11:32:29 From Vlad_NBERstaff : Welcome to SI 2020 Industrial Organization! 11:52:47 From Vlad_NBERstaff to liran einav(Privately) : Liran, if you would like to give me back the host authorization now, I could give you it back later whenever you need. I think it's time to begin livestream. 11:55:47 From Vlad_NBERstaff : We are now live on YouTube! 12:00:13 From Vlad_NBERstaff to liran einav(Privately) : Liran, can I have a host permission and give you back to you when you have break out rooms? Looks like YouTube livestream doesn't work when I'm a co-host. 12:04:03 From liran einav to Vlad_NBERstaff(Privately) : is this true? it worked last week …. I can't assign people to rooms if I'm not the host. Can you double check on this? I get request for assignment all the time, so need to be host for most of the time. 12:05:32 From Vlad_NBERstaff to liran einav(Privately) : Yes, only hosts can do livestream. 12:05:57 From Vlad_NBERstaff to liran einav(Privately) : Sorry about this. 12:05:58 From liran einav to Vlad_NBERstaff(Privately) : yes, but once it started, do you need to be host other than to stop it? 12:06:50 From Mo Xiao : Does the shared screen freeze? Or Emanuele didn’t move slides? 12:07:02 From liran einav to Vlad_NBERstaff(Privately) : I made you host, but I would need to be host again for breakout room assignment. let's talk about it in the first break (after the first paper). 12:07:16 From Mo Xiao : Problem solved 12:07:17 From Vlad_NBERstaff to liran einav(Privately) : Sounds good. Thank you. 12:07:47 From Vlad_NBERstaff to liran einav(Privately) : YouTube is now live again. 12:14:51 From Francine Lafontaine : Is the SIP connection of an independent bank revealed to customers - I.e. some “name” that is common to all members of a SIP 12:23:15 From Sergio Mayordomo : During the period covered in our analysis savings banks being part of a SIP were operating with the name they had before the consolidation process 12:25:05 From Jason Allen : NPL is 90 days late and chargeoffs are 180+. do you see same result for chargeoffs? And do we expect that both types of organizations have the same recovery rates? 12:26:15 From Mar Reguant : Do you have a sense of customer base overlap (geographical) between the acquiring and the acquired bank? 12:26:38 From Gloria Sheu : What should we think of as cost efficiencies in this context? I haven't seen a ton of documentation on the existence of cost efficiencies in banking due to mergers in the US, but this could be different in other markets. 12:27:06 From sofiavillas-boas : why are cost efficiencies known to be only realized two weeks after in banking? 12:27:23 From Mar Reguant : Gloria, Sergio knows better, but I think Spain used to have the highest amount of bank branches per capita (after bars), so they trimmed a lot of this. 12:27:56 From Sergio Mayordomo : @ Jason Allen: We have not checked it but we can deal with this issue and of course we will. Thanks a lot for this comment 12:28:32 From sofiavillas-boas : i meant two years 12:28:36 From Gloria Sheu : Are the merging banks located such that they have a lot of branch network overlap? We usually think of network overlap as occurring if they are located within the same county or MSA in markets in the US. 12:31:27 From Nicola Pavanini : @ Sofia Villas-Boas: There are a number of empirical papers documenting a reduction in operational costs for banks after around 2-3 years from consolidation (Focarelli, Panetta AER 2003 for example) 12:31:41 From Sergio Mayordomo : @ Gloria Sheu: Yes, Mar is right, it is mainly due to the high number of bank branches 12:36:36 From Sergio Mayordomo : @ Gloria Sheu: Savings banks had branches not only in the regions where the headquarters were located but they were spread over the whole country because regulation allowed for this 12:40:26 From Vlad_NBERstaff to liran einav(Privately) : Liran, I was told to give you back the host permission, so you are a host now. 12:41:30 From liran einav to Vlad_NBERstaff(Privately) : great. thanks! I checked with my students who are on the livestream and they didn't see any issue when we moved host duties back and forth around the beginning, so I think it's working fine.. let me know if you hear otherwise. 12:41:58 From Yizhou Jin : 05 12:42:57 From Vlad_NBERstaff to liran einav(Privately) : Awesome! Feel free to make me a co-host whenever you can. 12:44:08 From Sergio Mayordomo : @ Mar Reguant: In the case of SIP, it was larger than in the case of M&A. For instance, there were some SIP groups where all the constituents had the headquarters some km from each other. We will provide some figures on exact overlap in our sample in the next version of the paper given that we observe borrower-lender relationships. Thanks for this 12:51:59 From Thomas Wollmann : If you have questions for the authors, you can message now. 12:57:08 From sofiavillas-boas : thomas, my question was answered already. thank you 12:57:50 From Jean-Francois Houde : In the structural section of the paper, it would be interesting (perhaps in a different paper) to open the box of the cost efficiencies from consolidation: (i) reduction in origination cost (e.g. branch duplicates), and (ii) economies of scale in risk management (diversification). The second element would reduce the effect of adverse selection effect on MC (i.e. heterogeneous and less positive c1). 13:08:17 From Nicola Pavanini : @ Jean-Francois Houde: Thanks a lot for the comments, you are very right. At this stage we are just trying to show that marginal costs backed out from the FOC are increasing in quantity lent, consistent with the reduced form results, but we can definitely do more. 13:20:02 From Tobias Salz : So weights are not only unknown but also time varying? 13:21:23 From Eric Richert : @ Tobias Salz yes, the weights are unknown and auction specific 13:26:47 From liran einav to Vlad_NBERstaff(Privately) : how many people are tuned in to the youtube? 13:27:25 From liran einav to Vlad_NBERstaff(Privately) : also, just so I don't forget, once we are done, can you send me the chat log for the session so that I can share it with speakers? 13:29:37 From Thomas Wollmann : For those that joined late and missed the reminder: if you'd like to join a breakout room at the end of the day, send a chat message to Liran letting him know which paper's discussion you'd like to join when you've decided. 13:29:53 From Vlad_NBERstaff to liran einav(Privately) : Right now, 146 people are watching us on YouTube. Yes, sure. What email should I send it to? 13:30:46 From liran einav to Vlad_NBERstaff(Privately) : thanks! my email is leinav@stanford.edu 13:31:24 From Vlad_NBERstaff to liran einav(Privately) : Great 13:31:25 From Zach Brown : Is it possible that FDIC has preferences over bidding banks (and they choose weights ex-post to rationalize winners)? 13:32:36 From Robert Clark : They are not supposed to and there are annual audits to guard against this. 13:33:29 From Gloria Sheu : If the bidding banks are regulated by someone besides the FDIC (the Fed or the OCC), the FDIC will confer with that agency as to whether that agency has any objections to that bidder being involved. That could disqualify a bidder. 13:33:52 From Robert Clark : Rules changed in 1991 with the FDIC Improvement Act to have more transparency 13:35:08 From Robert Clark : That’s right Gloria. The FDIC establishes a list of qualified bidders and can consult with regulators when doing so. 13:35:57 From Gloria Sheu : Do you know whether a bidder was interested but was then disqualified? 13:36:14 From Robert Clark : No. That’s not something we can see. 13:36:19 From Brad Larsen : You probably already said this and I just missed it, but what motivates bidders to bid multiple times in a given auction here? what unobservable is varying across bid attempts of a given bidder with a given auction? 13:36:23 From Jeremy Fox : For a bidder making multiple bids, does the presence of one’s other bids as an argument to the probability of a focal bid winning result in an endogeneity problem, as your other bids are chosen to maximize the same profit function as the focal bid whose probability the term G is calculating? 13:36:39 From Alexander MacKay : I'm curious about the distribution of the discrete “switches” in the data. Do they appear to be (conditionally) random? Or are they clustered by auction and/or bank? 13:37:47 From Robert Clark : Brad: The uncertainty over the scoring rule motivates multiple bidding. 13:39:01 From Greg Lewis : Flip of Brad's question - why not infinite bids? Selection on # of bids as a problem? 13:39:52 From Chad Syverson : What does the FDIC think it is getting by keeping scoring rules shrouded? Is it just wrong about the relative sizes of the tradeoffs you're measuring, or does it think there are non-financial reasons for it? (Sorry if this was already addressed and I missed it.) 13:40:17 From Robert Clark : Greg: the substitution effect imposes a cost of bidding on extra packages. Plus we impose that they don’t bid on packages with less than 1% chance of winning. 13:40:50 From Robert Clark : Chad: with the FDIC Improvement Act they moved away from a more flexible system to one in which they were required to select the least-cost offer. but nothing was stipulated about the level of transparency required and so the FDIC just didn't reveal the exact rule it uses. From conversations with them it also seems like they believed that uncertainty would increase revenues. Moreover, the FDIC has a back-log of LS agreements, VAI agreements, etc, and they affect the weights in future auctions, but they don't want to reveal to the market all of this information. 13:40:51 From Mo Xiao : The other side of coin of Zack’s question: what if FDIC observes something about failing banks and set different weights in each auction to optimize some goal? 13:41:54 From Eric Richert : @Alexander MacKay we also try specifications including controls for bank level observables but they do not seem to make a big difference in the cost estimates 13:43:00 From Robert Clark : Mo: again, they are not supposed to according to the least-cost rule 13:48:07 From ekrasno1 : How do we know if they are even using a scoring rule? I the sense, do they use the same weight to evaluate different bids? 13:48:47 From Jan De Loecker : nice paper — have you guys put any thought in the political economy aspects of keeping the weights secret, to give the agency the freedom to act differently depending on the case in front of them (sorry if this was asked before) 13:51:01 From Jason Allen : @Elena: the LC formula is public. just not the weights. 13:53:13 From Robert Clark : Jan: Our understanding is that according to the FDIC Improvement Act they are not supposed to, but we agree that there may be stability vs market concentration issues at play. 13:54:35 From Ellie Prager : Do you know whether a second reason for multiple bids is that bidders are trying to learn (for next time) what weights the FDIC is using? Not suggesting you add dynamics to an already difficult problem, just curious. 13:56:12 From Eric Richert : Since the weights are auction specific draws they cannot learn them over time but we also did some reduced form testing for this and found no evidence of that 14:00:37 From Ryan Kellogg : Comments above were wondering if a motivation for not publicly committing to weights would be to favor a particular firm. Might it also be that the FDIC might learn from the bids how the discrete characteristics are valued, and then update its own valuation for those characteristics? 14:10:19 From Ellie Prager : Thanks Brett re: why learning is unlikely. Makes sense. 14:29:07 From liran einav to Vlad_NBERstaff(Privately) : Vlad, after the Q&A I will say a few works and then open the breakout rooms and then will give you host duties so that you can end the livestream, but then give host back to me because there would still be some people needing assignment (based on the experience last week). 14:29:41 From Jeremy Fox : Is a violation of fiduciary duty addressed using a lawsuit or regulatory action? 14:30:03 From Gaston Illanes : A lawsuit or arbitration 14:30:43 From Jesse Shapiro : Looking at equation (1), is there a reason why it wouldn't make sense to estimate the fully interacted model that allows the FEs to interact with adviser type, or equivalently to estimate separate models for RIA and BD advisers? 14:32:14 From Vlad_NBERstaff to liran einav(Privately) : Sounds good. 14:32:17 From Gaston Illanes : I think we could 14:39:12 From Chris Conlon : is it necessarily true a 14:39:43 From Chris Conlon : is it necessarily true that FC channel has the property even if number of firms adjusts? 14:40:24 From Brad Shapiro : Could a "high quality advice" firm sign a contract with a customer promising something like fiduciary duty in the absence of a law? Or otherwise credibly signal their high quality? Or is it entirely opaque? 14:40:55 From Gaston Illanes : @chris, we show in the paper that this holds under general conditions 14:42:37 From Gaston Illanes : @brad, they also say they’re high quality, of course. this + products that are hard to value make it quite opaque. 14:42:47 From Gaston Illanes : * they all say 14:42:57 From Mark Egan : @Brad Shapiro. In practice some firms/adviser groups have tried to do this, but the empirical evidence suggests their claims are not credible 14:44:11 From Jeremy Fox : On lawsuits, could there be class action lawsuits in this sector, making violating fiduciary duty at scale quite expensive? 14:44:44 From Manisha Padi : Jeremy, not under common law fiduciary duty, but this is what the department of labor rule tried to do (increase stringency) 14:45:24 From Brad Shapiro : @Gaston I see. Then in this sense, the true high quality firms might be in favor of FD rules to lessen competition from the lower quality firms? I wonder if you could find anecdotal evidence of something like that. It also seems that this opacity would make it difficult to prove FD violations in court? 14:47:37 From Gaston Illanes : @brad, that’s something I haven’t seen, but also not looked for. We could look. Re: your second point, it does, and it also makes it less likely someone would know to file an arbitration claim or a lawsuit to begin with. 14:48:20 From Gaston Illanes : an extreme version of that argument is a pure FC channel 14:49:23 From Andrea Szabo : Do you see differences in the commission structure offered to the advisors across states with different laws? 14:50:01 From Paul T Scott : I understand this might be difficult to answer concisely, but I can't help wondering: what do breaches of fiduciary duty look like in practice? 14:50:19 From Gaston Illanes : @andrea, We unfortunately do not observe commissions - this is why we focus on returns. the set of offered products, however, does not vary 14:52:07 From Manisha Padi : Paul, a typical situation arises when an investor experiences a loss, claims the adviser is to blame and takes them to court or to an arbitrator. We see lots of arbitrations and some, but not many, lawsuits. 14:53:05 From Paul T Scott : Is there usually a claim about conflict-of-interest, incompetence, something else? 14:54:28 From Manisha Padi : Both in lawsuits, we don’t observe arbitrators’ decisions so it’s hard to tell what proves a breach there 14:57:41 From Jing Li : Do you know if consumers travel across state borders to work with a fiduciary? 14:58:50 From Gaston Illanes : Yes - we do observe cross-border shopping, and its one of the observables that is balanced in the DID 15:15:46 From Panle Barwick : If some advisers practice in multiple states, could you exploit this source of variation? 15:21:40 From sofiavillas-boas : bye all