This last week, I conducted a linear regression of block group land value and land value change and proximity to transit and specific transit capital expansions. I was unsatisfied with the inconclusiveness of my previous analyses, which had created a somewhat platitudinous conclusion that the effects of transit and proximity to downtown are difficult to disaggregate. Regression analysis was developed precisely for this purpose, and, though the relationship between both transit and centrality and land values appears more explained by a logarithmic or inverse power relationship, a simple linear regression of these two variables was able to explain around 50% of the variance in land value/value change for most of the time periods analyzed. The analysis revealed that proximity to downtown is obviously the dominant variable, though transit proximity was statistically significant for: proximity to all rail and 2015 value; proximity to all frequent transit and appreciation since 1998; proximity to the Yellow line and appreciation from opening to 2015; and proximity to the eastside streetcar and appreciation in the three years since its opening. There is thus some evidence that transit modestly increases land value (a driver and expression of spatial class upgrading) beyond the extent to which its provision in Portland is simply correlated with proximity to downtown.
This week, I also edited my poster for Festival of Scholars. I added a brief introduction before the gentrification theory/typology section, replaced my earlier variegated correlation with the linear regression results, did some miscellaneous edits, and split conclusions into discussion and larger implications. I added more explicit treatment of why I highlight gentrification typology in the introduction; discussed some of the broad spatial differences in my results; and attempted to draw the results out more broadly by adding some information on municipal policies, integrating my framework of transit’s link to gentrification through its direct effects on land value and its densification justification and amenities indicator functions; and commented on the wider applicability of my results.