中国高校课件下载中心 》 教学资源 》 大学文库

中南财经政法大学:《金融计量学》课程教学课件(双语讲稿)Chapter 04 Causal Inference framework 2

文档信息
资源类别:文库
文档格式:PDF
文档页数:21
文件大小:484.02KB
团购合买:点击进入团购
内容简介
中南财经政法大学:《金融计量学》课程教学课件(双语讲稿)Chapter 04 Causal Inference framework 2
刷新页面文档预览

Financial Econometrics Chapter 4.Causal Inference framework 2 Channel and Mechanism identification Jin Ling School of Finance,Zhongnan University of Economics and Law

Financial Econometrics Chapter 4. Causal Inference framework 2 Channel and Mechanism identification Jin Ling School of Finance, Zhongnan University of Economics and Law 1

Outline The Definition of Channel and Mechanism Causal Directed Acyclic Graph The Methodology for Channel and Mechanism Identification 2

• The Definition of Channel and Mechanism • Causal Directed Acyclic Graph • The Methodology for Channel and Mechanism Identification 2 Outline

The Definition of Channel and Mechanism ·The definition: Channel and mechanism depict how treatment leads to changes of outcome variables. ·作用渠道或作用机制刻画因果效应何以存在、如何存在。 Managers'education and corporate operation;Bank competition and financial constraints;Retail investors trading and market stability. Why the identification of channel or mechanism is important? Convince us that the Causal effect does not originate from coincidence. To verify the financial theories. Further applications. 3

• The definition: • Channel and mechanism depict how treatment leads to changes of outcome variables. • 作用渠道或作用机制刻画因果效应何以存在、如何存在。 • Managers' education and corporate operation; Bank competition and financial constraints; Retail investors trading and market stability. • Why the identification of channel or mechanism is important? • Convince us that the Causal effect does not originate from coincidence. • To verify the financial theories. • Further applications. 3 The Definition of Channel and Mechanism

The Definition of Channel and Mechanism The key of Rubin Causal Model: ·Treatment-→Dutcomes. The treatment and non-treatment outcomes under same condition(assignment mechanism). Counterfactual outcomes. ·What will disturb unconfoundedness(非混杂性)? (Yo,Yu)IL D:Xi Covariates. D (a)基本因果模型I (b)基本因果模型Ⅱ (e)基本因果模型Ⅲ

• The key of Rubin Causal Model: • Treatment→Outcomes. • The treatment and non-treatment outcomes under same condition (assignment mechanism). • Counterfactual outcomes. • What will disturb unconfoundedness (非混杂性)? • Covariates. 4 The Definition of Channel and Mechanism

The Definition of Channel and Mechanism The correlation and causal effect: Causal inference:Causal identification and statistical inference. Causal identification:if A is not correlated with B,there is no causal effect between A and B.If A is corelated with B and there is only a causal model to identify the correlation,A leads to B or B leads to A. Statistical inference:Significant or insignificant correlation. If the coefficient for variable A in a regression on B is insignificant,can we conclude A has no causal effect on B? If the coefficient for variable A in a regression on B is significant,can we conclude A has a causal effect on B?

• The correlation and causal effect: • Causal inference: Causal identification and statistical inference. • Causal identification: if A is not correlated with B, there is no causal effect between A and B. If A is corelated with B and there is only a causal model to identify the correlation, A leads to B or B leads to A. • Statistical inference: Significant or insignificant correlation. • If the coefficient for variable A in a regression on B is insignificant, can we conclude A has no causal effect on B? • If the coefficient for variable A in a regression on B is significant, can we conclude A has a causal effect on B? 5 The Definition of Channel and Mechanism

The Definition of Channel and Mechanism The purpose of causal inference: The relationship between financial variables. The foundation:Financial theories. To understand Causal effect.The financial theory determines the Causal effect. Exploring the channel and mechanism can help ensure unconfoundedness. Why?If the Causal effect originates from the differences in unobservable covariates,what will happen when we explore the channel and mechanism? Consumer coupon and consumer spending. 6

• The purpose of causal inference: • The relationship between financial variables. • The foundation: Financial theories. • To understand Causal effect. The financial theory determines the Causal effect. • Exploring the channel and mechanism can help ensure unconfoundedness. • Why? If the Causal effect originates from the differences in unobservable covariates, what will happen when we explore the channel and mechanism? • Consumer coupon and consumer spending. 6 The Definition of Channel and Mechanism

The Definition of Channel and Mechanism Further application: Insider trading and stock returns-information processing and stock returns. Bank competition and financial constraints-Fund accessibility and financial constraints. Environmental court and green development-Law enforcement efforts and green development. Theoretical contributions

• Further application: • Insider trading and stock returns→ information processing and stock returns. • Bank competition and financial constraints→ Fund accessibility and financial constraints. • Environmental court and green development→ Law enforcement efforts and green development. • . • Theoretical contributions. 7 The Definition of Channel and Mechanism

Outline The Definition of Channel and Mechanism Causal Directed Acyclic Graph The Methodology for Channel and Mechanism Identification 8

• The Definition of Channel and Mechanism • Causal Directed Acyclic Graph • The Methodology for Channel and Mechanism Identification 8 Outline

Causal Directed Acyclic Graph 。 Essential element: ·Nodes(点):Variables ·Edges(箭头):Causal effect. ·子变量、孙变量 ·父变量、祖变量 图4.1因果图 9

• Essential element: • Nodes (点): Variables. • Edges (箭头): Causal effect. • 子变量、孙变量 • 父变量、祖变量 9 Causal Directed Acyclic Graph

Causal Directed Acyclic Graph 。Essential element: ·Common Cause(共同原因):D→/and X→Y. ·Confounder(混杂因素):X,D→Y ·Common effect(共同结果)or Collider(交汇变量):Y,D and X. ·,X,Z,andD? Z D 图4.1因果图 10

• Essential element: • Common Cause (共同原因): D→Y and X→Y. • Confounder (混杂因素): X, D→Y. • Common effect (共同结果) or Collider (交汇变量): Y, D and X. • X, Z, and D? 10 Causal Directed Acyclic Graph

共21页,试读已结束,阅读完整版请下载
刷新页面下载完整文档
VIP每日下载上限内不扣除下载券和下载次数;
按次数下载不扣除下载券;
注册用户24小时内重复下载只扣除一次;
顺序:VIP每日次数-->可用次数-->下载券;
相关文档