Practice 01 - Regression Discontinuity Designs in Development Economics

class: center, middle, inverse, title-slide

.title[
# Practice 01 - Regression Discontinuity Designs in Development Economics
]
.subtitle[
## Theory and Practice
]
.author[
### Bruno Conte
]
.institute[
### Barcelona School of Economics
]
.date[
### 08/Jul/2024
]

---

# RDD in Development Economics: Practical Classes

- Goal of theses classes: put in practice the lecture contents

- Structure:

- 20-40 minutes: exposition
  
  - ~ 1 hour: "hands-in" practice (monitored by me)
  
  - Mostly in `R`, `Stata` material will be also provided
  
- Main advantage: most of "hands-in" activities are part of assignment (but not all)

- Any <o>**questions**</o>?

---

# RDD in Development Economics: Practical Class 01

**For today**:

- Introduction to `R`

- Basics treatment effects (randomized treatment and IV)

- Reporting with `rmarkdown`

Download the material [here](https://www.dropbox.com/scl/fi/n43d50i7oiuwbwqlnopfx/00_practice01.zip?rlkey=vw4vozvz37i0pvskm3m02xpdr&st=s3v4huxt&dl=1).

Next slides: questions for assignment

---

.center[

# Take-Home/Assignment Exercises

]

---

# Take-Home/Assignment Questions (1/4)

Part 1: "*Indirect Effects of an Aid Program: How Do Cash Transfers Affect Ineligibles’ Consumption?", Angelucci and De Giorgi, <g>American Economic Review</g>, 2009*

1.a Using the data in `data_angelucci.csv`

- Replicate the results of the rows *Control*, *Treatment* and *ITE (No Controls)* of Table 1

- Use both difference in sample averages and regressions

[Hint: *ITE* here does not necessarily mean *Intentions to Treat Effects*; read Section III]

1.b Interpret the outcome of the row *ITE*. Is this estimate causal? Motivate your answer;

---

# Take-Home/Assignment Questions (1/4)

Part 2: "*Lifetime Earnings and the Vietnam Era Draft Lottery: Evidence from Social Security Administrative Records", Angrist, <g>American Economic Review</g>, 1990*

2.a The dataset `data_angrist.csv` contains information on cohorts of 1950, 1951 and 1952

- Run a regression of "future earnings" (i.e. earnings in years 1981-1984) on "veteran status" dummy. Discuss the potential bias of the estimated effects.

- Run an IV estimation instrumenting "veteran status" with "eligibility" dummy, as done in class (using `ivregress`). How does the point estimate of the variable "veteran status" compare to the one obtained before?

2.b Explain why this bias disappears if "veteran status" dummy is instrumented with the "eligibility" dummy described above. State which are the necessary identifying assumptions and interpret them in this context.

---

# Take-Home/Assignment Questions (1/4)

Part 2: "*Lifetime Earnings and the Vietnam Era Draft Lottery: Evidence from Social Security Administrative Records", Angrist, <g>American Economic Review</g>, 1990*

2.c Describe which are the structural, reduced form and first stage equations in this setup. Analytically show that, together, they provide the 'formula' for the IV estimator and Wald Estimand, i.e.
`$$\hat{\tau_{IV}} = \frac{\hat{\rho}}{\hat{\alpha}} = \frac{\mathbb{E}\left[Y_i | D_i = 1\right] - \mathbb{E}\left[Y_i | D_i = 0\right]}{\mathbb{E}\left[T_i | D_i = 1\right] - \mathbb{E}\left[T_i | D_i = 0\right]}$$`

---

.center[
# Thank you and

# see you tomorrow!
]

---

# References

- Angelucci, M. and De Giorgi, G., 2009. Indirect effects of an aid program: how do cash transfers affect ineligibles' consumption?. *American economic review*, 99(1), pp.486-508.