PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Transportation Economics & Management Systems, Inc. March 2015
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
TABLE OF CONTENTS 1 2
3
4
5
6
7
8
INTRODUCTION (STUDY PURPOSE AND CORRIDOR) .................................................................................. 1 ECONOMIC ANALYSIS FRAMEWORK ............................................................................................................ 4 2.1 The Character of the Overall Economy .................................................................................................4 2.2 Economic Rent Methodology .................................................................................................................6 THE ECONOMIC EVALUATION DATABASES .............................................................................................. 11 3.1 Introduction ......................................................................................................................................... 11 3.2 Economic Benefits Study Process ........................................................................................................ 11 3.3 Transportation Networks: General Overview ...................................................................................... 12 3.4 Auto Network ....................................................................................................................................... 13 3.5 Transit Network ................................................................................................................................... 15 3.6 Zone System ......................................................................................................................................... 17 3.7 Socioeconomic Database ..................................................................................................................... 18 3.8 Trip Database ...................................................................................................................................... 20 3.9 Conclusion ........................................................................................................................................... 21 THE TRAVEL DEMAND MODEL AND TRAFFIC FORECASTS ...................................................................... 22 4.1 Travel Demand Model ......................................................................................................................... 22 4.2 Socioeconomic Variables ..................................................................................................................... 24 4.3 Demand Model Calibration ................................................................................................................. 24 4.4 Travel Demand Forecast ..................................................................................................................... 25 TEMS ECONOMIC RENT MODEL ............................................................................................................... 26 5.1 Model Calibration ............................................................................................................................... 26 5.2 Assessment of the Impact of Economic Growth ................................................................................... 32 ECONOMIC RENT RESULTS ....................................................................................................................... 36 6.1 Overall Economic Rent Results ........................................................................................................... 36 6.2 Economic Rent for Montgomery, Prince George’s and Washington DC ............................................ 38 INPUT-OUTPUT ANALYSIS: CONSTRUCTION OF THE PURPLE LINE ......................................................... 43 7.1 Scope of Evaluation ............................................................................................................................. 43 7.2 Input-Output Methodology .................................................................................................................. 43 7.3 Application of Regional Input-Output Modeling System ..................................................................... 44 7.4 Results .................................................................................................................................................. 48 CONCLUSION ............................................................................................................................................... 60
APPENDICES
APPENDIX A: DATA SOURCES AND METHODS FOR RIMS II ................................................................................ 61 APPENDIX B: STATED PREFERENCE SURVEY ........................................................................................................... 66 REFERENCES
........................................................................................................................................................ 80
Transportation Economics & Management Systems, Inc.
March 2015
ii
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
1 INTRODUCTION (STUDY PURPOSE AND CORRIDOR) The purpose of this study is to update the Purple Line Economic Analysis developed by Transportation Economics & Management Systems, Inc. (TEMS) in 2010. The aim of Purple Line light rail is to provide better connectivity between Bethesda in Montgomery County and New Carrollton in Prince George’s County. As it is proposed by Maryland Transit Administration (MTA), U.S. Department of Transportation (USDOT) Purple Line ‘would provide a direct connection to the Metrorail Red, Green and Orange Lines; at Bethesda, Silver Spring, College Park, and New Carrollton. The Purple Line would also connect to MARC, Amtrak (Northeast corridor), and local bus services’1. A map showing the proposed Purple Line Preferred Alternative2 along with stations and other metro, MARC and rail lines is shown in Exhibit 1.1. Exhibit 1.1: Purple Line Route and Station Map
The approach of the Economic Impact Study is to measure supply side benefits generated by the Purple Line Project. To do this, two impact analyses have been used by Transportation Economics & Management Systems, Inc. (TEMS) in evaluating the Economic Benefits for the Study. These are as follows: Long-Term Supply Side Impacts: Used to identify the economic impacts of the 16-mile light rail line to the communities in the region, especially Montgomery and Prince George’s Counties. Since the implementation of the Purple Line light rail project calls for the investments by both the public and private sectors it is clear
1
See: http://purplelinemd.com/about-the-project for details.
2
According to MTA, The Preferred Alternative is a refined version of the Locally Preferred Alternative ( LPA) identified in AA/DEIS 2009 (see http://www.purplelinemd.com/images/studies_reports/feis/volume_01/013_PL%20FEIS_Vol-I_Ch%209%20Evaluation-ofAlts.pdf for more information).
Transportation Economics & Management Systems, Inc.
March 2015
1
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE that a project needs to be fully evaluated in economic terms, which includes the economic impact of the project on the local economy. Typically, federal and state governments have required a demand-side analysis of the economic benefits of a project which may be characterized as traveler or user benefits that measure time, cost and energy savings, and also external resource savings associated with making trips in the study region. However, of real interest to policy makers and individuals who live in a corridor are the supply side benefits of a project which may be characterized as community benefits or productivity benefits such as increased jobs, income and wealth that result from implementing the project. Economic Rent Analysis of supply side community economic benefits can be estimated using the TEMS RENTS™ model. This methodology is fully documented in the academic literature of transportation economics [1], [2], [3], and [4], and it is well established that while Consumer Surplus measures benefits from the demand side, Economic Rent measures benefits from the supply side. Because the economy is circular and demand must equal supply for the economy to be in equilibrium, Economic Rent analysis, is the “mirror image” of consumer surplus. Economic Rent is however a less well-used methodology. This is because it is more difficult to measure economic rent than to measure consumer surplus. The work on specific measurement methodologies for economic rent has only been conducted in the last twenty years. This reflects the growth of computer power and the ability of modern computers to handle the large number of calculations associated with conducting an Economic Rent Analysis. As documented in the literature [5] through [6] the initial work on economic rent grew out of urban economics and in particular the measurement of property prices and commuting activity. This work was later supplemented by the development of transportation analysis methodologies that greatly enriched the economic rent measurement process. This included transportation access measurement (by measuring transportation utility) and traffic movement databases (showing market interaction) that are so critical to economic rent [7]. The final formulation of economic rent methodology required the inclusion of the Economic Theory of Location and, specifically, Central Place Theory [8], [9] to provide a structure of “markets” to which the general economic rent proposition could be applied. This then provided an effective application method. Essentially the economic rent measures the improvement in income, employment and property value that individuals receive as a result of a transport investment (improved accessibility) which allows the market to work more efficiently and to raise individuals’ productivity. Temporary Construction Impacts: The construction impacts of the Purple Line light rail will constitute a significant investment in the region, especially Montgomery and Prince George’s Counties by the Federal Transit Administration (FTA). This Federal investment would comprise a major transfer payment to the region that would significantly increase total spending within the Montgomery and Prince George’s Counties economy. The spending of federal dollars cannot be expressed as a benefit to the U.S. economy, since the investment might well have been made elsewhere, rather than in the Purple Line project. However, it is clear that such an investment choice on the part of the Federal government will have a significant economic impact on the local regional economy. To estimate the economic impact of the additional federal construction spending in the region, especially Montgomery and Prince George’s Counties, an analysis was performed using the Bureau of Commerce, BEA, Regional Input-Output Modeling System II (RIMS II) economic model. The following report updates these analyses to a 2015 base, and applies the update to the Purple Line Impact Study. The update is applied to databases, methodology and the results of the study. The report includes the following: Chapter 1: Introduction - The study purpose and corridor Chapter 2: Economic Analysis Framework – This includes a brief assessment of the overall economic framework and the relationship between Consumer Surplus and Economic Rent. 2.1: The Character of the Overall Economy in the Purple Line corridor. Transportation Economics & Management Systems, Inc.
March 2015
2
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE 2.2: Economic Rent Methodology – covers theoretical and technical issues of the developed methodology – Economic Rent (evaluating the supply side of the Study). Chapter 3: The Economic Evaluation Databases – This describes the process of developing socioeconomic and transportation databases, as well as different methodologies necessary to perform both parts of economic benefits evaluation study. Chapter 4: The Travel Demand Model and Traffic Forecasts describe the model developed by TEMS for this study. The chapter also shows how the model was calibrated for this study, illustrates the forecasting process, and the forecast results that were obtained. Chapter 5: TEMS Economic Rent Model Chapter 6: Economic RENT™ Results Chapters 5 through 6 present the economic benefits associated with the Purple Line corridor and were evaluated by the Economic Rent model. Economic evaluation results for Montgomery and Prince George’s Counties that especially benefit from the Purple Line project were analyzed. Transfer payments in terms of tax benefits are particularly discussed here. Chapter 7: Input-Output Analysis: Construction of the Purple Line – This Chapter presents an InputOutput analysis to identify the transfer payment benefits of a major investment like the Purple Line Preferred Alternative (cost $1.9 billion) on the economy in terms of temporary construction and permanent operating jobs. Chapter 8: Conclusion: This chapter describes the overall benefits to the study region of the Purple Line project. Appendix A: Data Sources and Methods for RIMS II – This discusses an Input-Output (I-O) Matrix which is an accounting framework that mainly underpins the RIMS II methodology. Appendix B: Stated Preference Survey – This describes the methodology, the purpose, the process and the results of the Stated Preference Survey undertaken by TEMS in 2010.
Transportation Economics & Management Systems, Inc.
March 2015
3
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
2 ECONOMIC ANALYSIS FRAMEWORK In order to estimate the economic impact of the Purple Line project, it is important to understand the character of the different economic benefits to be quantified. These benefits arise from developing and operating the system and have a substantial impact on the productivity of the local and state economy.
2.1 THE CHARACTER OF THE OVERALL ECONOMY A model of the economy [10] shows that an economy is circular in character, with two equal sides (see Exhibit 2.1). On one side of the economy is the consumer side – the market for goods and services – in which consumers buy goods and services by spending the income earned by working for a commercial enterprise. For example, a transportation investment improves individuals travel times and costs and, therefore, increases consumer (traveler) surplus (travel and money savings). An analysis of the impact of a transportation investment in the market for goods and services quantifies the level of consumer surplus generated by a project, by showing how much money and travels individuals save because a given project (i.e., a transportation improvement) reduces their cost of travel, or makes their travel more efficient. Exhibit 2.1: Simple Model of the Economy Revenue
Spending
Markets for Goods and Services Firms sell Households buy
Goods and Services Sold
Inv es t m ent
FIRMS Produce and sell
HOUSEHOLDS Buy and consume
goods and services Hire and use factors of production
Inputs for Production
Wages, Rent and Profit
Goods and Services Bought
Markets for Factors of Production
Firms buy Households sell
Sav ings
goods and services Own and sell factors of production
Labor, Land and Capital
Income
The notion that a transportation project can be worthwhile if travel is made more cost effective is based on the idea that not only the cost, but also the time of a trip, has value. This axiom is agreed to by most transportation companies and by business travelers as well as by both academia and important transportation authorities such as USDOT Federal Highway Administration (FHWA)3 and Maryland State Transit Administration (MTA). In addition, academic and empirical research has shown that this concept holds true for commuters and social, recreational travelers as well [11]. Considerable research has been carried out to both identify the theoretical justification for value of travel time and to quantify its value.
3
See: http://ops.fhwa.dot.gov/tolling_pricing/value_pricing/tools/truce_st_devpfrom3_0.htm
Transportation Economics & Management Systems, Inc.
March 2015
4
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE On the other side of the economy is the market for factors of production. Most importantly, it is the market for land, labor and capital, which individuals provide to firms in exchange for wages, rent and profit. From the perspective of policy makers and the local community, this side of the economy is very interesting as it shows how investment in a new transportation infrastructure changes the efficiency of the economy and how the investment increases transportation efficiency, creates new jobs, opportunities, and, therefore, increases income and wealth and expands the tax base. One of the most important aspects of the circular economy model is that it shows that any project has two impacts, one in the consumer market – the benefits to travelers; the second, in the factor markets or supply side of the economy – the benefit to the community in terms of improved welfare such as increases in jobs, income and wealth (see Exhibit 2.2). Exhibit 2.2: Relation between Consumer Surplus and Economic Rent in the Economy Reduced Transportation Prices Increased Consumer ConsumerSurplus Surplus
Demand
Supply
Shoes
Shoes
Prices on Goods Market
Housing Transportation
Housing Transportation Transportatio n
Demand Market
Consumer
Service Retail
Agriculture Mining
Household
Supply Market
Owner of Assets Labor Land Capital Goods
Supply
Investment
Prices on Factor Markets (Wages, Rent, Interest)
Manufacturing Labor Exports Land
Imports
Capital Goods
Demand
Reduced Transportation Costs
Increased Economic Rent Increased
For the economy to reach equilibrium, both sets of benefits must be realized. As such, the benefits of a project are realized twice, once on the demand side and once on the supply side. As a result, there are two ways to measure the productivity benefits of a transportation project and, theoretically, both measurements must equal each other [12]. This is a very useful property since in specific analysis one can be used to check the other, at least at the aggregate level. This is very helpful and provides a check on the reasonableness of the estimates of project benefits. However, in assessing the benefits of a transportation project, it is important not to double-count the benefits by adding supply side and demand side benefits together. It must be recognized that these two sets of benefits are simply different ways of viewing the same benefit. The two markets are both reflections of and measure the same thing. For example, if both sets of benefits equal $50 million, the total benefit is only $50 million, but expressed in two different ways: travelers get $50 million of travel benefits
Transportation Economics & Management Systems, Inc.
March 2015
5
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE and the community gets $50 million in jobs, income, increased profits. As a ripple effect (or transfer payment) the economy also gets an expanded tax base. Therefore, if a given transportation project is implemented, equivalent productivity benefits will be seen in both the consumer market for goods and services (as the economy benefits from lower travel times and costs), as well as in the supply side factor markets. In the supply side market, improved travel efficiency is reflected in more jobs, income and profit. For a given transportation investment, therefore, the same benefit occurs on both sides of the economy. In the consumer markets, users enjoy lower travel costs and faster travel times. On the supply side of the economy, the factor markets take advantage of the greater efficiency in transportation. As a result, both sides of the economy move to a new level of productivity in which both sides of the economy are balanced in equilibrium. Improved efficiency will generate supply-side spending and productivity benefits that have a very real impact on the performance of the local economy. The method that develops estimates of productivity jobs and wealth creation is an Economic Rent Methodology. It measures how the performance of a new transportation facility raises the efficiency of the economy. This efficiency improvement creates jobs and income, and raises local property values to reflect the improved desirability of living or working in the area. Specific methods for applying Economic Rent economic theory will be identified and appropriate measurement methodologies will be developed. In particular, the issue of measuring the quality of the transportation system will be addressed.
2.2 ECONOMIC RENT METHODOLOGY The concept of Economic Rent is derived from basic Ricardian economic theory and provides a means of explaining the increased value of economic resources (land, labor and capital) and their change in value in different circumstances or market conditions. For a transportation investment, accessibility is a key spatial variable that affects the likely uses of economic resources and, therefore, their value. Changes in accessibility result in changes in the economic rent (or monetary value) that economic resources can command and, therefore, the value and character of the economic activities that take place at any location. As a result, for important economic welfare criteria (such as employment, household income, and property values) an evaluation can be made of the likely change in economic rent that will be associated with an improvement in accessibility generated by a given transportation investment [13]. Economic Rent may be defined as the difference between what the factors, or productive services, of a resource-owner earn in their current occupation and the minimum sum he/she is willing to accept to stay there. It is then a measure of the resource-owner’s gain from having the opportunity of placing his factors in the chosen occupation at the existing factor price, given the prices his factors would earn in all other occupations. It is the proper counterpart of consumers’ surplus when this is regarded as the consumers’ gain from having the opportunity of buying a particular good at the existing price, where all other prices are given. And like a change in the consumers’ surplus, it is a measure of the change of his welfare when the relevant prices in the market are altered. Whereas the increase of consumers’ surplus is a measure of his welfare gain for a fall in one or more product prices, the increase in that person’s economic rent is a measure of his welfare gain from an increase in the price or the volume of the sale of his factors, i.e., increased sales should generate increased profit. Typically, the level of economic rent can be calculated as follows: Economic Rent (ER) = f (Pt, It, Et, Ct, Tt) Where: Pt is a measure of Population structure of an area in year t; Transportation Economics & Management Systems, Inc.
March 2015
6
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE It is a measure of Industrial structure of an area in year t; Et is a measure of Education level of an area in year t; Ct is a measure of Cultural characteristics of an area in year t; Tt is a measure of Transportation efficiency of an area in year t. Any analysis region (area) has its own ‘Economic Rent Profile’. Economic rent profile shows the spatial distribution of the economy in terms of key factors such as income, property value and wealth. Each of the characteristics listed above can have a significant impact on economic rent profile. Population Structure: The population structure can affect the economic potential of an area positively or negatively. For example, an aging population could have a negative effect on the economy as the number of workers in the work force may fall. This can reduce productivity and, as a result, reduce the economic rent profile. The U.S. might experience this problem in the second quarter of this century as baby boomers age if technology improvements and increased output do not raise productivity sufficiently. Typically, the more productive the adult population of an area is, the higher the economic rent profile. Industrial Structure: The nature of the industrial structure and resource base defines the potential economic rent profile of an area, e.g., manufacturing, commercial, agricultural, residential, and service industry. The higher the value added by industry, the higher the area’s economic rent profile. For example, the “new economy” jobs in biotech, computers and finance all have very high incomes and economic rent profiles associated with them. Education Level: Educational levels can have a dramatic impact on economic rent potential of an area. Typically, a higher education level (especially Ph.D.’s or other high degrees) will increase the wealth generated by the population. The Baltimore-Washington region, for example, boasts one of the highest concentrations of Ph.D.’s in the U.S., which supports the growth of high tech industry in the region4. Cultural Characteristics: Differences in cultural, ethnic and other social characteristics of an area can impact its economic potential. For example, cultural belief systems can impact the ability of a population to work at certain jobs or in a certain way and, therefore, the level of economic rent that can be attained. A survey by the United Nations of the economic growth potential of Arab countries found that the low level of freedom, limited Internet use and the absence of women in the workforce have had a marked negative impact on economic productivity [14] Transportation Efficiency: Transportation efficiency can greatly affect the economic potential of an area. The more effective a transportation system in moving people and goods, the greater its ability to generate wealth if the economy is responsive to the opportunity presented. It is no coincidence that most of the U.S.’s large east coast cities grew as ‘break of bulk’ ports at locations that had good harbors and good routes to the interior resources and markets of the U.S. Since the quality of a transportation system is a management variable and can be changed in the short term, investment in the transportation system can generate economic development if the investment is made in a growing and vibrant economy. The level of response that the economy will have to a transportation investment is measured by the economic rent profile.
4
According to the data assembled by the Metropolitan Washington Council of Governments (source: US Census Bureau), 20.6 percent of individuals over the age of 25 residing in the Baltimore-Washington region have a graduate or professional degree. This is well above the national average of 8.9 percent.
Transportation Economics & Management Systems, Inc.
March 2015
7
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Where it is important to recognize that education, population, industry, structure, and culture can change over time changing the economic rent profile. However, these are not characteristics that typically change rapidly. Only if an area experiences a significant dislocation or migration associated with rapid and dramatic population and industrial base shifts will it experience a radical change in its economic rent profile. For example, the influx of Hong Kong residents to Vancouver, Canada, in the 1980s dramatically changed the economic rent profile of several areas of the city’s downtown. The effect was largely due to the wealth and “entrepreneurial” capability of this new population. In the United States one of the issues for the Midwest is the fact that while it has some of the country’s leading academic institutes, it is still losing much of this talent because it is not developing the New Economy businesses at a sufficient rate. In the absence of a major dislocation, we can assume that the economic rent factors It, Et, Pt, and Ct will remain largely unchanged in the short term (10-20 years). However, transportation efficiency can change significantly in the “short term.” Major transportation infrastructure projects can dramatically change the accessibility of markets and the opportunity for economic growth. This can apply to the measurement of goods in a manufacturing-dominated economy or to the movement of people in a service industrydominated economy. The economic rent generated by transportation improvements (Tt) has driven the desire to move people more quickly and cost-effectively over time. As a result, if population, industrial structure, education levels, and cultural characteristics remain constant, the Economic Rent (ER) model reduces to: ER = f(Tt)
By using socioeconomic variables (SEi) as a proxy for economic welfare and generalized cost (GCi) as a specific metric for transportation efficiency measured in terms of time and cost the economic rent equation can be rewritten as: SEi = ß oGCi 1
(3)
Where: SEi – Economic rent factors – i.e., socioeconomic measures, such as: employment, income, property value of zone i; GCi - Weighted generalized cost of auto travel for all purposes from (to) zone i to (from) other zones in the study area; ß o and 1 - Calibration parameters. The resulting curve generated by this function is the economic rent profile for transportation accessibility. The generalized cost of auto travel includes all aspects of travel time (in-vehicle time), travel cost (tolls, fuel costs and parking charges). The generalized cost of travel is typically defined in travel time rather than dollars. Costs are converted to time by applying appropriate conversion factors. The generalized cost of auto travel between zones i and j for purpose p is calculated as follows:
GCijp TTijp
TCijp VOTp
Transportation Economics & Management Systems, Inc.
(4)
March 2015
8
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Where: TTijp
=
Travel time between zones i and j for purpose p5;
TCijp
=
Travel cost between zones i and j and purpose p (toll and operating costs for auto);
VOTp
=
Value of Time for purpose p
The Economic Rent theory builds from the findings in Urban Economics, and Economics of Location that support Central Place Theory6. Central Place Theory argues that in normal circumstances places that are closer to the “center” have a higher value or economic rent. This can be expressed in economic terms, particularly jobs, income, and property value. There is a relationship between economic rent factors (as represented by employment, income, and property value) and impedance to travel to market centers (as measured by utility or generalized cost). As a result, lower generalized costs associated with a transport system improvement lead to greater transportation efficiencies, and increased accessibility. This, in turn, results in lower business costs/higher productivity and, consequently, in an increase in economic rent. This is represented by moving from point A to point B in Exhibit 2.3. Exhibit 2.3: Economic Rent Illustration
5
Issues of travel time calculation, including the weighting factor for travel time is broadly discussed in the literature. See, for example: [15] and [16]. 6
See: [7] and [8].
Transportation Economics & Management Systems, Inc.
March 2015
9
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE It should be noted that the shape of the economic rent curve reflects the responsiveness of the economy to an improvement in accessibility. Large cities typically have very steep curves, which indicate more significant economic impacts due to a transportation improvement; smaller communities have less steep curves, and rural areas have very flat curves that indicate lower economic responsiveness (see Exhibit 2.4). Exhibit 2.4: Types of Economic Rent Curve Large Economic Rent
Economic Rent Factors (e.g. (e.g., Income, Property Value)
Low Economic Rent
Little Economic Rent
Increasingly Poor Accessibility Urban
Transportation Economics & Management Systems, Inc.
Rural
March 2015
10
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
3 THE ECONOMIC EVALUATION DATABASES 3.1 INTRODUCTION The purpose of the Purple Line Economic Impact Study update is to explore the economic productivity impacts that will result from the Purple Line rail corridor. The tool being developed to estimate these impacts is TEMS Economic Rent model (RENTS™) which generates producer impacts. To meet this need a series of databases calculation processes were developed for the study. The following section outlines the development and calibration process adopted by the study.
3.2 ECONOMIC BENEFITS STUDY PROCESS The modeling and calibration process for the Economic Rent assessment is shown in Exhibit 3.1. This overall process has five main stages as follows: Stage 1: Auto and transit networks, origin-destination and socioeconomic databases are developed in order to provide input to the evaluation tools, so that they can meet the assessment requirements. Those databases are related to a comprehensive zone system that defines specific geographic areas. Stage 2: Stated Preference Surveys were undertaken. The surveys aimed at the better understanding the transportation needs of the local business and providing the data needed to evaluate travel demand functions for transportation demand model. Stage 3: A transportation demand analysis applies the calibrated demand functions in the COMPASS™ travel demand model to provide traffic volumes and the cost of travel (generalized cost) that are used in Economic Rent Model. Stage 4: Economic Rent modeling and supply curve calibration is developed using the RENTS™ model. Stage 5: Economic Rent analysis with producer benefits results and community benefits are generated.
Transportation Economics & Management Systems, Inc.
March 2015
11
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 3.1: Economic Impact Study - Modeling and Calibration Process
Auto/Transit Transportation Networks
Stated Preference Surveys
Origin Destination Travel Database
COMPASS™ Demand Model Calibration
Travel Demand Model Run
Community & Community Station Benefits Benefits Results Results
Socioeconomic Database
Base Year Trip Matrix
Forecast Year Trip Matrices
Producer Benefits Results
Zoning System
RENTS™ Database Development
RENTS™ Model Calibration
Economic Rent Analysis
3.3 TRANSPORTATION NETWORKS: GENERAL OVERVIEW Industry-standard procedures were implemented to develop the road networks that represented travelers’ route choices. The current transportation networks and future networks for the study area were developed based on the latest transportation networks provided by Parsons Brinckerhoff, Inc. (PB) and the two regional Metropolitan Planning Organizations – Metropolitan Washington Council of Governments (MWCOG) and Baltimore Metro Council (BMC). The networks for the study region were developed based on Transportation Analysis Zones (TAZ) system either from MWCOG or BMC, depending on the area. The networks were constructed using a node numbering system to reflect the relevant auto and transit travel segments within the study area. The segments were coded using network-mapping software to obtain a latitude/longitude location mapping and segment distances for all relevant highways and arterial roads. Equivalency files developed for the PB and MPOs network attributes were used to extract the segment attributes for the aggregated road link types in the system. A two-mode network system was developed for the Purple Line study. Auto and transit networks that are parts of this system are described in more detail on the following page.
Transportation Economics & Management Systems, Inc.
March 2015
12
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
3.4 AUTO NETWORK The 2015 network database developed based on PB, MWCOG and BMC contained latitude-longitude coordinates, node-to-node coded links, distance, lane numbers, speed codes, and capacity codes. The base and future network maps were provided to highlight all major highways segments within the study region of MWCOG. An aggregate equivalency file was developed to transform the urban networks into an interurban context that more closely reflected the purpose of this study. This equivalency file also allowed the filtering of urban links that would not be relevant to the study. Exhibit 3.2 illustrates the major segments that were included in the network. Exhibit 3.2: Major MWCOG Highway Links Coded within Network Highway Description
Segment Description
Interstate-95
Fredericksburg to I-195
Interstate-495
Entire Beltway
Interstate-66
Linden to Downtown Washington DC
Interstate-395
Entire Highway
Interstate-270
I-495 to Frederick, MD
Interstate-295
I-495 South of Downtown Washington DC to I-195
Interstate-70
Frederick, VA to I-695 Beltway
SR 267
I-495 to Leesburg, VA
George Washington Parkway
Entire Parkway
US 29
Downtown Washington DC to I-70
US 1
Downtown Washington DC to I-495 Beltway
US 301
US 50 to Waldorf, MD
US 50
Downtown Washington DC to Annapolis, MD
The BMC highway network was developed for specified links within its study area. The BMC network files contained latitude-longitude coordinates, node-to-node coded links, distance, lane numbers, speed codes, and capacity codes. An aggregate equivalency file was built to filter relevant links within the study area. The BMC network attributes were applied to highway links within its modeling jurisdiction. Exhibit 3.3 illustrates the major segments from the BMC network that were used in the TEMS Purple Line network. Exhibit 3.3: Major BMC Highway Links Coded within Network Highway Description
Segment Description
Interstate-95
I-495 to Aberdeen, MD
Interstate-695
Entire Beltway
Interstate-895
Entire Highway
Interstate-795
I-695 to Reisterstown, MD
Interstate-83
Downtown Baltimore, MD to Hunt Valley, MD
Interstate-295
I-495 to Downtown Baltimore, MD
Interstate-70
Frederick, MD to I-695 Beltway
Interstate-97
I-695 to US 50
Interstate-195
I-95 to BWI
Transportation Economics & Management Systems, Inc.
March 2015
13
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE MD 100
US 29 to I-97
US 29
I-495 to I-70
US 32
US 29 to I-97
US 40
I-695 to Downtown Baltimore, MD
The highway network was an aggregated amalgamation of the PB, MWCOG and BMC networks plus additional roads in the study area. Exhibit 3.4 shows the highway network developed by TEMS for the Purple Line Study. Exhibit 3.5 shows the detailed network developed for the improvement area under study. The auto operating costs were adjusted using common procedures to reflect the perceived costs that different markets exhibit behaviorally. The auto cost was allocated as 57.5 cents/mile for the business travel purpose based on the 2015 IRS standard mileage reimbursement rate for business travel7. This rate reflects the full vehicular operating cost and includes indirect costs such as depreciation, maintenance, wear and tear, and insurance as well as direct costs such as gasoline. The assumption is that business travelers’ decisions are based on the full cost of travel that reflects the anticipated amount that will be reimbursed for the trip. The non-business markets, however, do not consider all the indirect costs of operating a vehicle when making a decision to drive and typically only consider the direct costs. The non-business auto operating costs are therefore factored by 0.3 to yield an operating cost of 17 cents/mile. This cost better reflects the perceived cost that this market behaviorally responds to. This adjustment is a typical and standard procedure implemented within the modeling process. Exhibit 3.4: Study Region Auto Network
7
http://www.irs.gov/uac/Newsroom/New-Standard-Mileage-Rates-Now-Available;-Business-Rate-to-Rise-in-2015
Transportation Economics & Management Systems, Inc.
March 2015
14
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
3.5 TRANSIT NETWORK Base (2015) transit network was developed using mainly PB, WMATA and MTA networks. PB provided information on proposed station and links of the transit services. Since PB network base year was 2005, such information as fare and frequency was updated using WMATA data for the year 2015. In order to develop the networks outside PB or WMATA study regions (mainly in Baltimore metro and surrounding areas) the information from MTA and BMC was used. The transit services covered in the transit networks included the Metro, MARC, light rail, and local bus. Exhibit 3.5 provides the details of transit services coded in the network. The transit network developed for the study is presented in Exhibit 3.6. In the future transit network, the Purple Line Preferred Alternative was incorporated into the base transit network. The Purple Line Preferred Alternative is principally the medium investment light rail transit alternative of the AA/DEIS with several features taken from the high investment alternative, and several other refinements. The total length of the Preferred Alternative is 16.2 mile and total travel time is 60/63 minutes (60 minutes during off peak hours and 63 minutes during peak hours). Exhibit 3.7 provides information on distance, travel time and average speed between stations of Preferred Alternative. All this information has been coded to the future network. Exhibit 3.8 shows the Purple Line corridor in the future transit network.
Exhibit 3.5: Study Region Network
Transportation Economics & Management Systems, Inc.
March 2015
15
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Exhibit 3.6 Major Transit Links Coded within Network Transit Services Description
Segment Description
Washington Metro
Red, Green, Blue, Yellow, Orange Line
MARC
Penn Line, Camden Line, Brunswick Line
Washington Metro Bus
WMATA J1, J2, J3, J4, C2, F4, F6, UM Shuttle, etc.
Amtrak VRE
Fredericksburg Line, Manassas Line
Baltimore Transit
Baltimore Metro, Baltimore Light Rail
Exhibit 3.7: Travel Attributes of the Purple Line Preferred Alternative Station-to-Station Distance Time Average (feet) (decimal Speed minutes) (Mph)
Station-to-Station
Bethesda - Connecticut Avenue
7150
2.9
28
Connecticut Avenue - Lyttonsville
7200
2.8
29
Lyttonsville – Woodside
5400
2.2
27
Woodside - Silver Spring Transit Center
2800
1.8
18
Silver Spring Transit Center – Silver Spring Library
2200
3.5
7
Silver Spring Library – Dale Drive
3050
2.5
14
Dale Drive – Manchester Place
2900
2.6
13
Manchester Place – Long Branch
2250
1.8
14
Long Branch – Piney Branch Road
2500
3.8
7
Piney Branch Road – Takoma/Langley Transit Center
3400
2.8
14
Takoma/Langley Transit Center – Riggs Road
3250
2.6
14
Riggs Road – West Campus
7350
3.8
22
West Campus – UM Campus Center
3200
2.6
14
UM Campus Center – East Campus
2750
2.9
11
East Campus – College Park Metro
4800
5.3
10
College Park Metro – M Square
4450
2.7
19
M Square – Riverdale Park
4250
2.8
17
Riverdale Park – Beacon Heights
5150
4.5
13
Beacon Heights – Annapolis Road
6250
2.5
29
Annapolis Road – New Carrollton
5250
5.2
12
Exhibit 3.8: Transit Network with Purple Line Highlighted
New Carrollton Silver Spring Bethesda
College Park
Transportation Economics & Management Systems, Inc.
March 2015
16
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
3.6 ZONE SYSTEM The major step in the development of travel demand model and the economic benefits study was establishing a transportation zoning system. The zone system provides a reasonable representation of the market area where travel would occur between origins and destinations. 299 traffic zones were developed for the Purple Line economic analysis study area (see Exhibit 3.9). The zoning system covers Washington DC, North Virginia, the majority of Maryland counties, as well as Jefferson County of West Virginia. Exhibit 3.9: Purple Line Regional Study Area Zoning System
Transportation Economics & Management Systems, Inc.
March 2015
17
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Zones from 1 to 273 were developed using boundaries of the traffic analysis zones (TAZ) developed by MWCOG8 and with consideration to zones developed by PB. Where it was reasonable for study purposes TAZ were aggregated. For zones from 274 to 293 the zoning system was developed using boundaries of the TAZ by BMC9. Zones from 294 to 299 are developed using county boundaries. In order to establish the highest possible level of accuracy in regional economic analysis zones in Montgomery and Prince George’s Counties, especially in the areas located within 5 miles from the Purple Line rail corridor, were developed without the minimum level of aggregation and often correspond to MWCOG TAZ (see Exhibit 3.10). Exhibit 3.10: Purple Line Regional Study Area Zoning System – Purple Line Corridor Zoom In
3.7 SOCIOECONOMIC DATABASE In the COMPASS™ model forecasting travel demand between the model’s zones require base year estimates and forecasts of three socioeconomic variables – population, employment and household income – for each of the zone in the study area. Socioeconomic database with the data for the base-year (2015) and the forecast years up to 2040 at five-year intervals was developed using the most recent data, as well as the latest economic forecasts. To allow for assessment of the financial and operational feasibility of the system over 25 years, socioeconomic variables were forecasted through the year 2040. For the Metropolitan Washington region, population and employment projections up to the year 2040 were developed using MWCOG Cooperative forecasts by traffic analysis zones (TAZ)10. For the Baltimore region (city and county) as well as Harford County population and employment forecasts up to the year
8
MWCOG - Metropolitan Washington Council of Governments, http://www.mwcog.org
9
BMC - Baltimore Metropolitan Council, http:/www.baltmetro.org
10
Metropolitan Washington Council of Governments. Information Center. Cooperative Forecasting; Population and Household Forecasts by Traffic Analysis Zone. http://www.mwcog.org/publications
Transportation Economics & Management Systems, Inc.
March 2015
18
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE 2040 were developed using BMC Cooperative Forecasts by corresponding transportation analysis zones11. Projections beyond 2035 are based on 2000-2035 trend lines. Income projections for zones were developed using multiple sets of data from the official government and private sources. Among the main datasets are: historic 1969-2008 Bureau of Economic Analysis data on per capita income by county12; data on average household income and households by county and census tract from U.S. Census data and Applied Demographic Solutions; U.S. Census American Community Survey data on aggregate household income and number of households by county13; MWCOG and BMC population and household projections by TAZ. Bureau of Labor Statistics Consumer Price Index (CPI) data was used to bring historical average household income data from current into constant 2014$14. Exhibits 3.11 and 3.12 summarize the base-year and forecast-year socioeconomic data for the Purple Line study area. Exhibit 3.11: Purple Line Region Population, Employment and Households Projections
12 10
Million
8 6 4 2 Population
Employment
Household
0 2015
11
2020
2025
2030
2035
2040
Baltimore Metropolitan Council. Cooperative Forecasts by Jurisdiction. Community profiles http://baltometro.org
12
Local Area Personal Income and Employment, Regional Economic Accounts, Bureau of Economic Analysis, U.S. Department of Commerce, http://www.bea.gov/regional/reis/ 13 14
American FactFinder database, http://factfinder.census.gov and MapPoint demographic database. Bureau of Labor Statistics, http://bls.gov/pub/special.requests, ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt
Transportation Economics & Management Systems, Inc.
March 2015
19
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 3.12: Purple Line Region Average Household Income Projections
160
Thousand 2014$
140 120 100 80 60 40 20 0 2015
2020
2025
2030
2035
2040
It can be seen that the economy will expand with population annual growing by 0.9%, employment growing 1.2% and income growing 1.1%.
3.8 TRIP DATABASE Forecasting the future trends of the travel market within a region hinges upon the ability of the models used to depict the present travel patterns within the system. Capturing traveling characteristics requires proper market segmentation and travel demand volumes for each market segment. The trip databases from PB, MWCOG and BMC are the main data sources for developing the trip database for this study. The trip databases of PB and MWCOG divide trips based on the local urban travel market characteristics:
Home-based work (HBW) Home-based other (HBO) Non-home based (NHB)
Though this classification is used in studying urban auto travel market, it is found that such classification is insufficient to provide an effective interpretation of traveler behavior for the more complex transit travel market. TEMS’ market segmentation/purpose formulation was established to account for the major markets that would likely be affected by changes in the travel conditions in the study area. For example, business travelers are typically more concerned with travel time, regularity and availability of service than personal travelers, shopping and social travelers, who are more concerned with cost. For the purposes of this study, the desired market segmentation included:
Business travel - all who were on work-related business for one trip-end Commuter - all who reported home-to-work trips or vice versa Other (shopping/leisure/social) travelers
Transportation Economics & Management Systems, Inc.
March 2015
20
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Direct equivalency and interpretive methods were used to convert the MPO trip databases into the designated TEMS trip purposes. The business purposes were not explicitly defined within the MPO’s purpose splits since the total non-commuter market contained a mixture of business and non-business markets. The business share and non-business shares were obtained by applying breakout factors to the total non-commuter market trip segments. This breakout process used MPO trip share data, regional trip databases and survey information collected to determine the final distribution shares. The conversion of the urban trip purposes into an inter-urban aggregate trip purpose is illustrated in Exhibit 3.13. Exhibit 3.13: Purpose Split Equivalency TEMS Purposes Business
MPO Equivalents Interpolated share of NHB, HBO
Commuter Other
HBW Interpolated share of NHB and HBO
The conversion of the urban zonal MPO trip files into the inter-urban zonal trip files involved an aggregation process. This aggregation process resulted in the exclusion of internal trips from the intraurban trip file. These internal trips constitute all trips made within the individual aggregate zones and represented short distance trips. The exclusion of short-distance trips will not affect the analysis of study area subject to improvement, because the zones in that area have been developed in a way so that all trips that use the route would be captured. In addition, the exclusion of short-distance trips will make the calibrated travel demand model more appropriate for this study due to the inter-urban nature of it.
3.9 CONCLUSION It was found that the developed socioeconomic and converted transportation databases provided a solid basis for the evaluation of Economic Rent. The use of these two databases will allow an evaluation of both demand and supply sides of the economic benefits of the project.
Transportation Economics & Management Systems, Inc.
March 2015
21
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
4 THE TRAVEL DEMAND MODEL AND TRAFFIC FORECASTS 4.1 TRAVEL DEMAND MODEL The travel demand model used to forecast volume of traffic is based on the COMPASS™ Demand Model that was specifically developed for analyzing the statewide traffic flows in Maryland. The COMPASS™ Demand Model is a flexible forecasting tool that models the decision-making characteristics of travelers resulting from changes in the underlying travel network. The model system combines the various socioeconomic variables, attitudinal parameters, network attributes and origin-destination data, and calibrate them all against observed traveling distributions and traffic volumes. The COMPASS™ Demand Model has two parts: Total Demand Model and Hierarchy Model Split model. The two models are connected interactively. The Total Demand Model provides a mechanism for assessing overall growth in the travel market. The role of the Hierarchical Modal Split Model is to estimate relative modal shares, given the Total Demand Model estimate of the total market.
TOTAL DEMAND MODEL For the Purple Line Regional Economic Analysis Study, the travel demand models were developed and calibrated separately for three trip purposes, i.e., Business, Commuter and Other (shopping/leisure/social). The models were calibrated for 2010 base year data. In applying the models for forecasting, an incremental approach known as the "pivot point" method is used. By applying model growth rates to the base data observations, the "pivot point" method is able to preserve the unique travel flows present in the base data that are not captured by the model variables. Details on how this method is implemented are described below. The Travel Demand Model used in this analysis is
Tijp e
0 p
SE
1 p
e
2 p GCijp
(5)
where: Tijp = Volume of trips between zones i and j for purpose p SEijp = Socioeconomic variables for zones i and j for purpose p determined by population, employment, and average household income of zones i and j GCijp = Generalized cost of travel for zones i and j for purpose p B0p, B1p, B2p = Calibration parameters for purpose p As shown in equation (5), the total number of trips between any two zones for all routes of travel, segmented by trip purpose, is a function of the socioeconomic characteristics of the zones and the generalized cost of the transportation system that exists between the two zones. The generalized cost provides a logical and intuitively sound method of assigning a value to the travel opportunities provided by the overall transportation system. It is a measure of the quality of the transportation system in terms of the times, costs, and level of service provided by all routes for a given trip. The Travel Demand Model equation may be interpreted as meaning that travel between zones will increase as socioeconomic factors such as population and employment rise or as the utility (or quality) of the transportation system is improved by providing new facilities and services that reduce travel times and costs. The Travel Demand Model can then be used to evaluate the effect of changes in both socioeconomic and travel characteristics on the demand for travel by using the incremental “pivot point” method described below. Tijp
f
Tijp
b
SEijp f SE b ijp
1 p
exp( 2 p ( GCijp
Transportation Economics & Management Systems, Inc.
f
GCijp ) ) b
March 2015
(6)
22
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Where: Tfijp = Number of Trips between zones i and j for trip purpose p in forecast year f Tfijp = Number of Trips between zones i and j for trip purpose p in base year b SEfijp = Socioeconomic variables for zones i and j for trip purpose p in forecast year f SEbijp = Socioeconomic variables for zones i and j for trip purpose p in base year b GCfijp = Generalized Cost of the transportation system for zones i to j for trip purpose p in forecast year f GCbijp = Generalized Cost of the transportation system for zones i to j for trip purpose p in base year b
HIERARCHY MODAL SPLIT MODEL The role of the Hierarchical Modal Split Model is to estimate relative modal shares, given the Total Demand Model estimate of the total market. The relative modal shares are derived by comparing the relative levels of service offered by each of the travel modes. The COMPASS™ Hierarchical Modal Split Model uses a nested logit structure, which has been adapted to model the intercity modal choices available in the study area as shown in equation 7. The utility of travel between zones i and j by mode m for trip purpose p is a function of the generalized cost of travel. The coefficients of this utility function are based on the MWCOG model split model15. As shown in Exhibit 4.1, two levels of binary choice are used for auto and transit modes. Pijmp
exp(U ijmp / ) exp(U ijmp / ) exp(U ijnp / )
(7)
where: Pijmp = Percentage of trips between zones i and j by mode m for trip purpose p Uijmp, Uijnp = Utility functions of modes m and n between zones i and j for trip purpose i ρ is called the nesting coefficient
Exhibit 4.1 Hierarchical Modal Split Structure Total Demand
Auto
Transit
15
Based on Figure 6.11 of TPB Travel Forecasting Model, Version 2.3: Specification, Validation, and User’s Guide. http://www.mwcog.org/transportation/activities/models/files/V23_Draft_Jun30_2008_Report.pdf
Transportation Economics & Management Systems, Inc.
March 2015
23
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
4.2 SOCIOECONOMIC VARIABLES The socioeconomic variables in the Travel Demand Model show the impact of economic growth on travel demand. The COMPASS™ Model System, in line with most behavioral modeling systems, uses three variables (population, employment and average household income) to represent the socioeconomic characteristics of a zone (see Chapter 3). In the development of the original gravity models, these terms were customarily referred to as production and attraction terms, since they quantify the production and attraction that specific origins or destinations exert on trip volumes. A more detailed discussion on the incorporation of these terms in the theory of demand modeling can be found in [17] and [18] Denote (i,j) as an origin-destination pair, and let Pi, Ei and Ii, be the population, employment, and average household income of zone i, respectively. These quantities are combined in order to produce socioeconomic variable SEij that is used in the travel demand model. Different combinations were tested in the calibration process and it was found that the most theoretically justifiable and empirically stable relationship consists of the following set of formulae shown in equation (8) through (10): Business:
SEij Ei E j
Commuter:
SEij
Other:
SEij Pi Pj
Ii I j
(8)
2
Pi E j Pj Ei I i I j 2
(9)
2
Ii I j 2
.
(10)
The Business formula in equation (8) consists of a product of employment in the origin zone, employment in the destination zone, and the average household income of the two zones. Since business trips are usually made between places of work, the presence of employment in the formulation is reasonable. The role of income reflects the fact that higher paid employees make more intercity trips than lower paid employees. The Commuter formula in equation (9) consists of the product of population in the origin zone and employment in the destination zone and the product of population in the destination zone and employment in the origin zone for returning commuter trips. The average household income is included for the Commuter trip purpose because higher-income people tend to make more auto trips. The Other (shopping/leisure/social) formula in equation (10) consists of a product of population in the origin zone, population in the destination zone and the average household income of the two zones. Trips for Other purposes encompass many types of trips, but the majority is home-based and thus, greater volumes of trips are expected from zones from higher population, and higher income.
4.3 DEMAND MODEL CALIBRATION For the purpose of extrapolating a statistically valid distribution of trips in relation to demographic characteristics and trip disutility, we have assembled a base year database including the base year trip volume, base year demographic indicators (Population, Employment and Average Household Income), and generalized costs of travel between zones as obtained from TEMS’ generalized cost-based networks. We performed calibrations of the travel demand model defined above in equation (5). In evaluating the validity of the statistical regression of the calibrations, there are two key statistical measures: t-statistics and R2. The t-statistics refers to each coefficient to be estimated, while the R2 refers to the model equation (5). The t-statistics is a measure of the significance of the model coefficients; absolute values above 2 imply that the variable has significant explanatory power in estimating trip levels. The R2 is Transportation Economics & Management Systems, Inc.
March 2015
24
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE a statistical measure of the “goodness of fit” of the model to the data. The higher it is, the better the model equations represent the distribution of data points. Given the large volume of data, and thus degrees of freedom, a value of 0.4 and higher is considered a valid result. The calibration results are shown in Exhibit 4.2. It can be seen that, all t-statistics are large and highly significant, and the R2 is good for the travel demand models calibrated for all trip purposes. Exhibit 4.2: Calibration Results of Base Travel Demand Model 2
Purpose
B0
t-stat
B1
t-stat
B2
t-stat
R
Business
-4.9263
-35
0.4791
69
-0.0561
-127
0.58
Commuter
-19.1426
-206
0.8283
250
-0.0887
-209
0.71
Other
-7.9123
-59
0.3783
81
-0.0542
-110
0.58
These coefficients represent the trip elasticities with respect to the socioeconomic variables and the generalized cost of travel between zones within the study area. From the coefficients associated with socioeconomic variables, it can be seen that 1% growth in socioeconomic variables leads to 0.3% ~ 0.7% increase in travel demand based on trip purpose. For the coefficients associated with generalized cost, 1% improvement in generalized cost leads to 0.05% ~ 0.08% increase in travel demand based on trip purpose. These elasticities are in line with travel demand models developed earlier for the travel market of Maryland.
4.4 TRAVEL DEMAND FORECAST The calibrated travel demand models shown in Exhibit 4.1 were then used in travel demand forecast with future year socioeconomic variables and generalized cost network updates by using the incremental method described in equation (6). The travel demand forecast was performed for the existing Purple Line infrastructure (‘No Build’) and the Preferred Alternative (PA) –the medium investment light rail transit alternative of the AA/DEIS with several features taken from the high investment alternative, and several other refinements. PA is a 16-mile light rail line that would extend from Bethesda in Montgomery County to New Carrollton in Prince George's County, providing a fast and reliable transit service for people moving east-west in the corridor. It will link both branches of the Washington Metrorail Red Line at Bethesda and Silver Spring, the Green Line at College Park, and the Orange Line at New Carrollton. The PA would also connect to all three MARC Train lines, Amtrak, and local bus services. New generalized costs of PA were computed from the updated network for each purpose, and then they were used to forecast future travel demand in the study area together with projected socioeconomic variables. Exhibit 4.3 shows year 2030 trips of the corridor by each travel purpose in the study area under the PA option. Exhibit 4.3: Year 2030 Trip Forecasts Purpose
2030 Trips
% of Total Trips
Business
10,177
15%
Commuter
16,302
24%
Other
40,512
60%
Total
66,990
100%
Transportation Economics & Management Systems, Inc.
March 2015
25
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
5 TEMS ECONOMIC RENT MODEL 5.1 MODEL CALIBRATION In the Purple Line study area, we have three types of trip purposes p: business, commuter, and other. For each zone i of the zone system, the accessibility, measured in generalized cost is estimated as follows:
GCi = ∑∑GCijp * Tijp , i=1,N
(11)
p j
Where: GCijp - generalized cost of travel from zone i to zone j for purpose p; Tijp - number of trips from zone i to zone j for purpose p; N - total number of transportation zones in the network. The economic rent function in equation (3) shown in Section 2.3 can be transformed into a linear function (linear regression model) by applying the natural logarithm16 (ln) to both parts of the original economic rent function: ln (SEi ) = ß o + 1 ln (GCi)
(12)
or simply17: ln (SEi ) = o + 1 ln (GCi)
(13)
Where: SEi - Economic rent factor (socioeconomic variable) of zone i; GCi - Weighted generalized cost of travel for all purposes from (to) zone i to (from) other zones in the zone system (GCi is calculated using formula in equation (11)); o and 1 - Regression coefficients. In the regression equation (13) logarithm of (SEi) is the criterion (dependent) variable, while logarithm (GCi) is the predictor (independent) variable. o and 1 are the coefficients of the regression line (o is the intercept and 1 is the slope). Regression coefficients o and 1 are to be estimated in the regression model.
16
Natural logarithm is a logarithm to base ‘e’ of a given number, where ‘e’ is an irrational constant approximately equal 2.71828183. The natural logarithm of x is written: ln (x) or ln x. See, for example: http://www.mathwords.com/n/natural_logarithm.htm or http://www.themathpage.com/aPreCalc/logarithms.htm 17
o = ln (ß o )
Transportation Economics & Management Systems, Inc.
March 2015
26
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Application of regression analysis to the equation (13) allowed developing the Purple Line Economic Rent Model. In this process we established the mathematical relationship between the measure of accessibility (generalized cost of travel) and the economic rent socioeconomic variables (employment, income density and residential property value density) for each transportation zone. Exhibit 5.1 through 5.3 shows the observed values for natural logarithm (LN) of socioeconomic variable (employment, income density and residential property value density) versus natural logarithm (LN) of generalized costs of travel. The regression line reflects the relationship between socioeconomic indicators in each transportation zone included in the zone system and corresponding generalized costs, calculated using formula in equation (11). By the tight clustering of data points around the regression line, it can be seen that in each case a very strong relationship was identified. In order to identify the strength of the relationships, statistical methods were used to analyze the values of the coefficient of determination (R2) and Student’s t statistics (t). Exhibit 5.1: Employment Density as a Function of Accessibility
LN(Eemployment Density)
12
y = -3.30x + 17.67 R2 = 0.50
10
8
6
4
2
0 2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9
4.1
4.3
LN(Generalized Cost)
Exhibit 5.2: Income Density as a Function of Accessibility
LN(Income Density)
24
y = -3.79x + 30.73 R2 = 0.53
22
20
18
16
14
12
10 2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9
4.1
4.3
LN(Generalized Cost)
Transportation Economics & Management Systems, Inc.
March 2015
27
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 5.3: Residential Property Value Density as a Function of Accessibility
LN(Property Value Desnsity)
26
24
y = -3.94x + 32.93 R2 = 0.51
22
20
18
16
14 2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9
4.1
4.3
LN(Generalized Cost)
The value of the coefficient of determination (R2) shows how much the criterion (dependent) variable is influenced by predictor variable chosen in the study [19]. In other words, the coefficient of determination measures how well the model explains the variability in the dependent variable. As a result, the coefficient of determination illustrates the strength of the relationship between the criterion and predictor variables. Performing ‘t-test’ and calculating Students’ t statistics [19] for both regression coefficients - 0 (the intercept) and 1 (the slope), we analyze how significant the regression coefficients are. Assuming a Normal distribution, a t-statistics that equals two in absolute value is generally accepted as statistically significant. Regression statistics for each of the three socioeconomic indicators used in the model, as well as statistical measures of confidence are presented in Exhibit 5.4. Exhibit 5.4: Economic Rent Coefficients for Employment, Income Density and Residential Property Value Density
Intercept (βo)
Tstatistic s for βo
Slope (β1)
Tstatistics for β1
Coefficient of Determination – ‘R square’ 2 (R )
Employment
17.68
24.22
-3.30
-14.63
0.50
Household Income Density
30.73
45.36
-3.79
-18.15
0.53
Residential Property Value Density
32.93
44.83
-3.94
-17.40
0.51
Economic Rent Factor
It can be seen that the calibration was successful and regression coefficients in each equation were shown to be significant. This proves that the economic rent profiles are well developed for the Purple Line corridor. Each equation has highly significant ‘t’ values and coefficients of determination (R2). This reflects the strength of the relationship and, given the fact that there is a strong basis for the relationship, shows
Transportation Economics & Management Systems, Inc.
March 2015
28
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE firstly that the socioeconomic variables selected provide a reasonable representation of economic rent, and, secondly, that generalized cost is an effective measure of market accessibility. Given the performance of the models, the next step in developing the Economic Rent model is to determine the change in socioeconomic indicators as a result of accessibility improvement. In order to calculate change in economic rent factors, we differentiate the economic rent function in equation (13) with respect to generalized cost. The result of such differentiation is present in equations (14) through (16). It is easy to see that slopes 1E, 1I and 1PV in each regression equation represent economic rent elasticities. Each particular elasticity shows how much each economic rent factor changes when generalized cost of travel changes 18.
Empi = Empi = IncDi =
RPvDi
Empi GCi 1E Empi GCi
(14)
IncDi GCi 1I IncDi GCi =
(15)
RPvDi GCi 1pv RPvDi GCi
(16)
Where: GCi
-
Empi, -
Weighted generalized cost of zone i; Employment of zone i;
IncDi
-
Income density of zone i;
RPvDi
-
Residential property value density of zone i;
1E 1I 1pv -
Slope coefficients in regression equations.
It is seen from equations (14) through (16) that the relative absolute change in employment ( Empi ), relative change in household income density ( IncDi ) and relative change in residential property value (
RPvDi ) for each particular zone i equals the relative change in generalized cost GCi multiplied by GCi regression coefficient 1E , 1I or 1PV , respectively.
18
More about the role of elasticity in a measurement of economic rent profile change see: [20], [21].
Transportation Economics & Management Systems, Inc.
March 2015
29
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE The value for each slope coefficient (1) is obtained from the corresponding regression equation. Since area (size - S ) of each transportation zone remains constant, absolute change in household income ( i
Inci
) and residential property value ( RPv ) will be obtained from the following equations: i
Inci
=
RPv i
GCi IncD * IncD i * S i = 1I Si i GCi =
RPvDi * S i 1Pv
GCi RPvDi GCi
(17)
*
(18)
Si
In order to calculate the impact of accessibility improvement on average household income and average residential property value we also had to determine how the improvement in accessibility influences the number of households/housing units that are supported by any given area. To do this we used the Economic Rent model to predict household density/housing units’ density that is supported by any given level of market access. The results of regression analysis are shown in Exhibit 5.5 and 5.6 and economic rent coefficients are given in Exhibit 5.7. Again, it can be seen that good statistical relationships are derived with strong ‘t’ values and coefficients of determination R2. Exhibit 5.5: Household Density as a Function of Accessibility
LN(HouseHold Density)
12
y = -3.83 + 19.39 R2 = 0.53
10
8
6
4
2
0 2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9
4.1
4.3
LN(Generalized Cost)
Transportation Economics & Management Systems, Inc.
March 2015
30
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 5.6: Housing Units Density as a Function of Accessibility
LN(House Unit Density)
10 9
y = -3.97x + 19.02 R2 = 0.51
8 7 6 5 4 3 2 1 0 2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9
4.1
4.3
LN(Generalized Cost)
Exhibit 5.7: Economic Rent Coefficients for Household and Housing Units Density Economic Rent Factor
Intercept (βo)
T-statistics for βo
Slope (β1)
Tstatistic s for β1
Coefficient of Determination – ‘R square’ 2 (R )
Household Density
23.39
28.04
-3.83
-17.94
0.53
Housing Units Density
19.02
25.81
-3.97
-17.47
0.51
Change in average household income ( AvInci ) in zone i is calculated as follows:
AvInci
Inc i , ( Hhi Hhi )
(19)
And change in average residential property value (AvResPvi ) in zone i was calculated as follows:
RPv i , ( Hui Hui )
(20)
GCi GCi Hu HhDi * S i and Hui 1 * HuDi * S i GCi GCi
(21)
AvResPvi Where:
Hhi 1
Hh
*
ResPvi - the change in residential property value in zone i as a result of accessibility improvement Hhi / Hui - the change in the number of households/housing units in zone i as a result of accessibility improvement
Transportation Economics & Management Systems, Inc.
March 2015
31
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Hhi / Hui - the base number of households / housing units in zone i HhDi / HuDi - the base household density / housing units density in zone i
1Hh / 1Hu - the regression coefficients for household/housing units’ density are obtained from Exhibit 5.7. Total change in property values is calculated by applying the share of residential property in the particular county obtained from the statistical source19 to the total residential property value benefits in this county obtained from the Economic Rent model. The results of the analysis show that a statistically powerful Economic Rent model is developed that reflects the responsiveness of the economy to improved transportation access.
5.2 ASSESSMENT OF THE IMPACT OF ECONOMIC GROWTH A key assumption in the Economic Rent analysis is the impact of economic growth on the economic rent profile20. Economic growth will cause the economic rent profile to grow as each component that supports the economic rent profile, land, labor and capital becomes more valuable. As the economy expands, labor wages increase, so space becomes more valuable, and assets become more expensive. This increase in factor prices results in a rise in the economic rent profile. If the rise in the economic rent profile is constant as shown in Exhibit 5.8, then the increase in economic rent associated with an improvement in market accessibility (i.e., a reduction from GC1 to GC2) for the region is the same. As a result, in Exhibit 5.8 area A is equal to area B. This means that economic growth will not change the economic rent benefits of the project. This is the assumption made in this study. Exhibit 5.8: Impact of Economic Growth. Type 1 - Constant Profile
19
2007 Assessment Ratios Survey Report. State of Maryland Department of Assessment and Taxation. See: http://www.dat.state.md.us/sdatweb/stats/07rr_rpt.pdf 20
Economic Rent profile as it was defined in Chapter 2 shows the special distribution of the economy in terms of its key factors income, property values and wealth.
Transportation Economics & Management Systems, Inc.
March 2015
32
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Under most economic conditions, however, the growth in economic rent is not the same over the region and the profile will not grow proportionally along its entire length. For example, in Exhibit 5.9 there is a decline in the forecast year economic rent profile at the market center while in the more peripheral areas (surrounding the market center) there is economic growth, i.e., growth occurs in the suburbs, but not the market center. In this environment the forecast year benefits, as measured by area A, are smaller than the base year economic benefits, measured by area B. This would suggest that using the base year economic rent profile would overstate benefits. Exhibit 5.9: Impact of Economic Growth. Type 2 - Decrease in Profile
This type of growth, however, does not occur in normal markets, but rather in markets that suffer economic dislocations. For example, in the United States both Detroit and Buffalo experienced this type of growth impact when their downtown businesses failed in the 1970’s. In Buffalo the issue was the decline of metal industries, while in Detroit it was more related to social demographic pressures. In this case a forecast of economic benefits based on a base year assessment will be an overstatement of the benefit. Certainly if any city market areas along the Purple Line corridor and, particular, the improvement area suffer a major dislocation (such as experienced by Buffalo) during the life of the project, then the forecasts prepared for the Purple Line corridor study would be overstated. Under a normal economic growth situation in which the economy expands for a corridor, the typical impact is for growth to expand much faster at the market center than in the periphery. This reflects the fact that the market center provides the greater opportunities for growth in a normal economy and market. For example, the flood of Hong Kong Chinese into Victoria, Vancouver in the 1990s increased economic growth and income across the city. However, the impact was most severely felt in the city center with the development of new high-rise buildings, restaurants and businesses within the downtown area. This increased the economic profile of the downtown area more that it did in suburban areas. In this case the measurement of economic benefit using the base year economic profile will understate the size of the Transportation Economics & Management Systems, Inc.
March 2015
33
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE benefits to be derived from the project. In this case Area A will be larger than area B (see Exhibit 5.10). Since this is the usual impact of economic growth on a market center, and as our study suggests ongoing long-term economic growth it is likely that using area B to estimate economic rent benefits understates the overall economic benefits to be derived from an Economic Rent analysis. Exhibit 5.10: Impact of Economic Growth. Type 3 - Increase in Profile
As a result, it can be seen in Exhibit 5.11 that there are three conditions that can exist in the forecast year. Exhibit 5.11: Types of Economic Growth
Transportation Economics & Management Systems, Inc.
March 2015
34
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Type 1 has constant growth. This means that base and forecast year impacts along the economic rent are the same, and the base year analysis understates the benefits. Type 2 has negative growth at the market/city center. This typically results from a dislocation to the economy due to a loss of the economic base of the region. If this occurs the economic rent results, particularly in market centers, would be less than those that would be achieved if a base year economic rent profile is used. Using the base year economic rent profile will overstate the benefits. Type 3 has increased positive economic growth at the market center. As a result, the future year benefits are higher than suggested by measuring the economic rent profile in the base year. While Type 3 is the normal situation for a city or market center, we have selected Type 1 as the basis for estimating economic benefits, which we believe is a reasonable and conservative assumption. In most towns a Type 3 environment will generate benefits greater than those estimated in this study. In one or two towns it is possible that a Type 2 conditions could prevail and lower economic benefits would be generated from the project. However, it is worth noting that such a weak performance would not be consistent with the current economic projections for the Purple Line corridor study used in the Purple Line Regional Economic Analysis (see Exhibits 3.11 and 3.12).
Transportation Economics & Management Systems, Inc.
March 2015
35
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
6 ECONOMIC RENT RESULTS 6.1 OVERALL ECONOMIC RENT RESULTS For the Purple Line Economic Impact Study, the Preferred Alternative scenario of 16 mile improvement will create more than a 27 thousand jobs; will increase property value by 12.8 $ billion and the household income is estimated to increase by $2.2 billion. Two Maryland counties – Montgomery and Prince George’s Counties and Washington DC, will obtain most economic benefits from the undertaken transportation improvement21. Overall state, local and federal tax benefits are estimated about $635 million. Exhibit 6.1 represents the overall annual Economic Rent results, while Exhibit 6.2 represents these results over 30-year life of the project. It should be noted that the increase in employment, income and property value for the study region represents a growth in their overall economy of between 0.5 and 0.6 percent on current levels (see Exhibit 6.3). Employment Benefits: 27,183 annual jobs as estimated will be added to the regional economy as a productivity impact of the transportation project. This is a 0.5 percent increase. These jobs derived from the Purple Line Preferred Alternative improvement are productivity jobs and not temporary construction jobs associated with building the project (see Exhibit 6.1). While construction jobs actually result from transfer payments (relocating government money from one area to another), the jobs estimated in the frame of Economic Rent model result from productivity improvement – increase in the regional Economic rent profile and attracting new businesses to the region. Assuming 30-year life of the project, the productivity jobs in the region are estimated to be more than 815,000 person years of work (see Exhibit 6.2). Taking into account that 30-year life of the project is a very conservative estimate of the project’s life (which might be 40 and even 100 years) the real productivity jobs in the area could be doubled. Income Benefits: $2.2 billion a year or more than $43 billion over 30-year life of the project is estimated to be obtained in terms of income by the households in the region (see Exhibits 6.1 and 6.2). This represents a 0.6 percent increase. These income benefits are derived from the increased attractiveness of the region due to the accessibility improvement. Income benefits result from both, the increase in the number of households in the area (by 18,382 households) and the increase in the average household income per household (by over $661 for the study region). Development Potential (Real Property Value Benefits): The Purple Line project is estimated to provide more than $12.8 billion in development potential. This represents an increase of 0.6 percent. The total property value benefits (see Exhibit 6.2) result from the increase of the number of properties in the region as well as increase in the average value of commercial and residential buildings. Transfer Payments (Tax Benefits): Transfer payments play an exceptional role in the overall project evaluation. As shown in Exhibit 6.2, the Purple Line project should bring an expansion of the tax base by $12.4 billion over 30-year life of the project. Since total costs over the life of the project are estimated at the level of $1.9 billion22, the total government spending on the project would be far less than the benefits that the project should return in taxes. As a result the Federal and State government will receive a full payback directly for their investment over and above the benefits to the local communities.
21
Multiple Economic Rent studies performed by TEMS show that the maximum economic benefits are achieved in the radius of 5 miles from the improvement area. (For example, see [22] or [23] for more details). 22
Purple Line Capital Costs Technical Report, Aug., 2013
Transportation Economics & Management Systems, Inc.
March 2015
36
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 6.1: Overall Annual Economic Rent Benefits for Study Region Derived from the Purple Line Preferred Alternative Economic Rent Factor
Benefit
Direct Productivity Benefits: Employment (Number of Productivity Jobs) Income (billion 2014 $)
27,183 2.2
Average Household Income (2014 $)
$661
Average Residential Housing Value (2014 $)
$2,150
Number of Households
18,382
Transfer Payments – Tax Values: Federal Income Tax
23
(million 2014$)
State and Local Income Tax 25
Real Property Tax 26
Sales Tax
24
(million 2014$)
((million 2014$)
(million 2014$)
Total Tax Values (million 2014$)
286 153 90 106 635
Exhibit 6.2 Overall Economic Benefits over 30 years Project Life Benefit Parameter
Discount Rate @3%
Supply Side Benefits: Income Benefits: (Billions 2014 $)
$43
Employment (Thousands of person years of work)
815 $12.8
Residential- Property Value (Billions 2014 $) Transfer Payments - Tax Values (Billions 2014 $): State and Local Income Tax
$3.0
Federal Income Tax
$5.6
Residential Property Tax
$1.8
Sales Tax
$2.1
Total Tax Values
$12.4
23
Federal income tax rate are calculated based on the Tax Foundation document http://www.taxfoundation.org/files/ff231.pdf
24
State and Local income tax benefits are calculated by applying the median of personal income tax rates of Maryland, Virginia, West Virginia and Washington DC State. The state income tax rates are from Federation of Tax Administrators http://www.taxadmin.org/fta/rate/ind_inc.pdf . Local income tax only applied to the counties in Maryland and the rates are from 25 Real Property tax benefits are calculated by applying real property tax rates from Property Taxes on Owner-Occupied Housing by State from Tax Foundation http://www.taxfoundation.org/taxdata/show/24078.html 26 Sales Tax benefits are calculated by applying sales tax rate from Federation of Tax Administrators http://www.taxadmin.org/fta/rate/sales.pdf
Transportation Economics & Management Systems, Inc.
March 2015
37
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 6.3: Economic Benefits as a % to the Current Levels Benefit Parameter
Benefit (Increase from the Current level)
Employment (productivity jobs)
0.5%
Income
0.6%
Residential- Property Value
0.6%
The increase in Economic benefit is less than one year’s growth for the Baltimore-Washington regional overall economy, and close to one year’s growth for Montgomery, Prince George’s counties and DC. This provides a strong enhancement of the corridors economies. In terms of the time scale associated with the presented benefits it is likely that these benefits will begin to be achieved within two to five years after the completion of the Purple Line Preferred Alternative improvement. The benefits will be proportional to the development of the system routes and schedules. It should be noted that the benefits of the system are likely to increase over time in line with the growth in the economy, as the analysis used the base year economic rent profile not the forecast year economic rent profile. Increases in the economic rent profile will significantly expand these results. If the economy grows by 50 percent by 2050, the estimated benefits will at least increase accordingly.
6.2 ECONOMIC RENT FOR MONTGOMERY, PRINCE GEORGE’S AND WASHINGTON DC An important feature of the Purple Line improvement is its role to the communities especially in Montgomery and Prince George’s Counties and Washington DC. Successful implementation of the project will provide a local, regional and national gateway for the communities in these areas. With the introduction of Purple Line,” considerable employment opportunities and development potential will exist. Such activity will generate both commercial and residential development. Industries looking for a home in the Purple Line corridor will see it as a good “seeding” ground for business. To estimate county level results, the results obtained at a zone level were aggregated to county levels. Exhibit 6.4 and Exhibit 6.5 show the estimated benefits for Montgomery and Prince George’s Counties in Maryland and Washington DC – major recipients of such benefits. While Exhibit 6.4 presents the absolute values of estimated benefits, Exhibit 6.5 illustrate the increase in socioeconomic indicators as a percentage to the current levels. As shown in Exhibits 6.4 and Exhibit 6.5 both Montgomery and Prince George’s Counties derive significant benefits from the Purple Line improvements. The zones that include 16-mile improvement area have the major share of the benefits obtained by their counties. In accordance with the Economic Rent model the absolute value of each particular type of benefit primarily depends on the strength of the economy and the level of accessibility improvement resulted from the Purple Line project. Exhibit 6.6 shows the benefits received in these areas on the map.
Transportation Economics & Management Systems, Inc.
March 2015
38
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 6.4: Economic Rent Analysis by Counties – Montgomery, Prince George’s and Washington DC Employment (# Jobs)
Household Income (Millions 2014 $)
Property Value (Millions 2014 $)
Mean Househol d Income (2014 $)
Mean Residential Household Value(2014 $)
Montgomery
10,335
1,066
6,062
2,363
7,742
Prince George’s
4,962
468
2,500
1,548
2,306
Washington DC
7,736
311
2,429
454
4,703
Total
23,033
1,845
10,991
-
-
County
Exhibit 6.5: Economic Rent Benefits by Community – Montgomery, Prince George’s and Washington DC : Increase from the Current Levels (%) Benefit (% to the current levels) County
Employment
Household Income
Property Value
Montgomery
1.9%
2.0%
2.2%
Prince George’s
1.4%
1.4%
1.5%
Washington DC
0.9%
1.1%
1.1%
Exhibit 6.6: Major Recipients of the Economic Benefits in the Purple Line Regional Study Area
Transportation Economics & Management Systems, Inc.
March 2015
39
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE The Exhibit 6.7 shows the employment increase on different categories for Montgomery and Prince George’s Counties and Washington DC (major industries were obtained from NAICS classification27). Exhibit 6.8 shows the tax benefits received by these areas. Exhibit 6.7: Employment Benefits by Areas – Montgomery, Prince George’s and Washington DC Employment Increase by Categories
County
Professional Service
Health Care
Accommodation/Food Service
1,818
1,550
785
Prince George’s
499
634
486
Washington, DC
1,503
1,019
1,005
Total
3,820
3,203
2,275
Montgomery
Exhibit 6.8: Tax Benefits by Areas – Montgomery, Prince George’s and Washington DC Tax Benefits (Millions 2014 $) County
Federal Income Tax
State +Local Income Tax
Property Tax
Sales Tax
Montgomery
133.0
79.4
46.7
52.1
Prince George’s
58.5
34.9
19.3
22.9
Washington, DC
49.5
19.4
10.4
15.2
Total
240.9
133.7
76.4
90.3
Exhibits 6.9 and 6.10 illustrate the comparison between socioeconomic indicators for Montgomery and Prince George’s Counties and Washington DC28. Absolute numbers for population and employment are higher in Montgomery County than Prince George’s County and Washington DC. This explains why the Montgomery County receives the most benefits. The current employment of Washington DC is higher than the Prince George’s County so Washington DC receives more employment benefits, though the increase in employment rate is less than that of Prince George’s County. As it is shown in Exhibit 6.10, the Median household income in Montgomery County is 33 percent higher than that in Prince George’s County and median property value is 65 percent higher than Prince George’s County. This shows that Montgomery County is wealthier than Prince George’s County and this is the reason for it getting higher economic benefits than Prince George’s County.
27
The North American Industry Classification System (NAICS) is the standard used by Federal statistical agencies in classifying business establishments for the purpose of collecting, analyzing, and publishing statistical data related to the U.S. business economy, see for more details: U.S. Census Bureau, http://www.census.gov/eos/www/naics/ 28
US Census Bureau, http://www.census.gov; MWCOG, BMC, US Census Bureau State & County QuickFacts
Transportation Economics & Management Systems, Inc.
March 2015
40
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 6.9: 2015 Population and Employment in Montgomery, Prince George’s and Washington DC
Montgomery Prince George's Washington DC
1,200,000 1,000,000 800,000
600,000 400,000 200,000 Employment
Population
Exhibit 6.10: 2009-2013 Average Median Household Income and Median Value of Owner-occupied Housing Units for Montgomery and Prince George’s Counties
$500,000
Montgomery
$400,000 Prince George's
$300,000 $200,000 $100,000 $0 Median Value of Owner-occupied Housing Units
Transportation Economics & Management Systems, Inc.
Median Household Income
March 2015
41
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE As to the difference in levels of transportation interaction, analysis of auto network shows that improvement in accessibility as a result of the Purple Line improvements is higher for Montgomery County than that of Prince George’s County and Washington DC. With the light rail option for Montgomery County, average relative improvement in accessibility (measured in relative change in weighted generalized costs29) is 1.1 percent, while for Prince George’s County the corresponding improvement is 0.8 percent and DC is 0.6 percent. As we can see from this analysis, difference in the economic benefits estimated for Montgomery and Prince George’s Counties and Washington DC are mainly explained by the socioeconomic characteristics of the counties, and the level of improved connectivity to other areas using Purple Line project.
29
Relative improvement in accessibility in each zone is measured as the fraction between the absolute change in generalized costs (after implementation of the improvements) and the base generalized costs (before their implementation).
Transportation Economics & Management Systems, Inc.
March 2015
42
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
7 INPUT-OUTPUT ANALYSIS: CONSTRUCTION OF THE PURPLE LINE From a regional, state and local perspective, the Purple Line construction program will have significant economic and financial impact in terms of short term job creation, income and output. Regional input-output (I-O) multipliers, which account for inter-industry relationships within specific regions, have proven to be a very potent and useful tool for evaluating economic stimulus of construction expenditures in a region.
7.1 SCOPE OF EVALUATION In order to understand the economic impact of the Purple Line project construction on Montgomery and Prince George’s Counties and Washington DC, an analysis was made of the temporary supply-side benefits that are derived by the analyzed region. This includes an assessment of both the temporary direct and indirect jobs created by construction of the project, as well as income and economic output. An input-output methodology was used to identify the number of temporary jobs, both direct and indirect, that will be created in Montgomery and Prince George’s Counties and Washington DC during each of 5 years of construction of the project. The input-output analysis measures the short-term economic stimuli that are created in this region as a result of the additional construction spending on the project. In our study we assume that Federal government will fund the capital cost of the project. As a result we will measure the influence of construction spending on output, job creation and correspondent income. Although an FHWA cost-benefit analysis treats the capital investment as a cost rather than as a benefit of the project, according to BEA methodology the construction cost creates job and income benefits to Montgomery and Prince George’s Counties and Washington DC, because the money is spent in its states rather than elsewhere.
7.2 INPUT-OUTPUT METHODOLOGY In the 1970’s, the Bureau of Economic Analysis (BEA) developed a method for estimating regional I-O multipliers known as Regional Industrial Multiplier System (RIMS) In the 1980’s, BEA completed an enhancement of RIMS, known as RIMS II, the Regional Input-Output Modeling System30. A second edition of the RIMS II handbook based on more recent data and an improved methodology was issued in 1992. A third edition was made available in 1997. The main underpinning of the RIMS II methodology is an accounting framework known as an I-O matrix, which is discussed in detail in the Appendix A. The I-O matrix is an exhibit that shows the distribution of inputs purchased and outputs sold for each industry. There are two main data sources for the I-O matrix in RIMS II. First is the BEA’s national I-O exhibit, which provides the input and output structure of nearly 500 detailed U.S. industries (in accordance with NAICS codes) and of 20 aggregated industries. Second, is represented by BEA’s regional economic accounts, used to adjust the national I-O exhibit in order to reflect a region’s industrial composition and trading patterns. The fundamental idea behind the input-output model is that of the multiplier effect, whereby new money entering the economy has a ripple effect with spillover benefits for the entire community through direct and indirect impacts. To cite an example, when the government buys $10 billion worth of goods from a major industry, the purchase (notwithstanding the immediate effect of raising employment and profits in that
30
For a detailed discussion on the data sources and methods underlying the use of RIMS II, the Reader is referred to the technical Appendix A.
Transportation Economics & Management Systems, Inc.
March 2015
43
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE industry) has repercussions leading to higher overall incomes, which in turn lead to even higher demand, thereby triggering a positive feedback loop. The total impact on the quantity of goods and services demanded is much larger than the initial impulse felt from higher government spending. The factor by which the initial impulse is multiplied will be determined by the individuals’ marginal propensity to consume: the fraction of extra income that a household consumes rather than saves. Exhibit 7.1, 7.2 and 7.3 illustrate how a single dollar of additional spending on auto production, for example, benefits the plastics, electricity, instruments and rubber industries, among others31. As shown in the Exhibit 7.1, a single dollar spent on auto production translates into 14 cents spent on plastics, 5 cents on electricity, 11 cents on instruments, 7 cents on rubber, 21 cents on local industries, 17 cents in earnings for local employees and 25 cents leakage. In the second ripple shown in Exhibit 7.2, the 14 cents earned by the plastics industry feeds 9 cents to the chemicals industry, 2 cents earnings for local employees and 3 cents of leakage. Similarly, in the third Exhibit 7.3, 21 cents spent on other local industries re-enters the economy in the form of 1 cent for utilities, 5 cents on autos, 4 cents for other local industries, and 4 cents income for local employees and 7 cents leakage. It should be noted that depending upon the type of project and its location, the multiplier effects from the additional investment, jobs, income and workers’ spending decisions would differ. This is because the characteristics of the local economy (i.e., the types of industry present) determines exactly how much extra impact an investment will generate in that region.
7.3 APPLICATION OF REGIONAL INPUT-OUTPUT MODELING SYSTEM The main advantage of RIMS II is that multipliers can be estimated for any region composed of one or more counties, and for any industry or group of industries in the national I-O exhibit. In order to obtain multipliers especially for the Purple Line corridor a description of the region was provided to BEA. The RIMS II multipliers were calculated for the Montgomery and Prince George’s Counties and Washington DC. A systematic analysis of the regional economic benefits that will be accrued from the Purple Line transportation project calls for detailed information about inter-industry relationships not only at County level in general, but in the different parts of the Purple Line corridor region. Using Capital cost of 1.9 billion with construction cost of 1.7 billion (excluding vehicles cost from Capital cost), we calculate the economic impact of construction spending on output, income and job creations in different parts of the Purple Line corridor. Because the analysis is based on cash flows, we can identify only the employment generated during construction period.
31
The given example illustrates the methodology of RIMS II multiplier calculations made by BEA and does not relate to Midwest region.
Transportation Economics & Management Systems, Inc.
March 2015
44
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY Exhibit 7.1: Multiplier Mechanism 1 S tart
Plastics
Electr icity
$0.1 4 of auto industry spending on plastics ree nters: P lastics
$0.14
E lectric ity
$0.05 $0.01
Chemica ls
$0.09
Autos
$0.05
Instruments
$0.11
Rubbe r
$0.07
Other Local Industries
$0.21 $0.04
Local E mployees
Leak age
Chemicals
Au tos $1 of a dditional spendi ng on auto production initiates s pendi ng on:
In str umen ts
R ub ber
O ther L ocal In du str ies $ 0.21 of a uto indus try s pe nding on other loc al industries reenters :
$0.17 $0.02 $0.04 $0.25 $0.03 $0.04
Transportation Economics & Management Systems, Inc.
March 2015
45
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY
Exhibit 7.2: Multiplier Mechanism 2 Start
Plastics $0.14 of auto industry spending on plastics reenters: Plastics
$0.14
Electricity
$0.05 $0.01
Chem icals
$0.09
Autos
$0.05
Instrum ents
$0.11
Rubber
$0.07
O ther Local Industries
$0.21 $0.04
Local Em ployees
Leakage
Electricity
Chem icals
Autos $1 of additional spending on auto production initiates spending on:
Instrum ents
Rubber
Other Local Industries $0.21 of auto industry spending on other local industries reenters:
$0.17 $0.02 $0.04 $0.25 $0.03 $0.04
Transportation Economics & Management Systems, Inc.
March 2015
46
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY Exhibit 7.3: Multiplier Mechanism 3 Start
Plastics $0.14 of auto industry spending on plastics reenters: Plastics
$0.14
Electricity
$0.05 $0.01
Chemicals
$0.09
Autos
$0.05
Instruments
$0.11
Rubber
$0.07
Other Local Industries
$0.21 $0.04
Local Employees
$0.17 $0.02 $0.04
Leakage
$0.25 $0.03 $0.07
Transportation Economics & Management Systems, Inc.
Electricity
Chemicals
Autos $1 of additional spending on auto production initiates spending on:
March 2015
Instruments
Rubber
Other Local Industries $0.21 of auto industry spending on other local industries reenters:
47
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
7.4 RESULTS Using RIMS II output and employment multipliers for annual capital infusion in the construction industry over 5-year construction period we estimate how different industries are expected to benefit from the Purple Line project in terms of output and jobs creation. The capital cost is taken as $1.9 billion with construction cost of $1.7 billion for a five-year period in 2014 dollars32. Following the methodology developed by BEA for RIMS II we calculate economic impacts on output, income and job creation for each stage (year) of construction period for Montgomery and Prince George’s Counties and Washington DC. Exhibits 7.4 through 7.6 show the economic impact on job creations, output, and income by county during the construction period. As seen in Exhibit 7.4 through 7.6, 53% of temporary jobs, 37% of output and 53% of income are created in Montgomery County, 34% of temporary jobs, 35% of output and 33% of income are created in Prince George’s County, and 14% of temporary jobs, 29% of output and 15% of income are created in Washington DC. Exhibits 7.7 through 7.15 show the economic impact on temporary job creations, output, and income by Industry for Montgomery and Prince George’s Counties and Washington DC. We can see from Exhibit 7.7 through 7.15, construction industry itself will benefit from project implementation more than other industries. It will obtain around 60% of temporary jobs and 65% of output created in the Montgomery and Prince George’s Counties, and 77% of temporary jobs and 82% of output created in Washington DC. Exhibit 7.4: Temporary Job Creation Person Years Year 1
Year 2
Year 3
Year 4
Year 5
Average # of Jobs
Total Person Years of Work
Montgomery County
2,134
2,134
2,134
2,134
2,134
2,134
10,670
Prince George’s County
1,357
1,357
1,357
1,357
1,357
1,357
6,785
Washington DC
557
557
557
557
557
557
2,785
4,048
4,048
4,048
4,048
4,048
4,048
20,240
Total
Exhibit 7.5: Gross Regional Product (Millions 2014 $)
Montgomery County Prince George’s County Washington DC Total
Year 1
Year 2
Year 3
Year 4
Year 5
Total Output
$511.80
$511.80
$511.80
$511.80
$511.80
$2,558.98
$486.78
$486.78
$486.78
$486.78
$486.78
$2,433.89
$403.14
$403.14
$403.14
$403.14
$403.14
$2,015.70
$1,401.71
$1,401.71
$1,401.71
$1,401.71
$1,401.71
$7,008.57
32
According to Purple Line FEIS, the total project capital cost is 1.847 billion in 2012 dollars. By multiplying CPI factors from Bureau of Labor Statistics, it is converted to 2014 dollars.
Transportation Economics & Management Systems, Inc.
March 2015
48
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Exhibit 7.6: Personal Income (Millions 2014 $) Year 1
Year 2
Year 3
Year 4
Year 5
Average Income
Total Income
Montgomery County
$103.19
$103.19
$103.19
$103.19
$103.19
$103.19
$515.94
Prince George’s County
$64.05
$64.05
$64.05
$64.05
$64.05
$64.05
$320.27
$28.90
$28.90
$28.90
$28.90
$28.90
$28.90
$144.48
$196.14
$196.14
$196.14
$196.14
$196.14
$196.14
$980.68
Washington DC
Total
Besides construction industry, noteworthy employment gains in Montgomery County will be mainly in retail trade (9% of new temporary jobs), professional, scientific and technical services (6%), administrative and management services (4%), whereas in Prince George’s County besides construction industry, retail trade (10%), professional, scientific and technical services (4%), administrative and management services (4%). In Washington DC, besides construction, noteworthy employment gains will be mainly in professional, scientific and technical services (7%), administrative and management services (5%), and retail trade (2%). (See Exhibit 7.7 through 7.9) Among the industries that are estimated to receive significant share of output besides construction industry, we can point out on professional, scientific and technical services (6% in Montgomery County, 5% in Prince George’s County and 7% in Washington DC), real estate and rental and leasing (5% in Montgomery County, 4% in Prince George’s County and 2% in Washington DC), retail trade (5% in Montgomery County, and 5% in Prince George’s County), finance and insurance (4% in Montgomery County), and manufacturing (3% in Prince George’s County). (See Exhibit 7.10 through 7.12) As seen in Exhibit 7.13 through 7.15, among the industries that are estimated to receive significant income besides construction industry, we can point out on professional, scientific and technical services (9% in Montgomery County, 7% in Prince George’s County and 11% in Washington DC), and retail trade (5% in Montgomery County, 6% in Prince George’s County and 1% in Washington DC). Exhibit 7.7: Economic Impact by Industry Grouping - Temporary Job Creation in Montgomery County NAICS Descriptions 23 Construction
Year Year Year Year Year 1 2 3 4 5
Average # of jobs
1288 1,288 1,288 1,288 1,288
Total Person years of work
1,288
6,440
44 Retail trade
192
192
192
192
192
192
960
54 Professional, scientific, and technical services
130
130
130
130
130
130
650
56 Administrative and waste management services
80
80
80
80
80
80
400
62 Health care and social assistance
77
77
77
77
77
77
385
53 Real estate and rental and leasing
74
74
74
74
74
74
370
52 Finance and insurance
57
57
57
57
57
57
285
72 Food services and drinking places
54
54
54
54
54
54
270
Transportation Economics & Management Systems, Inc.
March 2015
49
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE 46
46
46
46
46
46
230
42 Wholesale trade
24
24
24
24
24
24
120
31 Manufacturing
23
23
23
23
23
23
115
51 Information
23
23
23
23
23
23
115
48 Transportation and warehousing*
15
15
15
15
15
15
75
55 Management of companies and enterprises
10
10
10
10
10
10
50
71 Arts, entertainment, and recreation
10
10
10
10
10
10
50
72 Accommodation
9
9
9
9
9
9
45
61 Educational services
8
8
8
8
8
8
40
H0 Households
8
8
8
8
8
8
40
21 Mining
4
4
4
4
4
4
20
11 Agriculture, forestry, fishing, and hunting
1
1
1
1
1
1
5
22 Utilities*
1
1
1
1
1
1
5
81, S0 Other services*
TOTAL
2,134 2,134 2,134
Transportation Economics & Management Systems, Inc.
2,134
2,134
March 2015
2,134
10,670
50
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.8: Economic Impact by Industry Grouping - Temporary Job Creation in Prince George’s County NAICS Descriptions
Year Year Year Year Year 1 2 3 4 5
Average # of jobs
Total Person years of work
23
Construction
843
843
843
843
843
843
4,215
44
133
133
133
133
133
133
665
58
58
58
58
58
58
290
56
Retail trade Professional, scientific, and technical services Administrative and waste management services
52
52
52
52
52
52
260
53
Real estate and rental and leasing
38
38
38
38
38
38
190
72
Food services and drinking places
35
35
35
35
35
35
175
48
Transportation and warehousing*
32
32
32
32
32
32
160
62
Health care and social assistance
31
31
31
31
31
31
155
81, S0
Other services*
30
30
30
30
30
30
150
31
Manufacturing
26
26
26
26
26
26
130
42
Wholesale trade
19
19
19
19
19
19
95
52
Finance and insurance
18
18
18
18
18
18
90
51
Information
11
11
11
11
11
11
55
61
Educational services
8
8
8
8
8
8
40
71
Arts, entertainment, and recreation
6
6
6
6
6
6
30
H0
Households
6
6
6
6
6
6
30
72
Accommodation
4
4
4
4
4
4
20
21
Mining
3
3
3
3
3
3
15
11
Agriculture, forestry, fishing, and hunting
2
2
2
2
2
2
10
22
Utilities*
1
1
1
1
1
1
5
55
Management of companies and enterprises 1
1
1
1
1
1
5
1,357
1,357
1,357
6,785
54
TOTAL
1,357 1,357 1,357
Transportation Economics & Management Systems, Inc.
March 2015
51
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.9: Economic Impact by Industry Grouping - Temporary Job Creation in Washington DC NAICS Descriptions
Year Year Year Year Year 1 2 3 4 5
Average # of jobs
Total Person years of work
430
430
430
430
430
430
2,150
37
37
37
37
37
37
185
26
26
26
26
26
26
130
44 Retail trade
13
13
13
13
13
13
65
72 Food services and drinking places
10
10
10
10
10
10
50
53 Real estate and rental and leasing
9
9
9
9
9
9
45
62 Health care and social assistance
6
6
6
6
6
6
30
51 Information
4
4
4
4
4
4
20
52 Finance and insurance
4
4
4
4
4
4
20
4
4
4
4
4
4
20
72 Accommodation
3
3
3
3
3
3
15
31 Manufacturing
2
2
2
2
2
2
10
42 Wholesale trade
2
2
2
2
2
2
10
48 Transportation and warehousing*
2
2
2
2
2
2
10
61 Educational services
2
2
2
2
2
2
10
71 Arts, entertainment, and recreation
2
2
2
2
2
2
10
H0 Households
1
1
1
1
1
1
5
11 Agriculture, forestry, fishing, and hunting
0
0
0
0
0
0
0
21 Mining
0
0
0
0
0
0
0
22 Utilities*
0
0
0
0
0
0
0
55 Management of companies and enterprises
0
0
0
0
0
0
0
23 Construction Professional, scientific, and technical 54 services Administrative and waste management 56 services
81, S0 Other services*
TOTAL
557
557
Transportation Economics & Management Systems, Inc.
557
557
557
557
March 2015
2,785
52
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.10: Economic Impact by Industry Grouping - Gross Regional Product in Montgomery County (Millions 2014 $) NAICS Descriptions 23 Construction Professional, scientific, and technical 54 services
Year 1
Year 2
Year 3
Year 4
Year 5
$332.63 $332.63 $332.63 $332.63 $332.63
Total Output $1,663.13
$31.48
$31.48
$31.48
$31.48 $31.48
$157.40
53 Real estate and rental and leasing
$26.11
$26.11
$26.11
$26.11 $26.11
$130.56
44 Retail trade
$23.92
$23.92
$23.92
$23.92 $23.92
$119.62
52 Finance and insurance
$20.48
$20.48
$20.48
$20.48 $20.48
$102.39
51 Information
$12.43
$12.43
$12.43
$12.43 $12.43
$62.13
62 Health care and social assistance
$11.10
$11.10
$11.10
$11.10 $11.10
$55.50
31 Manufacturing
$9.54
$9.54
$9.54
$9.54
$9.54
$47.72
42 Wholesale trade Administrative and waste management 56 services
$9.21
$9.21
$9.21
$9.21
$9.21
$46.06
$7.95
$7.95
$7.95
$7.95
$7.95
$39.76
$7.79
$7.79
$7.79
$7.79
$7.79
$38.94
55 Management of companies and enterprises
$4.41
$4.41
$4.41
$4.41
$4.41
$22.04
72 Food services and drinking places
$4.27
$4.27
$4.27
$4.27
$4.27
$21.37
48 Transportation and warehousing*
$2.85
$2.85
$2.85
$2.85
$2.85
$14.25
21 Mining
$2.65
$2.65
$2.65
$2.65
$2.65
$13.25
72 Accommodation
$1.66
$1.66
$1.66
$1.66
$1.66
$8.28
22 Utilities*
$1.19
$1.19
$1.19
$1.19
$1.19
$5.96
61 Educational services
$1.03
$1.03
$1.03
$1.03
$1.03
$5.14
71 Arts, entertainment, and recreation
$0.89
$0.89
$0.89
$0.89
$0.89
$4.47
11 Agriculture, forestry, fishing, and hunting
$0.20
$0.20
$0.20
$0.20
$0.20
$0.99
H0 Households
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
$511.80 $511.80 $511.80 $511.80 $511.80
$2,558.98
81, S0 Other services*
TOTAL
Transportation Economics & Management Systems, Inc.
March 2015
53
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.11: Economic Impact by Industry Grouping - Gross Regional Product in Prince George’s County (Millions 2014 $) NAICS Descriptions 23 Construction Professional, scientific, and technical 54 services
Year 1
Year 2
Year 3
Year 4
Year 5
$332.49 $332.49 $332.49 $332.49 $332.49
Total Output $1,662.47
$25.88
$25.88
$25.88
$25.88
$25.88
$129.40
44 Retail trade
$23.39
$23.39
$23.39
$23.39
$23.39
$116.97
53 Real estate and rental and leasing
$19.82
$19.82
$19.82
$19.82
$19.82
$99.08
31 Manufacturing
$16.40
$16.40
$16.40
$16.40
$16.40
$82.01
42 Wholesale trade
$12.59
$12.59
$12.59
$12.59
$12.59
$62.96
52 Finance and insurance
$9.87
$9.87
$9.87
$9.87
$9.87
$49.37
51 Information
$8.85
$8.85
$8.85
$8.85
$8.85
$44.24
81, S0 Other services*
$6.99
$6.99
$6.99
$6.99
$6.99
$34.96
48 Transportation and warehousing* Administrative and waste management 56 services
$6.63
$6.63
$6.63
$6.63
$6.63
$33.14
$6.26
$6.26
$6.26
$6.26
$6.26
$31.31
62 Health care and social assistance
$6.16
$6.16
$6.16
$6.16
$6.16
$30.82
72 Food services and drinking places
$3.58
$3.58
$3.58
$3.58
$3.58
$17.89
55 Management of companies and enterprises
$2.02
$2.02
$2.02
$2.02
$2.02
$10.11
22 Utilities*
$1.66
$1.66
$1.66
$1.66
$1.66
$8.28
21 Mining
$1.56
$1.56
$1.56
$1.56
$1.56
$7.79
61 Educational services
$1.06
$1.06
$1.06
$1.06
$1.06
$5.30
72 Accommodation
$0.73
$0.73
$0.73
$0.73
$0.73
$3.65
71 Arts, entertainment, and recreation
$0.60
$0.60
$0.60
$0.60
$0.60
$2.98
11 Agriculture, forestry, fishing, and hunting
$0.23
$0.23
$0.23
$0.23
$0.23
$1.16
H0 Households
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
TOTAL
$486.78
Transportation Economics & Management Systems, Inc.
$486.78 $486.78 $486.78 $486.78
March 2015
$2,433.89
54
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.12: Economic Impact by Industry Grouping - Gross Regional Product in Washington DC (Millions 2014 $) NAICS Descriptions 23 Construction Professional, scientific, and technical 54 services
Year 1
Year 2
Year 3
Year 4
Year 5
Total Output
$331.53 $331.53
$331.53 $331.53 $331.53
$1,657.66
$26.44 $26.44
$26.44 $26.44 $26.44
$132.22
53 Real estate and rental and leasing
$9.84
$9.84
$9.84
$9.84
$9.84
$49.21
51 Information
$6.36
$6.36
$6.36
$6.36
$6.36
$31.81
52 Finance and insurance Administrative and waste management 56 services
$5.87
$5.87
$5.87
$5.87
$5.87
$29.33
$4.31
$4.31
$4.31
$4.31
$4.31
$21.54
44 Retail trade
$3.51
$3.51
$3.51
$3.51
$3.51
$17.56
81, S0 Other services*
$2.35
$2.35
$2.35
$2.35
$2.35
$11.76
62 Health care and social assistance
$2.29
$2.29
$2.29
$2.29
$2.29
$11.43
31 Manufacturing
$2.25
$2.25
$2.25
$2.25
$2.25
$11.27
42 Wholesale trade
$2.05
$2.05
$2.05
$2.05
$2.05
$10.27
72 Food services and drinking places
$1.49
$1.49
$1.49
$1.49
$1.49
$7.46
48 Transportation and warehousing*
$1.23
$1.23
$1.23
$1.23
$1.23
$6.13
72 Accommodation
$1.06
$1.06
$1.06
$1.06
$1.06
$5.30
22 Utilities*
$0.89
$0.89
$0.89
$0.89
$0.89
$4.47
55 Management of companies and enterprises
$0.73
$0.73
$0.73
$0.73
$0.73
$3.65
61 Educational services
$0.46
$0.46
$0.46
$0.46
$0.46
$2.32
71 Arts, entertainment, and recreation
$0.46
$0.46
$0.46
$0.46
$0.46
$2.32
11 Agriculture, forestry, fishing, and hunting
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
21 Mining
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
H0 Households
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
TOTAL
$403.14
Transportation Economics & Management Systems, Inc.
$403.14
$403.14
$403.14 $403.14
March 2015
$2,015.70
55
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.13: Economic Impact by Industry Grouping - Income in Montgomery County (Millions 2014 $) NAICS Descriptions 23 Construction Professional, scientific, and technical 54 services
Year 1
Year 2
$66.70 $66.70
Year 3
Year 4
Year 5
Average Income
Total Income
$66.70
$66.70
$66.70
$66.70
$333.52
$9.41
$9.41
$9.41
$9.41
$9.41
$9.41
$47.05
44 Retail trade
$5.40
$5.40
$5.40
$5.40
$5.40
$5.40
$27.01
52 Finance and insurance
$3.98
$3.98
$3.98
$3.98
$3.98
$3.98
$19.88
62 Health care and social assistance Administrative and waste management 56 services
$3.71
$3.71
$3.71
$3.71
$3.71
$3.71
$18.56
$2.05
$2.05
$2.05
$2.05
$2.05
$2.05
$10.27
53 Real estate and rental and leasing
$1.86
$1.86
$1.86
$1.86
$1.86
$1.86
$9.28
42 Wholesale trade
$1.76
$1.76
$1.76
$1.76
$1.76
$1.76
$8.78
81, S0 Other services*
$1.49
$1.49
$1.49
$1.49
$1.49
$1.49
$7.46
51 Information
$1.46
$1.46
$1.46
$1.46
$1.46
$1.46
$7.29
31 Manufacturing
$1.36
$1.36
$1.36
$1.36
$1.36
$1.36
$6.79
55 Management of companies and enterprises
$1.03
$1.03
$1.03
$1.03
$1.03
$1.03
$5.14
72 Food services and drinking places
$0.99
$0.99
$0.99
$0.99
$0.99
$0.99
$4.97
48 Transportation and warehousing*
$0.63
$0.63
$0.63
$0.63
$0.63
$0.63
$3.15
21 Mining
$0.36
$0.36
$0.36
$0.36
$0.36
$0.36
$1.82
72 Accommodation
$0.30
$0.30
$0.30
$0.30
$0.30
$0.30
$1.49
61 Educational services
$0.27
$0.27
$0.27
$0.27
$0.27
$0.27
$1.33
71 Arts, entertainment, and recreation
$0.20
$0.20
$0.20
$0.20
$0.20
$0.20
$0.99
22 Utilities*
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.50
H0 Households
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.50
11 Agriculture, forestry, fishing, and hunting
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.17
TOTAL
$103.19 $103.19 $103.19 $103.19 $103.19
Transportation Economics & Management Systems, Inc.
March 2015
$103.19
$515.94
56
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.14: Economic Impact by Industry Grouping - Income in Prince George’s County (Millions 2014 $) NAICS Descriptions 23 Construction
Year 1
Year 2
$43.64 $43.64
Year 3
Year 4
Year 5
Average Income
Total Income
$43.64
$43.64
$43.64
$43.64
$218.21
54 Professional, scientific, and technical services
$4.18
$4.18
$4.18
$4.18
$4.18
$4.18
$20.88
44 Retail trade
$3.74
$3.74
$3.74
$3.74
$3.74
$3.74
$18.72
31 Manufacturing
$1.39
$1.39
$1.39
$1.39
$1.39
$1.39
$6.96
56 Administrative and waste management services
$1.39
$1.39
$1.39
$1.39
$1.39
$1.39
$6.96
42 Wholesale trade
$1.36
$1.36
$1.36
$1.36
$1.36
$1.36
$6.79
62 Health care and social assistance
$1.36
$1.36
$1.36
$1.36
$1.36
$1.36
$6.79
48 Transportation and warehousing*
$1.29
$1.29
$1.29
$1.29
$1.29
$1.29
$6.46
53 Real estate and rental and leasing
$1.16
$1.16
$1.16
$1.16
$1.16
$1.16
$5.80
52 Finance and insurance
$1.13
$1.13
$1.13
$1.13
$1.13
$1.13
$5.63
81, S0 Other services*
$0.96
$0.96
$0.96
$0.96
$0.96
$0.96
$4.80
51 Information
$0.73
$0.73
$0.73
$0.73
$0.73
$0.73
$3.65
72 Food services and drinking places
$0.66
$0.66
$0.66
$0.66
$0.66
$0.66
$3.31
61 Educational services
$0.27
$0.27
$0.27
$0.27
$0.27
$0.27
$1.33
21 Mining
$0.23
$0.23
$0.23
$0.23
$0.23
$0.23
$1.16
71 Arts, entertainment, and recreation
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.66
72 Accommodation
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.66
22 Utilities*
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.50
55 Management of companies and enterprises
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.50
H0 Households
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.33
11 Agriculture, forestry, fishing, and hunting
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.17
TOTAL
$64.05
$64.05
Transportation Economics & Management Systems, Inc.
$64.05
$64.05
$64.05
March 2015
$64.05
$320.27
57
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit 7.15: Economic Impact by Industry Grouping - Income in Washington DC (Millions 2014 $) NAICS Descriptions 23 Construction Professional, scientific, and technical 54 services Administrative and waste management 56 services
Year 1
Year 2
Year 3
Year 4
Year 5
$22.27 $22.27 $22.27 $22.27 $22.27
Average Income
Total Income
$22.27
$111.34
$3.28
$3.28
$3.28
$3.28
$3.28
$3.28
$16.40
$0.70
$0.70
$0.70
$0.70
$0.70
$0.70
$3.48
44 Retail trade
$0.40
$0.40
$0.40
$0.40
$0.40
$0.40
$1.99
52 Finance and insurance
$0.40
$0.40
$0.40
$0.40
$0.40
$0.40
$1.99
51 Information
$0.33
$0.33
$0.33
$0.33
$0.33
$0.33
$1.66
53 Real estate and rental and leasing
$0.27
$0.27
$0.27
$0.27
$0.27
$0.27
$1.33
62 Health care and social assistance
$0.27
$0.27
$0.27
$0.27
$0.27
$0.27
$1.33
72 Food services and drinking places
$0.20
$0.20
$0.20
$0.20
$0.20
$0.20
$0.99
31 Manufacturing
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.66
42 Wholesale trade
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.66
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.66
72 Accommodation
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.50
48 Transportation and warehousing*
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.33
61 Educational services
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.33
71 Arts, entertainment, and recreation
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.33
22 Utilities*
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.17
55 Management of companies and enterprises
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.17
H0 Households
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.17
11 Agriculture, forestry, fishing, and hunting
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
21 Mining
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
81, S0 Other services*
TOTAL
$28.90
$28.90
Transportation Economics & Management Systems, Inc.
$28.90
$28.90
$28.90
March 2015
$28.90
$144.48
58
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE The Input-Output analysis shows that the Purple Line Project will have a general sizeable impact on the economies of Montgomery and Prince George’s Counties and Washington DC. The results are summarized in Exhibit 7.4 through 7.15. Since contractors on the project will buy materials and services from other Purple Line corridor region businesses, the RIMS II analysis predicts a multiplier effect on the initial construction expenditure of $1.7 billion. All Purple Line corridor region industries are expected to benefit by approximately USD $7 billion and over 20,240 person-years of work during the construction period. Over a five-year construction period, this would be equivalent to adding approximately 4,048 temporary jobs annually in Montgomery and Prince George’s Counties and Washington DC, occurring as a direct result of the construction expenditures alone. It should be noted that depending upon the type of project and its location, the multiplier effects from the additional investment, jobs, income and workers’ spending decisions would differ. This is because the characteristics of the local economy determines exactly how much extra impact an investment will generate.
Transportation Economics & Management Systems, Inc.
March 2015
59
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
8 CONCLUSION The improvement of the Purple Line Preferred Alternative will provide an integrating force for the communities in Montgomery and Prince George’s Counties and Washington DC. It will also support the opportunities to fundamentally change the character of business in the area while expanding the level of social, personal and tourist interaction. The project will create a new business environment and will ensure the stability of existing service, transportation and manufacturing industries, while fostering the growth of new small businesses in the area because of the improved access to smaller communities along the Purple Line corridor. Economic Rent: The long-term supply side benefits are considerable because a very high level of east-west access is being provided between the communities along the Purple Line corridor and because the corridor has a very high level of economic vitality. The Economic Rent analysis shows supply side benefits for the economy of study region. 1.0 to 2.2 percent growth in the economy of Montgomery and Prince George’s Counties and Washington DC. 23,033 long-term person years of work per year for Montgomery and Prince George’s Counties and Washington DC. Increase in income to local households of $2.2 billion per year or over $40 billion over the life of the project (overall). The property value increase, assuming full advantage is taken by local communities of the Purple Line, is $12.8 billion (overall) or $11.0 billion for Montgomery and Prince George’s Counties and Washington DC. Total expected tax benefits (Federal, State, Local and Sales Tax) from the Purple Line project implementation are in the range of at least $635 million per year (overall) or $541 million for Montgomery and Prince George’s Counties and Washington DC. The development of the Purple Line Preferred Alternative will also result in significant economic impact in Montgomery and Prince George’s Counties and Washington DC during the construction of the project. Input-Output Analysis: The regional use of federal construction dollars to build the system will generate a substantial economic impact in the region. During the construction period it will – Create 20,000 person years of work or the equivalent of 4,000 full-time jobs annually during the 5-year construction period (this includes construction plus other industry jobs) for Montgomery and Prince George’s Counties and Washington DC. Increase in income in Montgomery and Prince George’s Counties and Washington DC by $980 million during the 5-year construction period. Increase in regional output in these two counties and Washington DC by 7.0 billion during the 5year construction period.
Transportation Economics & Management Systems, Inc.
March 2015
60
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
APPENDIX A: DATA SOURCES AND METHODS FOR RIMS II Most regional economists agree that the most accurate method of estimating a regional input-output (I-O) model is to survey the businesses in a region in order to determine which goods and services are purchased by industries in the region and whether these goods and services are purchased from other industries in the region or from industries in other regions. However, because these surveys are costly, few regional I-O models in the United States are based on survey, or primary source, data. As a result, the estimation of the regional input-output modeling system (RIMS II), like that of most regional I-O models, is based on data from the national I-O accounts and other secondary data. It is assumed that the national I-O data can be used to represent the composition of inputs purchased in the region. The national data are then adjusted by regional data, because the industries in a region cannot obtain all of their inputs from within the region. The RIMS II model and its multipliers are prepared in three major steps.33 First, an adjusted national industry-by-industry direct requirements table is prepared. Second, the adjusted national table is used to prepare a regional industry-by-industry direct requirements table.34 Third, a regional industry-by-industry total requirements table is prepared, and the multipliers are derived from this table.
A.1 THE ADJUSTED NATIONAL DIRECT REQUIREMENTS TABLE The adjusted national industry-by-industry direct requirements table is derived from the make and use tables in BEA’s 1987 benchmark I-O accounts for the U.S. economy.35 The use table is adjusted so that it includes only the use of domestically produced commodities. The data in a use table for imported commodities are subtracted from the data in the total commodity use table. After this adjustment, a national industry-by-industry direct requirements table is prepared by means of standard I-O procedures.36 An industry-share matrix, which shows each industry’s share of the production of a commodity, is calculated by dividing each entry in each column of the make table by the respective column total. Next, a commodity-by-industry direct requirements matrix, which shows the dollar’s worth of each commodity that is required to produce a dollar’s worth of each industry’s output, is calculated by dividing each entry in each column of the use table by the respective column total. A national industry-byindustry direct requirements table is then estimated by multiplying the industry-share matrix by the commodity-by-industry direct requirements matrix. Unlike the national I-O accounts, RIMS II includes households as both suppliers of labor inputs to regional industries and as purchasers of regional output, because it is customary in regional impact analysis to
33
This discussion is mainly for users who are familiar with I-O theory and linear algebra. For a more detailed discussion of I-O theory and the use of regional I-O models in impact analysis, see “Suggested Reading” at the end of this appendix. 34 In RIMS II, a region consists of the county or counties that are specified by the user. 35 The make table shows the dollar value, in producers’ prices, of each commodity produced by each industry. The use table shows the dollar value, in producers’ prices, of each commodity used by each industry and by each final user. See Benchmark InputOutput Accounts for the U.S. Economy, 1987, SURVEY OF CURRENT BUSINESS 74 (April 1994): 73–115;and U.S. Department of Commerce, Bureau of Economic Analysis, Benchmark Input-Output Accounts of the United States, 1987 (Washington, DC: U.S. Government Printing Office, 1994). 36 See Ronald E. Miller and Peter D. Blair, Input-Output Analysis: Foundations and Extensions (Englewood Cliffs, NJ: Prentice-Hall, 1985), 149–199.
Transportation Economics & Management Systems, Inc.
March 2015
61
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE account for the effects of changes in household earnings and expenditures. Thus, both a household row and a household column are added to the national direct requirements table before the table is regionalized.37
A.2 THE HOUSEHOLD ROW Each entry in the household row shows the earnings received by households per dollar of output of the column industry corresponding to the entry. In impact analysis with RIMS II, earnings is defined as the earnings that are received by households from the production of regional goods and services and that are available for spending on these goods and services. Thus, earnings is calculated as the sum of wages and salaries, proprietors’ income, directors’ fees, and employer contributions for health insurance less personal contributions for social insurance.38 In equation form, the household row is -
HSHRj = (W&Sj + PRPj + DFj + ECHIj - PCSIj) / TIOj , where the subscript j is industry j (column j in the direct requirements table), HSHR is the household row, W&S is wages and salaries, PRP is proprietors’ income, DF is directors’ fees, ECHI is employer contributions for health insurance, PCSI is personal contributions for social insurance, and TIO is total industry output. The estimates of wages and salaries by I-O industry are from the national I-O accounts. The other earnings components are not available by I-O industry and must be estimated. The estimates of non-farm proprietors’ income by I-O industry are made by multiplying non-farm proprietors’ income at the two-digit Standard Industrial Classification (SIC) level by each I-O industry’s share of wages and salaries in the corresponding two-digit SIC industry. The data source for non-farm proprietors’ income and for wages and salaries at the two-digit SIC level is BEA’s Regional Economic Information System (REIS). The estimates of farm proprietors’ income for 17 I-O agricultural industries are calculated by multiplying total farm proprietors’ income by the shares of total farm cash receipts accounted for by each of the agricultural industries. The data source for farm proprietors’ income and cash receipts is REIS.39 The estimates of directors’ fees by I-O industry are calculated by multiplying directors’ fees at the twodigit SIC level by each I-O industry’s share of wages and salaries in the corresponding two-digit SIC industry. The data source for directors’ fees is REIS. The estimates of employer contributions for health insurance by I-O industry are prepared in two steps. First, employer contributions to private pension funds and private welfare funds at the two-digit SIC level are multiplied by the all-industry ratio of employer contributions for health insurance to employer contributions to private pension funds and private welfare funds to yield estimates of employer contributions for health insurance at the two-digit SIC level. These estimates are then multiplied by each IO industry’s share of wages and salaries in the corresponding two-digit SIC industry. The source for the all-industry data is the national income and product accounts, and the source for the two-digit SIC data is REIS.
37
I-O theory requires that the sum of the entries in each column of the direct requirements table be less than, or equal to, one. Because this condition is not met for all industries after the household row is added, nine industries must be combined with similar industries. 38 Earnings include employer contributions for health insurance, because personal consumption expenditures data in the national I-O accounts include expenditures on health care. Earnings exclude personal contributions for social insurance, because these contributions are usually deducted from an employee’s wages and salaries and therefore are unavailable for spending on regional goods and services. 39 For agriculture, the estimates of proprietors’ income by I-O industry are not based on wages and salaries, because the share of total employment accounted for by wage-and-salary workers in agriculture is substantially smaller than that in other industries.
Transportation Economics & Management Systems, Inc.
March 2015
62
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE The estimates of personal contributions for social insurance by I-O industry are calculated by multiplying personal contributions for social insurance for all industries by each I-O industry’s share of wages and salaries for all industries. The data source for personal contributions for social insurance is REIS.
A.3 THE HOUSEHOLD COLUMN Each entry in the household column shows the expenditures per dollar of household earnings on the product of the row industry corresponding to the entry. The estimation of the household column is based on personal consumption expenditures (PCE) data from the national I-O accounts. PCE data in the importedcommodity use table are subtracted from PCE data in the overall use table to yield a column that shows PCE for domestically produced commodities. After each column entry is expressed as a share of total PCE, the column is multiplied by the industry-share matrix (discussed earlier) to yield the PCE shares by I-O industry.40 The PCE shares by industry are then multiplied by the ratio of personal income less taxes and savings to personal income in order to account for the dampening effect of taxes and savings on expenditures.
A.4 THE REGIONAL DIRECT REQUIREMENTS TABLE The regional industry-by-industry direct requirements table is derived from the adjusted national industryby-industry direct requirements table. Location quotients (LQ’s) are used to “regionalize” the national data.41 The LQ is used as a measure of the extent to which regional supply of an industry’s output is sufficient to meet regional demand. If the LQ for a row industry in the regional direct requirements table is greater than, or equal to, one, it is assumed that the region’s demand for the output of the row industry is met entirely from regional production. In this instance, all row entries for the industry in the regional direct requirements table are set equal to the corresponding entries in the adjusted national direct requirements table. Conversely, if the LQ is less than one, it is assumed that regional supply of the industry’s output is not sufficient to meet regional demand. In this instance, all row entries for the industry in the regional direct requirements table are set equal to the product of the corresponding entries in the adjusted national direct requirements table and the LQ for the industry. The household row and the household column that were added to the national direct requirements table are also adjusted regionally. The household row entries are adjusted downward, on the basis of commuting data from the Census of Population, in order to account for the purchases made outside the region by commuters working in the region. The household column entries are adjusted downward, on the basis of tax data from the Internal Revenue Service, in order to account for the dampening effect of State and local taxes on household expenditures.
A.5 THE REGIONAL TOTAL REQUIREMENTS TABLE AND THE MULTIPLIERS A regional industry-by-industry total requirements table is prepared by calculating the Leontief inverse from the regional direct requirements table.42 The regional total requirements table shows the regional final-demand output multipliers. In I-O terminology, the multipliers account for the sum of the direct,
40
The last entry of the column is purchases of domestic services by households, which equals earnings received by domestic service workers. 41 For most industries in RIMS II, LQ’s are based on 1992 wages and salaries by industry at the four-digit SIC level. The LQ for wages and salaries is the ratio of the industry’s share of regional wages and salaries to that industry’s share of national wages and salaries. For some industries, the LQ’s are adjusted, because wages and salaries in these industries, in comparison with proprietors’ income, accounts for a relatively small share of total earnings. 42 The Leontief inverse is defined as (I-A)-1, where I is an identity matrix, A is the regional industry-by-industry direct requirements matrix, and -1 indicates a matrix inversion.
Transportation Economics & Management Systems, Inc.
March 2015
63
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE indirect, and induced effects of a change in final demand. The final-demand, direct-effect, and outputdriven multipliers can be derived from the total requirements table.43
A.6 FINAL-DEMAND EARNINGS AND EMPLOYMENT MULTIPLIERS Final-demand earnings multipliers are derived by multiplying each final-demand output multiplier in the total requirements table by the household row entry in the direct requirements table that corresponds to the row industry for the output multiplier. This calculation is expressed as
ci,j = bi,j * a471,i , where ci,j is the entry in row i and column j of the final-demand earnings multiplier table, bi,j is the finaldemand output multiplier in the total requirements table, and a471,i is the household row entry in the direct requirements table.44 Final-demand employment multipliers are derived by multiplying each entry in the final-demand earnings multiplier table by the employment-to-earnings ratio for each row industry.45 This calculation is expressed as
ei,j = ci,j *Gi , where ei,j is the entry in row i and column j of the final-demand employment multiplier table, ci,j is the finaldemand earnings multiplier, and Gi is the employment-to-earnings ratio for row industry i.
A.7 DIRECT-EFFECT EARNINGS AND EMPLOYMENT MULTIPLIERS Direct-effect earnings multipliers are derived by dividing each household row entry in the total requirements table by the corresponding household row entry in the direct requirements table. This calculation is expressed as
Dj = b471,j / a471,j , where Dj is the direct-effect earnings multiplier for industry j, b471,j is the household row entry for industry j in the total requirements table, and a471,j is the household row entry for industry j in the direct requirements table. Direct-effect employment multipliers are derived by dividing the final-demand employment multiplier for each industry by the product of the corresponding household row entry in the direct requirements table and the employment-to-earnings ratio for each column industry. This calculation is expressed as
Hj = Fj / (a471,j * Gj) ,
43
For the discussion of the use of these multipliers in regional impact analysis, see the section “RIMS II Multipliers for Output, Earnings, and Employment.” 44 The sum of all the entries in column j of the final-demand earnings multiplier table is equal to the household-row entry in column j of the total requirements table. The last row of the final-demand earnings multiplier table represents earnings received by households that have domestic service jobs. 45 Employment is measured on a job-count basis for both wage-and-salary workers and proprietors. Estimates of employment by I-O industry are made by allocating REIS employment data by two-digit SIC industry in proportion to Bureau of Labor Statistics wageand-salary employment data by I-O industry.
Transportation Economics & Management Systems, Inc.
March 2015
64
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE where Hj is the direct-effect employment multiplier for industry j, Fj is the final-demand employment multiplier for industry j, a471,j is the household row entry for industry j in the direct requirements table, and Gj is the employment-to-earnings ratio for industry j.46
A.8 OUTPUT-DRIVEN MULTIPLIERS Output-driven multipliers can be calculated from the total requirements table. The table entry for which the row entry i equals the column entry j is called the “diagonal” entry for column j. The output-driven multiplier for industry j is defined as the ratio of each entry in column j to the diagonal entry for that column. This ratio is expressed as oi,j = bi,j / bj,j , where oi,j is the output-driven multiplier i for industry j, bi,j is the final-demand output multiplier i for industry j in the total requirements table, and bj,j is the diagonal entry for industry j in the total requirements table.
A.9 REFERENCES Hewings, Geoffrey J.D. Regional Input-Output Analysis. Scientific Geography Series vol. 6. Beverly Hills, CA: Sage Publications, 1985. Miernyck, William H. The Elements of Input-Output Analysis. New York, NY: Random House, 1965. Miller, Ronald E., and Peter D. Blair. Input-Output Analysis: Foundations and Extensions. Englewood Cliffs, NJ: Prentice-Hall, 1985. Otto, Daniel M., and Thomas G. Johnson. Microcomputer-Based Input-Output Modeling. Boulder, CO: Westview Press, 1993. Richardson, Harry W. Input-Output and Regional Economics. New York, NY: Halsted Press, 1972.
46
The final-demand employment multiplier for industry j is the sum of all the entries except the household-row entry in column j of the final-demand employment multiplier table.
Transportation Economics & Management Systems, Inc.
March 2015
65
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
APPENDIX B: STATED PREFERENCE SURVEY B.1 INTRODUCTION The Stated Preference Analysis was based on results from a broad range of collected stated preference survey forms. Stated Preference Survey method uses a quota sampling approach as a fast and effective way of gathering consumer information on the importance of different travel decisions. This includes such issues as how travelers value travel time (for auto and transit modes) and how they value frequency of service and access time (for transit modes). A quota survey, as opposed to a random survey or a focus group study, is particularly effective in ensuring that all the important travel attributes are measured for the whole population at minimum cost. The quota survey, which has been widely adopted for public opinion surveys, is based on the development of representative “quotas” of the traveling public. The TEMS analysis requires that, two sets of data be collected: (1) the data that define the “travel behavior” quota and (2) the data that define the “personal profile” quota for the individuals surveyed. This allows the data to be stratified by such factors as trip length, income, and group size.
B.2 SURVEY FORMS Specifically for the Purple Line Economic Impact Study, TEMS has developed three types of Stated Preference Survey forms - the Auto Mode Stated Preference Survey with VOT questions (see Exhibit B.1), the Transit Mode Stated Preference Survey with VOT and VOF questions (see Exhibit B.2) and the Transit Mode Stated Preference Survey with VOA questions (see Exhibit B.3). One part of the survey contains profile questions while the other part contains questions that aim on defining the travel behavior of the surveyed individuals. The profile data collected from the surveys is used in conjunction with origindestination and census data to ensure that the stated preference survey can be effectively expanded to properly represent the total population. The collected travel behavior data provides the critical part of the data needed to estimate the generalized cost of travel. In terms of the size of the survey for each of the quota groups identified - usually up to 10 primary groups - it has been shown that a sample as small as 20 individuals47 is statistically significant to define the behavioral choices of each group. These primary groups are based on correspondence between 2-3 mode groups - auto and transit (that includes bus and rail) to 3-4 purpose groups (commuter, business, and other (that includes shopping and social). To improve statistical reliability, TEMS typically seeks 40 to 100 respondents per quota. This means that between 500 and 1,200 surveys are needed for a stated preference survey analysis. There were 6 total quota groups identified for the Purple Line study on the base of selected 2 mode groups (auto and transit) and 3 purpose groups (commuter, business and other.) The minimum of 1,000-1,200 surveys was set as a goal.
47
Spiegel, M.R., Theory and Problems of Probability and Statistics, NY McGraw Hill, pp. 112-113, 1992
Transportation Economics & Management Systems, Inc.
March 2015
66
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.1: Survey Form for Auto Users with VOT questions
Transportation Economics & Management Systems, Inc.
March 2015
67
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.2: Survey Form for Transit Users with VOT and VOF Questions
Transportation Economics & Management Systems, Inc.
March 2015
68
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
Exhibit B.3: Survey Form for Transit Users with VOA Questions
B.3 SURVEY LOCATIONS A very important part of the survey process is to identify the desirable survey locations. Exhibit B.4 shows the Purple Line Stated Preference Survey locations map with the Purple Line rail corridor. Prince George’s and Montgomery Counties were the main locations chosen for the surveys, as the proposed Purple Line stations are located in these counties. The surveys were conducted both electronically and also in the field. The main aim of the surveys was to target all six quota groups i.e., Business, Commuters, and for other purpose such as shopping and other social events for both auto and transit users. The survey locations are listed below: Metro & Rail Stations: Greenbelt, and New Carrollton, and Bus stop at Prince George’s Community College; Social Events: Art Fest at Riverdale, Engineers Field Day event at University of Maryland at College park, Fenton Street Market near Silver spring Metro station, and Langley Park shopping Plaza; and
Transportation Economics & Management Systems, Inc.
March 2015
69
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Work Places and Educational Institutions: National Institutes of Health at Bethesda, and Rockville, Guardian Life Insurance at Bethesda, American Institute of Physics at College Park, and Montgomery Community College at Rockville. TEMS collected 3,374 surveys, and exceeded their target range. As shown in Exhibit B.5 each quota group has a statically significant representation.
Exhibit B.4: Survey Locations Map
Exhibit B.5: Stated Preference Surveys Collected by Mode and Purpose Business
Commuter
Other
Total
Auto
60
1863
328
2251
Transit
111
538
474
1123
Total
171
2401
802
3374
Transportation Economics & Management Systems, Inc.
March 2015
70
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
B.4 PERSONAL PROFILE DATA Exhibit B.6 shows the trip purposes of the potential Purple Line users. Commuting to/from work is the main purpose of travel for both auto users (84%) and transit users (52%), followed by travel to school (7% for auto users and 31% for transit users). It is easy to see that the total share of travelers who go either to work or to school comprise no less than 80% of auto as well as transit travelers. 5-7% of travelers indicate their trip purpose as ‘other’, which might include vacation/recreation, tourism/sightseeing or visiting family/friends. The share of shoppers is 2% for both auto and transit users, while the share of business travelers is higher for transit (8%) than for auto (2%). Exhibit B.6: Trip Purpose of Survey Respondents (by Mode)
AUTO Users: Commuter 84%
Other 5%
Business 2% Travel to Shopping School 2% 7%
TRANSIT Users: Business 8% Commuter 52% Travel to School 31% Other Shopping 7% 2%
Transportation Economics & Management Systems, Inc.
March 2015
71
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Employment status of survey respondents is presented in Exhibit B.7, separately for auto and transit users. It is seen that, for auto users 90% are employed, while for transit users only 62% have jobs. However, 33% of transit users are students, while only 7% of auto users are students. Workers and students represent over 95 percent of travelers for both modes. Also, in the poll of Transit users more respondents indicated themselves as unemployed (3%) than in the poll of auto users (1%). In both auto and transit poll of respondents 1% identified themselves as retired and 1% did not provide any identification of the employment status. Exhibit B.7: Employment Status of Survey Respondents (by Mode)
Auto Users: Student 7%
Unemployed 1% Uknown 1%
Retired 1%
TRANSIT Users: Unemployed 3%
Employed 90%
Uknown 1%
Student 33%
Retired 1%
Transportation Economics & Management Systems, Inc.
Employed 62%
March 2015
72
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.8 illustrate employment distribution along the Purple Line corridor in comparison with statistical and survey data. It is seen in the Exhibit that Montgomery and Prince George’s Counties have highest share of employed individuals and lowest share of retired and unemployed individuals based on statistical data from U.S. Census 2006-2008 American Community Survey data on 3-year estimates on Employment status48. The survey data is very similar to the statistical data showing the representativeness of the stated preference survey. Exhibit B.8: Employment Distribution – Statistical Data and Survey Data
90% 80%
Survey Results
70%
Percent
60%
Statistical Data
50% 40% 30% 20% 10% 0% Employed
Student Only
Retired
Unemployed
Employment Status
Exhibit B.9 and B.10 illustrate household income groups (in Montgomery and Prince George’s Counties respectively) distributed in accordance with statistical from U.S. Census 2006-2008 American Community Survey data on 3-year estimates on household income and survey data. It is easy to see in Exhibit B.9 that in Prince George’s County the share of households in each of four groups calculated based on statistical and survey data is almost the same (the difference is 5% or less). In Montgomery County the shares of two medium income groups, based on statistical and survey data are almost identical. Also in Montgomery County, there was a higher share of survey respondents from high-income households and lower share of those who belong to low-income households in comparison with the average statistical data. This is because some of the proposed Purple Line stations – such as Bethesda or Connecticut Avenue, are located in the center of the luxury communities in Montgomery County. Thus, in 2008 median household income was $130,637 in Bethesda and $211,349 in Chevy Chase (Connecticut Avenue station). For comparison, in 2008 median household income for Montgomery County was $94,319.
48
American FactFinder database, http://factfinder.census.gov and MapPoint demographic database.
Transportation Economics & Management Systems, Inc.
March 2015
73
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.9: Household Income Distribution in Prince George’s County – Statistical and Survey Data
40% Survey Results 35%
% of Households
30% Statistical Data 25% 20% 15% 10% 5% 0% Less than $55,000
$55,000 - $74,999
$75,000 - $99,999
$100,000 or more
Exhibit B.10: Household Income Distribution in Montgomery County – Statistical and Survey Data Range of Household Income
60% Survey Results
% of Households
50%
40%
Statistical Data
30%
20%
10%
0% Less than $55,000
$55,000 - $74,999
$75,000 - $99,999
$100,000 or more
Range of Household Income
A comparison of the results of the stated preference survey such as employment and household income data with that of the U.S. Census statistical data shows that the employment and income distributions by mode are representative of statistical data. This means that the database will be representative of the Purple Line corridor value of time, values of frequency and values of access.
Transportation Economics & Management Systems, Inc.
March 2015
74
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
B.5 BEHAVIORAL ATTRIBUTES Behavioral attributes reflect the behavior of the respondent when travel conditions change. For the purpose of this study using stated preference surveys we collected information necessary to identify the Value of Time (VOT) for all travelers, the Value of Frequency (VOF) and the Value of Access (VOA). Each of these three variables (VOT, VOF and VOA) has been analyzed using the “trade-off’ method. This method aims at deriving VOT (VOF or VOA) by examining the pattern of trades a respondent makes among the various situations described in a travel behavior part of a questionnaire form. In this way a VOT (VOF or VOA) estimate can be assigned to each respondent. Exhibit B.11 through B.13 provides the example of the respondent’s trading behavior and illustrates how VOT is calculated using ‘trade-off’ method. The VOT is calculated for the ‘neutral point’ located in the intersection between the line indicating ‘no preference’ and the line connecting the points indicated by the respondent. As seen in Exhibit B.11 the neutral point or no preference line is located at the fourth row indicating that the respondent is willing to spend $12 more for 45 minutes less. This implies the respondent is willing to spend $16 more for 1 hour of time saving. Thus, the respondent has a VOT value of 16 dollars per hour. Not all survey respondents illustrated perfect trading behavior (similar to those shown in Exhibit B.11 or B.12). About 6% of the respondents were identified as ‘non-traders’. VOT calculated based on the example shown in Exhibit B.13 is assumed to be $45 for 3 hours (15 dollar per hour) or less as there is no trading, and the individual is showing a preference to spend time rather than money.
Transportation Economics & Management Systems, Inc.
March 2015
75
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.11: VOT calculation based on “Trade-Off” Method: “Trading Behavior”- Example # 1
Exhibit B.12: VOT calculation based on “Trade-Off” Method: “Trading Behavior” - Example # 2
Transportation Economics & Management Systems, Inc.
March 2015
76
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.13: VOT calculation based on ‘“Trade-Off” Method: “Non Trading Behavior”
VOT, VOF and VOA calculated on the base of Stated Preference Survey were carefully checked for statistical significance. In accordance with the general statistical theory, results are considered statistically significant if they are normally distributed and 68% of the results fall in the range: (m-; m+), where m is mean and is standard deviation49. All travel demand variables calculated using data from the Purple Line Stated Preference Survey, were found statistically significant. Exhibit B.14 illustrates the VOT distribution for auto commuter travelers. It is easy to see that 80% of the results fall in the required range. Exhibit B.15 provides the information about the number of survey forms used in calculation of VOT, VOF and VOA by each mode and purpose. Since one transit survey form could be used for both transit VOF and VOT analysis, the total number of forms used for calculation as shown in Exhibit B.14 are higher than the total number of forms received as shown in Exhibit B.5. The Purple Line Stated Preference Survey results of VOT, VOF and VOA calculated for two modes (auto and transit) and three types of purpose (commuter, business and other) are presented in Exhibit B.16 and b4.17. Based on the calculations, the following observations were made:
49
The VOT, VOF and VOA values are larger for auto than for transit; Business trips have larger VOT, VOF and VOA values than commuter and other trips; The VOT, VOF and VOA values are consistent with those of previous studies (e.g., Bay Bridge Travel Survey, 2006) after adjusting to 2010 dollars for similar trip length.
T.T. Soong, Fundamentals of Probability and Statistics for Engineers, John Wiley & Sons Inc., 2004, p204.
Transportation Economics & Management Systems, Inc.
March 2015
77
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE Exhibit B.14: VOT Distribution: Auto Users (Commuter)
Mean VOT=12.12 12% 12%of ofresults resultsare are outside outside negative negativeSD
80% of results are in the In onerange: SD (m - , m+ )
one- SD m
8% of 8% of results resultsare are outside outside positive positive SD
m +
Exhibit B.15: Stated Preference Surveys Used to identify VOT, VOF and VOA (by Mode and Purpose) Business
Commuter
Other
Total
Auto
60
1863
328
2251
Transit
192
900
801
1893
Total
252
2763
1129
4144
Exhibit B.16: Purple Line Stated Preference Survey Results on Value of Time - VOT (Auto and Transit) Value of Time - VOT
Business
Commuter
Other
Auto
$16.83
$12.12
$11.60
Transit
$11.31
$8.91
$7.74
Exhibit B.17: Purple Line Stated Preference Survey Results for Transit for Value of Frequency VOF and Value of Access -VOA (Transit) Transit
Business
Commuter
Other
Value of Frequency - VOF
$8.45
$7.17
$6.17
Value of Access - VOA
$21.62
$17.46
$15.36
Transportation Economics & Management Systems, Inc.
March 2015
78
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE The comparison between Auto VOT values from the 2010 Purple Line and TEMS 2006 Bay Bridge Stated Preference Surveys are shown in Exhibit B.18. The Bay Bridge Auto VOT is inflated from 2006 dollars to 2010 dollars using corresponding CPI index and factored by trip length (i.e., in the Bay Bridge survey the average trip length was 25% longer than that of the Purple Line survey). As seen in Exhibit B.18, factored Bay Bridge Auto VOT values are close to Auto VOT values obtained from the Purple Line survey. Exhibit B.18: Comparison between Auto VOT from 2010 Purple Line and 2006 Bay Bridge Stated Preference Surveys Auto Value of Time (VOT)
Business
Commuter
Other
Bay Bridge Survey, $2006
$19.17
$15.98
$12.46
Bay Bridge Survey, $2010
$20.75
$17.29
$13.48
Bay Bridge Survey, $2010 (factored by Trip Length)
$15.56
$12.97
$10.11
Purple Line Survey, $2010
$16.83
$12.12
$11.60
Transportation Economics & Management Systems, Inc.
March 2015
79
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE
REFERENCES 1. McCarthy, P.S. Transportation Economics Theory and Practice: A Case Study Approach. Oxford: Blackwell Publishers Ltd., 2001. 2. Beesley, M. Urban Transport: Studies in Economic Policy. London: Butterworth, 1973. 3. Mishan, E.J. Cost-benefit Analysis: An Informal Introduction. 4th ed. London: Rutledge, 1994. 4. Mishan, E.J. ‘Rent as a Measure of Welfare Change.’ American Economic Review. 49, June 1958, pp.386-394. 5. Beckmann, M.J. ‘On the Distribution of Urban Rent and Residential Density.’ Journal of Economic Theory. Vol. 1, N1, 1969. 6. Solow, R.M. ‘Congestion, Density and the Use of Land in Transportation.’ Swedish Journal of Economics. Vol. 74, N1, 1972. 7. Lancaster, K.J. ‘A New Approach to Consumer Theory. Journal of Political Economy, 74.2, April 1966, pp. 132-157. 8. Christaller, W. Central Places in Southern Germany. (Translated by Baskin, C. W.) Englewood Cliffs: Prentice Hall, 1966. 9. Loesh, A. The Economics of Location. New Haven: Yale University Press, 1954. 10. Samuelson, P. & Nordhaus, W. Economics. 14th Edition. New York: McGraw-Hill. 1992. 11. Quarmby, D. ‘Travel Mode for the Journey to Work,’ Journal of Transportation Economics and Policy, September 1967 12. Mishan, E. ‘Cost-Benefit Analysis,’ New York, NY: Praeger Publishers, 1976. 13. Metcalf, A.E. ‘Economic Rent: A New Dimension in the Economic Evaluation Process’, Transportation Research Board, 71st Annual Meeting, January 12-16, Washington, DC, 1992. 14. Arab Human Development Report. New York: United Nations Development Program, 2002. 15. Manski, C. and McFadden, D. Structural Analysis of Discrete Data with Econometric Applications. London: MIT Press, 1981. 16. Glaister, S. Fundamentals of Transport Economics. New York: St. Martin’s Press, 1981. 17. Torrens, P.M. ‘How Land-Use Transportation Models Work’, CASA Working Paper Series, 2000, Paper 20, http://www.casa.ucl.ac.uk/. 18. Wilson, A.G. ‘Land-use/Transport Interaction Models’, Journal of Transport Economics & Policy, 1997, Volume 32, pp. 3—26. 19. Kachigan, S.K. Multivariate Statistical Analysis: A Conceptual Introduction. Second Edition. New York: Radius Press, 1991, pp. 176-179. Transportation Economics & Management Systems, Inc.
March 2015
80
PURPLE LINE PRELIMINARY ECONOMIC IMPACT STUDY: UPDATE 20. Metcalf, A. ‘Transport and Regional Development in Ireland.’ Transport and regional development (Ed. By Blonk, W.A.G.) Westmead, England: Saxon House, 1978, pp. 190-208 21. Metcalf, A., Markham, J., Fenney, B. ‘A Price Resistance Model for Personal Travel.’ Regional Studies, Journal of the Regional Studies Association. 10, 1976. pp. 79-88. 22. ‘Ohio Hub Passenger Rail Economic Impact Study’ Prepared for Ohio Rail Development Commission, Ohio DOT. TEMS, Inc. May, 2007. Chapter 7, p. 51. 23. Metcalf, A. ‘Economic Rent: The Supply-Side’s Answer to Consumer Surplus’, 2013
Transportation Economics & Management Systems, Inc.
March 2015
81
