Does China’s Renewable Energy Subsidy Leads to Emission Reduction?
Chi Gao
Introduction
Motivation: what policy options works?
Decisionmakers have been exploring policy options to accelerate the pace of decarbonization
There are three broad categories of policy options:
Command & Control Policy (C&C): forced closure of coal fired power plants
Market-based policy: Cap and Trade
Industrial Policy: Cheap loan for renewable energy generation
Literature: market-based policies is the way to go, or is it?
In the past, economists established a firm theoretical foundation in support of market-based policies.
For instance, in a cap and trade emission market, theoretically the market can find out the most cost-effective way to limit emission
However, recent works have suggested that C&C and industrial policies have their merits too.
This paper: explores the effect of subsidy policies in emission reduction
This research investigates China’s renewable energy(RE) feed-in tariff (FIT) as a case in point to see how effective are the subsidy policies at emission reduction
Highlight: geographic analysis at the district level
Logic chain: RE subsidy -> growing RE generation capacity -> total energy demand keeps constant, so thermal plants work less hours -> less emission
Result: RE FIT has no effects on emission reduction, but they do encourage RE generation capacity expansion
Question: RE FIT does lead to increase in RE generation capacity. What went wrong with the latter half of the logic chain?
Turns out, the “crowd out” effect is not does not come naturally. That is, at the current scale, RE energy generation has no direct local impact on fossil fuel generation
Data
FIT
Feed-in tariff - subsidize RE generation per unit of energy generated
Why use FIT? RE energy generation cost used to be higher than that of thermal plants. FIT allows RE energy to be more competitive
RE FIT is linked to the level of benchmark on-grid coal generated electricity price [标杆上网电价].
标杆上网电价
- 标杆上网电价 is set administratively for different regions. NDRC is in charge of adjusting them over time.
FIT & 标杆上网电价
- Total FIT formula for a RE project located in location \(i\):
\[ FIT_{i,t} = (P^{RE}_{i,t} - P^{coalBM}_{i,t})\times E \]
\(P^{RE}_{i,t}\) - unit price of electricity generation for plant \(i\) in year \(t\).
\(P^{coalBM}_{i,t}\) - local 煤电标杆电价 at plant \(i\)’s province
\(E\) is the amount of RE generated by this project
P_RE | Unit RE price rules
Set administratively by resource zones, supposedly reflects resource endowment
Trend: price is pressed downward, as RE gets cheaper and cheaper. Since 2021, there is no subsidy provided for new plants
P_RE | FIT price data
State Grid publishes all the subsidized RE projects served by its power grid, as required by law.
Data includes:
RE Resource type (solar vs wind)
Date at which the project is connected to the grid
RE On-grid price
RE Generation capacity
Location (at a district level)
P_coalBM | 标杆电价
- Data range: 2004 to 2017.
E | amount of energy generated
\[ E_i = cf_{i,t} \times genCap_i * 8760 \]
\(cf_{i,t}\) - capacity factor for project i at time t (what percentage of time is the project in operation?)
\(genCap_i\) - generation capacity of project i
Wrinkles
Data availability only at the provincial level.
Have to combining physical estimates and official reported \(cf\).
Now we are ready to calculate FIT!
\[ FIT_{i,t} = (P^{RE}_{i,t} - P^{CoalBM}_{i,t})\times cf_{i,t} \times genCap_i \times 8760 \]
\(FIT_{i,t}\) - amount of FIT for RE project \(i\) at year \(t\)
\(P^{RE}_{i,t}\) - on-grid price for RE project \(i\) at year \(t\)
\(cf_{i,t}\) - capacity factor for RE project \(i\) at year \(t\)
\(genCap_i\) - generation capacity for RE project \(i\)
Now we know the FIT of each RE project over the years, and their location (at a district level)
Emission Data
CO2 emission from 1997 to 2017 at a county level
Provincial level energy combustion, disaggregated via nightlight intensity share
Functional Form
\[ CO2Emission_{i,t} = \beta_1 genCap_{i,t} + \beta_2 subsidy_{i,t} + \\ \beta_3 genCap_{i,t-1} + \beta_4 subsidy_{i, t-1} + \\ \beta_5 genCap_{i,t-2} + \beta_6 subsidy_{i, t-2} + \\ \sum_i \alpha_i + \sum_t \gamma_t + \mu_{i,t} \]
Two-way fixed effects regression at the district level with lagging 3 years of lagging.
Explores the relationship between CO2 emission and the amount of subsidy and RE generation capacity 3 years prior to the current year
Results
Previous years’ FIT subsidy has a significant positive effect on generation capacity
However, generation capacity has limited impact on total emission
Discussion
Why “crowd-out” effect failed?
Amount of thermal generation is somewhat fixed due to inflexible contracting and administratively set on-grid price
RE scale is still too small to put a dent the coal industry – coal lock-in effect. Coal industry is particularly tenacious in countries where the sector is governed by State-owned enterprises by leaders who have connections within the government body.
Takeaways
Subsidy policy can foster the growth of an emerging industry, but it is too naive to think the incumbent would simply recede and let the challenger to take over
To ensure the gradual phase out of coal, perhaps more direct policies need to be in place to curb thermal plants to reduce emission