Question

1. An article entitled “The Income Elasticity of Health Care Spending in Developing and Developed Countries” by Farag et al, 2012 finds the following results (from Table 2 of the paper on page 153):

Explanatory Variables	Model 1: OLS (1)	Model 2: Fixed Effects
GDP per capita Voice and Accountability Government Effectiveness GINI index over the time period Avg life expectancy at birth % births attended by skilled attendant Physicians per 100,000 Primary school net enrollment ratio	1.093***	0.902*** 0.019*** 0.027*** 0.006*** 0.005*** 0.003*** 0.0007*** 0.001***

The authors ran two separate models. The first only controls for a country’s income (measured by gross domestic product). The other adds in country-specific attributes. All of the results are significant as indicated by ***. Answer the following questions: (30 points)

1. Are health care commodities normal goods? y/n
2. Would you say health care services as it relates to income is elastic based on Model 1? y/n
3. Once we control for indicators of how developed the country is (including any unobservable differences across countries by controlling for fixed effects) the income effect goes down. In other words, by controlling for more factors that explain health care spending, and are also related to GDP (which measures the country’s income), we come closer to the isolated effect of income. This provides support for which of the following economic concepts used repeatedly in our modeling?
   1. Complements
   2. Elasticities
   3. Ceterus parabus
   4. Consumer sovereignty

4. Would you classify health care as a necessity, or a luxury based on the income elasticity in Model 2 (the preferred model)?

Answer 1

a)

Normal goods are those goods whose demand rises as the income of the consumer goes up. In other words, as you earn more income, you'd demand more of these goods.

In order to check how the demand for healthcare will be corresponding to the income per capita (or GDP per capita), we have to check the partial / marginal effect of GDP per capita from model (1). Note that this is given by the estimated coefficients. This gives us a coefficient of 1.093, which is positive. Thus when GDP per capita rises by 1 unit, there is a positive change in the demand for healthcare, i.e.. demand for healthcare rises when income rises.

Hence, healthcare is a normal good.

b)

The marginal effect of per capital GDP on healthcare in model (1) is 1.093. This means that as GDP per capita increases by 1 unit, the demand for healthcare rises by 1.093 units (which is >1). Thus, healthcare rises by more than 100% (109.3%) when GDP rises by 100%, which suggests that the income elasticity > 1.

An elasticity > 1 it implies that healthcare is a relatively elastic good.

c)

Controlling for indicators of how developed the country is (including any unobservable differences across countries by controlling for fixed effects) the income effect goes down. This means, we are keeping the effect of other variables constant or fixed when we want to derive the isolated effect of income on healthcare consumption. This is the idea of ceteris paribus, which basically means "keeping other things fixed or constant"

d)

Note that luxury goods have an income elasticity >1, meaning that as income changes by 1 unit, the quantity demanded of those goods change by more than 1 unit. On the other hand, income elasticity <1 means as income changes by 1 unit, the quantity demanded of those goods change by less than 1 unit.

As in model (2), we have the marginal effect of income on healthcare in model (2) being 0.902, which is clearly less than 1. It means that for a unit change in the GDP per capita, the healthcare consumption changes by 0.902 units (<1). Hence healthcare in this case is a necessity good.