Dina Pomeranz and Dave Donaldson (MIT) talked with Tim Phillips (Vox) about their work on ghost firms along with Paul Carrillo and Monica Singhal. The group has been investigating how ghost firms operate in Ecuador, evaluating the government’s efforts to shut them down.

This podcast was originally published on “VoxTalks Economics” from the Center for Economic Policy Research, 9 September 2022. Transcribed and edited by the UBS Center.

Tim Philips, Interviewer (I): Welcome to VoxTalks Economics for the Center for Economic Policy Research. My name is Tim Phillips. Every week we bring you the best new research in economics. And remember to subscribe and follow us on Instagram as well.

Ghost firms issue fake receipts that allow their clients to claim fraudulent tax deductions. It’s widespread, relatively hard to stop, and it’s costing governments everywhere billions in lost revenues. Dave Donaldson of MIT and Dina Pomeranz of the University of Zurich have been on the case in Ecuador, where they, along with Paul Carrillo and Monica Singhal, have been investigating how ghost firms operate. They evaluated the government’s efforts to shut them down. Dave and Dina are joining me today. Dave, hello.

Dave Donaldson (DD): Hi, Tim

I: And Dina, welcome as well.

Dina Pommeranz (DP): Hi, Tim.

I: First of all, Dina, explain how this all works. What do these ghost firms actually do? Where do they come from?

DP: Ghost firms essentially are fake firms that are set up with the purpose to help other firms cheat on their taxes. It’s as simple as that. And how do they do that? They issue receipts pretending that they sold something to a client firm, but they haven’t actually sold anything except that they have issued a receipt. The client firm can then deduct this cost from their corporate income tax and from their value-added taxes because it looks like the client firm had less profit than it actually did. And therefore, it’s paying fewer taxes than it should.

I: Why is this such a big problem for governments? Does this affect all governments?

DD: Well, of course, all governments need to tax in one form or another. They need to do that because they try to fund public services that improve people’s lives, like health, education, public safety, and so on. Raising tax revenue can be extremely hard in some countries, like in Ecuador where we’re studying. I think we sometimes take that for granted. Building a state apparatus capable of enforcing a certain tax rate – whatever that rate may be and whoever is to be taxed – is really hard. A lot of countries around the world struggle to do that. And the form of evasion embodied in ghost firms and the ghost clients that work with those firms is just one example of such evasion. Another thing that’s distressing here is that governments try to raise revenue in a way that minimizes distortions, that minimizes kind of economic waste, and also does it in hopefully in an equitable way. One of the problems that ghost firms impose on governments is that they get in the way of that kind of relatively distortion, minimizing way of raising taxation and an equity enhancing way of raising tax revenues.

I: But we know they exist because we’re talking about them at the moment. So why are they so hard to get rid of?

DP: They are fake. They often don’t have any real operations, so even detecting a ghost firm is not that easy. But now imagine, you know exactly the following tax ID is a ghost firm, now try to punish that firm. You go to the location, maybe there’s nothing there. You want to find the person who registered that firm. But maybe there are three other firms in between, or the person is long deceased.

DD: Or a victim of identity fraud or identity theft.

DP: Exactly. If you manage to shut one down, you can always block a tax ID and say, “Okay, deductions from this tax ID are no longer valid.” But tomorrow a new ghost firm could pop up, so it’s essentially a game of “Whac-a-mole”. You shut one down, and a new one opens. That’s why we are excited to tell you about a very cool intervention that the Ecuadorian tax authority has implemented to tackle this problem with a new approach.

I: You said, you’ve been working with the tax authorities in Ecuador, so how big is the problem they have with ghost firms?

DP: We were quite surprised by how widespread this phenomenon is. At least 10% of firms in Ecuador use these types of transactions. That’s probably an underestimate because as we just discussed, these firms are very difficult to find. But we know that at least 10% of firms use these types of fake deductions. And for these companies, over 10% of the costs they deduct are from such ghost firms. So, it’s not just here or there, that little receipts are misfiled. We are talking about 2.1 billion dollars in fake deductions. So that’s a very large amount as you can imagine. The Ecuadorian economy is very small. Both because it doesn’t count that many people and it’s much poorer than a country like the United States or Switzerland.

I: And are these transactions conducted with ghost firms typically similar to real transactions?

DP: No. We have found that they are very different, and that also makes us confident that the companies that the tax authority has classified as ghost firms are not regular companies that have been mislabeled. First of all, there is a common forensic strategy that people look for, which is a suspiciously large amount of bunching at round numbers. Transactions have a certain type of distribution, and the transactions that firms make with ghost firms are much more likely to bunch at round numbers. In addition, these types of transactions rise strongly at the end of the year, which is consistent with firms that try to “optimize” the tax payment by buying more fake receipts at the end of the year. But I think the most important indication is that in Ecuador there is a threshold above which you must go through the formal banking system. We find a very strong bunching just below that amount, so there’s actually no real transaction going on here, right? If you were using the banking system, you would have to make that transaction. We find very strong bunching just below that threshold, and the same company does not have that kind of bunching for the rest of its transactions at that threshold.

I: If these transactions have a different character, what about the firms – the clients of those ghost firms, the ones that are committing these frauds – do they tend to be like the average firm?

DD: This honestly surprised me. I had the idea that these would be really small companies, maybe a couple of employees, that may or may not be on the borderline of the formal tax system. My idea was that these would be the companies that would be the ghost customers buying from the ghost firms. That was exactly wrong. We find that even Ecuador’s large firms throughout the size distribution – actually increasing in the size of firms –all those larger firms not only do more ghost transactions but have a larger share of ghost transactions in their total deductions that they claim on their tax forms. In other words, the main characteristic that surprises about these ghost client firms is that they’re on average larger than the rest of the firms in the economy.

I: Wow.

DD: And the second thing we were delighted about is, that we were able to link the ghost client firms, who work with the firms to evade taxes, to their owners. The actual person or persons behind the firm on the ownership side, and this fact sort of continues. So, it tends to be the relatively wealthy owners who do this on average. This is another characteristic of these ghost client companies: They tend to be owned by relatively wealthy people.

I: We have mentioned before that the tax authority in Ecuador has a cunning plan to eliminate these ghost firms. What is this plan? What is its strategy?

DP: The first step of the plan was to collect the data and find out who might be a ghost firm.

DD: Actually, there is a backstory to this. I am not a public finance or tax economist. About 10 years ago I started on this collaboration with Dina, Monica, and Paul. I was excited about the potential of Ecuador and the growing number of countries around the world, similar to Ecuador, that have these amazing tax databases in their tax authorities that track astounding things for the formal economy and the tax economy. Using the value-added tax system, for example, we can track the flow of money in an economy. Every time a company buys something from another company, they pay value-added tax on it. They report the type of transaction, who they are, who they buy it from, and what the value was, etc. Therefore, it’s this big network of a linked economy where we can see who trades with whom from firm to firm. As mentioned before, another crucial point for the study is, that we can link the firms to the owners. This gives us an important input about the firm in terms of the type of entrepreneurial and financial capital that flows into the firms. We can also link, as is possible in many countries nowadays, between workers and companies. Just imagine the large number of linkages that are possible in an economy like Ecuador thanks to these administrative records. I would like to emphasize the fact that this is, of course, thanks to the Ecuadorian tax authorities, and also to the collaboration they have built up with my co-authors, especially Paul Carrillo, whom I would like to highlight here. He is a professor at George Washington University and has worked closely with Ecuadorian tax authorities for at least 15 years now, both to help them with their problems and to help researchers access the data that enables a study like this one.

I: So, given this data, Dina, what are the tax authorities in Ecuador trying to do? What is their intervention here?

DP: We mentioned earlier that there is this Whac-a-mole problem: if the government shuts down one firm, another one of these ghost firms pops up. So, the policy intervention that the Ecuadorian government undertook is really creative, and I think it could potentially be a role model also for other countries. They flipped the script and went directly after the ghost clients instead of the ghost firms. This was the start to find out who might be a ghost firm. Since all of this is illegal, forbidden, and hidden, it’s not that easy. First, a list of suspected ghost firms was compiled. Firms that other firms claim a lot of deductions from them but that themselves pay very little or no income tax of those supposed sales, or firms they found in audits that are suspicious and potentially fake. People from the tax authority were sent out to confirm that those are not actually real stores. They also sent emails to the registered owners to find out if someone would respond. And if it seemed that this company did not carry out any real economic activity, but only passed on the receipts to others, they got listed on a website as a suspicious ghost firm. In a next step, they were asked to declare by phone that they are a legitimate business to be removed from the list. If neither them nor their staff would react and their location couldn’t be found, the tax authority concluded that it looks very much like a ghost firm. In the data I mentioned before, we see that transactions with these firms are weird and very different, and it looks like the tax authority was right in classifying them as ghost firms. Now, those were most likely not all the ghost firms in the country. There are probably many more. In the study, we used a set of about 800 firms that we assume are likely to be ghost firms. The tax authority then checked which regular established businesses file a lot of cost deductions from things that were supposedly bought from one of these ghost firms. And then there are companies that deduct hundreds of thousands or other huge amounts that come from fake firms. These are the companies the tax authority went after. So these are real and large companies, the tax authority sent a notice to saying: "Look, we have all these transactions that we think are from companies that don’t exist. Please justify and explain that these are real transactions or pay back the taxes on them.”

I: I see. So, this gets rid of the Whac-a-mole problem, doesn’t it? Because you’re dealing with real firms. When they were sent these notices, which ones complied? How much did they pay back?

DP: As in other studies also here, more than half of the firms did not respond. This points to a problem that has been mentioned in many countries. Indeed, in most countries around the world, it is really difficult to get taxpayers to comply. You can’t take a million people to court, the firms know that, so they wait their turn. Some people just took the risk of not responding. Those who responded answered very strongly. You would have imagined a lot of them picking up the phone to explain themselves. No, they rather paid back on average over $200,000 by filing additional costs they had previously deducted, stating, “Oh oops, sorry, I didn’t mean to deduct that.” So that’s a rare thing. You don’t usually find firms refiling a tax return with less costs than before. Usually, people forget to submit something. This has had a very big effect. Firms that responded paid on average over $40,000. They had a $40,000 higher tax liability than they had before. This sums up to almost double the amount of tax that they had before, or 81% more. Disproportionately coming again from firms that are owned by high-income individuals. This is interesting if you think about tax enforcement in developing countries. There is often a tradeoff: by raising more taxes we are making the poor even poorer. This is a big concern. The intervention clearly seems to be able to target firms owned by the richest. The additional tax paid as a result of this intervention, among the top 5% of the owners in terms of their income was 56 times more than the bottom 80% of the population. The firms owned by the richest people disproportionately paid a lot of this back and those are big amounts.

I: Now, I know from what we have done on VoxTalks before that when we try to raise more taxes, firms tend to cross over into the informal sector. I would imagine Ecuador has quite a large informal sector. Is there any evidence that they are doing that by, instead of paying twice as much in taxes, just not paying any taxes at all?

DD: That is indeed a very legitimate concern as you mentioned in settings like this one where the informal sector is large. But we looked at that and we think that there is very little evidence for that going on. If you think about it, it is actually quite natural. I have already told you that these ghost clients tend to be big, and those targeted by the intervention among ghost clients tend to be even bigger. In fact, they tend to be “the big of the big”. Taking those two factors together, it may not be terribly surprising that exit is not an attractive option for such a large company. In a sense, you could say that this kind of exit is far from the boundary between informality and formality and is therefore a legitimate strategy.

I: Is there anything that prevents this from being applied in other countries? What would be the upside if we manage to crack down on ghost firms in many places?

DP: I think it is a really great strategy that has the potential to be replicated elsewhere. As we mentioned it was quite effective at raising more taxes and building this tax capacity strength that Dave was alluding to. It also helped to strengthen tax fairness, because it closed the loophole that was used particularly by firms owned by very rich people. It can potentially have positive spillovers on the rest of the tax system. In another work, I’ve talked about self-enforcement in the value-added tax, where asking for a receipt, the supplier pays more taxes. But obviously having these fake firms in the system undermines the whole self-enforcement because if the receipts are fake, it doesn’t work. There might be positive spillovers on the rest of the system. If you’re asking about what stops it, I think there are two things. One is capacity. As Dave mentioned, this requires a lot of data. You must have this transaction data to find out who is taking transactions from these ghost firms. You must verify which firms are ghost firms. Ecuador is an upper-middle-income country with a lot of capacity. So, they were on the brink of being able to do something like that. I also work with tax authorities in Tanzania or the Democratic Republic of Congo. For them, that would be a big stretch. Another measure that can stop that kind of thing, not just in Ecuador but also in even richer countries, is the political economy. If it’s a practice that is used proportionately by companies that are owned by very rich people who may have very good relationships with the political parties or the people holding power, then they may not have much appetite for that type of intervention. I think there is a fixed cause of setting up digitalization. If people have tax authorities, have everything on the books, moving to a system where everything’s online has obviously a fixed cost and being able to do this transaction analysis. And if countries are struggling with much more basic things, like registering firms and knowing who’s where, not every country thinks that that’s the right step right now to invest just in digitalization. But I think we can be relatively confident that this intervention is appropriate for Ecuador. Now, it always depends how you value the social benefit of an additional tax dollar raised. It gets very complicated with welfare analysis. Because one of the issues is whether this will increase government revenue or, on net, is it going to keep government revenue the same? Are we going to be able to lower taxes for other people so that we have less distortions? If it is going to increase government revenue in total, what is it going to be spent on? If it goes to the jet of the president, then that’s not that valuable. I mean, the welfare impact on any intervention on tax enforcement depends a lot on the local context. I think that it would be a very interesting additional show where we could talk about under which condition tax enforcement is welfare increasing. Also, how much do you care about inequality? Will it affect the calculus in our case? This would be a very interesting debate.

I: I think you’ve given us about four extra podcasts to record Dina, that we won’t do today, but we will be back for them. This is very fascinating, and it is not an often-discussed subject. There is not a lot of research on this. So, this is fantastic to have it. I especially appreciate having such interesting and practical conclusions, Dina. Thank you.

DP: Thank you very much.

I: And Dave, thank you as well.

DD: Thank you, Tim.