Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

International Diffusion of Renewable Energy Technologies by cio18038

VIEWS: 10 PAGES: 39

									   International Diffusion of
Renewable Energy Technologies


   Nigel Meade and Towhidul Islam
                     University of Guelph
                     Department of Marketing and Consumer Studies
                                                              1
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              




     Overview
     • Renewable Energy Technologies
           – Types of RETs considered
     • Modelling the diffusion of innovations
           –   Diffusion
           –   Bass model
           –   Extensions to international model
           –   Model construction for RET diffusion
     • Data
           – Scale of use of fossil fuels in electricity generation
           – Incentives
     • Analysis
           – Hypotheses
           – Estimation
           – Forecasting
     • Conclusions


 Nigel Meade and Towhidul Islam                                                            2
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              




     Renewable Energy Technologies (RETs)
     • The RETs we consider are:
           –Wind turbines
           –Biomass (burnt instead of fossil fuel)
           –Solar energy via photo voltaic cells
     • We do not consider:
           –Marine energy (tidal and wave)
           –Hydro-electric
           –Carbon capture
           –Nuclear energy

 Nigel Meade and Towhidul Islam                                                            3
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



 Wind power
 • most successful of RETs
 • particularly in Denmark
   and Germany
 • „unit‟ of wind power
   generation is the turbine
 • capacity depends on site
   and size
 • current UK installations
   average around 2 to 3
   megawatts




 Nigel Meade and Towhidul Islam                                                            4
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



 Biomass
 • Longest established of
   RETs
 • Classes of biomass
    –Virgin wood
    –Energy crops
    –Agricultural residues
    –Industrial waste and co-
    products
 • Burnt alone or co-fired
   with fossil fuels



 Nigel Meade and Towhidul Islam                                                            5
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



 Photo-volatic cells (Solar
   energy)
 • Least established of RETs
 • Basic unit is small
 • Industrial use via „farms‟




 Nigel Meade and Towhidul Islam                                                            6
International Diffusion of Renewable Energy
Overview    Renewable Energy Technologies     Modelling Diffusion       Data      Analysis      Conclusion
                                                        




     Renewable Energy Technologies
     • not regarded as a better technology
     • established electricity suppliers prefer large systems
     • new entrants to liberalised markets prefer gas turbine
            – less capital intensive
            – can generate electricity continuously.
     •     wind farms by contrast
            – are capital intensive
            – produce output according to the vagaries of the weather
     •     benefits of RETs
            – low carbon dioxide emissions
            – low fuel costs (wind and PV)
     •     currently energy prices do not reflect the environmental damage done by
           fossil fuels
            –“environmental externalities” should be explicitly incorporated into the electricity tariff
            (Owen, 2006)
            – would lead to RETs becoming financially competitive
            – incentive schemes discussed later




 Nigel Meade and Towhidul Islam                                                                            7
International Diffusion of Renewable Energy
Overview     Renewable Energy Technologies    Modelling Diffusion         Data       Analysis       Conclusion
                                                           


    Incentive Schemes:
    Payment System Advantages                                       Disadvantages
    (Price based)           Fixed prices encourage:                 - less incentive to reduce costs
    Feed in tariff          - Investment                            - difficult for governments to control
    FIT                     - R & D leads to increased surplus      costs and capacity
    Tax incentives          - stimulate investment                  - needs high equity stake
                                                                    - increases finance costs
    (Quota based)           - incentive to lower costs              - limited reward for technical change
    Renewables              - easier for governments to control     - producer may buy in innovative
    Portfolio Standard      installed capacity                      technology
    RPS                     - encourages use of least costly        - may discourage fledgling
    a certificate trading   sources                                 technology
    scheme                  - futures market would help project     - volatility or illiquidity in certificate
                            planning                                price may discourage investors




 Nigel Meade and Towhidul Islam                                                                              8
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              




                                                 1

    A population of
    potential Adopters
    •Each potential adopter is
    subject to a pressure to
    adopt
    •The pressure may
    increase as others adopt
    Key:
    Non-adopter
    Adopter                                      0
                                                     0                                     1




 Nigel Meade and Towhidul Islam                                                            9
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion        Data    Analysis    Conclusion
                                                    



 • Idealised view of the adoption of a new technology


    Cumulative
     Adoption




    Adoption
    per period



              0       10      20     30       40     50          60   70      80     90    100
                                                    Time



 Nigel Meade and Towhidul Islam                                                                  10
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



 A selection of models of innovation diffusion
  Models for cumulative adoption
    •   Bass model:                        Bass (1969)
    •   Cumulative lognormal:              Bain (1963)
    •   Cumulative normal:                 Rogers (1962)
    •   Gompertz:                          Gregg, Hassel and Richardson (1964)
    •   Logistic:                          Gregg, Hassel and Richardson (1964).
    •   Log reciprocal:                    McCarthy and Ryan (1976)
    •   Modified exponential:              Fourt and Woodlock (1960)
    •   Weibull:                           Sharif and Islam (1980)
    Linearised trend and non-linear autoregressive models
    •   Floyd:                             Floyd (1962)
    •   Harvey:                            Harvey (1984)
    •   KKKI:                              Kumar and Kumar (1992)
    •   Sharif and Kabir:                  Sharif and Kabir (1976)
    •   SBB:                               Sharma, Basu and Bhargava (1993)

 Nigel Meade and Towhidul Islam                                                            11
International Diffusion of Renewable Energy
Overview    Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                               



 The Bass model
    • Adopters driven by
        – Desire to innovate
              • p is coefficient of innovation (external influence)
        – Desire to imitate
            • q is coefficient of imitation (internal influence)
    •   m potential adopters
    •   Number adopted by time t: Yt
    •   Probability of a person adopting at time t: p + q (Yt-1/m)
    •   Expected number of adopters at time t:
        –      m(1 – (Yt-1/m) ) (p + q (Yt-1/m) )




 Nigel Meade and Towhidul Islam                                                             12
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              

 Example using the Bass model
 • Adoption of steam and motor powered merchant ships in the US




    Wind power succumbs
    to fossil fuels




 Nigel Meade and Towhidul Islam                                                            13
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies                      Modelling Diffusion             Data          Analysis          Conclusion
                                                                                 

 • Adoption of steam and motor powered merchant ships in the US
                                  1
              Cumulative
                                        Proportion of Wind
              Adoption (Y)              Powered Ships




                                            Proportion of Self
                                            Powered Ships



                                  0
                                  1800                 1820           1840     1860    1880      1900         1920     1940

                                      0.1
               Adoption                          Annual Changes in
                                                 Proportion of Self
             per period (y)                      Powered Ships



    Bass type:                    0.05




    y = (p + qY)(m – Y)
    p: external influence              0
                                        1800              1820          1840    1860   1880          1900     1920     1940
    q: internal influence
    m: proportion of
        potential achieved        -0.05




 Nigel Meade and Towhidul Islam                                                                                                      14
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies                      Modelling Diffusion            Data          Analysis              Conclusion
                                                                                

 • Adoption of steam and motor powered merchant ships in the US
                                  1

              Cumulative
                                        Proportion of Wind
              Adoption (Y)              Powered Ships




    Forecast Origin 1880                                                                                    Forecast
                                                                                                            Origin 1880
    p = 0.0034
    q = 0.0186                              Proportion of Self
                                            Powered Ships


    m = 1.0000                    0
                                  1800                 1820           1840    1860    1880      1900         1920         1940


                                      0.1
               Adoption                          Annual Changes in
                                                 Proportion of Self
             per period (y)                      Powered Ships


                                  0.05




    p: external influence              0
                                        1800              1820         1840    1860   1880          1900     1920         1940
    q: internal influence
    m: proportion of
        potential achieved        -0.05




 Nigel Meade and Towhidul Islam                                                                                                         15
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies                      Modelling Diffusion            Data          Analysis              Conclusion
                                                                                

 • Adoption of steam and motor powered merchant ships in the US
                                  1

              Cumulative
                                        Proportion of Wind
              Adoption (Y)              Powered Ships




    Forecast Origin 1890                                                                                    Forecast
                                                                                                            Origin 1890
    p = 0.0024
    q = 0.0323                              Proportion of Self
                                            Powered Ships


    m = 1.0000                    0
                                  1800                 1820          1840     1860    1880      1900         1920         1940

                                      0.1
               Adoption                         Annual Changes in
                                                Proportion of Self
             per period (y)                     Powered Ships


                                  0.05




    p: external influence              0
                                        1800              1820         1840    1860   1880          1900     1920         1940
    q: internal influence
    m: proportion of
        potential achieved        -0.05




 Nigel Meade and Towhidul Islam                                                                                                         16
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies                      Modelling Diffusion           Data          Analysis              Conclusion
                                                                               

 • Adoption of steam and motor powered merchant ships in the US
                                  1

              Cumulative
                                        Proportion of Wind
              Adoption (Y)              Powered Ships




    Forecast Origin 1920                                                                                   Forecast
                                                                                                           Origin 1920
    p = 0.0005
    q = 0.0518                              Proportion of Self
                                            Powered Ships

    m = 1.0000                    0
                                  1800                 1820          1840    1860    1880      1900         1920         1940

                                      0.1
               Adoption                         Annual Changes in
                                                Proportion of Self
             per period (y)                     Powered Ships


                                  0.05




    p: external influence              0
                                        1800              1820        1840    1860   1880          1900     1920         1940
    q: internal influence
    m: proportion of
        potential achieved        -0.05




 Nigel Meade and Towhidul Islam                                                                                                        17
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              




     Stylised facts in forecasting diffusion
     • Forecasting accuracy tends to improve as more data
       become available
     • Forecasting using an origin before the point of inflection
       is fraught with danger
     • Estimates of
           –p (external influence) increase
           –q (internal influence) decrease
           as more data is used for estimation (according to Van den Bulte
           and Lilien,1997)
     • The opposite occurs in this example

 Nigel Meade and Towhidul Islam                                                            18
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



    Developments in diffusion modelling


    • Introduction of marketing variables into the parameterisation of the
      models
    • Robinson and Lakhani (1975)



    • Multinational models: considering innovations at different stages of
      diffusions in different countries
    • Gatignon et al (1989)




 Nigel Meade and Towhidul Islam                                                            19
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion        Data    Analysis   Conclusion
                                                   


    Modelling approach
    Standardise data to electricity generated per 1m of population in 2001
 additional electricity generated                    estimated electricity generated
 - by technology i                                   by installed capacity
 - in country c                                      - at time t-1
 - at time t

                                       ˆ                 ˆ
                       yict  pic  qicYict 1 mic Pic  Yict 1   ict     
 Bass model coefficients
                                                                 maximum theoretical
 External influence:                                             potential
 constant pressure to
 increase RET generation                                                             noise term

 Internal influence:
 pressure proportional
                                       proportion of
 to the installed capacity
                                       potential achieved
 Nigel Meade and Towhidul Islam                                                                   20
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion        Data    Analysis   Conclusion
                                                   


    Modelling approach
    Estimation is by Maximum Likelihood


           error terms are assumed to be Gaussian




                                                               
                                                                  
                                         ˆ                           
                     ict   ~ N  0, max Yict 1 ,1                 
                                                                     

      Estimated cumulative generation may be very small

                                           Heteroscedastic data
                                           - variance associated with amounts generated
                                           - estimation of the parameter eta


 Nigel Meade and Towhidul Islam                                                                 21
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies    Modelling Diffusion        Data       Analysis      Conclusion
                                                       


    Modelling approach
    Parameterise the Bass coefficients

 constant pressure to                           Hypothesis 1
 increase RET generation                        Adoption driven by incentive schemes

                                                                        4                               
                                  p 'ic  exp   0  1 w   2 PV    2 j c , INC  j    7Tic 
                                                                       j 1                             
                        constant
                                                                         Incentive
                        wind dummy                                       dummies
                        PV dummy
                                                                  thermal as proportion of
                                                                  total (only appears in p)
    q and m parameterised similarly

 Nigel Meade and Towhidul Islam                                                                          22
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              


    Modelling approach
    Parameterise the Bass coefficients

 constant pressure to                      Hypothesis 2
 increase RET generation                   Adoption driven by national characteristics


                                  p 'ic  exp  c  1 w   2 PV 

                        constant for
                        each nation

                        wind dummy

                        PV dummy

    q and m parameterised similarly

 Nigel Meade and Towhidul Islam                                                            23
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              


    Modelling approach
    Do not parameterise the Bass coefficients

 constant pressure to                      Hypothesis 3
 increase RET generation                   Adoption driven by common characteristics


                                              p 'ic  exp  

                        constant




    q and m parameterised similarly

 Nigel Meade and Towhidul Islam                                                            24
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              




     Adoption of renewable energy technology by
     power generators in the EU

     • Focus on 14 EU countries
     • Adoption influenced by a range of incentive schemes
     • Renewable energies considered are:
           – Wind, Biomass and Photo-voltaic cells
     • Focus is due to
           – Power generation data available from Eurostat
           – Incentive scheme details from Haas et al (2004)
           – Maximum potential per technology per country from de Noord,
             Beurskens and de Vries (2004)



 Nigel Meade and Towhidul Islam                                                            25
 International Diffusion of Renewable Energy
 Overview        Renewable Energy Technologies           Modelling Diffusion                Data        Analysis        Conclusion
                                                                             


                          Who uses which incentive scheme

                                                                                       Sweden            Finland




                                                                          Nether-             Denmark
                                                                           lands
                           Ireland          United
                                           Kingdom

                                                                                    Germany
                                                                  Belgium


                                                                                                   Austria


                                Portugal         Spain           France             Italy
                                                                                                                   Greece
RPS
RPS + FIT
FIT
Subsidy
Subsidy + Tax Credit




  Nigel Meade and Towhidul Islam                                                                                               26
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data     Analysis   Conclusion
                                               




    An indication of the pressure on countries to reduce fossil fuel consumption



                                  Proportion of                         Proportion of
                                  Thermal                               Thermal
           Country                electricity      Country              electricity
           Sweden                       0.09       Portugal                   0.70
           France                       0.10       United Kingdom             0.78
           Austria                      0.37       Italy                      0.83
           Belgium                      0.43       Greece                     0.86
           Finland                      0.57       Denmark                    0.86
           Spain                        0.62       Ireland                    0.90
           Germany                      0.64       Netherlands                0.94



 Nigel Meade and Towhidul Islam                                                             27
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies           Modelling Diffusion          Data    Analysis          Conclusion
                                                             


                               The beginning of an S shaped „diffusion‟ curve?


                       35000


                       30000                                   Wind is dominant
                                                               renewable technology
                       25000

                                  Germany
           Power GWH




                       20000


                       15000


                       10000
                               Biomass is longer
                       5000
                               established

                           0
                           1990     1992    1994   1996      1998        2000    2002   2004       2006

                                                            Photo-voltaics still at an early stage

 Nigel Meade and Towhidul Islam                                                                                  28
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies                          Modelling Diffusion           Data          Analysis    Conclusion
                                                                                   


                                   S shaped curve – some way short of an inflection point

                               700000
                               40000
                               180000

                               160000
                               35000
                               600000
                               140000       Net Electricity Maximum Potential
                                            TheoreticalGenerated
                               30000
                               500000
              Wind Power GWH




                               120000
                               25000
                Power GWH




                               400000
                               100000
                               20000
                                80000
                               300000
                               15000
                                60000
                               200000      Theoretical Maximum Potential
                               10000
                                40000

                               100000
                                            Germany
                                            Germany
                                5000
                                20000                                              Electricity Generated by Wind

                                   00
                                    1990
                                   1990          1992
                                                1992       1994
                                                          1994       1996
                                                                     1996      1998
                                                                               1998      2000
                                                                                         2000          2002      2004    2006


                                           Growth of Power generated by Wind Turbines

 Nigel Meade and Towhidul Islam                                                                                                  29
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies     Modelling Diffusion          Data          Analysis   Conclusion
                                                             



    Parameter Estimation

    Hypothesis 1
    Adoption driven by incentive schemes: p values

                        Variable                          p          q             m

                        Wind dummy                    0.08         0.00          0.21

                        PV dummy                      0.07         0.00          0.15

                        RPS                           0.03         0.06          0.47

                        FIT                           0.43         0.69          0.77

                        Subsidy                       0.72         0.43          0.06

                        Tax credit                    0.28         0.00          0.65

                        Proportion thermal            0.13

                       Log likelihood Function = -1703

 Nigel Meade and Towhidul Islam                                                                           30
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



    Hypothesis 1
    Adoption driven by
    incentive schemes:
    plot of countries in
    p q space (wind)




 Nigel Meade and Towhidul Islam                                                            31
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



    Hypothesis 2                           Likelihood Function = -1679
    Adoption driven by
    national
    characteristics:
    plot of countries in
    p q space (wind)




 Nigel Meade and Towhidul Islam                                                            32
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data      Analysis   Conclusion
                                                



    Comparison of Hypotheses 1 and 2
    Adoption (of Wind) driven by country characteristics
    versus incentive schemes
                     Euclidean distance in (p, q, m) space



               Austria              0.62   Denmark                      0.07
               France               0.45   UnitedKingdom                0.06
               Greece               0.42   Belgium                      0.06
               Ireland              0.33   Netherlands                  0.05
               Spain                0.25   Finland                      0.03
               Portugal             0.20   Italy                        0.02
               Germany              0.13   Sweden                       0.01
       Captured less well                      Captured reasonably well
       by incentive scheme                     by incentive scheme


 Nigel Meade and Towhidul Islam                                                              33
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



    Hypothesis 3
    Adoption process is common to all countries
    Log likelihood Function = -1784
    p = 0.00003; q = 0.20; m = 1.0; eta = 1.38 (all p values are less than 0.1%)




                                                                  Graph shows p and q
                                                                  from Hypothesis 2




 Nigel Meade and Towhidul Islam                                                            34
International Diffusion of Renewable Energy
Overview     Renewable Energy Technologies             Modelling Diffusion           Data            Analysis      Conclusion
                                                                          


Examples of fitted models

                            Austria (Wind)                                               France (Wind)
    90                                                          25

    80

                                                                20
    70

    60
                                                                15
    50
         (2) Country model                                           (2) Country model
    40
                                                                10
         (1) Incentive model                                         (1) Incentive model
    30

    20   (3) Simple model
                                                                 5

    10

    0                                                            0
         1    2   3   4     5   6   7   8    9   10   11   12        1   2   3   4   5   6   7   8     9   10 11 12 13 14




 Nigel Meade and Towhidul Islam                                                                                             35
International Diffusion of Renewable Energy
Overview       Renewable Energy Technologies          Modelling Diffusion                 Data         Analysis       Conclusion
                                                                            


Examples of fitted models

                             Portugal (PV)                                                  Finland (Biomass)
    0.25                                                        350

                                                                      (1) Incentive model
                                                                300
     0.2                                                              (2) Country model

                                                                250


    0.15
           (1) Incentive model
                                                                200
           (2) Country model
                                                                150
     0.1


                                                                100

    0.05
                                                                50


      0                                                          0
           1     2   3   4     5   6    7    8   9   10   11          1   2   3   4   5    6   7   8   9 10 11 12 13 14 15 16




 Nigel Meade and Towhidul Islam                                                                                                 36
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion    Data     Analysis   Conclusion
                                                




    Forecasting accuracy

                                         Horizon (years)
                                                 1                          6
                             Incentive          45                         92
                rmse         Country            42                         95
                             Simple Bass        52                         60
                             Incentive         544                       1365
                mape         Country           464                       1470
                             Simple Bass       786                       1160
                             Incentive          56                         92
                mdape        Country            55                         98
                             Simple Bass        66                         82
                                                                 Log Likelihood
    Accuracy of forecasts                          Incentive                 -1703
    corresponds with accuracy of fit               Country                   -1679
                                                   Simple Bass               -1784
 Nigel Meade and Towhidul Islam                                                              37
International Diffusion of Renewable Energy
Overview      Renewable Energy Technologies                Modelling Diffusion           Data            Analysis      Conclusion
                                                                              


Examples of forecasts from the country model

                                  Finland (Biomass)                                          France (Wind))
    350                                                             30
          (2) Country model                                              (2) Country model
          Forecasts                                                      Forecasts
    300
                                                                    25


    250
                                                                    20

    200
                                                                    15
    150

                                                                    10
    100


                                                                     5
     50


     0                                                               0
          1   2   3   4   5   6    7   8   9 10 11 12 13 14 15 16        1   2   3   4   5   6   7   8     9   10 11 12 13 14


    The reason why the mape and
    mdape are so high
 Nigel Meade and Towhidul Islam                                                                                                 38
International Diffusion of Renewable Energy
Overview   Renewable Energy Technologies   Modelling Diffusion   Data    Analysis   Conclusion
                                              



    Conclusions from this study
    • Bass model can be used beneficially with data that is not
      strictly adoption data – but reflects an adoption process
    • incentive model is most interesting from a policy point of
      view
           – worth developing by adding regional variables or other „cultural‟
             variables
    • long term forecasts poor due to lack of inflection point
    • information about planned capacity would (hopefully)
      improve longer term forecasts
           – might act as a proxy for the inflection point
    More generally
    • parameterisation of the Bass model coefficients
           – useful tool for policy evaluation
           – gives this model an edge over competitors like the Gompertz
             where the coefficients are more opaque

 Nigel Meade and Towhidul Islam                                                            39

								
To top