优博-政策工具的挤出效应与挤入效应研究

作者姓名：杨子晖

论文题目：政策工具的挤出效应与挤入效应研究

作者简介：杨子晖，男，1979年7月出生，2004年9月师从于中山大学陈浪南教授，于2007年6月获博士学位。

中 文 摘 要

私人投资与私人消费是一个国家总需求重要的组成部分，对国民经济的发展有着举足轻重的影响。货币政策与财政政策是宏观调控的两大基本工具，它们在需求管理中的运用是否能达到预期的效果常常成为人们关注的问题。然而，自九十年代中后期以来，我国经济在转轨过程中呈现出以内需不足为主要特征的基本态势，为了刺激有效需求，我国政府实施了一系列旨在促进经济增长的扩张性政策。扩张性政策的实施是否对私人部门支出产生挤出（或挤入）效应？此外，作为货币政策和财政政策结合点的国债政策在调节社会资金供求方面的作用也日益增强，特别是自九十年代中后期以来，随着扩张性政策的实施，作为政府重要融资方式之一的国债发行量更是大幅增长。国债融资是否引起政府与民间部门在资金需求上的竞争进而挤出了私人投资？国债融资规模的大幅增长是否引起理性个人对未来税负增加的预期，并由此减少了现期消费，从而引发挤出效应？所有的这些都是我们目前亟需研究的重大问题，对此问题的回答都将直接关系到未来的政策性选择与安排，而国内外学术界对此问题还没完全达成一致的结论。有鉴于此，本文的研究目标就是结合我国实际经济条件，考察政策工具对私人部门支出（私人投资与私人消费）是否产生挤出或挤入效应。

为了克服国内外现有研究文献分析框架的局限性，增进本文结论分析的可靠性与合理性，在研究方法上，本文运用了一系列最新发展的现代计量经济学方法，主要包括：（1） 为了有效解决传统时间序列单位根检验以及协整检验的小样本问题，本文在协整框架内对政策工具是否引发挤出（或挤入）效应的问题进行国际比较研究时，采用了最新发展的面板单位根检验、面板协整检验以及面板协整估计方法，其中面板单位根检验主要包括IPS检验（Im、Pesaran和Shin，2003），Fisher-ADF和Fisher-PP检验（Maddala和Wu，1999；Choi，2001）以及Breitung检验（Breitung，2000），面板协整检验主要采用了Pedroni（1999，2004）检验方法，面板协整估计则采用组间面板FMOLS估计方法（Pedroni，2000，2001）；（2） 为了克服传统的Granger因果检验和传统方差分解方法的局限性，本文在考察政策工具对私人部门支出的动态影响时，采用了最新发展的“有向无环图技术”（DAG；Pearl，1995，2000；Spirtes et

al.，1993；Swanson和Granger，1997），以研究政策工具在长短期对私人投资的动态影响；（3） 在传统的时间序列分析中，本文也采用了一些较为前沿的分析方法，其中在协整估计中，除了采用传统的VECM方法，本文运用了完全修正OLS（Phillips和Hansen，1990）、动态OLS（Stock和Watson，1993），并采用了“递归协整分析方法”（Hansen和Johansen，1999）对协整关系的稳定性进行检验。此外，本文还采用了递归方差分解分析、非参数相关性检验等一系列较为前沿的计量经济学方法。

为了全面地、系统地阐述政策工具是否引发挤出（或挤入）效应，本文结合最新的理论模型的发展，主要从以下七个方面展开深入的分析与研究：

其一，本文结合新的理论模型的发展，从期内替代与跨期替代的角度，对我国财政支出是否挤出居民消费的问题进行研究，并考察转轨时期我国消费行为的结构性转变。其二，本文结合最新发展的面板协整技术，分别从总体样本以及单位样本这两个不同角度，对财政支出是否挤出私人消费的问题进行跨国研究。其三，本文结合非参数检验等相关性分析方法，对财政支出与私人消费关系形成的影响因素进行研究。其四，本文结合我国实际经济条件，在对世代交叠模型进行有益拓展的基础上，从理论分析与经验分析的角度深入地考察了我国货币融资、国债融资等不同融资方式对私人投资的挤出或挤入效应，其中重点考察了不同经济条件下的国债融资对于私人投资所具有的不同效应。其五，本文结合新的理论模型的发展，考察国债对私人消费是否产生非线性的影响，其中重点考察了债务规模的扩大是否导致私人消费的减少，从而引发挤出效应。其六，本文在货币－财政政策的框架内，采用最新发展的“有向无环图技术”（DAG），考察了政策工具在长短期对私人投资的动态影响，在对政策工具是否引发挤出效应作进一步讨论的基础上，对货币政策与财政政策这两种政策工具在总需求管理中的有效性进行比较。在此研究过程中，本文结合新的理论发展，从货币政策影响私人投资的“货币渠道”、“信贷渠道”以及财政政策的直接作用机制等角度进行分析。最后是对本文结论的一个全面的概括总结，并在此基础上提出了现阶段完善我国货币政策、财政政策以及国债政策协调机制、增强宏观调控能力的政策建议。

本文研究结果表明：

（1）关于国债融资是否挤出私人投资，本文研究发现，在不同经济条件下，国债融资对我国私人投资具有不同的效应。在1980～2003年总体样本时期内，国债融资在一定程度上减少民间的资金供给，挤出了私人投资。进一步的分析显示，当经济有效需求不足，社会存在着大量闲置资金时，国债的发行并不对私人投资产生挤出效应，且在一定程度上拓展了财政支出的融资空间，增强了财政政策宏观调控的能力。实证结果同时显示，即使在国债产生挤出效应情况下，只要我们把国债资金进行以公共投资为主的经济建设，它的净效应依然为正，因此，我们政府当局不仅仅要在不同的经济条件下把握好国债融资的适当规模，更重要的是

坚持好国债资金正确的使用方向。

（2）关于国债融资对私人消费的宏观效应，本文研究发现，公共债务对私人消费有着非线性的影响：在低债务的国家中，国债发行所引发的财富效应增强了居民的消费意愿，从而使得公共债务对私人消费的影响显著为正；然而，在高债务的国家，公共债务规模的扩大引发了理性消费者对未来税负增加的预期，因此，公共债务对私人消费的影响显著为负。这就意味着对于通过大规模发行国债而实施的扩张性政策，我们必须加以审慎对待，债务规模的扩大可能将导致私人消费的减少，从而引发挤出效应，进而降低了政策工具在需求管理中有效性。

（3）关于财政支出是否挤出私人投资，本文研究发现，我国政府的公共投资在提高私人资本边际产出的同时挤入了私人投资，社会文教费的支出则对私人投资有着负影响。因此，当我们分析财政政策、特别是赤字性财政政策对经济增长的影响时，我们必须对各项的经常性支出和资本性支出加以区别对待，如果大规模减少公共投资支出，在长期可能将影响私人投资的健康增长，特别是在内需不足、非理性投资热潮退却（如房地产市场投资泡沫破灭）的情形下，由大规模减少公共投资而对私人投资造成的影响可能将进一步凸现。

（4）关于货币政策对私人投资的影响，本文研究表明，扩张性政策在刺激总需求、促进投资时，最终都体现为货币购买力的增加，因此，货币的适度增加有利于促进私人投资的增长。此外，本文结合货币政策的不同传导途径作进一步的深入剖析，研究表明相比较“货币渠道”而言，“信贷渠道”仍是我国现阶段货币政策影响私人投资的主要传导途径，但由于货币到信贷传导环节的断裂，使得“信贷渠道”自身存在着较大的政策局限性。

（5）由于货币政策缺乏相对的独立性、“信贷渠道”的自身局限性以及利率机制的僵化等原因而造成了货币政策对私人投资等实体经济部门影响乏力；相比较而言，由于政府通过财政支出进行的基础设施建设与私人投资成“互补”关系，从而在很大程度上直接挤入了私人投资，此外，由于财政支出并非真实利率上升的潜在因素，而公共支出在总体上并无“挤出”私人投资，因此，我国的财政政策对私人投资有着巨大影响作用，财政政策在总需求管理中更具有效性。

（6）关于我国财政支出是否挤出居民消费，本文研究发现，虽然政府与居民消费的期内替代关系在改革后进一步凸现，但由于居民消费的跨期替代弹性较大，改革后居民与政府消费仍保持着互补关系，财政支出挤入了居民消费，此外，本文还发现，1978年后我国政府与居民消费行为发生了结构性转变；关于财政支出是否挤出私人消费的国际研究，本文分析表明，在本文研究的样本国家及时期内，总体上政府与私人消费成替代关系，政府消费支出在一定程度上挤出了私人消费，降低了乘数效应，与此同时，在对单位样本国家的比较分析中我们发现，不同的样本国家的政府消费与私人消费呈现出不同的互补（或替代）关系，这就

为各国未来财政政策的选择与安排进一步提供了参考依据。

（7）关于财政支出与私人消费关系的影响因素分析，本文结合非参数相关性检验研究表明，替代弹性与政府规模（国防支出比重）没有显著的相关关系，这就意味政府规模与国防支出比重并非替代程度大小的决定性因素，Karras（1994）以及Evan和Karras（1996）有关替代弹性与政府规模（国防支出比重）成正（负）相关关系的结论并不成立。

目前，结合我国实际经济条件对政策工具是否引发挤出效应的问题展开综合性、系统性的研究仍较少。本文运用最新发展的现代计量经济学方法，结合我国转轨经济的特点，从各个崭新的角度对该命题展开深入地阐述与分析，并对现有研究文献的做了有益的拓展和补充。此外，本文不仅增进了学术界关于政策工具是否引发挤出效应问题的理解，而且有助于我们重新审视并检验债务中性理论、新古典理论、凯恩斯理论以及货币学派等相关理论在中国的适用性。因此，本文对于拓展转轨经济条件下政策工具的挤出与挤进效应的研究具有重要的学术价值。

与此同时，本文的结论与政策建议为政策当局对我国未来货币政策、财政政策及国债政策的选择与安排提供理论分析与实证检验的参考依据。因此，本文对于完善我国宏观需求调控机制，提高宏观经济决策的科学性和系统性具有重要的实用价值。

关键词： 货币政策 财政政策 国债融资 挤出效应 挤入效应

Research on Crowding-Out and Crowding-In Effects of Policy Tools

Yang Zihui

ABSTRACT

As the essential components of domestic demand, private investment and private consumption have very important influence over the economy development; monetary and fiscal policy are the key macro-control policies and the effectiveness of aggregate demand policies has become the focus of public attention, however, since the middle and latter 1990s, there was underdemand of growth in the macroeconomy of China during economic transition. To stimulate domestic demand, Chinese government has implemented the expansionary policy. Could the expansionary policy crowded out (or crowed in) private sector spending? In addition, as the link between monetary and fiscal policy, government debt plays a more important role in controlling the demand and supply of society capital, especially, with the implementation of expansionary policy from the middle and latter 1990s, as the important channel of government financing, government debt scale enlarges on a large scale. Could debt financing result in the competition of demand for capital between government and private sector then crowd out private investment? When government debt reaches high levels, could expectations about future taxes result in crowding out by a fall in private consumption? All these are the problems we have to solve urgently and the answers to these questions could provide the references for us to choose and arrange the future policies, however, until now there appears to be no clear consensus among research works on these questions. In light of this, this dissertation investigates whether policy tools crowd out (or crowd in) private sector spending (private investment and private consumption) during China’s economic transition.

To improve the reliability and rationality of the analysis framework, this dissertation applies a lot of recently developed modern econometrics methods as follows: firstly, to solve the low power problem of conventional unit root test and cointegration test, this dissertation uses the recently developed panel unit root tests, panel cointegration tests and panel cointegration estimation when we investigate whether policy tools cause crowd-out effect among the international countries. Three panel unit root tests are applied here: IPS test (Im, Pesaran and Shin, 2003), Fisher-type tests using ADF and PP tests (Maddala and Wu, 1999; Choi, 2001) and Breitung test (Breitung, 2000). We test for cointegration data employing the heterogeneous panel cointegration test developed by Pedroni (1999, 2004) and then applies between-group FMOLS (Pedroni, 2000, 2001) to panel cointegration estimation; secondly, to overcome the limitation of Granger causal tests and forecast error variance decomposition, this dissertation applies the new Directed Acyclic Graphs (DAG, Pearl, 1995, 2000; Spirtes et al., 1993; Swanson and Granger, 1997) technology to investigate the short- and long-term dynamic effects of policy tools on private investment; thirdly, this dissertation uses the FMOLS (Phillips and Hansen, 1990) and DOLS(Stock and Watson, 1993) to conventional cointegrated series besides the conventional VECM and then employs recursive cointegration analysis technique

(Hansen and Johansen, 1999) to test the stability of cointegration relation. In addition, we also use other advanced econometrics methods such as recursive forecast error variance and non-parametric correlation coefficient tests.

The dissertation consists of nine chapters that can be grouped into seven main parts.

The first part investigates the relationship between government and household consumption in China by analyzing intratemporal and intertemporal elasticity of substitution, and then studies the structural change of Chinese consumption in economic transition; the second part is an international investigation on whether government spending crowds out private consumption. It employs the recent developed panel cointegration techniques to make the individual and panel analysis; the third part applies non-parametric correlation coefficient tests to investigates the factors affecting the relationship between fiscal spending and private consumption; the fourth part, based on extending overlapping generation model and China’s actual economic condition, theoretically and empirically investigates whether the government financing such money financing and debt financing crowds out (or crowds in) private investment, especially studying the different effect of debt financing on private investment under different economy conditions; the sixth part, based on the recent developments of theory, investigates the non-linear effect of government debt on private consumption; under the framework of monetary and fiscal policy, applying the new Directed Acyclic Graphs (DAG) technology the sixth part investigates the short- and long-term dynamic effects of policy tools on private investment, and analyzes whether policy tools crowd out private investment then examines whether monetary or fiscal policy is the more "efficient" instrument for influencing aggregate demand. It analyzes from the monetary channel, the credit channel and the direct transmission mechanism of fiscal policy. The last part of this dissertation summarizes the main conclusions and then puts forward some suggestion to strengthen macro-control effectiveness and to improve the co-ordination among monetary, fiscal and debt policy.

The conclusions of this dissertation are as follows:

(1) When we investigate whether government debt crowds out private investment, the results indicate that government debt has different impact on private investment under different economic conditions. During 1978～2003 debt financing decreases the supply of fund for private sector spending and then crowds out private investment, however, the results also reveal that when there are plenty of idle funds in society due to insufficient domestic demand, debt financing could not crowd out private investment. Instead, it extends the channel of government financing to a certain extent and strengthens the macro-control ability of fiscal policy. We also finds that even when debt induces crowding out effect, if we spend the funds financed by debt on economic construction such as public investments, the net effect of debt financing is positive. So we must keep an appropriate scale of debt based on different economic conditions, what’s more important, we must make appropriate use of the funds financed by debt.

(2) When we investigate the macroeconomic effects of government debt, the results indicate that the non-linear effect of government debt on private consumption. In low debt countries, wealth effects induced by debt strengthen household’s willingness to consume, so government debt has a significant positive impact on private consumption, however, in high debt countries, expectations

about future taxes would result in crowding out by a fall in private consumption, so government debt has a significant negative impact on private consumption. This means that an expansionary fiscal policy financed by debt will be partly crowded out by private consumption and maybe less effective in stabilising business cycle fluctuations.

(3) When we investigate whether fiscal spending crowds out private investment, the results indicate that Chinese government’s public investment raises the marginal productivity of private investment and then crowds in private investment, however, government spending on social, cultural and educational development has a significant negative impact on private consumption. This means that when we analyze the effect of fiscal policy (especially deficit fiscal policy) on private investment, we must distinguish carefully between current expenditure and capital expenditure. If we decrease public investment on a large scale, the healthy increase of private investment will be affected in long term.

(4) When we investigate the effect of monetary policy on private investment, the results indicate that when an expansionary policy is implemented to stimulate domestic demand, the demand for money will go up, so a seemly increase of money supply could promote private investment. In addition, we also find that the monetary policy affects private investment through both the monetary channel and the credit channel, but the latter plays a more important role in transmitting monetary policy, however, the credit channel has low effectiveness due to the obstructions of transmitting from credit to money.

(5) We also find that monetary policy has less influence over private investment because of no independence of monetary policy, the limitation of credit channel and the control of interest rate system; on the other hand, since government spending on infrastructure construction improves the environment of private investment and is complements for private investment, moreover, fiscal spending is not the factor leading to the increasing of real interest and do not crowd out private investment on a whole, so fiscal policy has more influence over private investment and is more effective than monetary policy.

(6) When we investigate whether fiscal policy crowds out household consumption in China, the results indicate that since intertemporal elasticity of substitution is greater than the intratemporal elasticity of substitution, government and household consumption are best described as complemental after reform, implying there will be direct crowding in of household consumption by fiscal spending. In addition, we also find that there was a structural change of China’s consumption after 1978. When we investigate whether government spending crowds out private consumption for ten Asian countries, panel cointegrating regression uncovers that a significant degree of substitutability between government consumption and private consumption. Country-by-country analysis, however, reveals diversity in the relations we studies, so this can further provide the references for countries to choose and arrange their future fiscal policies.

(7) When we use the non-parametric correlation coefficient tests to investigate the factors affecting the relationship between fiscal spending and private consumption, the results indicate the government size and the share of government expenditure that goes to national defense are not the key factors to explain the degree of substitutability between them, implying there are no further

evidences to support the hypothesis of Karras(1994) and Evan and Karras(1996).

Until now very little academic research has been done systemically and synthetically on surveying whether policy tools cause crowd-out effect based on particular economic conditions in China. In light of this, based on particular economic conditions in China’s economic transition and applying the methods of modern econometrics, this dissertation has extended this line of research by making a thorough research. In addition, this dissertation not only extends the academia’s understandings about whether policy tools cause crowd-out effect but also helps us to re-analyze the applicability of debt neutrality, Keynesian, neo-classical and monetarist theory in China, so this research has an important theoretical significance.

This research has provided the theoretical and empirical references for the authorities to choose and arrange their future monetary, fiscal and debt policy, so it has an important practical significance in improving macro-control policies and making our macroeconomic decision more systematic and more scientific.

Key words: Monetary Policy; Fiscal Policy; Debt Financing; Crowd out; Crowd in