Methodology for Determining the Economic Development Impacts of Transit Projects (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

100 7. CONCLUSIONS Our empirical analysis found that metropolitan-area-level associations between transit capacity, agglomeration, and productivity are strong. But in the four regions selected for firm analysis and for qualitative case studies, there is mixed evidence on whether and under what conditions improving transit capacity will cause densification of employment near stations, or influence the industry mix along corridors. In Portland, Oregon, we did find some evidence that there was significant densification and firm clustering patterns near transit stations, but in Dallas the opposite result appeared to be occurring. In the qualitative case studies, we did not find strong evidence of strategies to encourage densification near transit stops, or of actual employment densification occurring. In fact, there was more evidence of population densification occurring near rail corridors. This suggests that any agglomeration-related productivity increases may be more large-scale than what is possible to readily observe with the methods we used for our firm-level analysis, or with the case study approach. This is consistent with the relatively large-scale measures that we found to be significant in our metropolitan area-level analysis: principal city employment density (not limited to employment density near rail stops, but throughout the main cities in the metro area) and metropolitan area population. We investigated the role of physical metropolitan area-level measures of agglomeration as the mediator between transit investments and productivity impacts. We were not able to analyze the potential effects of “effective density” (i.e., employment accessibility) or other local measures of agglomeration that might also be affected by transit investments. However, as discussed earlier in the report, metropolitan area-level physical measures may be most appropriate in the case of rail investments in the US. The firm-level analysis results and the case study results are helpful in providing a caveat to the national level estimates. The spreadsheet tool simply employs the national estimates because there is no straightforward way to incorporate additional information, and because the national level estimates provide the only quantitative findings directly related to productivity. We would recommend that these estimates be combined with an evaluation of the likely regulatory barriers to local or regional densification and growth. Regulatory policy includes parking requirements, building density, maximum building heights, infrastructure provision, and short duration of the approval process. All tend to push against the realization of agglomeration benefits. Local policy should be designed to facilitate agglomerations rather than to prevent them. There could also be policies to proactively encourage agglomeration, including transit inducements like the Transit Hub Tax Credits in New Jersey. Similarly, there have been instances, such as in Arlington, Virginia, where the relative unavailability of developable land elsewhere likely intensified development along transit corridors. Although we have not evaluated the efficacy or value of these particular policies, our case studies suggest such policies could play an important role in agglomeration economies. Application in practice: Spreadsheet tool We created a simple companion spreadsheet tool that can be used by transit agencies, MPOs, or others to calculate the wage and GDP benefits of proposed rail projects after entering one or more of five possible measures of new additional transit capacity: new track miles, added total seat capacity, added bus seat capacity, expected total additional revenue miles, and/or expected additional rail revenue miles. As noted above, these terms are defined as follows. Track miles are “route miles” of track, not accounting for direction or number of tracks (therefore only

101 a fraction of the “directional route miles” reported by agencies to the NTD). Rail revenue miles and seating capacity are both reported by agencies to the NTD, so they are familiar with both terms. Rail revenue miles is defined as the miles that vehicles travel while in revenue service, including layover time but excluding deadhead. Seat capacity is the total number of seats in the vehicles owned by the operator. Projecting rail revenue miles is subject to some judgment by the proposing agency, while the other two measures (rail mileage and seat capacity) are less subjective. Applying our estimates using the spreadsheet tool is very straightforward from the transit agency’s perspective, although the population and employment density figures should be updated annually for the greatest accuracy. The spreadsheet contains a row for each metropolitan area in the United States. Note that for users who have access to Excel Version 2010, a drop-down list enables the metropolitan area to be chosen. For those with an earlier version, the MSA names must be copied from the second tab in the spreadsheet. A range of possible agglomeration- productivity benefits is provided in the box on the right-hand side of the spreadsheet. These estimates are best used to compare transit agency proposals to each other, rather than as absolute dollar values that can be compared to cost-benefit analysis results. The spreadsheet tool is user-friendly in the sense that all the metro-level data is built in, and the user only needs to enter two or three summary pieces of data about the proposed transit investment to see a range of estimated agglomeration impacts. From that standpoint, it should be attractive for use by a range of stakeholders, even those less technically inclined. The downside is that one of the three input types, revenue miles, can be tricky to forecast. But all of the measures are required to be provided by transit agencies and reported to the FTA for the National Transit Database, so they are not new to those agencies responsible for those reports. Updating the spreadsheet tool yearly with current population, employment density, and other metropolitan characteristics would be required in order to have the most current estimates, because these factors affect the net estimates of productivity benefits. However, the accuracy of the estimates will likely not suffer very much for at least five years, since there are not likely to be large shifts in any of the underlying metropolitan area data over that time period. Population estimates, wages, and payroll can be downloaded from the Census and BEA websites and readily entered into the spreadsheet. Updating the central city employment density figures is more complicated because the Census LEHD data have to be aggregated up from the block level to the central city level; this task took our research team a substantial amount of time. Spreadsheet tool limitations A minor limitation of the spreadsheet tool is that it cannot easily be used to project values into the future, or allow anticipated values to be input (for example, future expected levels of wages and GDP per worker, population, and number of workers). Making changes of this kind, however, would be relatively straightforward. A more significant improvement to the tool would be to extend our existing models to incorporate characteristics of projects other than only track mileage, rail revenue miles, and seat capacity. Such measures could include connectivity/accessibility, alignment quality, projected ridership, or other within-MSA information. (Ridership could not be used as an input to these models because it is endogenous: agglomeration also increases ridership.) However, conducting the analysis needed to support such measures would require substantial additional resources. The most significant limitation of the tool is that we could not construct a dataset that would enable us to distinguish one proposed project from another within the same metropolitan

102 area (e.g., two different corridors for a new rail service). Future work could try more fine-grained analysis, but its success is somewhat uncertain because the data do not support detailed modeling; there are too few total rail investments in the US in the last two decades or so (about 80). Thus the value of the spreadsheet tool, following the value of the metro-level analysis on which it is based, is in distinguishing the agglomeration impacts of transit investments in different metropolitan areas from each other, not in distinguishing different alignments within the same metropolitan area. Use of spreadsheet tool outputs The spreadsheet tool would be most productively used for comparative analysis of proposed projects from different regions. This was one of the main goals of the H-39 research effort. The outputs could also be used by agencies to help inform public outreach and information campaigns about the possible benefits of transit investments. Another use for the tool is to compare different rail service scenarios within a region, without depending on their spatial characteristics—only depending on their total track mileage, seating capacity, and projected revenue miles. For example, there might one scenario in which five or six proposed rail lines are to be constructed in a region, and another where only one or two new lines are built. In this case the agency can sum the new track, revenue miles and seat capacity for all of these, enter them and compare the outputs from the spreadsheet. A better service might, in theory, have more revenue miles but less track; the range of estimates would reflect this and might yield roughly equivalent estimated agglomeration effects for the two scenarios, or might yield a significantly higher upside for the higher-revenue-mile scenario. The analysis on which the spreadsheet tool is based was carried out in wage and GDP terms, not in terms of jobs. So the outputs cannot be directly discussed in terms of job creation, although the effects on the wage side are all job-related. Metropolitan area-wide wage and GDP increases can take many forms—including more jobs, higher wages for existing jobs, shorter unemployment spells, and greater firm profits—and most likely, a combination of all of these. Future research Further research is certainly needed on this topic. The obvious next research steps are to collect and analyze historical data on transit, agglomeration, and productivity over several decades, and to use more advanced statistical methods to better understand the mutual causality between agglomeration, transit and productivity. Other research needs include: testing other measures of transit capacity; examining how firm formation may occur in response to transit investments; and investigating whether the residential focus of TOD efforts may dampen employment-related agglomeration impacts of transit. Our data are limited in that we are not able to distinguish the effects of rent increases on gross domestic product (GDP) or wages. But the literature is clear, as we note in Appendix A, that wages and rents both reflect some portion of agglomeration benefits and that therefore both should be investigated simultaneously. We do use both wages and GDP as measures of productivity; both capture some different and some similar elements of productivity. Future research should bring rents into the analysis. Much more could be done to discriminate among different types of cities so as to better understand how metropolitan areas themselves may vary in terms of agglomeration effects. This

103 is readily achievable within the current dataset, although it would benefit highly from a decadal (10 year census) data approach. It would also be desirable to have a companion highway input-output model to fairly evaluate the total scenario impact, although our estimates do include highway miles as an input. It is unclear from our analysis over what period of time agglomeration benefits can be expected to accrue after a transit investment. Among other methods, it might be fruitful to qualitatively investigate the perceptions of agglomeration benefits by tenants and other business and residential location decision-makers. As noted in our case studies, market acceptance may play a role in the lag time between the new or improved transit facilities’ completion and the visibility of agglomeration benefits. The metro-level estimates could be tested retroactively in a region where transit investments have been made and compared with a region where a new investment is proposed. This would require overcoming limitations in the current dataset, which has only about eight years of data.