基于智能体的货币政策对行业非对称效应研究

发布时间：2018-11-21 07:34

【摘要】：在2011年底举行的中央经济工作会议中,政府传达了今后一个时期经济工作的基调与路线。这些宏观经济目标的实现依赖于经济政策合理与有效地制定和实施,因此中央重点提出要加强宏观经济调控,体现经济政策的灵活性、针对性和前瞻性。中国的货币政策具有特殊性。与国外的利率高度市场化的情况相比,我国货币政策的利率渠道并不通畅,因此,国家干预金融市场更多地依靠货币供应量这个指标,它同时也是我国货币政策的中介目标。此外,利率、存款准备金率、公开市场业务等货币政策工具由政府管控,它将以何种方式、何种程度影响国民经济的运行,也是我们重点关心的问题。以往常用的分析经济(货币)政策的方法,主要是系统动力学、投入产出分析、联立方程模型、可计算一般均衡模型等。它们基于建立起宏观经济系统或变量间的时间序列,或通过求解方程组的方式,得到了经济(货币)政策有效性的实证结论。相比较于这些传统方法,本文发展了智能空间中的B-4问题求解方法,提出了解耦Agent、联合问题求解方法、自主强度等概念,并运用该理论在研究货币政策领域中进行了如下一些探索： (1)建立了基于智能体的货币政策模拟系统(MPSS)。通过2007年42部门投入产出表中的消耗系数设计出生产部门的投入产出关系,以智能体的自主决策、相互协作进行仿真,设计出追求自身利益最大化的功能型智能体、对资源配置起调节作用的协调型智能体以及间接干预经济活动的控制型智能体。其中,详细刻画了对货币政策传导起重要作用的中央银行以及商业银行智能体。他们之间通过要素流、资金流以及信息流进行物质与消息的传递,实现多目标下的共同演化,这样可以得到与实际经济相符合的结果。 (2)强调了货币政策的外生性,强化了商业银行的信贷功能。存款性金融机构在充当资金池的作用时,一方面接收中央银行的政策信息,做出相应的政策反应；另一方面,对于贷款产生的预期风险与收益,科学合理地进行决策,控制各行业的再生产规模。 (3)模拟了货币政策的非对称效应,表现出行业产值及价格受该政策影响的差异特性,并能够与实际时间相对应,具有一定参考价值。依据双向经济建模方法,将宏观货币政策的作用通过信贷渠道影响微观行业,同时又从微观智能体的决策与协作结果涌现出宏观现象。由于设定了与实际时间相一致的步长,因此可以同一年的经济周期对应起来。 (4)反映了能源、电力部门的重要指标。通过经济模拟和对其中能源行业的投入产出结果分析,可以得到经济政策作用下的行业发展情况,进而提炼出能源、电力部门的产量、用电量等信息。
[Abstract]:At the Central Economic work Conference held at the end of 2011, the government conveyed the tone and route of economic work in the coming period. The realization of these macroeconomic objectives depends on the rational and effective formulation and implementation of economic policies, so the central government puts forward to strengthen macroeconomic regulation and control to reflect the flexibility, pertinence and foresight of economic policies. China's monetary policy has its particularity. Compared with the high marketization of interest rate in foreign countries, the interest rate channel of monetary policy in China is not smooth. Therefore, the national intervention in financial market depends more on the index of money supply, which is also the intermediate target of monetary policy in our country. In addition, interest rates, reserve requirements, open market operations and other monetary policy instruments are controlled by the government, and how and to what extent will they affect the operation of the national economy are also our key concerns. The commonly used methods of analyzing economic (monetary) policy are system dynamics, input-output analysis, simultaneous equation model, general equilibrium model and so on. Based on the time series of macroeconomic systems or variables, or by solving equations, the empirical conclusions on the effectiveness of economic (monetary) policies are obtained. Compared with these traditional methods, this paper develops the B-4 problem solving method in intelligent space, and puts forward the concepts of decoupling Agent, joint problem solving method, autonomous strength and so on. Using this theory in the field of monetary policy, the following explorations are carried out: (1) an agent based monetary policy simulation system, (MPSS)., is established. The input-output relation of production department is designed through the consumption coefficient of 42 departments input-output table in 2007. The functional agent is designed to maximize its own benefit by the independent decision-making of the agent and the simulation of cooperation with each other. The coordination agent which regulates the allocation of resources and the control agent that indirectly interferes with economic activities. Among them, the central bank and commercial bank agent which play an important role in monetary policy transmission are described in detail. By means of factor flow, capital flow and information flow, they can transmit material and message to achieve the co-evolution of multi-objective, so as to get the result that is consistent with the actual economy. (2) it emphasizes the externality of monetary policy and strengthens the credit function of commercial banks. When the depository financial institution acts as the fund pool, on the one hand, it receives the policy information of the central bank and makes the corresponding policy response; On the other hand, for the expected risks and benefits of loans, scientific and reasonable decisions are made to control the reproduction scale of various industries. (3) the asymmetric effect of monetary policy is simulated, which shows the different characteristics of industry output value and price influenced by the policy, and can correspond to the actual time, which has certain reference value. According to the bidirectional economic modeling method, the role of macro-monetary policy is used to influence the micro-industry through credit channels, and at the same time, the macro-phenomenon emerges from the decision-making and cooperation results of micro-agent. By setting a step that is consistent with the actual time, it is possible to match the economic cycle of the same year. (4) it reflects the important index of energy and power sector. Through the economic simulation and the analysis of the input-output results of the energy industry, we can get the development of the industry under the action of economic policy, and then extract the information of energy, power sector output, electricity consumption and so on.
【学位授予单位】：华北电力大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F822.0;F224

【参考文献】