| 摘要: | 既有文獻顯示，美國公司投資商用不動產竟有顯著「高買低賣」的現象；其解釋為公司於經濟擴張時期購置不動產，於衰退期時售出，隱含其添置/處分不動產目的不為獲利，反之可能為公司資產配置的調整。由於公司理財文獻中對廠商如何配置資產的理論研究付之闕如，本計畫擬整合並延伸兩靜態模型，亦即Cochrane (1991) 的生產面資本資產定價模型、與考慮耐久性財貨（如持有不動產）消費的Grossman and Laroque (1990) 消費面資本資產定價模型，以嚴謹理論模型及實證分析探討影響公司資產配置的產品市場因素。第一年計畫中，本人擬依循 Cochrane，假設廠商選擇產量、固定資本投資等以使其利潤極大；並將 Grossman and Laroque 模型中的代表性消費者轉換為廠商，延伸模型中的資產至三項：風險性/無風險資產及耐久性財貨。本人亦將依循並延伸 Dixit (1991) 的模型設定，假設廠商面臨固定彈性之需求函數，其生產函數為寇布─道格拉斯型式，且風險性/固定資產的價格均遵循幾何布朗寧模式。模型的優點為 (1) 考量資產價格的動態變化，假設廠商短期追求最大生產利潤，長期則在最適產量下極大化公司價值，其目標函數符合 Grossman and Laroque 模型解的均值─變異數效率性；(2) 假設特定函數型態可簡化計算，進而導出分析解及比較靜態結果。 模型推導的初步結果如下：當廠商面臨產品市場競爭強度較高/或其生產方式較為資本密集，配置在固定資產比例可能以遞減的速率增加。這是因為，此二產品市場特性降低廠商整體投資組合的變異數，因而其可增加具風險的固定資產配置，以提昇公司價值。第二年的計畫目的在於驗證理論模型推導之結論。本人擬採用紐約證券交易所等公開發行公司、由1987年到2016年的縱橫資料進行實證。本人將利用公開網站如華爾街日報取得股市大盤的趨勢和波動度資料；並擬由 Compustat 及 CRSP資料庫取得公司財務報表資料，以計算公司配置在固定資產比例及其他控制變數；至於商用不動產交易及特性、市場的趨勢和波動資料則擬自CoStar COMPs® 及 Properties® 等資料庫求取。實證方法為 (1) 以公司固定資產比例為應變數進行最小平方法迴歸，及 (2) 以不動產交易為應變數，利用序列 probit 以及多項式 logit 對前述假說進行統計測試。 ;Empirical literature shows that U.S. corporations tend to “buy high and sell low” when it comes to commercial real estate transactions, and explains that American firs tend to acquire (dispose) real estate properties during the economic expansion (downturn), implying a motive of asset reallocation. As in the corporate finance field there exists no analytical models can be used to investigate corporate asset allocation, the purpose of this research project is to examine this important issue by integrating the production-based CAPM by Cochrane (1991), and the consumption-based CAPM by Grossman and Laroque (1990), so as to discuss how asset and product market characteristic affect corporate asset allocations – especially on real estate holdings – with a theoretical model and corresponding statistical analysis.In the first year under the proposal, I assume the firm has a goal similar to Cochrane which assumes a firm chooses production quantity and amount of capital investment so as to maximize the profit of the firm. I will replace the representative consumer in Grossman and Laroque with a representative firm, and then extend the two assets in their paper so that the firm considers investing in three assets: risky/riskless assets and a durable good. I will also follow Dixit (1991) to assume that the firm faces a constant-elasticity demand function, has a Cobb-Douglas production function, and the evolution of prices for the risky asset and the durable good follow a geometric Brownian motion, respectively. As the firm in my model maximizes its value while employing an optimal quantity of production at each instant, the long-run equilibrium in my model is then mean-variance efficient, similar to Grossman and Laroque. Under a simplified assumption that the firm does not face uncertainty nor incur sunk costs when purchasing capital, it will allocate more value on the durable good (including its real estate holdings) but with a decreasing speed when the firm either faces more competitive pressure or uses capital more intensively. In the second year under the proposal, I aim to statistically test if the derivations in my theoretical model are supported by using empirical data. I plan to us the panel data of the U. S. firms listed on the NYSE, AMEX, and NASDAQ exchanges during the period from 1987 to 2016. I plan to employ stock market data drawn from publicly available web sites such as the Wall Street Journals to obtain the trend and volatility data of the risky assets. For calculating the real estate holdings and other control variables of public companies, I plan to use data drawn from Compustat and CRSP data bases. For accessing the prices and other features of commercial real estate transitions, I plan to employ CoStar COMPs® and Properties® databases. For the econometric analysis I will use (1) the OLS method to run regressions with corporate fixed assets holdings as the dependent variable; and (2) both ordered probit and multinomial logit methods to run regressions with a dependent variable which indicates the purchase, inactive, and dispose of real estates of a firm. |
