【作者】 刘康兵；

【导师】 尹伯成；

【作者基本信息】 复旦大学 ， 经济思想史， 2007， 博士

【摘要】 微观层面的企业投资是经济理论研究的中心课题之一，对厂商投资决策的分析在宏观经济学、公共经济学、产业组织理论和公司金融等学科中占据着非常突出的地位。在现实经济活动中，微观厂商的投资构成了整个社会总投资的主体，厂商的投资行为对经济增长和经济波动将产生重要影响，因此，分析影响厂商投资波动的因素，将有助于理解经济活动周期性波动的原因。本文从微观层面着手，对公司投资行为进行了深入分析，其中所关注的核心问题是：哪些因素导致了微观层面的厂商投资支出的波动?本文系统地研究了资本市场不完美和不确定性对微观厂商投资决策的影响，包括理论建模和实证分析两个层面，本文共分为七章：第一章为导论，主要介绍了本文的研究背景与动机、研究方法、思路和逻辑框架，并简要说明了论文的可能贡献和不足之处。第二章是文献回顾，对前人关于资本市场不完美和不确定性与厂商投资关系的两个系列的文献进行了综述分析，以作为本文建立理论模型和进行实证分析的基础和依据。第三章从理论上探讨了不对称信息、信贷配给和公司投资的关系。笔者把信贷市场信息不对称的意义与投资项目的可分性结合起来，通过一个最优信贷契约模型分析信贷配给和企业家净财富对公司投资的影响。在这一框架下，削减投资规模可被视为提供更多内部融资的备择方案。首先，在项目规模不可分条件下，笔者证实了斯蒂格利茨和韦斯的结论：如果银行将经济中所有企业家视为一组完全相同的借款人，存在一个随机信贷配均衡；如果银行能够根据某种可观察到的特征将企业家分成若干不同的类别，则均衡中还存在贷款歧视。其次，更为重要的是，文章发现均衡信贷配给意味着企业家使用外部融资的边际成本为无穷大。给定单位资本边际回报是有限的，此时削减一单位货币的投资，从而减少一单位货币的贷款需求，所获得的收益是无穷大的。因此，本文结合信息不对称和投资项目规模的可分性减少证明，在信贷配给均衡中，企业家有不可抗拒的激励选择削减投资支出，这种行为导致信贷配给消失，信贷市场出清。文中类似的分析还证明信贷歧视也不会成为一种均衡现象。最后，本文证明企业家对初始计划投资的削减幅度，进而其实际资本支出取决于企业家拥有的净财富水平，也就是说，厂商的投资受到融资约束(即厂商的投资支出对内部净财富水平具有过渡敏感性)，这一结论有着丰富的宏观经济含义，一个小的不利冲击可能通过影响厂商的净财富水平和外部融资溢价而加剧投资的波动，甚至形成大的经济周期。第四章以我国大陆制造业上市公司面板数据为样本，基于标准Q模型对融资约束假说进行了系统检验。资本市场不完美是否会对经济波动产生影响?第三章的理论模型在一定程度上对这一问题作了肯定回答，当然还需要经验证据的支持。近年来宏观经济方面的实证研究将这一问题指向了公司投资领域，即通过考察那些能够自由进入资本市场的厂商与受融资约束厂商在投资行为方面是否存在差异来寻求证据。我们使用三种分类指标将样本总体划分为“受融资约束”和“不受融资约束”的子样本，对融资约束假说进行检验：上市公司的国有股比例大小实质上代表国家信用所占比重的高低，股利支付政策反映公司的经营绩效和未来盈利前景，资产规模大小则体现公司可用于抵押的资产的多少，从而其举债的安全性高低。显然，第一个指标属于我国特有的制度因素，后二者属于市场因素。根据上市公司国有股比例进行分类的经验分析表明，现金流对公司投资存在普遍影响，并且国有股比例较小的上市公司其投资行为对现金流的敏感性大于国有股比例较高的公司，尽管前者的盈利能力等较差。这种敏感性的差异无论在统计上还是经济上都非常显著，一系列的稳健性检验还证明分析结果具有很强的稳定性。这一结论与融资约束理论的基本预测完全一致。根据上市公司的红利支付比率对样本总体进行的分类研究与理论预测恰好相反，红利支付比率高的厂商受融资约束的程度反而更大。最后，虽然基准模型的估计显示按资产规模大小进行分类时，资产规模大的上市公司其投资支出对现金流的敏感性更大，不过这一结果对模型的设定和估计方法比较敏感，稳健性检验表明资产规模小的公司其现金流估计系数大于资产规模较大的公司，尤为重要的是，所有情况下的Wald检验均不能拒绝这两组样本的现金流估计系数相等的虚拟假设，说明根据资产规模进行的分类研究并不能显著地识别出“受融资约束”和“不受融资约束”的厂商类别，从而也不能得出资产规模是否与企业所受融资约束程度直接相关的结论。这说明我国证券市场受制度因素影响非常大，市场化特征并不明显。企业的外部投资者，包括银行等金融结构，在决策时最优先考虑的因素是国家信用所提供的“隐性担保合同”的作用而非市场化指标所传递的信号。国家信用过度倾斜与行政性制度安排引致证券市场制度缺陷，使得既定体制下信息难以发挥其内在的传导作用，造成市场行为的扭曲，以致证券市场运行的效率过低，因而难以做到社会资源的有效配置。第五章通过在Q模型中引入一个通常被忽视的变量——营运资本投资——为融资约束假说及其数量效应提供新的证据。营运资本不仅是厂商融资资金的占用者，而且也是厂商流动性的一个重要来源，当厂商面临融资约束时它可被用于平滑现金流冲击产生的固定资本投资波动。在Q模型引入营运资本投资变量有两个作用：第一，它可用于检验实证分析中显著为正的现金流估计系数究竟反映的是流动性约束还是未来投资机会；第二，应用类似第四章的简化型Q模型可能会低估融资约束对投资的总量影响，因为这类模型没有控制厂商的投资平滑行为，为全面评估融资约束对厂商投资决策的影响，需要在模型引入营运资本的变化内生化厂商的投资平滑行为。本章的实证分析发现：在包含营运资本投资的Q模型中，营运资本投资的估计系数显著为负，这证明正的现金流估计系数反映的是融资约束而非投资需求的变化；与第四章的结果相比，固定投资对现金流的敏感性有着显著的提高，这证明标准的Q模型由于忽略营运资本的作用而低估了融资约束对厂商固定投资的长期影响。此外，两组样本厂商在现金流和营运资本投资估计系数方面的显著差异也进一步证实了融资约束和投资平滑假说。第六章集中分析了厂商面临不可逆约束时的投资行为，特别是在不确定性和资本市场不完美条件下的投资动态学。本章首先建立一个部分不可逆和部分可扩张性模型同时考察了投资决策的时序和最优投资水平的决定问题，通过引入投资买权和卖权概念，我们证明这种情况下根据传统的NPV原则决定的最优投资水平，实际上只是一种次优行为，因为这一投资决策标准忽略了边际买权和卖权的存在性及其对厂商投资行为的影响。同时还证明不确定性、部分不可逆性和可扩张性导致一个无任何投资发生的区域，并且厂商进行正投资、负投资还是“按兵不动”，取决于高需求状态值发生的概率和低需求状态值发生的概率，而与条件分布函数的具体形状无关。本章以厂商层面的成本和产品需求状态为例考察了不确定程度的变化对不可逆投资的影响，证明根据均值保留展形标准衡量的不确定性程度增加会导致厂商当期投资减少。本章最后将资本市场不完美引入包含不确定性的投资不可逆模型，考察了信息不对称对投资等待的期权价值和不确定性对边际风险溢价的影响以及这种情况下厂商投融资决策的相互关系。我们证明，在不对称信息条件下，不确定性的增加通过降低资本预期边际价值而使得边际风险溢价上升，导致外部融资溢价上升从而加剧厂商受融资约束的程度。第七章从实证角度考察了第六章所证明的资本市场不完美和不确定性的交互作用对公司投资支出的影响，即检验是否面临较高程度不确定性的厂商其投资支出受融资约束的程度更大。本章以上市公司股票价格的波动性作为不确定性的衡量指标，并以不确定性程度大小为标准对样本总体进行分类，不同于第四章和第五章的是，在那里，我们应用的是外生的或先验的分类阈值，但在本章，我们通过把不确定性引入实证模型中，应用汉森估计法内生地由模型决定阈值，实证分析结果表明，股票价格波动性越大的样本，其投资对现金流的敏感性越大，这种差异说明不确定性会加剧厂商的融资约束程度，本章最后应用阿雷拉诺—邦德GMM估计法进行的进一步的检验也支持了这个结论。更多还原

【Abstract】 Firm-specific investment is one of the central topics of study in economic theory. Analysis of the investment decisions of firms occupies a prominent place in research programs in macroeconomics, public economics, industrial organization, and corporate finance. The purpose of this dissertation is to explore the determinants of firm-level investment, since an understanding of this problem may shed light on the source of cyclical fluctuations of aggregate output. So the primary question that we focus on in this thesis is which factors affect the firm-level investment fluctuations.This dissertation analyzes the effects of capital market imperfections and uncertainty on firm investment, theoretically and empirically. The structure is organized as follows:Chapter one is introduction, it represents the research background, motivation, methodology and logic framework of this paper, and also its shortcomings and potential theoretical innovations.As our analytical basis, we review the existing literature that studies the relationship among capital market imperfections, uncertainty and corporate investment in chapter two.Chapter three analyzes corporate investment behavior in a credit market with asymmetrically informed lenders and borrowers and costly state verification. In a standard lending model, we replicate the results of Stiglitz and Weiss (1981) that there exist equilibrium rationing and/or redlining in credit markets. More importantly, however, we further demonstrate that equilibrium credit rationing implies that the marginal cost of funds raised externally to the entrepreneur is infinite. Henceforth, when project scale divisibility is introduced, as we do, entrepreneurs have an overpowering incentive to cut back their investment spending and thus loan demand even by one unit of currency. This action guarantees that lenders will be on the rising part of their return function. As a result, credit rationing disappears in equilibrium and the credit market clears. The similar analysis also shows that redlining is not an equilibrium phenomenon. Finally we show that the extent to which borrowers will cut their project, and therefore entrepreneurs’ actual capital spending depends on the level of net worth of firms. That is, firms are financing constrained, which means a small adverse shock may generate large investment fluctuations or even big business cycles through its effects on firm’s net worth and external finance premium.Chapter four tests the financial constraint hypothesis using the panel data of China mainland listed manufacturing companies and the standard Q model. Do imperfections in capital market play a role in output fluctuations? Recent research in empirical macroeconomics has directed this question to the area of investment, asking in particular whether firm with free access to capital markets have different investment behavior from those who do not. We classify the total sample into "constrained" and "unconstrained" firms by three different criteria. The ratio of state-hold shares means how much government credit is put in the firm. A company’s dividend policy reveals its operating performance and future profit prospects. The asset size of a firm represents its safeness in debt financing. Estimates from sample split by state-hold shares ratio show that financial effects are generally important for investment in all firms. But the results consistently indicated a substantially less sensitivity of investment to cash flow in firms with a ratio of state-hold shares greater than 50% than that of firms with a ratio equal to or less than 50%, although the former have a relatively poor performance. This statistically and economically significant difference was robust to a wide variety of model specifications and estimation techniques. These empirically important differences across firms are consistent with financial constraints arising from capital market imperfections. Contrary to the prediction of liquidity constraints hypothesis, the high-dividend firms are more cash constrained than low-dividend firms. Although estimates from the bench empirical model imply that the investment of big firms is more sensitive to fluctuations in their cash flow than that of small firms, the robust tests reverse these results. Most important, the Wald test cannot reject the null hypothesis that the cash flow coefficient of big firms equals the corresponding coefficient of small firms. All the abovementioned empirical evidence suggests that the outside investors, including banks and other financial institutions, put more emphasis on government credit than on signals generated by market factors when they make decisions.Chapter five conducts further tests for financial constraints and its quantitative effects on firm investment by introducing the often-ignored variable——working capital investment——into Q model. In the standard reduced form Q model, cash flow effects on investment have often been criticized as proxies for factors that shift investment demand rather than as evidence of financial constraints. Working capital is current assets less current liabilities, which acts as both an input and a readily reversible store of liquidity. Developing the role of working capital leads to two empirical predictions. First, when included as an endogenous variable in a fixed investment regression, working capital investment should have a negative coefficient if firms face financing constraints. Second, the reduced form Q model may underestimate the full impact of financing constraints on investment since it neglects the fact that firms smooth fixed investment in the short run with working capital. Consistent with our predictions, the coefficient of endogenous working capital is negative in the Q model. Moreover, controlling for the smoothing role of working capital results in a much larger estimate of the impact of finance constraints than reported in chapter four and other studies. These findings address the criticism and provide new evidence on financing constraints.Chapter six investigates firm investment behavior under irreversibility. In particular, we discuss the impact of uncertainty and capital market imperfections on the dynamics of investment. We first develop a model of partial irreversibility and expandability and examine the timing of investment decisions and the sequential determination of the optimal amount invested simultaneously. By introducing the concepts of call option and put option of investment, we show that firms that make their investment decision according to the naive NPV rule would be behaving sub-optimally since it ignores the effects of the two options on the expected marginal value of capital. We also demonstrate that irreversibility, uncertainty and expandability lead to a zone of inaction, and the firm’s decision to invest, disinvest or remain inactive at a point of time is affected by the probability of high demand and low demand state, but not by the shape of conditional distribution function of demand state. We next explore the effects of firm level cost uncertainty (the price of capital) and demand uncertainty on corporate investment in a model with complete irreversibility and show that a mean preserving spread in conditional distribution of demand state and the price of capital tends to depress the expected future marginal value of capital and raise the marginal endogenous adjustment cost, leading to a decrease in current investment. Finally, we consider a model of irreversible investment with uncertainty and capital market imperfections. It shows that a rise in uncertainty lowers the expected shadow value of capital and increases the marginal risk premium, thereby raising the external finance premium, which means firms have to face more severe financing constraints.Chapter seven uses a panel of China listed manufacturing companies to empirically explore the relationship between capital market imperfections, uncertainty and corporate investment, that is, we focus on whether firms that are confronted with higher degree of uncertainty will suffer more from financial constraints than those with relatively low uncertainty. In this chapter, stock price volatility is used as uncertainty measure. Firms are classified by a threshold level for uncertainty. More importantly, unlike chapter four and five, here we appeal to Hansen estimator to integrate the threshold level into the empirical model and determine the cut-off value of uncertainty simultaneously with the parameters of the model. The estimates show that the cash flow coefficient for high stock price volatility observations is substantially larger than the corresponding parameter for low volatility observations. The further test using Arellano-Bond GMM estimator obtains similar results. Therefore, we can conclude that there is a differential impact of cash flow on investment depending on the degree of uncertainty that firms are facing, suggesting that financial constraints arising from capital market imperfections are more severe for the relatively high volatility finns.更多还原