Offshore wind in South Carolina: an estimate of impacts on the regional economy and electrical rates
This article outlines steps required to estimate the economic and fiscal impact of a demonstration scale offshore wind farm on a state.
Authors: Kenneth Sercy; Robert T. Carey; Ellen W. Saltzman
This article outlines steps required to estimate the impact of a demonstration scale offshore wind farm on a state economy. First, the authors estimate current and potential economic impact on the state from construction and operation of a 40 MW offshore wind farm, including impacts on output, employment, wages and salaries, disposable income, and state and local government revenues. Second, they estimate the offshore wind farm's net impact on electric rates. This model takes into consideration the financing of wind farm construction costs over 20 years, as well as the anticipated costs of operating conventional generating facilities, some of whose output would be offset by power from the offshore wind farm.
This article outlines steps required to estimate the economic and fiscal impact of a demonstration scale offshore wind farm on a state. We also develop a model to estimate the offshore wind farm's net impact on electric rates. The proposed offshore wind farm is sited off the coast of South Carolina (SC), USA; economic, fiscal, and electric rate impacts are estimated for the entire state using state-specific inputs.
Economic and Fiscal Impact Analysis of a 40 MW Offshore Wind Farm
Economic Impact Modelling
In the US, three models are commonly used to estimate the economic impact of an activity on a regional economy: Regional Input–Output Modeling System (RIMS-II), Impact Analysis for Planning (IMPLAN), and Regional Economic Models, Inc. (REMI) [1–3]. Each model uses industry input–output data published by the US Department of Commerce's Bureau of Economic Analysis as a foundation. Input–output data quantifies production relationships among industries and commodities, which allow it to be used to estimate the impact in the change in demand for a particular commodity on all other industry sectors that are affected by production of that commodity. REMI adds the dimension of inter-regional trade and dynamic price effects.
Economic Impact Estimate Definitions
Employment is the number of jobs in the economy that are attributable to the operation and capital expenditures of firms involved in the actual production, construction, and operation and maintenance (O&M) of the wind farm.
Total compensation is the change in aggregate income from wages and salaries (including fringes) paid by all firms in the state to workers employed in the state.
Output is the dollar value of all goods and services produced in the state in a given year. This is similar to regional gross domestic product, but is not limited to final goods.
Net state or local government revenue is the revenue to state, county, and municipal governments throughout the state from all sources, including taxes, fees, and intergovernmental transfers, less expenses.
Direct effects are the workers employed in the actual production, installation, and O&M of the wind farm, their wage income, and the involved firms’ actual output.
Indirect effects are the jobs, wages, and output of second- and third-tier suppliers located within SC.
Induced effects are the ‘ripples’ expanding into the broader economy from the direct and indirect effects of spending of wage income by employees of the firm and its suppliers.
To estimate the economic and fiscal impacts of construction and operation of a 40 MW offshore wind farm on the state of SC, we used the REMI PI+ economic modelling engine. Changes to employment, income, or demand for products or services by either the private or the public sector can be used as input to the PI+ model. On the basis of these inputs, the model generates an estimate of the resultant variation from the projected baseline (status quo), as well as the effects on every industry sector. All REMI estimates include direct, indirect, and induced effects. The impact analysis described here is for the state of SC; however, the REMI modelling engine can also be used to model single or multi-county impacts.
Model Assumptions and Data Sources
The model used in this analysis assumes a 40 MW offshore wind farm constructed in 2016 and beginning operation in 2017. Estimated costs associated with this scenario assume:
- offshore installation of 3–5 MW turbines,
- 25 m water depth at the site,
- 100 miles between site and staging ports,
- 50 miles to land electrical interconnection, and
- <30 miles to the servicing port [4, 5].
On the basis of data provided by Santee Cooper, an SC state-owned electric utility, the total installed cost of turbines in the modelled offshore wind farm is assumed to be $6.46 million per megawatt (MW), or ∼$258 million for a 40 MW facility.
The economic impact of spending on O&M is modelled through 2036 in order to capture the first 20 years of facility operational life. All costs and impacts are reported in constant 2012 US dollars. O&M cost assumptions are:
- fixed O&M costs of $66.16 per kilowatt (kW) year in the first year,
- variable O&M costs of 0.73 cents per kilowatt-hour (kWh) in the first year, and
- fixed and variable O&M costs increasing at a rate of 2% per year beginning in 2017 to account for replacement parts and general wear and tear on equipment.
Component Manufacturing and Installation
The wind turbine component portion of the model estimates the economic impact on the state from the production of individual wind turbine components. Each component's production was assigned to 1 of 12 North American Industry Classification System (NAICS) 2-digit–5-digit sectors .
The offshore wind farm installation model estimates the economic impact of labour and port services, land, and marine transportation, and other activities. Proportional cost estimates for each were derived from the National Renewable Energy Laboratory's (NREL) Offshore Jobs and Economic Development Impact (JEDI) model and from data provided by Santee Cooper . For example, the local share of Cement Manufacturing (NAICS 32731), used in turbine foundations, was estimated to be 77% for SC, whereas the local share for Ferrous Metal Foundries (NAICS 33151), used for hubs, isolation mounts, and support structures, was estimated at zero.
Given a total installed cost per MW of $6.46 million, the assumed percentage of in-state provision of services of each activity was determined using 2009 IMPLAN regional purchase coefficient tables  and in consultation with industry sources.
Operations and Maintenance Activities
The O&M activities model estimates the impact of ongoing wind farm O&M on the state. This model includes the impact from technician and engineering jobs and water transport. It also contains a levellised estimate of replacement parts costs (these costs increase over time as turbines age). The proportional cost of each of these O&M activities was extracted on a per-MW basis from the NREL Offshore JEDI model and from consultation with industry sources .
The total cost of operations and maintenance activities per installed MW was estimated to be $88 500 in 2017, the first year of wind farm operation. This figure includes fixed and variable per-MW costs. For subsequent years, O&M costs are assumed to increase over the life of the wind farm at a rate of 2% per year. The in-state share of replacement part manufacturing and O&M services were estimated using the same method as in the previous model.
Economic and Fiscal Impacts: Turbine Component Manufacturing and Installation
Table 1 shows the average annual economic impact on the state resulting from wind turbine component manufacture and turbine installation off the SC coast. For the proposed offshore wind farm, we assume that 40 MW of turbine components will be manufactured, purchased, and installed in 1 year, 2016. In that year, this activity would generate about 959 total jobs in SC, or about 24 jobs per MW of turbine components installed. An estimated 58 cents of every dollar invested in manufacture and installation of wind farm components would remain in SC through direct investment and indirect and induced effects.
|ESTIMATED IMPACT||ESTIMATED IMPACT, MW|
|total jobs||959 jobs||24 jobs|
|total compensation||$46.3 million||$1.2 million|
|output||$148.4 million||$3.7 million|
|net state revenue||$2.4 million||$60 450|
|net local revenue||$1.1 million||$28 340|
Table 1: Impact of turbine component manufacture and installation, 2016
The economic activity associated with production and installation of turbine components generates both government revenue (taxes and fees) and costs (demand on infrastructure). The model estimates that the 1 year increase in combined state and local government revenues outweighs the increase in government costs for a net impact of $3.5 million. This model does not assume any financing of industry inducements using state or local government general revenue funds or through tax increases.
In this model, construction and installation activities occur in only 1 year. However, these large economic impacts would continue for a number of years during build out of a commercial scale offshore wind farm.
Economic and Fiscal Impacts: Offshore Wind Farm O&M
Economic and fiscal impact estimates for the O&M phase begin in 2017, the year after the installation of the 40 MW facility. As shown in Table 2 , O&M activities are estimated to generate ten jobs per year, on average, or about 0.26 jobs per MW installed over the 20 years of operational life modelled.
|AVERAGE IMPACT, YEAR||IMPACT, MW/YEAR|
|total jobs||ten jobs||0.26 jobs|
|total compensation||$934 000||$23 300|
|output||$2.8 million||$70 900|
|net state revenue||−$115 000||−$2875|
|net local revenue||−$107 000||−$2675|
Table 2: Average annual impact of O&M activities, 2017–2036
Wind farm O&M activities are estimated to generate average annual output valued at $2.8 million a year during the two decades. Aggregated local government net revenue is estimated to decrease by ∼$107,000 per year on average. This loss will be spread across multiple counties and municipalities, depending on the geographic distribution of new residents and economic activity within the state. State government net revenue is projected to decrease by about $115 000 per year on average.
These small negative impacts in net revenue for state and local governments are due to the increase in demand for government services by new residents projected to relocate to the state due to the positive impact on relative wages from construction and the subsequent ongoing smaller wage impact from O&M. These increased demands, coupled with the small, short-term negative impact that higher relative wage rates from the 2016 construction boom is predicted to have on employment in the following years, results in increased outlays that exceed the increase in revenue.
The model estimates that about 67 cents out of every dollar spent on O&M will remain in SC. Since O&M activities continue after completion of wind farm installation, these economic impacts will persist as long as the wind farm continues to operate.
Electric Rate Impact of a 40 MW Offshore Wind Farm
This section estimates electric rate impacts to residential, commercial, and industrial energy users resulting from construction and operation of a 40 MW offshore wind farm sited off the coast of SC. Our approach is directly applicable to a range of renewable energy project types in traditionally regulated jurisdictions, and could be adapted for applications to other technology types and to projects in jurisdictions with alternative market structures.
A project rate impact assessment should incorporate three factors: (i) technology capital and O&M costs; (ii) avoided fuel and other production costs due to project electricity generation; and (iii) allocation of capital costs, O&M costs, and avoided production costs to customer classes.
Wind Farm Capital Costs and O&M Costs
For an electric generating project to move forward, the revenue expected to be received by the owner and operator of the project must be sufficient to cover the costs of construction and operation and the cost of capital. Thus, a capital recovery model is required. Large utilities and public institutions may maintain in-house capital recovery models and/or may have resources to create such models as needed. Public models also exist and can be used directly or adapted for the purpose of a rate impact assessment. We used the Cost of Renewable Energy Spreadsheet Tool (CREST), a cash flow model developed under the direction of NREL .
The CREST model computes capital and O&M costs per kWh for a facility, using generating capacity, project lifetime, installed cost, financing parameters, and available incentives. It also computes the total cost of energy production each year over the life of the facility. The CREST model accounts for tax liability, asset depreciation, debt service, and equity investor return requirements. Inputs such as installed cost may be obtained from public sources such as the US Energy Information Agency (EIA). Financing inputs may be obtained from public filings for regulated vertically integrated utilities, or potentially sourced from developers for projects expected to be owned by third parties.
Key assumptions in the capital recovery model are:
- Capital cost of construction is financed over 20 years from 2017 to 2036.
- Operations and maintenance expenditures occur during years 2017–2036.
- The proposed wind farm is jointly owned by SC's electric utilities; accordingly, the project's weighted-average cost of capital is a blended rate based on these utilities’ recent capital structures and cost of debt and equity financing.
- Financial incentives are excluded from the capital recovery model.
We obtained installed cost per MW of generating capacity and annual O&M cost estimates from an in-state utility. These figures were based on internal research and equipment vendor contacts developed as part of ongoing investigations into offshore wind development in SC. Key inputs to the CREST model are provided in Table 3. Fig 1 shows the annual capital recovery and O&M costs of the proposed wind farm over its assumed 20 year life.
Fig 1: Project capital recovery and O&M costs
|generator nameplate capacity||40 MW|
|project useful life||20 years|
|total installed cost||$6459 kW −1|
|fixed O&M cost||$66.16 kW −1 year|
|variable O&M cost||$0.0073 kWh −1|
|annual O&M cost inflation||2% year −1|
|blended after-tax weighted-average cost of capital||6.11>#/td###|
Table 3: Capital recovery model inputs
Avoided Production Costs
Secondary rate impacts occur when wind-generated electricity allows the utility system to avoid purchasing power or burning fuel and incurring other variable production costs of running fossil fuel-based (coal, oil, and gas) generating units. These avoided power purchases and/or production costs offset some of the rate impacts from capital recovery and O&M costs.
Avoided power purchases, fuel burn, and variable O&M may be only part of the total avoided cost value of a project to a utility system. Additional benefits such as capacity value, transmission and distribution line loss avoidance, and risk hedging related to fuel purchases or regulatory requirements may apply depending on the technology used. Variable generating resources such as solar and wind also may impose integration costs on the system. We excluded these factors given the small scale of the SC offshore wind farm project relative to utility system size.
As with capital recovery modelling, large utilities and public institutions may maintain in-house models for projecting system production costs. However, these sophisticated models require considerable time and resources to run. Here we estimated the savings from avoided production costs by creating a simple, spreadsheet-based production cost model that dispatches generating units based on the marginal cost of generation of available units on the system during hourly segments of customer demand. Since additional constraints such as unit ramp rates, minimum up and down times, minimum output, and provision of operating reserves would raise total production costs, this approach is expected to yield conservative estimates of cost savings from displacement of conventional generation by wind farm production.
Marginal Cost of Generation
The marginal cost of generation for each generating unit during each hour of customer demand was calculated as follows, excluding unit conversion factors:
where i is the generating unit I; t is the time period t (h); MC i , t is the marginal cost of generating electricity for unit i during time period t, in $/MW hour (MWh); HR i is the heat rate of unit i, in British thermal unit (Btu)/kWh; FP i,t is the fuel price applicable to unit iduring time period t, in $/mega Btu (MBtu); CEF i is the carbon dioxide (CO 2) emissions factor for the fuel type applicable to unit i, in lb/MBtu; CP t is the price of a CO 2 emissions allowance during time period t, in $/metric ton, and OM i, t is the non-fuel variable O&M cost for unit i during time period t, in $/MWh.
Thus, for each hour of customer demand, marginal unit costs are calculated and the lowest cost units are dispatched first, followed by progressively more costly units until customer demand for that hour is satisfied.
We ran scenarios with and without wind power production for each hour of customer demand, over the 20 year project life. The difference in hourly costs of these two scenarios is the cost savings from displacement of conventional generation by wind farm production.
An hourly avoided production cost simulation requires two sets of load data: a baseline hourly load dataset, and a net load dataset for which the hourly generation of the project of interest is subtracted from the baseline load in each hour. The project generation dataset must be time-synchronised with the baseline load dataset for the model to accurately identify the generating units displaced in each hour.
We derived system load inputs using historical hourly load data that are publicly available through the US Federal Energy Regulatory Commission . Since the majority of SC's electric load is summer peaking and exhibits daily and seasonal demand patterns that are broadly similar, we used load data from a single SC utility as an adequate generic representation of demand patterns within the state.
On the basis of the expected load growth rates reported by SC utilities in their 2012 and 2013 Integrated Resource Plans (IRPs) , we assumed a 1% annual growth rate in summer and winter peak and off-peak demand and applied these growth rates to historical load data.
Wind Output Profile
Net load input to the production cost model requires an hourly project generation dataset. In 2011, AWS Truepower LLC created wind generation output data for offshore locations in the southeastern US . We used Block 6 data corresponding to waters off the coast at Georgetown, SC. We averaged the 10 min net power data into hourly values, then scaled these values to equivalent output for a 40 MW offshore wind farm.
The portfolio of generating units used as inputs to the production cost model is broadly representative of expected future capacity mixes of North Carolina (NC) and SC utilities. In an offshore wind farm shared ownership scenario, hourly power output would likely be divided proportionately among utilities based on ownership share. Thus the wind power would displace some fossil fuel generation from each utility system.
We modelled a simplified system in which the full output of the wind farm displaces conventional generation from a single generic Carolinas utility. This generic utility system is composed of existing and planned generating units located in NC and SC.
The initial 2017 capacity mix was created using the US Environmental Protection Agency's National Electric Energy Data System (NEEDS) database, version 4.10 . NEEDS contains US generating unit IDs, locations, capacities, technology and fuel types, heat rates, and other key unit data. We summed existing NC and SC generation capacities by technology type and identified the percentage contribution of each technology to the full portfolio. We then selected individual generating units to populate our generic utility system such that:
- total capacity of the model utility could meet 2017 system peak load input plus a 15–20% reserve margin and
the percentage contribution of each technology type reflected the actual NC–SC portfolio as represented in NEEDS, as adjusted for completed or expected unit additions and retirements through 2016 (see Table 4).
|NC–SC GENERATION||MODEL UTILITY|
|GENERATING TECHNOLOGY||CAPACITY, MW||PERCENTAGE OF TOTAL, >#/TH###||CAPACITY, MW||PERCENTAGE OF TOTAL, >#/TH###|
|coal steam||20 642||40.5||2144||36.7|
Table 4: NC–SC capacity mix vs. model utility capacity mix
Next, we created a roadmap of unit additions for our generic utility system based on expected capacity additions in the NC–SC system over 20 years, as indicated in utility IRPs. The unit additions maintain a 15–20% system reserve margin as peak demand grows annually by 1%. The production cost model adds new units to the system according to the roadmap and system dispatch is altered based on available capacity and marginal cost.
Fuel and CO 2 price assumptions
Numerous energy industry consulting firms and government agencies maintain fuel price projections and potential CO 2 allowance price scenarios. We used delivered fuel price projections for the power sector in the South Atlantic region of the US developed by the EIA for its Annual Energy Outlook publication . For CO 2 prices, we used EIA's medium case trajectory from the Annual Energy Outlook and assumed CO 2 compliance begins in 2017.
Production cost modelling results
Fig 2 shows the estimated annual production cost savings resulting from wind farm operation, broken down by cost category. Annual savings range from just over $6 million in the first year of wind farm operation to over $10 million in 2036.
Fig 2: Avoided costs of conventional electric power generation
The final component of a project rate impact analysis accounts for how capital recovery and operating costs and savings are allocated among utility customer classes. In the US, regulated utilities use cost allocation formulas to divide the costs of capital assets and fuel fairly among customer classes. Then, they design rates such that electric sales allow the utility to collect the appropriate amount of revenue from each class .
Where detailed utility cost allocation data are not accessible, are not directly applicable (for example, in a multi-jurisdictional analysis such as ours), or are too complex for the rate impact assessment at hand, simple cost allocation methods can serve as substitutes.
Our study assumed that capital asset revenue requirements are allocated among customer classes in a non-uniform manner based on class equipment usage, whereas fuel revenue requirements are allocated evenly among all kWhs consumed on the system, regardless of customer class. We used statewide electric utility revenue data from the EIA [17,18] to derive capital asset cost allocators for average SC residential, commercial, and industrial customers, in keeping with our statewide focus. Fuel cost savings were allocated evenly among all system kWhs on a statewide basis.
State electric sector fuel expenditure data were used along with statewide sales and revenue data to derive total rates, fuel-only rates, and non-fuel rates for each customer class. Non-fuel rates were used to derive capital asset allocators and fuel-only rates were used to derive fuel allocators. Table 5 shows the capital asset and fuel savings allocators for each customer class. Note that various alternative allocation approaches (for example, a per-customer allocation) could be investigated in order to further inform decision-makers about potential project rate impacts across different customer classes.
|RATE CLASS||CAPITAL ASSET||FUEL SAVINGS|
Table 5: Cost allocators
Offshore Wind Farm Rate Impacts
Once the annual capital cost, annual avoided costs, and cost allocators are estimated, the total project costs and benefits are divided by projected annual sales to obtain annual project rate impacts on a cost-per-unit basis (Table 6).
|RATE CLASS||RATE CHANGE, $/KWH|
Table 6: Est. rate impacts by rate class
Table 7 illustrates how these estimated rate changes would impact individual customer electric bills. Monthly kWh consumption and electric bill charges were calculated for the average customer in each class using consumption and revenue data from the EIA . For example, residential customers are estimated to pay an additional $0.42 month −1 on average over the life of the wind farm. This would be an increase of about 0.3% over the average 2012 residential electric bill.
|RATE CLASS||AVERAGE, KWH/MO||AVERAGE, BILL/MO||DOLLAR INCREASE FROM WIND||PERCENTAGE INCREASE FROM WIND|
|industrial||534 380||$32 173||$57.02||0.2|
Table 7: Est. rate impact on customer bill, by rate class
Conclusion and lessons learned
Electric utility planning processes and broader public policy decisions often assess emerging electric generating technologies on a cost-per-unit basis such as $/MWh, or levellised cost [19,20]. This approach is relatively simple and transparent and can serve as a technology deployment screening tool. However, decision-making in the public arena could benefit from additional information, such as the anticipated impacts to the state or regional economy resulting from facility construction and operation as well as the resulting electric rate impacts. This case study demonstrates how such broader assessments can be made.
The strength of any economic impact analysis is a combination of the model used and the quality of the inputs. Component and installation costs will vary over time as technologies change. Moreover, because of the lack of extant offshore wind in the US, the full cost of installation is yet unknown. An additional factor is the local share of component manufacture and installation, which is crucial in estimating the regional impact; these shares are different for each region and similarly change over time. Both the economic impact and the impact on electric rates will be sensitive to such variation in installation cost. A decrease in component costs, for example, will result in less cost that needs to be offset by an increase in electric rates while decreasing the dollar amount of the economic impact from installation due to the lower quantity of dollars injected into the economy. The electric rate impact will also be sensitive to inputs related to project financing and incentives claimed.
In the electric rate impact analysis, the most resource-intensive component was the avoided production cost estimate. Multiple large datasets from different sources were required, and a spreadsheet tool with numerous formulas and macros was designed from scratch to compute hourly avoided costs, broken down by cost category. For researchers who seek a granular analysis without access to sophisticated production cost models, a spreadsheet-based hourly dispatch model may be appropriate. However, researchers also may find simpler proxy-unit methods appealing, as these can yield estimates of project avoided costs that are reasonable as first-order approximations. Regardless of the specific approach used to estimate avoided production costs, these cost savings estimates will be sensitive to inputs related to system generator operation, such as fuel prices and emissions allowance costs over the life of the project.
Note: Estimates in 2012 dollars.
- ‘RIMS II Online Order and Delivery System’. Available at https://www.bea.gov/regional/rims/rimsii/, accessed January 2015.
- ‘IMPLAN’. Available at http://www.implan.com/, accessed January 2015.
- ‘Regional Economic Models, Inc.’. Available at http://www.remi.com/, accessed January 2015.
- ‘Jobs and economic development impact models’. Available at http://www.nrel.gov/analysis/jedi/, accessed June 2014.
- Hamilton B. Lantz E. Paidipati J.: ‘.
- See ‘North American Industry Classification System’. Available at http://www.census.gov/eos/www/naics/, accessed January 2015.
- ‘Jobs and economic development impact models’. Available at http://www.nrel.gov/analysis/jedi/, accessed June 2014.
- See ‘IMPLAN’. Available at http://www.implan.com/, accessed January 2015.
- ‘Jobs and economic development impact models’. Available at http://www.nrel.gov/analysis/jedi/, accessed June 2014.
- ‘Form No. 714 - Annual Electric Balancing Authority Area and Planning Area Report’. Available at http://www.ferc.gov/docs-filing/forms/form-714/data.asp, accessed December 2014.
- ‘Utilities’. Available at http://www.energy.sc.gov/utilities/, accessed December 2014.
- AWS Truepower LLC: ‘ .
- ‘EPA's IPM Base Case v.4.10’. Available at http://www.epa.gov/airmarkets/programs/ipm/BaseCasev410.html, accessed January 2015.
- US Department of Energy, Energy Information Administration (EIA): .
- Regulatory Assistance Project: ‘.
- EIA: ‘State Energy Data System’. Available at http://www.eia.gov/state/seds/, accessed December 2014.
- EIA: ‘Form 861’. Available at http://www.eia.gov/electricity/sales_revenue_price/, accessed December 2014.
- ‘Levelized Cost of New Generation Resources in the Annual Energy Outlook 2013’. Available at http://www.eia.gov/forecasts/aeo/er/pdf/electricity_generation.pdf, accessed December 2014.
- Branker K. Pathak M. J. M. Pearce J. M.: ‘ A review of solar photovoltaic levelized cost of electricity’, .