Exam 1

If we increase price from 0 to 20 Kenyan shillings, what is % decrease in quantity demanded?

(-0.803)(20) = 16.06 units; 16.06/41 = 39.2% reduction in quantity demanded

What do microeconomic studies estimate the impact of malaria to be?

0.6% of GDP.

Macroeconomic studies find that malaria maybe responsible for reducing growth by _____ percentage point a year.

1 percentage point The magnitude is huge because if a country is supposed to grow at 3% a year, reducing 1 percentage point means the country is only growing at 2% a year, or GDP growth rate is reduced by 33%. GDP growth rate= β*malaria rate + controls maybe we are over-estimating the effect of malaria on poverty.

Cohen and Dupas, ITN experiment in Kenya. What did they examine?

1. elasticity of demand with respect to price 2. elasticity of usage with respect to price 3. the vulnerability of the consumer and the price 4. externalities

Arguments against cost-sharing:

1. reduce program coverage by dampening demand. 2. selection effect: if people who cannot afford to pay even at a reduced price are also more likely to be sick, and therefore need the good more, then charging a positive price excludes the neediest and could significantly reduce the benefits of the partial subsidy.

Arguments for cost-sharing:

1. selection effect: a positive price could select out those who do not value it 2. psychological effect: a positive price could induce people to use it more effectively. 3. if people interpret higher prices as higher quality, then this may encourage usage.

________ of the world is relatively rich and has a high standard of living, the rest of the world is poor and the gap between rich and poor countries is growing. Though this may be realistic of the past 30-50 years, it's not true if a longer time period is considered. Why?

20% No economic development occurred before the Industrial Revolution. All countries had the majority of their populations near subsistence level. Modern economic development is a process of technological development and institutional change.

What's the reality of nutrition in China?

30-35 million school-aged children are estimated to be suffering from malnutrition 60,000 children across China are anemic 25% of children were myopic, few have glasses Huge percentage of intestinal worms in young children

What is the nature of China's human capital today? In poor rural areas?

45% of school children live in poor rural areas

Malnutrition in the first 100 days

60% of infants are seriously sick 20% are stunted/wasted 20-30% of Chinese children are in danger of becoming permanently physically and mentally handicapped.

confidence level: suppose the t-statistic is 1.96 and the associated p-value is 5%. How confident are we?

95% confident that the true beta lies within beta hat ±1.96*se.

What is the effect of the eradication program on the student enrollment rate for counties that had 20% infection rate?

= (0.20)*(0.56) = 0.112 Counties with a 20% infection rate increased the probability of student enrollment after eradication by 11.2%.

The poster children for those who advocate industrial policy are the East Asian "miracle" economies.

A famous one is how the government's chief economic planner K.Y. Lin in Taiwan decided on the basis of a report by a USA consultant J.G. White Engineering Corporation, that plastics was a suitable industry to develop (Wade, 1990, p. 80). He then identified Y.C. Wang, a local businessman, as someone with the resources to do this, apparently through bank records. He then told him to start the business! The first factory was built under government supervision and given to Wang in 1957. Wang, subsequently head of the Formosa Plastic's Group, went on to become one of the leading entrepreneurs in the country.Both Singapore and Hong Kong had long histories of mercantile capitalism. In addition, in the Korean and Taiwanese cases there was an extensive bureaucratic tradition which played a key role in adopting and monitoring the policies. The politicians had to allow them to do this, but the fact that this capacity existed clearly could have influenced the success of the policy and thus the incentive of the politicians to adopt it. Finally these countries also had a lot of human capital. These circumstances suggest that the policy of promoting industry had a good chance of succeeding, and other things equal this would encourage any government to adopt it.

Results for ITN experiment with elasticity of usage:

Among people who pay for ITNs, usage is not significantly affected by change in prices. Usage rate is about 60% for all groups. (common reasons given by the pregnant women: waiting for the old one to wear out, waiting for the birth of the child to use the ITN).

What is "our estimate is significantly different from 0 at x%"?

An easy way to think about it is: our estimate is different from 0 1-x% of the time, i.e. our estimate is wrong (therefore equal to 0) only x% of the time. We want t-statistic to be big, and p-value to be small. So our estimate is wrong only x% of the time.

Vulnerability of the consumer and price:

Are pregnant women who have lower hemoglobin levels willing to pay higher prices for ITNs

Why is such a large portion of today's manufactured goods produced in China?

Because wages were so low in the 1980s and 1990s. Prior to this, most things were made in South Korea, Taiwan, Hong Kong, Singapore.

Using the enrollment dataset, suppose you would like to estimate the following equation using ordinary least squares (OLS) regression method: y_ijt=α*Post+δ*Infection Rate+ β*Post*Infection Rate+error where subscript i indicates student i, j indicates county j, and t indicates year t, and y is the enrollment status of student i in county j at time t. Suppose the estimated β ̂=0.56. How do we interpret this number? What is the effect of the eradication program on the student enrollment rate for counties that had 100% infection rate?

Beta is how much enrollment is increasing as a result of the eradication program. If you are a student in the infection county, your probability of being enrolled increases by 56% after the eradication program.

Elasticity of usage:

Check usage a few months later at their homes.

The stories of SK and Mexico provide the backdrop for what's happening in China today. While low wages and labor intensive manufacturing fueled economic growth in the 80s and 90s -- China is entering a new age. What are the implications?

China continues to grow so there is a rising demand for labor. However, labor supply is falling. Wages are rising and industrial structures are changing. Can they move up the productivity ladder?

Education is important. What challenges does rural China face in the near future?

China needs to be able to educate its workers to perform the higher-paying, more technologically advanced job its growing economy demands. Rural Chinese schools have trouble educating students, in large part due to poor nutrition and health. Rural areas experience high rates of anemia, vision impairment, and intestinal worms in young children. Malnutrition and lack of access to vitamins is thought to seriously impair cognitive function in malnourished infants, impairments which can handicap them for life. Rural China needs access to better nutrition and technology in order to supply educated students to the work force.

Chlorine water treatment (3 groups):

Control group: free distribution to your door Treatment group 1: voucher Treatment group 2: 50% of market price for Chlorine.

Is correlation identical to causal relationship? Why do we want to determine if there is a causal relationship? What are the challenges in indentifying a causal relationship?

Correlation is not causation. Economists want to determine causal relationships so that policies that are successful in one country or sector may be implemented in another, or so that detrimental economic policies or actions can be avoided. Causal relationships are difficult to identify because there could be outside factors contributing to outcomes other than the policy of interest. Causal relationships are also hard to determine empirically, as it may be hard establish a control.

Cost-benefit analysis of free ITN distribution:

Cost: 40 Ksh vs. 0 for each ITN. Ignore other management costs under free distribution. Benefit: child mortality rate reduced, positive externality. Overall: free distribution is better in most scenarios.

Using the enrollment dataset, in sheet 1 of the excel file, what is the average enrollment rate before and after the disease eradication program takes place for counties with 100% infection rate vs. counties with 0% infection rate. How would you use this information to get the DID estimate of the effect of the disease eradication program on enrollment rate? (Y1) Average enrollment rate before policy for county with 100% infection rate = 44.4% (Y3) Average enrollment rate after policy for county with 100% infection rate = 100% (Y2) Average enrollment rate before policy for county with 0% infection rate = 100% (Y4) Average enrollment rate after policy for county with 100% infection rate = 100%

DID estimate = (Y3-Y1) - (Y4-Y2) = (100-44.4) - (100-100) = 55.6% increase in the probability of enrollment in diseased county after the eradication.

What is the difference-in-differences (DID) method? What is it used for? Write down the equation for the DID method. Why do we take the first difference? Why do take the second difference?

DID method can be used to estimate the impact of a policy on a variable in question by comparing before and after policy results for a treatment group and a control group. The DID method allows a researcher to see the difference in outcomes for the two groups while controlling for variables that have nothing to do with the policy change. Equation for the DID method: (treatment group after policy - treatment group before policy) - (control group after policy - control group before policy) We take the first difference to see the impact of the policy on the treatment group and the second difference to see the impact of the policy on the control group.

We may be under-estimating the effect of malaria on poverty due to negative externalities - additional cost of malaria in a community/society.

Discourage trade and foreign investment. Demographic structure is affected, since malaria mostly affects young children under 5. High fertility and high mortality also discourage investment in women's human capital, reduce women's productivity, labor force participation etc. Human capital of children is affected by malaria by its effect on school attendance and performance. Other chronic diseases as a consequence of malaria, blood transfusion and HIV.

What is economic development (or the lack thereof?) How to measure it? What can we learn from the recent "success stories" in development?

Economic development occurs when countries experience economic growth (ongoing GDP/ GNP growth) accompanied by an increase in the standard of living. Increases in standard of living are typically measured by increases in literacy and education, increase in life expectancy and decrease in infant mortality, better health outcomes, etc. The recent economic development "success stories" in East Asia can lend some insight into what policies can be used to foster economic development in low-income countries. It is often hard to determine casual relationships between economic policy and resulting development, but there are some commonalities across these countries that may help to assess policy effects. Some general takeaways are that governments might use industrial policies to develop some sectors of the economy, try to increase exports and develop technology, encourage cooperation between government and business, and increase employment and redistribute income to create more equality.

Why randomly select people into different groups?

Ensures on average, people in different groups have the same observable characteristics (such as age, weight, height, gender), and unobservable characteristics (such as ability, intelligence, perseverance, etc), so there is no longer selection bias. Then you can compare the results.

Elasticity of demand with respect to price:

Estimate quantity demanded at these clinics at each price level. Price ranged from 0 to 40 Ksh ($0.60). 40 Ksh is at 90% subsidy already and is what most malaria-endemic African countries use.

Why is it hard to determine a causal relationship?

Example: The government of "East Asian miracle" countries subsidized certain industries for export and growth. Does subsidizing an industry (call this A) necessarily lead (cause) growth (call this B)? Not really. Maybe those countries chose to subsidize industries that were already growing fast. This is the so called selection bias. So maybe it's because those industries were growing fast, and this caused the government to subsidize them, which then led to even faster growth. i.e. B->A->B. This is why other countries that tried to subsidize certain industries saw zero to no growth (i.e. they couldn't replicate the success stories of East Asia).

What are the types of market failures?

Externalities, public goods or missing markets, information failure, adverse selection, moral hazard

Growth comes from factor accumulations and productivity growth. What are factor accumulations and what constitutes productivity growth?

Factor accumulations are increase in the capital stock and labor force Productivity growth is technological change, improvement in efficiency such as specialization and the use of assembly lines.

P-value

For each t-statistic value, there is an associated P-value. For example, when t-statistic=1.96, the associated p-value is 0.05. This means our beta hat is significantly different from 0 at 5%. For higher t-statistic, the p-value will be smaller, meaning our beta hat is significantly different from 0 at a smaller %.

What happened when low-wage factories moved out?

Foreign direct investment (FDI) declined, unemployment, and gang and cartel violence rose.

Joseph Ferrie used the Georgia's Cherokee Land Lottery of 1832 to analyze how a sudden influx of wealth affects families over time. Why does he not use the data from lottery winners today? What is the benefit of using the data from Georgia's Cherokee Land Lottery? What are his main findings?

He didn't use the data from lottery winners today because people with certain characteristics are more likely to participate in the lottery (buy lottery tickets). If you have taken a course in probabilities, you should know that the expected pay off from a lottery is less than the cost of buying a lottery ticket, therefore a rational person should not purchase lottery tickets. Lottery buyers today are more likely to be have a relatively low level of income and education. Looking at the outcomes of winners of lottery today and compare to those of non-winners, if the lottery winners do not have an increase in wealth 20 years from today, we don't if it's because the lottery has no effect or because the winners are simply different from the pool of non-winners due to their relatively low level of education. The Georgia's Cherokee Land Lottery on the other hand, has universal participation, i.e. any male resident of Georgia was given a chance to win the land lottery, and the chance of winning was ¼, so we have a considerable size of the treatment group and the control group. Because participation is universal, there should be no underlying differences between the control (non-winners) and the treatment (winners) group. Comparing the outcomes of the treatment vs. control group allows us to draw a causal conclusion on the effect of winning a lottery. He found that winning a lottery on average did not increase a household's wealth 20 years later, winners did not spend more on their children's education, winners did not have more children, and only the lottery winners that had been wealthy before winning the lottery experienced an increase in wealth. This suggests that maybe the poor and the average households simply did not know how to handle an increase in wealth and therefore wasted it and did not see an increase in wealth years later. The results of this study suggest that giving money to the poor probably cannot get them out of the poverty trap, because the poor may not know how to invest the money.

Results of the Ozier study:

He finds that finishing secondary schools increases the vocabulary test scores, decreases the probability of low-skill self-employment, and decreases the probability of teen pregnancy for females

1789, Samuel Slater, known as the "Father of the American Industrial Revolution":

He memorized the designs of textile factory machinery and brought the technology to America.

What is the impact of vitamins on students?

High hemoglobin, low anemia, higher math test scores

Pricing and Coverage

If price is high, coverage will be low. If you want to have maximum coverage, you may have to set price to 0. Thus, the paper by Cohen and Dupas on free distribution or cost sharing.

What role does nutrition play in developing human capital?

If students are sick and malnourished, they won't be able to learn.

Why do we do regression discontinuity analysis?

If there exists a cutoff, then people who are far above and below the cutoff may be very different. On the other hand, people who are around the cutoff should be very similar, and it's almost like it's a random assignment of people above and below the cutoff. Example: cutoff score is 60. People whose test scores around the cutoff have similar abilities. It's almost just dumb luck (like guessing a multiple choice question right, which is like random assignment) that landed some people barely above the cutoff and some barely below the cutoff. So it's better to compare the outcomes of people who are just above the cutoff to the outcomes of people who are just below the cutoff. There is a way to choose the optimal bandwidth around the cutoff based on some test statistics.

In Cohen and Dupas (2010), in column (3) Table III, the estimated effect of price in Kenyan shillings on the # of ITNs sold is -0.803 and is significant at 1% level. Based on this estimate, how do we reach the conclusion that raising the price from 0 to 40 Kenyan shillings reduces demand by 78%?

If you increase the price of the ITN from 0-40 Ksh, the demand for ITNs decreases by (0.803)(40) = 32 units. If the clinic is distributing 41 free ITNs per week, then 32/41 = 78% reduction in demand due to the price increase

Anne Case and Hanushek and Woesmann discussed the role of education on economic growth and income. If we run the following regression:Income = β*education+other controls, and find the estimated β to be positive, what are the problems with interpreting this result as higher education leading to (causing) higher income? What should be done instead to infer a causal relationship?

If β is estimated to be positive, all we can say is that we know education is positively correlated with income. It's possible that education is causing higher income, or higher income is causing people to get more education, or some third factor, like a person's ability, that causes both higher income and higher education. To determine a causal relationship, studies in the past have used identical twins and examine the effect of education on income. The idea behind using data limited to identical twins is that because identical twins have the same DNA, therefore they should have the same ability. If they have different education and different income, then we can say the differences in income are caused by the differences in education.

Given limited resources, how should countries deploy resources to increase education attainment? Is it more important to reduce class sizes, or to increase teachers' salaries? Do student outcomes respond more to free textbooks or to enhanced teacher incentives?

In the past, in the developed countries, researchers find a non-negative relationship between class size and students performance. Smaller class size does not increase students performance.

How did South Korea make this transformation?

Labor force was highly educated. Almost everyone in rural and urban SK graduated from high school.

During the same time frame, Mexico was a candidate for developing success. What happened?

Manufacturing was growing, wages were rising, and low-wage factories shut down and moved. The hope was that employers would invest in high-productivity jobs that would support rising wage rates. But Mexico's school systems did not succeed in educating a large share of the work force to support the developing economy.

Potential problems with Push programs:

Moral hazard - grant recipients may divert effort to get more grants, instead of focusing on research. Adverse selection - researchers have more information than funders about probability of their researching leading to successful products. Therefore, funders may not be able to determine which research projects are worth pursuing. Governments may make funding decisions partly based on political, rather than scientific considerations. There maybe political pressure to allocate funds to particular regions or countries. Tax credits also face similar programs. Firms have incentive to relabel their R&D as eligible for the targeted credits.

Three ways to conduct economic analysis:

Natural experiment, Randomized Controlled Experiments (RCTs), or field experiments, Regression analysis

What are other ways to determine causal relationships?

Natural experiment, regression analysis.

Results for ITN experiment on the psychological effect?

No psychological effect, i.e. no price effect on usage rates.

Are more vulnerable people willing to pay a higher price for ITN?

No. Women who pay a higher price do not appear to be sicker than the average prenatal clients in the area. In fact, women who receive free nets are healthier than the average prenatal women. Reason: women who receive free nets are more likely to be coming for a repeat prenatal visit. The free ITNs induced women to come to prenatal visits, therefore they are more likely to be healthier.

Is an increase in GDP per capita equivalent to development?

Not necessarily -- example: Qatar. Very high GDP per capita, but in terms of education and life expectancy rates they don't rank as highly. Development occurs when income increases along with other standards of living (reduced mortality, lower illiteracy, increase in education, increase in life expectancy)

What is economic development?

Occurs when the standard of living of a large majority of the population rises, including income and other dimensions like health and literacy.

Time inconsistency problem:

Once firms have sunk their R&D costs, governments can set low prices for firms' products, then firms have little incentive to invest in R&D in the beginning. Could potentially be overcome through reputation formation.

What are the outcome variables for secondary schooling in Kenya?

Outcome variables are vocabulary test scores (as an indicator of human capital), low-skill self-employment, teen pregnancy for females.

Ozier (2011) studies the impact of secondary schooling in Kenya.

People who score above the cutoff are more likely to be admitted into secondary school. Ozier wants to compare the outcomes of people who finish secondary school to the outcomes of people who don't go to secondary school.

Why is the development of vaccines for poor countries a market failure?

Poor countries cannot afford to pay for vaccines, therefore pharmaceutical companies are not willing to invest in developing vaccines for those diseases. Lack of intellectual property protection may also prevent firms from investing in R&D in developing a vaccine.

Why does the real source of the problem begin before high school?

Poor quality of education in grades 1-9 and poor nutrition.

What are the benefits and costs of using Randomized Controlled Trials (RCTs) or field experiments to determine a causal relationship.

Randomized controlled trials randomly assign study participants into experimental and control groups. The expected difference between the experimental and control group is what is attributable to the variable being studied. This allows researchers to test for causal relationships because all other variables have been controlled for (to the best of their extent). Uses large numbers in the study to try and ensure that all participants are as similar in all attributes as possible. Drawbacks to these studies are that they can be costly to conduct.

Randomized Controlled Experiments (RCTs), or field experiments:

Randomly select people into control and treatment groups, and evaluate their outcomes. Law of large numbers ensure that people in each group, on average, are the same in every observable characteristics (age, height, weight, income, etc), and unobservable characteristics (ability, IQ, perseverance, discipline, effort).

How much of a drop in quantity demand is this?

Since the average number of ITNs in clinics distributing them for free is 41 per week, this reduces quantity demanded to 9 per week when price is 40. This is 32/41=78%≈80% drop in quantity demanded.

The second set of analysis in Bleakly (2003): Outcome variable: adult earnings. Treatment intensity: death rates for malaria (between 0 and 1, denoted by Trj), for hookworm the authors used a proxy for the infection rate at the state level, the number of individuals per capita treated by the RSC. The length of treatment: depending on the year a person is born, the years of exposure to the eradication campaign can be computed, denoted by Expik. Hookworm campaign started in 1913, and was targeted at children 16 or younger, so if one is born 17 or more years before 1913, one has 0 zero year of exposure to the eradication campaign. If one is born say in 1910, one has 16 - (1913-1910) = 13 years of exposure to the eradication campaign. If one is born after 1913, one has 16 years of exposure to the eradication campaign.

So we are 95% confident that the true effect of eradication campaign lies within 0.09±1.96*0.023=[0.044, 0.135]. As you can see, 0 is outside of this confidence interval, so we are 95% confident that the eradication campaign has a non-zero and positive impact on school attendance.

Do the authors recommend free distribution or cost-sharing in the case of ITNs? and Why?

The authors recommend free distribution because an increase in price of ITNs leads to a significant reduction in demand and usage doesn't decrease significantly when people receive the nets for free. There are hugely positive externalities in the decrease of infant and child mortality, indicating that benefits outweigh the costs of free ITN distribution.

Some limitations of pull programs:

The effectiveness of a pull program depends on its credibility and its design. Purchase commitment has to be credible. The design should specify the eligibility conditions.

Results for ITN experiment on elasticity of demand with respect to price:

The estimated coefficient for ITN price in Ksh is -0.803. Interpretation: increase price from 0 to 40 Ksh, the quantity demanded is reduced by 40*(-.803)=-32.

Failure story of industrial policies:

The footwear factory ... would have linked the Meat factory in the North through transportation of the hides to the South (for a distance of over 500 miles) to a tannery (now abandoned); the leather was to have been backhauled to the Footwear factory in Kumasi, in the centre of the country and about 200 miles north of the tannery. Since the major footwear market is in the Accra metropolitan area, the shoes would then have to be transported an additional 200 miles back to the South.""Projects were begun without feasibility studies and without competitive tendering. New enterprises were distributed among party functionaries as private fiefs, enabling them to give patronage to relatives, friends, and supporters."

However, Lazear (1999) pointed out, with a simple model, that class size is chosen, and not some randomly determined thing.

The idea is less disruptive students are put into large classes, and less disruptive students also tend to have better grades. So it's not that large class size causes better grades, but rather selection of better students into larger class causes betters grades. However, class size is a choice variable chosen by schools to maximize students performance. i.e. less disruptive students are put in bigger class rooms. This is why the regression (1) gives rise to a non-negative β. However, this does not mean bigger classes CAUSE students to perform better. Random assignment of students into classes of different size shows that smaller classes increase students performance. However, not enough experiments of this type are done.

What is the malaria gap? Why does it exist?

The malaria gap is the difference in results conducted on a macroeconomic and a microeconomic on what the true impact of malaria is on the GDP of a nation. Some macroeconomic studies find that in countries where malaria is endemic, GDP growth may be reduced by as much as 1% a year. Macroeconomic studies may overestimate the impact of malaria on GDP for two reasons: reverse causality and spurious regression. Reverse causality makes it difficult to determine if poverty is caused by malaria, or if the opposite is true. Spurious regression causes problems for macroeconomic studies because other factors that are difficult to control for may be at play. Microeconomic studies may underestimate the impact of malaria on GDP because of the difficulty in accounting for externalities associated with the disease. Ultimately, researchers get different results using different methods.

Why do we want to use the regression discontinuity analysis? How does this help us find a causal relationship? What problems can this approach avoid?

The regression discontinuity analysis is based on the idea of comparing the treatment group and the control group around some margin or cutoff. In the case of the Kenyan study, where they want to look at the effect of secondary education on vocabulary scores and teen pregnancy, they used regression discontinuity analysis because admission into secondary school depends on some cutoff test score. For people who are just above and just below the cutoff score, they should be very similar in their abilities, whereas people who are way above the cutoff may be very different from people who are way below the cutoff. For people who are close to the cutoff, it's almost random assignment that put some above and some below the cutoff (maybe just dumb luck guessing a question right), so they are not different in their abilities. Comparing the people who are just above the cutoff who get to go to secondary school (the treatment group) to the people who are just below the cutoff who don't get to go to secondary school (the control group) allows us to determine the true effect of going to secondary school.

What is the problem if we simply compare children's enrollment in counties with Ij=1 to children's enrollment in counties with Ij=0 after the eradication of the disease? i.e. if we only look at (Yijt | Ij=1, Post=1) - (Yijt | Ij=0, Post=1) Therefore, we need to check if there are any differences in children's enrollment before the eradication of disease. i.e. (Yijt | Ij=1, Post=0) - (Yijt | Ij=0, Post=0)

Thus, the word difference in difference, because we look at the difference in children's enrollment after the treatment, but also taking into account the difference in children's enrollment before the treatment, to truly identify the effect of the treatment on children's enrollment. i.e. [(Yijt | Ij=1, Post=1) - (Yijt | Ij=0, Post=1)] - [(Yijt | Ij=1, Post=0) - (Yijt | Ij=0, Post=0)]

Why do we need the two-stage price design? Why not just look at the usage rate for women paying 0 to 40 Ksh?

To separate out the selection effect and psychological effect. If we just look at usage rate for women who pay 0 vs. women who pay positive amount, we don't know how much of the difference in usage rate is driven by selection, and how much by psychological effect.

What is Kuznets curve? What phenomenon is it trying to explain? How does the use of time series data enable us to find out more about the Kuznets curve? What have economists found using time series data?

U-shaped curve. GDP per capita v. inequality. Kuznets curve plots inequality on the vertical axis and gdp per captia on the horizontal axis, it's inverted-U shaped curve, indicating as an economy grows, its inequality will first rise, then fall. At first, Kuznets curve was constructed using a cross-section of different countries, so each point on the diagram above represents a different country. This was due to data limitations. Using time series data, we can track each country over time, and when we plot the inequality and the GDP per capita for each country over time, we no longer find this inverted-U shaped curve. i.e. for some countries, as they grow, they are able to keep their inequality low. For other countries, they are poor and not growing, but their inequality is still very high.

South Africa, Case and Deaton (1999) use variation in school quality (class sizes) between districts in South Africa to estimate the impact of school quality on children's progress through school.

Under apartheid, blacks were severely limited in their residential choice and thus their choice of schools. Black parents were forced to send their children to black schools, whose funding decisions were made by White controlled entities. Some districts had 20 students per teacher, but others 80 students per teacher. They find strong and significant effect of pupil-teacher ratios on enrollment, educational achievement, and on test scores.

What are some ways to address the educational issues in China?

Vitamins, eyeglasses, de-worming, early childhood education, computer assisted learning, counseling programs, making high school free

What transformation took place in South Korea in the 80s and 90s?

Wages began to grow rapidly, went from a low-wage, labor-intensive economy to a high-productivity, service-based, and innovative based economy.

What is economic growth?

When there is sustained (ongoing for 2-3 years) increase in a country's output (as measured by GDP or GNP) or in the per capita output (GDP or GNP per person)

Adverse selection occurs when a. when buyers and sellers have asymmetric information. b. one party gets involved in a risky event knowing that the other party will bear the cost. c. there is negative externality. d. the market is missing.

a

Economic growth occurs when there is a sustained a. increase in GDP per capita. b. increase in standards of living. c. increase in the share of the population living in rural areas. d. increase in adult literacy rate.

a

In Bleakly (2003), the author used the Difference in Difference method to estimate the effect of the eradication of disease on school attendance by regressing: Yijt= β*Ij*Post + α*Post+ δ*Ij + other controls, where Yijt is individual i's school attendance (0 if individual i is not attending school at the time, and 1 if individual i is attending school at the time) in county j at time t, Ij is the infection rate in county j, after the eradication of disease, Post=1, and before the eradication of disease, Post=0. Suppose the estimated β=0.09 and is significant at 1% level. What is the effect of the eradication of disease on school attendance in counties with infection rate 50% vs. counties with infection rate 100%? a. school attendance is increased by 4.5 percentage points in counties with an infection rate of 100% compared to counties with an infection rate of 50%. b. school attendance is increased by 9.0 percentage points in counties with an infection rate of 100% compared to counties with an infection rate of 50%. c. school attendance is decreased by 4.5 percentage points in counties with an infection rate of 100% compared to counties with an infection rate of 50%. d. school attendance is decreased by 9.0 percentage points in counties with an infection rate of 100% compared to counties with an infection rate of 50%.

a

Kuznets speculated on the relationship between growth and in equality, arguing that inequality rose when: a. GDP per capita rose. b. everyone remained in the agriculture sector. c. a country opened up to trade. d. GDP per capita fell.

a

The absence rate in teachers and health workers in developing countries is higher in a. teachers and health workers in higher positions. b. female workers. c. richer regions. d. concentrated on a small number of "ghost" workers.

a

To break the poverty trap, Sachs in his article "Breaking the Poverty Trap" recommended a. investments on raising agriculture productivity. b. increase in government spending to increase employment. c. reduce corruption. d. less government interventions.

a

Moral hazard:

a situation in which one party gets involved in a risky event knowing that it is protected against the risk and the other party will incur the cost. Example: investment banks, fund managers.

Market failure:

a situation in which the allocation of goods and services is not efficient. That is, there exists another conceivable outcome where an individual may be made better-off without making someone else worse-off. (The outcome is not Pareto optimal.) Government intervention or regulation is called for in times of market failures.

Pareto efficiency:

a state of allocation of resources in which it is impossible to make any one individual better off without making at least one individual worse off.

Microeconomic studies:

aggregate cost per case, find a smaller impact, generally less than 1% of GDP per capita a year. This is done by calculating the Cost of Illness = private medical costs+ non-private medical costs + foregone income + pain and suffering. Alternatively, the cost of illness can be calculated using willingness-to-pay (WTP) approach. How much you would like to pay to avoid this illness. Then multiply the cost by the number of cases of infection, then divide it by GDP.

Economic analysis at Micro level:

at individual or at household level.

Economists think that economic growth depends fundamentally on all of the following EXCEPT: a. factor accumulation. b. inequality. c. efficiency. d. technological change.

b

Economists think that economic growth depends fundamentally on all of the following EXCEPT: a. factor accumulation. b. inequality. c. efficiency. d. technological change.

b

Estimating the following regression: Y=a+b*X, where a is a constant, X is a control variable, and Y is the outcome variable. If b is estimated to be 0.5, then we can interpret this as: a. one unit increase in X causes Y to increase by 0.5 units. b. one unit increase in X is associated with 0.5 units increase in Y. c. one unit increase in X causes Y to decrease by 0.5 units. d. one unit increase in X is associated with 0.5 units decrease in Y.

b

If we run the following regression: Student's grades=β*class size+ other controls. Studies in the past have found β to be 0 or slightly positive. Can we say bigger class size leads to better student's grades? a. Yes. b. No. c. Uncertain.

b

Market failure is a situation in which the allocation of goods and services a. is pareto efficient. b. allows one individual to be better off without making someone else worse off. c. makes it impossible to make one individual better off without making someone worse off. d. is pareto optimal.

b

The malaria gap exists because: a. the cost of illness approach does not include foregone income when estimating the cost of a disease. b. the positive externalities associated with the eradication of tropical diseases cannot be estimated using the cost of illness approach. c. the willingness to pay approach is not accurate. d. the cost of illness approach does not include non-private medical costs when estimating the cost of a disease.

b

What are the effective programs on reducing absence rate in developing countries? a. higher teacher salaries for better test performance by students. b. girls' merit scholarships and contract teachers. c. installing time-stamping machines. d. none of the above.

b

GDP divided by total population provides a measure of: a. the wellness index. b. the gap between the rich and poor. c. per capita income. d. the poverty index.

c

In the case of public goods, markets fail to form because a. the good is too cheap. b. the good is too expensive. c. individual consumer is not willing to pay for the good on her own. d. the market is pareto efficient.

c

Using difference-in-difference method, suppose we find that the average income for each group is the following: Before the intervention After the intervention Treatment group 45,000 55,000 Control group 50,000 55,000 We can then conclude that the intervention leads to: a. an increase in income of $10,000. b. no increase in income. c. an increase in income of $5,000. d. a decrease in income of $5,000.

c

Regression analysis:

can be used on natural experiment and RCTs used to test hypotheses, derived from economic theory, against real-world data.

A Push program for subsidizing research is associated with following difficulties: a. requires specifying output in advance. b. may not work well to encourage basic research that may not have an immediate use. c. purchase commitment is not credible. d. moral hazard of grant recipients.

d

Bleakley's study on the eradication of disease in the American South is important because: a. the eradication of disease leads to increase in school enrollment and earnings. b. the American South a century ago was substantially poorer, much like many tropical African countries today. c. we can interpret the results as causal relationships. d. all of the above.

d

Economic development occurs when a. adult illiteracy rate falls. b. infant mortality rate falls. c. GDP per capita rises. d. all of the above.

d

In Cohen and Dupas (2010), the estimated effect of price in Kenyan shillings on the # of ITNs sold is -0.803 and is significant at 1% level. Based on this estimate, raising the price from 0 to 20 Kenyan shillings: a. reduces quantity demanded by 0.8 ITNs. b. reduces quantity demanded by 8 ITNs c. reduces quantity demanded by 1.6 ITNs. d. reduces quantity demanded by 16 ITNs.

d

In addition to the private benefit of education, such as increased in one's future income, the following is(are) example(s) of positive externality of education: a. reduced teen pregnancy rate. b. reduced crime rate. c. technological innovation. d. all of the above.

d

Underdevelopment of vaccines for diseases in developing countries is a result of: a. market failures. b. lack of intellectual property protection. c. time inconsistency problem. d. all of the above.

d

What have we learned from Cohen and Dupas study on free distribution and cost sharing using evidence from a randomized malaria prevention experiment? a. people are less likely to use the product if they receive it for free. b. people's quantity demanded of the product is not significantly affected by the product's price. b. people who pay higher prices for the product are sicker than the average buyer. d. none of the above.

d

Which of the following is not a market failure: a. negative externalities. b. moral hazards. c. public goods. d. business cycle.

d

Growth accounting. A country's labor force grows 1.2 percent and its capital stock grows 3 percent. Assume that labor share of output is 0.25 and capital share of output is 0.75. If the economy grows 4.55 percent, how much does total factor productivity grow? a. 0.5 percent b. 1 percent c. 1.5 percent d. 2 percent

d 0.0455 = (0.75)(0.03) + (0.25)(0.012) + tech => 0.0455 = 0.0255 + a, a = 0.02

Natural experiment:

evaluate the effect of an historical event to answer relevant questions today.

How do you measure the growth rate, for example, between 2014 and 2015?

growth = GDP 2015 - GDP 2014 / GDP 2014

What is the equation for growth?

gy = (wk*gk) + (wl*gl) + a wk = share of national income generated by capital wl = share of national income generated by labor gk = growth rate of capital gl = growth rate of labor a = technology

t-statistics measures

how close our estimated β given the data we have is to the true β if it's >1.96, we claim the our β (hat) is significantly different from 0 at 5%. Therefore we can reject the null hypothesis that β = 0.

Standard error (se) measures

how close the number you get is to the true number. For example, if the standard error of the estimated β is small, then we know the estimate we get is precise (close to the true β). What is close, and how close are we to the true β?

Standard deviation (sd) measures

how dispersed (spread out) the data points are.

Maimonides' rule:

lass size is allowed to rise until there are 40 students in a class. When the 41st student enters, the class is split into two classes, one of 20, and one of 21 students. Angrist and Lavy (1999) evaluate the effect on students' test scores of being just below the cutoff (40 students), and of being just above the cutoff(41 students split into two classes), and find that students with smaller classes score significantly higher on tests. They find reducing class size by 10 is associated with 0.25 standard deviation increase in fifth graders' test scores.

Reverse causality:

lower growth rate (and therefore poor) countries cannot afford malaria interventions, therefore this causes more cases of malaria. It's hard to separate how much malaria is caused by poverty, and how much poverty is caused by malaria.

Negative externality:

markets may fail to control the production and sale of goods like cigarettes, alcohol, and pollution goods, which have less merit than consumers perceive.

Public goods or missing markets:

markets may fail to form, resulting in a failure to meet a need or want, such as the need for public goods, such as defense, street lighting, and highways.

Positive externality:

markets may fail to produce enough merit goods, such as education and healthcare

Information failure:

markets may not provide enough information because, during a market transaction, it may not be in the interests of one party to provide full information to the other party.

Impact of computer assisted learning:

much higher test scores

Results of chlorine water treatment experiment:

no difference in usage rate for control and voucher group. Voucher group can greatly reduce waste.

Spurious regression:

other factors such as climate may drive both poverty and malaria. Not including these factors in the regression may lead us to wrongfully conclude the extent of the effect of malaria on poverty.

Bleakly (2003) paper: County j has a value of Ij between 0 and 1, with 1 indicating 100% of infection rate, and 0 indicating 0% of infection rate. You can think of Ij as treatment intensity. Post =1 indicates after the eradication of the disease (the treatment), Post=0 is before. Yijt= β*Ij*Post + α*Post+δ*Ij + other controls Again, β is the DiD method of estimating the treatment effect. The interpretation of β depends on the value of Ij, example: if Ij=50%=0.5, then if the estimate β= 0.09, this implies that:

relative to a county with 0% infection rate, children's enrollment on average will increase by 0.09*0.5=0.045 or 4.5 percentage points in county with 50% infection rate.

Potential problem with macroeconomic regression to estimate the impact of malaria:

reverse causality and spurious regression

Pull programs:

reward research outputs. othing is paid unless a viable vaccine is developed.

Push programs:

subsidize research inputs.

Use the regression analysis in Excel to estimate β ̂, is β ̂ significantly different from 0 at 5%?

t-statistic = 2.55 > 1.96 = yes, statistically significant outcome at 5%.

To determine the psychological effect:

two-stage pricing design. Among clinics with a positive price, women who already decided to buy the ITN were given a chance to win a lottery for an additional discount, from 0 to the posted price. Among these women, any change in usage with the actual price paid can be interpreted as psychological effect.

Economic analysis at Macro level:

using national aggregates, such as GDP, unemployment rate, etc.

Adverse selection:

when buyers and sellers have asymmetric information (access to different information); the "bad" customers are more likely to apply for the service. Example: health insurance.

The estimated coefficient β identifies the effect of the treatment, why?

β is the DiD method of estimating the treatment effect. If the estimated β=0.09, we can say that in a region with Ij=1, or 100% infection rate, the eradication of the disease will on average increase the children's enrollment by 0.09*100%=0.09, or 9 percentage points.