Econometrics Final Research Project

Literature Review

Our research explores the question of whether the value of a bachelor’s or master's degree

varies by geographical location within the United States. We answer this by measuring the value

of a degree by income. Several research papers have been published on this topic or similar

elements. Not only is this an interesting question for curiosity’s sake, but it is also worth

answering in a practical sense for individuals deciding where to live. If a person is interested in

increasing how much income they can earn with a given occupation, they might be interested to

know where they should live to maximize the value of their master’s degree.

To understand the question and its context we need to understand the value of education

in a static environment; in other words, without geography being a factor. “The Big Payoff:

Educational Attainment and Synthetic Estimates of Work-Life Earnings. Special Studies. Current

Population Reports” (Day, Newberger 2002) has already established valuable information on this

topic. This economic research paper goes in depth to explain the effect of education level on

work-life earnings (and how that effect increases with time) and takes into consideration several

other significant social factors, including race, sex, and work experience. Using 5-year age

groups if categorized by sex and 10-year age groups if categorized by race or ethnic origin, Day

and Newberger ultimately show that differences in earnings between education levels increased

between 1975 and 1999. A worker with a graduate level degree earned 1.8 times more in 1975

than a worker with a high school degree and 2.6 times more in 1999. This strengthens our own

understanding of the factors that change income for individuals based on variables like race and

sex. Our study extended this research by examining the effect of geography on income.
 “Education and Lifetime Earnings in the United States” is a research paper by

Christopher R. Tamborini, ChangHwan Kim, and Arthur Sakamoto (2016) which approaches the

relationship between educational attainment and income through the lens of lifetime earnings.

This paper examines how level of earnings changes based on stages of life. The authors do this

by looking at four cohorts of individuals born within a range of seven years, and those cohorts

being ten years apart. For example, cohort one was born from 1962 to 1969 and cohort two was

born from 1952 to 1959. This allowed them to compare the different groups while they were in

different stages of life. In 1982, cohort one was 13 to 20, cohort two was 23 to 30, cohort three

was 33 to 40, and cohort four was 43 to 50. As the years moved on, they were then able to see

both what the cohort had earned in past periods of life and current periods, and compare them to

the other cohorts who were in different present stages. They found that people who are later

degree holders earn less than those who never earn a degree during the beginning stages of their

earnings (early 20s) because the that is the time degree holders are earning their degree rather

than fully participating in the workforce. Even so, these people who later earn their degree make

a significant amount more in their later earning years, especially those with graduate degrees.

While most people continue increasing their earnings until peaking in their 40s, plateauing, and

declining in their 50s, people with graduate degrees extend their rising median earnings until

their mid-50s. This study has informed our research on age and income, especially when

considering the effect of educational attainment. Again, our study is set apart by its distinct

consideration of geography as a factor of earnings.

In “The Long-Run Economic Impact of an Institution of Higher Education,” (Johansen,

Arano 2016) the relationship between educational institutions and the community around them is

brought to the center of attention. This paper uses the human capital contribution approach to
further the short-run expenditure approach. They do this to examine the impact a university has

on its state economy through their alumni’s skills which they learned at the university. These

skills are part of human capital. They also aim to observe the counterfactual, which is migrants

or people who attended colleges in other states, not attended college at all, or if attended trade

school. This is because they constrain the study to alumni who graduated specifically from the

college in the area who still live in the area, to accurately observe the expected economic and

income changes. They find two benefits of the university’s impact in an area. The first is that the

existence of a college in an area creates additional economic activities simply from the university

community. The long run effect is for alumni: they find evidence that additional work life

earnings is higher for alumni but again this is only measured for alumni that live and work in the

area they completed their degree(s). They also state that these findings show growing human

capital as a benefit of institutions in an area. Johansen and Arano applied their study to a public

regional university. In our own study, we will be examining the value of degrees in cities with

universities and will use this economic study to inform our own findings.

“Education Level Drives City Growth,” (Davis 2004) is a comprehensive digest that

points towards the relationship between education level and city development. It references two

economic research studies that demonstrate the effect that an educated population has on its

community. The first study is called “The Rise of the Skilled City” (Glaeser, Saiz 2003) and

points out the trend of cities with a higher percentage of educated citizens progressing faster than

cities with less educated citizens. They show that education and skills increase the growth and

productivity of a city, and they research to explore the compatibility of this assertion with the

Information City hypothesis and the Reinvention City hypotheses. The Information City

hypothesis says that skills should predict growth among all cities. The Reinvention City
hypothesis says that skills should only matter towards growth in cities that have experienced

negative shocks. Glaeser and Saiz run several regressions to show impact on city growth of

various variables including a percent college educated variable, climate factors, and human

capital. They find results that align with the Reinvention City outlook: human capital and growth

have a strong relationship in declining cities but have a weaker relationship relationship in

growing cities.

The second study referenced in “Education Level Drives City Growth,” (Davis 2004) is

called “Reinventing Boston: 1640-2003” (Glaeser 2003). The majority of this paper is spent

informing the audience of Boston’s economic history: its highs and its many lows and its

recoveries. In the 1990s, Boston suffered under threat of irrelevance as a declining American

city. Now, it has one of the hottest housing markets in the country. This paper finds a clear

relationship in Boston of education and urban success, as measured by population growth,

income growth, and housing price growth. It asserts that housing prices reflect the desire of the

public to live in a city, which is representative of the city’s success. Ultimately, the paper

concludes that higher education is correlated with urban success, and that success leads to

increased population as people desire to live there. With new construction being hard to come by

in the very regulated Boston economy, lack of new supply of houses translates into increased

housing prices.

“Education Level Drives City Growth,” (Davis 2004) ties together these two papers and

builds on them by asserting that the percentage of workers with a college education is an

indicator of city growth and local development in those areas. Both articles are important in

affirming our own research while also providing us with more insight on the cities we are
examining. Reading these papers also inspired ideas for what research we could expand upon in

this paper.

The most notable and valuable study we will use is “What You Make Depends on Where

You Live: College Earnings Across States and Metropolitan Areas” (Winters 2020). This study

serves to help people make decisions on what to invest their money in (education, residential

location) by informing them with helpful data. For the first part of his research in this paper,

Winters uses three years of data that focuses on full time, full year workers. This makes the set

more comparable when dividing the data into groups, which allows for a wider range of

outcome. The study examines four levels of education: high school diploma but no college, some

college but no degree, associate degree, and bachelor’s degree but no advanced degree. The

report focuses on mean earnings as the main measure of earned income. Winters finds that as the

education level increases income does as well. Notably, there is a smaller increase between some

college and associates and a large increase in income between associates and bachelor's degrees.

For the discussion around regional effects, they collect for the same education levels and income

data but add the variables of eight regions defined as Far West, Rocky Mountain, Plains,

Southwest, Great Lakes, South East, Mid East and New England. They do not go into their

findings extensively on this matter but they do find premiums of income for those with

bachelor’s over all other levels observed.

The last hypothesis Winters explores is metropolitan area. They continue to use the

previous data and education levels mentioned but add variables for twenty-five cities across the

US. They create a set of categories to observe four groups of metropolitan population levels: less

than 0.5 million, 0.5 million—1.5 million , 1.5 million—4 million, and more than 4 million.

They also look into mean income and similarly to the results found throughout regions, they find
that a bachelor's degree makes a premium of income no matter the city or metropolitan size. This

study essentially performs the same tasks we aim for while using CPS data, just like we do. This

will be a very valuable resource, especially as it already establishes the question we are trying to

answer. This study is set apart by the ones that come before it by its emphasis on geographic

location. Other studies like those previously mentioned simply study the relationship between

education level and income. In the same way we do, Winters wonders whether the relationship

changes based on where one lives in the United States.

Our study does not venture too far from “What You Make Depends on Where You Live,”

(Winters 2020) but it does approach the geographic aspect from a regional perspective rather

than a detailed state by state perspective. We explore deeper into patterns that occur based on

regions rather than state and hope to find other variables that contribute to these trends. They

investigate variables such as age and gender at a state-by-state examination. Our research is also

different from this study because we examine the income differences in graduates specifically

with a master’s degree. In contrast, Winters actually includes high school graduates in their

study, along with people with associate’s degrees. Although Winter examines a different sample,

their findings still add valuable comparative information to our research.

The existing pool of literature relevant to our research is concentrated mostly on earnings

and the effect of education. The third most important factor of our study being geography, we

found little research on this subject in comparison. The study of the relationship between income

and our three focuses in tandem(education, number of colleges, and geography) is what sets our

research apart and provides a foundation for further exploration.