Extended: Revise paper to prepare for JUE

[santiagohermo]

NEW extended description of issue:

In this issue I will rewrite the intro, and maybe update some other sections, to prepare for the Journal of Urban Economics submission. The key goals, outlined in https://github.com/diegogentilepassaro/min_wage_rent/issues/266#issuecomment-1616713116, are to

1. Rewrite the intro saying that the paper is about the spillover effects of the MW within the city, and how the residence and workplace MW helps us estimate that.
2. Stress the importance of commuting patterns and of granular and high-frequency data for our research question. Drop the estimates using county and year variations.
3. Frame the estimates of the "share pocketed" as an illustration of the importance of commuting patterns.

---

OLD:

Revise text to discuss data unavailability to test channel of local prices

In this issue I will implement one of the requests in the AER referee reports.

Extracts from the cover letter submitted to AEJ

the AER editor suggested that we “oversell [our] results regarding the negative impact of the residence MW on rents,” and indicated that more evidence on the price channel that according to our model underlies said negative impact is needed. R2 made comments in the same direction, which we discuss below. To our knowledge aggregate data on prices of local consumption at the ZIP code level is not publicly available. We will revise the paper to note that due to data constraints we cannot test this channel directly and that more work is needed to conclusively establish that the negative estimated effect of the residence MW on rent arises from changes in local prices.

R2 notes that the residence MW does not take into account where residents of a ZIP code shop, it simply assumes that they shop in their own ZIP code. (In footnote 11 we mention a possible extension to the model where this would be allowed.) We are not aware of origin-destination data on consumption intensity between ZIP codes, but we do have access to data on the amount of retail activity in each ZIP code, which we can use to check for heterogeneity in our findings.

[santiagohermo]

Should we thank the anonymous referees from the AER?

[diegogentilepassaro]

Should we thank the anonymous referees from the AER?

Claro que si @santiagohermo !

[santiagohermo]

Thanks @diegogentilepassaro! The title page is looking like this

[santiagohermo]

Maybe we have to use "Across Local Housing Markets" instead of "On Local ..."?

[diegogentilepassaro]

Maybe we have to use "Across Local Housing Markets" instead of "On Local ..."?

Me gusta across también! Un par de comentarios:

• Dejé otro comment acá.
• Agradecería a Oded también. Aunque haya metidos pocos comments, fue a las presentations e igual lo revisó para mi tesis.
• Podría ser una buena idea agradecer a algunos de nuestros compañeros que lo hayan leído o metido comments porque no cuesta nada.
• Si querés me podés cambiar el affiliation a Amazon ya a esta altura.

[santiagohermo]

Gracias @diegogentilepassaro!

Me gusta across también! Un par de comentarios:

Le pregunté a GPT4 y dice que "on" es mejor jaja.

• Dejé otro comment acá.

Metemos "Residence" nomás. Me gustaba que "workplace" rima con "workplace" y la frase quedaba catchy, pero tenés razón que residence es más claro y no genera confusiones con el resto del paper.

Implemented in c00bb9b

• Agradecería a Oded también. Aunque haya metidos pocos comments, fue a las presentations e igual lo revisó para mi tesis.

Sabelo.

Implemented in c00bb9b

• Podría ser una buena idea agradecer a algunos de nuestros compañeros que lo hayan leído o metido comments porque no cuesta nada.

Sabelo x2. Quién tenés en mente? También puedo agregar a algunos pibes que nos hicieron comentarios via email (creo que yamagishi y alguno más)

• Si querés me podés cambiar el affiliation a Amazon ya a esta altura.

Sure! Should I use Amazon Pharmacy? Btw, @gabrieleborg has Amazon Web Services. Is this still accurate gabri?

[diegogentilepassaro]

@santiagohermo:

• Del affiliation mete Amazon nomás :)
• De a quienes agradecer: a los que enviaron comentarios por email seguro a todos! De los pibes, me acuerdo de Joao, Marcella, y JP Colombia pero quizás me estoy olvidando de alguien. Santi Peréz nos metió algo de feedback? No me acuerdo bien

[santiagohermo]

@diegogentilepassaro:

• Del affiliation mete Amazon nomás :)

Done! I also changed @gabrieleborg's to "Amazon".

• De a quienes agradecer: a los que enviaron comentarios por email seguro a todos! De los pibes, me acuerdo de Joao, Marcella, y JP Colombia pero quizás me estoy olvidando de alguien. Santi Peréz nos metió algo de feedback? No me acuerdo bien

A algunos pibes ya agradecemos cuando decimos "Applied Lunch at Brown". Creo que habría que mencionar a gente con la que tuvimos una reunión especial sobre el paper tema, o leyó la intro e hizo comentarios.

Gente que me leyó el draft anterior:

Yamagishi

Hi Hermo, I finally got to skim through your paper, this is very interesting and I enjoyed it!! A few thoughts came to my mind: I think it would be helpful to clarify upfront that your approach assumes away any "(dis)employment effect" of minimum wages. As you know the MW literature has always centered around this issue. Even in the smaller literature of the MW + spatial economics, many (e.g., Cadena 2014 JUE; Monras 2018 JLabor E; and my RSUE paper) are motivated by inferring something about the potential disemployment effect. Let me be clear: I think it's ok to assume it away given the recent evidence and that this assumption gives you a lot of fruits. But I think positioning your paper out of the disemployment effect early on would help readers correctly understand your contributions. I am not very sure about the time horizon in which you model can be applied. On the one hand, you are assuming away changes in workplace residence, giving it a flavor of the short-run analysis (, and I think this was mentioned in paper here and there). On the other hand, your analysis assumes that the housing demand is endogenous on the intensive margin, which creates the link between the MW and the housing rents. But I think the adjustment in the individual housing demand is something that takes time to appear: one needs to build or scrap houses, redraw the boundaries of land plots etc. That is, assuming the endogenous housing consumption gives some flavor of the medium-run or long-run analysis. I think it would help to think a bit more about what situations or time horizons can best fit your model The incidence of MW hikes on home/land owners is something you could compare to previous studies. For instance, my RSUE paper estimates that landlords pocket around 7.5-13.5 cents for every dollar increase in worker income, which is surprisingly close to your headline numbers despite many differences of my paper and yours. To me, the incidence of MW is one of primary reasons why we need more studies on MW and housing markets, so I'm simply interested in this. Putting your number in the previous estimates while further emphasizing the rich heterogeneity uncovered in your paper perhaps helps for readers. I hope you find these comments helpful, at least somewhat! Thanks so much again for sharing your draft. Best regards, Atsushi

Sam Hughes

Hi Santiago & Diego, I’m happy to hear about your new draft (and also congrats on presenting it at UEA, if I read the uea schedule correctly). Your estimation/identification strategy is a really nice innovation. Several years ago I noticed in the Census ACS data that low-wage workers who work in a higher minimum wage state but live across a nearby border in a lower minimum wage state spend more on rent, but didn’t have a good identification strategy or framework for thinking about individual income elasticity of demand versus higher prices. It’s great that you all took a stab at this—particularly for studying city-level minimum wages. It might be nice to more clearly separate city and state-level MW changes—my simple estimates suggested that MW will have different housing effects in cities with more inelastic housing supply/stringent land use regulation (which likely look different on many state borders). I didn’t see it cited in your current version, but I’d encourage you to add a little discussion of Jorge Perez Perez’ jmp which has a similar strategy, but looks at changes in the number of low-wage commuters between central tracts and those outside the city. Again, perhaps I didn’t read carefully enough—but it didn’t look like you all estimated whether the number of commuters/workers was changing between tracts depending on differences in residential/weighted place of work MW. In an event-based MW change approach (a la Cengiz, Dube et al), it should be possible to look at whether the number of commuters by income group is also changing. I think this connects back to the issue of whether you are going to take a strong stand on the negative coefficient on residence MW—it took me a while to figure out the argument about density of MW workers working in the zip leading other prices to be higher and rents to be lower. I think you could expand on this more, or find a more refined test of this hypothesis, because it looked like the heterogeneity/interaction regression were not significantly different from 0 (or maybe I misinterpreted). In particular, is this higher other price channel coming directly from operating rental housing or from non-housing goods (i.e. the grocery store price channel)? A difficult issue to think through is how demand shifts across types of homes & types of neighborhoods—people whose incomes have increased will presumably demand some combination of higher quality homes, higher quality neighborhoods, and perhaps locations closer to work (depends on the elasticity of commute time to own income). In particular, I thought Agarwal, Ambrose & Diop made a nice attempt in the first version of their paper to distinguish between the effects of the MW change on the same tenant whose income increased subsequently increasing the quality of housing they live in, versus an increase in the price of constant quality housing (I also don’t have a great way to get at it—besides the crude DDD-style regressions in my paper). My understanding is that Perez Perez (2022) suggests that people (or jobs) shift away from commuting from low wage workers’ tracts after minimum wage increases. Likewise, my interpretation of Eriksen & Ross (2015) is that as housing vouchers increase, rental prices near the threshold for voucher recipiency increase, but rents far below FMR decline (consistent with a decrease in demand for very low-quality housing as vouchers allow low-income household to buy higher quality housing). Your paper would be an even greater advance if you could show some evidence of these heterogeneous effects consistent with this type of intuition. The last thing I thought of (naturally) is the extent to which non-homotheticity matters for your welfare calculations. As a Jessie Handbury student, this is something I’ve thought about quite a bit, and my MW paper was a crude attempt to show that the non-homotheticity of housing demand/curvature of the utility function in housing consumption could matter quite a bit in studying the effects of MW. It would be a nice addition if you allowed for changes in rent-to-income ratios and showed whether/how that influenced your estimates of incidence between landlords/tenants. Sorry if some of this was abbreviated/a little scattered—I’m about to leave the office to go out of town for a few days. Hope the thoughts were helpful and again congrats on the great draft, Sam

Ben Hyman

Discussant de la [presentación en UEA](https://github.com/diegogentilepassaro/min_wage_rent/wiki/Presentations#uea-meeting-april-2022). Capaz haya que mencionarlo.

Juan Pereira


Kyle Butts

[min_wage_rent_kyle_comments.pdf](https://github.com/diegogentilepassaro/min_wage_rent/files/11940197/min_wage_rent_kyle_comments.pdf)

We sent the below to Giacomo but he never replied, so I wouldn't add him.

Giacomo R


Vos tenés algunos que cumplen este criterio?

[diegogentilepassaro]

@diegogentilepassaro:

• Del affiliation mete Amazon nomás :)

Done! I also changed @gabrieleborg's to "Amazon".

• De a quienes agradecer: a los que enviaron comentarios por email seguro a todos! De los pibes, me acuerdo de Joao, Marcella, y JP Colombia pero quizás me estoy olvidando de alguien. Santi Peréz nos metió algo de feedback? No me acuerdo bien

A algunos pibes ya agradecemos cuando decimos "Applied Lunch at Brown". Creo que habría que mencionar a gente con la que tuvimos una reunión especial sobre el paper tema, o leyó la intro e hizo comentarios.

Gente que me leyó el draft anterior:

Yamagishi Sam Hughes Ben Hyman Juan Pereira Kyle Butts We sent the below to Giacomo but he never replied, so I wouldn't add him.

Giacomo R Vos tenés algunos que cumplen este criterio?

Gran compilada @santiagohermo, muchas gracias! Solo Oded, así que demosle con esos nomás :D

[santiagohermo]

Excelente @diegogentilepassaro! En los commits de arriba reformulé la thankyou note. Avisá si ves mejoras! Así van quedando la title page:

[santiagohermo]

Continues in #269

[santiagohermo]

Summary: In this issue we started changing the paper for a new submission. See the opening comment of the PR for details on the changes implemented in this branch.

Changes merged to master in https://github.com/diegogentilepassaro/min_wage_rent/commit/64ee9191d7d04a9ee9f46430f879aaf388ae2968, updated PDF as of the merging of this task is here