11:31:16 From Vlad_NBERstaff : Welcome to SI 2020 Industrial Organization! 11:56:55 From Vlad_NBERstaff to liran einav(Privately) : Could you please make me a co-host? 11:56:59 From Vlad_NBERstaff to liran einav(Privately) : Thanks! 11:57:08 From liran einav to Vlad_NBERstaff(Privately) : yes. just did 12:26:05 From Sofia’s iPad : Meredith. still in title page 12:29:17 From Paul L Joskow : It is not just coal units that have retired in New England. Nuclear units have retired (Pilgrim and Vermont Yankee) and many oil plants (often dual fuel oil/gas) have retired as well. The latter were primarily older steam plants with relatively high heat rates. The new are very efficient CCGTs. There are older less efficient CCGT units that are slowly retiring as well. 12:57:14 From Arshia Hashemi : What is your concrete measure of "exposure" to the supply side shock? 12:57:35 From Dan Sacks : we look at chains that have presence in the state that raised its excise taxes 12:57:57 From Dan Sacks : in some analyses we stratify on their pre-tax-increase revenue or revenue share in the state raising its excise tax 12:58:36 From Carlos Noton : Does the nature of permanent or transitory cost shocks matter here? 12:58:48 From Arshia Hashemi : thanks. so there's a dummy exposure variable, and also a continuous exposure variable. 12:59:07 From Ellie Prager : This is a really nice idea. Is there a particular reason to focus on just ~40 products? UPC codes are super granular, right? 12:59:13 From Dan Sacks : yes, Arshia 12:59:28 From Dan Sacks : Carlos I think it might. the tax changes we look at are basically permanent (or last multiple years) 12:59:31 From Andrew Butters : @carlos yes it certainly does. 12:59:58 From Dan Sacks : we focus on these few products because they're widely sold so prices are never missing - you might worry with rarely sold products that the composition changes after the tax change 13:00:02 From Andrew Butters : especially in the context of connecting our results with other findings in the pass-through literature 13:00:26 From Jing Li : On the specific set of products you are studying that are exposed to excise taxes, are you able to replicate the uniform pricing results in the literature? It’s possible that for these products, stores respond to the cost shocks that you’ve identified as well as to demand shocks? 13:00:34 From Chris Conlon : Don’t retailers for alcoholic beverages have to buy from in-state wholesalers (the three tier system). One might think this is a feature unique to alcohol that applies to basically no other goods sold in stores? 13:00:42 From Dan Sacks : we do replicate the uniform pricing for these products, in the paper's appendix 13:01:30 From Ellie Prager : @Chris this wouldn't be a problem for cigarettes, right? 13:01:38 From Dan Sacks : we will see similar patterns for cigarettes also. it is true that the wholesale market is state specific here though - we think that's one reason for uniform pricing (mostly wholesale don't vary across states) 13:01:42 From Andrew Butters : @Chris Conlon we do study cigarettes as well 13:03:27 From Jean-Pierre Dube : do you have any examples for categories that don't have such inelastic "category" demand? Alcohol and cigarettes are outliers (i.e., consumption relatively inelastic). 13:04:04 From Chris Conlon : I admit I have no idea how wholesale distribution of cigarettes is organized. Does anyone know? 13:04:19 From Sofia’s iPad : did you ever ask the chains about pricing. from previous research and data access to large retail chains in the us we learned, I learned that some chains use division level pricing, sama prices promotions and display and variety for stores for which they classify a pricing division. the division typically stays inside the same date but they have different divisions within states. and also different product categories have different pricing divisions. 13:04:20 From Andrew Butters : @Jean-Pierre Dube - we do not currently have any results beyond alcohol and cigarettes. 13:04:27 From Dan Sacks : we focused on these categories because they're subject to excise taxes (possibly they're taxed precisely because they have inelastic demand) , so unfortunately no toher categories, but if you suggestions we're very interested 13:05:06 From Dan Sacks : we have not talked to chains but we would be very happy to 13:05:17 From Jeremy Fox : On JP’s point, store choice might be very sensitive to price in terms of alcohol and cigarettes as these are focal products for some consumers of them. So stores must be careful about setting these prices versus, say, candy you just grab on a whim. 13:05:19 From Sarah Moshary : Have you considered how assortments change in response to different types of shocks? 13:05:22 From Sofia’s iPad : so it is not uniform model or non uniform model. they have division local pricing and not overall uniform 13:06:22 From Dan Sacks : if the divisions are at the state level, yes I think that's right 13:06:46 From Andrew Butters : @Sarah on assortment: are you thinking the difference between cost vs. demand shocks? 13:07:29 From Paul Scott : is it possible that this market is almost perfectly competitive? that is a form of uniform markups 13:07:51 From Jean-Francois Houde : Changes in sales taxes can be a demand shock for stores in unexposed states, located near the boundary of exposed states (assuming that consumers shop across state boundaries). This seems testable within the same framework. 13:08:08 From Matthew Backus : Do you have any sense of the scale of the (potential) mispricing as compared to prior work, in terms of profits? 13:08:24 From Dan Sacks : we did look at spill overs (for example in Portland, which is acorss the river form wa) and we found basically none for beer, although people do find spillovers like this for cigarettes. we do some robustness to excluding border counties 13:09:38 From Dan Sacks : we don't estimate lost profits because we're not sure what optimal pass-through should be, basically because we don't know how elasticity varies as price varies (we have tried to estimate this without much success) 13:10:54 From Paul Scott : To clarify the case for this market being perfectly competitive: of course I know that retail prices are higher than wholesale prices. But do retail prices correspond to wholesale prices plus operating costs plus the opportunity cost of shelf space? 13:11:25 From Paul Scott : This theory seems to make sense of passthrough=1 and prices not being related to demand elasticities 13:11:47 From Sarah Moshary : @Andrew, yes, to see what other levers firms use to respond to different types of shocks 13:12:13 From Dan Sacks : I agree that theory would rationalize our results, and we could think harder about this. I can tell you that demand has finite elasticity, and we find that across chains (not within chains) price is correlated with income 13:12:21 From Dan Sacks : (responding to Paul) 13:12:49 From Paul Scott : I think when people estimate demand elasticity, it's a demand elasticity with respect to all retailers raising their prices in similar ways 13:12:54 From Andrew Butters : @sarah: we have been able to replicate the DG (2019) QJE results in relation to differences of income across geographies 13:13:01 From Jesse Shapiro : I upvote Paul's point that competitive retail is an interesting benchmark. 13:13:01 From Andrew Butters : (as a demand) 13:13:08 From Paul Scott : so it's not the partial demand elasticity that you would need to assess competitiveness 13:13:32 From Jesse Shapiro : As Dan says one piece of evidence against that is the within vs. between chain difference. Another is DV-G's finding that prices are not just similar across outlets within a chain, but really really similar. 13:13:33 From Dan Sacks : right, we haven't estimated the industry-wide elasticity 13:14:10 From Jeremy Fox : Paul’s comment relates to mine if they make their profit off other products bought by consumers who pick a store based on alcohol and cigarette prices (high elasticity on alcohol, cigarettes and lower elasticities for other products) 13:15:00 From Dan Sacks : yes we should think more about that hypothesis, thanks Jeremy 13:16:54 From Jing Li : Following your rationalization of uniform markups, should we expect to find that firm-wide cost shocks are uniformly passed through rather than differentially passed through based on local demand elasticities? Would you be able to test this, and maybe that will allow you to expand the set of products you study beyond alcohol and tobacco? 13:18:24 From Dan Sacks : that is a nice idea! we're not sure what the right benchmark is since we're not totally comfortable with the constant elasticity model as a benchmark. so we're not sure how much pass through should vary acorss markets. but I think that would be a really nice test 13:19:13 From Wesley Blundell : One set of products to consider besides alcohol and cigarettes might be bottled sports drinks or water. Recent per unit plastic taxes like those in CA would function very much like the excise taxes you study but those products may not have such inelastic demand. This would help to understand whether your result is generalizable across consumer products. 13:19:16 From Andrew Butters : @Jing Li: to fully evaluate that hypothesis you would want to have local measures of demand curvature. But our interpretation would be that a firm-wide cost shock would be passed through at a similar rate for all the firms stores. 13:20:12 From Dan Sacks : thanks Welsey that is very helpful. we wanted to look at soda taxes but they were too recent to show up in the data when we started the project. they might be there now though! 13:20:13 From Ellie Prager : Following up on @Wesley, there are also sugar taxes on drinks in many localities (sometimes but not always on a per-unit basis) 13:20:23 From Ellie Prager : Oops, never mind 13:21:37 From Mar Reguant : What about packaging taxes? (e.g., associated to refunds on returned bottles?) Those tend to be state by state? 13:22:03 From Sofia’s iPad : the pass through on sugar taxes is also at wholesale and is not always completely passed through. the soda taxes are in the Nielsen data now, we are looking at those and passhtough is not perfect 13:22:04 From Andrew Butters : @Mar that's a great idea. Thank you. We'll have to investigate that. 13:22:08 From Dan Sacks : that is a great idea! We should look into how much those vary over time but yes packaging sounds great 13:23:19 From Mar Reguant : @Sofia This is interesting! I think sugared vs non-sugared are much closer substitutes than alcohol and cigarettes, so it seems like an interesting and relevant comparison 13:23:50 From Mar Reguant : e.g., uniform pricing between coke and coke zero contraints 13:25:44 From Sofia’s iPad : another way to pass through that I found is not to change the shelf price but to offer the product on sale less often 13:26:08 From Dan Sacks : ah so we would be missing that channel by focusing only on the commonly sold products 13:26:31 From Sofia’s iPad : you can look at whether percent times products on sale changes 13:26:43 From Dan Sacks : yes that's a nice approach 13:26:56 From Nancy Rose : on Glenn 13:27:12 From Chris Conlon : many states ban temporary sales on alcoholic beverages… 13:28:07 From Nancy Rose : on Glenn's final point- there is a literature on this oligopoly coordination for cigarette tax hikes. Makes it more interesting to see if these results generalize to other products, non-excise taxes 13:28:50 From matthew gentzkow : Dan — Quick question, apologies if you addressed this and I missed it — We know that sales taxes are effectively passed through to prices ~1:1 because we observe ~uniform shelf prices and sales taxes are added on top of them at the register. To what extent could I think of this as firms treating exise taxes like sales taxes? I.e., having an explicit policy of adding exise taxes on top of a base price. 13:30:12 From Jean-Pierre Dube : The Kilts data includes some Promodata which gives information about wholesale prices (retailer's marginal cost), albeit only for packaged goods. 13:30:30 From Dan Sacks : Matt - I think that could rationalize what we're seeing, because the distribution of pass-through rates is centered around 1 - it does look like they're just adding in the excise tax 13:32:01 From matthew gentzkow : @Dan, thanks. I guess an implication of that is that this might be specific to taxes and not apply (as much) to other wholesale cost shocks. 13:32:51 From Dan Sacks : it might be, we think it's striking though that excise taxes are one case where there is clear and salient geographic variation in whole costs. 13:35:25 From matthew gentzkow : @Paul S: We show in my paper w/ Stefano that elasticities are < 1 only makes sense if demand is log-convex — usually in intermediate micro we assume log concavity — we don’t need to but it makes life easy... 13:40:16 From Sofia’s iPad : do you also have cost drop variations and find asymmetries? 13:41:09 From Sharat Ganapati : @Chris; is there any good evidence on the concavity of demand? 13:41:21 From Dan Sacks : @ mar, thanks, yes that was a potential concern. 13:41:46 From Dan Sacks : @sofia, we do have 2 cost drops, it's a bit noisy but we see prices fall with a pass through rate a bit less than 1 14:03:14 From liran einav to Vlad_NBERstaff(Privately) : Vlad, like yesterday: (a) what's our livestream count today? (b) don't forget to send me the chat log after it ends. thanks!! 14:06:10 From Vlad_NBERstaff to liran einav(Privately) : Liran, 256 people are watching us on YouTube. Yes, I remember. Did you receive my email with the chat yesterday? 14:07:44 From liran einav to Vlad_NBERstaff(Privately) : great. thanks. yes, I got it yesterday (I replied (with a thanks) to you and Rob, but I think you sent it from some weird zoom/nber account so it bounced back). I then clean it up from the private chats and share it with the speakers, which they find very useful. 14:09:27 From Vlad_NBERstaff to liran einav(Privately) : Awesome. Yes, it was one of the NBER emails. I'm glad you find it useful. 15:33:11 From Thomas Wollmann : Let's all hope the inputs immediately adjust to demand for the vaccine!