Economies of Scale, Learning Curves and Offshore Wind Development - - PowerPoint PPT Presentation

▶

Nov 04, 2023 267 likes •715 views

Economies of Scale, Learning Curves and Offshore Wind Development Costs Presentation by Greg Upton David Dismukes and Gregory Upton Jr. Louisiana State University Greg Upton (LSU) Offshore Wind July 29, 2013 1 / 22 Introduction The worlds

SLIDE 1

Economies of Scale, Learning Curves and Offshore Wind Development Costs

Presentation by Greg Upton David Dismukes and Gregory Upton Jr. Louisiana State University

Greg Upton (LSU) Offshore Wind July 29, 2013 1 / 22

SLIDE 2

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark.

Greg Upton (LSU) Offshore Wind July 29, 2013 2 / 22

SLIDE 3

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark. Since 1991, over 40 additional OSWs have been constructed, none of which are located in the United States.

Greg Upton (LSU) Offshore Wind July 29, 2013 2 / 22

SLIDE 4

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark. Since 1991, over 40 additional OSWs have been constructed, none of which are located in the United States. Recently, though, there has been interest in developing offshore wind in the U.S., as there are currently nine OSW projects totaling over 2,300 MW of total capacity in the permitting and development process (USDOE 2011).

Greg Upton (LSU) Offshore Wind July 29, 2013 2 / 22

SLIDE 5

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark. Since 1991, over 40 additional OSWs have been constructed, none of which are located in the United States. Recently, though, there has been interest in developing offshore wind in the U.S., as there are currently nine OSW projects totaling over 2,300 MW of total capacity in the permitting and development process (USDOE 2011). It is still uncertain if any of these projects will be completed and two reasons are cited for this holdup. (USDOE 2011)

Relatively high cost of offshore wind compared to other forms of energy.

Difficulty in receiving permitting.

Greg Upton (LSU) Offshore Wind July 29, 2013 2 / 22

SLIDE 6

Introduction

These two issues are interrelated, though, as relatively expensive projects are less likely to receive approval than relatively less expensive projects.

Greg Upton (LSU) Offshore Wind July 29, 2013 3 / 22

SLIDE 7

Introduction

These two issues are interrelated, though, as relatively expensive projects are less likely to receive approval than relatively less expensive projects. Another potential reason for why projects are having difficulty receiving permitting is because there is currently no consistent methodology available for comparing the cost of a proposed off-shore wind project to other similar off-shore wind projects around the world as this is not straightforward for a variety of reasons.

Different areas have different physical characteristics, and these heterogenous conditions can have a potentially large impact on costs.

Economic environments in which existing OSWs were built are also heterogeneous.

Greg Upton (LSU) Offshore Wind July 29, 2013 3 / 22

SLIDE 8

Introduction

These two issues are interrelated, though, as relatively expensive projects are less likely to receive approval than relatively less expensive projects. Another potential reason for why projects are having difficulty receiving permitting is because there is currently no consistent methodology available for comparing the cost of a proposed off-shore wind project to other similar off-shore wind projects around the world as this is not straightforward for a variety of reasons.

Different areas have different physical characteristics, and these heterogenous conditions can have a potentially large impact on costs.

Economic environments in which existing OSWs were built are also heterogeneous.

The forty-one OSWs examined in this study were built in eight different countries over a twenty one-year period.

Greg Upton (LSU) Offshore Wind July 29, 2013 3 / 22

SLIDE 9

Introduction

This paper will combine three different literatures:

1 Economies of Scale 2 Learning Curves 3 Overnight Cost Greg Upton (LSU) Offshore Wind July 29, 2013 4 / 22

SLIDE 10

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004).

Greg Upton (LSU) Offshore Wind July 29, 2013 5 / 22

SLIDE 11

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004). There has been very little research on economies of scale in either the

n-shore or off-shore wind market.

Greg Upton (LSU) Offshore Wind July 29, 2013 5 / 22

SLIDE 12

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004). There has been very little research on economies of scale in either the

n-shore or off-shore wind market.

◮ USDOE (2011) finds evidence of economies of scale in relatively small

n-shore windfarms (less than 20MW).

Greg Upton (LSU) Offshore Wind July 29, 2013 5 / 22

SLIDE 13

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004). There has been very little research on economies of scale in either the

n-shore or off-shore wind market.

◮ USDOE (2011) finds evidence of economies of scale in relatively small

n-shore windfarms (less than 20MW).

◮ Junginger & Turkenburg (2005) find that for orders of over 100

turbines, there is approximately a 30-percent reduction in the list price.

Greg Upton (LSU) Offshore Wind July 29, 2013 5 / 22

SLIDE 14

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004). There has been very little research on economies of scale in either the

n-shore or off-shore wind market.

◮ USDOE (2011) finds evidence of economies of scale in relatively small

n-shore windfarms (less than 20MW).

◮ Junginger & Turkenburg (2005) find that for orders of over 100

turbines, there is approximately a 30-percent reduction in the list price.

◮ Snyder & Kaiser (2009) find a positive relationship between total cost

and total capacity, but do not empirically test for the presence of economies of scale in their specification.

Greg Upton (LSU) Offshore Wind July 29, 2013 5 / 22

SLIDE 15

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

n the presence of learning curves in offshore wind has been

conducted.

Greg Upton (LSU) Offshore Wind July 29, 2013 6 / 22

SLIDE 16

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

n the presence of learning curves in offshore wind has been

conducted. Junginger & Turkenburg (2005) empirically analyze experience curves in wind farms worldwide and find significant learning effects, but limit their analysis to on-shore wind.

Greg Upton (LSU) Offshore Wind July 29, 2013 6 / 22

SLIDE 17

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

n the presence of learning curves in offshore wind has been

conducted. Junginger & Turkenburg (2005) empirically analyze experience curves in wind farms worldwide and find significant learning effects, but limit their analysis to on-shore wind. USDOE (2011), on the other hand, finds no evidence of a learning curve in on-shore wind in the United States.

Greg Upton (LSU) Offshore Wind July 29, 2013 6 / 22

SLIDE 18

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

n the presence of learning curves in offshore wind has been

conducted. Junginger & Turkenburg (2005) empirically analyze experience curves in wind farms worldwide and find significant learning effects, but limit their analysis to on-shore wind. USDOE (2011), on the other hand, finds no evidence of a learning curve in on-shore wind in the United States. Snyder & Kaiser (2009) test for the change in cost of off-shore windfarms over time holding other factors such as distance to shore, turbine size, capacity, and water depth constant, but do not find a decline in cost over time.

Greg Upton (LSU) Offshore Wind July 29, 2013 6 / 22

SLIDE 19

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

◮ The forty-one off-shore windfarms analyzed in this paper were built

ver a twenty one year period in eight countries that have different

exchange rates, interest rates, and inflation rates over time.

◮ Some of these wind farms were constructed in just a few months while

ther projects took multiple years to complete.

Greg Upton (LSU) Offshore Wind July 29, 2013 7 / 22

SLIDE 20

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

◮ The forty-one off-shore windfarms analyzed in this paper were built

ver a twenty one year period in eight countries that have different

exchange rates, interest rates, and inflation rates over time.

◮ Some of these wind farms were constructed in just a few months while

ther projects took multiple years to complete.

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

◮ The forty-one off-shore windfarms analyzed in this paper were built

ver a twenty one year period in eight countries that have different

exchange rates, interest rates, and inflation rates over time.

◮ Some of these wind farms were constructed in just a few months while

ther projects took multiple years to complete.

Overnight Costs

Step 1: Cumulative Percentt =

1 − cos

t − si fi − si × π 2 αβ (1) α is assumed to be 4.082 and β is 3.25, which is consistent with previous literature (EIA 1998) Step 2: Next, we calculate the percent of the expenditures incurred in each month

ver the course of the construction period. This can be done by taking the

derivative of (1) with respect to time, or more simply, the discrete difference between Cumulative Percentt and Cumulative Percentt−1. Percent Expenditurei,t = Cumulative Percenti,t − Cumulative Percenti,t−1 (2)

Greg Upton (LSU) Offshore Wind July 29, 2013 8 / 22

SLIDE 23

Greg Upton (LSU) Offshore Wind July 29, 2013 9 / 22

SLIDE 24

Overnight Costs

Step 3: Discount Factori,t = Percent Costi,t(1 + ri,t)(1 + inflationi,t) (3) Where ri,t is the interest rate of country i in time t and inflationi,t is the inflation rate in country i in time t. Finally, the overnight cost is calculated as follows. Step 4: Overnight Costi = Total Costi T

t=1 Discount factori,t

(4)

Greg Upton (LSU) Offshore Wind July 29, 2013 10 / 22

SLIDE 25

Empirical Specification

The following empirical specification will be used to test for economies of scale: ln(Overnight Costi) = α + β ln Capacityi + X

′

i δk + εi

(5) Overnight Cost is in 2012 U.S. Dollars. Capacity is the total capacity in MW of windfarm i. ˆ β is the estimated elasticity of cost with respect to capacity. If β is estimated to be less than 1, then we will have evidence of economies of scale in off-shore windfarms. X

′

i is a list of covariates that include the distance to shore, water

depth and country fixed effects.

Greg Upton (LSU) Offshore Wind July 29, 2013 11 / 22

SLIDE 26

Empirical Specification

The following empirical specification will be used to test for learning curves: ln(Overnight Costi) = α+γ ln Cumulative Capacityi+β ln Capacityi+X

′

i δk+εi

(6) ˆ γ is the estimated percent change in overnight cost associated with respect to a one percent change in cumulative capacity. If learning curves are present, then we expect γ < 0. X

′

i includes the same list of control variables as seen above in the

economies of scale specification.

Greg Upton (LSU) Offshore Wind July 29, 2013 12 / 22

SLIDE 27

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide.

Greg Upton (LSU) Offshore Wind July 29, 2013 13 / 22

SLIDE 28

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide. Data on total cost, capacity, water depth and distance to shore are from 4COffshore.

Greg Upton (LSU) Offshore Wind July 29, 2013 13 / 22

SLIDE 29

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide. Data on total cost, capacity, water depth and distance to shore are from 4COffshore. Overnight cost calculations are completed with the overnightcost Stata program that accompanies this paper.

Greg Upton (LSU) Offshore Wind July 29, 2013 13 / 22

SLIDE 30

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide. Data on total cost, capacity, water depth and distance to shore are from 4COffshore. Overnight cost calculations are completed with the overnightcost Stata program that accompanies this paper. Interest rates and inflation rates are from the World Bank.

Greg Upton (LSU) Offshore Wind July 29, 2013 13 / 22

SLIDE 31

Data

Table : Windfarm Information

Windfarm Nation Year MW Depth Distance Belwind Phase 1 Belgium 2010 165 22.5 46 Gunfleet Sands UK 2010 173 6.5 7 Thanet UK 2010 300 18.5 12 EnBW Baltic I Germany 2011 48 17.5 16 Greater Gabbard UK 2011 504 20.5 36 Sheringham Shoal UK 2011 317 18.5 23 Walney Phase 1 UK 2011 184 21 14

Greg Upton (LSU) Offshore Wind July 29, 2013 14 / 22

SLIDE 32

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 15 / 22

SLIDE 33

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 16 / 22

SLIDE 34

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 17 / 22

SLIDE 35

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 18 / 22

SLIDE 36

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 19 / 22

SLIDE 37

Results

Greg Upton (LSU) Offshore Wind July 29, 2013 20 / 22

SLIDE 38

Conclusions

There is little evidence of economies of scale or learning curves in the

ffshore wind market.

Greg Upton (LSU) Offshore Wind July 29, 2013 21 / 22

SLIDE 39

Conclusions

There is little evidence of economies of scale or learning curves in the

ffshore wind market.

These results are robust when wind farm characteristics as well as country fixed effects are included.

Greg Upton (LSU) Offshore Wind July 29, 2013 21 / 22

SLIDE 40

Conclusions

There is little evidence of economies of scale or learning curves in the

ffshore wind market.

These results are robust when wind farm characteristics as well as country fixed effects are included. This result can potentially have substantial policy implications for the

ffshore wind market

Greg Upton (LSU) Offshore Wind July 29, 2013 21 / 22

SLIDE 41

Conclusions

There is little evidence of economies of scale or learning curves in the

ffshore wind market.

These results are robust when wind farm characteristics as well as country fixed effects are included. This result can potentially have substantial policy implications for the

ffshore wind market

There is no evidence to suggest that investing in more offshore wind will lead to a decrease in costs of future projects.

Greg Upton (LSU) Offshore Wind July 29, 2013 21 / 22

SLIDE 42

Conclusions

There is little evidence of economies of scale or learning curves in the

ffshore wind market.

These results are robust when wind farm characteristics as well as country fixed effects are included. This result can potentially have substantial policy implications for the

ffshore wind market

There is no evidence to suggest that investing in more offshore wind will lead to a decrease in costs of future projects. Due to the absence of evidence of economies of scale, larger projects do not appear to have lower average costs than smaller projects. For this reason we do not find evidence that increasing capacity on the margin, has made projects more economically viable.

Greg Upton (LSU) Offshore Wind July 29, 2013 21 / 22

SLIDE 43

The End

Thank You!

Greg Upton (LSU) Offshore Wind July 29, 2013 22 / 22

Economies of Scale, Learning Curves and Offshore Wind Development Costs

Presentation by Greg Upton David Dismukes and Gregory Upton Jr. Louisiana State University

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark.

Introduction

The worlds first offshore windfarm (OSW), Vindeby, was completed in 1991 in Ravnsborg, Denmark. Since 1991, over 40 additional OSWs have been constructed, none of which are located in the United States.

Introduction

Introduction

Relatively high cost of offshore wind compared to other forms of energy.

Difficulty in receiving permitting.

Introduction

These two issues are interrelated, though, as relatively expensive projects are less likely to receive approval than relatively less expensive projects.

Introduction

Different areas have different physical characteristics, and these heterogenous conditions can have a potentially large impact on costs.

Economic environments in which existing OSWs were built are also heterogeneous.

Introduction

Different areas have different physical characteristics, and these heterogenous conditions can have a potentially large impact on costs.

Economic environments in which existing OSWs were built are also heterogeneous.

The forty-one OSWs examined in this study were built in eight different countries over a twenty one-year period.

Introduction

This paper will combine three different literatures:

Economies of Scale

Economies of scale in electric power generation specifically have been studied extensively both in the United States (Christensen & Greene 1976) as well in other countries around the world (Filippini 1996, Franquelli, Piacenza, & Vannoni 2004).

Economies of Scale

Economies of Scale

Economies of Scale

turbines, there is approximately a 30-percent reduction in the list price.

Economies of Scale

turbines, there is approximately a 30-percent reduction in the list price.

and total capacity, but do not empirically test for the presence of economies of scale in their specification.

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

conducted.

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

conducted. Junginger & Turkenburg (2005) empirically analyze experience curves in wind farms worldwide and find significant learning effects, but limit their analysis to on-shore wind.

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

conducted. Junginger & Turkenburg (2005) empirically analyze experience curves in wind farms worldwide and find significant learning effects, but limit their analysis to on-shore wind. USDOE (2011), on the other hand, finds no evidence of a learning curve in on-shore wind in the United States.

Learning Curves

Learning curves have been empirically estimated for electricity generation (Jamsab 2007; Zimmerman 1982), but very little research

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

exchange rates, interest rates, and inflation rates over time.

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

exchange rates, interest rates, and inflation rates over time.

Similar cost-comparison problems have arisen when analyzing nuclear plants (Ellis & Zimmerman 1980; Marshall & Navarro 1991).

Overnight Costs

In order to test for economies of scale and learning by doing, it is imperative to get an “apples-to-apples” comparison of costs.

exchange rates, interest rates, and inflation rates over time.

Similar cost-comparison problems have arisen when analyzing nuclear plants (Ellis & Zimmerman 1980; Marshall & Navarro 1991). We will therefore borrow the methodology used in this literature in

Overnight Costs

Step 1: Cumulative Percentt =

t − si fi − si × π 2 αβ (1) α is assumed to be 4.082 and β is 3.25, which is consistent with previous literature (EIA 1998) Step 2: Next, we calculate the percent of the expenditures incurred in each month

derivative of (1) with respect to time, or more simply, the discrete difference between Cumulative Percentt and Cumulative Percentt−1. Percent Expenditurei,t = Cumulative Percenti,t − Cumulative Percenti,t−1 (2)

Overnight Costs

Step 3: Discount Factori,t = Percent Costi,t(1 + ri,t)(1 + inflationi,t) (3) Where ri,t is the interest rate of country i in time t and inflationi,t is the inflation rate in country i in time t. Finally, the overnight cost is calculated as follows. Step 4: Overnight Costi = Total Costi T

t=1 Discount factori,t

(4)

Empirical Specification

The following empirical specification will be used to test for economies of scale: ln(Overnight Costi) = α + β ln Capacityi + X

i δk + εi

(5) Overnight Cost is in 2012 U.S. Dollars. Capacity is the total capacity in MW of windfarm i. ˆ β is the estimated elasticity of cost with respect to capacity. If β is estimated to be less than 1, then we will have evidence of economies of scale in off-shore windfarms. X

i is a list of covariates that include the distance to shore, water

depth and country fixed effects.

Empirical Specification

The following empirical specification will be used to test for learning curves: ln(Overnight Costi) = α+γ ln Cumulative Capacityi+β ln Capacityi+X

i δk+εi

(6) ˆ γ is the estimated percent change in overnight cost associated with respect to a one percent change in cumulative capacity. If learning curves are present, then we expect γ < 0. X

i includes the same list of control variables as seen above in the

economies of scale specification.

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide.

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide. Data on total cost, capacity, water depth and distance to shore are from 4COffshore.

Data

The analysis considers 41 offshore wind farms located in eight different countries worldwide. Data on total cost, capacity, water depth and distance to shore are from 4COffshore. Overnight cost calculations are completed with the overnightcost Stata program that accompanies this paper.

Data

Data