时空视角下各省居民消费对经济发展水平的影响——基于动态函数系数自回归时空模型的实证分析

The Impact of Residents’ Consumption in Each Province on the Economic Development Level from the Perspective of Time and Space—Empirical Analysis Based on the Autoregressive Spatio-Temporal Model of Dynamic Function Coefficients

摘要: 国内生产总值作为衡量经济发展水平的通用指标，消费是其核心核算要素，因此确定消费和国内生产总值存在的关联机制，是分析经济发展的重要依据。本文借助动态函数系数时空自回归(DyFAST)模型对各省居民消费与国内生产总值这两个经济指标进行分析，计算结果表明，与仅考虑时间维度或空间维度的传统线性模型相比，采用空间权重矩阵融合策略的DyFAST模型具有更小的均方预测误差(MSPE)。此外模型分析结果表明；居民消费通过重新构建区域竞争对时空滞后因子的响应模式，实现对经济发展的影响，这为中国经济发展模式向内循环转型提供实证基础。

Abstract: Gross Domestic Product (GDP) as a common indicator for measuring the level of economic development, consumption is a core element in the accounting of gross domestic product. Therefore, identifying the correlation mechanism between consumption and economic development level is an important basis for adjusting economic development. This paper conducts an empirical analysis through dynamic varying-coefficient spatiotemporal modeling (DyFAST) and describes the spatial proximity effect of regional economies through two spatial weight matrix setting schemes. The case study shows that compared with the traditional linear models that only consider the time dimension or the spatial dimension, the DyFAST model with the fusion strategy shows the smallest prediction deviation value. In addition, the model analysis results indicate that residents’ consumption affects economic development by reconstructing the response mode of regional competition to spatiotemporal lag factors, which provides an empirical basis for the transformation of China’s economic development model to internal circulation.