Land I: Incentives and Effort

This subtitle section introduces key motivations for studying land markets and land reform, which correspond directly to several foundational themes in the “Slides_06_Land.pdf”. Here's a breakdown and explanation, section by section, matched to the content of the slides:

1. Why Study Land Markets?

"Agriculture is still the biggest industry in the world..."

This aligns with Slide 2–3:

Slide 2 discusses the relationship between farm size and productivity, noting the stylized fact that small farms are often more productive per acre than large ones.
Slide 3 addresses the surprising nature of this observation in light of increasing returns to scale from mechanization, better capital access, and more.

2. Policy Relevance of Land Redistribution

"Repeated discussion of intervention... political pressure to redistribute land..."

Matches Slides 16–18:

Slide 16 raises the question: What are the implications of redistributing land?
Slide 17–18 introduce tenancy reforms (e.g., Operation Barga in India) and their empirical motivations: increasing both equity and efficiency through redistribution or tenancy security.

3. Importance of Property Rights and Insecurity

"Land is a traditionally held property and often insecure..."

Corresponds to Slides 29–30:

These slides focus on property rights as central to development economics, particularly regarding expropriation risk, tenure insecurity, and squatting (as mentioned by the speaker).
Examples from Peru and urban squatters (later covered in Field 2007, Slide 33–34) support this point, where insecure rights discourage investment or labor supply.

4. Empirical Evidence of Size-Productivity Inversion

"Striking table from Brazil, Pakistan, Malaysia... large farms produce much less per acre..."

Directly related to Slide 2:

This slide cites the empirical finding that small farms outperform large farms in output per acre, with stark contrasts like small farms in Brazil producing 5x more per hectare.
This is the motivating puzzle: Why do small farms seem more efficient, despite the theoretical advantages of scale?

5. Equity and Efficiency Trade-off — or Not

"If causal, then redistribution is a win-win..."

Discussed in Slides 4–6:

Slide 4 lays out potential reasons for this paradox: agency problems (hired vs. family labor), land quality, and selection effects.
Slide 5–6 present studies (Binswanger & Rosenzweig; Shaban) showing that owner-cultivated plots are more productive, suggesting agency/incentive issues matter and giving plausibility to the causal interpretation.

6. Causality Question and Land Quality Confound

"Is this causal or just a land quality issue?"

This is crucial in Slides 4–6 again:

The lecture recognizes that land quality may confound the observed relationship.
Hence the importance of empirical strategies (e.g., fixed effects, controlling for plot quality) to isolate the true effect of ownership or size on productivity.

Summary:

This introductory segment of the video is a conceptual overview that:

Motivates the importance of studying land and agriculture.
Frames the equity-efficiency debate in land redistribution.
Introduces the size-productivity paradox.
Emphasizes the role of property rights and tenure security in development.
Previews empirical strategies needed to move from correlation to causation — a recurring theme throughout the lecture.

You're ready to share the next subtitle segment whenever you're ready.

This subtitle passage corresponds very closely to Slides 2–4 and 5–6 in the lecture PDF. Here's how the discussion breaks down and what it means:

🔹 1. Adjusting for Land Quality with Profit-to-Wealth Ratio

“So what this does is it looks at the profit to wealth ratio...”

This is directly discussed in Slide 2:

Rather than just looking at output per acre, which could be biased due to land quality, the lecture proposes using profit-to-wealth ratio (return per dollar invested).
Wealth = land value + capital, so this method indirectly controls for land quality by assuming better land costs more.

📊 Finding: Poorer (smaller) farms earn a much higher return per dollar invested than richer (larger) ones — about 3× more in some comparisons.

🔹 2. Risk and Poor Farmers

“Poorer people do particularly worse when risk is high...”

This aligns with the right side of Slide 2:

The horizontal axis is weather variability (e.g., monsoon onset variability), a proxy for risk.
Poorer farmers are more vulnerable to weather uncertainty due to:
- Lack of savings or access to insurance
- Greater sensitivity to income shocks
- Inability to smooth consumption during bad years

💡 Even though they earn more per dollar on average, their returns are more volatile, reflecting the risk–return trade-off.

🔹 3. Why Big Farms “Should” Be More Productive

“Technology of farming has fixed costs... can’t even turn a tractor...”

These are the theoretical arguments for increasing returns (see Slide 3):

Fixed costs of technology (tractors, combines)
Better capital access for large/rich farmers
Political power for input subsidies (e.g., from Bates' research in Africa)
Higher education and tech-savviness of wealthier landowners

🧠 But despite these advantages, large farms are less productive per unit of land.

🔹 4. The Key Missing Factor: Incentives and Monitoring

“Anybody who owns a big farm has to hire labor... harder to monitor...”

This explanation is found in Slides 4–6:

Large farms rely on hired labor, which often has lower effort incentives.
Agency problem: Employers can’t perfectly monitor worker effort.
In contrast, small farms are owner-cultivated — no incentive issues.
Even wealthy landowners may under-cultivate if managing labor is too costly or time-consuming.

✅ Conclusion / Takeaways:

Profit-to-wealth ratio shows small farms are more efficient even when controlling for land value.
Risk matters more for the poor, making their high returns more volatile.
Large farms have theoretical advantages, but incentive problems (monitoring labor) can offset them.
Owner cultivation is a big reason why small farms often outperform large ones.

This passage reinforces that simple comparisons of farm size and output are misleading unless we account for land quality, risk, capital access, and labor incentives — which the PDF and lecture try to carefully disentangle.

This subtitle section corresponds closely to Slides 5–6 of the lecture PDF, which present the empirical strategy used by Binswanger and Rosenzweig, and later Shaban (1987), to estimate the effect of land ownership on productivity — while addressing two major challenges: selection bias and land quality.

🔍 Main Idea: Does owning the land make a farmer more productive?

The goal is to answer: Is land ownership (vs. renting) causally related to productivity, or are we just observing correlation driven by unobserved factors?

✅ Problem 1: Selection Bias (Who becomes an owner?)

“People who become owners might be different... maybe they are just better businessmen...”

This is the concern that ownership isn’t randomly assigned.
More motivated, skilled, or risk-tolerant individuals may become landowners, so their higher productivity might reflect who they are, not whether they own the land.

📊 Represented in the equation as:
ηᵢ = unobserved fixed characteristics of farmer i

✅ Problem 2: Land Quality

“Maybe the owners hold onto the best land and sell all the bad land…”

If owners are cultivating higher-quality plots, then their higher productivity may be due to land characteristics, not ownership.

🧪 Empirical Strategy: Controlling for Individual Fixed Effects

“So what Binswanger and Rosenzweig point out is that... some farmers cultivate both owned and rented plots...”

This is the key insight used in Slide 5:

Some farmers operate both their own land and rented land.
By comparing productivity across these plots for the same person, we control for ηᵢ, the farmer’s fixed characteristics.

Equation from Slide 5:
Πᵢⱼ = α + βRᵢⱼ + ηᵢ + νᵢⱼ
Where:

Πᵢⱼ = profit for farmer i on plot j
Rᵢⱼ = 1 if rented, 0 if owned
ηᵢ = farmer fixed effect (hardworking, risk tolerance)
νᵢⱼ = other random factors (e.g. good neighbor, luck)

To eliminate ηᵢ, they run a within-farmer regression:

Πᵢ₂ − Πᵢ₁ = β(Rᵢ₂ − Rᵢ₁) + (νᵢ₂ − νᵢ₁)

This lets them compare productivity on owned vs. rented plots within the same individual, removing selection bias.

✅ Key Result (from Slide 6 and the lecture):

“People work harder on their own land.”

Shaban (1987), building on this, also controls for land quality.
He finds:
- Farmers work ~40% more on their own land.
- Productivity is 15–30% higher on owned land than rented, even after controlling for plot quality.

✅ Additional Clarification on Contract Types:

“I just put a 0/1 here... but contracts vary (fixed wage, sharecropping, etc.)...”

This is an important caveat:

The Rᵢⱼ = 1 (non-owned land) category includes many types of contracts (e.g. wage labor, fixed rent, sharecropping).
These all offer different incentive levels.
This will be analyzed more deeply later when the lecture turns to sharecropping models (Slides 7–14).

✅ Summary Takeaways from This Section:

Selection bias and land quality confound comparisons between owner-cultivated and rented land.
The within-farmer comparison strategy (Binswanger & Rosenzweig) helps isolate the causal effect of ownership.
Results show: farmers work harder and achieve higher productivity on their own land, supporting the agency/incentives hypothesis.
Not all rented land arrangements are equal — contracts vary in how much incentive they provide, which will be explored next.

This subtitle passage dives deeper into the empirical identification strategy used to test whether ownership causes higher productivity — and it connects directly to Slides 5–6 of the lecture.

🔍 What's being done here?

The goal:

Is the difference in productivity between owned and rented land caused by ownership itself, or by something else (like land quality or farmer ability)?

✅ Step 1: Difference within the same farmer

“I look at the difference between what you do on your two plots of land... eta i has vanished because I took the difference.”

This is a within-individual comparison:

Each farmer i cultivates two types of land: owned and rented.
By computing the difference in outcomes (e.g., profits, inputs used) within the same farmer, you cancel out their fixed characteristics (ηᵢ).
This helps isolate the causal effect of ownership (β), since all individual-level factors are held constant.

Π_{i1} - Π_{i2} = β(R_{i1} - R_{i2}) + (ν_{i1} - ν_{i2})

This eliminates selection bias due to farmer-level traits.

✅ Step 2: Control for Land Quality

“So it could very well be that land quality, the own land is better... and that would be misleading...”

The speaker highlights a further potential bias:

Even within the same farmer, they may reserve better land for themselves and rent out the worse plots.
So, land quality needs to be controlled for to isolate the true effect of ownership.

➡️ That’s what the Shaban (1987) study does: uses data from farmers who cultivate both types and also includes objective land quality measures (soil, irrigation, slope, etc.).

✅ Step 3: Regression and Results

“You have several hundred farmers who each do both... I take the difference between their outcomes on owned vs. rented land...”

The regression is run on differences, not levels:

Each observation is a farmer-level difference in outcomes (like profit, labor input, output).
This helps average out and test whether, on average, farmers:
- Work more on their own land
- Produce more on their own land

🟢 Result (as summarized in the table):

40% more labor used on owned land
40–50% more output on owned land
Suggests a strong ownership effect

🔁 Why this works:

Problem	How it’s solved
Selection bias (ηᵢ)	Within-farmer comparison eliminates it
Land quality differences	Controls for land characteristics

✅ Conclusion:

“People, if you own your own land, you put in more effort on it. And that’s very clear.”

This confirms the agency effect:

Ownership incentivizes effort.
Productivity increases are not just because better people or better land are involved, but because of ownership itself.

This section introduces the economic model of incentives under limited observability — a foundational concept in contract theory and development economics — and maps directly to Slides 7–10 in the lecture PDF.

Let’s walk through what’s going on, piece by piece.

🔑 Main Question:

Why do tenants put in less effort on land they don’t own?

This is not just about laziness — it’s a result of incomplete contracts and unobservable effort, which create imperfect incentives.

✅ Structure of the Incentives Model (See Slides 7–8)

🎭 The Setup:

There is a landowner who owns 2 plots, but can only farm 1 himself.
He hires a tenant to work the other.
The tenant chooses effort level: $e$

❗ Key Assumption:

Effort is unobservable to the landowner.
- The landlord cannot verify or contract on effort, only on output.
- That’s the core of the agency problem.

📉 Effort and Cost:

The cost of effort is increasing: $\text{Cost}(e) = ce^2$ (quadratic form).
As effort rises, the marginal cost rises too — this is standard in microeconomic modeling.

🎯 Output is Probabilistic:

The model simplifies real-world uncertainty into two outcomes:
- High output = $H$
- Low output = $0$
Probability of success depends on effort: $P(\text{success}) = e$
(Think of $e$ as normalized between 0 and 1)
So the expected output = e × H

💸 Contracting Options (See Slides 9–10)

Since effort can’t be observed, the landlord must offer a contract based on output:

\text{Payment to tenant} = \begin{cases} h & \text{if output = } H \\ l & \text{if output = } 0 \end{cases}

Different contracts correspond to different values of $h$ and $l$ :

Contract Type	$h$	$l$	Description
Wage contract	= $l$	= $h$	Fixed wage, independent of output
Rent contract	$h = H - R$ , $l = -R$	Tenant pays landlord a fixed amount $R$ , keeps all output
Sharecropping	e.g., $h = \frac{H}{2}, l = 0$	Tenant keeps a share of the output

📌 Outside Option: $w$

The tenant has alternatives, so he must receive at least utility $w$ to accept the contract.
This introduces the participation constraint.

🧠 Why Sharecropping Exists Despite Being Inefficient?

This is a central puzzle the model aims to address:

A rent contract gives perfect incentives (tenant keeps all output) but puts risk entirely on the tenant.
A wage contract removes all incentives, but guarantees income.

🟰 Sharecropping is in between — it shares output and risk.

It balances:
- Incentives (keep part of what you earn)
- Insurance (don't bear the full risk)

This helps explain why sharecropping is common in real-world settings with:

Limited observability of effort
Risk-averse tenants
Poor monitoring technologies

✅ Summary Takeaways:

Effort is unobservable, so landlords can’t contract on it directly.
Different contracts offer different combinations of:
- Incentives (more if you keep more output)
- Risk exposure (higher if income depends entirely on output)
Sharecropping emerges as a compromise: some incentive, some insurance.
The model simplifies reality, but is powerful for understanding why owner-cultivated land often sees higher effort (since the owner keeps all the output and bears the risk directly).

This section explains the core logic of optimal contract design under unobservable effort — a key insight in development and labor economics — and connects directly to Slides 9–11 in the PDF. Here's a breakdown:

🧩 Problem Being Solved: Incentivizing Effort When Effort Can’t Be Observed

Landlord’s Goal: Maximize the expected output (and his own income) by getting the tenant to exert optimal effort, even though he cannot monitor it directly.

✅ Step 1: Benchmark — What Would the Owner Do Himself?

"If I work the land myself, what’s my optimal effort?"

Expected benefit: $e \cdot H$
Cost: $\frac{1}{2} \cdot c \cdot e^2$

Maximizing utility:

\max_e \left[ e \cdot H - \frac{1}{2} c e^2 \right]

Take the derivative:

H - c e = 0 \Rightarrow e^* = \frac{H}{c}

📌 So, the optimal effort level (if the landlord worked the land himself) is:

e^* = \frac{H}{c}

✅ Step 2: Getting the Tenant to Choose the Same Effort

"Now imagine I hire a tenant and pay him depending on whether output is high or low..."

If output = H, the tenant gets $h$
If output = 0, the tenant gets $l$

Tenant’s expected utility:

e(h - l) + l - \frac{1}{2} c e^2

Take derivative with respect to effort $e$ :

(h - l) - c e = 0 \Rightarrow e = \frac{h - l}{c}

So, to get the tenant to work at effort level $e = \frac{H}{c}$ (same as landlord), we need:

\frac{h - l}{c} = \frac{H}{c} \Rightarrow h - l = H

🪙 Residual Claimant Solution (Rent Contract)

“This is called making the tenant a residual claimant...”

Set:
- $h = H - R$
- $l = -R$

➡️ The tenant keeps all the output, but pays a fixed rent $R$ to the landlord.

This gives him:

Strong incentives: Every extra unit of output benefits him directly.
Same effort level as the landlord would choose himself.

✅ Outcome: He behaves as if he owns the land — because he bears both the risk and reward.

⚠️ Caveat: Is This Contract Feasible?

"The question is, can the tenant afford to pay R up front?"

In theory, rent contracts provide perfect incentives.
In practice, many tenants lack liquidity or access to credit, so they can’t afford to pay the fixed rent upfront.
This constraint is what makes sharecropping or wage contracts more common in the real world — even though they’re less efficient.

✅ Summary Takeaways:

Contract Type	Incentive Strength	Risk to Tenant	Practical Issue
Rent (residual claimant)	⭐⭐⭐⭐ (perfect)	High	May be unaffordable
Sharecropping	⭐⭐ (moderate)	Shared	More feasible
Wage	❌ (none)	None	Feasible, but inefficient

So, the model shows that perfect incentives are theoretically possible — but real-world constraints like credit access limit their use.

Enlighten Mind

Categories