Can a Policy Mix Achieve a Collaborative Effect? Exploring the Nested Implementation Process of Urban Carbon Emission Reduction Policies
Abstract
:1. Introduction
2. Literature Review
3. Policy Background and Hypotheses
3.1. LCP and CCEEFP Policies
3.2. Main Regressions Hypotheses
3.2.1. The LCP Policy and Carbon Emission Reduction
3.2.2. The CCEEFP Policy and Carbon Emission Reduction
3.2.3. The Nested Pilot of the LCP and CCEEFP and Carbon Emission Reduction Performance
4. Methods and Data
4.1. Methodology
4.2. Variables
4.2.1. Dependent Variable: Indictors of Carbon Emissions and Pollutant Emissions
4.2.2. Control Variables
4.3. Data Collection
5. Empirical Results
5.1. Baseline Model Results
5.2. Robustness Test
5.2.1. Parallel Trend Test
5.2.2. Placebo Test
5.2.3. Other Robustness Tests
5.3. Heterogeneity Test
5.4. Influencing Mechanism and Evolution Path
5.4.1. The Carbon Emission Reduction Mechanism Analysis
5.4.2. The Nested Evolution Path
6. Conclusions and Policy Implications
6.1. Conclusions
6.2. Policy Implications
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Qian, Y.; Chakraborty, T.C.; Li, J.; Li, D.; He, C.; Sarangi, C.; Leung, L.R. Urbanization impact on regional climate and extreme weather: Current understanding, uncertainties, and future research directions. Adv. Atmos. Sci. 2022, 39, 819–860. [Google Scholar] [CrossRef] [PubMed]
- Chen, X.; Di, Q.; Jia, W.; Hou, Z. Spatial correlation network of pollution and carbon emission reductions coupled with high-quality economic development in three Chinese urban agglomerations. Sustain. Cities Soc. 2023, 94, 104552. [Google Scholar] [CrossRef]
- Huo, H.; Zhang, Q.; Wang, M.Q.; Streets, D.G.; He, K. Environmental implication of electric vehicles in China. Environ. Sci. Technol. 2010, 44, 4856–4861. [Google Scholar] [CrossRef]
- Wang, K.; Zheng, L.J.; Zhang, J.Z.; Yao, H. The impact of promoting new energy vehicles on carbon intensity: Causal evidence from China. Energy Econ. 2022, 114, 106255. [Google Scholar] [CrossRef]
- Banister, D.; Anderton, K.; Bonilla, D.; Givoni, M.; Schwanen, T. Transportation and the environment. Annu. Rev. Environ. Resour. 2011, 36, 247–270. [Google Scholar] [CrossRef]
- Van der Heijden, J. Experimentation in policy design: Insights from the building sector. Policy Sci. 2014, 47, 249–266. [Google Scholar] [CrossRef]
- Crespi, F.; Ghisetti, C.; Quatraro, F. Environmental and innovation policies for the evolution of green technologies: A survey and a test. Eurasian Bus. Rev. 2015, 5, 343–370. [Google Scholar] [CrossRef]
- Ossenbrink, J.; Finnsson, S.; Bening, C.R.; Hoffmann, V.H. Delineating policy mixes: Contrasting top-down and bottom-up approaches to the case of energy-storage policy in California. Res. Policy 2019, 48, 103582. [Google Scholar] [CrossRef]
- Greco, M.; Germani, F.; Grimaldi, M.; Radicic, D. Policy mix or policy mess? Effects of cross-instrumental policy mix on eco-innovation in German firms. Technovation 2022, 117, 102194. [Google Scholar] [CrossRef]
- Heilmann, S. From local experiments to national policy: The origins of China’s distinctive policy process. China J. 2008, 59, 1–30. [Google Scholar] [CrossRef]
- Eckert, S.; Börzel, T.A. Experimentalist governance: An introduction. Regul. Gov. 2012, 6, 371–377. [Google Scholar] [CrossRef]
- Cheng, C.; Li, C. Laboratories of democracy: Policy experimentation under decentralization. Am. Econ. J. Microecon. 2019, 11, 125–154. [Google Scholar] [CrossRef]
- Eisenhardt, K.M.; Graebner, M.E. Theory building from cases: Opportunities and challenges. Acad. Manag. J. 2007, 50, 25–32. [Google Scholar] [CrossRef]
- Sabel, C.F.; Zeitlin, J. Learning from difference: The new architecture of experimentalist governance in the EU. Eur. Law J. 2008, 14, 271–327. [Google Scholar] [CrossRef]
- McFadgen, B.; Huitema, D. Experimentation at the interface of science and policy: A multi-case analysis of how policy experiments influence political decision-makers. Policy Sci. 2018, 51, 161–187. [Google Scholar] [CrossRef] [PubMed]
- Pathak, M.; Shukla, P.R. Co-benefits of low carbon passenger transport actions in Indian cities: Case study of Ahmedabad. Transp. Res. Part D Transp. Environ. 2016, 44, 303–316. [Google Scholar] [CrossRef]
- Heilmann, S. Policy experimentation in China’s economic rise. Stud. Comp. Int. Dev. 2008, 43, 1–26. [Google Scholar] [CrossRef]
- Montinola, G.; Qian, Y.; Weingast, B.R. Federalism, Chinese style: The political basis for economic success in China. World Politics 1995, 48, 50–81. [Google Scholar] [CrossRef]
- Lau, L.J.; Qian, Y.; Roland, G. Reform without losers: An interpretation of China’s dual-track approach to transition. J. Political Econ. 2000, 108, 120–143. [Google Scholar] [CrossRef]
- Heilmann, S.; Shih, L.; Hofem, A. National planning and local technology zones: Experimental governance in China’s Torch Programme. China Q. 2013, 216, 896–919. [Google Scholar] [CrossRef]
- Eckerberg, K. Implementation of environmental protection in Swedish forestry: A policy perspective. For. Ecol. Manag. 1986, 17, 61–72. [Google Scholar] [CrossRef]
- Huque, A.S.; Watton, N. Federalism and the implementation of environmental policy: Changing trends in Canada and the United States. Public Organ. Rev. 2010, 10, 71–88. [Google Scholar] [CrossRef]
- Grossman, G.M.; Krueger, A.B. Economic growth and the environment. Q. J. Econ. 1995, 110, 353–377. [Google Scholar] [CrossRef]
- Zhao, Y.; Zhang, X.; Wang, Y. Evaluating the effects of campaign-style environmental governance: Evidence from Environmental Protection Interview in China. Environ. Sci. Pollut. Res. 2020, 27, 28333–28347. [Google Scholar] [CrossRef]
- Zhang, P.; Zhou, D.; Guo, J. Policy complementary or policy crowding-out? Effects of cross-instrumental policy mix on green innovation in China. Technol. Forecast. Soc. Change 2023, 192, 122530. [Google Scholar] [CrossRef]
- Li, X.; Yao, X. Can energy supply-side and demand-side policies for energy saving and emission reduction be synergistic?—A simulated study on China’s coal capacity cut and carbon tax. Energy Policy 2020, 138, 111232. [Google Scholar] [CrossRef]
- Dhanshyam, M.; Srivastava, S.K. Effective policy mix for plastic waste mitigation in India using System Dynamics. Resour. Conserv. Recycl. 2021, 168, 105455. [Google Scholar] [CrossRef]
- Costantini, V.; Crespi, F.; Palma, A. Characterizing the policy mix and its impact on eco-innovation: A patent analysis of energy-efficient technologies. Res. Policy 2017, 46, 799–819. [Google Scholar] [CrossRef]
- Zheng, J.; Shao, X.; Liu, W.; Kong, J.; Zuo, G. The impact of the pilot program on industrial structure upgrading in low-carbon cities. J. Clean. Prod. 2021, 290, 125868. [Google Scholar] [CrossRef]
- Zhuang, G. Policy design logic of low-carbon city pilots in China. China Popul. Resour. Environ. 2020, 30, 19–28. (In Chinese) [Google Scholar]
- Tian, S.; Sun, L.; Xu, Z.; Fan, Z. Research on the promoting function of fiscal policy for enterprise transformation and upgrading under green and low carbon development goal. Public Financ. Res. 2022, 474, 79–96. (In Chinese) [Google Scholar]
- Huang, H.; He, G.; Xiao, Y. Carbon emission reduction effect of the low-carbon city construction pilot program. Resour. Sci. 2023, 45, 1044–1058. (In Chinese) [Google Scholar] [CrossRef]
- Khanna, N.; Fridley, D.; Hong, L. China’s pilot low-carbon city initiative: A comparative assessment of national goals and local plans. Sustain. Cities Soc. 2014, 12, 110–121. [Google Scholar] [CrossRef]
- Jing, W.; Zhang, L. Environmental regulation, economic opening and China’s industrial green technology progress. Econ. Res. J. 2014, 9, 34–47. (In Chinese) [Google Scholar]
- Pu, Z.; Fu, J. Economic growth, environmental sustainability and China mayors’ promotion. J. Clean. Prod. 2018, 172, 454–465. [Google Scholar] [CrossRef]
- Chen, G.; Xu, J.; Qi, Y. Environmental (de) centralization and local environmental governance: Evidence from a natural experiment in China. China Econ. Rev. 2022, 72, 101755. [Google Scholar] [CrossRef]
- Ozusaglam, S.; Kesidou, E.; Wong, C.Y. Performance effects of complementarity between environmental management systems and environmental technologies. Int. J. Prod. Econ. 2018, 197, 112–122. [Google Scholar] [CrossRef]
- He, L.; Zhang, X. The distribution effect of urbanization: Theoretical deduction and evidence from China. Habitat Int. 2022, 123, 102544. [Google Scholar] [CrossRef]
- Yang, Z.; Gao, W.; Han, Q.; Qi, L.; Cui, Y.; Chen, Y. Digitalization and carbon emissions: How does digital city construction affect China’s carbon emission reduction? Sustain. Cities Soc. 2022, 87, 104201. [Google Scholar] [CrossRef]
- Chen, J.; Gao, M.; Cheng, S.; Hou, W.; Song, M.; Liu, X.; Liu, Y.; Shan, Y. County-level CO2 emissions and sequestration in China during 1997–2017. Sci. Data 2020, 7, 391. [Google Scholar] [CrossRef]
- Lee, K.; Leung, J.Y.; Pinedo, M.L. Coordination mechanisms with hybrid local policies. Discret. Optim. 2011, 8, 513–524. [Google Scholar] [CrossRef]
- Yu, Y.; Zhang, N. Low-carbon city pilot and carbon emission efficiency: Quasi-experimental evidence from China. Energy Econ. 2021, 96, 105125. [Google Scholar] [CrossRef]
- Jin, H.; Yang, J.; Chen, Y. Energy saving and emission reduction fiscal policy and corporate green technology innovation. Front. Psychol. 2022, 13, 1056038. [Google Scholar] [CrossRef] [PubMed]
- Laporte, A.; Windmeijer, F. Estimation of panel data models with binary indicators when treatment effects are not constant over time. Econ. Lett. 2005, 88, 389–396. [Google Scholar] [CrossRef]
- Chetty, R.; Looney, A.; Kroft, K. Salience and taxation: Theory and evidence. Am. Econ. Rev. 2009, 99, 1145–1177. [Google Scholar] [CrossRef]
- Wang, S.; Zhang, R.; Wan, L.; Chen, J. Has central government environmental protection interview improved air quality in China? Ecol. Econ. 2023, 206, 107750. [Google Scholar] [CrossRef]
- Heyman, F.; Sjöholm, F.; Tingvall, P.G. Is there really a foreign ownership wage premium? Evidence from matched employer–employee data. J. Int. Econ. 2007, 73, 355–376. [Google Scholar] [CrossRef]
- Li, L.; Zheng, Y.; Zheng, S.; Ke, H. The new smart city programme: Evaluating the effect of the internet of energy on air quality in China. Sci. Total Environ. 2020, 714, 136380. [Google Scholar] [CrossRef]
- Liu, Y.; Dong, F. How technological innovation impacts urban green economy efficiency in emerging economies: A case study of 278 Chinese cities. Resour. Conserv. Recycl. 2021, 169, 105534. [Google Scholar] [CrossRef]
- Wang, Y.; Wang, C.; Hong, J.; Nian, M. Natural conditions, administrative hierarchy and urban development in China. J. Manag. World 2015, 1, 41–50. (In Chinese) [Google Scholar]
- Cartier, C. A political economy of rank: The territorial administrative hierarchy and leadership mobility in urban China. J. Contemp. China 2016, 25, 529–546. [Google Scholar] [CrossRef]
- Wang, Y.; Zhao, Y. Is collaborative governance effective for air pollution prevention? A case study on the Yangtze river delta region of China. J. Environ. Manag. 2021, 292, 112709. [Google Scholar] [CrossRef]
- Tan, R.; He, Q.; Zhou, K.; Song, Y.; Xu, H. Administrative hierarchy, housing market inequality, and multilevel determinants: A cross-level analysis of housing prices in China. J. Hous. Built Environ. 2019, 34, 845–868. [Google Scholar] [CrossRef]
- Fan, Z.; Zhao, R. Does rule of law promote pollution control? Evidence from the establishment of the environmental court. Econ. Res. J. 2019, 54, 31–37. (In Chinese) [Google Scholar]
- Liu, X.; Li, Y.; Chen, X.; Liu, J. Evaluation of low carbon city pilot policy effect on carbon abatement in China: An empirical evidence based on time-varying DID model. Cities 2022, 123, 103582. [Google Scholar] [CrossRef]
- Zeng, S.; Jin, G.; Tan, K.; Liu, X. Can low-carbon city construction reduce carbon intensity? Empirical evidence from low-carbon city pilot policy in China. J. Environ. Manag. 2023, 332, 117363. [Google Scholar] [CrossRef]
- Xue, F.; Chen, X. The carbon emission reduction effect of green fiscal policy: Evidence from the “National Comprehensive Demonstration City of Energy Saving and Emission Reduction Fiscal Policy”. J. Financ. Econ. 2022, 48, 79–93. (In Chinese) [Google Scholar]
- Wan, A.; Huang, Y. Carbon reduction technology selection mechanisms for low-carbon cities and innovative cities. China Polpulation Resour. Environ. 2024, 34, 23–31. (In Chinese) [Google Scholar]
Pilot Policies | Policy Target | Policy Implementation | Policy Evaluation | Policy Termination |
---|---|---|---|---|
LCP | Carbon emission reduction | Weak incentive and soft constraint (publicity, communication and evaluation, capacity building) | Task decomposition; Self-assessment | Batch piloting |
CCEEFP | Carbon emission reduction; Pollutant emission reduction | Strong incentive and strict measures (financial support and top-down environmental regulation) | Refining evaluation indictors; Top-down assessment; Public supervision. | Competitive adjustment; Deduction of funds |
Variables | Category | Obs | Mean | S.D. | Min | Max |
---|---|---|---|---|---|---|
co2_den | DV | 4245 | 0.0257 | 0.017 | 0.002 | 0.127 |
co2 | 4245 | 27.304 | 24.134 | 1.845 | 230.712 | |
so2_den | 4245 | 80.777 | 172.114 | 0 | 4930.267 | |
DID_LCP | KV | 4245 | 0.218 | 0.413 | 0 | 1 |
DID_CCEEFP | 4245 | 0.0427 | 0.202 | 0 | 1 | |
DID_nested | 4245 | 0.0247 | 0.155 | 0 | 1 | |
pop_den | CV | 4245 | 430.61 | 331.529 | 4.7 | 2648.11 |
gdp | 4245 | 1585.386 | 2258.147 | 52 | 25,388.3 | |
foreign_gdp | 4245 | 29.419 | 28.539 | 0.02 | 311.13 | |
second | 4245 | 0.489 | 0.107 | 0.148 | 0.91 | |
energy_coms | 4245 | 20.618 | 15.284 | 0.366 | 94.494 | |
coal_coms | 4245 | 0.998 | 0.364 | 0.0692 | 2.345 |
Variable | Model 1 | Model 2 | Model 3 |
---|---|---|---|
DID_LCP | −0.00143 *** | ||
(−5.32) | |||
DID_CCEEFP | −0.00173 *** | ||
(−3.74) | |||
DID_nested | −0.00118 ** | ||
(−1.98) | |||
pop_den | −1.69 × 10−7 | −6.87 × 10−8 | −2.78 × 10−8 |
(−0.16) | (−0.06) | (−0.03) | |
gdp | −1.27 × 10−6 *** | −1.35 × 10−6 *** | −1.39 × 10−6 *** |
(−13.66) | (−14.89) | (−15.18) | |
foreign_gdp | −0.00003 *** | −0.00003 *** | −0.00003 *** |
(−6.58) | (−5.93) | (−5.96) | |
second | 0.0143 *** | 0.0135 *** | 0.0138 *** |
(8.92) | (8.37) | (8.56) | |
energy_coms | 0.00069 *** | 0.00071 *** | 0.00071 *** |
(54.07) | (57.87) | (58.12) | |
coal_coms | −0.00137 * | −0.00163 ** | −0.00168 ** |
(−1.93) | (−2.31) | (−2.37) | |
Constant | 0.00871 *** | 0.00879 *** | 0.00864 *** |
(6.32) | (6.36) | (6.24) | |
City FE | Yes | Yes | Yes |
Year FE | Yes | Yes | Yes |
Observations | 4245 | 4245 | 4245 |
R-squared | 0.1961 | 0.1888 | 0.1881 |
Replacing Dependent Variable | PSM-DID | Four Municipalities Excluded | |||||||
---|---|---|---|---|---|---|---|---|---|
Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | Model 8 | Model 9 | |
DID_LCP | −0.933 *** | −0.0014 ** | −0.00082 *** | ||||||
(−3.59) | (−2.09) | (−3.02) | |||||||
DID_CCEEFP | −1.28 *** | −0.00299 *** | −0.00098 ** | ||||||
(−3.04) | (−2.93) | (−2.47) | |||||||
DID_nested | −1.578 *** | −0.00254 * | −0.00297 *** | ||||||
(−2.81) | (−1.77) | (−5.65) | |||||||
Control variables | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
City FE | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Year FE | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Observations | 4245 | 4245 | 4245 | 3298 | 4057 | 4137 | 4185 | 4185 | 4185 |
R-squared | 0.6798 | 0.6927 | 0.6816 | 0.0825 | 0.1421 | 0.2439 | 0.6634 | 0.1785 | 0.4219 |
Administrative Grade | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 |
---|---|---|---|---|---|---|
Low Cities | High Cities | Low Cities | High Cities | Low Cities | High Cities | |
DID_LCP | −0.00065 ** | −0.00121 ** | ||||
(−2.21) | (−2.54) | |||||
DID_CCEEFP | −0.00097 * | −0.00297 *** | ||||
(−1.68) | (−5.65) | |||||
DID_nested | −0.000386 *** | −0.00336 *** | ||||
(−2.64) | (−5.03) | |||||
pop_den | 1.01 × 10−6 | 6.41 × 10−6 *** | −1.78 × 10−6 | 7.17 × 10−6 *** | 4.22 × 10−7 | 7.35 × 10−6 *** |
(1.38) | (5.42) | (−1.49) | (5.68) | (0.71) | (5.56) | |
gdp | −2.91 × 10−6 *** | −9.06 × 10−8 | −2.76 × 10−6 *** | −6.90 × 10−8 | 1.58 × 10−7 | −1.96 × 10−8 |
(−14.48) | (−0.70) | (−17.29) | (−0.60) | (1.04) | (−0.16) | |
foreign_gdp | −0.00004 *** | −0.00003 *** | −0.00004 *** | −0.00002 *** | −0.00003 *** | −0.00002 *** |
(−6.69) | (−4.70) | (−7.70) | (−4.44) | (−6.20) | (−4.37) | |
second | 0.0097 *** | 0.0124 ** | 0.0108 *** | 0.0131 ** | 0.00224 | 0.00953 |
(4.90) | (1.97) | (6.45) | (2.13) | (1.07) | (1.57) | |
energy_coms | 0.00069 *** | 0.00273 *** | 0.00066 *** | 0.00259 *** | 0.000314 *** | 0.00262 *** |
(28.51) | (6.98) | (49.83) | (7.60) | (9.42) | (7.79) | |
coal_coms | −0.00749 *** | 0.00545 *** | −0.00277 *** | 0.00493 *** | −0.00867 *** | 0.0054 *** |
(−7.41) | (3.56) | (8.99) | (3.28) | (−9.02) | (3.53) | |
City FE | Yes | Yes | Yes | Yes | Yes | Yes |
Year FE | Yes | Yes | Yes | Yes | Yes | Yes |
Observations | 3720 | 525 | 3720 | 525 | 3720 | 525 |
R-squared | 0.6845 | 0.3696 | 0.6784 | 0.4219 | 0.1677 | 0.4234 |
Carbon Emission Reduction Mechanism | Infrastructure | Green | Infrastructure | Green | Infrastructure | Green |
---|---|---|---|---|---|---|
Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | |
DID_LCP | 1.157 | 2.238 *** | ||||
(1.33) | (2.62) | |||||
DID_CCEEFP | 2.631 *** | 2.517 *** | ||||
(3.52) | (2.93) | |||||
DID_nested | 2.054 ** | 3.101 *** | ||||
(2.22) | (2.98) | |||||
Control variables | Yes | Yes | Yes | Yes | Yes | Yes |
City FE | Yes | Yes | Yes | Yes | Yes | Yes |
Year FE | Yes | Yes | Yes | Yes | Yes | Yes |
Observations | 4245 | 4245 | 4245 | 4245 | 4245 | 4245 |
R-squared | 0.1316 | 0.5218 | 0.1185 | 0.3071 | 0.2024 | 0.4146 |
Nested Evolution Mechanism | CO2_den | SO2_den | ||
---|---|---|---|---|
LCP First, Then CCEEFP | CCEEFP First, Then LCP | LCP First, Then CCEEFP | CCEEFP First, Then LCP | |
Model 1 | Model 2 | Model 3 | Model 4 | |
DID | −0.00396 *** | −0.00155 * | −37.154 *** | −20.858 |
(−5.88) | (−1.72) | (−5.13) | (−1.45) | |
Control variables | Yes | Yes | Yes | Yes |
City FE | Yes | Yes | Yes | Yes |
Year FE | Yes | Yes | Yes | Yes |
Observations | 900 | 315 | 900 | 315 |
R-squared | 0.6418 | 0.5574 | 0.6515 | 0.4930 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhao, Y.; Zhang, Y.; Wang, S. Can a Policy Mix Achieve a Collaborative Effect? Exploring the Nested Implementation Process of Urban Carbon Emission Reduction Policies. Sustainability 2024, 16, 6529. https://doi.org/10.3390/su16156529
Zhao Y, Zhang Y, Wang S. Can a Policy Mix Achieve a Collaborative Effect? Exploring the Nested Implementation Process of Urban Carbon Emission Reduction Policies. Sustainability. 2024; 16(15):6529. https://doi.org/10.3390/su16156529
Chicago/Turabian StyleZhao, Yihang, Yuanyuan Zhang, and Shengyu Wang. 2024. "Can a Policy Mix Achieve a Collaborative Effect? Exploring the Nested Implementation Process of Urban Carbon Emission Reduction Policies" Sustainability 16, no. 15: 6529. https://doi.org/10.3390/su16156529
APA StyleZhao, Y., Zhang, Y., & Wang, S. (2024). Can a Policy Mix Achieve a Collaborative Effect? Exploring the Nested Implementation Process of Urban Carbon Emission Reduction Policies. Sustainability, 16(15), 6529. https://doi.org/10.3390/su16156529