Skin tone penalties:

class: center, middle, inverse, title-slide

.title[
# Skin tone penalties:
]
.subtitle[
## Skin color gaps and discrimination
]
.author[
### Guillermo Woo-Mora
]

---

<style>

.center2 {
  margin: 0;
  position: absolute;
  top: 50%;
  left: 50%;
  -ms-transform: translate(-50%, -50%);
  transform: translate(-50%, -50%);
}

</style>

.center2[
# Inequalities are not colorblind
]

---

---

---

---

.center2[
# How do you clasiffy yourself?
]

---

.center2[
# How do others clasiffy you?
]

---

.center2[
# How do we actually look?
]

---

## Identity

- People think of themselves in terms of certain groups that they belong to

> A white female French economist may identify as white, as a female, as French, or as an economist (or all or none of the above) (Shayo 2020)

- Individuals choose their identity (Akerlof and Kranton 2000, Shayo 2020)

`$$U_{iJ}(a) = \pi_i (a) - \beta_i d_{iJ}(a) + \gamma_i S_J(a)$$`
--

- Inequalities through differences between groups `$J$`

## Physical attributes (phenotype)

- Some individual aspects are not easy to change (i.e. stature, biological sex, skin tone)

- Identities may depend on physical attributes

- Inequalities can be directly linked to physical attributes (independent of identity)

---

.center2[
# Ethnicity ≠ Race
]

---
# Ethno-racial inequalities

<img src="imgs/chetty2014.png" width="70%" style="display: block; margin: auto;" />
.center[
Chetty et al. (2014)
]

---
# Ethno-racial inequalities

.center[
Chetty et al. (2014)
]

---

.center2[
# Other latitudes?
]

---
# Latin America

---
# Latin America

.center[
Woo-Mora (2023)
]

---
# Latin America

.center[
Woo-Mora (2023)
]

---
# Income gaps

.center[
Woo-Mora (2023)
]

---
# Educational gaps

.center[
Woo-Mora (2023)
]

---
# Intergenerational mobility

.center[
Woo-Mora (2023)
]

---
# Intergenerational mobility gaps

.center[
Woo-Mora (2023)
]

---

.center2[
# What causes these skin tone gaps?
]

---

.center2[
# Discrimination?
]

---
.center[
# Discrimination?
]

> members of a minority group (women, blacks, Muslims, immigrants, etc.) are treated differentially (less favorably) than members of a majority group with otherwise identical characteristics in similar circumstances (Bertrand and Duflo 2017)

.center[
#### Cirmcustances may differ from the origin and cumulate through time
]

.center[
Claudia Goldin
]

.center[
#### Difficult to find contexts where groups have *identical characteristics in similar circumstances*
]

---

.center2[
## One context where individuals with different skin tones do the same task but are treated less favorably
]

---

.center[
<img src="https://e0.365dm.com/16/06/2048x1152/panenka_3478366.jpg?20160604162257" width="90%" style="display: block; margin: auto;" />
]

---
# Kamel and Woo-Mora (2023)

.center[
### We exploit football as a lab
]

- High quality headshots to infer skin tone

- Players valuation (market value, ratings, wages)

- Detailed information on performance

- Situations where players do the same task (and perform equally well) but are treated differently

.center[
<img src="imgs/sofifa.png" width="47.5%" style="display: block; margin: auto;" />
]

---

.center[
<img src="imgs/uk_kane_penalty.png" width="47.5%" style="display: block; margin: auto;" />
]

---

.center[
<img src="imgs/uk_saka_penalty.png" width="47.5%" style="display: block; margin: auto;" />
]

---
# Skin tone segmentation and classification

.center[
<img src="imgs/color_classification_exp1.jpg" width="85%" style="display: block; margin: auto;" />
]