 This is a joint work with Yasin Lefwili from T.C. And I need to say that this paper, it is a policy paper that is partially based on a report commissioned by the European Commission, but the opinion expressed in this report are only those of the authors. And as I was mentioning, this is going to be a policy paper that perhaps and hopefully is going to foster some food for thought and create a path for new research questions to be investigated. So in this audience, we are all familiar about platforms, platforms are intermediaries that allow interactions between different group of agents and some of these agents, buyers and sellers, users and advertisers sometimes can engage in some illegal activities. You can think, for example, at counterfeit items that are sold in several e-commerce platforms and this has become a big problem in recent years that even the U.S.C.D. in a policy report a couple of years ago wrote that e-commerce platforms now represent an ideal storefront for counterfeiters. But you can also think, for example, at the high-pin fringing material that is present on an hosting platform or eight-speech on social media platforms like Facebook, Instagram or Twitter. And this has also created some concerns among policymakers in recent years. Platforms might have incentive to stop illegal misconduct by third parties. These are third parties, these are direct from doers. And they might have incentives to do so because we can think of screening and monitoring as a part of a platform governance. Platform governance has been a recent investigator in the platform economic literature, for example, to control competition in market places. So there are some recent paper by Ben Kasner in a GIO or Tao Tao Teh that is for coming in the American Economic Journal Micro in a social media platform. So you might think, for example, platforms are being interested in ensuring in protecting advertisers from the presence of unsafe material that can damage their reputation. And this is something that we investigate in a recent paper with Martin Quinn or platform screening can be part of a marketing strategy of a platform to position itself as a content as a more moderate or extreme content platform. There is a very nice paper by Louis D. Riemann Zang as a working paper. And finally, platforms might engage in screening in order to protect some of its participants like right holders and brain donors. And this is something that we try to investigate in a working progress with the Doshin Zhiyun and I just see the food. More generally, platforms maybe willing to take some measure to stop illegal activities because reputation in this environment matters a lot and reputation matters because it impacts the strength of the network externalities. However, even if platforms are active and engage in screening, the platforms incentives might differ from those of a welfare maximizing social planner. And indeed are likely to be lower in a wide range of circumstances to make an example, let us think for example at the platform that needs to decide about its screening policies. In doing so, it will decide on the basis of its market reach the potential market reach operating at the margin about the next users to attract while instead the social planners also cares about the externalities that are exerted on those agents that are not part of the platform ecosystem. Again, you can think of social media platforms with users that are not present on the social media that can be hard by its pitch. We have several examples about this possibility. Or for example, right holders and brain donors that are outside the platform ecosystem and therefore the platform is feeling to take into account their presence. At the moment in Europe and in the US the platforms benefit from a safe harbor provision. In the US, for example, under section 230 on the communication decency act that was approved in 1996, a platform are considered immune and they should not be held responsible for third party misconduct and therefore they can benefit from a safe harbor provision. Something very similar applies in Europe under some restrictions and under the Commerce Directive of 2000 platforms can benefit from immunity, this liability exemption as long as they are passive and I will go back to this notion later on and under the so-called knowledge standard platforms are exempt, they can benefit from liability exemption, they can have immunity as long as they are not aware of illegal conduct that is committed on the platform. And the moment they become aware of this illegal conduct they should act immediately otherwise they are going to lose the immunity. And these liability regimes that are mainly liability exemption regime were designed more than 20 years ago with a specific rationale, protect platforms from endless litigations, thereby stimulating investments by platforms in the very early phase of the internet. But in the last 20 years things have changed even radically and there are now several proposals on the table of policymakers. In Europe in the last December the European Commission proposed the Digital Services Act. Something very similar has been proposed by the UK government in the last April with the UK online safety bill that mainly tailors social media platforms. And in the US there is a discussion regarding changing upgrade section 230. There was also, there were also tweets and by Donald Trump in the last months before his ban. And there are also proposals that mainly concern e-commerce platforms and the presence of unsafe products being sold on these platforms. And the basic idea that the common understanding is that the digital landscape changed in the last 20 years for several reasons. Some platforms now operate cross-border and therefore they allow interactions between vendors and buyers that are present, for example, in different jurisdictions. And this makes harder for victims to sue wrongdoers that are present in another jurisdictions. Similarly, some platforms are no longer mere intermediaries now. They are active in the market and they provide complimentary services. They recommend products, they sponsor products. And more generally, there are some platforms that are quite big, they have big pockets and they also have the resources, financial and technical resources to engage in screening and potentially stop online misconduct. So if we want to upgrade and change the current liability regime, we may want to go back to the law and economic literature that has studied liability mostly in traditional markets where you have bilateral relationship between buyers and sellers or manufacturers. And this literature has mainly focused on how to incentivize sellers to exert a higher level of care, more effort, more safety. And this literature has identified benefits of including introducing a liability regime. For example, when there are market failures, asymmetric information is a leading example, consumers might not be aware of the level of safety of a product and therefore liability in this case can be beneficial. Or again, when there are some harbored parties that are not participating in the transaction or when there is market power. And my discussion, Katrin Speer has a very nice paper with Hua last year in RAND that tackles this dimension. But while liability can bring some social benefits, there are also some costs that are associated with the introduction of liability. For example, inefficient litigation costs. Parties may simply not be able to reach out of court settlements and therefore might need to go to court leading therefore to inefficient enforcement and litigation costs. Or simply given that liability will entail more effort and therefore additional costs, these costs can be passed on to consumers and therefore the benefits of liability should be weighted against the cost that liability brings. This literature is also, and this is where there are some important connections with the platform environment. This literature has also identified condition for example, for imposing liability to parties that are not let's say directly involved into a crime or into something that can damage consumers. And again, Katrin has a paper in the yard with Hai in 2005, which she shows that consumer, it might be optimal to impose a liability on manufacturers, even when a product is sold to consumers and these consumers are intentionally causing arms to other consumers. And one way to do so is to make sure that there are some exempt incentives for the manufacturer and also to compensate victims is to impose, to complement the primary liability on consumers with residual liability on the manufacturer. But there might be reasons to impose a liability between direct from doers also when there are multiple infringers or more generally when, and this goes back to a very important paper by Crackmite in the Journal of Law and Economics when you have an intermediary, you have a person, you have a party that is in the condition to prevent accidents, to stop illegal activities by withholding support. And this is precisely what platforms can do even though platforms are not directly the wrong doers are intermediaries that because of their position can potentially stop some misconduct. And in a platform environment which we have multiple parties, multiple incentives there is not much investigation. And this is where it is important to potentially contribute with our research on platform liability. And the economic effects of platform liability in multi-sided markets are indeed rather unexplored. And so far there is only an unformalized contribution very important contribution by we can, that's Feria Martin-Pais last year that tries to tackle some of the problems that are also discussed today. Before continuing with the potential effects that is stricter liability regime for platform semi until I might pause for a couple of minutes if there are some questions. If there are. Okay, so no questions. So let us now try to identify what are the potential effects that the introduction of a stricter liability regime for platform semi-entained. And to do so let us identify a stricter platform liability regime as a liability regime that is based on a liability exemption that becomes simply costly, more costly to maintain. You can think for example at an exemption regime that requires platforms to comply with additional obligations otherwise they are going to lose the immunity. And this additional obligation will entail additional precaution, additional screening and therefore additional screening cost. And we might expect to have a two-fold effect. For sure we are going to have some direct effect the usual suspect given that the platform will have to comply with a stricter liability regime. We might expect a reduction in the harm caused to consumers to users to third parties but they are also indirect effects because we are living in a second best environment and policymakers will only be able to, or regulators will only be able to control one instrument the screening intensity, leaving all the other instruments that are part of a platform governance completely free. And these are the effects, the indirect effects that we plan to discuss in the next slides. And in particular we might expect some indirect effects on the pricing strategies of the platform, on the level of the activity on the platform, on the business models employed by the platform, on the creation of barriers to entry, terms of condition and also platforms and third-party investments. Now to start with, let us consider the canonical textbook example of a multi-sided market in which there is a platform that is allowing intermediation between users, buyers and sellers. And among sellers there might be some sellers that are legitimate and some sellers that can be malicious. They can sell unsafe products, they can sell underfit items, they can sell low quality mutations that are simply violating IP rights. Let us suppose that the platform is charging a membership fee to sellers and there is no cost to buyers. Now, if there is a stricter liability regime, this stricter liability regime will imply that the platform will need to start screening sellers more closely. For example, by acquiring additional documentation verifying their identity, verify the business data and so on and so forth. And this implies indeed that there might be some additional marginal costs of serving the sellers and therefore by increasing liability by making stricter liability for the platforms we might expect that the commission fee, the membership fee charged to sellers is going to increase and therefore this liability costs are going to be passed onto sellers at least partially, therefore leading fewer sellers to join the platform and abstracting from any value that buyers might put on the screening activity of the platform this might lead to fewer buyers participating in market. So you can see that there are these the cross network externalities, the usual cross network externalities that might amplify some of the effects that the liability might create. But let us now consider what might happen if for example buyers are also charged the commission fee. Now the platform might have an additional instrument and might try to attract buyers with a discount and that depending on the strength, the relative strength of the network externalities participation of buyers might also increase and therefore this might mitigate the reduction in the number of sellers that are present on the platform because of a higher commission fee. This is just to say that there might be many moving parts, business model and new neutrals and multi-sided markets effect are not often very clear. Now, let's try to dig a bit more and let us consider any commerce platform, again in which you have original products that are selling high quality products and you have vendors that are selling low quality limitations and these limitations are limitations that are violating IP rights. Now, let us suppose that buyers do not want counterfeits, high pin fringes on the platform because for example, these low quality limitations can be considered as unsafe or simply because buyers have no taste for limitations. Now, in this environment, a stricter liability regime that entails more screening of counterfeiters would lead to have more, a higher utility for buyer participation on the platform and therefore more participation by buyers and therefore more legitimate seller participation. My face can be different if instead buyers to some extent benefit from the presence of counterfeiters and there might be high pin fringes, for example, that are illegal simply because they are infringing IPs but are selling products that are safe for consumers and consumers have a taste for low quality safe limitations. In this case, a stricter liability regime managed to have a situation in which there is a lower expected utility for buyer participation on the e-commerce platform and therefore we may expect fewer participation by buyers and therefore also fewer participation by legitimate sellers. Let us consider now a different platform environment and let us focus on another funded platform that can be, for example, a hosting platform like YouTube or a social media platform like Facebook, Instagram. And in this platform, there are some advertisers, there is no price for the consumers but the consumers perceive these ads as a nuisance, as a cost. Now, if there is a stricter liability regime because of the presence, for example, with legal content that are uploaded are generated by the users, then for the platform, each user will become more costly because basically the platform will have to screen the user activity. And this will reduce the net expected value of an additional user on the platform creating the incentive for the platform to compensate for the screening cost by increasing the intensity, the number of ads displayed to users. And again, there might be this indirect effect and given that users dislike ads, it is not clear what is going to be the welfare impact at least on the consumer side of the market. Another way to think at this problem is to consider a situation in which advertisers care about the environment in which they are placed because they would like to be protected from brain safety issues. And there have been many scandals in the past years with the so-called adipocalypse in YouTube or many advertisers that last year decided to pull ads from Facebook because they were associated with illegal content or a speech. And if we think at a liability regime that entails more screening, then more screening is going to reduce the brand risk that advertisers might see my face, therefore giving more surface to advertisers. And the platform might try to extract this surface by charging advertisers a higher price. And the paper that we have with Martin Quinn, we try to look at this dimension. And we also show that not only there is a higher price charged to advertisers, but there will be also more ads being displayed to users. And therefore, also in this case, the net effect for the consumers is not going to be exactly clear. While advertisers will be better off because of the reduction in the brand risk, consumers might eventually benefit from a stringent content moderation as a consequence of the stringent liability regime, but they will be exposed to a higher number of ads that creates a disability for them. In the medium-long run, platforms can also respond to a change in the liability regime by also changing their business model. Let us consider the two polar cases of agency model versus the wholesale model. In the agency model, the platform is an intermediary. Vendors are selling their products directly to the consumers. They are taking care of the logistic and there is no direct control of the product being sold by the platform owner, simply because the product is not in the facility in the warehouse of the platform. While instead, under the wholesale model, products are directly produced by the platform owner or they are sold by the platform owner, even if by working as a retailer. And therefore, there may be the possibility for the platform owner to inspect, to verify, for example, the authenticity, the safety, the documents concerning a product. And indeed, we may expect that a stricter liability regime, other things to see well, will make platforms moving towards more integration, towards a hybrid business model, simply because it may lower the liability cost. It might be easier for the platform owner to inspect product and engage in screening activity. And also in this case, the social desirability of a dual business model is this question, there is a discussion on the social desirability of such a model. But as an alternative to more integration, platform owner might also react by changing the way they act in a decent market. Some platforms, for example, Amazon, not only provide allow intermediations between buyers and sellers, but they also offer some complimentary services like fulfillment services or logistics. And again, if a product being sold by third parties is actually in the premises, in the availability of the platform owner because the platform owner is going to take care of the logistics, a stricter liability regime might incentivize the platform to impose or to require sellers to join this fulfillment program because this is going to entail an increased control over the product being sold in the market because it's going to be easier or cheaper for the platform owner to inspect, to screen out the presence of illegal products, illegal items. And we may have pros and cons also in this case. The pros are the usual are the pros coming from a reduction in the double marginalization because the platform will be active in two markets and it will internalize the externality between the complimentary logistics service and the primary one, but there will be cons because there will be less competition for fulfillment services because the platform will engage in, will foreclose rival by engaging in time. Still on competition, a stricter liability regime is also likely to impact platform fixed costs. On the one hand, if my impact marginal cost of serving users, sellers to control content, but it might also be the case that platforms will have to develop some technology in order to carry out some extent of verification. And this is a usual increase in the body of the entry that can benefit large incumbents because they have been operating for years under a different liability regime, but also because they will be able to cope with the higher costs and they will better cope with the higher costs than entrances. But we also benefit existing platforms, especially those platforms that rely on a data-driven business model because these platforms may have already data regarding, for example, notices and takedown or flags by third parties. And this material can be used to better train the algorithm, therefore reducing the liability cost compared to entrances. And therefore, if a stricter liability regime is not differentiated according to size, platforms of different sizes and competitive positions will not be affected equally, creating an income-based advantage as it was shown, for example, for the GDPR. But there might be other ways for platforms to react to respond to a stricter liability regime. For example, let's privacy, platforms might start removing online anonymity because they might have incentives to control, to better control users. And this is going to create an interplay with the current privacy regulations or it might be possible that the screening costs are passed onto users, especially on zero pricing platforms, the additional data monetization and less privacy. And this is something that we are investigating in a paper with Nestor Dutch Brown, you'll see that really a match like a sobolesque. There might be also other ways for platforms to adapt their terms of condition. And one way could be to neutralize liability by writing terms of condition to waive liability for victims, at least those that are participating in the platform ecosystem. But this requires market power because otherwise platforms might be hardly attractive for users, but we might expect, for example, that courts and regulators might challenge or might prevent these practices, but we need to keep an eye on this. And finally, we might expect that a stricter liability regime is going to affect third party and platform investments. For example, it might lead to more or less innovation by third parties like brain donors. And in any commerce platform, for example, stricter liability might help brain donors to invest, to innovate, to increase their excellent investment because of protecting them from forced competition by counterfeiters and from the reduced exposed litigation costs. But things can go the other way around. For example, if consumers value the presence of IP infringers of counterfeiters because they are providing, for example, some safe low-quality mutations. And if this leads to a demand contraction, there might be fewer incentives for brain donors to develop new products and new innovation. And these negative effects are likely to be amplified if liability also entails a higher permission fee for the brain donors. It is also important to take into account that there might be some, there might be more odds of screening because having a platform investing more in screening might crowd out right-holders incentives to screen and to monitor illegal conduct by third parties. And also in this case, the net effect is not going to be exactly clear. It's going to be dependent on whether the investment in screening technology by third parties and platforms are going to be complementary or duplicative instead. But when I also expect an amplified positive direct effect and this might come, especially when platforms might be, might have liable for products that are unsafe. And one way to think at this problem is that suppose that the platform needs to invest, needs to screen out unsafe products. Now, given the fact that the platform will have to screen out unsafe products, manufacturers might have an incentive exante to further increase their level of care. And therefore there's going to be an amplified positive direct effect, not only because consumers will have exposed safer product, but also so product will be safer as a consequence of the fact that the manufacturer will internalize platforms incentives. And finally, if we want to think about overall, the overall impact of stringent liability regime on platform investments, we may think of the investment in screening technology. There might be a new market for screening technologies. And these screening technologies can be developed in-house by the platforms. For example, YouTube has been running for years content ID for content recognition. But there are also, there is also a market with major parties that are operating, that are selling, that are licensing their technology to many online platforms, again for content recognition. But if we want to think more broadly about the effect on innovation, there might be an effect on innovation via the channel of competition. And we know from seminal papers that reduced competition as a consequence, for example, of a higher bar into entry caused by higher, but by stricter liability regime might have an effect that can be detrimental or even positive on the level of innovation. And finally, we might expect also changes on the direction of innovation because platforms will have to allocate their resources towards their investments. And there might be a bias towards those investments that lower liability costs. And also in this case, it might be unclear whether this is going to be socially desirable. Now, having identified what are indirect effects that should be taken into account when designing a liability regime along with the direct effect of a higher screening intensity has on the level of care accepted by the platform. There are some important issues that should be considered by policy makers. And we discussed this in the policy report. First of all, it is important to make sure that platforms, the tab incentive to monitor are actually able to do so. And for instance, as I was mentioning at the very beginning in the European Union, the e-commerce directive allows platforms to benefit from immunity, from like a built exemption as long as they are not aware of illegal conduct. And this creates this incentives for platforms to take some proactive measures because this way they are not going to know whether some illegal activities are carried out on the platform. And it is important indeed that a good summary and clause that is present in the US legislation that has been discussed also in the report by every 10, the Australian Martin Pies is included. And this is something that finally now is present in the digital services act proposed by the European Commission. There is another aspect that we have not discussed so far that concerns, for example, the fact that technologies are not perfect. There are errors and accuracy indeed is not perfect. And this might create incentives for platforms to over remove. Indeed creating a type of one error. And given the fact that platforms might over remove especially borderline content, borderline product, there might be also an incentive by for right holders, competitors or users that have different opinions to strategically flag to strategically notice the platform regarding some conduct committed by others. And the digital services act has a very important article on this topic because grants platforms the possibility to oblige platform actually to suspend those users that are frequently submitting unfunded complaints and notices. The second aspect that I'm going to discuss is that it would be important to introduce a layer to maintain the current liability exemption regime, but it would be important to maintain this exemption regime subject to some procedural obligations that make platforms accountable to the general public. And one way to do this is to make sure that platforms have a simplified and transparent system to report illicit material. But at the same time, there is control over the activity of the platform with the possibility to request feedback and to solicit the platforms. And at the same time, the platform should be responsive, should be, should timely respond to this solicitation. It is also important that platform engage in transparency reporting. And there is a very important closing the DSA that not only includes some of these obligations, but also imposes data sharing with authorities and researchers to very large platforms. So those platforms that for their dimension and for their role are quite big. And this is something that can foster research, especially empirical research. Given the potential availability of data if the DSA is finally approved. And finally, there is a final aspect that I would like to discuss that is partially related to the idea of having some large platforms and tries to consider to tackle the problem of having the competition problem from having a one-size-fits-solve regime. Having a one-size-fits-solve regime is likely to amplify current asymmetries in the market and increase market power of leading platforms or dominant platforms. So it would be important to include some additional clauses to big platforms as the DSA and the UK online safety deal do, especially for what concerns that is communication activities. But something that we discussed in the report is that it might be important to make sure that these platforms also take some pro-social, pro-welfare actions, like for example, sharing their past data and technology with small entrants. This might allow competition in the market, might also have entrants to prevent online misconduct. And the reason is that if platforms that are big are subject to additional obligations and they monitor their environment more closely, there might be an incentive of wrongdoers to move, to migrate towards small platforms that are not required to fulfill all these obligations. And therefore, if we want to make sure that we are living in a safer environment, it would be important for this platform, small platforms to have access to this data and indeed, why preventing misconduct also maintaining the contestability of this market. And then it is time to wrap up. So I hope that this has created some interest for a rather unexplored area of research which concerns liability for platforms. What we wanted to discuss is that there are many incentives given that there are many parties, different business models and therefore there are many theoretical and empirical issues to be investigated. From a policy point of view, we conclude that designing a liability regime is not an easy task because we are living in a second best environment. Policy makers and regulators can only affect the one strategy variable while leaving all the other instruments free for the platform. And therefore there might be unintended effects that should be taken into account. And also there is a lot of interplay with other dimensions that do not concern illegal activities. There is an interplay with competition policy, privacy regulation and consumer protection. And finally, the discussion today mostly focused on the economic dimension. But in social media platforms and hosting platforms, we should also take into account a fundamental right which is freedom of speech. This is to say thanks a lot for your attention and I look forward for Catherine discussion and any other comments. Thank you Leonardo, very interesting presentation. So let's move on to the discussion. So today we have Catherine Speer joining us from Cambridge US. So thanks for agreeing to give a discussion shortly in the morning for five minutes and then I'll open up for questions from the floor. Catherine, the floor is yours. Well, thank you very much. I'm so pleased to be here and I appreciate the opportunity to discuss this very interesting paper. In the five minutes that I've been allotted I'd like to do two things. First, I'd like to say a bit about the law and economics literature of liability and then offer some general reactions to the paper. There is a long literature going back several decades on whether firms should be held liable for the harms that their economic activities cause and the form that liability should take. I think it's useful to distinguish two branches of the literature. In one branch, the victims of the firm's activities are bystanders and so liability there would be a mechanism for controlling negative externalities on the bystanders. In another branch of the literature the victims are the consumers of the firm. So they're willing participants in the economic activity and in those settings liability is a mechanism for overcoming market imperfections and frictions and contracting. I'm going to discuss each of these in turn. So let's start with the setting where firms while engaging in an economic activity cause harm to bystanders. So for example, we might think about an oil company that causes an oil spill that harms the environment and harms local fishermen or other businesses. Now, of course there's a strong economic argument for holding firms strictly liable in these types of settings. Strict liability where the firms have to compensate the fishermen and the environment for all the harms caused will induce the firm to take cost justified precautions to avoid the harm. Moreover, since the firms are internalizing the harms to the victims and the costs of their activity they're going to scale back on the level of economic activity. And so liability in this setting can operate as a Paguvian tax forcing harmful industries to internalize those harms. Now, there also are downsides to using liability in these types of settings. First, liability is administratively costly and indeed with strict liability there would be grounds for a lawsuit whenever a victim is harmed. So one could get a lot of volume of litigation. Secondly, and I think maybe relevant for platforms if a firm has market power if it's an imperfectly competitive market then the market may already be producing too little. And so imposing strict liability on a market could actually make the monopoly distortion worse it can increase the deadweight loss. And so in situations like that, fault-based liability or the negligence rule may be preferable. I've just talked about firms harming bystanders. I think it's also relevant to think about another setting where the consumers of the firms harm bystanders while using the firm's product. Okay, so the firms are not harming the victims directly it's indirectly through the consumers themselves. I think this is highly relevant broadly think about gun manufacturers in the United States there are more than a hundred thousand deaths and injuries from guns every year. So the question is should gun manufacturers should be held liable? Well, there's a clear case for holding the consumers themselves liable when they've harmed others with the guns. Facing greater liability, consumers will take higher precautions and maybe lock away their guns to prevent accidents. And since liability increases the overall cost of gun ownership consumers will buy fewer guns. So the activity level will tend to fall. Okay, we might want to hold gun manufacturers liable too. We may want to extend liability in situations where consumers do not have deep pockets so they're judgment proof or they're otherwise not fully internalizing all the harms that are caused by their guns. If firms are held liable for the residual damages caused by guns then the price of guns will rise and fewer guns will be sold. And so in that way the consumers and the firms will jointly internalize the harms to the bystanders. Now I should also mention that holding firms liable for the harms caused by consumers can be problematic. Imagine a world where consumers are heterogeneous. There's some harmful consumer types, maybe the criminals who are using the guns and there are other safe consumer types, hunters who operate them safely. If the criminals have a higher willingness to pay for guns, well then holding the manufacturers of guns liable can be problematic. Insofar as the price of guns will rise to reflect the average harm caused by guns, as the prices of guns rises, it's going to drive out the hunters, the safe users and then the criminals will be the ones who buy the guns. And so liability actually in the extreme can destroy a socially valuable market. And so this is something that one needs to be concerned about some of these unintended consequences of liability. Now I've just talked about settings where it's the bystanders who were harmed. There's also a literature, and this was the focus within much of the paper where consumers themselves or users on the platform or platform participants are harmed. Okay, imagine that victims are consumers of the firms. Okay, so it's consumers who are harmed by a firm's product. The economic benefits and harms are now within the value chain. And so in a well-functioning market, the harms to consumers should be largely internalized and the case for liability on the firm is not as clear. Consumers are willing to pay a higher price for a safer product and firms have an incentive to maximize joint value and deliver optimally safe products in these types of settings where it's the participants themselves or consumers who are harmed by products, then market failures, contracting frictions, other information asymmetries, that may suggest that product liability is necessary. It's necessary, for example, if consumers misperceive product risks or product safety is an experience good and not observed by consumers at the time of sale. Okay, I realized that let me just take a couple of minutes and give comments on the paper. First, I thought this was a really interesting project and it's provocative paper is overflowing with good ideas. And even though there are no formal models or empirical analysis, the paper really is helping to set the agenda for future academic research in this area. This paper is highlighting the economic effects and trade-offs that policymakers should be taking into account when they're designing liability regimes. So some questions and some comments and suggestions. I think the paper might do a better job distinguishing between harms to platform bystanders and harms to the platform participants. This is an important theoretical distinction. And I think it would be helpful to a reader to understand and to see which of these the authors are talking about at various points in the paper. If the victims are the participants, then the paper might be more explicit and then underscore which market failures and contracting frictions are preventing the platform from achieving this joint efficiency. The authors might also think about tweaking some of the terminology in the paper. The phrase liability cost in the paper is used very broadly to include litigation costs but also compensatory transfers to the victims. If the victims are themselves part of the platform, if they're participants, then this could be neutralized through membership fees or transactions fees. And so theoretically, it's really important to, I think, break out those compensatory transfers when we're talking about harms to platform participants. Just to wrap up, I think a lot of the ideas in this paper can be developed into their own papers, hopefully with formal models and perhaps empirical support. This is such an exciting research agenda and I am very much looking forward to seeing what else the authors do and how they pursue this project further. Thanks.