AI summary by TwinMind

Lecture 18: Land Markets (Part 3) / Public Finance (part 1)

MIT OpenCourseWare

1h 21m 10s

Land Titling Conclusion

  • Final point on land titling: collateral effect less strong than expected
    • Original motivation: land titles as collateral for borrowing
    • Research shows investment effects through time horizon channel, not collateral channel
    • Legal system imperfections can make land seizure difficult, potentially backfiring

Farm Size and Agricultural Economics

  • Discussion of optimal farm size: small vs. large farms
    • Small farms optimal due to owner-operator incentives (low moral hazard)
    • Large farms benefit from economies of scale in technology (tractors, combines, pumps)
  • Data shows correlation between GDP per capita and farm size across countries
    • Richer countries have larger farms due to capital-labor substitution
    • US example: small farms have higher value-added per acre, large farms higher value-added per worker
  • Foster and Rosenzweig (2017) find U-shaped productivity curve
    • Small farms work well with intensive owner-operator farming
    • Large farms efficient with industrial equipment and technology
  • Technology exhibits returns to scale (declining cost per unit of output with size)
  • Equipment adoption higher on larger farms (tractors, sprayers)

Public Finance in Development Context

  • Key difference from developed countries: poor information environment and low state capacity
  • Challenges of redistribution and taxation look different in developing contexts
  • Substantial inequality exists within developing countries
    • Indonesia example: 10th percentile $1/day, 98th percentile $5/day per capita
  • Government programs massively larger scale than NGO programs due to tax funding capacity

Social Safety Net Programs Scale

  • Cash transfer programs in 40+ sub-Saharan African countries
  • Conditional cash transfers in 64 countries (2014)
  • Large program examples:
    • Brazil CCT: 49 million people
    • Mexico CCT: 26 million people
    • China Dibao: 75 million people
    • Indonesia cash transfer: 61 million people
    • India MGNREGA: 58 million people (now 100+ million)

Targeting Challenges in Development

  • Core problem: government doesn't observe people's income
    • Unlike developed countries with income tax systems
    • Most developing countries exempt 80-85th percentile from income tax
    • No third-party reporting for informal sector workers
  • Three approaches to address targeting problem:
    1. Change prices of particular goods (subsidies)
    2. Universal programs
    3. Direct targeting mechanisms

Price Subsidies Approach

  • Easier for government to intervene at market level than individual level
  • Optimal to subsidize inferior goods for targeting
  • Political problems: people don't like subsidizing low-quality goods
  • Countries often subsidize normal goods instead (inefficient redistribution)
  • Price subsidies create consumption distortions
  • Still prevalent: half of countries spend more on subsidies than targeted programs

Universal Basic Income in Development Context

  • Requires two infrastructure components:
    1. Unique identifier system (deduplicated IDs with authentication)
    2. Cash distribution mechanism reaching rural/remote areas
  • Countries investing heavily in national ID systems (India Aadhaar, Pakistan, others)
  • Government payment systems can catalyze broader financial sector development
  • UBI more expensive in developing countries due to limited tax base
    • Same net transfer given much higher up income distribution
    • 80% of people outside tax net means limited ability to tax back UBI

Targeting Mechanisms

  • Proxy Means Testing (PMT): most common approach
    • Government collects observable asset data (house type, roof, floor, TV, land)
    • Runs regression using survey data to predict income from assets
    • Applies coefficients to census data to get predicted income (y-hat)
    • Targets based on predicted rather than actual income
  • Challenges with PMT:
    • R-squared much less than 1 creates inclusion/exclusion errors
    • Formula kept secret to prevent gaming
    • Colombia example shows manipulation over time when formula leaked
    • Creates horizontal equity problems (similar households treated differently)

Community-Based Targeting Research Design

  • Randomized trial in 600 Indonesian villages comparing three methods:
    1. Proxy means test
    2. Community ranking
    3. Hybrid approach
  • Community ranking process:
    • String method with cards in random order
    • Complete rank ordering from most to least well-off
    • Piloting crucial: bin method failed due to strategic behavior
  • Cross-randomized design testing additional factors:
    • Elite vs. broad participation
    • Meeting timing (afternoon=more women, evening=more men)
    • Random ordering effects
  • Experimental design allows learning multiple dimensions from same sample size

TwinMind

Get summaries like this for your own meetings

TwinMind takes notes in the background during meetings, lectures, and calls, then writes the summary for you.

Free plan to get started

No bot joins your calls

Phone, desktop & Chrome apps

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome