Power plants and structural transformation

class: center, middle, inverse, title-slide

.title[
# Power plants and structural transformation
]
.author[
### Sunghun Lim <br> <font size = "5"> LSU </font>
]
.author[
### Joshua D. Merfeld <br> <font size = "5"> KDI School and IZA </font>
]
.date[
### December 2023
]

---

/* .remark-slide table{
      width: 100%;
    } */

/* Change the background color to white for shaded rows (even rows) */

.title-slide h3:nth-of-type(2) {
      color: red;
    }

.remark-slide thead, .remark-slide tr:nth-child(n) {
      background-color: #A7A9AC;
    }

.remark-slide table {
      background-color: #A7A9AC;
    }
    
    tfoot {
      font-size: 80%;
    }

table{
      border-collapse: collapse;
      border-color: transparent;
      background-color: #A7A9AC;
    }

/* .hljs-github {
      background-image: url("logo.png");
      background-position: bottom left;
      background-size: 10%;
    }

.inverse {
      background-image: url("");
      background-position: bottom left;
      background-size: 10%;
    } */

.title-slide {
      background-image: url("logo_title.png");
      background-position: bottom left;
      background-size: 20%;
    }

</style>

## This paper
<br>
- Following long literature on the structural transformation process
<br><br>
- Previous literature has focused on things like urbanization, technological change, trade, etc. <font size = "2"> (e.g. Michaels et al., 2012; Herrendorf et al., 2015; Samaniego and Sun, 2016; Teignier, 2018; Lim, 2021; Nguyen and Lim, 2023; Bustos et al., 2020) </font>
<br><br>
- More recently, focus on importance of infrastructure <font size = "2"> (e.g. Lakshmanan, 1989; Perez- Sebastian and Steinbuks, 2017; Timilsina et al., 2020; Moneke, 2020; Raifu et al., 2021) </font>
  - Much of this focused on roads <font size = "2"> (Datta, 2012; Ghani et al., 2016; Giobbons et al., 2019; Banerjee et al., 2020) </font>

---

## This paper
<br>
- Our focus: how does power plant infrastructure affect local development patterns?
<br><br>
- Coal plants
  - Predominant form of electricity in India
  - Plants over 30MW in this paper

---

## Main findings
<br>
- Increased structural transformation after opening of coal plants
  - Driven by manufacturing, not services
<br><br>
- Increased agglomeration
  - More people, more firms, more jobs
  - More electricity (and more reliable electricity)
<br><br>
- Areas located downwind of plants have starkly different patterns
  - Less employment than other areas
  - Remaining employment concentrated in larger firms

---

## Data
<br>
- Power plants
  - Coal plants over 30MW
  - Global Energy Monitor
<br><br>
- Indian Economic Census
  - 1990, 1998, 2005, 2013
<br><br>
- Indian Population Census
  - 1991, 2001, 2011
<br><br>
- Nightlights data from VIIRS
  - 2013-2020
<br><br>
- Village shapefiles matched to censuses from SHRUG databaset <font size = "2"> (Asher et al., 2021) </font>

---

## Coal plants over time

---

## Methods
<br>
- Match the opening of coal plants to village-level census data from SHRUG
<br><br>
- Two-way fixed effects
  - Village and year fixed effects
<br><br>
- Use recent advancements in TWFE designs for consistent estimates
  - Use `did2s` package in `R`
  - Implements Butts and Gardner (2022), which calculates identical coefficients to Borusyak et al. (2021)

---

## Event study (standardized outcome)

---

## Event study (standardized outcome)

---

## Effects empirically, employment rates (standardized)

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Emp. rate </th>
   <th style="text-align:center;"> Manuf. </th>
   <th style="text-align:center;"> Services </th>
   <th style="text-align:center;"> In firm of 21+ people </th>
   <th style="text-align:center;"> In private firm </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 4cm; "> Coal plant within 50km </td>
   <td style="text-align:center;width: 2.5cm; "> -0.014 </td>
   <td style="text-align:center;width: 2.5cm; "> 0.031*** </td>
   <td style="text-align:center;width: 2.5cm; "> -0.177*** </td>
   <td style="text-align:center;width: 2.5cm; "> 0.477*** </td>
   <td style="text-align:center;width: 2.5cm; "> 0.056*** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; ">  </td>
   <td style="text-align:center;width: 2.5cm; "> (0.008) </td>
   <td style="text-align:center;width: 2.5cm; "> (0.009) </td>
   <td style="text-align:center;width: 2.5cm; "> (0.019) </td>
   <td style="text-align:center;width: 2.5cm; "> (0.049) </td>
   <td style="text-align:center;width: 2.5cm; "> (0.007) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; "> Observations </td>
   <td style="text-align:center;width: 2.5cm; "> 1,604,998 </td>
   <td style="text-align:center;width: 2.5cm; "> 1,604,998 </td>
   <td style="text-align:center;width: 2.5cm; "> 1,522,711 </td>
   <td style="text-align:center;width: 2.5cm; "> 1,522,313 </td>
   <td style="text-align:center;width: 2.5cm; "> 1,522,711 </td>
  </tr>
</tbody>
</table>

---

## Effects empirically, population census

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Power </th>
   <th style="text-align:center;"> Pop (log) </th>
   <th style="text-align:center;"> Lit. rate </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 4cm; "> Coal plant within 50km </td>
   <td style="text-align:center;width: 3.5cm; "> 0.075*** </td>
   <td style="text-align:center;width: 3.5cm; "> 0.020*** </td>
   <td style="text-align:center;width: 3.5cm; "> 0.018*** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; ">  </td>
   <td style="text-align:center;width: 3.5cm; "> (0.004) </td>
   <td style="text-align:center;width: 3.5cm; "> (0.001) </td>
   <td style="text-align:center;width: 3.5cm; "> (0.004) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; "> Observations </td>
   <td style="text-align:center;width: 3.5cm; "> 1,164,672 </td>
   <td style="text-align:center;width: 3.5cm; "> 1,434,084 </td>
   <td style="text-align:center;width: 3.5cm; "> 1,434,084 </td>
  </tr>
</tbody>
</table>

<br>

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Ag. (prop) </th>
   <th style="text-align:center;"> Ag. (log) </th>
   <th style="text-align:center;"> HH Manu. (prop) </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 4cm; "> Coal plant within 50km </td>
   <td style="text-align:center;width: 3.5cm; "> -0.085*** </td>
   <td style="text-align:center;width: 3.5cm; "> -0.050*** </td>
   <td style="text-align:center;width: 3.5cm; "> -0.048*** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; ">  </td>
   <td style="text-align:center;width: 3.5cm; "> (0.008) </td>
   <td style="text-align:center;width: 3.5cm; "> (0.004) </td>
   <td style="text-align:center;width: 3.5cm; "> (0.006) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; "> Observations </td>
   <td style="text-align:center;width: 3.5cm; "> 1,433,671 </td>
   <td style="text-align:center;width: 3.5cm; "> 1,408,509 </td>
   <td style="text-align:center;width: 3.5cm; "> 1,394,511 </td>
  </tr>
</tbody>
</table>

---

## Nightlights... more consistent power?

<br>
<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Nightlights (mean) </th>
   <th style="text-align:center;"> Nightlights (log sd) </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 5cm; "> Coal plant within 50km </td>
   <td style="text-align:center;width: 4cm; "> 0.056*** </td>
   <td style="text-align:center;width: 4cm; "> -0.013*** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 5cm; ">  </td>
   <td style="text-align:center;width: 4cm; "> (0.005) </td>
   <td style="text-align:center;width: 4cm; "> (0.002) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 5cm; "> Observations </td>
   <td style="text-align:center;width: 4cm; "> 3,042,032 </td>
   <td style="text-align:center;width: 4cm; "> 3,042,032 </td>
  </tr>
</tbody>
</table>

---

## Exposure to coal plant pollution, three examples

---

## Empirical results, economic census

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Emp. rate </th>
   <th style="text-align:center;"> Manuf. </th>
   <th style="text-align:center;"> Services </th>
   <th style="text-align:center;"> In firm of 21+ people </th>
   <th style="text-align:center;"> In private firm </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 4cm; "> Top quartile of exposure </td>
   <td style="text-align:center;width: 3cm; "> -0.060*** </td>
   <td style="text-align:center;width: 3cm; "> 0.003 </td>
   <td style="text-align:center;width: 3cm; "> -0.211*** </td>
   <td style="text-align:center;width: 3cm; "> 0.502*** </td>
   <td style="text-align:center;width: 3cm; "> 0.027** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; ">  </td>
   <td style="text-align:center;width: 3cm; "> (0.012) </td>
   <td style="text-align:center;width: 3cm; "> (0.011) </td>
   <td style="text-align:center;width: 3cm; "> (0.027) </td>
   <td style="text-align:center;width: 3cm; "> (0.067) </td>
   <td style="text-align:center;width: 3cm; "> (0.013) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; "> Observations </td>
   <td style="text-align:center;width: 3cm; "> 505,278 </td>
   <td style="text-align:center;width: 3cm; "> 505,278 </td>
   <td style="text-align:center;width: 3cm; "> 478,985 </td>
   <td style="text-align:center;width: 3cm; "> 478,985 </td>
   <td style="text-align:center;width: 3cm; "> 478,985 </td>
  </tr>
</tbody>
</table>

---

## Empirical results, population census

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:center;"> Power </th>
   <th style="text-align:center;"> Pop (log) </th>
   <th style="text-align:center;"> Lit. rate </th>
   <th style="text-align:center;"> Agriculture </th>
   <th style="text-align:center;"> Marg. Workers </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;width: 4cm; "> Top quartile of exposure </td>
   <td style="text-align:center;width: 3cm; "> -0.023*** </td>
   <td style="text-align:center;width: 3cm; "> -0.012*** </td>
   <td style="text-align:center;width: 3cm; "> -0.030*** </td>
   <td style="text-align:center;width: 3cm; "> -0.117*** </td>
   <td style="text-align:center;width: 3cm; "> 0.091*** </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; ">  </td>
   <td style="text-align:center;width: 3cm; "> (0.006) </td>
   <td style="text-align:center;width: 3cm; "> (0.002) </td>
   <td style="text-align:center;width: 3cm; "> (0.006) </td>
   <td style="text-align:center;width: 3cm; "> (0.011) </td>
   <td style="text-align:center;width: 3cm; "> (0.009) </td>
  </tr>
  <tr>
   <td style="text-align:left;width: 4cm; "> Observations </td>
   <td style="text-align:center;width: 3cm; "> 321,457 </td>
   <td style="text-align:center;width: 3cm; "> 416,498 </td>
   <td style="text-align:center;width: 3cm; "> 416,498 </td>
   <td style="text-align:center;width: 3cm; "> 416,376 </td>
   <td style="text-align:center;width: 3cm; "> 411,787 </td>
  </tr>
</tbody>
</table>

---

## Wrapping up
<br>
- Still a work in progress, but...
<br><br>
- Coal plants have large effects on local development patterns
  - Increased structural transformation, driven by increase in manufacturing
  - Possible evidence that power is more consistent
<br><br>
- Areas downwind of plants have starkly different patterns
  - Less employment than other areas around coal plants
  - Remaining employment concentrated in larger firms

---

## Wrapping up

<br>
- Didn't present today, but new results:
  - Manufacturing increases most for metal products, consistent with increase in power supply
  - Large decrease in employment related to construction, which makes sense
  - No change in employment in transport-related industries

---

class: center, middle

<font size = "40"> Thank you! </font>

[https://joshmerfeld.github.io](https://joshmerfeld.github.io) <br>
[https://github.com/JoshMerfeld](https://github.com/JoshMerfeld) <br>
Twitter: [@Josh\_Merfeld](twitter.com/Josh_Merfeld)