Tackling Tax Evasion via Ghost Firms: Insights from Ecuador

Key Question Being Explored in This Insight:

How do ghost firms contribute to tax evasion, and what insights does transaction-level tax data from Ecuador provide about their prevalence and impact?

View the page of: Paul Carrillo

Institution: George Washington University

Explore more insights in:

Economics Microeconomics Social Sciences

Media Literacy Toolkit

Faculti's Media Literacy Toolkit helps viewers critically engage with academic insights by analyzing the research context, identifying perspectives, and encouraging thoughtful evaluation.

Critical Questions to Consider ▼

What assumptions does the research make?
Are there alternative perspectives not explored?
What are the limitations of the research method?

Bias and Perspective Awareness ▼

This research comes from George Washington University . Reflect on how the institution's academic focus and research partnerships may shape the questions being explored.

Further Reading ▼

Read the Study

How do ghost firms contribute to tax evasion, and what insights does transaction-level tax data from Ecuador provide about their prevalence and impact?

Image courtesy of interviewee. February 10, 2025

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month. Login to view 8000+ figures on the platform.

Generating Workspace...

Export as PDF

This video discusses the issue of ghost firms, which are fake entities created for fraudulent purposes, particularly in the context of tax evasion and fake transactions. It highlights the prevalence of such firms in developing countries and the challenges faced by tax authorities in identifying and combating them. This video presents a detailed analysis of how ghost firms operate and discusses a specific case study in Ecuador where a policy intervention targeted the clients of ghost firms to increase tax compliance. The study emphasizes the importance of collecting firm-to-firm transaction data and implementing effective policies to address the issue of ghost firms and tax evasion.

Language:

Export as PDF

Export as PDF
Based on the video , here are 5 key points:

1. **Identification of Fake Firms**: The video discusses the existence of fake firms that are incorporated by criminal organizations in developing countries with the sole purpose of issuing fake receipts for other firms to claim deductions on their taxes.

2. **Extent of the Issue**: Between 2010 and 2015, about 10% of firms in the sample traded at least once with a ghost firm. These ghost transactions accounted for approximately 14% of total purchase deductions among these firms.

3. **Correlation with Firm Size**: Contrary to initial expectations, there is a notable correlation between the size of a firm and the likelihood of trading with a ghost firm, with larger firms being more likely to engage in such transactions.

4. **Characteristics of Ghost Transactions**: Analysis of transaction data with ghost firms revealed suspicious patterns, such as transactions clustered at the end of the fiscal year, the prevalence of round numbers and multiples of 500, and an avoidance of transactions just above the $5,000 threshold.

5. **Policy Intervention**: The Ecuadorian tax authority implemented a targeted policy where they focused on the clients of ghost firms and requested them to amend past tax returns in cases of suspected fake transactions. This intervention was found to be successful, increasing total tax liability by about $20 million.

Please let me know if you need further clarification or additional information.

Export as PDF
Key Quotes from the Text:

1. "Those firms essentially are fake firms. These are firms that do not produce any products, do not give any services. But these are firms that have been incorporated mostly by criminal organizations with just one purpose to become invoice mails, to issue fake receipts."
2. "If we just focus, let's say on one year, 2015, we found that they [firms] traded at some point with a ghost firm. And those ghost transactions comprised about 14% of the total purchase deductions among these firms. So this is large, this is big, right?"
3. "So if you, at least in this convideo, in Ecuador, if you wanted to purchase something from a firm above $5,000, well, you cannot use cash. You had to go through the bank or just a credit card or something that, and then you have a financial trail, you actually have to make a payment."
4. "We find that, well, first finding is that about one fourth of firms respond to the notification. Among the firms that actually amended the tax returns, the policy was extremely effective in boosting the tax liability by almost 80%, which collectively increased total tax liability by about $20 million."
5. "Collecting firm-to-firm transaction level data is very important for many reasons, but also because it allows you to spot and pinpoint cases where this type of tax evasion occurs."

These key quotes summarize the issues surrounding ghost firms, the impact of their transactions, the methods of tax evasion, and the effectiveness of targeted policies in addressing the problem.

X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:90000

1
00:00:00.055 --> 00:00:02.645
Those firms essentially are fake firms.

2
00:00:03.375 --> 00:00:07.365
These are firms that do not produce any products,

3
00:00:07.385 --> 00:00:08.565
do not give any services.

4
00:00:09.225 --> 00:00:11.725
But these are firms that have been incorporated mostly

5
00:00:11.825 --> 00:00:14.605
by criminal organizations with just one purpose

6
00:00:15.185 --> 00:00:18.725
to become in invoice mails, to fake,

7
00:00:19.025 --> 00:00:20.445
to issue fake receipts.

8
00:00:21.025 --> 00:00:24.565
And those fake receipts are going to be used by other firms

9
00:00:25.105 --> 00:00:28.125
to claim deductions on their value added taxes

10
00:00:28.625 --> 00:00:30.605
or on their corporate income taxes.

11
00:00:31.425 --> 00:00:33.965
So this is a criminal, criminal enterprise.

12
00:00:34.455 --> 00:00:35.525
These are fake firms,

13
00:00:36.665 --> 00:00:41.005
and it's a big problem, particularly

14
00:00:41.545 --> 00:00:42.845
in developing countries.

15
00:00:43.545 --> 00:00:46.045
Um, if you talk to tax authorities all over the world,

16
00:00:46.475 --> 00:00:48.805
they're gonna tell you that they are aware of this problem.

17
00:00:49.215 --> 00:00:52.885
Coast firms or invoice mails are a topic of concern.

18
00:00:52.885 --> 00:00:55.805
International organizations know about this problem,

19
00:00:57.265 --> 00:01:00.965
and surprisingly, there's little academic research on it,

20
00:01:01.105 --> 00:01:03.685
and that's why we were so interested in this topic.

21
00:01:03.945 --> 00:01:05.445
We have two main objectives,

22
00:01:05.445 --> 00:01:08.325
and I'm going to elaborate more, more on that, as we,

23
00:01:08.785 --> 00:01:10.965
as we talk about the details of the paper.

24
00:01:10.985 --> 00:01:14.525
But we want to, we're gonna use Ecuador as cases study

25
00:01:15.435 --> 00:01:16.935
to learn about the ghost economy.

26
00:01:17.275 --> 00:01:19.535
You know, how big the ghost economy is,

27
00:01:20.195 --> 00:01:24.655
how common transactions from firms to ghost firms are.

28
00:01:25.675 --> 00:01:27.175
What's the magnitude of this problem?

29
00:01:28.235 --> 00:01:31.575
Eh, that's the first, that's the first objective.

30
00:01:31.725 --> 00:01:34.855
Give you a descriptive overview of the coast economy.

31
00:01:35.715 --> 00:01:39.295
And the second big objective is try to evaluate a policy

32
00:01:39.325 --> 00:01:40.735
that tells us, well, what do we do?

33
00:01:41.225 --> 00:01:43.735
We're gonna try to enforce, we're gonna try to do something

34
00:01:44.195 --> 00:01:45.375
to stop this practice.

35
00:01:45.915 --> 00:01:49.215
And in the context of the Ecuadorian economy, we are going

36
00:01:49.215 --> 00:01:54.015
to evaluate one specific policy that was designed

37
00:01:54.275 --> 00:01:55.375
to alleviate this problem.

38
00:01:55.615 --> 00:01:57.215
I wanna tell you a little bit more about the data.

39
00:01:57.215 --> 00:02:00.855
Because we have a unique op, we had a unique opportunity

40
00:02:00.915 --> 00:02:03.015
to collaborate with the Ecuadorian tax authority

41
00:02:03.315 --> 00:02:06.935
and to collect a vast amount of magnificent data

42
00:02:06.965 --> 00:02:09.375
that allow us to actually target this question.

43
00:02:10.315 --> 00:02:14.295
So, first of all, we, we were able to work

44
00:02:14.295 --> 00:02:17.015
with the tax authority and they gave us a list

45
00:02:17.675 --> 00:02:20.535
of ghost firms that they had identified

46
00:02:21.105 --> 00:02:23.455
going back all the way to 2010.

47
00:02:23.795 --> 00:02:27.255
And I think we, we started this study in 2017.

48
00:02:27.355 --> 00:02:29.575
So the list actually comprises all ghost firms

49
00:02:29.695 --> 00:02:32.775
that were identified between 2010 and 2015.

50
00:02:33.555 --> 00:02:36.585
And remember, it sounds easy, but it's not.

51
00:02:37.745 --> 00:02:40.665
Identifying ghost firms requires detective work,

52
00:02:41.405 --> 00:02:43.425
tax authority use, you know, auditors,

53
00:02:44.095 --> 00:02:47.585
they use tax specialists, they use accountants,

54
00:02:48.125 --> 00:02:50.545
whistleblowers to identify these firms.

55
00:02:51.335 --> 00:02:54.705
They also use big data to try to find discrepancies

56
00:02:54.705 --> 00:02:58.825
between deductions claimed by some firms and reported

57
00:02:58.925 --> 00:03:00.285
Revenues by other firms, right?

58
00:03:00.285 --> 00:03:03.605
If these discrepancies were very large, if some firms,

59
00:03:04.505 --> 00:03:07.045
if some firms never declared the tax returns,

60
00:03:07.185 --> 00:03:09.765
but other firms were claiming big deductions from these

61
00:03:09.765 --> 00:03:11.565
firms, that raises a big flag, right?

62
00:03:12.425 --> 00:03:15.885
And they made attempts, they sent inspectors

63
00:03:16.185 --> 00:03:17.845
to verify the addresses

64
00:03:17.925 --> 00:03:19.965
and to send notifications to these coast firms.

65
00:03:20.995 --> 00:03:23.695
And in many cases, they just couldn't find them, right?

66
00:03:23.695 --> 00:03:24.855
They would knock at their doors

67
00:03:25.035 --> 00:03:27.175
and they, they were fake addresses.

68
00:03:27.325 --> 00:03:28.935
They couldn't find the, the owners.

69
00:03:29.755 --> 00:03:34.215
And finally they posted a list on their website

70
00:03:34.235 --> 00:03:37.015
and they say, Hey, we are, we think that all this list

71
00:03:37.035 --> 00:03:38.855
of firms correspond to ghost firms.

72
00:03:39.275 --> 00:03:40.735
If you are not, if you're on the list

73
00:03:40.735 --> 00:03:42.615
and you are not a ghost firm, please come by

74
00:03:42.615 --> 00:03:44.775
and verify that you actually exist.

75
00:03:45.645 --> 00:03:48.385
And through this big process, they actually came up

76
00:03:48.385 --> 00:03:51.425
with this big long list of 811 ghost firms

77
00:03:51.425 --> 00:03:53.865
that were operating in the economy during this period.

78
00:03:54.845 --> 00:03:56.665
So that's the first piece of the puzzle.

79
00:03:56.975 --> 00:04:00.185
Then we were able, again, to work with the Ian Tax Authority

80
00:04:00.485 --> 00:04:04.865
and to link together a vast amount of administrative data,

81
00:04:05.005 --> 00:04:09.145
the universe of the firm's tax reports, you know,

82
00:04:09.165 --> 00:04:11.105
the income and earn their earnings

83
00:04:11.105 --> 00:04:12.585
statements, the tax reports.

84
00:04:13.165 --> 00:04:15.185
We also serve the characteristics of the firms.

85
00:04:16.285 --> 00:04:18.705
We also know who owns those firms.

86
00:04:18.965 --> 00:04:22.945
So we have a data set that tells us that the, the ownership

87
00:04:23.665 --> 00:04:25.265
structure of each one of these firms.

88
00:04:26.125 --> 00:04:30.425
And we also know how much these owners earn both, you know,

89
00:04:30.425 --> 00:04:31.465
the labor income

90
00:04:31.465 --> 00:04:33.465
and their capital income related to the firms.

91
00:04:34.725 --> 00:04:39.185
And importantly, we are also able to work with firm

92
00:04:39.365 --> 00:04:41.185
to firm transaction level data.

93
00:04:41.645 --> 00:04:44.385
So the Ecuadorian tax Authority at that time,

94
00:04:44.385 --> 00:04:47.705
they required all firms to report every single transaction,

95
00:04:47.705 --> 00:04:49.985
every single invoice that they made.

96
00:04:50.525 --> 00:04:51.665
And you have a link

97
00:04:51.665 --> 00:04:54.025
and you can link, you know, the whole ECU economy.

98
00:04:54.705 --> 00:04:56.745
I mean, we're talking about hundreds of millions

99
00:04:56.745 --> 00:04:57.865
of transactions per year,

100
00:04:58.205 --> 00:04:59.825
but that's information is very rich

101
00:04:59.825 --> 00:05:02.545
because it will allow us to trace, you know,

102
00:05:02.845 --> 00:05:03.985
who is trading with who.

103
00:05:04.845 --> 00:05:08.315
So armed with this data, we can

104
00:05:09.315 --> 00:05:12.555
actually start describing the first question, right?

105
00:05:12.635 --> 00:05:13.795
I mean, tackling the first question,

106
00:05:14.505 --> 00:05:16.275
what is the ghost economy like?

107
00:05:16.495 --> 00:05:18.835
The first question that we're going to ask ourselves is,

108
00:05:19.175 --> 00:05:20.475
how big is it?

109
00:05:21.105 --> 00:05:23.895
Right? We, we just didn't know, is it going

110
00:05:23.895 --> 00:05:25.855
to be super large, is going to be super small?

111
00:05:25.855 --> 00:05:30.015
Is this really a big problem? And we found that between 2010

112
00:05:30.015 --> 00:05:34.175
and 2015, about 10%

113
00:05:34.595 --> 00:05:36.295
of firms in the sample,

114
00:05:36.965 --> 00:05:39.775
they at least once traded with a ghost firm.

115
00:05:40.435 --> 00:05:44.695
If we just focus, let's say on one year, 2015, we found that

116
00:05:45.725 --> 00:05:48.675
about 4.7 of firms, they're,

117
00:05:49.225 --> 00:05:51.395
they traded at some point with a ghost firm.

118
00:05:52.055 --> 00:05:56.755
And those ghost transactions comprised about 14%

119
00:05:57.105 --> 00:05:59.285
Of the total purchase deductions among these firms.

120
00:05:59.785 --> 00:06:01.485
So this is large, this is big, right?

121
00:06:01.485 --> 00:06:03.445
Conditional on trading with a ghost.

122
00:06:03.985 --> 00:06:07.725
On average, 14% of your costs came from ghost firms.

123
00:06:08.615 --> 00:06:09.995
So question number one, this is big,

124
00:06:10.335 --> 00:06:11.435
at least in the context of equa.

125
00:06:11.435 --> 00:06:13.395
This is big. And this is a lower bound, right?

126
00:06:13.905 --> 00:06:16.275
Because you have to imagine that the Ecuador

127
00:06:16.275 --> 00:06:18.715
and tax authority didn't identify

128
00:06:18.735 --> 00:06:19.835
the universe of ghost firms.

129
00:06:19.835 --> 00:06:23.145
So there are more out there, right? So it's large.

130
00:06:24.475 --> 00:06:27.595
Then we ask ourselves, is there any relationship

131
00:06:27.625 --> 00:06:29.635
between the size of the firm

132
00:06:29.695 --> 00:06:33.695
and the likelihood that you trade with a ghost firm?

133
00:06:34.875 --> 00:06:37.185
And before we,

134
00:06:37.365 --> 00:06:38.825
before we analyze this question,

135
00:06:38.825 --> 00:06:40.065
it wasn't obvious to us, right?

136
00:06:40.345 --> 00:06:42.705
'cause one can think, the smaller firms, the ones

137
00:06:42.705 --> 00:06:45.905
that are close to the informal sector, the ones that have,

138
00:06:46.285 --> 00:06:48.265
you know, fewer tax controls,

139
00:06:48.265 --> 00:06:50.305
maybe they are the ones taking advantage

140
00:06:51.045 --> 00:06:53.305
of these tax evasion practice.

141
00:06:53.805 --> 00:06:54.985
But we find the opposite.

142
00:06:55.575 --> 00:06:59.225
It's, there is a very, very big, correlation

143
00:06:59.225 --> 00:07:00.385
between the firm size

144
00:07:00.645 --> 00:07:01.985
and the likelihood that you trade

145
00:07:01.985 --> 00:07:03.185
with ghost, with a ghost firm.

146
00:07:03.405 --> 00:07:05.305
And of course, you can say, well, larger firms,

147
00:07:05.415 --> 00:07:06.825
they just have more transactions.

148
00:07:06.845 --> 00:07:08.665
That's why they're more likely to trade with a ghost.

149
00:07:08.965 --> 00:07:13.545
But even when you look at the share at the share

150
00:07:13.545 --> 00:07:15.545
of ghost transactions as a fraction

151
00:07:15.545 --> 00:07:16.825
of your total reported costs,

152
00:07:17.405 --> 00:07:18.985
you see this positive relationship.

153
00:07:19.085 --> 00:07:22.105
So it's the larger firms, the ones that are actually trading

154
00:07:22.105 --> 00:07:26.065
with ghosts, except at the very, very, very top, except

155
00:07:26.065 --> 00:07:28.785
that, I mean the, the firms that are very, very, very large,

156
00:07:29.445 --> 00:07:30.465
you see a huge drop.

157
00:07:30.775 --> 00:07:31.985
They don't trade with ghost.

158
00:07:32.045 --> 00:07:33.425
And that actually is interesting,

159
00:07:33.425 --> 00:07:35.905
and it, it makes sense, at least in the context of Ecuador,

160
00:07:36.835 --> 00:07:40.925
because these are the large taxpayer units that are subject

161
00:07:40.945 --> 00:07:42.525
to additional tax controls.

162
00:07:42.985 --> 00:07:45.125
And maybe, I mean, the literature also suggests

163
00:07:45.125 --> 00:07:48.845
that they might be able to use other,

164
00:07:49.035 --> 00:07:52.205
more sophisticated practices to avoid paying taxes.

165
00:07:52.705 --> 00:07:54.085
So that's the fact number two, right?

166
00:07:54.185 --> 00:07:55.445
So this is large number one.

167
00:07:55.445 --> 00:07:58.285
Second is that there's a correlation between firm size

168
00:07:59.345 --> 00:08:00.565
and the likelihood

169
00:08:00.565 --> 00:08:03.725
and the amount of transactions we trade with those firms.

170
00:08:05.635 --> 00:08:10.585
Then we also try to understand the relationship

171
00:08:10.975 --> 00:08:12.865
between those deductions

172
00:08:13.845 --> 00:08:17.105
and firms owned by high income individuals.

173
00:08:18.085 --> 00:08:21.185
And we, because we can, you know, we, we know the network of

174
00:08:21.365 --> 00:08:23.705
of owners and we can assess, right?

175
00:08:23.705 --> 00:08:27.785
Whether, whether firms owned by higher income individuals

176
00:08:28.325 --> 00:08:29.825
are the ones that are more likely

177
00:08:29.845 --> 00:08:31.865
to take advantage of this practice.

178
00:08:32.485 --> 00:08:35.645
And this is important because we want

179
00:08:35.645 --> 00:08:37.685
to assess whether this practice is

180
00:08:38.055 --> 00:08:39.405
regressive in nature, right?

181
00:08:39.405 --> 00:08:43.485
Whether higher income folks are being benefited the most

182
00:08:43.485 --> 00:08:45.485
from this evasion practice,

183
00:08:46.025 --> 00:08:49.365
or alternatively, if we find a way to close this loophole,

184
00:08:49.655 --> 00:08:51.565
maybe this policy is going to help, is going

185
00:08:51.565 --> 00:08:53.405
to target richer individuals,

186
00:08:53.465 --> 00:08:55.645
and it's going to be progressive, right?

187
00:08:56.145 --> 00:08:59.165
And we find what we expected at this case, right?

188
00:08:59.235 --> 00:09:01.765
Because I mean, there's a correlation between firm size

189
00:09:02.695 --> 00:09:04.715
and being connected with the ghost.

190
00:09:04.895 --> 00:09:07.515
And because higher income individuals are more likely

191
00:09:07.535 --> 00:09:10.235
to own larger firms, we also see that, you know,

192
00:09:11.105 --> 00:09:15.945
individuals, richer individuals are more likely to,

193
00:09:17.035 --> 00:09:18.825
lemme, lemme, lemme put it straight.

194
00:09:19.445 --> 00:09:23.745
We also find that firms, that firms that are owned

195
00:09:23.745 --> 00:09:25.705
by higher income individuals are more likely

196
00:09:26.405 --> 00:09:27.665
to trade with ghosts.

197
00:09:28.525 --> 00:09:33.165
So these are the, these are the three descriptive facts

198
00:09:33.175 --> 00:09:35.725
about the ghost economy that we uncover.

199
00:09:36.825 --> 00:09:41.485
But we also wanted to exploit the firm to firm transaction,

200
00:09:41.785 --> 00:09:45.685
data at another level to convince ourselves

201
00:09:45.995 --> 00:09:47.285
that these transactions

202
00:09:47.355 --> 00:09:50.605
with ghosts were indeed fake transactions, right?

203
00:09:50.605 --> 00:09:52.765
Because that, that was our hypothesis.

204
00:09:52.765 --> 00:09:54.965
That's what all the tax authorities are telling you, right?

205
00:09:54.965 --> 00:09:56.925
You have ghost firms, these transactions are fake.

206
00:09:57.265 --> 00:10:00.645
But we wanted somehow a smoking guy, we wanted to see

207
00:10:01.185 --> 00:10:03.365
are these transactions suspicious in any way?

208
00:10:03.665 --> 00:10:05.525
Are these transactions different than others?

209
00:10:05.945 --> 00:10:10.205
So in order to do that, we focused on the sample

210
00:10:10.745 --> 00:10:14.045
of ghost clients, the sample of firms that traded

211
00:10:14.045 --> 00:10:17.725
with ghosts, and we compared the distribution

212
00:10:17.725 --> 00:10:20.885
of transactions with ghost firms, with the distribution

213
00:10:20.885 --> 00:10:22.485
of transactions with other firms.

214
00:10:23.975 --> 00:10:26.235
And first let me focus about the distribution

215
00:10:26.235 --> 00:10:28.435
of transactions across the calendar year.

216
00:10:29.445 --> 00:10:32.495
When you look at the transactions with other firms,

217
00:10:32.645 --> 00:10:35.015
this distribution is basically uniform, right?

218
00:10:35.175 --> 00:10:38.215
I mean, firms buy

219
00:10:38.475 --> 00:10:41.535
and they purchase things from other firms, you know,

220
00:10:41.535 --> 00:10:43.655
basically at the uniform rate, it doesn't really matter.

221
00:10:43.915 --> 00:10:46.335
But when you look at ghost transactions,

222
00:10:47.135 --> 00:10:49.665
they are bunched at the end of the calendar year,

223
00:10:49.795 --> 00:10:51.425
which coincides with the fiscal year.

224
00:10:51.925 --> 00:10:53.785
And I mean, that's very suspicious

225
00:10:53.785 --> 00:10:56.145
because that is the time where, you know, accountants

226
00:10:56.165 --> 00:10:58.305
and maybe owners are trying to figure out, okay,

227
00:10:59.005 --> 00:11:00.585
how much tax do we wanna pay, right?

228
00:11:00.745 --> 00:11:02.745
I mean, what's the, what's our targeted tax liability?

229
00:11:03.205 --> 00:11:06.025
And where can we cut, where, where, where,

230
00:11:06.115 --> 00:11:11.065
where we can we add deductions to target, that amount

231
00:11:11.205 --> 00:11:12.745
of taxation that we wanna pay?

232
00:11:13.445 --> 00:11:15.825
So that's very suspicious. That's number one.

233
00:11:16.935 --> 00:11:20.545
The second very interesting thing that we found is that,

234
00:11:21.715 --> 00:11:24.975
these ghost transactions are more likely

235
00:11:24.975 --> 00:11:28.965
to bunch at round numbers, especially at multiples of 500

236
00:11:30.155 --> 00:11:32.955
compared to, nongo transactions.

237
00:11:33.655 --> 00:11:36.195
And that's consistent with representing fake flow

238
00:11:36.195 --> 00:11:37.515
of non non-existent goods.

239
00:11:37.515 --> 00:11:40.155
So there's a literature out there that says that, you know,

240
00:11:40.705 --> 00:11:44.875
over reporting round numbers is correlated with, you know,

241
00:11:44.915 --> 00:11:47.915
fraudulent practices or taxation or manipulation of data.

242
00:11:49.175 --> 00:11:52.755
And finally, I think the, the, the smoking gun for us

243
00:11:53.825 --> 00:11:57.995
Was the fact that these ghost transactions

244
00:11:58.615 --> 00:12:01.915
disproportionate, I mean, they, they budge just below

245
00:12:02.475 --> 00:12:04.795
a threshold, the $5,000 threshold.

246
00:12:04.835 --> 00:12:06.715
I mean, if you see the distribution of,

247
00:12:06.855 --> 00:12:10.675
of Nongo transactions is, you know, it's declines with

248
00:12:11.735 --> 00:12:15.035
as you, it, it declines with the, the number of,

249
00:12:15.145 --> 00:12:16.795
with the level of transaction,

250
00:12:16.935 --> 00:12:19.355
but it's basically smooth around $5,000.

251
00:12:19.735 --> 00:12:21.675
But if you look at the distribution

252
00:12:21.735 --> 00:12:24.155
of ghost transactions, they just budge.

253
00:12:24.235 --> 00:12:26.715
I mean, there is a, there's a very large mass right

254
00:12:26.895 --> 00:12:28.195
before $5,000.

255
00:12:28.775 --> 00:12:31.315
And you then you say, why $5,000?

256
00:12:31.315 --> 00:12:33.835
Well, that's a threshold at which firms are required

257
00:12:33.855 --> 00:12:37.235
to make payments via the formal financial system.

258
00:12:37.695 --> 00:12:40.195
So if you, at least in this context, in Ecuador,

259
00:12:40.195 --> 00:12:42.195
if you wanted to purchase something from a firm

260
00:12:42.365 --> 00:12:44.275
above $5,000, well, you cannot use cash.

261
00:12:44.455 --> 00:12:46.675
You had to go through them, through, through the,

262
00:12:46.675 --> 00:12:49.115
through a bank or just a credit card or something that,

263
00:12:49.175 --> 00:12:50.395
and then you leave a trade, right?

264
00:12:50.395 --> 00:12:51.475
You leave a financial trade,

265
00:12:51.475 --> 00:12:52.715
you actually have to make a payment.

266
00:12:53.175 --> 00:12:54.875
So if you have a big, a fake transaction,

267
00:12:54.875 --> 00:12:56.435
you don't wanna make a real payment, right?

268
00:12:56.975 --> 00:13:01.515
So the, if, if we get, you know, the, these three facts

269
00:13:02.235 --> 00:13:05.675
together, I think we have very strong evidence that indeed

270
00:13:06.285 --> 00:13:08.115
these transactions were not real.

271
00:13:08.335 --> 00:13:09.595
So the problem is big,

272
00:13:10.545 --> 00:13:13.205
and we think that the,

273
00:13:14.115 --> 00:13:16.245
they indeed represent fake transactions.

274
00:13:16.425 --> 00:13:17.765
So we have this big problem, right?

275
00:13:18.105 --> 00:13:20.445
As I, as I as I described the, the next question,

276
00:13:20.445 --> 00:13:22.965
the obvious question is what do we do about it, right?

277
00:13:23.425 --> 00:13:25.805
We have these, those firms are now there.

278
00:13:25.865 --> 00:13:28.285
The problem is big. They, we have fake transactions.

279
00:13:28.465 --> 00:13:29.525
How do we solve this problem?

280
00:13:31.345 --> 00:13:34.245
One, one option is to go after the ghost firms.

281
00:13:34.345 --> 00:13:36.405
And I think this has been a traditional policy

282
00:13:36.425 --> 00:13:37.605
of many tax authorities,

283
00:13:38.305 --> 00:13:40.005
but there's a big problem with it, right?

284
00:13:40.295 --> 00:13:43.685
These, these ghost firms have been incorporated

285
00:13:43.785 --> 00:13:45.445
by criminal organizations.

286
00:13:45.475 --> 00:13:46.925
They are very sophisticated.

287
00:13:47.065 --> 00:13:49.445
So by the time the tax authority finds

288
00:13:49.595 --> 00:13:51.765
that there is one ghost firm, they go

289
00:13:51.765 --> 00:13:53.085
after it and it has disappeared.

290
00:13:53.355 --> 00:13:54.525
They go after the owners.

291
00:13:54.635 --> 00:13:57.405
Well, the owners are victims of identity fraud,

292
00:13:57.465 --> 00:13:58.885
or they are deceased individuals.

293
00:13:58.885 --> 00:14:00.285
They go to the accountant, the accountant has

294
00:14:00.285 --> 00:14:01.725
disappeared, nowhere to be pumped.

295
00:14:02.185 --> 00:14:03.965
So, and at that moment,

296
00:14:04.235 --> 00:14:05.885
some other ghost through just pops up.

297
00:14:06.305 --> 00:14:07.765
So it's a guacamole game, right?

298
00:14:07.785 --> 00:14:10.045
You are trying to go, it, it's a never ending game.

299
00:14:10.315 --> 00:14:12.165
It's very hard to actually target this

300
00:14:12.165 --> 00:14:13.565
problem with that strategy.

301
00:14:14.715 --> 00:14:18.445
What the Ecuadorian tax authority did is they employed,

302
00:14:19.445 --> 00:14:23.005
a very smart, a very clever alternative where they went

303
00:14:23.135 --> 00:14:27.045
after the clients of ghost firms

304
00:14:27.825 --> 00:14:29.005
and they targeted them

305
00:14:29.025 --> 00:14:32.935
and they say, Hey, you know, what, if five years ago,

306
00:14:32.935 --> 00:14:35.655
three years ago, you made these deductions,

307
00:14:35.655 --> 00:14:36.935
you made these purchases, you claim

308
00:14:36.935 --> 00:14:38.175
to make these purchases from this firm

309
00:14:38.175 --> 00:14:39.695
that we identified as a ghost firm.

310
00:14:40.355 --> 00:14:43.735
Now we ask you to amend your past tax return

311
00:14:44.835 --> 00:14:47.795
and either, you know,

312
00:14:49.195 --> 00:14:50.405
take away the deduction

313
00:14:50.745 --> 00:14:52.245
And pay the correspondent tax,

314
00:14:52.425 --> 00:14:55.165
or prove to us that that transaction was real indeed.

315
00:14:56.785 --> 00:14:59.925
And they, what they did essentially is, you know,

316
00:14:59.925 --> 00:15:01.405
they looked at all the discrepancies,

317
00:15:01.425 --> 00:15:04.645
and I think they targeted about 1500 firms

318
00:15:06.595 --> 00:15:08.245
sent formal notifications to them.

319
00:15:09.445 --> 00:15:12.605
And we evaluated that policy.

320
00:15:12.805 --> 00:15:15.445
I mean, we used formal econometric methods

321
00:15:15.745 --> 00:15:18.565
to estimate the causal effect of that intervention

322
00:15:19.505 --> 00:15:22.085
on tax liabilities, on reported tax liabilities.

323
00:15:23.225 --> 00:15:27.025
And this is one of the simplest papers that

324
00:15:27.775 --> 00:15:31.745
I've written where identification strategy is just,

325
00:15:31.805 --> 00:15:35.765
you know, comparing the post amended return versus the

326
00:15:35.765 --> 00:15:36.805
pre amended return, right?

327
00:15:36.885 --> 00:15:39.285
I mean, this is literally, you can estimate SAL effect,

328
00:15:39.285 --> 00:15:40.765
which is one OLS regression

329
00:15:40.785 --> 00:15:42.685
or just with a different simple means.

330
00:15:43.425 --> 00:15:47.005
And the identifying assumption here is that in the absence

331
00:15:47.005 --> 00:15:50.005
of the notification, firms would not have amended their

332
00:15:50.005 --> 00:15:51.725
returns, which is very plausible.

333
00:15:51.725 --> 00:15:54.445
And we show evidence of that, which makes sense, right?

334
00:15:54.525 --> 00:15:56.805
I mean, if you are a firm that has submitted your,

335
00:15:57.355 --> 00:16:01.245
your 2011 taxes four years ago, it's almost impossible

336
00:16:01.245 --> 00:16:03.005
that it's very unlikely that you're going to amend

337
00:16:03.005 --> 00:16:06.085
that return India if you did not receive a

338
00:16:06.085 --> 00:16:07.525
notification that told you to do so.

339
00:16:07.905 --> 00:16:12.325
We find that, well, first finding is that about one fourth

340
00:16:12.865 --> 00:16:15.125
of firms actually respond, 25%

341
00:16:15.125 --> 00:16:16.765
of firms respond to the notification.

342
00:16:16.985 --> 00:16:18.085
And this is not surprising.

343
00:16:18.085 --> 00:16:20.365
We also have some additional research on previous research

344
00:16:20.365 --> 00:16:22.205
that shows that at least in the context of Ecuador,

345
00:16:22.275 --> 00:16:25.085
when firms receive notifications,

346
00:16:25.765 --> 00:16:27.205
a large fraction just do nothing.

347
00:16:27.235 --> 00:16:30.575
They just wait and see for, for next steps

348
00:16:30.575 --> 00:16:31.815
that the tax authority might take.

349
00:16:32.635 --> 00:16:36.455
But among the firms that actually amended the tax returns,

350
00:16:36.915 --> 00:16:41.015
the policy was extremely effective in boosting,

351
00:16:41.955 --> 00:16:43.895
the, I mean, tax liability.

352
00:16:44.635 --> 00:16:48.655
So on average, firms increased, they reported tax liability

353
00:16:48.755 --> 00:16:53.535
by almost 80%, which collectively the intervention,

354
00:16:53.875 --> 00:16:57.535
increased total tax liability by about $20 million.

355
00:16:58.155 --> 00:17:00.655
So this is a very targeted, very small intervention,

356
00:17:00.835 --> 00:17:01.895
almost costless,

357
00:17:01.895 --> 00:17:03.815
because they were just sending notifications

358
00:17:04.075 --> 00:17:05.855
and they were able to collect $20 million.

359
00:17:05.955 --> 00:17:09.255
So it was very successful, it was very successful.

360
00:17:09.595 --> 00:17:13.295
And the Korian tax authority, I mean, was,

361
00:17:13.315 --> 00:17:14.455
was very happy with it.

362
00:17:15.155 --> 00:17:18.375
And I mean all, I mean, their, their instinct

363
00:17:18.375 --> 00:17:21.095
that this was successful was, is confirmed with

364
00:17:21.595 --> 00:17:24.655
or formal academic, analysis,

365
00:17:24.675 --> 00:17:26.935
if we think about external valid of these findings, right?

366
00:17:27.015 --> 00:17:28.295
I mean, what are the, what, what's the key

367
00:17:28.295 --> 00:17:29.415
message that we wanna take?

368
00:17:30.295 --> 00:17:31.375
I mean, the first key message

369
00:17:31.375 --> 00:17:33.615
that I would take from this study is that, you know,

370
00:17:33.615 --> 00:17:35.535
collecting that information is gold, right?

371
00:17:35.535 --> 00:17:38.295
Collecting firm to firm transaction level data is very

372
00:17:38.295 --> 00:17:40.935
important for many reasons, but also

373
00:17:40.935 --> 00:17:42.775
because it allows you to, to spot

374
00:17:42.795 --> 00:17:45.735
and pinpoint cases where ghost, I mean, where this type

375
00:17:45.735 --> 00:17:48.655
of taxation is, is being practiced.

376
00:17:50.255 --> 00:17:53.895
Um, the second big picture question is that

377
00:17:54.745 --> 00:17:58.635
conditional on identifying those firms, it is important

378
00:17:58.775 --> 00:18:01.915
or it seems to be successful to target those clients.

379
00:18:03.175 --> 00:18:06.515
And, you know, at least in the academic literature,

380
00:18:06.515 --> 00:18:10.635
we know very little about how to, how to,

381
00:18:11.495 --> 00:18:13.675
how to contain, how to target instances

382
00:18:13.725 --> 00:18:15.715
where firms are inflating their costs.

383
00:18:16.255 --> 00:18:18.435
And at least in this context, we're giving you evidence

384
00:18:18.435 --> 00:18:19.755
that it might be effective, right?

385
00:18:19.775 --> 00:18:24.475
If you don't allow firms to offset those, those, I mean the,

386
00:18:24.475 --> 00:18:27.075
the, the, the, the decrease in deductions in other line

387
00:18:27.195 --> 00:18:29.755
items, it seems to be an effective policy

388
00:18:30.215 --> 00:18:31.955
to boost tax collection.

389
00:18:33.135 --> 00:18:36.855
But when we think about the standard validity in dynamic

390
00:18:36.905 --> 00:18:40.535
terms in what should Ecuadorian economy do later, I think

391
00:18:40.535 --> 00:18:42.535
that it's, again, we go, we come back

392
00:18:42.535 --> 00:18:43.815
to this guacamole game, right?

393
00:18:43.815 --> 00:18:47.215
Because in equilibrium, you know, the tax authority

394
00:18:47.835 --> 00:18:50.215
can increase their efforts to, to spot

395
00:18:50.215 --> 00:18:51.335
and identify ghost firms.

396
00:18:51.435 --> 00:18:53.575
But criminal organizations are also evolving

397
00:18:53.995 --> 00:18:57.715
and they're going to come up with different ways of, of,

398
00:18:58.385 --> 00:19:00.725
you know, coming up with ghost firms or,

399
00:19:01.345 --> 00:19:05.605
or may, maybe now you may have, I mean, I'm, I'm just,

400
00:19:06.385 --> 00:19:10.345
I'm just, imagining,

401
00:19:10.575 --> 00:19:13.865
putting myself in the, in the shoes of a tax eva.

402
00:19:14.395 --> 00:19:17.655
Maybe they are, maybe now they have firms

403
00:19:17.765 --> 00:19:19.735
that have a very small amount of production,

404
00:19:19.735 --> 00:19:22.015
but they issue, they also should receipts

405
00:19:22.315 --> 00:19:24.135
and they, they can just evolve

406
00:19:24.135 --> 00:19:25.615
with more sophisticated practices.

407
00:19:25.615 --> 00:19:26.815
So it's, it's a hard question.

408
00:19:27.005 --> 00:19:30.105
It's a hard question, but I think collecting data,

409
00:19:30.615 --> 00:19:31.945
running these campaigns

410
00:19:32.765 --> 00:19:36.025
and limiting, you know, the deductions that you can make,

411
00:19:36.025 --> 00:19:39.105
making sure that they are real purchases isn't,

412
00:19:39.105 --> 00:19:42.665
is a very important first step towards, you know, this big.

Quick Answers

Combatting the Menace of Fake Firms: An In-Depth Analysis

00:00:00 - Those firms essentially are fake firms. These are firms that do not produce any products, do not give any services. But these are firms that have been incorporated mostly by criminal organizations with just one purpose to become in invoice mails, to fake, to issue fake receipts. And those fake receipts are going to be used by other firms to claim deductions on their value added taxes or on their corporate income taxes.

00:00:34 - These are fake firms, and it's a big problem, particularly in developing countries. if you talk to tax authorities all over the world, they're gonna tell you that they are aware of this problem. Coast firms or invoice mails are a topic of concern. International organizations know about this problem, and surprisingly, there's little academic research on it, and that's why we were so interested in this topic.

00:01:05 - and I'm going to elaborate more, more on that, as we, as we talk about the details of the paper. But we want to, we're gonna use Ecuador as cases study to learn about the ghost economy. You know, how big the ghost economy is, how common transactions from firms to ghost firms are. What's the magnitude of this problem? Eh, that's the first, that's the first objective. Give you a descriptive overview of the coast economy.

Evaluating Ecuadorian Tax Policies

00:01:39 - that tells us, well, what do we do? We're gonna try to enforce, we're gonna try to do something to stop this practice. And in the context of the Ecuadorian economy, we are going to evaluate one specific policy that was designed to alleviate this problem. I wanna tell you a little bit more about the data. Because we have a unique op, we had a unique opportunity to collaborate with the Ecuadorian tax authority and to collect a vast amount of magnificent data

00:02:10 - So, first of all, we, we were able to work with the tax authority and they gave us a list of ghost firms that they had identified going back all the way to 2010. And I think we, we started this study in 2017. So the list actually comprises all ghost firms that were identified between 2010 and 2015. And remember, it sounds easy, but it's not.

00:02:41 - tax authority use, you know, auditors, they use tax specialists, they use accountants, whistleblowers to identify these firms. They also use big data to try to find discrepancies between deductions claimed by some firms and reported Revenues by other firms, right? If these discrepancies were very large, if some firms, if some firms never declared the tax returns, but other firms were claiming big deductions from these

Fighting Ghost Firms: Unveiling the Fraudulent Transactions

00:03:12 - And they made attempts, they sent inspectors to verify the addresses and to send notifications to these coast firms. And in many cases, they just couldn't find them, right? They would knock at their doors and they, they were fake addresses. They couldn't find the, the owners. And finally they posted a list on their website and they say, Hey, we are, we think that all this list of firms correspond to ghost firms. If you are not, if you're on the list

00:03:42 - and verify that you actually exist. And through this big process, they actually came up with this big long list of 811 ghost firms that were operating in the economy during this period. So that's the first piece of the puzzle. Then we were able, again, to work with the Ian Tax Authority and to link together a vast amount of administrative data, the universe of the firm's tax reports, you know, the income and earn their earnings

00:04:13 - We also serve the characteristics of the firms. We also know who owns those firms. So we have a data set that tells us that the, the ownership structure of each one of these firms. And we also know how much these owners earn both, you know, the labor income and their capital income related to the firms. And importantly, we are also able to work with firm to firm transaction level data.

Title: Unveiling the Size and Impact of the Ghost Economy

00:04:44 - they required all firms to report every single transaction, every single invoice that they made. And you have a link and you can link, you know, the whole ECU economy. I mean, we're talking about hundreds of millions of transactions per year, but that's information is very rich because it will allow us to trace, you know, who is trading with who. So armed with this data, we can actually start describing the first question, right?

00:05:14 - what is the ghost economy like? The first question that we're going to ask ourselves is, how big is it? Right? We, we just didn't know, is it going to be super large, is going to be super small? Is this really a big problem? And we found that between 2010 and 2015, about 10% of firms in the sample, they at least once traded with a ghost firm. If we just focus, let's say on one year, 2015, we found that

00:05:49 - they traded at some point with a ghost firm. And those ghost transactions comprised about 14% Of the total purchase deductions among these firms. So this is large, this is big, right? Conditional on trading with a ghost. On average, 14% of your costs came from ghost firms. So question number one, this is big, at least in the context of equa. This is big. And this is a lower bound, right? Because you have to imagine that the Ecuador

Ghost Firms and Firm Size Correlation

00:06:18 - the universe of ghost firms. So there are more out there, right? So it's large. Then we ask ourselves, is there any relationship between the size of the firm and the likelihood that you trade with a ghost firm? And before we, before we analyze this question, it wasn't obvious to us, right? 'cause one can think, the smaller firms, the ones that are close to the informal sector, the ones that have,

00:06:48 - maybe they are the ones taking advantage of these tax evasion practice. But we find the opposite. It's, there is a very, very big, correlation between the firm size and the likelihood that you trade with ghost, with a ghost firm. And of course, you can say, well, larger firms, they just have more transactions. That's why they're more likely to trade with a ghost. But even when you look at the share at the share of ghost transactions as a fraction of your total reported costs,

00:07:19 - So it's the larger firms, the ones that are actually trading with ghosts, except at the very, very, very top, except that, I mean the, the firms that are very, very, very large, you see a huge drop. They don't trade with ghost. And that actually is interesting, and it, it makes sense, at least in the context of Ecuador, because these are the large taxpayer units that are subject to additional tax controls. And maybe, I mean, the literature also suggests that they might be able to use other,

Relationship Between Firm Size and Ownership in Ghost Economy

00:07:52 - So that's the fact number two, right? So this is large number one. Second is that there's a correlation between firm size and the likelihood and the amount of transactions we trade with those firms. Then we also try to understand the relationship between those deductions and firms owned by high income individuals. And we, because we can, you know, we, we know the network of

00:08:23 - Whether, whether firms owned by higher income individuals are the ones that are more likely to take advantage of this practice. And this is important because we want to assess whether this practice is regressive in nature, right? Whether higher income folks are being benefited the most from this evasion practice, or alternatively, if we find a way to close this loophole, maybe this policy is going to help, is going

00:08:53 - and it's going to be progressive, right? And we find what we expected at this case, right? Because I mean, there's a correlation between firm size and being connected with the ghost. And because higher income individuals are more likely to own larger firms, we also see that, you know, individuals, richer individuals are more likely to, lemme, lemme, lemme put it straight. We also find that firms, that firms that are owned

Detecting Ghost Transactions: A Deceptive Analysis

00:09:26 - to trade with ghosts. So these are the, these are the three descriptive facts about the ghost economy that we uncover. But we also wanted to exploit the firm to firm transaction, data at another level to convince ourselves that these transactions with ghosts were indeed fake transactions, right? Because that, that was our hypothesis. That's what all the tax authorities are telling you, right?

00:09:57 - But we wanted somehow a smoking guy, we wanted to see are these transactions suspicious in any way? Are these transactions different than others? So in order to do that, we focused on the sample of ghost clients, the sample of firms that traded with ghosts, and we compared the distribution of transactions with ghost firms, with the distribution of transactions with other firms. And first let me focus about the distribution

00:10:29 - When you look at the transactions with other firms, this distribution is basically uniform, right? I mean, firms buy and they purchase things from other firms, you know, basically at the uniform rate, it doesn't really matter. But when you look at ghost transactions, they are bunched at the end of the calendar year, which coincides with the fiscal year. And I mean, that's very suspicious because that is the time where, you know, accountants and maybe owners are trying to figure out, okay,

Title: Suspicious Round Number Ghost Transactions

00:11:00 - I mean, what's the, what's our targeted tax liability? And where can we cut, where, where, where, where we can we add deductions to target, that amount of taxation that we wanna pay? So that's very suspicious. That's number one. The second very interesting thing that we found is that, these ghost transactions are more likely to bunch at round numbers, especially at multiples of 500

00:11:33 - And that's consistent with representing fake flow of non non-existent goods. So there's a literature out there that says that, you know, over reporting round numbers is correlated with, you know, fraudulent practices or taxation or manipulation of data. And finally, I think the, the, the smoking gun for us Was the fact that these ghost transactions disproportionate, I mean, they, they budge just below

00:12:04 - I mean, if you see the distribution of, of Nongo transactions is, you know, it's declines with as you, it, it declines with the, the number of, with the level of transaction, but it's basically smooth around $5,000. But if you look at the distribution of ghost transactions, they just budge. I mean, there is a, there's a very large mass right before $5,000. And you then you say, why $5,000? Well, that's a threshold at which firms are required

Combatting Fake Transactions in Ecuador

00:12:37 - So if you, at least in this context, in Ecuador, if you wanted to purchase something from a firm above $5,000, well, you cannot use cash. You had to go through them, through, through the, through a bank or just a credit card or something that, and then you leave a trade, right? You leave a financial trade, you actually have to make a payment. So if you have a big, a fake transaction, you don't wanna make a real payment, right? So the, if, if we get, you know, the, these three facts together, I think we have very strong evidence that indeed

00:13:08 - So the problem is big, and we think that the, they indeed represent fake transactions. So we have this big problem, right? As I, as I as I described the, the next question, the obvious question is what do we do about it, right? We have these, those firms are now there. The problem is big. They, we have fake transactions. How do we solve this problem? One, one option is to go after the ghost firms. And I think this has been a traditional policy

00:13:38 - but there's a big problem with it, right? These, these ghost firms have been incorporated by criminal organizations. They are very sophisticated. So by the time the tax authority finds that there is one ghost firm, they go after it and it has disappeared. They go after the owners. Well, the owners are victims of identity fraud, or they are deceased individuals. They go to the accountant, the accountant has disappeared, nowhere to be pumped. So, and at that moment, some other ghost through just pops up. So it's a guacamole game, right?

Tax Authority Targeting Ghost Firms: Strategic Approach

00:14:10 - It's very hard to actually target this problem with that strategy. What the Ecuadorian tax authority did is they employed, a very smart, a very clever alternative where they went after the clients of ghost firms and they targeted them and they say, Hey, you know, what, if five years ago, three years ago, you made these deductions, you made these purchases, you claim to make these purchases from this firm

00:14:40 - Now we ask you to amend your past tax return and either, you know, take away the deduction And pay the correspondent tax, or prove to us that that transaction was real indeed. And they, what they did essentially is, you know, they looked at all the discrepancies, and I think they targeted about 1500 firms sent formal notifications to them.

00:15:12 - I mean, we used formal econometric methods to estimate the causal effect of that intervention on tax liabilities, on reported tax liabilities. And this is one of the simplest papers that I've written where identification strategy is just, you know, comparing the post amended return versus the pre amended return, right? I mean, this is literally, you can estimate SAL effect, which is one OLS regression

Targeting Ghost Firms: Successful Tax Intervention

00:15:43 - And the identifying assumption here is that in the absence of the notification, firms would not have amended their returns, which is very plausible. And we show evidence of that, which makes sense, right? I mean, if you are a firm that has submitted your, your 2011 taxes four years ago, it's almost impossible that it's very unlikely that you're going to amend that return India if you did not receive a notification that told you to do so. We find that, well, first finding is that about one fourth

00:16:15 - of firms respond to the notification. And this is not surprising. We also have some additional research on previous research that shows that at least in the context of Ecuador, when firms receive notifications, a large fraction just do nothing. They just wait and see for, for next steps that the tax authority might take. But among the firms that actually amended the tax returns, the policy was extremely effective in boosting, the, I mean, tax liability.

00:16:48 - by almost 80%, which collectively the intervention, increased total tax liability by about $20 million. So this is a very targeted, very small intervention, almost costless, because they were just sending notifications and they were able to collect $20 million. So it was very successful, it was very successful. And the Korian tax authority, I mean, was, was very happy with it.

Success in Tax Policy Analysis

00:17:18 - that this was successful was, is confirmed with or formal academic, analysis, if we think about external valid of these findings, right? I mean, what are the, what, what's the key message that we wanna take? I mean, the first key message that I would take from this study is that, you know, collecting that information is gold, right? Collecting firm to firm transaction level data is very important for many reasons, but also because it allows you to, to spot and pinpoint cases where ghost, I mean, where this type

00:17:50 - the second big picture question is that conditional on identifying those firms, it is important or it seems to be successful to target those clients. And, you know, at least in the academic literature, we know very little about how to, how to, how to contain, how to target instances where firms are inflating their costs.

00:18:18 - that it might be effective, right? If you don't allow firms to offset those, those, I mean the, the, the, the, the decrease in deductions in other line items, it seems to be an effective policy to boost tax collection. But when we think about the standard validity in dynamic terms in what should Ecuadorian economy do later, I think that it's, again, we go, we come back to this guacamole game, right? Because in equilibrium, you know, the tax authority

Evolving Strategies of Criminal Organizations in Creating Ghost Firms

00:18:50 - and identify ghost firms. But criminal organizations are also evolving and they're going to come up with different ways of, of, you know, coming up with ghost firms or, or may, maybe now you may have, I mean, I'm, I'm just, I'm just, imagining, putting myself in the, in the shoes of a tax eva. Maybe they are, maybe now they have firms that have a very small amount of production,

00:19:39 - and they, they can just evolve with more sophisticated practices. So it's, it's a hard question. It's a hard question, but I think collecting data, running these campaigns and limiting, you know, the deductions that you can make, making sure that they are real purchases isn't, is a very important first step towards, you know, this big.

Faculti AI can make mistakes

Research

Policy

Analysis

Data

Originals

Explainers

Podcasts

Watch the latest episode of The Exchange – the show where influential policymakers and leading researchers engage in insightful discussions shaping the future.

Episode: Brazil's Economic Resilience