Since this is my first blog post, a short introduction is probably warranted. My name is Manu Navjeevan and I’m an economics student at CMU. My honors thesis is focused on studying trends in income mobility in the U.S and specifically in Pittsburgh/Allegheny County. I chose this topic, partially because it is, I believe, extremely relevant in today’s political climate, but also because it is a less studied field of economics that I felt I could contribute to.
The last few weeks have been extremely constructive in terms of getting a more focused research question and getting a better idea of how to approach the problems I want to work on. When I came into this at the beginning of the summer, I had relatively little idea of what specific question I wanted to answer. I had chosen the topic of income mobility about halfway through the spring semester, with some input from my advisor, Prof. Laurence Ales and had initially thought I would look at specific Pittsburgh programs and see how they might affect income mobility. As I did a bit more looking into the subject over the back half of last semester, it became increasingly apparent to me, however, that I did not have the data to analyze these programs. To study the effect on a specific program on life outcome, one needs individual level data on a variety of variables and over a relatively long period of time for the people in the program. Even if the City of Pittsburgh kept this data, the odds that I would be given access to this data (privacy concerns, etc.) or that it would be robust enough to get significant results were slim. Also, I was having some trouble identifying programs unique to the City of Pittsburgh that I could analyze (though this was probably due to a lack of discipline in looking through the budget on my end). Because of this (and again, in the interest of transparency, a good deal of laziness in doing any real research or reading on my topic during the school year), I didn’t really know what I should be doing apart from reading papers when I got back to Pittsburgh.
However, through reading papers, I began to get a better idea of what problems I could reasonably expect to tackle in an honors thesis. My advisor, Prof. Laurence Ales, has also been particularly helpful in this regard, pointing me to a number of websites where I could find county and census tract level data. Also, with the help of the Dean’s and Associate Deans in Dietrich, I was able to get in touch with the office of City Councilman Dan Gilman and meet with his Chief of Staff Erika Strassburger this Tuesday to talk about city and county programs targeting income mobility. As it stands, I am currently studying income mobility via two approaches.
The first is to look at what county programs or attributes may be correlated with higher income mobility. Through the work of Prof. Raj Chetty at Stanford, we have estimates on the causal effects of living each county in the U.S on income mobility. We don’t know, however, what policies may drive the differences in income mobility between counties. By looking at data on demographic characteristics and the relative sizes of people on public assistance income or on differences in public spending in these counties we hope to study these differences. I’ve currently merged together census data with Chetty’s estimates and am in the stage of identifying what characteristics may be the best predictors and cleaning/transforming the data to perform inference on our regression estimates.
The second approach is through studying the effects of gentrification on income mobility, a problem salient to Pittsburgh. There is a considerable body of work out there that shows that growing up in a better neighborhood has positive effects on life outcomes for poorer children and there is some evidence that people in gentrifying neighborhoods may not move out at a higher rate than in non-gentrifying neighborhoods. Given this, we may want to study the extent to which (if at all) the gains for poorer children in gentrifying neighborhood caused by lower crime rates, etc. are offset by the detriments (less disposable income, more inequality, etc.). To this end, I again used census data at the census tract level to identify which of the over 50,000 census tracts in the U.S look like they’re gentrifying and to what extent. I was able to use this to generate a heat map (below) of which states in the U.S look like they are experiencing the most gentrification (weighted gentrifying neighborhoods as a % of total neighborhoods). The map shows a few interesting results. For example, there appears to be a lot of gentrification in the Dakotas as well as Montana, which runs contrary to where we might believe gentrification is happening. Some of this can probably be explained by the emergence of shale gas in those regions making oil towns in those states significantly better off (North Dakota has the nation’s lowest unemployment rate). When this is combined with the fact that those states have relatively few people, and therefore relatively few census tracts, it probably explains the high rate of gentrification we are seeing on the maps. The hope is now to study outcomes or other characteristics of gentrifying neighborhoods to get a better sense of their effects on life outcomes.
Research aside, life in Pittsburgh over the summer has been relaxing. It’s odd to be on campus without as many people but it means that finding a place to work on campus is nice and restaurants in the area are generally less crowded. Also, because there are no homeworks or midterms, there’s more time to run errands or catch up with people over the summer that, over the school year, you may not get as much time to see. The flexibility of independent research also allows me to go to events and fit my schedule around other things that I may want to do. I’m excited for the rest of the summer, both in terms of making progress on my thesis as well as being to do things in Pittsburgh that I haven’t made time to do yet.