Job Loss and Infrastructure Job Creation
Spending During the Recession
Linda Levine
Specialist in Labor Economics
October 2, 2009
Congressional Research Service
7-5700
www.crs.gov
R40080
CRS Report for Congress
P
repared for Members and Committees of Congress

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Summary
After the long economic expansion that characterized much of the current decade, the nation
entered its eleventh postwar recession in December 2007. The unemployment rate, which is a
lagging economic indicator, did not start to rise until May 2008 when it jumped 0.5 percentage
points to 5.5%. By December 2008, the unemployment rate exceeded 7.0% and well over
600,000 jobs were lost—the biggest monthly decrease since December 1974, when another deep
recession was taking place. These labor market indicators and comments equating the latest
recession to the Great Depression intensified congressional interest in passage of legislation early
in 2009 aimed at encouraging creation of new jobs and warding off further loss of jobs. (See CRS
Report R40655, The Labor Market During the Great Depression and the Current Recession.)
To mitigate all but one recession since the 1960s, Congress chose to increase federal spending on
infrastructure. (See CRS Report 92-939, Countercyclical Job Creation Programs.) But, there are
a number of issues associated with using expenditures on public works to quickly create jobs in
times of recession. (See CRS Report R40107, The Role of Public Works Infrastructure in
Economic Stimulus.)
Public works expenditures traditionally have gone chiefly to construction activities (e.g., building
highways and bridges, dams and flood control structures) which indirectly increase demand in
industries that supply their products to construction firms (e.g., manufacturing). Today, the
definition of infrastructure has been expanded to include green jobs, which include those in
industries that utilize renewable resources (e.g., electricity generated by wind), produce energy-
efficient goods and services (e.g., mass transit), and install energy-conserving products (e.g.,
retrofitting buildings with thermal-pane windows).
A question that typically arises during congressional consideration of economic stimulus
legislation is which approach produces the most bang for the buck. In the instant case, this means
how many jobs might be supported by federal expenditures on traditional and green infrastructure
projects. Once stimulus legislation is signed into law, the focus of Congress customarily turns to
estimates of the number of jobs that result as federal funds are allocated to specific activities.
Therefore, after briefly examining the trend in employment and unemployment since the
recession’s onset, the report turns to an in-depth look at estimates of job creation, including the
limitations of the methodology often used to derive them and the difficulties associated with
developing job estimates for green infrastructure in particular. The report closes with a review of
what is known to date about the number of jobs supported by infrastructure spending among other
provisions in the American Recovery and Reinvestment Act (ARRA, P.L. 111-5). Section 1512
requires entities that receive ARRA appropriations from federal agencies, totaling approximately
$271 billion, to include in quarterly reports the number of jobs created or maintained as a result.
Section 1513 requires the Council of Economic Advisors to report quarterly on the effect of
ARRA provisions on employment and other economic indicators.

Congressional Research Service

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Contents
Employment and Unemployment Through Job Loss.................................................................... 1
Infrastructure Spending and Job Creation Estimates .................................................................... 4
Job Creation Estimates: What Are They?............................................................................... 5
Some Caveats ................................................................................................................. 6
The Multiplier Effect....................................................................................................... 6
Job Estimates and Construction Spending ............................................................................. 7
The Federal Highway Administration .............................................................................. 7
BLS Employment Requirements Table ............................................................................ 8
BEA’s Regional Input-Output Modeling System (RIMS II) ............................................. 9
Job Estimates and Green Infrastructure Spending ................................................................ 10
Measuring Jobs Supported by Spending Provisions in the American Recovery and
Reinvestment Act ................................................................................................................... 12

Tables
Table 1. Payroll Jobs at Nonfarm Employers ............................................................................... 2
Table 2. Number of Payroll Jobs by Industry ............................................................................... 3
Table 3. Number of Direct and Indirect Jobs by State Dependent on an Expenditure of $1
Billion in the Construction Industry.......................................................................................... 9

Contacts
Author Contact Information ...................................................................................................... 13

Congressional Research Service

---

Job Loss and Infrastructure Job Creation Spending During the Recession

fter the long economic expansion that characterized much of the current decade, the
nation entered its eleventh postwar recession in December 2007. The unemployment rate,
A which is a lagging economic indicator, did not start to rise until May 2008 when it
jumped 0.5 percentage points to 5.5%. By December 2008, it exceeded 7.0% according to data
from the U.S. Bureau of Labor Statistics (BLS). Well over 600,000 jobs were lost in December
2008—the biggest monthly decrease recorded by the BLS Current Employment Statistics
program (CES) since December 1974, when another deep recession was taking place.
The Business Cycle Dating Committee of the National Bureau of Economic Research, the official
arbiter of peaks and troughs in the business cycle, announced at the end of November 2008 that a
substantial and widespread decline in economic activity had begun a year earlier. December 2007
marks both the end of the 73-month economic expansion that began in March 2001, and the
beginning of the latest recession. As part of its announcement, the committee noted that it “views
the payroll employment measure, which is based on a large survey of employers, as the most
reliable comprehensive estimate of employment. This series [the CES] reached a peak in
December 2007 and has declined every month since then.”
The committee’s announcement intensified congressional interest in passage of legislation aimed
at encouraging creation of new jobs and warding off further loss of jobs. So, too, did comments
equating the recession to the Great Depression. (See CRS Report R40655, The Labor Market
During the Great Depression and the Current Recession.)
To mitigate all but one recession since the 1960s, Congress chose to increase federal expenditures
on infrastructure (public works), thereby directly raising demand for goods and services to offset
the reduced demand of consumers. (See CRS Report 92-939, Countercyclical Job Creation
Programs.) But, there are a number of issues associated with using spending on public works to
quickly create jobs during a recession. (See CRS Report R40107, The Role of Public Works
Infrastructure in Economic Stimulus.)
When Congress considers spending on infrastructure to help stimulate a flagging economy, “how
many jobs are created” is a commonly asked question. After first briefly examining trends in job
loss since the latest recession began, this report focuses on job creation estimates associated with
increased spending on traditional and so-called green infrastructure, placing a heavy emphasis on
explaining the methodology often used to derive them and the difficulties associated with
developing estimates for green infrastructure in particular.
Once stimulus legislation is signed into law, the focus of Congress customarily turns to estimates
of the number of jobs that result as federal funds are allocated to specific activities. In the case of
the American Recovery and Reinvestment Act (ARRA, P.L. 111-5), Congress included language
requiring entities that receive ARRA appropriations from federal agencies to report the number of
jobs created or maintained as a result and requiring the Council of Economic Advisors to report
on the employment and other economic effects of ARRA provisions. The report closes with a
review of what is known to date about the number of jobs associated with the stimulus act.
Employment and Unemployment Through Job Loss
As shown in Table 1, employment on nonfarm payrolls has steadily declined since December
2007. The number of job cutbacks intensified starting in late 2008. Of the 7.2 million jobs lost
since the recession’s onset, the majority have disappeared since November 2008.
Congressional Research Service
1

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Table 1. Payroll Jobs at Nonfarm Employers
(seasonally adjusted employment in thousands)
Year by Month
Total Employment
Private Sector Employment
2007

December 138,152 115,783
2008

January 138,080
115,689
February 137,936 115,515
March 137,814
115,373
April 137,654
115,203
May 137,517
114,029
June 137,356
114,834
July 137,228
115,691
August 137,053
114,497
September 136,732 114,197
October 136,352 113,813
November 135,755 113,212
December 135,074 112,542
2009

January 134,333
111,793
February 133,652 111,105
March 133,000
110,457
April 132,481
109,865
May 132,178
109,573
June 131,715
109,182
July 131,411
108,936
August 131,210(p)
108,754
September 130,947(p)
108,544
Source: U.S. Bureau of Labor Statistics, data from the Current Employment Statistics program.
Notes: (p) = preliminary.
As is typical during economic downturns, employees in the goods-producing sector have been the
most adversely affected. They saw their ranks shrink by almost 3.6 million between December
2007 and September 2009. (See Table 2.) Workers in the sector’s construction industry began
experiencing job losses before the economy-wide downturn began. Nonetheless, between the
recession’s onset and September 2009, construction firms cut almost 1.5 million jobs. Across all
manufacturing industries, employment fell by 2.1 million as well. Although manufacturing job
losses have been widespread, two industries that produce durable goods—fabricated metal
products (e.g., hardware, wire, and screws) and transportation equipment (e.g., motor vehicles
and parts)—have been particularly hard hit.
Congressional Research Service
2

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Table 2. Number of Payroll Jobs by Industry
(seasonally adjusted employment in thousands)
Industry by
Employment,
Employment,
Sector
December 2007
September 2009 (p)
Goods-producing sector
22,043
18,465
Mining and logging
743
708
Construction 7,523
6,038
Manufacturing 13,777
11,719
Service-providing sector
116,109
112,482
Trade, transportation and
26,725 25,092
utilities
Wholesale trade
6,045
5,649
Retail trade
15,568
14,700
Transportation and
4,555 4,178
warehousing
Utilities 557
565
Information 3,025
2,826
Financial activities
8,243
7,702
Professional and business
18,109 16,597
services
Education and health services
18,570
19,311
Leisure and hospitality
13,551
13,154
Other services
5,517
5,397
Government 22,369
22,403
Source: U.S. Bureau of Labor Statistics, data from the Current Employment Statistics program.
Notes: (p)=preliminary.
Employment in the service-providing sector most recently peaked in December 2007, when the
recession began. Although some service-providing industries have continued to grow—utilities,
education and health services—cutbacks elsewhere have far outweighed their gains. As shown in
Table 2, these industries reported higher employment in September 2009 than at the start of the
recession. In contrast, the financial activities industry began to lose jobs before the advent of the
economy-wide downturn. This mirrors the above-mentioned trend in construction employment in
part because real estate is a component of financial activities and it, like construction, has been
hurt by the collapse of the housing market. Other components of financial activities, such as
brokerage firms that packaged high-risk mortgages and the investors (e.g., banks) that purchased
them, have been negatively affected by the housing market downturn as well.
Despite a marked slowdown in the pace of job loss in recent months and the widely expressed
belief that the recession ended this summer, the prospect of steady job growth beginning in the
near term looks dim according to a CRS analysis of employment trends after the end of the prior
ten recessions. In all but one instance,
Congressional Research Service
3

---

Job Loss and Infrastructure Job Creation Spending During the Recession

“the number of jobs on employer payrolls fluctuated for months.... Sustained job growth
occurred within three to five months of the start of seven recoveries. In sharp contrast, steady
job growth did not commence until March 1992—12 months after the July 1990-March 1991
recession ended—and not until September 2003—22 months after the March-November
2001 recession ended.”1
The unemployment rate in September 2009 rose to 9.8% from 5.0% in December 2007, according
to BLS data derived from the Current Population Survey.2 The Blue Chip Economic Indicators
reported that the latest consensus forecast among the nation’s leading business economists is that
the unemployment rate will continue to rise to 10% or higher and “recede from that level only
grudgingly” during the last six months of 2010. Most of these economists think it will not “be
until the second half of 2012, or later, before the unemployment rate falls below 7 percent on a
sustained basis.”3
Infrastructure Spending and Job Creation Estimates
When in response to a recession Congress has acted to create jobs by raising demand for goods
and services through increased federal spending, it often has chosen to direct the funds to
infrastructure (public works) activities. Other means of direct countercyclical job creation—
employment tax credits, state revenue-sharing, and public service employment—have been relied
on much less often.4
A more expansive definition of infrastructure than was used in the past is now under
consideration. Historically, public works has been synonymous with heavy and civil construction
activities (e.g., road and bridge building, flood control structures and dam building). Today, it
includes so-called green jobs. Although numerous studies on the emerging green economy have
been released in the last several years, no consistent definition of green jobs exists at present.
Green jobs seemingly are those in and related to industries that utilize renewable resources to
produce their outputs (e.g., energy generated by wind, solar, and geothermal technologies) and
jobs in and related to industries that produce energy-efficient goods (e.g., Energy Star appliances
and equipment) and services (e.g., intra- and inter-city mass transit).5 For this reason, the
following discussion focuses on what is known about the job-generating impact of infrastructure
spending broadly defined.
The section below begins with an in-depth examination of how job creation estimates usually are
developed. The focus then narrows to look at two models that can be used to calculate the number
of jobs nationwide dependent upon demand in the construction industry among other industries,
and one model that can be used to calculate the number of jobs by state dependent on the
construction industry among other industries. The section ends by reviewing the difficulties that
researchers encounter in estimating the number of jobs supported by expenditures on green

1 CRS Report R40798, Unemployment and Employment Trends Before and After the End of Recessions, by Linda
Levine.
2 Data from the Current Population Survey of households is available at http://stats.bls.gov/cps.
3 “Recession Finished, Say Blue Chip Analysts,” Daily Labor Report, September 10, 2009.
4 CRS Report 92-939, Countercyclical Job Creation Programs, by Linda Levine.
5 Related jobs include, for example, those in industries that manufacture wind turbines and install thermal-pane
windows.
Congressional Research Service
4

---

Job Loss and Infrastructure Job Creation Spending During the Recession

infrastructure and the consequent caution that should be taken when utilizing these estimates in
particular.
Job Creation Estimates: What Are They?
Interest in how many jobs are created by a particular type of economic activity has surfaced when
the economy is in a downturn and policymakers seek to compare the relative advantages of
different stimulus options. It also has arisen when policymakers want to know the impact of
shifting expenditures from one federal budget category to another (e.g., away from defense and
towards social services programs). Unless there is an increase in total spending, however, the
number of jobs in the labor market would remain largely unchanged.6
Although there are other bases upon which to develop estimates of the number of jobs created by
a given economic activity, an input-output (I-O) model of the economy often is utilized due to its
cost-effectiveness.7 An I-O model describes the interrelationships between industries in the
production process, showing how the dollar value of a sale is distributed across industries at a
particular point in time. It thus reflects how much of the purchased product comes from final and
supplier industries. An I-O table might show, for example, the dollar value of roof trusses
produced by the veneer, plywood, and engineered wood products manufacturing industry and the
dollar value of bricks produced by the clay product and refractory manufacturing industry used by
the construction industry to erect residential buildings.
The output requirements from each industry must then be converted to employment requirements.
Employment requirements are derived from productivity estimates for each industry at a
particular point in time. The total employment requirement associated with a given type of final
demand (e.g., a water reuse program) is the employment in the industry producing the final
product or service and in the supplier industries. In other words, it is an approximation of both the
direct and indirect employment dependent upon/supported by the economic activity. It commonly
is expressed as the number of jobs per billion dollars of expenditures valued in a particular year’s
dollars.
Like an I-O table, an employment requirements table is a matrix of hundreds of columns and
rows. Each column displays the number of jobs supported in each of the industry rows by an
expenditure of one billion dollars in the column industry. For example, one billion dollars spent in
the construction industry supports (direct) employment in the various components of that industry
(e.g., residential and commercial building, highway and bridge building) and (indirect)
employment in the many industries that supply their goods and services to the construction
industry (e.g., asphalt shingle manufacturing, fabricated metal bridge section manufacturing). An
employment requirements table thus permits estimation of the varying impact of an expenditure
on different industries and the varying impact of different kinds of expenditures.

6 Small differences in the total number of jobs could occur at the same spending levels if the economic activities to
(from) which funds were being shifted were more (less) capital-intensive, for example.
7 Another basis for estimating the impact of policy and other changes on the economy is conducting surveys. According
to the U.S. Bureau of Economic Analysis (BEA), the advantage of the I-O approach to making impact estimates is the
accessibility of the data sources required to develop the I-O model.
Congressional Research Service
5

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Some Caveats
I-O models freeze technology and productivity at a particular point in time. Thus, the job-
generating potential of an economic activity undertaken today could differ from that of an earlier
period if there were technological and productivity improvements in the intervening years.
Similarly, the estimates often are stated in terms of the number of jobs created for every billion
dollars of expenditures, but a billion dollars spent in one year could buy less (more) than a billion
dollars spent in another year depending on changes in price levels over time.
There also could be differences in estimated versus actual job creation because I-O models
assume that resources are unlimited. If, for example, the economy was performing at a fairly high
level with plants operating near full capacity and with fairly few workers unemployed, the actual
number of new jobs might fall short of the estimate due to capital and labor constraints. This is
less likely to matter during a broad-based economic downturn.
Further, I-O tables do not necessarily differentiate between imported and domestically produced
goods. As a consequence, the domestic employment impact of expenditures might be overstated
to the extent that inputs are imported. Similarly, I-O tables typically do not express employment
in terms of full-time equivalents (i.e., both full-time and part-time jobs are counted equally).
Thus, programs which draw upon industries that rely relatively more on part-time workers (e.g.,
retail trade) might appear to create more jobs than programs that draw to a greater extent on
industries employing relatively more full-time workers (e.g., manufacturing).
The Multiplier Effect
A more comprehensive estimate of the number of jobs created by a particular type of economic
activity has three components, namely,
• the number of jobs directly attributable to the activity,
• the number of jobs indirectly attributable to the activity, and
• the number of jobs induced throughout the economy as a result of the activity.
Induced jobs are those dependent upon the purchases of persons in direct and indirect jobs. For
example, workers who are directly or indirectly employed as the result of a highway construction
program might spend some portion of their wages in their communities at grocery stores, auto
repair shops, and movie theaters.
Estimates of induced jobs or the multiplier are considered tenuous. To calculate the multiplier
effect, one must estimate how much of the additional money earned by directly and indirectly
employed workers will likely be spent versus saved. The actual number of jobs created by this
added spending will further depend on economic conditions (e.g., the availability of labor, the
inflation rate). As a result, there are widely varying estimates of the multiplier effect and those job
creation studies that include induced employment utilize different multipliers.
Congressional Research Service
6

---

Job Loss and Infrastructure Job Creation Spending During the Recession

Job Estimates and Construction Spending
The Federal Highway Administration
Perhaps the most widely known estimate of the employment impact of federal spending on our
nation’s roads comes from the Federal Highway Administration (FHWA). Although the FHWA
twice updated its 1997 analysis, which estimated that $1 billion of federal-aid highway
expenditures plus a $250 million state match supported 47,575 jobs, some proponents of
stimulating job growth through increased federal spending on infrastructure continue to use this
figure. The most recent update by the FHWA to 2007 indicates that a $1.25 billion expenditure on
highway construction consisting of $1 billion from the federal government and $250 million from
state government could support 34,779 jobs. If a state match is not required, “then $1 billion in
Federal funds supports 27,800 jobs.”8 The jobs number has decreased over time in part because of
increases in the price of inputs, such as asphalt and diesel fuel.
The FHWA breaks down the estimate of 27,822 jobs per billion dollars of federal spending on
highways as follows:
• 9,536 construction-oriented jobs (i.e., jobs at construction companies working on
the projects and at businesses that provide direct inputs to the projects such as
asphalt, concrete, and guard rails);
• 4,324 jobs in supporting industries (i.e., employment at firms that provide inputs
to the industries directly providing the materials and equipment utilized in
highway construction such as producers of sheet metal who supply the
manufacturers of guard rails); and
• 13,962 induced jobs (i.e., jobs throughout the economy dependent upon
consumer expenditures from the wages of workers in “construction-oriented” and
“industry-supporting” jobs).
Thus, the multiplier effect accounts for one-half of the total estimate.
The FHWA notes one caveat about I-O analysis in addition to those mentioned above, that is, the
job estimate “utilizes the national average mix of construction materials and labor inputs. Specific
projects and local utilization ratios will alter the estimated number of jobs supported.”9 For
example, a different combination of materials and number of workers might be required for road
resurfacing projects compared to bridge building or commuter rail projects.
The FHWA also states that
[t]he employment figures have recently been used as a justification for including highway
spending in an economic stimulus package. But with the exception of short-term resurfacing

8 U.S. Department of Transportation, Federal Highway Administration, Employment Impacts of Highway Infrastructure
Investment, p. 1, http://www.fhwa.dot.gov/policy/otps/publications.htm.
9 U.S. Department of Transportation, Federal Highway Administration, Employment Impacts of Highway Infrastructure
Investment, p. 2, http://www.fhwa.dot.gov/policy/otps/publications.htm.
Congressional Research Service
7

---

Job Loss and Infrastructure Job Creation Spending During the Recession

and preservation projects, highway funds spend out slowly, with only 27% of a project, on
average, outlaying in the first year.10
BLS Employment Requirements Table
In recognition of the fact that “people want to assess the impact on employment of different
policies or actions,” the U.S. Bureau of Labor Statistics (BLS) makes available electronically
free-of-charge to the public the employment requirements tables it develops as part of its
employment projections program.11 I-O and employment requirements tables developed and
utilized by others often are proprietary and not made widely available.
The employment requirements tables are based on the official I-O tables for the nation that the
U.S. Bureau of Economic Analysis (BEA) develops every five years. BLS takes the latest
national I-O table available from BEA – in this case, 1997 – and updates it to reflect more recent
production and distribution technologies. It then utilizes the updated I-O table and recent labor
productivity data to develop an employment requirements table. Because the base year for the
most recently published employment projections is 2006, the latest employment requirements
table reflects 2006 technologies of production and distribution as well as labor productivity.
The BLS employment requirements table provides information for the construction industry as a
whole. The construction industry, according to the North American Industry Classification
System, is composed of three major subdivisions:
• construction of buildings (residential and nonresidential),
• heavy and civil engineering construction (highway, street, and bridge
construction; utility system construction; construction of flood control structures,
dams, and hydroelectric power generation facilities), and
• specialty trade contractors (foundation, structure, and building exterior
contractors; building equipment contractors; building finishing contractors).
The BLS employment requirements table shows 11,768 jobs directly and indirectly dependent
upon $1 billion of spending on construction. A majority of the jobs are in the construction
industry itself (i.e., 6,925 direct jobs).
The figure from the BLS employment requirements table for construction expenditures (11,768)
is somewhat lower than the direct and indirect jobs figure for highway expenditures from the
FHWA (13,860). Potential explanations for the disparity include differences in industry definition,
data sources, method of updating the model, and time period.
The employment requirements available from BLS do not break out other types of construction
that have been discussed as part of a federal job creation package (e.g., public school
construction). BLS formerly conducted surveys to estimate full-time year-long employment
associated with a variety of different construction activities, including new schools, hospitals,

10 U.S., Employment Impacts of Highway Infrastructure Investment, U.S. Department of Transportation, Federal
Highway Administration, p. 2, http://www.fhwa.dot.gov/policy/otps/publications.htm.
11 U.S. Bureau of Labor Statistics, Layout and Description for 201-order Employment Requirements Tables,
Washington, D.C., December 2007, p. 3, http://stats.bls.gov/emp/empind4.htm.
Congressional Research Service
8

---

Job Loss and Infrastructure Job Creation Spending During the Recession

water and sewer facilities, roads, mass transit, and maintenance and repair construction. The
survey information was last updated a few decades ago, however.
BEA’s Regional Input-Output Modeling System (RIMS II)
From its Regional Input-Output Modeling System (RIMS II), the BEA produces estimates by
geographic area of the employment, earnings, and output dependent on additional spending in
hundreds of different industries. 12 For a fee to most parties, BEA currently utilizes either the 1997
benchmark I-O for the nation or the 2006 annual I-O for the nation adjusted by 2006 data from its
regional economic accounts to provide these estimates at the subnational level.13
As shown in Table 3, the number of jobs directly and indirectly supported by an expenditure of
$1 billion in the construction industry in a given state ranges widely. The main reason for the
disparity in job creation estimates is that each state has a different mix of industries within its
borders. As a consequence, one state varies from the next in its capacity to supply all the
intermediate goods needed to carry out construction projects. A secondary explanation is that
earnings vary by state.
Table 3. Number of Direct and Indirect Jobs by State Dependent on an Expenditure
of $1 Billion in the Construction Industry
State
Number of Jobs
State
Number of Jobs
Alabama 15,851
Montana 16,127
Alaska 11,009
Nebraska
13,946
Arizona 12,238
Nevada 11,459
Arkansas 15,306
New
Hampshire
12,374
California 12,289
New
Jersey 11,118
Colorado 12,575
New
Mexico
14,279
Connecticut 10,709
New
York 10,106
Delaware 9,518
North
Carolina
15,555
District of Columbia
1,874
North Dakota
13,500
Florida 13,127
Ohio 14,391
Georgia 14,224
Oklahoma
16,232
Hawai 11,614
Oregon 13,184
Idaho 15,860
Pennsylvania
12,390
Illinois 11,916 Rhode
Island
10,767
Indiana 13,747
South
Carolina
15,319

12 For additional information on RIMS II see BEA, Regional Multipliers: A User Handbook for the Regional Input-
Output Modeling System, http://www.bea.gov/scb/pdf/regional/perinc/meth/rims2.pdf.
13 More specific detail by industry is available from the 1997 benchmark I-O than from the annual I-O. Therefore,
Table 1 (Number of Direct and Indirect Jobs Per $1 Million of Output Produced by the Water, Sewage and Other
Systems Industry) in CRS Report R40107, The Role of Public Works Infrastructure in Economic Stimulus, was drawn
from the 1997 benchmark I-O because the 2006 annual I-O provides data only for the utilities industry as a whole.
Congressional Research Service
9

---

Job Loss and Infrastructure Job Creation Spending During the Recession

State
Number of Jobs
State
Number of Jobs
Iowa 14,330
South
Dakota
15,316
Kansas 13,625
Tennessee
14,556
Kentucky 15,039
Texas
12,985
Louisiana 13,731
Utah
14,692
Maine 15,988
Vermont
14,883
Maryland 10,687
Virginia
12,085
Massachusetts 10,714
Washington
12,171
Michigan 13,354
West
Virginia
13,834
Minnesota 12,998
Wisconsin 13,673
Mississippi 15,357
Wyoming 13,091
Missouri 13,241
United
States
14,315
Source: Prepared by CRS from RIMS II estimates supplied by the BEA Regional Product Division.
Job Estimates and Green Infrastructure Spending
Estimating the number of jobs dependent upon green infrastructure activities presents a greater
challenge than estimates related to infrastructure projects as traditionally defined. The basis for
most data collection by U.S. statistical agencies is the North American Industry Classification
System (NAICS). It currently does not identify separately so-called green industries (e.g., those
that utilize renewable resources to produce their outputs, those that manufacture goods which
minimize energy use). For example, the NAICS disaggregates the electric utility industry into
hydroelectric, fossil fuel, nuclear, and other power generation, transmission, and distribution.
Such renewable sources of energy production as wind, solar, and biomass are not uniquely
recognized; they are included in the “other” category. If harnessing the wind to produce
electricity and plant material to produce biofuel requires a substantially different mix of inputs
than relying on coal and gasoline, for example, the conventional I-O model does not seem well-
suited as a basis for estimating the number of jobs supported by these green activities. Similarly,
within NAICS, the construction industry does not have a unique category for retrofitting (e.g.,
installing additional insulation, fluorescent lighting, or energy-efficient heating and air-
conditioning systems). Retrofitting likely requires a combination of inputs from supplier
industries that differs from the mix for the top-to-bottom construction of buildings, once again
making use of conventional I-O models problematic.
This recognized difficulty generally is either not mentioned, or how it is dealt with is not
described, in the analyses of green job creation. One study, commissioned by the Center for
American Progress that is discussed in more detail below, does address the problem. The
researchers explain that because “the U.S. government surveys and accounts that are used to
construct the input-output tables do not specifically recognize wind, solar, biomass, building
retrofitting, or new mass transit as industries in their own right,” they created synthetic industries
by combining parts of industries for which data are available. The researchers provided an
example in the case of the biomass “industry:” they constructed it by combining farming, forestry,
Congressional Research Service
10

---

Job Loss and Infrastructure Job Creation Spending During the Recession

wood products, and refining industries; then they “assigned relative weights to each of these
industries in terms of their contributions to producing biomass products.”14
Further complicating the matter is the context and manner in which estimates of green jobs
generally are presented. Studies often develop employment projections based on differing sets of
assumptions and time horizons. For example, the number of direct and indirect jobs some 10 or
more years in the future supported by an assumed increase in the demand for energy that is met
by an assumed shift during the projection period from coal to wind and geothermal power
generation. Some reports also include induced employment, but this is not always made clear. In
addition, some analyses relate to a particular state. Their results may not be generalizeable to
other areas because state economy’s have different mixes of industries and may not be able to
provide any or all of the inputs for a particular green output. Additionally, the assumptions and
methodologies underlying the job creation estimates often are not clearly articulated, which
makes thoughtful review of the results very difficult.
It should be noted that many of the studies by green economy proponents that were available
when Congress was crafting stimulus legislation had not been conceived for the purpose of
quickly stabilizing or increasing the number of jobs in the nation or in industries particularly hard
hit by the recession. Job creation estimates from two organizations that proposed broad-based
green economy strategies intended in part to stimulate the deteriorating labor market are briefly
described below.
• The September 2008 report, Green Recovery: A Program to Create Jobs and
Start Building a Low-Carbon Economy, was commissioned by the Center for
American Progress (a research and educational institute). It represents an
acceleration of a 10-year program included in a 2007 report (Capturing the
Energy Opportunity: Creating a Low-Carbon Economy). The 2008 report’s
authors at the Department of Economics and Political Economy Research
Institute (University of Massachusetts – Amherst), who relied on I-O analysis,
estimate that almost 2 million jobs (935,200 direct jobs, 586,000 indirect jobs,
and 496,000 induced jobs) could be created or preserved by a two-year $100
billion “green economic recovery program.” Of the $100 billion total, $46 billion
would be in the form of federal spending for such activities as public building
retrofits, mass transit and freight rail expansion, and smart electrical grid
development. Much of the remainder would be in the form of tax credits to
encourage businesses and homeowners to retrofit commercial and residential
buildings. The authors acknowledge that not all of the green activities
can contribute equally to a short-term green economic recovery program. Some ... strategies
are clearly capable of delivering within a year, while others will require as long as two years
to be implemented.15
• In December 2008, the Apollo Alliance (a coalition of labor, environmental,
business and community leaders) proposed The Apollo Economic Recovery Act.

14 Robert Pollin, Heidi Garrett-Peltier, and James Heintz, et al., Green Recovery: A Program to Create Good Jobs and
Start Building a Low-Carbon Economy, Center for American Progress, Washington, D.C., September 2008, p. 20,
http://www.americanprogress.org.
15 Robert Pollin, Heidi Garrett-Peltier, and James Heintz, et al., Green Recovery: A Program to Create Good Jobs and
Start Building a Low-Carbon Economy, Center for American Progress, Washington, D.C., September 2008, p. 5,
http://www.americanprogress.org.
Congressional Research Service
11

---

Job Loss and Infrastructure Job Creation Spending During the Recession

It is an initial step toward achievement of a 10-year $500 billion program to
create 5 million green-collar jobs, which had been released in September 2008.
The new initiative calls for federal spending of about $50 billion to create or
maintain more than 650,000 direct jobs and 1.3 million indirect jobs. The
derivation of these job creation figures is not always clear, appearing to rely
much of the time on spending-to-jobs relationships estimated by other
organizations. The proposed allocation of federal funds and associated job
estimates include $10 billion to improve the efficiency and reliability of the
electric transmission grid (131,000 direct and indirect jobs), $8 billion to repair
roads and bridges (278,000 direct and indirect jobs), and $8 billion to encourage
localities to replace aging buses and trains with U.S.-made clean-energy vehicles
(37,600 direct jobs in vehicle manufacturing and 167,000 indirect jobs).
Measuring Jobs Supported by Spending Provisions
in the American Recovery and Reinvestment Act
While crafting the American Recovery and Reinvestment Act (ARRA), Congress was concerned
about timely tracking of the number of jobs whose creation or maintenance results from the
legislation. The 111th Congress therefore addressed this matter in bill language much more than
prior Congresses had in countercyclical job creation legislation. At Title XV—Accountability and
Transparency of Division A—Appropriations Provisions, P.L. 111-5 requires entities that receive
ARRA appropriations from federal agencies (e.g., grant, loan, or contract recipients; states) to
include in their quarterly reports to the agencies estimates of the number of direct jobs created
and retained by infrastructure projects, for example. Recipients of recovery funds must make their
first submission of the required information in October 2009, and the agencies must post the
contents of the reports on websites 30 days after the end of each calendar quarter. The Office of
Management and Budget (OMB) is directed to provide guidance to help recipients prepare the
reports, including the development of job estimates.16 The act further charges the Congressional
Budget Office and the Government Accountability Office with commenting on the job estimates
contained in the reports within 45 days after they have been submitted to federal agencies.
P.L. 111-5 additionally tasks the Council on Economic Advisers (CEA) with submitting quarterly
reports to the Committees on Appropriations on the effect of ARRA-provisions on employment
and other economic indicators. The CEA’s mandate thus extends well beyond the above-described
reporting requirements, which apply only to $271 billion in direct government investment
spending out of a total of $787 billion.17 Accordingly, as well as utilizing the direct job estimates
provided by recipients of investment spending under ARRA, the CEA anticipates incorporating
into its macroeconomic model the data the Treasury and the OMB are collecting weekly on tax
cuts and other spending.

16 Office of Management and Budget, Implementing Guidance for the Reports on Use of Funds Pursuant to the
American Recovery and Reinvestment Act of 2009, M-09-21, Washington, D.C., June 22, 2009.
17 The remaining ARRA funds fall into five categories, as the Council of Economic Advisors states in Estimates of Job
Creation from the American Recovery and Reinvestment Act of 2009 (May 2009, p. 2): “individual income tax cuts; a
two-year patch to the alternative minimum tax; investment incentives; aid to people directly hurt by the recession; and
state fiscal relief.”
Congressional Research Service
12

---

Job Loss and Infrastructure Job Creation Spending During the Recession

The first quarterly report of the CEA was issued in September 2009. Based on two different
estimating procedures, it found that ARRA might have added some one million jobs to employer
payrolls in August 2009 compared to what employment would have been in the absence of the
legislation. The CEA focused on the impact of the state fiscal relief contained in ARRA because
$38.4 billion had been provided to the states by August, which “represents almost half of outlays
and one-quarter of total ARRA stimulus (that is, outlays plus tax reductions).”18 A positive
relationship was estimated between the fiscal relief thus far provided to states and employment in
state and local government as well as in the education and health care industries.

Author Contact Information

Linda Levine

Specialist in Labor Economics
llevine@crs.loc.gov, 7-7756





18 Council of Economic Advisers, The Economic Impact of the American Recovery and Reinvestment Act of 2009, First
Quarterly Report, Washington, DC, September 10, 2009, p. 33.
Congressional Research Service
13

---