Low Carbon Urban Infrastructure Investment in Asian Cities

RENEWABLE ENERGY INVESTMENT RISK ANALYSIS FOR LOW-CARBON CITY... 19

electricity demand, and monthly electricity price levels. There are three
models to estimate the forecasted data: a stochastic model, a regression
model, and a non-parameter model such as the multi-agent system and
advanced neural network models. In this study, estimations of these data
for periods up to 2032 are simulated using the mean-reverting process
of a stochastic process model (Huisman and Mahieu 2003 ; Shenoy and
Gorinevsky 2015 ; Daniels and Jonge 2003 ). We assume that electricity
prices and electricity demand converge at constant prices over the long
term [Eq. (2.3)].

SStte

bSt

+

= ́()- +

1

asloglog e

(2.3)

Where St+ 1 is the estimation of electricity prices and electricity use, a is the
regression rate (converges to b over a period if there is no change), b is the
regression level, σ denotes volatility, ε is the standard normal distribution,
and log is the natural logarithm.
To simulate future electricity prices via the mean-reverting process, his-
torical data are used from April 2002 to March 2013. Electricity demand
data are extracted from statistics on the city of Yokohama, and electricity
price data are extracted from the Tokyo Electricity Generation Company
(TEPCO), as electricity in Yokohama is provided by this entity. Monthly
electricity prices can be influenced by future changes in basic electric-
ity fees, fuel cost adjustment system unit prices (JPY/kWh), renewable
energy promotion levy unit prices (which were in effect from November
2009 to September 2014),^8 and PV power promotion levy under excess
electricity purchasing scheme (which become in effect from August
2012).^9 Thus, electricity prices are calculated by including these fees and
by estimating the future costs of these fees. In addition, solar power gen-
eration levels depend heavily on solar insolation levels. For this reason,
this paper uses the Japan Meteorological Agency observation data for
solar insolation levels in Yokohama, and it simulates solar insolation levels
until 2030 using historical data for 2002 to 2009. Finally, we estimate
the monthly solar insolation range and average. The beta distribution of
the optimum data assumption with parameters for Yokohama is used and
has a minimum value of 1.70, a maximum value of 6.18, an alpha value of
1.68, and a beta value of 2.07. In this analysis, we also compare the risks
and returns of 42 JPY/kWh FIT fixed rate in 2012 with 38 JPY/kWh
FIT fixed rate in 2013.