Dana Scott

Thanks for the question! I focus on workplace flexibility, encompassing temporal characteristics of jobs—when, where, and how flexibly one works. The literature thinks of amenities as any non-pay characteristics of jobs. As I argue in the paper, these are choices firms make—certainly not arbitrary!

Read the whole paper here. I'm on the market this year and looking forward to discussing this work!

These findings suggest that we can’t simply regulate our way to a future with more women-friendly workplaces. Thinking about ways to reduce coordination costs is more likely to stick—but certainly a lot more challenging to imagine.

Using the model, I can ask what would happen if we required firms to provide a minimum level of flexibility. While this would increase amenity provision, it would *increase* the gender pay gap by 1.8pp because it would deepen gendered sorting across jobs.

On the firm side, I find that the productivity cost of flexibility varies sharply by skill level. High-skill occupations in competitive markets (e.g., finance jobs in Paris) face 2x the penalty of mid-skill jobs. This explains why high-paying jobs often bundle good wages with demanding schedules.

Women's higher WTP for flexibility (30% higher than men's) combined with lower labor supply elasticities (25% lower) creates a double burden. Preference differences explain 20% of the gender wage gap—a substantial share, but certainly not everything!

On the labor supply side, I identify labor markets nonparametrically using E:E transitions, estimate rich substitution patterns across occupations and jobs, and recover job-specific labor supply elasticities WRT wages and amenities for men and women.

These suggest that there are important roles for both firms’ technological constraints and workers’ preferences. To explain these patterns, I estimate a model where firms choose wages and flexibility amenities subject to BOTH worker preferences AND productivity costs of providing the amenity.

I document three empirical patterns: (1) flexibility co-varies within firms across establishments and occupations; (2) some flexibility dimensions correlate positively and others negatively with pay; and (3) women sort into jobs with better flexibility within occupations.

Studying amenities is hard in part because they are difficult to observe. I link French matched employer-employee data to large-scale surveys to observe a range of amenities related to workplace flexibility. This addresses the measurement challenge and helps us earn about firms' endogenous choices.

Importantly, compensating differentials are *equilibrium* trade-offs between wages and amenities. Worker preferences are one side of the equation. But we know less about the firm’s side. With flexibility in particular, we might be concerned that flexibility may affect workers’ output productivity.

I study these questions by focusing on workplace flexibility—what Goldin (2014) defines as the “temporal aspects of work”: when, where, and how flexibly one works. Flexibility is especially important for gender inequality, because we know that women have a higher WTP for it than men do.

🚨 JOB MARKET PAPER ALERT! 🚨

Why do firms provide different workplace amenities? And how does amenity provision affect labor market inequality? 🧵

Giving a lunch talk today and my April Fools bit is referring to it as my "job market paper" as if there is going to be a job market in the fall 🤪

Yes this is a much clearer summary of what was bothering me about it than what I had — thank you!

Sounds ominous... thanks Vitor!!

As for estimand, probably effects of some occ-level treatment on Y (an ATE that averages across occs in some way is less interesting than the occ-specific treatment effects I think)

Yes, so in this example a nursing degree is informative for whether someone becomes a nurse rather than anything else, but a liberal arts degree is less informative about relative probabilities of most occ choices (aside from maybe "not nurse")

Suppose I have an instrument Z for some endogenous variable X. Z does a much better job predicting X for some X values than others. E.g. major better predicts occupation for nurses than for management consultants. What problems does this introduce? Is there a go-to reference for this issue?

we're all just Chris from Parks and Rec at the end of the day

Thanks! year 4 has the strongest middle-of-the-ocean vibe yet

Hello #EconSky 📉📈

I'm Dana! I'm a 4th-year economics PhD student at Yale. I study labor markets and dabble elsewhere: public, household finance, (aspirationally) applied econometrics. I race marathons and bikes in my spare time. Excited to see what is happening over here!