How can climate adaptation policy enhance urban climate resilience? Empirical evidence from China

PLOS One | https://doi.org/10.1371/journal.pone.0335736 November 6, 2025 7 / 21 ( ˆτ sdid, ˆμ, ˆα, ˆβ ) = arg min τ ,μ,α,β { N ∑ i=1 T ∑ t=1 (Yit – μ – αi – βt – W itτ )2 ˆωsdid i ˆλsdid t } (2) where ˆτ sdid denote the average treatment effect estimates from the Synthetic DID method. N and t denote the units and time of the balanced panel data. Treatment variables are denoted as Wit ∈ {0, 1}. αi and βt denote individual fixed effects and time-fixed effects. μ is a random error term. ˆωsdid i and ˆλsdid t denote individual weights and time weights, respectively. 3.2.2. Mediating effect model. To examine the mediating mechanism role of green bonds in the implementation of the CACCP policy, this paper constructs a mediating effect model based on the model and the Sobel test methods proposed by Wen et al. (2022) [51], as shown in Eqs (3) - (5). Eq (3) represents the total effect of the CACCP policy, while Eq (4) and (5) represent the CACCP policy's pathway to enhance urban climate resilience through the use of green bonds. In this model,GBdenotes green bonds and represents the total amount of municipal green bond issuance. The meanings of the other variables are consistent with those in the model (1). The stepwise regression method indicates the presence of a partial mediation effect if ˆβ1, ˆβ2, ˆβ3, and ˆδ are significant, and a full mediation effect if ˆβ3 is not significant and the others are all significant [52]. Resilienceit = α0 + β1treated it + 3 ∑ j=1 γj Xit + u i + λt + εit (3) GBit = α0 + β2treated it + 3 ∑ j=1 γj Xit + u i + λt + εit (4)

Resilienceit = α0 + β3treatedit + δGB it + 3 ∑ j=1 γj X it + u i + λt + εit (5) 3.2.3. Moderating effect model. To further test the moderating role of education, public environmental concern, and mineral resource dependence in the implementation of the policy, based on the moderation mechanism testing approach proposed by Baron and Kenny (1986) [52], this paper constructs a moderating effect model as follows: Resilienceit = α0 + β1piloti × post t × Mod + β2piloti × post t + β3piloti × Mod + β4post i × Mod + 3 ∑ j=1 γj Xit + ui + λt + εit (6) In the above model, pilot denotes the group dummy variable. It is assigned the value of 1 when it is a policy pilot, other- wise, it is assigned the value of 0. post denotes the time dummy variable, and it is assigned the value of 1 after the imple- mentation of the policy, otherwise, it is assigned the value of 0. Mod denotes the moderating mechanism variables, which are education (Edu), public environmental concern (Pec), and mineral resources dependence (Mrd ) in this study. β1 is the estimated coefficient we are interested in, which indicates that the mechanism positively or negatively moderates the CACCP policy’s effect if it is statistically significant as a positive or negative value. Otherwise, it indicates that the mecha- nism does not play a moderating role. 3.3. Variable measurement and data source 3.3.1. Outcome variable. In this paper, Resilience represents the outcome variable of the CACCP policy, i.e., urban

climate resilience. Based on this study’s conceptual definition of urban climate resilience, we constructed a system of