﻿
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF AIR AND RADIATION
NATIONAL VEHICLE AND FUEL EMISSIONS LABORATORY
2000 TRAVERWOOD DRIVE
ANN ARBOR, MI  48105-2498
   
   DATE, 2020
   MEMORANDUM
   SUBJECT:Final Assessment Analysis:  Amendments Related to Marine Diesel Engine Emission Standards
   FROM:  Jean Marie Revelt, EPS; Kenneth Davidson, PS; Margaret Zawacki, EE
   Office of Transportation and Air Quality
   Assessment and Standards Division
   
   TO:Amendments Related to Marine Diesel Engine Emission Standards  - 
   Docket EPA-HQ-OAR-2018-0638
   The Environmental Protection Agency (EPA) is amending the national marine diesel engine program with relief provisions to address concerns associated with finding and installing certified Tier 4 marine diesel engines in certain high-speed commercial vessels.  This relief is in the form of additional lead time for qualifying engines and vessels.
   EPA's 2008 Final Rule for Control of Emissions of Air Pollution from Locomotive Engines and Marine Compression-Ignition Engines Less than 30 Liters per Cylinder adopted Tier 4 emission standards for commercial marine diesel engines at or above 600 kilowatts (kW) (73 FR 37096, June 30, 2008).  These standards, which were expected to require the use of exhaust aftertreatment technology, phased in from 2014 to 2017, depending on engine power.   After the Tier 4 standards were fully in effect for all engine sizes, some boat builders informed EPA that there were no certified Tier 4 engines available with suitable performance characteristics for the vessels they needed to build, specifically for high-speed commercial vessels that rely on engines with rated power between 600 and 1,400 kW that have high power density.
   To address these concerns, EPA proposed, and through a final rule is adopting, provisions to provide additional lead time for implementing the Tier 4 standards for engines used in certain high-speed vessels (84 FR 46909, September 6, 2019). Through that rule EPA is also finalizing the proposed approaches for streamlining certification requirements to facilitate or accelerate certification of Tier 4 marine engines with high power density.  
   The new lead time provisions have two phases.  The first phase sets model year 2022 as the Tier 4 implementation deadline for propulsion engines with maximum power output up to 1,400 kW and power density of at least 27.0 kW per liter displacement, installed in high-speed vessels up to 65 feet in length (waterline) with total nameplate propulsion power at or below 2,800 kW.  These include pilot boats and some research boats.  The second phase sets model year 2024 as the Tier 4 implementation deadline for propulsion engines up to 1,000 kW and power density of at least 35.0 kW per liter displacement, installed in high speed vessels with fiberglass and other nonmetal hulls up to 50 feet in length (waterline), with a single propulsion engine.  These are expected to be primarily lobster or other fishing boats.  For both phases, engine manufacturers can certify, produce, and sell Tier 3 engines prior to the new Tier 4 implementation dates.  Finally, EPA is adopting waiver provisions that will be available beginning in 2024 for vessels meeting certain Phase 2 specifications; the waiver provision is intended to allow qualifying boat builders to continue building boats with Tier 3 engines if engine manufacturers have not yet certified suitable engines for those vessels.
   The economic impact, emission inventory, and human health and welfare assessments performed for the final rule that is the subject of this memorandum use the same methodologies as were used for the proposed rule.   The inventory and cost assessments rely on the data and methodologies developed to support our 2008 Final Rule.  The benefits assessment uses a simplified health benefits estimation method.  
   The additional lead time for implementing the Tier 4 standards for qualifying engines will delay the emission inventory and air quality benefits of the original standards adopted in 2008.  The estimated annual increase in NOx and PM10 emissions associated with the additional lead time is about 108 and 2.3 short tons, respectively, in 2020 and 2021, when both sets of engines are affected, decreasing to about 37 and 1 ton, respectively, in 2022 and 2023, when only those engines up to 1,000 kW are affected.  The lifetime inventory increase is estimated to be about 3,760 tons of NOx and 80 tons of PM10, assuming a 13-year engine lifetime.  This represents less than one-tenth of one percent of the national annual emissions for these pollutants from commercial Category 1 marine diesel emissions (i.e., engines below 7.0 liters per cylinder displacement).  
   While it is challenging to estimate the monetized value of these forgone emission reductions without performing detailed air quality and benefits modeling, this Memorandum examines potential forgone human health and welfare benefits associated with this Tier 4 relief rule using a simplified approach.  Using reduced form health benefit per ton values, we estimate that the annual PM2.5-related forgone benefits do not exceed a high-end estimate of $3.0 million in any given year (2015$, 3% discount rate, mortality effect estimate derived from Lepeule et al., 2012). The total present value of the stream of forgone benefits ranges from $9.8 million to $31 million.
    Consistent with the economic impact analysis prepared for EPA's 2008 rulemaking, the costs for this final rule were estimated using both a behavioral approach (in the intermediate-run after the adoption of new standards, producers pass only some compliance through to consumers), and a full-cost pass-through approach (in the long-run after the adoption of new standards, producers pass all compliance costs through to consumers).  This rule imposes no additional economic costs above those included in our 2008 rulemaking.  Instead, the additional lead time is expected to result in cost savings.  We estimate cost reductions of about $3.9 million, using a behavioral modeling approach, or $4.2 million, using a full-cost pass-through approach (2015$).  These are the estimated cost reductions from installing less expensive Tier 3 engines in new vessels during the relief period (2020 through 2023) and the associated operating cost reductions during the 13-year lifetime of those engines (2020 through 2035).
   The estimated cost and emission inventory impacts do not include the use of waivers; if engine manufacturers apply for and receive waivers post-2023, the estimated cost reductions and inventory impacts would increase and would extend for a longer period of time (the useful life of the engines produced subject to the waiver).  
   Finally, the proposed rule sought comment on a regulatory alternative which would adjust the Tier 4 compliance deadlines further for the second phase of proposed relief instead of providing a waiver provision, setting the new start date for Tier 4 at model year 2028 for engines 600 kW to 1,000 kW.  As explained in the preamble for the final rule, the Agency is not adopting this alternative.  Adoption of this regulatory alternative would have increased the estimated total forgone inventory benefits of the proposal by about 1,760 additional short tons of NOx and 37 additional short tons of PM10 above the estimated inventory increases associated with the final program as adopted by the Agency.  Using reduced form health benefit per ton values, we estimate that the annual PM2.5-related forgone benefits for this regulatory alternative could be up to a high-end estimate of $4.4 million in any given year (2015$, 3% discount rate, mortality effect estimate derived from Lepeule et al., 2012). The total present value of the stream of forgone benefits ranges from $13.5 million to $44.6 million.  We estimate cost reductions associated with the regulatory alternative to be about $7.2 million, using a behavioral modeling approach, or $7.8, using a full-cost pass-through approach (2015$).  These estimated costs are about $3.3 million and $3.6 million, respectively, more than the estimated cost savings associated with the final adopted program. 
   This remainder of this Memorandum describes the methodology used to develop these impact estimates and provides detailed results of the analyses.  Section A presents the inventory analysis, Section B describes the air quality impacts, Section C provides estimated monetized benefits, and Section D presents the estimated cost reductions.  Section E provides an analysis for the regulatory alternative.  Section F contains the E.O. 13771, Reducing Regulation and Controlling Regulatory Costs analysis prepared for this rule.
A. Inventory Impacts
 Methodology
   The inventory impacts of the additional lead time provisions are estimated using the same methodology employed in EPA's 2008 Final Rule for Control of Emissions of Air Pollution from Locomotive Engines and Marine Compression-Ignition Engines Less Than 30 Liters per Cylinder (73 FR 37096, June 30, 2008, the 2008 Rule).  A comprehensive explanation of the estimation methodology is contained in Chapter 3 of the Regulatory Impact Analysis (RIA) prepared for that rule.
   The emission inventories for NOx, and PM10 were estimated using the equation:
Equation 1     
where each term is defined as follows:
      I = the emission inventory (gram/year)
      N = engine population (units)
      P = average rated power (kW)
      L = load factor (average fraction of rated power used during operation; unitless)
      A = engine activity (operating hours/year)
      EF = emission factor (gram/kW-hr)
      
   This equation yields the number of grams of pollutants emitted per year; these results are then converted to metric tonnes (g * 1,000,000) and, in a final step, converted to short tons/year (1 tonne = 1.1023 tons) for ease of comparison with the 2008 Rule inventories.
Data Inputs
   The data inputs for the analysis are described in this section.  Note that some of the modeling inputs are different from those used to develop the emission inventories for the 2008 Rule analysis (see Table 3-3 of the 2008 RIA).  These changes generally reduce the inventory impact estimates compared to the 2008 Rule inventories.  Appendix 1 to this Memorandum provides estimated emission impacts using the 2008 Rule data inputs.
   Years of the Analysis.  This inventory analysis is performed for the years of relief, which are 2020 through 2023 for engines in the 600 to 1,000 kW range (4 years), and 2020 through 2021 for engines in the 1,000 to 1,400 kW range (2 years).  While the Tier 4 standards began to apply in 2017 for the affected engines, this analysis does not include impacts for 2017 through 2019. Some boat builders were able to continue building boats with Tier 3 engines during that period because the marine diesel engine program allows equipment manufacturers to use up inventories of new Tier 3 engines produced prior to the effective dates of the Tier 4 standards as long as those inventories were not stock piled.  Other boat builders adjusted their designs to use smaller engines that are not subject to the Tier 4 standards.  Because both strategies are allowed under the current regulatory program, it is not necessary to estimate their inventory impacts.  Engine manufacturers are expected to take advantage of the extended implementation dates once the regulatory changes go into effect to produce Tier 3 engines, and boat builders are expected to install those Tier 3 engines beginning in 2020.  Therefore, that year is used as the start year for the analysis.
   N = engine population (units).  This is the estimated number of high power density marine diesel engines that are expected to be affected by the additional lead time.  We estimate as many as 25 engines per year will qualify for relief annually, 21 in the range of 600 kW to 1,000 kW (these engines are like those used on high-speed lobster fishing boats), and 4 in the range of 1,000 to 1,400 kW (these engines are like those used on high-speed pilot boats).  
   
   For engines in the 600 kW to 1,000 kW range, the cumulative number of engines affected over the period of relief is estimated to be 84 engines.  The number of engines affected annually is a large share of the population of high power density engines for this power range used in the 2008 Rule inventory analysis for these years (about 20.5% of the population of 102 engines in 2002, and about 17%% of the population of 124 engines in 2024).  The assumption of 21 engines per year eligible for relief is reasonable given that the market demand for high power density engines of this size has changed since we finalized the 2008 Rule.  In 2008, most or all lobster boats used high power density engines below 600 kW.  Targeted lobster beds were typically located relatively close to shore and fishing could occur no farther than 20 miles from shore under federal licensing requirements; therefore, engines of this size were suitable for these boats.  In recent years, changes in the location of lobster beds, which now may be as far as 40 miles from shore, led to different operating needs and the use of larger, high power density engines for some new vessels.   
   
   For engines in the 1,000 kW to 1,400 kW range, the cumulative number of engines affected over the period of relief is estimated to be 8 engines.  The number of engines affected annually is consistent with the population of high power density engines for this power range used in the 2008 Rule inventory analysis for these years (less than 2% of the population of 214 engines in 2002 and 261 engines in 2024).  
   
   P = average rated power (kW).  This is the average rated power for the relevant power band for high power density engines.  This analysis uses 750 kW for engines in the 600 kW to 1,000 kW range, based on information received from industry stakeholders, and 1,200 kW for engines in the 1,000 kW to 1,400 kW range, based on the midpoint of that range.  The 750 kW value is higher than what was used in the 2008 Rule analysis, which was 678 kW for the 600 kW to 1,000 kW range.  In other words, in this analysis, qualifying lobster built during the relief period are expected to have higher power than for similar engines modeled in the 2008 Rule.  This this change is reasonable because it reflects the changes in boat power for the lobster boat market.  The use of a different, slightly higher rated power for the covered engines results in more emissions compared to the analysis performed for the 2008 Rule.  However, due to the small number of engines affected, the total national inventory would not be significantly affected.  The 1,200 kW mid-point value for the 1,000 kW to 1,400 kW range is similar to the 1,175 kW value used in the 2008 Rule analysis.  
   
   L = load factor (average fraction of rated power used during operation; unitless).  This refers to the average power level produced by the engine over its entire range of operation.  This analysis uses a 0.37 load factor for engines in the 600 kW to 1,000 kW, instead of 0.79 that was used in the 2008 Rule analysis, based on information received from industry stakeholders.  These people informed us that lobster boats are at idle for a large portion of their operating hours, as boat operators tend their lobster pots.  For engines in the 1,000 kW to 1,400 kW range, the analysis uses the same load, 0.79, as was used in the 2008 Rule, which was based on information from engine manufacturers and user stakeholders obtained at the time of that rule.  It should be noted that the use of a different, lower load factor for engines 600 to 1,000 kW results in less emissions compared to the analysis performed for the 2008 Rule.  However, due to the small number of engines affected, the total national inventory would not be significantly affected.    

   A = engine activity (operating hours/year).  This refers to how many hours an engine is operated.  This analysis uses 1,500 hours of operation per year for engines in the 600 kW to 1,000 kW range, instead of 4,503 hours that was used in the 2008 Rule analysis, based on information received from industry stakeholders.  This reflects the usage pattern of lobster boats in Maine: while they may spend 10 hours or more at sea per trip, the number of days they operate is limited by weather and other factors.  For engines in the 1,000 kW to 1,400 kW range, the analysis uses the same activity rate, 4,503 hours per year, as was used in the analysis performed for the 2008 Rule.  This is about 12.3 hours per day, every day of the year, which is reasonable for commercial vessels including pilot boats.  Again, the use of fewer annual operating hours for engines 600 to 1,000 kW results in less emissions compared to the analysis performed for the 2008 Rule.  However, due to the small number of engines affected, the total national inventory would not be significantly affected.         
      
   EF = emission factor (gram/kW-hr).  This refers to the rate of emissions that are generated from an engine.  The purpose of this analysis is to estimate the forgone emission reductions due to the additional lead time, which is the difference between the Tier 4 standards (the requirement) and the Tier 3 standards (engines must be certified to this level to qualify for relief).  Therefore, the EFs used in the analysis are the difference between the two standards, for each pollutant, NOx and PM10.  Because the rule is focused on relief for high power density engines, the Tier 3 standards for high power density engines (above 35 kW/l) from 40 CFR 1042.101, Table 1 are used; there are no separate standards for high power density engines for Tier 4.  It should be noted that the Tier 3 standards at 40 CFR 1042.101 are for NOx+HC; to obtain the NOx only standard, the Tier 3 NOx+HC standard is reduced by 0.2 g/kW-hr, which is slightly more than the 0.19 g/kW-hr HC standard for Tier 4.  
   
   Tables 1 and 2 summarize the data inputs used in this analysis.  

Table 1 - Inventory Analysis Data Inputs

Population (N)
Average Rated Power (P)
Load Factor (L)
Engine Activity (A)
NOx Emission Factor (EF) = ∆ NOx std (g/kW-hr)
PM10 Emission Factor (EF) = ∆ PM std (g/kW-hr)
600 <kW<=1000
21
750*
0.37*
1,500*
3.8
0.08
1000<kW<=1400
4
1,200*
0.79
4,503
3.8
0.08
* Values are different from those used in the 2008 Rule inventory analysis; see text for explanation and Appendix 1.

Table 2 - Emission Standards (g/kW-hr)

Tier 3 (HPD engines)
Tier 4 (all power densities)

NOx*
PM10
NOx
PM10
600 <kW<=1000
5.6
0.12
1.8
0.04
1000<kW<=1400
5.6
0.12
1.8
0.04
* Tier 3 standard is 5.8 g/kW-hr NOx+HC; reduced by 0.20 g/kW-hr HC for 5.6 g/kW-hr NOx limit

Inventory Impact Results
   Using the above data and methodology, the annual forgone emission inventory benefits are summarized in Table 3 and are about 108.1 tons of NOx and 2.3 tons of PM10 for those years in which engines are eligible for the delayed effective dates.  These estimates do not include the use of waivers.  If engine manufacturers apply for and receive waivers post-2023, the estimated emission inventory impacts would increase and would extend for a longer period of time (the useful life of the engines produced subject to the waiver).
Table 3 - Estimated Annual Forgone Emission Reductions (short tons)

Number of Engines/Year of relief
NOx /year
PM10/year
600 <kW<=1000
21
36.6
0.8
1000<kW<=1400
4
71.5
1.5
Total
25
108.1
2.3
   Table 4 shows the forgone emissions for each year's cohort of affected engines.  In each of 2020, and 2021, up to 25 engines can benefit from the revised implementation dates; the forgone emissions from these engines are estimated to be 108.1 tons of NOx and 2.3 tons of PM10.  In 2022 and 2023, 21 engines can benefit (only those up to 1,000 kW), and the forgone emissions from these engines is estimated to be about 36.6 tons of NOx and 0.8 tons of PM10.  
Table 4 - Estimated Forgone Emission Reductions by Cohort, 2020-2023 (short tons)
Year
NOx
PM10
2020
108.1
2.3
2021
108.1
2.3
2022
36.6
0.8
2023
36.6
0.8
   The affected engines will continue to have higher emissions than would otherwise have been the case for their entire service lives (estimated to be 13 years, as set out in the 2008 Rule analysis).  As shown in Table 5, the forgone lifetime emission inventory benefits are estimated to be about 3,764 tons of NOx emissions and 79 tons of PM10 emissions.


Table 5  - Estimated Forgone Emission Reductions, Engine Lifetime, 2020-2035 (short tons)


Year
600-1,000 kW
1,000-1,400 kW
Total

New Engines 
Population
NOx
PM10
New Engines 
Population
NOx
PM10
New Engines 
Population
NOx
PM10
2020
 21.0 
 21.0 
 36.6 
 0.8 
 4.0 
 4.0 
 71.5 
 1.5 
 25.0 
 25.0 
 108.1 
 2.3 
2021
 21.0 
 42.0 
 73.2 
 1.5 
 4.0 
 8.0 
 143.0 
 3.0 
 25.0 
 50.0 
 216.3 
 4.6 
2022
 21.0 
 63.0 
 109.8 
 2.3 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 71.0 
 252.9 
 5.3 
2023
 21.0 
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 92.0 
 289.5 
 6.1 
2024
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2025
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2026
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2027
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2028
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2029
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2030
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2031
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2032
 -   
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 -  
 92.0 
 289.5 
 6.1 
2033
 -   
 63.0 
 109.8 
 2.3 
 -   
 4.0 
 71.5 
 1.5 
 -  
 67.0 
 181.4 
 3.8 
2034
 -   
 42.0 
 73.2 
 1.5 
 -   
 -  
 -  
 -  
 -  
 42.0 
 73.2 
 1.5 
2035
 -   
 21.0 
 36.6 
 0.8 
 -   
 -  
 -  
 -  
 -  
 21.0 
 36.6 
 0.8 
Total
 84 
 1,092 
 1,904 
 40 
 8 
 104 
 1,860 
 39 
 92 
 1,196 
 3,764 
 79 


Discussion
   Inventory Impacts.  To put the estimated annual forgone emission inventory impacts in context, this section compares them to national inventories developed for the 2008 Rule.    
   Table 6 sets out the annual NOx and PM10 inventories estimated for the 2008 Rule, for all marine diesel engines below 30 l/cyl displacement, for each of the years of relief.,  The forgone emission reductions are estimated to result in a small incremental increase in annual emissions for the years of the program.  On a national basis, the increase is estimated to be less than one-tenth of one percent relative to all marine diesel engine emissions below 30 l/cyl displacement (Table 6), and to Category 1 commercial propulsion marine diesel engine emissions (Table 7).  It is about one half of one percent or less of the much smaller category of high power density commercial propulsion marine diesel engines (Table 8).  
Table 6  -  Estimated Annual Emission Increase Compared to All Marine Diesel Engines below 30 l/cyl displacement (short tons)

NOx
% Increase NOx, T4 Relief 
PM10
% Increase PM10, T4 Relief 
2020
576,707
0.02%
18,968
0.01%
2021
549,507
0.02%
18,211
0.01%
2022
523,322
0.01%
17,475
0.00%
2023
498,574
0.01%
16,732
0.00%
Table 7 - Estimated Annual Emission Increase Compared to Category 1 Commercial Propulsion Marine Diesel Engines (short tons)

NOx
% Increase NOx, T4 Relief 
PM10
% Increase PM10, T4 Relief 
2020
         185,242 
0.06%
         4,938 
0.05%
2021
         174,843 
0.06%
         4,562 
0.05%
2022
         164,971 
0.02%
         4,208 
0.02%
2023
         155,589 
0.02%
         3,873 
0.02%
Table 8  -  Estimated Annual Emission Increase Compared to Category 1 High Power Density Commercial Propulsion Marine Diesel Engines (short tons)

NOx
% Increase NOx, T4 Relief 
PM10
% Increase PM10, T4 Relief 
2020
           22,093 
0.49%
630
0.36%
2021
           20,480 
0.53%
575
0.40%
2022
           18,950 
0.19%
525
0.15%
2023
           17,496 
0.21%
477
0.16%
   Finally, the lifetime inventory impacts of the additional lead time are a small portion of total national Category 1 propulsion diesel marine engine NOx and PM10 inventories over the period 2020 through 2035, as illustrated in Table 9:  about two tenths of one percent of national NOx and PM10 inventories for all C1 commercial propulsion marine diesel engines, and less than 2 percent of high power density C1 commercial marine diesel engines.
Table 9 - Estimated Lifetime Emissions Compared to Category 1 Commercial Propulsion Marine Inventories, 2020-2035 (short tons)

NOx Inventory
Affected Engines %
PM10 Inventory
Affected Engines %
C1 Commercial Propulsion Engines
 2,108,335 
0.2%
 49,256 
0.2%
High Power Density C1 Commercial Engines 
 220,582 
1.7%
 5,920 
1.3%
Engines affected by Tier 4 relief rule
 3,764 

 79 

   Although the estimated inventory impacts are small when compared to national inventories, it is also important to look at the impacts in areas in which they are likely to occur.  To examine the inventory impacts on a more localized level, all of the annual forgone emissions for the engines from 600 kW to 1,000 kW, 36.6 tons per year of NOx and 0.8 tons per year of PM10, can be attributed to Maine (lobster fishing boats).  Similarly, all of the annual forgone emissions for the engines from 1,000 kW to 1,400 kW, 71.5 tons per year of NOx and 1.5 tons per year of PM10, can be attributed to Georgia (pilot boats operating out of the Port of Savannah).  
   EPA's Trends data reports off-highway mobile source NOx and PM10 emissions of 56,872 tons and 4,472 tons, respectively, for Georgia for 2017.  The forgone emissions from the Tier 4 relief amount to about 0.13 percent and 0.03 percent of those State-wide NOx and PM10 emissions.    
   Similarly, EPA's Trends data for off-highway mobile source NOx and PM10 emissions in Maine are 10,257 tons and 1,102 tons, respectively, for 2017.  In this case the forgone emissions from the Tier 4 relief amount to about 0.4 percent and 0.1 percent of those State-wide NOx and PM10 emissions.  
   
B.  Air Quality Impacts
   Full-scale photochemical air quality modeling is necessary to accurately project ambient concentrations of pollutants. The atmospheric chemistry related to ambient concentrations of PM2.5, ozone and air toxics is very complex, and making predictions based solely on emissions inventory changes is extremely difficult. We did not conduct photochemical air quality modeling for the small, localized increase in emissions from the boats subject to the relief provisions but we can look at previous air quality modeling analyses that have assessed the impact from marine emissions to onshore ambient PM2.5 and ozone concentrations.,,  These analyses demonstrate that emissions from different classes of marine vessel engines, both in port and underway, impact ambient concentrations of PM2.5 and ozone along the East coast and further inland.  These analyses modeled broad classes of marine vessel engines and were not specific to the lobster boats and pilot boats subject to the relief provisions.  Because the relief provisions result in small, localized increases in emissions, we expect that there will be only small increases in ambient concentrations of pollutants onshore. 
C.  Monetized Forgone Health Impacts
      As described in Section A, the relief provisions will result in forgone reductions of NOX (an ozone and PM precursor) and directly emitted PM2.5 emissions. Because human exposure to particulate matter is associated with increased incidence of premature mortality and morbidity outcomes, forgone emissions reductions will also lead to forgone improvements in human health. This section describes how forgone PM-related health benefits are estimated for this analysis.
1.Methodology and Data Inputs
      As noted above, full-scale photochemical air quality modeling is necessary to estimate projected ambient concentrations of pollutants on a spatial scale that could be used in a full-scale benefits analysis. However, since the scale of the impacts are projected to be small, this analysis instead uses a reduced form approach to estimating forgone health benefits. Specifically, the monetized value of forgone ambient PM2.5-related health benefits are estimated by applying mobile sector-specific benefit per ton (BPT) values for NOX emissions and directly-emitted PM2.5.
      
  The PM2.5-related BPT values provide the total monetized human health benefits (the sum of the economic value of the reduced risk of premature death and illness) that are expected from reducing one ton of directly-emitted PM2.5 or PM2.5 precursor such as NOX emissions. Recently, EPA updated its mobile source sector PM2.5-related BPT estimates using a source apportionment approach the Agency has used in the past.,  For this analysis, we utilized BPT values that were derived for "Category 1 and Category 2" (C1C2) marine vessels located in the Eastern U.S. We multiplied these Eastern C1C2 BPT values by the forgone emission reductions of NOX and directly-emitted PM2.5 (in tons) to estimate the forgone PM2.5-related health benefits associated with the amended program's relief provisions. The C1C2 BPT values were derived using detailed mobile sector source-apportionment air quality modeling, and apply EPA's existing method for using reduced-form tools to estimate PM2.5-related benefits., Compared to values EPA has used in the past, these BPT values provide better resolution by mobile sector (16 specific mobile sectors) and geographic area (national as well as Western and Eastern BPT estimates). The underlining source apportionment modeling is also based on a more recent emissions modeling platform (emissions data is from 2011 instead of 2005). We note, however, that an important limitation when using the BPT approach is that it is reliant upon air quality modeling input assumptions that do not reflect future conditions. Another limitation is that PM-related BPT estimates are unable to monetize the health impacts associated with exposure to ozone or other pollutants such as mobile source air toxics. Additional limitations are described in Section 3.
      Table 10 presents these unit values for C1/C2 marine vessel engines for each component of PM2.5, based on the underlying epidemiological study from which the PM mortality effect estimate is derived.
      
Table 10. Monetized value of mortality and morbidity benefits per ton of directly emitted PM2.5 and PM2.5 precursors by C1/C2 marine engines in 2025 (2015$); 
Results presented as average unit values for the Eastern U.S.a

Discount Rateb
NOx (per ton)
Direct PM2.5 (per ton)
Krewski et al., 2009
3%
$1,500
$140,000

7%
$1,400
$130,000
Lepeule et al., 2012
3%
$3,500
$320,000

7%
$3,200
$290,000
a "Eastern US" includes Texas and states to the north and east.
b The monetized value of PM2.5-related mortality accounts for a twenty-year segmented cessation lag. To discount the value of premature mortality incidence that occurs at future points along the distributed lag, we apply both a 3 and 7 percent discount rate consistent with current regulatory practices and guidance.
      To estimate the forgone health benefits in each year for which there are forgone emission reductions (2020-2035; see Table 5), this analysis applies the 2025 benefit per ton values in Table 10 to emissions in each calendar year. This assumption yields a slight overestimate of forgone benefits in years before 2025 and a slight underestimate of forgone benefits in years after 2025. For this analysis, it is assumed that the bias is offsetting and acceptable for characterizing the magnitude of potential impacts.
2.Results
      Tables 11 and 12 present the stream of forgone PM2.5-related health benefits related to the Tier 4 relief provisions.

Table 11. Forgone monetized PM2.5-related health benefits (2015$; 3% Discount Rate)a 

Krewski et al., 2009b
Lepeule et al., 2012c
Year
NOX
PM2.5d
Total
NOX
PM2.5
Total
2020
$160,000
$310,000
$470,000
$380,000
$710,000
$1,100,000
2021
$320,000
$620,000
$900,000
$760,000
$1,400,000
$2,200,000
2022
$380,000
$720,000
$1,100,000
$890,000
$1,600,000
$2,500,000
2023
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2024
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2025
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2026
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2027
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2028
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2029
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2030
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2031
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2032
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2033
$270,000
$520,000
$800,000
$630,000
$1,200,000
$1,800,000
2034
$110,000
$200,000
$310,000
$260,000
$470,000
$720,000
2035
$55,000
$110,000
$160,000
$130,000
$250,000
$380,000
Present Value
$13,000,000
Present Value
$31,000,000
a Totals may not sum due to rounding. A 3 percent discount rate is used to both account for a 20-year segmented cessation lag in the valuation of mortality and in the calculation of the present value of benefits.
b Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., ... & Thun, M. J. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality (No. 140). Boston, MA: Health Effects Institute.
c Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environmental health perspectives, 120(7), 965.
d PM2.5 emissions are assumed to be 97% of PM10 emissions.
Table 12. Forgone monetized PM2.5-related health benefits (2015$; 7% Discount Rate)a

Krewski et al., 2009b
Lepeule et al., 2012c
Year
NOX
PM2.5d
Total
NOX
PM2.5
Total
2020
$150,000
$290,000
$440,000
$350,000
$650,000
$990,000
2021
$300,000
$580,000
$900,000
$690,000
$1,300,000
$2,000,000
2022
$350,000
$670,000
$1,000,000
$810,000
$1,500,000
$2,300,000
2023
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2024
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2025
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2026
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2027
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2028
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2029
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2030
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2031
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2032
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2033
$250,000
$480,000
$700,000
$580,000
$1,100,000
$1,600,000
2034
$100,000
$190,000
$290,000
$230,000
$420,000
$660,000
2035
$51,000
$100,000
$150,000
$120,000
$230,000
$340,000
Present Value
$9,800,000
Present Value
$22,000,000
a Totals may not sum due to rounding. A 7 percent discount rate is used to both account for a 20-year segmented cessation lag in the valuation of mortality and in the calculation of the present value of benefits.
b Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., ... & Thun, M. J. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality (No. 140). Boston, MA: Health Effects Institute.
c Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environmental health perspectives, 120(7), 965.
d PM2.5 emissions are assumed to be 97% of PM10 emissions.
3.Discussion
      As shown in Tables 11 and 12, annual forgone benefits do not exceed a high-end estimate of $3.0 million in any given year (2015$, 3% discount rate, mortality effect estimate derived from Lepeule et al., 2012). The total present value of the stream of forgone benefits range from $9.8 million to $31 million.
      Reduced form tools, by their nature, are subject to uncertainty. BPT values reflect the geographic distribution of the underlying modeled emissions used in their calculation, which may not exactly match the geographic distribution of the emission reductions that would occur due to a specific regulatory action. Similarly, BPT estimates may not reflect local variability in population density, meteorology, exposure, baseline health incidence rates, or other local factors for any specific location. The photochemically-modeled emissions of the C1C2 sector-attributable PM2.5 concentrations used to derive the BPT values may not match the small change in localized air quality that might result from the relief provisions described above. For this reason, the forgone health benefits reported here may be larger or smaller than those that would be realized through this action. For example, the forgone benefits may be overstated in a location like Maine, since there will be some transport of emissions offshore or to areas external to the United States with different population and geographic characteristics.  .
      Given the uncertainty associated with analyses that rely upon reduced form tools, EPA systematically compared benefits estimated using its benefit per ton approach (and other reduced form approaches) to benefits derived from full-form photochemical model representation. This work is referred to as the "Reduced Form Tool Evaluation Project" (the Project), which began in 2017, and the initial results were available at the end of 2018. The Agency's goal was to better understand the suitability of alternative reduced form air quality modeling techniques for estimating the health impacts of criteria pollutant emissions changes in EPA's benefit-cost analysis. The Project analyzed air quality policies that varied in the magnitude and composition of their emissions changes and in the emissions source affected (e.g., on-road mobile, industrial point, or electricity generating units). The policies also differed in terms of the spatial distribution of emissions and concentration changes, and in their impacts on directly emitted PM2.5 and secondary PM2.5 precursor emissions (NOX and SO2). 
      For scenarios where the spatial distribution of emissions was similar to the inventories used to derive the benefits per ton, the Project found that total PM2.5 benefits were within approximately 10% to 30% of the health benefits calculated from full-form air quality modeling, though the discrepancies varied by regulated scenario and PM2.5 species. The scenario-specific emission inputs developed for the Project, and a final project report, are available online.  We note that the benefit per ton values used to monetize the forgone benefits of the relief provisions for Category 1 and Category 2 marine vessels were not part of the Project. However, analysis like that done for the Project is one way to better understand uncertainty associated with using reduced form tools for regulatory analysis. EPA strives to characterize the uncertainty from using reduced form modeling tools when possible.
D.  Economic Impacts
   The economic impacts of the additional lead time for the Tier 4 relief standards are the savings that will accrue to boat builders and boat purchasers from the installation and use of less expensive Tier 3 engines in boats built during the period of relief.  The cost impacts presented are estimated using a methodology consistent with the 2008 Rule.  Two sets of costs are presented:  one based on the results of the behavioral economic impact modeling performed for the 2008 Rule and one based solely on the engineering cost analysis prepared for the 2008 Rule.
   Economic Impact Modeling
   Economic impact modeling takes into account the dynamic reaction of the markets for marine diesel engines and vessels in response to a new regulatory requirement over different time horizons for market adjustments.  
   As explained in the 2008 Rule, competitive markets like those for marine diesel engines and vessels are not able to immediately respond to increased costs associated with a regulation.  In the very short run, labor or capital inputs cannot be adjusted and, as a result, all compliance costs are incurred by the manufacturer.  This is referred to as the "full-cost absorption" scenario.  Although there is no hard and fast rule for determining what length of time constitutes the very short run, it is inappropriate to use this time horizon for regulatory analysis because it assumes economic entities have no flexibility to adjust factors of production.  
   At the other extreme, in the long run, all factors of production are variable as the manufacturer adjusts production plans in response to cost changes imposed by a regulation, and all compliance costs are ultimately passed to the consumer.  This is referred to as the "full-cost pass-through" scenario.  The engineering costs analysis presented below reflects this approach.
   In the intermediate run, both the producer and consumer bear a portion of the compliance costs of a rule.  For the 2008 Rule, an economic impact model was developed to reflect the behavioral responses of these actors.  This behavioral model uses a combination of economic theory and econometric modeling to evaluate potential market changes associated with a new regulatory program, more specifically the impacts on market prices and quantities.  It also examines how those costs will be borne by producers and purchasers.  A key factor in this type of analysis is the responsiveness of the supply and demand sides of the market to the addition of compliance costs.  This relationship is captured the price elasticities of demand and supply.  As explained in Chapter 7 of the 2008 RIA, the model's methodology is rooted in applied microeconomic theory and was developed following the OAQPS Economic Analysis Resource Document.  
   Compliance costs associated with a regulatory program are categorized as fixed or variable costs.  Consistent with the 2008 Rule, the economic impacts of the additional lead time are based only on variable compliance costs.  Variable costs are the costs of new hardware required for an engine to meet the new Tier 4 emission standards or for a boat to incorporate Tier 4 engines.  Along with assembly and associated markups, these costs occur on a per unit basis as the engines and boats are produced.  Over time, variable costs per unit decrease because, as manufacturers gain experience in production, they can apply innovations to simplify machining and assembly operations, use lower cost materials, and reduce the number or complexity of component parts, all of which allows them to reduce the per-unit cost of production.  This learning curve is a well-documented phenomenon and is included in many mobile source cost estimates.  The progress ratio applied in the cost analysis for the 2008 Rule to reflect learning effects was 80 percent.  
   Fixed costs are treated differently when considering the economic impact of a rule.  Fixed costs are for research and development, tooling, certification, and equipment/vessel redesign, and occur prior to making a compliant engine available for sale.  Fixed costs for engine research are estimated to be incurred over the five-year period preceding introduction of the engine.  Fixed costs for engine tooling and certification are estimated to be incurred one year ahead of initial production.  Fixed costs for equipment redesign are also estimated to be incurred one year ahead of production.  Thus, all fixed costs are spread over the 5 years prior to initial sale.  
   As explained in the Economic Impact Analysis prepared for the 2008 Rule (see Chapter 7 of the 2008 RIA), an industry's supply curve in a competitive market is based on the market's marginal cost curve (costs for producing one extra unit; these are variable costs); fixed costs do not influence production decisions at the margin.  Therefore, and consistent with the economic impact analysis performed for the 2008 Rule, this analysis uses only variable costs and not fixed costs.  This approach assumes that manufacturers do not recover the production fixed costs associated with the control program by passing all or part of them to consumers through new engine price increases.  Instead, the analysis assumes that the Tier 4 fixed compliance costs displace the product development component of the existing prices.  In other words, manufacturers have ongoing product development programs the costs of which are already included in the current market price structure.  It is expected that the resources for those programs would be re-oriented toward compliance with the regulatory program until those costs are recovered for each manufacturer.  
   Nevertheless, because they are a cost to society in that they displace other product development activities that may improve the quality or performance of engines and equipment, the 2008 Rule accounted for these fixed costs separately in the economic impact analysis for that rule.  For the analysis for this Tier 4 relief rule, however, the focus is on estimating the economic impact of providing additional lead time for qualifying engines.  The fixed costs estimated in the 2008 Rule will still occur, either during the original time frame (five years prior to the effective date of the standards) for those engine manufacturers that have development programs in place, or the five years preceding the end of the additional lead time period.  These costs are not affected by the relief, and therefore are not included in this cost impact analysis.  
Methodology
   The economic impacts of extending the implementation date of the Tier 4 standards for the relevant engines are estimated by applying the estimated variable compliance costs of the 2008 Rule to the number of engines and vessels that are expected to qualify for relief.  These costs will not be incurred as a result of the relief.  Results are presented based on the estimated price impacts of the 2008 Rule for affected engines (behavioral approach) and on the entire engineering costs (full-cost pass-through approach).  
   Economic Impact = Compliance Cost per engine * Number of qualifying engines 
Data inputs
   The data inputs used in this analysis are from the economic impact modeling and engineering cost analysis performed for the 2008 rule.  Note that the data is reported in 2005$, for ease of convenience for comparing the results with the economic impact analysis performed for the 2008 Rule.  The conversion to 2015$ is performed as the last step in the analysis.
   The period of analysis is the same as that used for the inventory impacts analysis:  2020 through 2023 for engines in the 600 kW to 1,000 kW range, and 2020 through 2021 for engines in the 1,000 to 1,400 kW range.
   The quantities used for this analysis are the same as those used for the inventory impacts analysis:  21 engines per year in the 600 kW to 1,000 kW range and 4 engines per year in the 1,000 kW to 1,400 kW range.  This means 21 lobster boats per year (1 engine per vessel) and 2 pilot boats per year (2 engines per vessel).  
   The compliance costs used for this analysis are those developed for the 2008 Rule and are reported in Tables 13 and 14.  These include per engine variable costs, per vessel variable costs, and combined adjusted costs (price impacts) as modeled by the behavioral economic impact model.  All engines in the 600 kW to 1,000 kW range are assigned to the lobster boat market, which was modeled as "C1 Fishing Vessel" market in the 2008 Rule.  All engines in the 1,000 kW to 1,400 kW range are assigned to the pilot boat market, which was modeled as the "Other Commercial Vessel" market in the 2008 Rule.  
   In the 2008 Rule, these compliance costs were applied beginning in 2016 because the standards began to apply in 2016 for most of the engines in this category.  The costs in 2018 and later are reduced due to learning effect impacts.  This analysis applies the 2016 compliance costs beginning in 2020 to account for those early-year learning effects; this is appropriate because production has not yet begun or is only just beginning for the affected high-power density commercial engines.
Table 13 - Estimated Costs, C1 Fishing Vessel Market:  800-2,000 hp (Average Price per engine = $155,000; Average Price per Vessel = $1,085,000; one engine per vessel; 2005$)
Year
Variable Engineering Costs per Engine
Variable Engineering Costs per vessel
Total Variable Engineering Costs per Unit
Estimated Change in Price per Vessel+
2016
$15,196
$6,585
$21,781
$18,493
2017
$15,196
$6,587
$21,783
$18,493
2018
$11,618
$5,504
$17,122
$14,369
2019
$11,618
$5,501
$17,119
$14,366
2020
$11,618
$5,503
$17,121
$14,365
2021
$11,618
$5,504
$17,122
$14,363
2022
$11,618
$5,504
$17,122
$14,362
2023
$11,618
$5,504
$17,122
$14,360
+Estimated using the Economic Impact Model developed for the 2008 Rule.
Source:  2008 RIA, Tables 7A-3 and 7B-3

Table 14  -  Estimated Costs, Other Commercial Vessel Market:  800-2,000 hp (Average Price per Engine $155,000; Average Price per Vessel = $1,085,000; 2 engines per vessel; 2005$)
Year*
Variable Engineering Costs per Engine
Variable Engineering Costs Per Vessel
Total Variable Engineering Costs per Unit
Estimated Change in Price per Vessel+
2016
$15,196
$6,496
$36,888
$33,657
2017
$15,196
$6,731
$37,123
$33,780
2018
$11,618
$5,511
$28,747
$25,859
2019
$11,618
$5,462
$28,698
$25,664
2020
$11,618
$5,413
$28,649
$25,462
2021
$11,618
$5,365
$28,601
$25,264
2022
$11,618
$5,597
$28,833
$25,350
2023
$11,618
$5,547
$28,783
$25,158
+Estimated using the Economic Impact Model developed for the 2008 Rule.
Source:  2008 RIA, Tables 7A-3 and 7B-19

Results
   The economic impacts of the additional lead time are the sum of the savings associated with purchasing and installing Tier 3 engines and the reduction in operating costs associated with their use.  The estimated cost impacts do not include the use of waivers.  If engine manufacturers apply for and receive waivers post-2023, the estimated emission inventory impacts would increase and would extend for a longer period of time (the useful life of the engines produced subject to the waiver).
   
   The savings due to purchasing and installing Tier 3 engines are estimated to be about $1.5 million, using the behavioral approach, and about $1.8 million using the full-cost pass-through method.  These results are set out in Table 15.  
Table 15 Estimated Economic Impacts of Tier 4 Relief Rule (2005$)
Year
Estimate Cost Impacts - Behavioral Approach (price)
Estimated Cost Impacts - Full-Cost Pass-Through (engineering costs)

600-1,000 kW
1,000-1,400 kW
600-1,000 kW
1,000-1,400 kW
2020
 $388,353 
 $67,314
$457,401
$73,776 
2021
$388,353
 $67,560 
 $457,443 
 $74,246
2022
 $301,749 
 $   -   
 $359,562 
 $   -   
2023
 $301,686 
 $   -   
 $359,499 
 $   -   
Total
$1,380,141
$134,874
$1,633,905
$148,022
Grand Total
$1,515,015
$1,781,927
   In addition to the compliance cost savings, boat owners will also save on the operational costs associated with using the Tier 4 engines.  These include urea use, diesel particulate filter maintenance, and fuel consumption impacts.  In the 2008 Rule, these operational costs were estimated based on estimated fuel consumption for the relevant market segments.  
   To estimate the operating savings aspect of the economic impacts of additional lead time, the operating costs for the total Category 1 marine diesel engine sector reported in Table 5-48 of the 2008 RIA are attributed to the engines covered by the relief (21 engines in the 600 kW to 1,000 kW range and 4 engines in the 1,000 kW to 1,400 kW range) in proportion to the share of NOx emissions from all engines in those size categories to total C1 propulsion engines above 600 kW (the engines that are subject to the Tier 4 standards), for each year (2020-23).  This ratio is approximately the same for PM10 and NOx and averages about 0.26 percent for engines 600 kW to 1,000 kW, and 0.08 percent for engines 1,000 kW to 1,400 kW.  This results in an average operating cost of about $1,440 per engine.  It should be noted that this approach likely over-estimates the operating savings for lobster boats given their smaller number of hours of operation and engine load compared to assumptions used for the 2008 Rule inventory analysis operating cost analysis.
   Using an average annual operating cost of $1,440 per engine and a 13-year service life for C1 engines, the estimated lifetime operating cost impacts for the Tier 4 Relief Rule are about $1.7 million.  This is set out in Table 16.
Table 16 - Estimated Operating Cost Savings, 2020-35 (2005$)
Year
Annual No. Qualifying Engines
Population Qualifying Engines
Total Operating Cost Savings
2020
25
25
 $36,000 
2021
25
50
 $72,000 
2022
21
71
 $102,240 
2023
21
92
 $132,480 
2024
0
92
 $132,480 
2025
0
92
 $132,480 
2026
0
92
 $132,480 
2027
0
92
 $132,480 
2028
0
92
 $132,480 
2029
0
92
 $132,480 
2030
0
92
 $132,480 
2031
0
92
 $132,480 
2032
0
92
 $132,480 
2033
0
67
 $96,480 
2034
0
42
 $60,480 
2035
0
21
 $30,240 
TOTAL
   
   
$1,722,240
   Based on the above analyses, the total estimated economic impacts of the Tier 4 Relief Rule are estimated to be:
Total Equipment Costs + Total Operating Costs = Total Program Costs
For the behavioral approach, the estimated economic impacts during the period of relief are:
$1,515,015 + $1,722,240 = $3,237,255 in 2005$, or $3,880,329 in 2015$
For the full-cost pass-through approach, the estimated economic impacts during the period of relief are:
      $1,781,927 + $1,722,240 = $3,504,167 in 2005$, or $4,200,262 in 2015$
The conversion to 2015$ is achieved by applying the ratio of the 2015 GDP deflator index to the 2005 GDP deflator index (1.198648).


E.  Alternative Analysis
EPA sought comment on a regulatory alternative which, instead of providing a waiver process for builders of certain qualifying boats, would simply extend the Tier 4 effective date for engines from 600 to 1,000 kW for an additional 4 years.  This section provides estimated inventory, monetized forgone benefits, and cost impacts for this alternative approach using the methodology described above.  Note that this alternative was not adopted in the final rule.
The alternative approach would allow the production and sale of up to 84 additional engines in the 600-1,000 kW range (21 per year for 4 years), each of which would have a service life of 13 years. 
The estimated inventory impacts of the regulatory alternative program are set out in Table 17.  The additional four years of relief would increase the forgone lifetime emission inventory benefits to about 5,668 tons of NOx emissions and 119 tons of PM10 emissions (short tons).  This is an increase of about 1,904 tons of NOx and 40 tons of PM10 over the estimated inventory increase of the final program as adopted by the Agency, set out in Table 5.
The monetized forgone benefits of the emission increases associated with the regulatory alternative program, estimated as described in Section C, are set out in Tables 18 and 19.  Using reduced form health benefit per ton values, we estimate that the annual PM2.5-related forgone benefits do not exceed a high-end estimate of $4.4 million in any given year (2015$, 3% discount rate, mortality effect estimate derived from Lepeule et al., 2012). The total present value of the stream of forgone benefits ranges from $13.5 million to $44.6 million.    


Table 17 - Estimated Forgone Emission Reductions, Regulatory Alternative Program, Engine Lifetime, 2020-2039 (short tons)


Year
600-1,000 kW
1,000-1,400 kW
Total

New Engines 
 Population
NOx
PM10
New Engines 
Population
NOx
PM10
New Engines 
Population
NOx
PM10
2020
 21.0 
 21.0 
 36.6 
 0.8 
 4.0 
 4.0 
 71.5 
 1.5 
 25.0 
 25.0 
 108.1 
 2.3 
2021
 21.0 
 42.0 
 73.2 
 1.5 
 4.0 
 8.0 
 143.0 
 3.0 
 25.0 
 50.0 
 216.3 
 4.6 
2022
 21.0 
 63.0 
 109.8 
 2.3 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 71.0 
 252.9 
 5.3 
2023
 21.0 
 84.0 
 146.5 
 3.1 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 92.0 
 289.5 
 6.1 
2024
 21.0 
 105.0 
 183.1 
 3.9 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 113.0 
 326.1 
 6.9 
2025
 21.0 
 126.0 
 219.7 
 4.6 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 134.0 
 362.7 
 7.6 
2026
 21.0 
 147.0 
 256.3 
 5.4 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 155.0 
 399.4 
 8.4 
2027
 21.0 
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 21.0 
 176.0 
 436.0 
 9.2 
2028
 -   
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 -  
 176.0 
 436.0 
 9.2 
2029
 -   
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 -  
 176.0 
 436.0 
 9.2 
2030
 -   
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 -  
 176.0 
 436.0 
 9.2 
2031
 -   
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 -  
 176.0 
 436.0 
 9.2 
2032
 -   
 168.0 
 292.9 
 6.2 
 -   
 8.0 
 143.0 
 3.0 
 -  
 176.0 
 436.0 
 9.2 
2033
 -   
 147.0 
 256.3 
 5.4 
 -   
 4.0 
 71.5 
 1.5 
 -  
 151.0 
 327.8 
 6.9 
2034
 -   
 126.0 
 219.7 
 4.6 
 -   
 -  
 -   
 -   
 -  
 126.0 
 219.7 
 4.6 
2035
 -   
 105.0 
 183.1 
 3.9 
 -   

 -   
 -   
 -  
 105.0 
 183.1 
 3.9 
2036
 -   
 84.0 
 146.5 
 3.1 
 -   
 -  
 -  
 -  
 -  
 84.0 
 146.5 
 3.1 
2037
 -   
 63.0 
 109.8 
 2.3 
 -   
 -  
 -  
 -  
 -  
 63.0 
 109.8 
 2.3 
2038
 -   
 42.0 
 73.2 
 1.5 
 -   
 -  
 -  
 -  
 -  
 42.0 
 73.2 
 1.5 
2039
 -   
 21.0 
 36.6 
 0.8 
 -   
 -  
 -  
 -  
 -  
 21.0 
 36.6 
 0.8 
Total
 168.0 
 2,184.0 
 3,807.9 
 80.2 
 8.0 
 104.0 
 1,859.6 
 39.2 
 176.0 
 2,288.0 
 5,667.6 
 119.3 
Emission Increase Over Final Program (Table 5)
1,904.0
40.1



Table 18. Forgone monetized PM2.5-related health benefits, Regulatory Alternative Program (2015$; 3% Discount Rate)a 

Krewski et al., 2009b
Lepeule et al., 2012c
Year
NOX
PM2.5d
Total
NOX
PM2.5
Total
2020
$160,000
$310,000
$470,000
$380,000
$710,000
$1,090,000
2021
$320,000
$620,000
$900,000
$760,000
$1,400,000
$2,200,000
2022
$380,000
$720,000
$1,100,000
$890,000
$1,700,000
$2,500,000
2023
$430,000
$830,000
$1,300,000
$1,000,000
$1,900,000
$2,900,000
2024
$490,000
$930,000
$1,400,000
$1,100,000
$2,100,000
$3,300,000
2025
$540,000
$1,040,000
$1,600,000
$1,300,000
$2,400,000
$3,600,000
2026
$600,000
$1,140,000
$1,700,000
$1,400,000
$2,600,000
$4,000,000
2027
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2028
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2029
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2030
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2031
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2032
$650,000
$1,250,000
$1,900,000
$1,500,000
$2,800,000
$4,400,000
2033
$490,000
$940,000
$1,400,000
$1,150,000
$2,100,000
$3,300,000
2034
$330,000
$630,000
$960,000
$770,000
$1,440,000
$2,200,000
2035
$275,000
$520,000
$800,000
$640,000
$1,200,000
$1,840,000
2036
$220,000
$420,000
$640,000
$510,000
$960,000
$1,470,000
2037
$165,000
$310,000
$480,000
$380,000
$720,000
$1,100,000
2038
$110,000
$210,000
$320,000
$260,000
$480,000
$730,000
2039
$55,000
$100,000
$160,000
$130,000
$240,000
$370,000
Present Value
$19,370,000
Present Value
$44,600,000
a Totals may not sum due to rounding. A 3 percent discount rate is used to both account for a 20-year segmented cessation lag in the valuation of mortality and in the calculation of the present value of benefits.
b Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., ... & Thun, M. J. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality (No. 140). Boston, MA: Health Effects Institute.
c Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environmental health perspectives, 120(7), 965.
d PM2.5 emissions are assumed to be 97% of PM10 emissions.


Table 19. Forgone monetized PM2.5-related health benefits, Regulatory Alternative Program (2015$; 7% Discount Rate)a

Krewski et al., 2009b
Lepeule et al., 2012c
Year
NOX
PM2.5d
Total
NOX
PM2.5
Total
2020
$150,000
$290,000
$440,000
$350,000
$640,000
$990,000
2021
$300,000
$570,000
$900,000
$690,000
$1,300,000
$2,000,000
2022
$350,000
$670,000
$1,000,000
$810,000
$1,500,000
$2,300,000
2023
$410,000
$770,000
$1,200,000
$930,000
$1,700,000
$2,600,000
2024
$460,000
$870,000
$1,300,000
$1,040,000
$1,900,000
$3,000,000
2025
$510,000
$960,000
$1,500,000
$1,160,000
$2,100,000
$3,300,000
2026
$560,000
$1,060,000
$1,600,000
$1,280,000
$2,400,000
$3,600,000
2027
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2028
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2029
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2030
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2031
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2032
$610,000
$1,160,000
$1,800,000
$1,400,000
$2,600,000
$4,000,000
2033
$460,000
$870,000
$1,300,000
$1,050,000
$1,900,000
$3,000,000
2034
$310,000
$580,000
$890,000
$700,000
$1,300,000
$2,000,000
2035
$256,000
$490,000
$740,000
$590,000
$1,080,000
$1,670,000
2036
$205,000
$390,000
$590,000
$470,000
$870,000
$1,340,000
2037
$154,000
$290,000
$450,000
$350,000
$650,000
$1,000,000
2038
$103,000
$190,000
$300,000
$230,000
$430,000
$670,000
2039
$51,000
$100,000
$150,000
$120,000
$220,000
$330,000
Present Value
13,500,000
Present Value
30,400,000
a Totals may not sum due to rounding. A 7 percent discount rate is used to both account for a 20-year segmented cessation lag in the valuation of mortality and in the calculation of the present value of benefits.
b Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., ... & Thun, M. J. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality (No. 140). Boston, MA: Health Effects Institute.
c Lepeule, J., Laden, F., Dockery, D., & Schwartz, J. (2012). Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environmental health perspectives, 120(7), 965.
d PM2.5 emissions are assumed to be 97% of PM10 emissions.

The economic impacts of the proposed Tier 4 relief rule are the sum of the savings associated with purchasing and installing Tier 3 engines and the reduction in operating costs associated with their use.  
The savings due to purchasing and installing Tier 3 engines under the regulatory alternative program is estimated to be about $2.7 million, using a behavioral approach, and about $3.2 million, using the full-cost pass-through method.  These results are set out in Table 20.  These are the savings that would be incurred by purchasers of qualifying boats that have Tier 3 instead of Tier 4 engines
Table 20 - Estimated Economic Impacts, Regulatory Alternative Program (2005$)
Year
Estimate Cost Impacts - Behavioral Approach (price)
Estimated Cost Impacts - Full-Cost Pass-Through (engineering costs)

600-1,000 kW
1,000-1,400 kW
600-1,000 kW
1,000-1,400 kW
2020
 $388,353 
 $67,314 
 $457,401 
 $73,776 
2021
 $388,353 
 $67,560 
 $457,443 
 $74,246 
2022
 $301,749 
 $ -   
 $359,562 
 $ -   
2023
 $301,686 
 $ -   
 $359,499 
 $ -   
2024
 $301,665 
 $ -   
 $359,541 
 $ -   
2025
 $301,623 
 $ -   
 $359,562 
 $ -   
2026
 $301,602 
 $ -   
 $359,562 
 $ -   
2027
 $301,560 
 $ -   
 $359,562 
 $ -   
Total
$2,586,591
$134,874
$3,072,132
$148,022
Grand Total
$2,721,465
$3,220,154

Using an average operating cost of $1,440 per engine and a 13-service life for C1 engines, the estimated lifetime operating cost savings for the regulatory alternative program are about $3.3 million.  These savings are set out in Table 21.  
Table 21 - Estimated Operating Cost Savings, Regulatory Alternative Program, 2019-35 (2005$)
Year
Annual No. Qualifying Engines
Population Qualifying Engines
Total Operating Cost Savings
2020
25
25
 $36,000 
2021
25
50
 $72,000 
2022
21
71
 $102,240 
2023
21
92
 $132,480 
2024
21
113
 $162,720 
2025
21
134
 $192,960 
2026
21
155
 $223,200 
2027
21
176
 $253,440 
2028
0
176
 $253,440 
2029
0
176
 $253,440 
2030
0
176
 $253,440 
2031
0
176
 $253,440 
2032
0
176
 $253,440 
2033
0
151
 $217,440 
2034
0
126
 $181,440 
2035
0
105
 $151,200 
2036
0
84
 $120,960 
2037
0
63
 $90,720 
2038
0
42
 $60,480 
2039
0
21
 $30,240 
TOTAL
   
   
$3,294,720

   Based on the above analyses, the total estimated economic impacts of the regulatory alternative program would be:
Total Equipment Costs + Total Operating Costs = Total Program Costs
For the behavioral approach, the estimated economic impacts of the regulatory alternative program are:
$2,721,465 + $3,294,720 = $6,016,185 in 2005$, or $7,211,288 in 2015$
For the full-cost pass-through approach, the estimated economic impacts of the regulatory alternative program are:
      $3,220,154 + $3,294,720 = $6,514,874 in 2005$, or $7,809,040 in 2015$
Again, the conversion to 2015$ is achieved by applying the ratio of the 2015 GDP deflator index to the 2005 GDP deflator index (1.198648).
      These results are an increase of about $3.3 million and $3.6 million (2015$) for the behavioral and full-cost pass-through approaches, respectively, over the cost savings associated with the final program as adopted by the Agency.

F.  Analysis of 2020-2036 for E.O. 13771, Reducing Regulation and Controlling Regulatory Costs
   This final action is considered a deregulatory action under E.O. 13771, Reducing Regulation and Controlling Regulatory Costs.
   Table 17 presents the undiscounted compliance cost savings for the program, in 2016$.  These compliance cost savings are the estimated cost reductions that are expected to occur as a result of the additional lead time for implementing the Tier 4 standards for engines used in certain high-speed vessels (installing less expensive Tier 3 engines in new vessels during the relief period, 2020 to 2023).  As noted earlier, the impact of the waiver provision is not included in these cost reductions.  
   EPA calculated the present value of the estimated cost reductions using a seven percent discount rate.  EPA used an end-of-period discount rate for E.O. 13771 analysis.  The estimates for the program are presented in Table 22 and are from the perspective of 2016.  For purposes of E.O. 13771 accounting, the present value and equivalent annualized value estimates assume an infinite timeframe.  This is different from the cost analysis presented elsewhere in this memorandum, where the present values and equivalent annualized value estimates are for 2020 to 2035.  
   The present value of the stream of cost reductions is $2,346,106 when discounted at 7 percent (2016$).  The compliance cost reduction estimates represent the regulatory costs related to the regulatory allowance under E.O. 13771.  Table 23 also presents the equivalent annualized value.
Table 22  -  Compliance Cost Reduction for Additional Lead Time, 2020-35 (2016$)
Year
Compliance Cost Reduction
2020
$687,318.43 
2021
$731,564.53 
2022
$559,622.52 
2023
$596,191.72 
2024
$160,542.38 
2025
$160,542.38 
2026
$160,542.38 
2027
$160,542.38 
2028
$160,542.38 
2029
$160,542.38 
2030
$160,542.38 
2031
$160,542.38 
2032
$160,542.38 
2033
$116,916.73 
2034
$73,291.09 
2035
$36,645.54 


Table 23 -  Present Value and Equivalent Annualized Value of Compliance Costs for Additional Lead Time, 2020-35 (2016$)
Discount Rate
7%
Present Value (over infinity)
$2,346,106
Equivalent Annualized Value (over infinity)
$164,227



Appendix 1:  Inventory Estimates using 2008 Rule Assumptions
This Appendix presents inventory estimates using the 2008 Rule inventory data inputs.  As noted in Section A.2., the 2008 Rule inventory analysis used different inputs for average rated power, load factor, and engine activity for engines 600 kW to 1,000 kW, and different inputs for average rated power for engines 1,000 kW to 1,400 kW.  Table 20 contains the original 2008 Rule inventory data; Table 21 provides the estimated annual emission inventories using the 2008 Rule Inventory data, Table 22 provides the lifetime emission impacts, and Table 23 compares these results to the 2008 rule inventories for all marine diesel engines.
Table 20  -  2008 Rule Inventory Analysis Data Inputs
(original data in bold font; compare to Table 1)

Population (N)
Average Rated Power (P)
Load Factor (L)
Engine Activity (A)
NOx Emission Factor (EF) = ∆ NOx std (g/kW-hr)
PM10 Emission Factor (EF) = ∆ PM std (g/kW-hr)
600 <kW<=1000
21
678
0.79
4,503
3.8
0.08
1000<kW<=1400
4
1,176
0.79
4,503
3.8
0.08

Table 21 - Estimated Forgone Emission Reductions, 2020-2023, Based on 2008 Rule Data
(short tons; compare to Table 4)

600-1,000 kW
1,000-1,400 kW
Total
Year
NOx
PM10
NOx
PM10
NOx
PM10
2020
212.2
4.5
70.1
1.5
282.3
5.9
2021
212.2
4.5
70.1
1.5
282.3
5.9
2022
212.2
4.5
0.0
0.0
212.2
4.5
2023
212.2
4.5
0.0
0.0
212.2
4.5
Total tons 
848.6
17.9
140.2
3.0
988.8
20.8


Table 22- Estimated Forgone Emission Reductions, Engine Lifetime, 2020-2035, Based on 2008 Rule Data (short tons; compare to Table 5)
Year
New engines/year
Population
NOx emissions
PM10 emissions

2020
 25.0 
 25.0 
 282.3 
 5.9 
2021
 25.0 
 50.0 
 564.5 
 11.9 
2022
 21.0 
 71.0 
 776.7 
 16.4 
2023
 21.0 
 92.0 
 988.8 
 20.8 
2024
 -  
 92.0 
 988.8 
 20.8 
2025
 -  
 92.0 
 988.8 
 20.8 
2026
 -  
 92.0 
 988.8 
 20.8 
2027
 -  
 92.0 
 988.8 
 20.8 
2028
 -  
 92.0 
 988.8 
 20.8 
2029
 -  
 92.0 
 988.8 
 20.8 
2030
 -  
 92.0 
 988.8 
 20.8 
2031
 -  
 92.0 
 988.8 
 20.8 
2032
 -  
 92.0 
 988.8 
 20.8 
2033
 -  
 67.0 
 706.6 
 14.9 
2034
 -  
 42.0 
 424.3 
 8.9 
2035
 -  
 21.0 
 212.2 
 4.5 
Total
 92 
 1,196 
 12,855 
 271 

Table 23 - Estimated Lifetime Emissions Compared to National Marine Diesel Engine Inventories, 2020-2035, Based on 2008 Rule Data (short tons; compare to Table 9)

NOx Inventory
Affected Engines %
PM10 Inventory
Affected Engines %
C1 Commercial Propulsion Engines
 2,305,182 
0.7%
 54,643 
0.6%
High Power Density C1 Commercial Engines 
 244,509 
6.7%
 6,618 
5.2%
Affected Engines
 12,855 

271



Appendix 2:  Estimated Stream of Annual Total Costs, Full-Cost Pass-Through
(2005$, 2015$)

Year
Full-Cost Engineering
Operating Costs
Total (2005$)
Total (2015$)*

Lobster
Pilot



2020
 $457,401 
 $73,776 
 36,000 
 $567,177 
 $679,846 
2021
 $457,443 
 $74,246 
 72,000 
 $603,689 
 $723,611 
2022
 $359,562 
 $-   
 102,240 
 $461,802 
 $553,538 
2023
 $359,499 
 $-   
 132,480 
 $491,979 
 $589,710 
2024
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2025
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2026
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2027
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2028
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2029
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2030
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2031
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2032
 $-   
 $-   
 132,480 
 $132,480 
 $158,797 
2033
 $-   
 $-   
 96,480 
 $96,480 
 $115,646 
2034
 $-   
 $-   
 60,480 
 $60,480 
 $72,494 
2035
 $-   
 $-   
 30,240 
 $30,240 
 $36,247 
Total



 $3,504,167 
 $4,200,262 



NPV 3%
$3,107,870
$3,725,243



NPV 7%
$2,715,359
$3,254,759

*Stream of costs presented in both 2005$ and 2015$, to compare with stream of monetized benefits.  The stream of costs analysis assumes end of year discounting; the deflator used to convert from 2005$ to 2015$ is 1.198648.


Appendix 3:  Estimated Stream of Annual Total Costs, 
Full-Cost Pass-Through, Regulatory Alternative Program
(2005$, 2015$)

Year
Full-Cost Engineering
Operating Costs
Total (2005$)
Total (2015$)*

Lobster
Pilot



2020
 $457,401 
 $73,776 
 36,000 
 $567,177 
 $679,846 
2021
 $457,443 
 $74,246 
 72,000 
 $603,689 
 $723,611 
2022
 $359,562 
 $-   
 102,240 
 $461,802 
 $553,538 
2023
 $359,499 
 $-   
 132,480 
 $491,979 
 $589,710 
2024
 $359,541 
 $-   
 162,720 
 $522,261 
 $626,007 
2025
 $359,562 
 $-   
 192,960 
 $552,522 
 $662,279 
2026
 $359,562 
 $-   
 223,200 
 $582,762 
 $698,526 
2027
 $359,562 
 $-   
 253,440 
 $613,002 
 $734,774 
2028
 $-   
 $-   
 253,440 
 $253,440 
 $303,785 
2029
 $-   
 $-   
 253,440 
 $253,440 
 $303,785 
2030
 $-   
 $-   
 253,440 
 $253,440 
 $303,785 
2031
 $-   
 $-   
 253,440 
 $253,440 
 $303,785 
2032
 $-   
 $-   
 253,440 
 $253,440 
 $303,785 
2033
 $-   
 $-   
 217,440 
 $217,440 
 $260,634 
2034
 $-   
 $-   
 181,440 
 $181,440 
 $217,483 
2035
 $-   
 $-   
 151,200 
 $151,200 
 $181,236 
2036
 $-   
 $-   
 120,960 
 $120,960 
 $144,988 
2037
 $-   
 $-   
 90,720 
 $90,720 
 $108,741 
2038
 $-   
 $-   
 60,480 
 $60,480 
 $72,494 
2039
 $-   
 $-   
 30,240 
 $30,240 
 $36,247 
Total



$6,514,874
$7,809,040



NPV 3%
$5,457,764
$6,541,938



NPV 7%
$4,457,245
$5,342,667

*Stream of costs presented in both 2005$ and 2015$, to compare with stream of monetized benefits.  The stream of costs analysis assumes end of year discounting; the deflator used to convert from 2005$ to 2015$ is 1.198648.

