EPA Methodology for Power Sector-Specific Employment Analysis


EPA METHODOLOGY FOR POWER SECTOR-SPECIFIC

EMPLOYMENT ANALYSIS

MAY, 2018

-------
Contents

Part I: Introduction and Overview of Methodology	1

Objective	1

Scope of the analysis	1

Overall Methodology: 4 Steps	1

Step 1: Quantify projected actions from power sector modeling	2

Step 2: Estimate changes in resource requirements	2

Step 3: Estimate Labor Impacts	3

Step 4: Aggregate Labor Estimates	3

Part II: Detailed Methodology	4

1.	Quantify Projected Actions	4

1.1.	Retrofit Capacity	4

1.2.	Changes in new generation capacity	5

1.3.	Fuel use changes	5

1.4.	Changes in Existing Capacity	6

2.	Estimate Resource Requirements for Changes in Projected Actions	6

2.1.	Construction of New Generation Capacity	6

2.2.	Operation of New Generation Capacity	8

2.3.	Installation of New Pollution Control Retrofit Technology	8

2.4.	Operation of Newly Installed Retrofits	12

2.5.	Retirements	13

2.6.	Fuel Production and Pipeline Construction	14

3.	Estimate Labor Impacts	14

3.1.	Annual Job-Year	14

3.2.	Product Prices	14

3.3.	Labor productivity	15

4.	Aggregate Labor Impact Estimates	18

4.1.	New Generation Capacity	18

4.2.	New Pollution Retrofits	19

4.3.	Fuel Use Changes	20

4.4.	Pipeline Construction	20

4.5.	Changes to Existing Capacity	20

References	22

-------
Part I: Introduction and Overview of Methodology

Objective

Over the past decade, ICF has conducted numerous bottom-up employment analyses for EPA in
support of several rulemakings. While the methodology specific to each analysis has been
documented in conjunction with the relevant rulemaking, this document provides a
comprehensive summary of the general methodology for estimating the impact on labor
between any two modeling scenarios.

Scope of the analysis

EPA's labor analyses have focused on the potential impacts of air pollution regulations on the
direct changes in the amount of labor needed in the power generation sector and directly
related sectors (e.g., equipment manufacturing, fuel supply, and generating efficiency services).
The same approach can be applied to estimate labor changes between any two modeling
scenarios. In short, this approach converts changes in projections of the resources required for a
certain scenario (e.g., pollution controls) to estimates of changes in labor demand. This
approach does not address another potential change in labor demand, i.e. if a regulation causes
marginal cost to increase, placing upward pressure on output prices, decreasing quantity
demanded and production, this can cause a decrease in factor demands, such as labor.1
Generally, this conversion of resource requirements relies on the translation of non-monetary
units into monetary estimates, and subsequently those monetary estimates into labor demand.

This approach can be characterized as an evaluation of "first order employment impacts" using a
partial equilibrium modeling approach. It does not include the potential ripple effects of these
impacts on the broader economy. These ripple effects are generally classified as "multiplier"
impacts and include the secondary job impacts in both upstream and downstream sectors. This
approach also excludes the economy-wide effects of changes to energy markets (such as higher
or lower forecasted electricity prices) that would be included in a more general equilibrium
modeling context.

Overall Methodology: 4 Steps

The overall methodology for these analyses consists of four sequential steps:

1 Since electricity demand is relatively inelastic, and industry output may not change much, this output effect on
labor demand may be small, overall, for the power sector. The distribution of labor demand changes may vary
across facilities within the power sector.

-------
1) Quantify the change in projected actions between two modeling scenarios.

2) Estimate the resources required for those actions based on EPA's engineering estimates
of the per-unit resources required to implement those changes.

3) If resource needs are not already reported in terms of the labor required, estimate the
labor needs under each analysis by converting the resource needs into a monetary value
which is then converted to labor needs using labor productivity data.

4) Combine this information to generate the first-order labor impacts.

Each of these steps are summarized below and discussed in more detail in the numbered
sections in Part II.

This approach estimates labor changes measured as the change in job-years in each analysis
year. Job-years are not individual jobs, and they are not necessarily permanent or full-time jobs.
Job-years are the amount of work performed by one full time equivalent (FTE) employee in one
year. For example, 20 job-years may represent 20 full-time jobs or 40 part-time jobs in a given
year, or any combination of full- and part-time workers such that total is equivalent to 20 FTE
employees.

Step 1: Quantify projected actions from power sector modeling

The first step is to quantify the changes projected to occur between scenarios. To do so, we compare
projections of a two or more scenarios in order to estimate the changes projected to occur. These
actions typically include changes in pollution control retrofits, changes in new and existing generation
capacity, and fuel use changes.

Step 2: Estimate changes in resource requirements

After the changes in projected actions are quantified, the next step estimates the resources
required to implement these actions. These resources can be physical resources such as tons of
steel, monetary inputs such as dollars required to install or run a compliance technology, or the
labor hours required to implement a compliance change. In general, resource needs for many of
the compliance changes are broken down into two main categories: those needed for
construction or installation and those for operating the resources.

Construction-related resources are relevant for building new generation capacity or installing
new retrofits. To estimate these resources, we use data and assumptions on the specific types of
labor categories required (e.g., boilermakers) along with the physical quantities of total
resources needed, such as total labor hours required for a specified capacity of the new
equipment. In addition, we also use information on the indirect labor needs to construct and
install new equipment, such as those for steel or initial catalyst load in certain types of pollution
control equipment.

We also estimate the resources required to operate new generating capacity and retrofits. For
new generation capacity, the cost of operation and maintenance is estimated based on the

-------
projected capacity increase. For pollution control retrofits, the physical resources required for
operation are estimated based on assumptions regarding the necessary inputs for each type of
control (e.g., tons/MWh of limestone used in scrubbers).

This step also involves estimating resource changes related to generator retirements. Under this
step the resources no longer needed due to retirement are the reduction in operating and
maintenance expenditures on the facility retiring; any employment impacts such a retirement
has related to its fuel consumption are captured separately through our assessment of projected
differences in total fuel consumption under the policy being analyzed.

Estimates of fuel use changes are based directly on differences in model projections. We also
estimate the cost to construct additional pipeline capacity needed to accommodate projected
increases in natural gas demand.

Step 3: Estimate Labor Impacts

The next step is to convert the estimated resource requirements into estimated labor impacts, if
not already reported in terms of labor requirements. We first convert the estimated physical
quantities of required resources into monetary amounts, where necessary, using market prices
for the resource.

Next, we apply labor productivity values to the estimated monetary amounts in order to
estimate labor impacts. We derive labor productivity values based on the North American
Industry Classification System (NAICS)-based sectors that produce the relevant resources
estimated in the previous step. Labor productivity quantifies the amount of labor required to
produce a dollar amount of output. In addition, because labor estimates are being made for
future years, labor productivity growth rates were also calculated based on historical data. These
growth rates are used to project labor productivity to future years.

Step 4: Aggregate Labor Estimates

The final step combines the information developed in the previous steps to create the labor
estimate representing the first-order policy impact as direct labor changes resulting from
projected actions.

-------
Part II: Detailed Methodology

1. Quantify Projected Actions

The following table contains a list of all outputs obtained from the power sector modeling that
are used as inputs to this analysis.

Table 1: Power Sector Modeling Outputs

Modeling Output Category

Technology

Units

New Generation Capacity

Biomass

Cumulative GW

Solar

Wind

Coal

Natural Gas Combined Cycle
(CC)

Natural Gas Combustion
Turbine (CT)

Retirements

Natural Gas Combined Cycle

Cumulative GW

Coal

Pollution Control Retrofits

FGD

Cumulative GW

SCR

SNCR

ACI

DSI

HRI

Combustion Control

Installation Costs

Fuel Consumption

Appalachian Coal

Million tons

Interior Coal

Western Coal

Waste Coal

Natural Gas

Trillion cubic feet

1.1. Retrofit Capacity

One of the largest drivers of employment growth in policies analyzed to date is the installation of
new pollution control equipment. Retrofit technologies evaluated to date include:

4

-------
•	Flue gas desulfurization, or "scrubbers" (FGD)2

•	Selective Catalytic Reduction (SCR)

•	Selective Non-Catalytic Reduction (SNCR)

•	NOx Combustion Controls

•	Activated Carbon Injection (ACI)

•	DrySorbent Injection (DSI)

•	Fabric Filters (FF)

•	Electrostatic Precipitators (ESP)

•	Heat rate improvements (I IRI)

1.2.	Changes in new generation capacity

Projected changes in new electric generating capacity might include coal steam, natural gas
combined cycle, natural gas combustion turbine, solar, wind, and biomass.

1.3.	Fuel use changes

We account for changes in coal demand regionally because Appalachian and Interior coal require
more labor to mine the same amount of coal as Western coal, such that they have notably
different, labor productivity values (that are applied regionally in the following step).

Figure 1: Regional Coal Production in the U.S.

2 Note that this labor analysis does not make a distinction between wet and dry scrubbers,

5

-------
We evaluate projected changes in natural gas consumption at a national level. Additionally, since
increases in natural gas demand likely correlate with new pipeline demand, in this step we also
estimate potential pipeline construction that could accompany projected increases in natural gas
demand.

1.4. Changes in Existing Capacity
In order to estimate employment changes related to plant closures, we assume that the fixed
operating costs (FOM) that are no longer paid for retired units can be used as a proxy for ceased
economic output due to retirements.3

2. Estimate Resource Requirements for Changes in Projected Actions
In this step, "resource" refers to any inputs associated with changes in projected actions
between modeling scenarios. These resources can be physical resources such as tons of steel,
monetary inputs measuring cost of installing or running a pollution control device, or the direct
labor hours. This section describes the resources that are estimated, assumptions made, and
details how projected changes are converted into resource requirements. This section covers the
two main types of resources: resources for construction and resources for operation.

2.1. Construction of New Generation Capacity
The resource requirements for new capacity are expressed in terms of the costs of equipment,
material, labor, and special technical services, such as engineering and construction
management. Calculating these resource requirements occurs in two steps. The first step
converts projections of annual capacity (MW) into the cost of construction. Next, the total
construction cost value is broken down into shares for equipment, material, labor, and
specialized labor.

For each technology, estimates about construction cost and time to completion are based on
EPA's assumptions as documented in EPA (2018). To estimate an annual cost, the total
construction cost was divided by the number of years assumed to complete construction.

The annual construction costs of new generation capacity can be summarized as:

Capacity (MW) x Capital Cost x 1000

Annual Construction Cost = — -

Build Duration (years)

The assumed capital cost ($/kW) and build duration values are presented in the table below.

3 FOM costs are treated as a proxy for retirement-related job losses occurring at the power plant only.

-------
Table 2 Capital Cost and Build Duration Assumptions:

New Capacity Type

Capital Cost
(2016$/kW)

Build Duration

Combined Cycle

1,081

3 Years

Combustion Turbine

662

2 Years

Onshore Wind

1,404

3 Years

Solar PV

1,034

3 Years

Source: EPA (2018)

The total construction costs for each generation technology are then divided into the specific
capital and labor components for expenditures on equipment, material, labor, and engineering
and construction management. The breakdown into these categories is provided in the table
below.

Table 3: Capital and Labor Contents for New Generation Capacity

New Capacity Type

Equipment

Material

Labor

Engineering and
Construction
Management

Renewables +
Biomass

54%

6%

31%

9%

Combined Cycle

65%

10%

18%

7%

Combustion
Turbine

65%

10%

18%

7%

Sources: EPA (2002) and Staudt (2011)

Dividing the annual construction costs into the constituent categories results in estimates of
expenditures on equipment, material, and labor required for construction of new generation
capacity.

For illustrative purposes, an example calculation to estimate the resources required for a 500-
MW combined cycle unit is presented in the table below.

7

-------
Table 4: Estimate Required Resources: New Generation Capacity Example (500-MW Combined Cycle)

Assumptions

Calculation

Resource
Estimate

1,006
$/kW

3 Year Build
Duration

l,006-^-*500 AfW*l,000
kW

= $167,666,667
per year

65% Equipment

= $108,983,333

10% Material

= $16,766,667

18% Labor

= $30,180,000

3 Years

7% Engineering and
Construction
Management

= $11,736,667

2.2. Operation of New Generation Capacity
The operation of new generation capacity requires resources that can be estimated based on the
annual cost of operation and maintenance. We estimate annual resource requirements for
operating new generation capacity based on these annual fixed operating and maintenance
(FOM) costs. These assumptions are summarized in the table below.

Table 5: New Generation Capacity Fixed Operation and Maintenance Cost Assumptions

New Capacity Type

FOM Cost
(2016$/kW-yr)

Combined Cycle

9.9

Combustion Turbine

6.8

Onshore Wind

49.46

Solar PV

11.35

Source: EPA (2018)

Using the above FOM cost assumptions, we estimate operating costs by multiplying the FOM
cost by the projected capacity of new generation for each technology. The table below
illustrates this calculation for a 500 MW new combined cycle unit.

Table 6: Estimate Required Resources: New Capacity Operation Example (500 MW Combined Cycle)

Assumption

Calculation

Resource Estimate

15 $/kW-yr FOM

15 %/kW — yr * 500 MW * 1,000

= $7,500,000 Annual FOM Cost

2.3. Installation of New Pollution Control Retrofit Technology
Resources required for installation of pollution control retrofits consist of both the labor hours
required to install a retrofit (by specialized labor category) as well as other resources required to
install the retrofits (such as steel and any relevant reagent). In order to estimate these hours and
resources, the quantity of projected new retrofit capacity is combined with technology-specific

8

-------
assumptions discussed below. The methodology for performing these steps differs depending
on the retrofit technology.

For FGD, SCR, and ACI, assumptions regarding construction labor are based on a 2002 EPA study
that analyzed the resource requirements for installing various pollution control equipment (EPA,
2002). For DSI and FF, assumptions are based on a memo from Andover Technology Partners
(ATP), the author of which was also one of the contributors of the 2002 EPA study (Staudt,
2011a). To derive estimates for DSI, ATP scaled the labor requirements for AGs found in the EPA
(2002). Installation labor was further sub-divided into different labor categories, such as
boilermakers, engineers and "other installation labor," based on EPA (2002), and similarly scaled
for DSI and FF.

In addition to construction labor, we also estimated the additional labor necessary to produce
the steel required by these retrofits. The increased steel demand is estimated by multiplying the
per-MW steel demand from EPA (2002) by the estimated increases in retrofit pollution control
capacity. For DSI and FF, the same proportionality assumption was taken from Staudt (2011a) for
installation labor to estimate the steel needed for installation. In addition to steel, SCRs also
require an initial load of their catalyst at installation. Similar to the calculations for steel, this
initial load was calculated based on per-MW values. The per-unit labor needs, labor category
breakdown, and physical resource requirements for FGD, SCR, ACI, DSI, and FF technologies are
presented in the table below.

Table 7: New Retrofit Capacity Resource Assumptions

Resource Needs
(Units in parenthesis)

FGD

SCR*

ACI

DSI

Total Labor (labor
hours/MW)

760

700

398

Boilermaker (%)

40%

45%

50%

45%

Engineering (%)

20%

17%

Other Installation Labor
(%)

40%

48%

33%

48%

Steel (tons/MW)

2.25

2.5

0.35

2.20

1.42

Source: EPA, 2002 and Staudt, 2011a

*SCR also requires a catalyst initial load of 1.2 m3/MW, in addition to steel.

To illustrate how these assumptions are used to estimate the required resources to install a
retrofit, the following table has an example calculation for a 500-MW FGD unit.

-------
Table 8: Estimate Required Resources: New Retrofit Example (500 MW FGD)

Assumptions

Calculation

Resource
Estimate

760 labor hours per
MW

labor hours
760 * 500 MW

= 380,000
labor hours

40% Boilermakers

= 152,000 Hours

20 % Engineering

= 76,000 Hours

MW

40% Other
Installation Labor

=152,000 Hours

2.25 Tons Steel
per MW

Tons

2.25	* 500 MW

MW

= 1,125 Tons of
Steel

Resource requirements for HRI and NOx combustion controls were determined using a slightly
different methodology due to a lack of more detailed labor requirement data for those
technologies. For both HRI and NOx combustion controls, we estimate the annual cost of
installation and then estimate constituent capital and labor components.

For HRI, total labor and capital estimates were broken down into two specialized labor
categories as well as the equipment and material required for installation. These breakdowns
were applied to the estimated average cost of available HRI installations. 4 We assumed a 4-year
duration for HRI changes (Staudt, 2014) and thus estimated the annual cost as one-fourth of the
total cost for these improvements.

Table 9: Annual Labor and Capital Breakdown for Heat Rate Improvement

Category

HRI

Cost ($/kW)

$25/kW

Boilermaker/General Construction

40%

Management/Engineering

20%

Equipment

30%

Materials

10%

Source: Staudt (2014) Note that these are illustrative annual costs over a period of 4
years.

To illustrate how resource requirements are estimated, the table below demonstrates the
calculation of a HRI at a 500 MW unit.

4 The ability to improve heat rates and the cost of doing so is dependent on various unit-specific factors and varies
across the existing fleet. In a 2015 analysis, EPA assumed $100/kW, and presents that assumption here for
illustrative purposes.

10

-------
Table 10: Estimate Required Resources: New Retrofit Example (500 MW Heat Rate Improvement)

Assumption

Calculation

Resource
Estimate

25 $/kW
Annual Cost

25 — * 500 MW* 1,000

kW

40% Boilermaker/General
Construction

= $5,000,000

20% Management/Engineering

= $2,500,000

30% Equipment

= $3,750,000

10% Materials

= $1,250,000

For NOx combustion controls, estimates of required resources are based on the capital and labor
content of projected installation costs. Note that this differs from other controls, for which we
use capacity projections. The labor and capital components of total installation costs are based
on analysis of McAdams (2001), which estimates the capital cost and labor requirements to
upgrade to state-of-the-art NOx combustion controls. In that report, the costs of installing flue
gas recirculators and ultra-low NOx burners were estimated to be $400,000 for capital and
$180,000 for labor. These cost estimates were used to derive the capital and labor content at
approximately 70 percent and 30 percent, respectively.

Table 11: Capital and Labor Costs of Combustion Control Equipment



Capital

Installation Labor

Flue gas recirculation (FGR)5





Forced draft and FGR ductwork

$150,000



New forced draft/FGR fan

$50,000



Forced draft/FGR fan installation



$100,000

Next-generation low NOx burner6





Burners

$200,000



Installation



$80,00

Sum

$400,000

$180,000

Total Cost

$580,000

Percentage of Total

70%

30%

Capital and labor were then further split into boilermakers and general construction,
management and engineering, equipment, and material using the same ratios as for HRI (a 2:1
ratio of boilermakers and general construction to management and engineering). The table
below summarizes the breakdown of total NOx combustion control costs into specialized labor
and capital components.

5	Ibid., Table 2

6	Ibid, Table 7

11

-------
Table 12: Labor and Capital Breakdown for Combustion Control Equipment

Category

Percentage

Boilermakers/General Construction

20%

Management/Engineering

10%

Equipment

52%

Materials

18%

2.4. Operation of Newly Installed Retrofits
Resources required for operating pollution control retrofits included materials such as limestone,
catalyst, ammonia, activated carbon, and others. As with resources required for construction of
pollution controls, the resources required for operation of pollution controls is based largely on
EPA (2002) and Staudt (2011a). These resource estimates were multiplied by an estimated level
of generation for each type of pollution control in order to estimate the total (physical) quantity
of resources needed during operation.7

Table 13: Resource Need Assumptions for Retrofit Operation

Pollution Control Type

Resource
(Units in parenthesis)

Usage Estimates

FGD

Limestone (Tons/MWh)

0.036

SCR

Ammonia (lbs/lb NOx Reduced)

0.39

Operational Catalyst (m3/MWh)

0.00002

DSI

Sodium Bicarbonate/Trona (Tons/MWh)

0.03

ACI

Activated Carbon (tons/MWh)

0.000074 (with FF)
0.000914 (with ESP)

FF/ESP

Baghouse Material (Annual VOM Costs)

$95,000,000**

Sources: Usage: EPA (2002. DSI and FF: Staudt (2001a)

Ammonia: Development of Supply Curves for Abatement ofGHG from Coal-fired Utility Boilers, Air Pollution

Prevention and Control Division, US-EPA, RTF and NCState University, 2009

Catalyst: EPA Air Pollution Control Cost Manual, Sixth Edition, January 2002.

Sodium Bicarbonate: Communication with Andover Technology Partners, Feb 7 2011

**For Fabric filters, we used the projected VOM cost and converted this to labor hours using the direct

workers per million dollars output for the relevant manufacturing industry sector.

7 Total generation was estimated based on the total capacity of each type of retrofit and an assumed 85 percent
capacity factor. For ammonia, the usage was calculated based on the total predicted NOx reduced, consistent with
the EPA (2002) approach.

12

-------
In addition to the resources required for operating newly installed retrofits, under certain
situations, there may also be additional resources required for incremental changes in operating
existing retrofits. These changes are reflected in EPA's labor analyses as projected changes in
variable operating and maintenance costs.

The following table illustrates the calculation of resource requirements for a scrubber retrofit.

Table 14: Estimate Required Resources: New FGD Retrofit Operation Example (500 MW, $1 Mil FOM,

FGD)

Assumption

Calculation

Resource Estimate

0.036
Limestone
Tons/MWh

0.85
Capacity
Factor

500 MW * 0.85 * 8760 Hours *
0.036

134,028 Tons of Limestone

$1 Million FOM

2.5. Retirements

Projected retirement of existing capacity results in a decrease in required resources in operating
these units. To estimate resource requirements associated with these projections, the projected
change in generation capacity is multiplied by FOM costs. In order to convert retired capacity
into potential reductions in labor needed, we assumed that changes in operating costs for
electricity generation can be used as a proxy for the ceased economic output due to fossil
retirements. To convert projected retirement capacity into total changes in resource
requirements, we use the FOM ($/kW-yr) values in the table below.

Table 15: Resource Assumptions for Generator Retirements

Retirement Type

FOM Cost (2016$/kW-yr)

Coal (Average)

Oil and Natural Gas (Average)

Source: EPA 2018

These FOM costs in $/kW-yr are then multiplied by the incremental change in capacity in order
to estimate total incremental FOM costs related to projected retirements. An example of this
calculation is found in the table below.

Table 16: Estimate Required Resources: Retired Capacity Example (500 MW Combined Cycle)

Assumption

Calculation

Resource Estimate

23 $/kW-yr

23 %/kW-yr * 500 MW * 1,000

= $11,500,000 Annual FOM Cost

-------
2.6. Fuel Production and Pipeline Construction

Projected changes in fuel use and price are modeled directly, and do not require an intermediate
estimation of required resources. However, a significant projected increase in natural gas use
may require additional pipeline capacity. To estimate the resources required for pipeline
construction we multiply the projected change in natural gas use (TCF) by the capital cost of
pipeline capacity expansion, assuming that additional pipeline capacity is needed is linearly
proportional to the projected increase in natural gas consumption. Using data from EPA's power
sector modeling assumptions, we estimated the incremental capital cost for constructing new
pipeline capacity to be about $215 million per TCF of incremental gas capacity in any given year.
The resulting capital cost for pipeline construction is subsequently converted into labor
estimates using the labor productivity value for pipeline construction found in Table 18.

3. Estimate Labor Impacts

This section describes the other economic variables that are needed to estimate the final
employment impacts. These variables enable us to convert the estimated resource requirements
in the previous step to estimated impacts on labor. The economic variables used in the labor
analysis are job-years, product prices, and labor productivity.

3.1. Annual Job-Year

We assume 2,080 labor hours per year is the equivalent of one job. For the resource estimates
in the previous step that are estimated in labor hours, we divide total labor hours by 2,080 to
estimate total job-years.

3.2. Product Prices

Product prices are necessary to convert resource requirements estimated in physical units into
dollars. Where applicable, product prices used for each resource are summarized in the table
below. Product prices are multiplied by the resource estimates in the previous step to estimate
the dollar amount of each resource.

-------
Table 17: Product Price Assumptions

Resource

Price

Steel

$550 per ton

Limestone

$75 per ton

Ammonia (NH3)

$150 per ton

Catalyst

$8,475 per m3

Activated Carbon

$1,120 per ton

Trona

$120 per ton

Sources: Steel: Platts, Steel Markets Daily, Vol 3 Issue 209, October 29, 2009.

Limestone: FGD Tech Evaluation, March 2007.

Ammonia: Development of Supply Curves for Abatement ofGHG from Coal-fired
Utility Boilers, Air Pollution Prevention and Control Division, US-EPA, RTP and NC
State University, 2009.

Catalyst: EPA Air Pollution Control Cost Manual, Sixth Edition, January 2002.

Activated Carbon: Preliminary Cost Estimate of ACI for Controlling Hg Emission,

US DOE, NETL, November 2003.

Sodium Bicarbonate: Communication with Andover Technology Partners, Feb 7,

2011.

3.3. Labor productivity

In order to convert the estimated resource requirements into actual labor demand, we first
estimate labor productivity for each of the relevant resources or sectors. Estimating labor
productivity requires estimating both a base (historical) year labor productivity, as well as a
future year labor productivity. Since labor productivity generally increases over time (as the
number of workers needed to obtain a unit of output generally decreases over time), it is
important to adjust labor productively based on historical data to account for these changes
when evaluating impacts on future years.

3.3.1. Labor Productivity: Base year
This section includes the sources of base labor productivity estimates and the formulas used to
calculate them. It also describes how different sectors have different values and how NAICS-
based sector definitions are matched with the relevant sectors and their respective productivity
estimates.

To determine base year labor productivity, we first connect all resource estimates to their North
American Industry Classification System (NAICS) sectors. Each type of resource is matched to a
NAICS sector based on the types of goods produced by that sector. Once the NAICS sector is
identified, we collect value of shipments and total employees from the Economic Census and
Bureau of Labor Statistics (BLS) for each relevant sector. The NAICS sectors for each resource,
value of shipments and total employees data, the year that data was collected (data vintage),
and calculations used to determine labor productivity are presented in the table below.

-------
Table 18: Assumptions to Calculate Base Labor Productivity

Resource

NAICS
Sector

Value of
Shipments
($ Million)

Total
Employees

Labor Productivity
(Direct workers per
$ Millions of
Output)

Data Vintage





A

B

A/B



Steel

33121

$13,818

26,196

1.9

2012

Limestone

32741

$1,876

4,369

2.3

2007

Ammonia (NH3)

325188

$22,829

35,801

1.6

2007

Catalyst

331419,
331492

$13,911

19,666

1.4

2007

Activated Carbon

325998

$17,129

35,956

2.1

2007

Trona

212391

$1,708

3,711

2.2

2007

FF Resource

325211

$85,232

71,216

0.8

2007

Power Plant
Construction
(Including HRI
and combustion
controls)

237130

$44,270

222,684

5.0

2012

Equipment
Manufacturing

333

$407,669

1,063,392

2.6

2012

Engineering

54133

$211,936

994,363

4.7

2012

Power Plant
Operators

22111

$14,311

142,240

9.9

2012

Pipeline
Construction

237120

$41,153

172,311

4.2

2012

For estimated changes in fuel consumption we use labor productivity estimates that are
production-based. For coal, we use EIA data on regional coal mining productivity (in short tons
per employee hour).8 9 For natural gas, labor productivity per unit of natural gas was unavailable.
Since most secondary data sources (such as Census and EIA) provide estimates for the combined
oil and gas extraction sector, we use an adjusted labor productivity estimate for the combined oil

8	From US Energy Information Administration (EIA) Annual Energy Review, Coal Mining Productivity Data. Used 2008.

9	Unlike the labor productivity estimates for various equipment resources which were forecasted to future years
using BLS average growth rates, for fuel sectors we use the most recent historical productivity estimates (i.e.,
without forecasting to the future). In general, labor productivity for the fuel sectors (both coal and natural gas)
showed a significantly higher degree of variability in recent years than the manufacturing sectors, which would have
introduced a high degree of uncertainty in forecasting productivity growth rates for future years.

16

-------
and gas sector that accounts for the relative contributions of oil and natural gas in the total
sector output (in terms of the value of energy output in MMBtu).10

Table 19: Fuel Use Labor Productivity Estimates

Fuel Type

Labor

Productivity

Coal by Region (Short
Tons/labor hour)

Appalachia

2.32

Interior

4.73

West

17.09

Waste Coal

5.19

Natural Gas
(MMBtu/labor hour)

122

Notes: US national coal productivity is used for waste coal.

3.3.2. Labor Productivity: Estimating Future Labor Productivity
Since we are evaluating impacts of future policies, the base productivity estimates are adjusted
for future years in order to account for productivity growth. To do this, labor productivity growth
rates are estimated using historical trends in sectoral productivity growth rates (from BLS and
Census data). To calculate these growth rates, labor productivity estimates are calculated for
each sector from 1992 through the year of data that informs the productivity factors (see Data
Vintage column in Table 18). Calculated growth rates for each resource are presented below.

10 We converted 2012 EIA data for natural gas production (29.5442 TCF) and Crude Oil Production (2,377,806,000
barrels) into MMBtus of natural gas (30,194,172,400 MMBtu) and oil (13,793,274,800 MMBtu). The resulting sum of
MMBtu values were divided by the number of total employees in the oil and gas extraction sector (173,281 based
on 2012 economic census data) to calculate MMBtu per labor-year. This value was divided by 2,080 hours to
estimate MMBtu/labor-hour value of 122.

-------
Table 20: Estimated Annual Productivity Growth Rates by Resource / Sector

Resource / Sector

Growth Rate

Steel

1.0%

Limestone

2.0%

Ammonia (NH3)

3.3%

Catalyst

1.9%

Activated Carbon

3.1%

Trona

1.3%

FF Resource

4.2%

Power Plant Construction (Including
HRI and combustion controls)

0.0%

Equipment Manufacturing

2.7%

Engineering

0.8%

Power Plant Operators

1.7%

To estimate labor productivity for a given year the following formula is used:

LPv

id = i

7 (1 + LPt)yt~Yv

Where:

LPt = Labor Productivity in Year T

LPv = Labor Productivity in Year of Data Vintage

Yj = YearT

Yv = Year of Data Vintage
4. Aggregate Labor Impact Estimates

The final step in the employment analysis estimates the total impact on labor by combining
outcomes from the previous steps to produce an estimate of labor impact. This section discusses
how this is implemented for each type of projected compliance change.

4.1. New Generation Capacity
New generation capacity labor estimates are derived from demand for labor needed to produce
the materials and equipment used in construction as well as the labor to construct and operate
the new capacity. In Step 1 we quantified the amount of new capacity, then converted that
capacity amount into the total construction cost and the capital and labor components in Step 2.
In Step 3 we estimate the labor productivity values for appropriate sectors to convert Step 2
costs into labor needs. This final step combines these previous steps to produce final labor
needs. The schematic of the steps involved in these calculations are shown in the graphic below.

-------
Table 21: Estimating Final Labor Needs for New Generation Capacity

Step 1 (Quantify
Compliance Change!

Step 2 (Estimate Required
Resources)

Step 3 (Estimate Labor Need)*

Step 4

¦

MW of New
Generation Capacity

r

Construction
Costs

Equipment

LP Equipment Manufacturing

Labor for
Construction

Material

LP Steel

Construction
Labor

LP Power Plant Construction

Engineering/
Management

LP Engineering

Operating Costs (FOM)

LP Power Plant Operators

Labor for Operation

*LP refers to labor productivity

4.2. New Pollution Retrofits
For pollution control retrofits, we evaluate the labor necessary to both construct and operate
these controls. The quantity of new pollution control retrofits was estimated in Step 1. Under
Step 2 the installation resource requirements were calculated as a function of the required
amount of steel and the labor hours required for construction. In Step 3 we convert labor hour
to labor estimates, convert other resources estimated under Step 2 into dollars, and estimate
labor productivity values for industries associated with producing retrofit input material. This last
step combines these results to estimate the final labor necessary to produce the required
resources, install the retrofits, and operate the retrofits.

Table 22: Estimating Final Labor Needs for Retrofits

Step 1 (Quantify
Compliance
Change)

Step 2 (Estimate Required Resources)

Step 3 (Estimate Labor
Need)*

Step 4





Equipment

LP Equipment Manufacturing







Material***

LP Steel***

Labor for
Installation

MW of New
Retrofits

Installation
Costs**

Construction Labor

LP Power Plant Construction

Boilermakers

LP Boilermakers





Engineering/Management

LP Engineering









Resource

LP for



MWh of Retrofit

Resource Usage (physical quantity)

price ($/Phys.

Resource

Labor for

Operation





Qty.

Sector

Operation

FOM Costs

LP Power Plant Operators

* LP refers to labor productivity.

**For some technologies we calculate resource needs in terms of labor hours. For these we assume 2,080 hours
of work per year in Step 3 to derive final labor demand under Step 4.

***For SCRs, material also consists of initial catalyst load which uses the catalyst price to convert to expenditures
and then Labor Productivity for catalyst to convert expenditures into required labor

19

-------
4.3. Fuel Use Changes

Projected changes in fuel use were quantified under Step 1 and included an evaluation of coal
and natural gas. It was not necessary to complete step 3 since the projections were already a
measure of required resources (tons or MMBtu). In order to convert the physical quantities of
fuel into labor estimates in Step 3, we determine physical labor productivity estimates for coal
and natural gas. Under Step 4 we combine the previous steps to generate final labor estimates.

Table 23: Estimating Final Labor Needs for Fuel Use Changes

Step 1 (Quantify
Compliance Change)

Step 2 (Estimate
Required Resources)

Step 3 (Estimate Labor Need)

Step 4

Regional Coal

Coal (Tons)

Productivity (tons

2080 Hours

Labor for Coal

per labor hour)

Natural Gas (MMBtu)

—

Natural Gas
Productivity
(MMBtu per labor
hour)

2080 Hours

Labor for Natural
Gas

4.4. Pipeline Construction
In addition to labor to produce the fuel, we also estimate the labor required to build the pipeline
infrastructure necessary to deliver additional natural gas. In Step 1 we quantify the change in
natural gas consumption. In Step 2 we estimate the cost of constructing additional pipeline
capacity for the natural gas. These construction costs were then converted into labor impacts
using estimated labor productivity values for pipeline construction estimated in Step 3.

Table 24: Estimating Final Labor Needs for Pipeline Construction

Step 1 (Quantify
Compliance Change)

Step 2 (Estimate Required
Resources)

Step 3 (Estimate Labor Need)*

Step 4

Natural Gas (TCF)

Construction Costs ($MM)

LP Pipeline Construction

Labor for Pipeline
Construction

* LP refers to labor productivity.

4.5. Changes to Existing Capacity
The total capacity of retirements is estimated in Step 1 and that capacity is converted into
reduced expenditures using FOM in step 2. Labor productivity values for power plant operators
are estimated in step 3, and in step 4 we convert expenditures to labor using the labor
productivity factors.

-------
Table 25: Estimating Final Labor Needs for Generator Retirements

Step 1 (Quantify
Compliance Change)

Step 2 (Estimate
Required Resources)

Step 3 (Estimate Labor Need)*

Step 4

Coal (MW)

FOM Costs

LP Power Plant Operators

Labor for Coal
Retirements

Natural Gas and Oil
(MW)

FOM Costs

LP Power Plant Operators

Labor for Natural
Gas and Oil
Retirements

* LP refers to labor productivity.

21

-------
References

EPA. "Documentation for EPA's Power Sector Modeling Platform v6 Using the Integrated Planning
Model." 2018, https://www.epa.gov/airmarkets/clean-air-markets-power-sector-modeling

EPA. "ENGINEERING AND ECONOMIC FACTORS AFFECTING THE INSTALLATION OF CONTROL
TECHNOLOGIES FOR MULTIPOLLUTANT STRATEGIES (EPA-600/R-02/073)." October 2002.

McAdams, J.D., SD. Reed, D.C. Itse. "Minimize NOX Emissions Cost-Effectively." June 2001,
http://www.johnzink.com/wp-content/uploads/mimimize-nox-emissions.pdf

Staudt, Jim. "Memo to Bansari Saha Re: DSI and Fabric Filter System Installation Labor Estimate." January
2011.

Staudt, Jim. "Memo to Bansari Saha Re: Labor requirements for heat rate improvement, for retrofit of
CCS, and for construction of new units with/without CCS." August 2011.

Staudt, Jim. "Memo to Bansari Saha Re: Labor Requirements for Various Compliance Actions for the
Proposed EGU GHG NSPS Rule." September 2011.

Staudt, Jim. "Estimating Labor Effects of Heat Rate Improvements." March 2014.

22

-------