Development Flashcards

1
Q

What is positive selection bias?

A

Overestimate of the treatment effects.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is head count? The head count ratio?

A

HC is the number of people below the poverty line, HC ratio is the proportion of people below the poverty line.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the PG ratio? The Income gap ratio?

A

PG: Income needed to bring all people to poverty line / income of society

IGR: Income needed to bring all people to poverty line / income needed to bring all people to poverty line if the poor had no income at all.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is negative selection bias?

A

Underestimation of treatment effects.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

When does selection bias arise?

A

It arises when individuals are selected (or sefl-selected) based on characteristics that may affect their outcome.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are three ways in which the Gauss-Markov condition of zero correlation between mean and error term can be violated?

A
  1. Measurement error (difficult to address but not too strong)
  2. Omitted variable bias
  3. Reverse Causation
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the two types of randomisation?

A

Simple and Stratified.

  • Simple: randomly drawn from uniform distribution
  • Stratified: randomisation is performed separately within each stratum. For example median age.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are two tests that are used to check if the sample is balanced?

A

Individual tests
Joint-orthogonality tests

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are the two main problems with survey attrition?

A
  • Statistical power decreases
  • Differential survey attrition/ Differential compliance
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are the two types of treatment effects?

A

ITT- intention to treat
ToT - treatment on treated or LATE

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is Collier’s view on what should be done to reduce poverty?

A
  • We should help countries of the bottom billion for compassion and self-interest. (look at Marshall plan)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What are Duflo and Banerjee views on what should be done to reduce poverty?

A
  • It is most important that we learn IF and HOW aid can end poverty. We need to understand how people make decisions. This can only be done through RCTs.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What are Angues Deaton’s views on what should be done to reduce poverty?

A
  • Criticises the use of RCTs (Randomisation does not equalise the two groups, external validity VS generalisability, scalability…). His main concern with providing aid is passiveness and corruption.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

When is a poverty trap present?

A

When the scope for growing income or wealth at a fast space is limited or those who have too little to invest, but expands dramatically for those who can invest a bit more.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the condition for the absence of a poverty trap?

A

When the potential for fast growth is high among the poor, and slows down as one gets richer.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What shape does the income function with a poverty trap have?

A

An S shape

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What did the case of the public emergency in Bogota 1997 show about human behaviour?

A

Policy can undermine and nurture cooperative behaviour. (The two policies to reduce water consumption).
We are not fully rational and also act on emotions… (not like Micro 1)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Why can stereotypes be problematic?

A

It can lead to disadvantaged groups to underestimate their abilities, and perform worse in social situations where they are reminded of their status.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What are two things you need to consider when stratifying data for RCTs?

A
  • too many strata can be risky (creates misfits)
  • no guarantee balance on non-strata variables
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What are the main takeaways of the paper by Berg and Zia?

A
  • Financial edutainment improves viewers’ financial knowledge, especially regarding debt
  • Effect is driven by emotional connection to main character
  • Short-term effect on usage of debt helpline
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is present Bias?

A

In the present we are impulsive, governed in large part by emotions and immediate desire. We have a tendency to procrastinate.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What are the two ways to make nudges effective?

A
  • Default options
  • Small immediate incentives
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What is the condition for a positive spillover effect?

A

MSB>MPB (positive externality)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What are the advantages of subsidies?

A
  • Internalizes externality
  • Learning by doing
  • Relaxes households’ budget
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

What are the disadvantages of subsidies?

A
  • Anchoring effect
  • Dependency/entitlement effect
  • Improper use
  • Expensive
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is SUTVA?What is the danger associated with it?

A

Stable unit treatment value assumption. Spillover effects and general equilibrium effects.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

What are the advantages of using fixed effects instead of control variables?

A
  • Allows more precise estimate
  • Captures effect of all area-specific char. that impact outcome variable
  • Both observed and unobserved characteristics
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

What are the main takeaways of the paper by Dupas 2014?

A
  • Subsidies increase short run adoption
  • Subsidies increase WTP for malaria nets a year later
  • No anchoring effect
  • Large, positive learning effects
  • Malaria nets may be less effective than deworming pills or water disinfectant.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

What are the two types of spillover in the bed nets paper by Dupas?

A

Indirect: lower chance of transmission
Direct: lower chance of infection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

How does education affect income?

A

Education increases human capital, which increases productivity, which increases income.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
31
Q

What is a top-down approach to improving education?

A

For example:
-Providing infrastructure
- Making education compulsory
- Prohibit young children from working

All policies which increase access/supply

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
32
Q

What is a bottom-up approach to improving education?

A

It aims to fix the failures of the top down approach. It based on the idea that demand must come first (there is no point in increasing supply if there is no demand). In order to do this benefits to education need to be made clear and increased in order for people to be convinced that education is worth it.

This by itself is not enough, needs to be combined with top-down approach.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
33
Q

What is a CCT?

A

Conditional Cash Transfer

34
Q

What type of curve do parents in developing countries see for the returns of education? Is this an accurate view?

A

An S- shape, the first years of education pay much less than later ones . Not accurate, evidence suggests that returns are more or less proportional each year.

35
Q

When do we cluster standard errors? What happens if we do not?

A

When observations with a control group may be correlated with each other more than with observations from another group (think of students from Tilburg are more similar to each other than students from UVA). The standard errors will be too small because corr. within group >0, can lead to misleadingly small p-value. However it does not impact the coefficient.

36
Q

At what level to we cluster?

A

At the level of randomization.

37
Q

What is the difference between Fixed effects and clustered standard errors?

A

FEs capture heterogeneity between different groups, Clustered standard errors are relevant when observations within the same group are not i.i.d.

38
Q

What is the risk of using normalised differences to assess the effectiveness of an intervention?

A

It inflates the effectiveness done on more homogenuos groups (lower sd)

39
Q

What are the main takeaways of the paper by Banerjee et al. (2007)

A
  • unskilled teachers and computer improve student outcomes, both due to personalised nature
  • Long term results are disappointing
  • CRZAY things happen outside of researcher’s control
40
Q

What are aspiration windows?

A

People draw their aspirations from the lives, achievements, or ideas of those around them –> poverty may cause small windows.

41
Q

What are the implications of the aspiration gap?

A

If economic betterment is a important goal, the aspiration window must be open, otherwise there is no drive for self-betterment. Yet, it should not be open too wide or the curse of frustrated aspirations will be encountered.

42
Q

What are the main takeaways from Dalton’s model of aspirations?

A
  • A rational poor will choose a lower effort and aspire lower than a rational rich
  • Poverty exacerbates the likelihood that an aspirations failure arises –> more poor people are more likely to aspire below their potential.
43
Q

What are the policy implications of Dalton’s model?

A
  • Releasing external constraints (providing material endowments)
  • Exogenously boosting aspirations (exposure to role models)
  • Help to internalise the feedback from effort, achievement, aspirations (psychotherapy)
44
Q

What is an ANCOVA regression? Why do we use it?

A

It is when we include the baseline measure of outcome variable in the regression. We do this because the outcome variable may have persistence (e.g. aspirations today only differ little from aspirations yesterday).

45
Q

What are the effects of an ANCOVA regression?

A
  • does NOT change your point estimate (B1)
  • decreases standard error –> improves precision (larger T, smaller p, more significant)
46
Q

What is a randomised saturation design?

A

It is used to measure spillovers. It varies the intensity of treatment in the different groups, while keeping a pure placebo group.

47
Q

What are the main takeaways by the paper by Bernard et al. ?

A
  • 1-hour video with role models can increase one aspirations?
  • Increased aspirations have positive short- and long-term effects (higher labour supply, usage of agricultural inputs, better educational outcomes)
  • No spillovers of aspirations (hence interventions need to reach all)
  • Research can take a long time
48
Q

What were some of the problems with state-run banks providing credit programs in developing countries?

A
  • Lax management
  • Corruption: political motives played a role
  • Costly Bureaucracy
  • LR effects
49
Q

Why are lending rates so high for the poor?

A
  • Adverse Selection
  • Ex ante Moral Hazard
  • Ex post Moral Hazard
50
Q

What is a microcredit?

A
  • Basically requesting a loan as a group
51
Q

What is assortative matching?

A

Elimination of Adverse selection through microcredits:

Safe borrowers will form groups with other safe borrowers, risky borrowers will have no choice but to form groups with other risky borrowers.
Safe groups end up paying lower effective interest rate (less defaults), they will enter the market and improve efficiency.
It essentially acts as a screening device, lowering average interest rate.

52
Q

What are the disadvantages of microcredits?

A
  • Penalise good clients
  • Limits entrepreneurship
  • Too much pressure
53
Q

What are some recent innovations to micro credits?

A
  • Progressive lending
  • Weak collateral
54
Q

What is Differential Attrition?

A

When the remaining treatment and control groups are not comparable. (the study may make it more likely for control members to drop out)

55
Q

How can you check for differential attrition?

A

Regress attrition on a dummy variable which indicates whether the person was in the control or treatment group (also include controls which might affect attrition).

56
Q

What can you do if you have differential attrition?

A
  • Sampling weights (if attrition predicted by individual characteristics) -> assign lower weights to obs. with characteristics which reduce attrition
  • Lee bounds (if prediction predicted by Treatment status) ->create a range for the treatment effects
57
Q

How can outliers be dealt with?

A
  • Trimming
  • Winsorizing
  • Quantile treatment effects
58
Q

What is winsorizing? What effect does it have?

A

Data below the 10th percentile is set to the 10th percentile, data above the 90th percentile is set to the 90th percentile. Reduces SD.

59
Q

What is quantile treatment?What is the concern with it?

A

Dividing the sample into smaller groups, and re-estimating the regression for each subgroup.
The sample per quantile gets very small.

60
Q

What are the main takeaways of the paper by Banerjee et al. (2015)

A
  • Microfinance increases investment and profitability of existing businesses
  • Little impact other than that (much less than we thought)
61
Q

what is the difference between type I error and type II ?

A

Type I is a false positive (falsely rejecting the null hypothesis), Type II is a false negative (wrongly accept the null hypothesis).

62
Q

What factors determine the power of a test?

A

Sample Size
Outcome Variance
True effect size (regression coefficient B1)

63
Q

What are the steps to calculate the expected required sample size of an experiment?

A

1) Design intervention
2) Estimate effect treatment will have
3) Compute sample size required to detect effect

64
Q

What are the main takeaways of the paper by Lowe (2021)

A

-Inter-group contact is important for friendships AND trade:
- Collaborative: positive effect
- Adversarial: negative effect
- Sports can have social and economical implications
- Predictions can be very off

65
Q

What does the empirical evidence show about the poor and entrepreneurship?

A
  • Most poor are entrepreneurs
  • The ROI in the small businesses of the poor is very high
66
Q

What are the 5 dimensions of the Big Fiver Personality test?

A
  1. Openness to experience
  2. Conscientiousness
  3. Extraversion
  4. Agreeableness
  5. Neuroticism
67
Q

What does empirical evidence show about the characteristics of successful entrepreneurs?

A

High conscientiousness, self-control, self-confidence and competitiveness. Also open to experiences.

68
Q

What explains overall returns being so low for the poor’s businesses despite very high marginal returns?

A

The small scale of the business, also technological barriers.

69
Q

What is the S-shape of entreptrenurship?

A

After a certain point, MR is not enough to expand the business (e.g. cannot pay interest on loan to make your store bigger). Therefore, businesses ten dto stay small.

70
Q

What factors/barriers differ across people which can impact their ability to scale their business?

A
  • Personality
  • Preferences
  • Beliefs
  • Biases
  • Skills and know-how
71
Q

How does the Fox and the Grapes fable relate to entrepreneurship for the poor?

A

Failure could be due to a lack of enthusiasm or aspirations rather than a lack of knowledge. This can be solved by showing the successes of other businesses, ‘providing a ladder’. Facilitating knowledge diffusion is also important.

72
Q

What is the most important finding of Dalton’s paper in Jakarta?

A

There are both informational and behavioural constraints at play. (Only handbook is not enough)

73
Q

Why would you use log transfomation?

A

To dampen outliers if you have heavily skewed data

74
Q

What are the conditions for a valid instrument?

A

Relevance and Exclusion Restriction

75
Q

What are the main takeaways of the paper by Mel et al.?

A
  • Marginal return to capital is high in Sri Lanka
  • Heterogeneous Treatment Effects
76
Q

What are institutions?

A

The rules of the game in a society or, more formally, are the humanly devised constraints that shape human interaction.

77
Q

On what three premises does AJR’s theory rest on? What IV do they result in?

A
  • There were different types of colonisation policies which created institutions: Extractive VS Neo Europes
  • The colonisation strategy was influenced by the feasibility of settlements
  • Colonial state and institutions persisted even after independence

Using settler mortality as an IV.

78
Q

What are the main conclusions of the paper by Lowes et al. (2017)

A
  • effects of institutions are very persistent
  • Culture and institutions interact
  • Findings support the Tabellini model
79
Q

What are the main concerns with AJR’s IV?

A
  • Mortality rates of settlers cold be correlated with the current disease environment, which may have a direct effect on economic performance.
  • Omitted factors
  • Settler mortality may affect economic performance but through other channels
80
Q

What are potential confounders of the paper by Lowes et al. (2017)

A

Selective Migration
Different People Surveyed
Geography

81
Q

What is the oversubscription method?

A

Basically a lottery for who gets the treatment. This is done when there are limited resources.