Semi‐Parametric Generalized Additive Vector Autoregressive Models of Spatial Basis Dynamics
将半参数广义可加向量自回归模型应用于北卡罗来纳州玉米和大豆市场的基差联动分析，发现基差和价格关系呈非线性，且该模型比标准阈值向量自回归模型能揭示更多统计显著性和非线性调整模式。
Abstract An extensive line of research has examined linkages among spatially‐distinct markets. We apply semi‐parametric, generalized additive vector autoregressive models to a consideration of basis linkages among North Carolina corn and soybean markets. An extensive suite of linearity tests suggests that basis and price relationships are nonlinear. Marginal effects, transmission elasticities, and generalized impulse responses are utilized to describe patterns of adjustment among markets. The semi‐parametric models are compared to standard threshold vector autoregressive models and are found to reveal more statistical significance and substantially more nonlinearity in basis adjustments. Marginal effects are nonlinear and impulse responses suggest greater adjustments to extreme shocks and asymmetric adjustment patterns. The results provide evidence in favor of efficiently linked markets.