*4.1. Comparison of Returns under Different Quota Scales*

In order to clarify the scale of the balanced quota and the characteristics of revenue in different projects, all objects are divided into below 3.33 ha (50 mu), 3.33–6.66 ha (50–100 mu), 6.66–13.33 ha (100–200 mu), 13.33–33.33 ha (200–500 mu), and more than 33.33 ha (500 mu) for group descriptive statistics (Table 2). From the perspective of scale characteristics, the scale of the balanced quota generally does not exceed 33.33 ha, and the number of projects with 3.33, 6.66, and 13.33 ha as grouping intervals is relatively evenly distributed. From the perspective of revenue characteristics, when the scale of the balanced quota is within 33.33 ha, the revenue increases with the expansion of the quota scale while, when the scale exceeds 33.33 ha, the revenue shows a declining situation. In general, the change trend and fluctuation characteristics of the quota scale and revenue provide evidence for the existence of the optimal balanced quota scale. On this basis, the following will use the econometric model to further explore the quantitative relationship between the scale of the quota and the revenue.


**Table 2.** Comparison of the scale and revenue of the balanced quota in different projects.

#### *4.2. The Influence of Turnover Index Scale on Index Return*

Before estimating the nonlinear model, the variables such as the balanced quota scale, income and population, with large standard deviations, were logarithmically processed, for the purpose of eliminating heteroscedasticity 7. At the same time, all independent variables were multicollinearity tested by variance inflation factor method 8. The results show that the VIF values of all variables are less than 10, which means that there is no collinearity problem. Then, based on OLS, Stata 16.0 software is used to estimate the impact of the scale

of the balanced quota on revenue. In the process of estimating, considering that the sample data has mixed cross-sectional characteristics, the robust standard error regression method is used. The results are shown in Table 3, where model 1 is the estimated result without adding control variables and model 2 is the result of adding other control variables.


**Table 3.** Estimated results of the impact of the scale of project-based balanced quota on revenue.

Note: \*\*\* denotes a significance level of 10%.

According to Table 3, no matter whether control variables are added or not, when the revenue of the balanced quota is taken as the explained variable, the regression coefficients of the first and second terms of the quota scale are positive and negative, respectively, and both are significant at the level of 1%. This measurement result verifies the theoretical hypothesis of this paper, that is, there is an inverted U-shaped relationship between the scale of the balanced quota and the revenue in the project based IVDB implementation. Specifically, the inverted u-shaped relationship means that the revenue gradually increases and then decreases with the expansion of the quota scale. The gradually expanding scale makes all kinds of inputs close to the optimal combination ratio step by step, along with the increase in the revenue. When the scale and other factors of production such as labor and capital reach the optimal ratio, the maximum revenue is achieved, and the quota scale reaches the optimal size. If the scale exceeds the optimal size, the whole input will be faced with deviation from the optimal production state and this will lead to a decrease in revenue eventually. Moreover, the optimal scale of the balanced quota can be calculated from the regression coefficients of the first and second terms of the independent variable in the model estimation results. According to model 2 of Table 3, the logarithm of the appropriate scale of the balanced quota is 0.31 ha in Zhejiang province and the corresponding moderate scale is 7.19 ha. Combined with the 1907 IVDB projects completing inspection in Zhejiang, 22.44% of the approved balanced quota exceeds the appropriate scale, which means that the quota is inefficiently allocated.

Among the first group of control variables, the proportion of new construction in urban areas has a significant negative impact on the revenue of the balanced quota. Although this seems unexpected, it is actually reasonable. In this respect, our explanation is that, although the larger the scale urban new construction means a higher the demand for quota and makes it easier to increase the economic benefits theoretically, in reality the benefits are also limited by such factors as an underdeveloped economy and weak financial strength in some projects, resulting in the low transaction unit price of the balanced quota and low revenue. In the second group of control variables, fiscal revenue vs. expenditure ratio is positively correlated with the revenue of the balanced quota. The main reason for this is that a higher ratio represents a bigger surplus, meaning local governments have the financial capacity to pay for the balanced quota. Urbanization rate, GDP growth rate and proportion of the service production are negatively correlated with the revenue. A possible explanation is that the former three variables are key standards to measure the level of economic development, and the higher the level of economic development, the better the rural economic situation. Therefore, the time and cost of the demolition, transaction, resettlement and other aspects will be longer and more expensive. Further, the 1907 projects in this paper are counted according to the acceptance inspection data, which may lead to fewer projects not only being approved but also accepted in developed areas. All this can cause negative correlation. In addition, per capita disposable income, urban vs. rural, does not pass the significance level.

#### *4.3. Spatial Heterogeneity Analysis*

Due to the differences in resource endowment, economic development and implementation cycle of IVDB projects in different regions, the optimal quota scale for maximizing revenue contains distinctions. Zhejiang province is divided into the northeast and southwest region according to the urban spatial pattern of 'one bay, two cores, four poles and multiple clusters' in Zhejiang 9. OLS estimation is performed on the two sub-samples, respectively, in this study. The regression results of spatial heterogeneity are shown in Table 4.


**Table 4.** Estimated results of the spatial heterogeneity of the scale and revenue in Zhejiang IVDB projects.

Note: \*, \*\* and \*\*\* denote a significance level of 1%, 5% and 10%, respectively.

In the two sub-sample models of Table 4, the coefficient of the first term of the balanced quota scale variable is significantly positive and the secondary term is significantly negative, indicating that the scale has a significant inverted 'U' impact on the revenue of the balanced quota in different regions, which is consistent with the baseline regression results of Table 3. Furthermore, combined with the regression coefficient of core explanatory variable, the optimal scale of the balanced quota in northeast and southwest Zhejiang is 9.50 ha and 6.03 ha, respectively 10. The former scale is larger than that of the latter, which is closely related to the geomorphological factors of Hang-Jia-Hu Plain and Ning-Shao Plain distributed in the Hangzhou Bay area. In addition, there are two pivotal problems that need to be specially explained in the regression results to distinguish spatial heterogeneity. Firstly, the slope of the inverted 'U'-shaped curve in the northeastern Zhejiang is greater than that of southwestern Zhejiang, indicating that the marginal return of the balanced quota in northeastern Zhejiang is higher. This result is in line with the general rule that the more developed the region is, the higher the unit price of the quota. Secondly, the control variable of proportion of commercial and residential land has a significant positive correlation with revenue in southwest Zhejiang. This is consistent with the normal expectation that the transfer income of urban construction and the revenue of the balanced quota is positively correlated.

#### *4.4. Robustness Test*

(1) Change the selected model. Usually, different models may obtain different regression results. Considering the numerical feature that the dependent variable (R) is not less than 0, we choose the Tobit model, which can handle the tail-broken data to reregression. The estimate results are shown in Table 5 for model 3. It can be found that, regardless of the direction or significance, the influence of the scale of the balanced quota and its square term on the revenue is consistent with the results in Table 3, indicating that the inverted U-shaped relationship between the scale of the quota and the revenue is robust.


**Table 5.** Robustness test of model estimation results.

Note: \*, \*\* and \*\*\* denote a significance level of 1%, 5% and 10%, respectively.

(2) Replace the explained variable. The revenue of the balanced quota is replaced by the profit of the quota and the impact of the index scale and net income is estimated as robustness test. The profit of the quota is obtained by subtracting the actual total investment <sup>11</sup> from the revenue and the OLS regression is performed after the negative number is turned to the positive and the logarithm is processed. The results are shown in Table 5 for model 4. Similarly, the primary item of the balanced quota scale is negative and the secondary item is positive, with the significance tests of 10% and 5%, respectively, being passed to verify the existence of an appropriate scale of the project-based balanced quota.

#### **5. Discussion and Policy Implication**

#### *5.1. Problems Associated with the Scale and Revenue of the Balanced Quota*

In the context of spatial planning with tight constraints, as the incremental quota becomes less and less, the balanced quota will be more pivotal for local governments in acquiring urban construction land. At the same time, the revenue of the balanced quota provides indispensable capital for the revitalization of rural regions. The spatial relocation under the IVDB implication conforms to the development of urban-rural coordination [20]. Therefore, though the IVDB implicating process may impose on the welfare of vulnerable groups such as peasants [17], the quality of reclaimed farmland land may be poorer than the occupied land, and the hidden debt of local government may be at greater risk, we can still safely deduce that the IVDB policy will be implemented persistently in China over years and decades. Therefore, there are two pivotal issues to be further concerned combined with our theoretical analysis and empirical results on the relationship between balanced quota's scale and revenue.

The first issue is how to determine the optimal scale of the balanced quota under the project-based IVDB policy? As the 1907 projects of Zhejiang province show, there is an inverted 'U' curve relationship between the scale of the balanced quota and its revenue. In other words, the quota's scale follows the rule that 'the more is not the better'. Three perspectives need to be considered to achieve the optimal scale to maximize returns. One is the factor of natural perspective. The implementation of IVDB policy is strongly dependent on the regional location, topographic features, soil quality, irrigation conditions and other natural factors. From the perspectives of ecological benefit, requisition-compensation, balance of cultivated land and farmers' use of cultivated land, Yang, et al. (2015) analyzed the rationality of the project of IVDB in the China Mountain Area [36]. If local governments pursue land indicators excessively and promote rural demolition arbitrarily, without considering the objective limitations of natural conditions, the IVDB policy will be out of control and the overall interests will be damaged. The other is the production perspective. As the theory of Economies of Moderate Scale says, as a kind of production factors, land should be considered in relation to labor, capital and other inputs. Only when the proportion of all kinds of production factors are close to the optimal combination can the revenue of the balanced quota be maximized. Then the local governments' financial burden will be reduced and the economic incentive mechanism of the IVDB policy will be brought into full play. In addition, the market perspective is an indispensable consideration. According to some Chinese cases, building high-rise buildings can save 90% of the balanced quota to transfer to urban-supporting construction, while building 'small villas' can only transfer 60% of the total quota. In the process of the IVDB's implementation, the prefectural government also tends to save more transferable quota. Contrary to most humans' intuitive prediction, our measurement results of both the whole region and the spatial heterogeneity of Zhejiang province show that, the larger the scale of urban construction is, the lesser the quota's income. Essentially, this is a reflection of market behavior. When the quota supply exceeds the demand, the earnings will naturally decline. Therefore, rather than trying to reduce the living space of farmers to obtain a bigger urban construction quota, it is better to find ways to turn to a seller's market for the balanced quota.

The second issue is to what extent the provincial government can orient the project based IVDB policy. In China's vertical land management system, there exist three levels of government: central, provincial and prefectural [37,38]. In fact, in 2004 has clarified the division of land management powers between central and local government. That is, the power and responsibility to regulate the total amount of newly added land for construction belongs to the central government, while the power to revitalize existing land for construction belong to local government. The responsibility for protecting a rational use of land rests with the local government at all levels, with the provincial government bearing the primary responsibility. As the implication of the IVDB policy almost has no effect on incremental construction land, relevant power for the policy rests with provincial and prefectural government according to the above document. However, considering the

possibility of overuse of land resources by local government, the central government has been controlling the balanced quota before 2020. Under the new system of spatial planning, provincial government has the right to arrange the implication of the IVDB policy, in line with the tight constraints of resource utilization. The provincial government's role is shifting from that of a hub [37] to a decider, and neither the top-down [39–41] nor bottom-up [42–44] theories of the process of IVDB can explain well the administrative discretion of the role. Particularly, the provincial government and prefecture may have the possibility of engaging in collusive behaviors because of common economic interests [45]. Experience and lessons from "centralization-decentralization-recentralization" [46] may be conductive to understanding the function of the provincial government in the IVDB policy implementation.
