1:24:49 implemented in Uzbekistan,
1:24:51 in one of the regions of Brazil,
1:24:53 are these real time
1:24:55 back cashback regimes.
1:24:57 So I'll be interested in
1:24:59 your perspective and Gabriella is also on,
1:25:02 The feasibility of these
1:25:04 schemes in in
1:25:06 in low income countries and middle income countries,
1:25:10 particularly where,
1:25:11 The countries that
1:25:13 you would expect
1:25:14 to
1:25:15 have the technological capacity to implement these schemes,
1:25:20 Would likely also
1:25:23 be countries that
1:25:24 are able to effectively implement a traditional targeted cash transfer.
1:25:30 So I appreciate your views on that,
1:25:31 thank you.
1:25:34 And we have one more hand up,
1:25:35 so I'll just turn it over to Thomas.
1:25:37 Kanne,
1:25:38 can you unmute yourself?
1:25:45 Go ahead.
1:25:48 Uh,
1:25:49 we can't hear you.
1:25:52 Sorry,
1:25:53 OK,
1:25:53 so I'll turn it back to,
1:25:54 to Dario and Gabriela if you have anything to add.
1:25:57 Thank you.
1:25:58 Yeah,
1:25:58 thank you,
1:25:58 Alastair.
1:25:59 I think it's,
1:25:59 um,
1:26:00 yeah,
1:26:00 in some countries true that like Brazil,
1:26:02 Rio Grande do Sul,
1:26:03 and then in Pakistan it's.
1:26:06 They have the technical ability to have like a VAT cashback
1:26:09 program where they do rebates instead of having to use rates,
1:26:13 um,
1:26:14 I think it's feasible,
1:26:16 but I also,
1:26:16 I also think it's,
1:26:17 uh,
1:26:18 we,
1:26:18 we shouldn't expect it to be magic,
1:26:20 I mean,
1:26:20 the devil is
1:26:22 in the details,
1:26:23 and
1:26:23 I think these programs,
1:26:24 cashback programs,
1:26:25 they should be designed very carefully if they
1:26:28 really want to reach the target population,
1:26:30 um.
1:26:31 But it's,
1:26:32 it's true.
1:26:32 I have a,
1:26:33 I left in the,
1:26:34 the slides that will be posted online.
1:26:36 I put one in the appendix,
1:26:38 where I illustrate the,
1:26:40 the example of Argentina with a card
1:26:43 called Tartament which is for um.
1:26:46 The,
1:26:47 the
1:26:49 cash transfer recipients.
1:26:51 If they pay,
1:26:52 they receive the cash transfer in a debit card.
1:26:54 If they use a debit card in the in any like supermarket,
1:26:58 then there is like a cashback that they get automatically,
1:27:01 right?
1:27:01 So it's targeted to people that receive a cash
1:27:04 like a
1:27:05 cash transfer,
1:27:07 and then it's,
1:27:07 you know,
1:27:08 you're really reaching the people that
1:27:10 need it most,
1:27:11 and it's,
1:27:12 it's kind of automatic,
1:27:13 so we have
1:27:14 the systems to implement that.
1:27:17 So those are interesting policies and I'm.
1:27:19 I,
1:27:19 I,
1:27:20 I hope to see
1:27:21 much more in the coming years.
1:27:24 I,
1:27:25 I would second that.
1:27:26 I think that,
1:27:26 uh,
1:27:27 this is one way to get around the political economy difficulty of
1:27:32 eliminating tax exemptions.
1:27:34 And to the,
1:27:35 to the extent that you,
1:27:36 if you are a low-income person and you are a beneficiary of the tax
1:27:42 transfer system,
1:27:43 or sorry,
1:27:43 the cash transfer system,
1:27:45 so you receive social assistance,
1:27:47 there's a means,
1:27:48 means test somewhere that says that you,
1:27:50 You know,
1:27:51 you deserved a break.
1:27:53 Um,
1:27:54 instead of having to wait for,
1:27:56 you know,
1:27:57 paying the VAT and then getting,
1:27:59 getting cash,
1:27:59 um,
1:28:01 Through an alternative program you immediately get that
1:28:04 um
1:28:05 relief tax relief
1:28:07 uh in a targeted fashion so it it sort of ticks all the boxes in terms of efficiency.
1:28:12 It also gets to the political economy question very quickly
1:28:16 um
1:28:17 the the issue of course is
1:28:19 does the tax administration have the capacity to.
1:28:23 Merge
1:28:24 essentially the tax administration and the
1:28:25 and the social assistance administration databases in order to make that happen
1:28:32 and the and and here I would say
1:28:34 it's only gonna be the middle income countries that have
1:28:37 sophisticated tax authorities sophisticated
1:28:40 um
1:28:40 social assistance uh registries and and so on
1:28:43 and also the other
1:28:45 the other um.
1:28:46 I guess danger here is that a lot of the social assistance prop uh registries
1:28:52 even in upper middle income countries are not
1:28:55 going to fully target
1:28:57 everyone who's at the bottom of the distribution,
1:29:00 right?
1:29:00 So there are lots of,
1:29:01 uh,
1:29:01 leakages and and sort of.
1:29:04 People who should potentially be getting that relief that aren't going
1:29:07 to be getting relief just because of the eligibility criteria,
1:29:11 uh,
1:29:11 for that social assistance.
1:29:13 So there's that danger that if you have those
1:29:15 polls that you're not going to be getting it,
1:29:18 those,
1:29:18 uh,
1:29:18 populations,
1:29:19 but I think it it there it's,
1:29:20 there's a huge promise here.
1:29:21 I think the Brazilian example is beautiful
1:29:24 and,
1:29:25 and,
1:29:25 uh,
1:29:26 yeah,
1:29:26 hopefully more countries can do more of this.
1:29:29 Thank you.
1:29:31 Great,
1:29:31 thank you very much.
1:29:32 Um,
1:29:32 thank you everyone for joining.
1:29:33 Um,
1:29:35 the,
1:29:35 just a reminder,
1:29:36 the event recording will be,
1:29:37 uh,
1:29:38 posted online,
1:29:39 uh,
1:29:39 as well as the,
1:29:40 the slides and uh if you wanted to follow up.
1:29:42 For those of you who are on the World Bank's intranet.
1:29:46 Uh,
1:29:46 you can listen to,
1:29:48 uh,
1:29:48 the AI generated podcast of,
1:29:50 of Dario's paper on,
1:29:52 uh,
1:29:52 uh,
1:29:53 tax,
1:29:53 uh,
1:29:53 VAT tax cuts in Argentina
1:29:55 on the DC decoded,
1:29:57 uh,
1:29:57 podcast series.
1:29:59 Um
1:30:01 With that,
1:30:02 thank you Dario,
1:30:03 thank you,
1:30:03 Gabriela,
1:30:04 um,
1:30:04 for that really insightful,
1:30:05 um,
1:30:06 presentation and,
1:30:07 and discussion and thank you all.
1:30:09 See you.
1:30:20 You know
1:30:21 not.
1:30:23 Very.
1:30:27 No.
1:30:30 I.
- add-style
- lp-body-content
In many developing countries, tax policy is often driven more by pragmatism rather than the pursuit of optimal outcomes. Barriers such as informality, low tax capacity, political roadblocks, and limited data analytics often result in inefficient and inequitable tax structures that persist over time.
In this Policy Research Talk delivered on April 7, 2025, World Bank Economist Dario Tortarolo drew on recent research on both national and subnational taxes—including value added taxes (VAT), personal income taxes, wealth taxes, business income taxes, and property taxes— to show how new data-driven approaches are reshaping how governments design and evaluate tax systems in real time, enabling policymakers to move beyond educated guesses toward informed, evidence-based decisions.