Credit II Understanding Credit Constraints

This opening segment of Lecture 2: Credit Part 2 picks up from the earlier theoretical discussion and sets the stage for the empirical investigation into the nature of credit constraints for the poor.

📌 Section Summary: Framing the Supply vs. Demand Constraints in Credit Markets

🔁 Recap of the Previous Model

Last time, we saw that:
- Poor borrowers can only be lent very small amounts (due to incentive constraints).
- Fixed costs of monitoring make these small loans very expensive per dollar lent.
- That leads to very high interest rates, even in competitive, no-default settings.

🧠 Example:
If it costs $1 to monitor and you only lend $1, then you must charge at least 100% interest to break even.

❓ New Question: What Kind of Constraint Is This?

Banerjee now sets up the core empirical puzzle for Lecture 2:

Are poor people credit-constrained because they are unwilling to borrow at high rates?
→ Demand-side constraint
(“I don’t want to borrow at 110%.”)
Or are they being restricted by lenders who fear they might default at such high rates?
→ Supply-side constraint
(“I can only lend a small amount, or I fear I won’t get paid back.”)

🔍 Why This Distinction Matters

If it's demand-side, then poor people don’t want the credit being offered (perhaps because it’s too expensive).
If it's supply-side, then poor people want more credit, but can’t get it (even at high interest rates).

This helps determine what kind of interventions are effective:

Lowering interest rates? (if it's a demand constraint)
Reducing enforcement/monitoring costs? (if it's a supply constraint)

🎯 Goal of Lecture 2

To empirically investigate:

Are borrowers willing to borrow at high rates?
Do small business owners have high enough returns to justify borrowing, even at 80–100% interest?
Are loan supply constraints—not just rates—the main issue?

This leads into the RCT studies in Sri Lanka and South Africa, which provide direct evidence on borrower behavior and credit returns.

Let me know when you're ready to continue with the next video segment, or slide, and I’ll summarize/analyze it.

This segment deepens the core empirical and theoretical investigation of Lecture 2: Are the poor credit-constrained by demand-side reluctance or supply-side restrictions? Banerjee frames this with more precision, bringing in key concepts from information economics.

📌 Section Summary: Understanding Credit Constraints — Demand vs. Supply

🔄 Main Question

Why aren’t poor people borrowing more?

Is it because they don’t want to borrow at high interest rates? → Demand-side constraint
Or because lenders are unwilling to lend more, even if borrowers are willing to pay? → Supply-side constraint

🔍 Three Supply-Side Mechanisms for Lender Reluctance

Moral Hazard
- Higher interest rates → borrowers more tempted to default
- Lenders expect this and restrict loan size to reduce the risk
Default Aversion
- Even without bad intentions, higher rates can cause repayment to fail mechanically
  
  E.g., paying back $220 on a $100 loan is just harder than paying $150
- Lenders don’t want to deal with enforcement (costly, messy), so they lend smaller amounts that borrowers can safely repay
Adverse Selection
- As interest rates rise, repayers drop out, and risky (or dishonest) borrowers stay in
- Those with low cost of default (or no intent to repay) are unaffected by interest rates
  
  They’ll borrow at 1,000% if they never planned to repay
- This deteriorates the pool of borrowers, so lenders hold back lending

✅ These mechanisms explain why even when poor people are willing to pay high interest, lenders still lend small amounts.

📉 Demand-Side Logic

Banerjee also reviews the standard borrower logic under diminishing returns to capital:

If interest rate is high, borrower stops borrowing when marginal product of capital = interest rate
So, even if credit is available, high price = less demand

🔁 So both sides (borrowers and lenders) can rationally reduce borrowing at high interest rates — but we need data to know which is the main constraint.

🎯 What Comes Next

Banerjee is preparing us to interpret empirical evidence:

Are small loan sizes caused by borrower reluctance (high price)?
Or lender fear (default, moral hazard, adverse selection)?

This leads into randomized experiments and survey-based studies that test credit constraints in the field.

Let me know when you're ready to move on to the next part — it likely involves the Sri Lanka RCT or return to capital data.

This segment provides crucial empirical evidence on how poor entrepreneurs behave when they receive access to capital — and challenges the idea that high interest rates alone deter borrowing. Instead, it suggests that returns to capital are often higher than the cost of borrowing, and yet people still don't borrow much, implying credit constraints are real, but not demand-side.

📌 Section Summary: High Returns and Borrowing Behavior – Evidence from Sri Lanka

💸 Fact 1: Poor People Often Borrow at Extremely High Rates

Banerjee discusses:

Fruit vendors in Chennai pay:
- 5% interest per day for informal credit (i.e., paying ₹1,000 at end of day for ₹950 worth of fruit)
- This is ~4 million percent annualized
More common informal loans:
- 3–4% per month = ~60% per year (still enormous)
And yet — people borrow regularly at these rates

🔍 This rules out the idea that borrowers are unwilling to pay high rates. They're clearly willing and used to it — even when it's not for emergencies.

📈 Fact 2: The Sri Lanka RCT on Returns to Capital

Study by de Mel, McKenzie, and Woodruff
Sample: 405 small businesses, with < $1,000 in capital
Randomly assigned grants of $0, $100, or $200
Measured profit effects over time

💡 Key Findings:

$100 grant → ~6% increase in monthly profits
$200 grant → No additional profit compared to $100
- Suggests diminishing returns
No grant → No change in profits

✅ Interpretation: 6% monthly return on capital = 72% annual return
So borrowing at 4% per month (~60% annually) would still be profitable

🧠 Key Insight

These business owners can afford to borrow at high rates and still make a profit.

So why don’t they borrow more?

It's not because the interest rate is too high
This points to a supply-side constraint — lenders aren’t willing to lend, even when profitable
Or possibly, other frictions (e.g., risk aversion, behavioral factors, fear of debt)

⚠️ A Curious Puzzle

People who got $200 didn’t earn more than those who got $100

This suggests non-linear returns — maybe businesses can’t use large capital effectively
Or lack of complementary inputs (skills, labor, demand)
Possibly psychological factors (we’ll revisit this later)

✅ Conclusion

These business owners could handle high interest rates given their high returns
So they are not demand-constrained in the standard economic sense
Instead, they don’t have access to credit, or face other constraints

This strongly supports a supply-side credit constraint narrative.

Let me know when you're ready for the next segment — this likely leads to further exploration of why people don’t borrow even when it's profitable (e.g., risk aversion, trust, marketing, or behavioral explanations).

This extended segment reinforces the central insight of Lecture 2: Poor entrepreneurs often operate businesses with very high returns to capital—higher than the cost of borrowing—and yet do not borrow much. This paradox suggests that the main issue is not unwillingness to borrow (demand), but limited access to loans (supply) or deeper behavioral or structural barriers.

📌 Section Summary: High Borrowing Costs and High Returns – So Why Aren’t People Borrowing More?

🛒 1. Chennai Fruit Vendor Case (Karlan & Mullainathan)

Informal vendors pay:
- ₹950 in goods (cash price)
- ₹1,000 on credit (same-day repayment)
That’s 5% interest in less than a day
- Annualized: ~4 million percent
These are routine loans, not emergencies
- Vendors borrow like this daily, for years
- ⛔️ So it’s not desperation borrowing

✅ Takeaway: Poor vendors willingly borrow at extremely high rates for day-to-day business

📊 2. The Sri Lanka RCT (Capital Drop Experiment)

405 micro-businesses (<$1,000 capital)
Random assignment:
- Group 1: $200 grant
- Group 2: $100 grant
- Group 3: $0 (control)
$100 grant → ~6% increase in monthly profits
- i.e., 72% annual return
Even 4% monthly interest (60% annual) is affordable for these firms

🔄 Borrowing at 60% is profitable if you’re earning 72% return

🤔 3. Surprising Result: More Capital ≠ More Return

$200 grant did not lead to more profit than $100
- In fact, numerically slightly less
Possible explanations:
- Diminishing marginal returns
- Lack of complementary inputs (skills, labor, demand)
- Behavioral limits (e.g., underinvestment, misuse, risk aversion)
Still: both $100 and $200 grants significantly improved profits over the control group

✅ Interpretation: Access to small capital works, but there are limits to scale.

🧠 Key Economic Insight

If borrowers are earning more than the cost of borrowing, and yet not borrowing, then the issue is not price — it’s access, structure, or psychology.

Banerjee frames this as evidence that demand constraints are not binding — people could pay the rates, and borrowing would still be profitable.

📉 Policy Implication

Microcredit, subsidies, or savings interventions might be justified — but the problem is not simply high interest rates. Rather, the barriers are in access, enforcement, scale, or trust.

Let me know when you’re ready to continue. The next section likely moves toward explanations for why, despite high returns, people may still not borrow (including behavioral explanations or framing effects).

This segment presents strong empirical evidence that poor borrowers are not demand-constrained—they're supply-constrained. It brings in a second key study from South Africa, where interest rates were randomly varied, and finds minimal impact on loan take-up, even over a very wide range of rates.

📌 Section Summary: The South Africa Interest Rate RCT — Low Elasticity of Loan Demand

🔬 Study Design (Karlan & Zinman)

Sample: 55,000 former clients in South Africa
Randomized:
- Interest rates: from ~2% to ~12% per month
- Loan maturities
Objective: See how interest rate changes affect loan take-up

📉 Key Finding: Very Low Demand Elasticity

Despite a 10 percentage point increase in monthly interest rate, loan demand fell by only ~4%
Borrowers showed little sensitivity to price
Even 12% per month didn't deter most borrowers
People seem willing to borrow at any rate offered

🧠 Interpretation: Borrowers are not making decisions based on interest rate, at least not within this high range

🔁 What Does This Suggest?

If borrowers don't respond to price, they may not be constrained by price (demand)…
Instead, they may be supply constrained:
- Banks/lenders limit the amount they offer, regardless of demand
- Borrowers would take more even at high rates — but can't

🧪 Connection to Sri Lanka Study

Recall: Microentrepreneurs earned 6% per month returns
Interest rate was ~4% per month
→ Borrowing at 4% would still be profitable
Yet they didn’t borrow — they only invested more when given cash

✅ This is very strong evidence of supply constraint:

If return > cost of capital, they should borrow until those equalize
That they don't → market isn’t letting them

📊 Economic Framework Revisited

In standard theory:
Borrow until marginal product of capital = interest rate
Here:
Marginal product > interest rate → should keep borrowing
But they don’t → they’re blocked from borrowing more

🧠 Final Insight

Low elasticity of loan demand + high returns to capital
→ Suggests: Borrowers would like to borrow more, but are not allowed to
→ A clear case of supply-side constraints

Let me know when you’re ready to continue — likely the next section covers alternative explanations (e.g., behavioral barriers, psychological framing, or institutional frictions).

This segment continues the theme of why people don’t borrow more, even when loans are profitable, and introduces two major new insights:

Borrowers strongly care about repayment duration, more than interest rates.
Borrowing behavior can be irrational or influenced by irrelevant factors, which challenges the assumption that borrowers make purely economic (rational) decisions.

📌 Section Summary: Loan Duration, Misunderstanding Interest, and Behavioral Biases

🕒 1. Loan Duration Matters More than Interest Rate

Extending loan duration by 1 month → 15% increase in loan take-up
Even though borrowers still pay interest, they value flexibility
Interpretation: Borrowers want loans that match their repayment capacity
- E.g., investments take time to yield returns
- A 3-month loan may feel too short, while 4+ months is acceptable

✅ Conclusion: People want credit, but on terms that work for them

📉 2. Demand is Not Responding to Price (Again)

Interest rate changes (e.g., 2% to 12%) → minimal change in demand
Suggests borrowers either don’t understand interest rates, or don’t care
Challenges standard assumption of rational, price-sensitive borrowers

🧠 3. Behavioral Limitation: Misunderstanding Interest Rates

Banerjee introduces a key caveat:

All this assumes borrowers understand interest rates

Often, interest rates are quoted misleadingly:
- "15%" might mean 15% flat, not annualized or compounded
- Loans with weekly repayments are effectively much more expensive
Borrowers may not grasp the true cost of credit, especially when repayments are front-loaded
Even Muhammad Yunus (Grameen Bank founder) has criticized non-transparent lending practices

🧪 4. Evidence from Framing Effects: South Africa Brochure Study

People were mailed loan offers with randomized interest rates AND pictures
One group got:
- Picture of a man
Another got:
- Picture of a woman

🎯 Result:

Just changing the picture to a woman increased loan take-up by 0.4%
That’s more than reducing the interest rate by 1 percentage point per month

(Which is a massive discount at 12% baseline rates)

✅ Implication: Borrowers are influenced by irrelevant visual cues

This defies profit-maximizing behavior
Borrowing decisions might reflect:
- Solidarity
- Attractiveness
- Trust
- Emotional associations

🧩 Final Insight

The assumption that borrowers rationally weigh interest rates and returns may be flawed.

Borrowers may:
- Misunderstand financial terms
- Be influenced by non-economic factors
- Respond more to loan terms (like duration) or framing than to cost

This shifts the policy discussion toward:

Better financial education
Simplified, transparent loan products
Understanding the psychology of borrowing

Let me know when you're ready for the next section — the lecture may now shift toward policy implications, savings behavior, or alternative interventions like commitment savings or framing.

This section of the lecture explains an ingenious experimental design by Karlan and Zinman to separate two key credit market frictions:

🎯 Main Goal: Distinguish Between Adverse Selection and Moral Hazard

✅ Definitions Recap:

Adverse Selection: The type of borrower who chooses to take a loan differs depending on the interest rate offered.
→ High interest rates attract less trustworthy borrowers.
Moral Hazard: Once a borrower has the loan, they may choose not to repay, especially if the cost of repayment is high.
→ Incentives change after receiving the loan.

🧪 Experimental Design: Triple Randomization

Randomized Offer Rate
- Different people are sent letters with random interest rates (e.g., 8%, 12%).
- This affects who chooses to apply for the loan → helps identify adverse selection.
Randomized Contract Rate (Lucky Day Surprise)
- When applicants show up, they are told:
  - "Today is your lucky day. You get a lower contract rate than what was in the letter!"
- Separates actual interest paid (contract rate) from the advertised offer.
- If two people pay the same, but applied under different offer rates, then any difference in repayment behavior is due to selection.
Randomized Repayment Incentive ("Carrot" for Repayment)
- Some borrowers were told:
  - If you repay, you can renew the loan at the same low rate.
- Others were told:
  - You will not get a preferential renewal rate.
- This isolates whether borrowers respond to incentives, i.e., tests moral hazard vs. pure inability to repay.

📊 Key Identification Strategies:

📌 Adverse Selection:

Compare borrowers who pay the same contract rate, but who applied at different offer rates.

Same repayment terms → different type of person.
Difference in default? → Points to selection (some types self-select under low/high interest offers).

📌 Moral Hazard:

Compare borrowers who applied under the same offer rate, but are randomly assigned different contract rates.

Same borrower type → different repayment behavior?
Suggests behavior changes due to incentives = moral hazard.

📌 Incentive Effects (Carrot Test):

Compare borrowers with/without promise of renewal at same low rate.
If repayment rises in the "renewal" group → this is evidence of strategic default, not just inability to pay.

🧠 Final Point:

Banerjee emphasizes that this design is powerful because it:

Controls who applies (selection)
Controls what they pay (contract rate)
Tests incentive-driven repayment behavior
Separates mechanical default (can’t pay) from strategic default (won’t pay)

Let me know when you’re ready to move on — the next part likely presents findings from this experiment and how much of default is due to selection vs. behavior.

This section of the lecture focuses on two key ideas: the findings from a clever credit market experiment in South Africa, and a broader institutional challenge in lending — particularly moral hazard from the lender's side.

1. Findings from the South Africa Experiment

Professor Banerjee discusses an experiment by Karlan and Zinman which randomized both:

the interest rate offered (what borrowers were told before applying), and
the actual contract rate (what they paid upon receiving the loan).

This dual randomization allowed the researchers to isolate:

Adverse selection: whether the type of borrower changes when interest rates change.
Moral hazard: whether behavior changes based on the actual rate paid.

Key findings:

No evidence of adverse selection: Changing the offer rate had no effect on default rates.
Some evidence of moral hazard: When borrowers were told they would be eligible for loan renewal at the same rate (i.e., a longer-term lending relationship), default rates dropped significantly. This implies people respond to incentives and care about long-term access to credit.

This aligns with earlier evidence that borrowers are sensitive to loan maturity but not very sensitive to interest rates, reinforcing the idea of supply constraints: borrowers want more credit but are not getting it.

2. Institutional Moral Hazard and Lending Constraints

Banerjee then pivots to an institutional explanation for limited credit supply. He describes a different form of moral hazard — one that affects loan officers, not borrowers. Examples include:

A loan officer accepting a bribe to approve a risky loan.
Favoritism toward friends or relatives.

Bank response: To prevent such corruption or discretion-based errors, banks impose rigid, rule-based lending systems:

Credit scoring models decide loan limits.
Hierarchy: Larger loans require approval at higher levels of management.

Consequences:

These rules reduce flexibility. Even if a borrower seems promising (e.g., talented but poor), a loan officer may be unable to approve a larger loan.
Talent or business potential — which is difficult to quantify — is often ignored, while observable traits like income or collateral dominate decisions.

Conclusion:

There’s substantial evidence that credit constraints are supply-driven.
But even beyond standard economic models (e.g., moral hazard/adverse selection from the borrower's side), real-world institutional constraints and bureaucratic risk-aversion play a major role in excluding potentially successful borrowers, especially among the poor.

Enlighten Mind

Categories