ABSTRACT. In the United States, the Federal Register publishes, among other things, regulations. These regulations are continuously updated and are rarely removed, leading to an accumulation of regulations. We examine how regulation growth in one sector spillover across the different sectors. We use the Industry-Specific Regulatory Constraint Database (RegData), which quantifies federal regulations at the industry level from 1997 to 2020 based on text analysis and machine learning. We then develop Connectedness Indices measures to examine information transmission of regulation from one sector spillover across the different sectors.

ABSTRACT. We revisit the question of the economic consequences of state-level business climate indexes in a dynamic context and a more rigorous treatment of selection issues than previously undertaken. Specifically, we evaluate the effects of a high enactment rate for model bills advocated for by a prominent lobbying group, the American Legislative Exchange Council (ALEC). Using a treatment effects approach, we find that, holding the existing business climate constant, there is no effect on overall Gross State Product, with a possible negative effect on employment and higher average wages. States both raise more tax revenue and increase expenditures, while poverty rises. The income distribution is compressed in the short run, but there is no long-term effect. We also offer some speculation as to the mechanisms underlying these dynamics.

ABSTRACT. The literature is mixed as to whether veterans receive a wage penalty or wage premium in the private sector. In many studies, context and demographics matter. In line with previous research, a recent paper by Renna and Weinstein (2019) find that on average women veterans and black veterans tend to experience a veteran wage premium whereas white veterans tend to face a veteran wage penalty. They also find location choice matters too. Upon a more detailed examination of the full distribution of wages, Renna and Weinstein find that in the left tail of the wage distribution, all demographic groups experience a veteran wage premium (but the premium eventually becomes a penalty as you move along the distribution of wages). Their results support the theory that military service is especially beneficial for individuals from socioeconomically disadvantaged backgrounds, the “bridging hypothesis”. This provides some evidence to alleviate concerns that military service may fall unfairly and more heavily on socioeconomically disadvantaged individuals. Still, studies on the veteran wage differential and specifically the bridging hypothesis may miss the more fundamental question on the economic wellbeing of veterans. That is, how does military service affect the intergenerational mobility of veterans? We examine and compare how military service impacts intergenerational mobility using panel data from the National Longitudinal Survey of Youth (NLSY) and the Panel Study of Income Dynamics (PSID). Renna and Weinstein (2019) use birth location to show exposure to military service increases the likelihood of serving in the military. If military service affects intergenerational mobility, then this may shed some light on one of the mechanisms behind the geographic disparity in intergenerational mobility across counties in the U.S. (Chetty, Hendren, Kline, Saez, 2014). Geocoded NLSY data also allow us to measure exposure to military service at age 16, around the age many teenagers may be considering military service. Initial results from the NLSY suggest that military service may increase the mobility of veterans and weaken the link between the permanent income of parents and their children.

ABSTRACT. The proposed research exploits differences in the relative intensity of incentives across industries and cities, combined with individual-level data which includes information on wages and employment, to understand how industry-targeted EDI affect individuals’ wages and how differential targeting affects workers by race, sex, and position in the wage distribution. We use exogenous variation in available incentives and changes in local industry composition in an instrumental variable approach to answer the following unanswered research questions:

1) How does the extent to which locations incentivize specific industries vary by the industries’ position in the wage distribution as well as racial and gender composition? 2) To what extent does variation in industry-level incentives affect wages for workers in the targeted industries relative to wages for workers in the same occupation in non-targeted industries both in the same location (city) and in the same industries in other locations? 3) Are there unintended consequences in terms of segregation or lower wages for workers in untargeted industries? 4) How do incentive-induced wage changes vary by position in the wage distribution, race, and gender?

ABSTRACT. In a society where people of all races are perceived equally, the neighborhood racial composition should not impact house price, ceteris paribus. Prior work has robustly identified house price discounts associated with increasing proportions of Black neighborhood residents in some housing markets, while in other housing markets statistically significant price discounts are not apparent. We investigated one factor that may explain the variation: quantile effects. We found that after considering within market heterogeneity, 94 of the 95 housing markets studied was characterized by statistically significant price discounts in at least one quantile of the house price distribution. If price discounts are interpreted as evidence for willingness to pay to avoid Black neighbors, our results suggest that this prejudicial behavior is widespread across US housing markets.

ABSTRACT. This paper studies the effect of craft breweries on the residential property values. Using a hedonic Difference-in-differences approach and investigating about 250 thousand housing transactions in Denver, Colorado from 1990 to 2016, we found that the opening of a new craft beer brewery contributes to a 10-15% premium to the housing prices nearby. Meanwhile, the premium decreases as the Euclidean distance of a home to the newly opened brewery increases. We also found that different breweries (micro-brewery vs. brewpub & taproom) have different effects, and the effects vary across different properties (single-family home, condo, and rowhouse). Last but not least, we also measure the distance by walking time, and the results are robust. The findings suggest that when policy makers enact location specific policies for economic spurring such as opening craft brewery districts, the heterogeneity of specific business and neighborhood characteristics should be considered.

ABSTRACT. This study examines variation in the rise of U.S. housing values between January 2020 and September 2021. In that period, the U.S. experienced unprecedented growth in the housing sector. While drivers have not yet been fully explored in the empirical literature, media accounts point to a confluence of economic and social factors that came to a head as the economy began to reopen following the height of COVID-19 closures; notably, however, the adoption of remote working may have allowed urban residents to relocate to peri-urban and rural communities, leveraging the sale of high value urban properties to purchase housing in non-metro areas at a premium. Here, we examine differences in growth rates of housing value data between urban, peri-urban, and rural settings. We test the hypothesis that values in non-metro areas grew faster than urban areas during the identified window. Data utilized is the Zillow Housing Value Index at both county and ZIP code levels. The ability to disaggregate to the zip-code level is important, as county level analysis may overshadow the distinction of urban, suburban, and rural areas. To designate geographic areas, we utilize two different metrics: at the county level, we use the rural-urban continuum code (RUCC) from USDA’s Economic Research Service is used to classify counties into nine types, while at the ZIP code level a new index of urban-rural perception provided by the U.S. Department of Housing and Urban Development's Office of Policy Development and Research is employed. Both OLS and quantile regression techniques are used to estimate differences in home value growth between area types. While OLS can test differences at the mean value, it may be problematic under circumstances where home value growth has risen well above the mean value, thus the need to incorporate quantile regression analysis. Models are estimated for each month from January 2020 to September 2021. This allows us to visualize housing value growth between different area types and identify whether growth trends widen or narrow. Results indicate that suburban home values grew at a faster rate than urban housing beginning in January 2021, and that compared to suburban and rural values, growth rates in urban areas plateaued around July 2021. Spatial analysis indicates that the largest overall growth rates were in the Western U.S., the peri-urban border of the east coast’s metropolitan corridor, and Florida.

ABSTRACT. This paper examines the impact that the COVID-19 recession had on states and regions in the United States. The COVID-19 recession was the strangest recession this country has ever experienced. The national unemployment rate went from 3.5 percent to 14.7 percent in just two months. Real GDP fell by 31.2 percent in the second quarter of 2020, and during the third quarter of 2020 real GDP increased by 33.8 percent. The NBER determined that the recession lasted only two months the shortest on record since 1854. During March and April of 2020 over 22 million establishment jobs were lost, almost three times as many as were lost during the Great Recession. While the national numbers are staggering, the state impacts were quite varied. In April of 2020 state unemployment rates ranged from 7.9 percent to 28.2 percent. Similar differences occurred in real GDP disparities and in job losses. This paper looks at two different measures of state recession impacts; the initial decline in real GDP, employment, and unemployment rates and the length of decline. The paper will also use regression analysis to investigate why some states faired poorly while other states suffered little during this downturn. This paper uses state data from 2005 through 2021. The regression model used includes both qualitative variables and fixed effect variables to explain the differing performance of the 50 states during the downturn and subsequent recovery.

Louis A. Amato UNC Charlotte Richard Cebula George Mason University John E. Connaughton UNC Charlotte

ABSTRACT. We use the pandemic as an opportunity to analyze what happens to traffic patterns when there is an abrupt change to one variable. This era provides an experimental environment in which we can develop a model to detect changes in spatial patterns under the influence of this change. Using pedestrian crashes as our traffic pattern, our results suggest that this abrupt change - a global pandemic and its consequences - does not alter existing neighborhood crash patterns. We are able to determine that pedestrian crashes in 2019 are statistically indistinguishable from 2020 pedestrian crashes, despite fewer total occurrences and the impact of the pandemic on traffic.

ABSTRACT. The ABC Science Collaborative has been an influential voice in the discussion about how schools should respond to the COVID-19 pandemic. Their influence has been underwritten by a pair of peer-reviewed publications written by ABC members that have appeared in the prestigious journal Pediatrics: Zimmerman, Akinboyo, et al. (2021) and Zimmerman, Brookhart, et al. (2021). Unfortunately, the main conclusion of each paper - that COVID-19 transmission in North Carolina schools was relatively low in the covered time periods - was dependent on problematic contract tracing. In this paper I perform an easy robustness check on the Pediatrics results by using North Carolina data to calculate COVID-19 rates for age-matched cohorts of the schools studied. This analysis is not nearly so favorable to the central claim of each paper. This calls into question reliance on contact tracing data and complicates the ABC Science Collaborative’s endorsement of face-to-face schooling. I explore the methodological oversight of the ABC papers using the psychological concept of motivated reasoning.

ABSTRACT. This study examines the location decisions of manufacturing foreign direct investment (FDI) firms in the United States, focusing on taxes and incentives relative to agglomeration as determinants. Using a panel Poisson regression with random effects, we model the probability of site selection in U.S. states and counties. The results suggest that localization and urban agglomeration economies clearly exert the most influence on FDI location. The localization estimate, as captured by the number of domestic manufacturing establishments, has an elasticity of 0.92. Urbanization economies, measured by the area's wage premium, have an elasticity of 1.31. Among taxes and incentives, the investment tax credit (as a share of value added) is statistically significant, with an elasticity of 1.56. Further analysis reveals that this incentive is only significant in counties that rank in the highest quartile of agglomeration. In areas falling in the lowest quartile of agglomeration, our estimates indicate that job training subsidies may attract FDI. Property tax, the job creation tax credit, and research and development tax credit have no measurable effect on the location decisions of foreign manufacturers. In addition, the distance from the foreign-owned plant to an airport appears to be an attractive determinant.

ABSTRACT. We investigate the offshoring effect on local productivity and capital investment. We find that offshoring can increase local productivity and capital investment, specifically in non-offshoring industries. Both MSA (urban) and non-MSA (rural) counties receive benefits of productivity expansion and capital investment from offshoring. The increased capital investment from offshoring could be a channel of local productivity expansion.

ABSTRACT. This paper investigates whether spending on economic development programs converges across U.S. states. States use a wide array of tax and subsidy programs to try to attract firms in a highly competitive environment. If states engage in strategic interaction as previous literature indicates (Wang, 2018), we should expect economic development spending to converge over time. We provide a novel test of this implication using a national database of state-level economic development program expenditures and the panel convergence method developed by Phillips and Sul (2007, 2009). We uncover three spending convergence clubs as well as factors associated with club membership. Even if states aspire to emulate each other in terms of economic development efforts, we find that they flock together in groups based on socioeconomic characteristics and geographic proximity. This implies that global spending may be limited by underlying state determinants of de facto spending.

ABSTRACT. The transmission of pollution across countries has been studied through the lens of atmospheric chemical transport or through its content in international trade. The few studies that consider both channels concurrently do not highlight what the key drivers of the change in pollution production and transmission are. Based on a structural decomposition method this paper uncovers which changes in target pollutants emanate from technological changes, structural changes, final demand changes or household activities taking place locally, in the trade partners, or in the upwind countries. We apply our approach to a five-region model and focus on carbon monoxide (CO) for its capacity to promote the formation of secondary pollutants. Our results provide solid scientific evidence for the US, the European Union and South Korea to request changes from China because the large increase of its domestic demand is the main driver of the growth in CO they experienced over 1990-2014. By providing new insights into the interconnected sources of air pollution, this paper suggests more nuanced global emission abatement policies than the consumer-focused or producer-focused approaches currently used.

ABSTRACT. This article addresses the question of the impact of culture on transport choices. We test the hypothesis that Australia is an exception for the link between density and gasoline use (the Newman and Kenworty curve). Actually, Sydneysiders should drive more than they do if they were following the Density/energy curve (on the VMTs versus density link, Sydney differs from US cities with a p value of 0,054). The share of transit for journey to work is significantly higher in Sydney (p value 0,008) and to a lesser extent, in Melbourne (p value 0,11) than in the average American cities (N.Y. excluded). This might be due to differences in culture, lato sensu. So, among many factors, geographical history, transit and land use policy, culture is a co-factor of urban passengers behaviours.

ABSTRACT. Commuting zones introduced by Tolbert and Killian (1987) and recently re-examined by Foote et al. (2017) are groups of counties used to delineate local geographic labor markets across all counties in the United States. They are appealing geographic markets as they are based on actual commuting patterns and not political boundaries and they are commonly applied to look at a variety of economic activity and shocks. However, commuting zones are not the most appropriate grouping for other economic activities, including household consumption, which similarly spans across county borders (Dunn and Gholizadeh (2020)). We introduce consumption zones as an alternative to commuting zones for analysis of household consumption. For constructing these consumption zones we use unique card data from Fiserv, a major card transaction intermediary, which are used to estimate spending flows across more than 3,000 counties and 15 NAICS industries in the United States. Using these spending flows we construct consumption zones for each of the 15 industries, as well as for aggregate consumption, following the same clustering methodology used to form commuting zones.

ABSTRACT. We adapt a machine learning method to provide new estimates for land valuation in the United States, pairing this approach with “big data” from Zillow. Because this data includes detailed information from hundreds of millions of property transactions covering much of the US, the heterogeneous nature of this data serves as fertile ground for highlighting some of the practical limitations of linear hedonic regression techniques for land valuation, a common method for mass appraisal of land. We first construct traditional hedonic estimates of land value at the parcel-level for most of the US as a baseline, focusing on single-family residential properties in our initial analysis. We then modify the hedonic approach by using a machine learning method, gradient boosting trees, for comparison. The results demonstrate how a machine learning approach can more effectively address issues of sparse data with spatial controls or thin cells at fine levels of geography (like census tracts or block groups). Our initial estimates also show that the machine learning method offers a substantial improvement in prediction of single-family sale prices (i.e., a 75% reduction in RMSE, on average) and great potential for further applications in constructing aggregate measures of land value beyond the cases we pilot here.

ABSTRACT. I discuss recent and on-going research projects at BEA aimed at improving, expanding, modernizing and exploring BEA regional statistics. Feedback from data users is welcome.

ABSTRACT. In this paper, we introduce the bivariate penalized spline estimation over triangulation (BPST) method to investigate spatial non-stationarity in the operational expenditure of U.S. county governments. Unlike some of the existing methods, this method does not impose spatial interaction as the source of spatial non-stationarity and therefore is more general for the exploration of spatial (non-)stationarity. As an application, this method is used to identify how local wage, population, health and crime affects county administration costs differently over space. Then, we explore some contributors to the observed spatial non-stationarity in the data.

ABSTRACT. Nearly half of farmland in the United States is rented to other operators. Counties periodically adjust tax rates on farmland, which translates into unintended effects of higher tax rates on rented land. Landowners subsequently pass on the additional burden to renters as a tax incidence. Previous literature estimates tax incidences for cropland in the Midwest, but the tax incidence on pastureland and cropland in the Southern Great Plains has not been studied. This study calculates the tax incidence for rented cropland and pastureland in Oklahoma using dynamic spatial panel regression with county-level data spanning from 2008 to 2019. The analysis estimates how much of the tax burden is shifted from landowners to renters by increasing rental rates of farmland using a Ricardian model of land rental rates. We hypothesize that the tax incidence will be higher (more elastic) for ranted cropland relative to pastureland because of soil quality and other land quality characteristics. Implications for new and beginning farmers and others whose farm operation depends on renting land is consider, in addition to location-specific effects of the tax incidence on producer and consumer welfare.

ABSTRACT. Tax scholars have long recognized that the income tax code is unable to treat all families ``equally" (Dickert-Conlin and Houser, 1998; Whittinger and Alm, Alm & Leguizamon). The relative earnings of spouses is the primary driver of this differential, with one-earner married couples seeing the largest "marriage bonus" and equal earning couples seeing the largest "marriage tax". While this differential treatment has received consideration in the literature, less attention has been paid to the disparate impact of this differential treatment that may be experienced by Black families in particular. Although the tax code is race-blind by design, the effects of these structures may disproportionately penalize Black households, primarily because Black households are more likely to have relative spousal earning structures that increase the likelihood that these households will experience a marriage penalty relative to white households. To our knowledge, our paper is the first to consider the average marriage tax/bonus experienced by Black households through time. We use detailed individual level data from the Current Population Survey (CPS) to calculate differences in the marriage tax for Black households relative to white households and how this differential varies by region. Implications of this regional variation are discussed.

ABSTRACT. The advent of fully autonomous vehicles (AVs) and shared autonomous vehicles (SAVs) in the market may boost long-distance passenger travels across the United States in the coming years. This paper forecasts their impacts on the frequency, destination, and chosen mode, party size and scheduling of long-distance (over 75-miles one-way) passenger trips within the US. Two national travel surveys are used to derive equations for such choices with and without AVs (and SAVs), using Poisson, negative binomial, zero-inflated negative binomial distributions, multinomial and nested logit models.

Results suggest that a one-standard-deviation (1 SD) increase in variables like the number of licensed drivers (in a household) and home ownership increase a household’s vehicle ownership levels by 39 and 18 percent, respectively. Such 1-SD change in the number of vehicles per adult in the household increases the number of long-distance trips (per survey respondent) by 57.9%. A respondent’s age and education notably increase the likelihood of making long-distance trips for recreational purposes (like shopping, visiting friends and family, and dining). Long-distance trips’ party sizes fall by 21%, on average, when the commute-purpose variable becomes more likely (rising by 1 SD), and by 13.6% when the business trip variable is increased. On the other hand, a one-standard deviation rise in worker density at the destination track increases the party size by 15.5%. Party size is notably bigger on long-distance trips during the summer and fall (relative to spring and winter trips).

AVs and SAVs are more often preferred for such trips by younger and more educated persons, males, drivers, and full-time workers. But air travel still dominates other modes for trips over 500 miles. Destination choice models show that entertainment, retail, health, and education jobs are important contributors to the destination choice of non-work trips. However, the number of industrial, and public administration jobs are important explanatory variables for work-trip destinations.

ABSTRACT. Shared fully-automated vehicles (SAVs) will provide different services in the future, including door-to-door (D2D) service (Childress et al., 2015; Fagnant & Kockelman, 2018; Narayanan et al., 2020), first-mile last-mile (FMLM) connections to transit stations (Farhan et al., 2018; Gurumurthy et al., 2020; Pinto et al., 2020; Shen et al., 2018), and low-cost public transit service (Quarles et al., 2020). This study integrates SAVs’ D2D service, FMLM service, and the SAV-based transit service, and reveals the possible mode shares, fleet performance, and social welfare change for the 5% population sample across the Chicago network. A large-scale multi-agent activity-based travel demand model, POLARIS, was leveraged to simulate the detailed behavior of agents, with novel functions added that focuses on the integrated modeling of multimodal routing and the transfer behavior between SAVs and transit. Since most of the previous transit-related simulations do not optimize the multimodal routing for a mixed-use of SAVs and transit lines, the multimodal routing in this study ensures the best routes are considered by taking the travel time, cost, and number of transfers between different modes into account. The POLARIS model is initialized with a population synthesis module (Auld & Mohammadian, 2010), which includes home, school, and work location choices for synthesized households and individuals. All activities, along with start times and durations, are generated by each agent in the 24-hr period, and further updated to include an activity location and mode. Agents then schedule their travel, incorporating four different travel choices: destination choice, mode choice, departure time choice, and travel party choice (Auld & Mohammadian, 2011; Gurumurthy et al., 2020). A rescheduling model is used to manage conflicts among activity plans and travel schedules, and the trips are loaded onto the network with dynamic traffic assignment. Different FMLM and transit services were simulated for the Greater Chicago region. The baseline scenario is the year 2018 Chicago run using 5% of the total synthesized population, which ended up with 201k households and 520k persons. The business as usual (BAU) case is set up when no SAV services but only taxi service is provided, which has a $3.3 fare for taxi service with a further $1.5 per mile. Based on the BAU case, scenarios involving SAVs were designed to have one additional SAV service each time, so one can see the incremental changes of the new SAV service brought to the whole network. The first change was to use SAV D2D service to replace traditional taxi service across the whole region, the second one added SAV’s FMLM service, and the last one further added SAV-based transit to replace the regular bus service (CTA and PACE bus lines). All the scenario runs simulated a 24-hour weekday, starting from midnight. SAV D2D service accounts for 15% of the mode share under the assumption of $0.50 per mile fare and households’ willingness to relinquish their vehicle for future years. A fleet of 12k SAVs serving 5% of the Chicago population, or 1 SAV every 40 residents, could offer 15-minute service for trips averaging 4.6 miles. Operating for more than 4 hours, on average, each SAV served nearly 20 requests per day. Most SAV riders preferred to use the SAV D2D service for relatively short-distance trips, but the trip could be longer than 50 miles given the large nature of the region. Based on the distribution of the social welfare change, residents in the suburban area benefited most from the SAV D2D service, followed by those in the urban area. When the same SAV fleet offered both D2D service and FMLM service at the same time, the SAV fleet was more utilized, by serving 12% more requests per day per SAV with only a 4% increase in VMT. The transit use was also brought up from 5.4% to 6.3%, with a stable mode split among other modes compared to only using D2D service. The average trip distance of FMLM service was also shorter, most of which were between 1.7 to 1.9 miles. This indicated a prominent expansion of the transit catchment area, from a typical 0.25-mile average walking distance. The spatial patterns of SAV FMLM service also indicated such improvement, as many more boardings were observed in the TAZs along the transit lines (e.g., PACE suburban bus and the METRA commuter rail). Downtown Chicago is also the busy zone for FMLM trips, due to the CTA bus service. FMLM service boarding happened mostly across the day, especially during morning peak and midday. Trips to/from rail lines dominated the FMLM trips, compared to the bus stations, with a ratio of 6:1. When adding the FMLM service, the social welfare does not change much in the downtown area, because of the multiple travel choices. TAZs near transit stations in the suburban areas are more likely to have welfare gain. Lastly, when the SAV-based transit service was added to the scenario, the performance of the on-demand SAV fleet did not change much since the FMLM service mainly focused on connecting to rail. The social welfare change also showed a mixed pattern in both the urban and suburban areas. The reason for this is likely to be riders skipping SAVs due to small-size Abuses and the road congestion in the transit corridor caused by more frequently dispatched SAVs, although some riders enjoyed lower fares and more frequent service. Based on the various service options tested here, SAVs can provide promising integration with future public transportation systems. The low fare D2D service will be key to reducing vehicle ownership, encouraging more shared rides, and gaining social welfare in the suburban area, while the FMLM service can increase transit ridership and catchment area. The SAV-based transit will also offer a cost-efficient service, and the network and fleet performance may be improved through integrations with the on-demand service fleet and new pricing strategies.

ABSTRACT. Shared autonomous vehicles (SAVs) will likely emerge in many urban settings over the coming decade and may significantly impact passenger travel. SAV fleet managers, the public, and policymakers may be attracted to all-electric drivetrains’ lower operating costs and environmental benefits, but will need to account for charging times and range limitations of EV battery packs. This study investigates a variety of potential electric SAV (SAEV) fleet designs and charging strategies from the literature. The agent-based transportation tool, POLARIS, is used to simulate several scenarios serving passenger travel across Illinois’ Greater Chicago region. Results show a mixed fleet of short (100-mi) and long (300-mi) range SAEVs performs better than a homogenous short-range fleet. SAEVs can stay in place longer (1 hr versus 15 min) before charging to keep eVMT low, but only if long-range SAEVs are in the fleet. SAEVs in large regions are exposed to location-specific trip requests when staying in place and need to have high average SoC across the fleet to serve all incoming requests, necessitating careful downtime management. Smart siting of EVCS, availability of fast chargers, and charging during maintenance remain key to keeping response times low.

ABSTRACT. Demand for transportation network companies (TNCs), such as Lyft and Uber, is expected to grow in the upcoming years. Most shared autonomous vehicle (SAV) studies assume that vehicles are allowed to idle in place after completion of a trip, which adds to the congestion by taking up a lane. Some other studies remove SAVs from the network and assign them to fake links while idling, which is not a realistic assumption. Thus, this study focuses on allowing SAVs to park on existing on-street (free and metered) and off-street parking while idling. This capability was applied to POLARIS, which is an agent-based simulation platform. Three parking configurations were tested for the Bloomington network to test the code. In the first scenario, SAVs were allowed to simply idle in their place after drop-offs. In the second scenario, SAVs move to the nearest available parking zone. The third parking search scenario finds a list of the 50 closest spaces and sends the SAV to the parking zone with the best combination of price and distance, including the cost of moving to the parking location and the parking fee.

These parking search strategies were applied to the Bloomington network with an area of 74 sq miles and almost 66,000 population. Random costs and spaces were generated for parking locations on this network, as parking fees and spaces were not available. Therefore, 5 scenarios were tested for this network that differ in the parking search strategy, parking space allocation, and parking fee. The nearest distance parking search strategy was tested with unlimited parking spaces (i.e., nearest distance with unlimited space scenario) and randomly generated spaces between 10 and 20 for all parking locations (i.e., nearest distance with random space scenario). The minimum cost parking search strategy was also implemented for randomly generated costs between $1 and $4 (i.e., random cost scenario) and fixed cost of $2 for parking locations (i.e., fixed cost scenario). Finally, the base scenario was without any parking implementation. The SAV fleet was assumed to be 880 in all scenarios. The results investigate the impacts of parking strategy implementation on the SAV mode share, the percentage of empty VMT, and idle time. In addition, parking time and costs were compared for different parking search and cost/space scenarios. By applying the parking strategy, the SAV mode share increased up to 2.7% in different parking scenarios relative to the base scenario without parking. For example, the nearest distance scenario with unlimited space resulted in 7.67% SAV mode share, while without parking, SAV mode share was 7.47%. In addition, the percentage of empty VMT slightly decreased by parking implementation in this network. The empty VMT for the nearest distance with unlimited and random space scenarios and the fixed cost scenario was 30% while having an average vehicle occupancy of around 1.25. Parking implementation also slightly decreased the average idle time of SAVs. For example, the parking strategy with unlimited space decreased the idle time of SAVs by 1.7%. Finally, the comparison of different parking strategies and cost/space scenarios shows that in the nearest distance parking search strategies, parking costs are much higher than the scenarios with the minimum cost search strategy. For instance, the random cost scenario resulted in $8,553 parking cost for all SAVs, including parking fees and the cost of moving to the parking location, while the nearest distance scenario with random spaces resulted in $21,520 parking costs. Finally, parking allocation for SAVs is expected to decrease VMT and empty VMT in the network compared to a strategy that vehicles move around after completing trips. In addition, the results of this study showed that compared to the scenario without any parking and letting SAVs idle on their drop-off place, which is not possible, the empty VMT still decreased and SAV mode share increased on the Bloomington network. The parking search strategy, which leads to different parking costs for TNCs, can change depending on the preference of the TNCs and the parking availability in the network.

ABSTRACT. The objective of this paper is to estimate the effects of the Colombian 2020 strike on the housing prices in Cali. We use a database of adversitised houses obtained using web scrapping techniques on the website www.fincaraiz.com. We evaluate the impact of proximity to the location of blockages and protests. We apply a difference-in-difference hedonic price model to estimate pre and post impacts(Linden y Rockoff, 2008; Nowak y Sayago, 2019).

ABSTRACT. We use secondary data primarily from the Quarterly Census of Employment and Wages to model the factors that affect the employment of building contractors at a county level. We first present maps showing the distribution of building contractors per 10K household across the US counties using annual data. We then develop regression models – both in levels and changes – to investigate what factors affect the employment of building contractors. Key regressors include median home age, median household income, unemployment rate, the share of bachelor’s degree, among others. Finally, we explore what counties have a shortage or surplus of building contractors from regression residuals. Our results show a strong U-curve effect of median home age on the density of building contractors. We find the turning point is about 60 years, which means – controlling for net migration – that places with older homes have fewer building contractors until the median home age reaches 60, when the effects become positive. We also map the areas with shortages and surpluses of building contractors. For example, San Francisco County in California as expected has a very high shortage, but Storey County in Nevada has a substantial surplus of building contractors.

ABSTRACT. There are multiple definitions of rural areas used in the federal government. The definition used for a particular government program/researcher depends on the various geographies and population, different aspects of rurality in terms of socioeconomic characteristics, and Congressional mandate. Using HUD longitudinal data, we investigate how some of the most commonly used rural definitions could affect the number of HUD-assisted population and HUD funding in each major program spent in rural areas as a consequence. We analyze the differences by definitions, degrees of overlapping areas and rural HUD coverage. Then, we investigate the geographic, demographic, and socioeconomic differences among HUD-assisted rural households by diverse rural definitions. Main implications of defining rural is how it could be used to distribute the resources (or, how to measure the resources spent in the rural areas) and weigh the varying degrees of rural housing challenges. Understanding which definition is best suited for specific purposes could be beneficial to researchers/policymakers.

ABSTRACT. Rural hospital closure remains an important issue, but most predictors focus solely on the financial health of the facility. We use a panel dataset from 2009-2019 to test whether changes in county-level demographics (population, poverty, insurance rates, etc.) can help predict closures in rural hospitals that are classified as financially distressed.

ABSTRACT. When policy discussions center on rural health care, the focus tends to be on rural hospitals, and to a lesser extent access to specific services such as mental health and dentistry. One key element of the rural health milieu that is largely overlooked is access to pharmacies. In this exploratory analysis we compare and contrast the concentration of and ownership nature of pharmacies across U.S. counties annually from 2011 to 2020. Particular attention is paid to differences across the urban-rural spectrum.

ABSTRACT. Rural U.S. populations face particular challenges in terms of maternal and obstetric care. Women living in rural areas have more children than metropolitan women per capita, and report an earlier age at first birth. Unfortunately, rural populations also suffer from elevated infant mortality, maternal mortality, and serious complications. Given this context, policymakers need to understand the demographic, economic, and geographic differences in access to obstetric care. Furthermore, existing literature points to negative outcomes in both health and economic development in areas losing health care facilities This presentation provides a descriptive overview of economic changes underway in counties that lost obstetric care facilities between 2012 and 2019.