公司的投資決策是否會受到同行股價的影響：公司是否對同行股價有學習效應

Foucault and Fresard - JFE - 2014 - Learning from peers’ stock prices and corporate investment

主要關注點

公司的投資會受到同行的市場估值（股票價格）的影響。具體而言，我們檢驗了一個假設，即公司同行的市場估值會影響其投資，因爲這種估值可以告知管理者公司的增長機會，從而補充管理者可獲得的其他信息。例如，管理人員可以從專注於此活動的公司的股票價格中瞭解特定活動中增長機會的其他信息。

實證方法

檢驗公司投資和自己公司股價以及同行公司股價的關係

To estimate the covariation between a firm’s investment and (i) the stock prices of its peers and/or (ii) its own stock price, we include peers’ characteristics in a standard linear investment equation. Specifically, we consider the following baseline specification:
$I _ { i , t } = \alpha _ { i } + \delta _ { t } + \eta Q _ { - i , t - 1 } + \beta Q _ { i , t - 1 } + \varphi \mathbf { X } _ { - i , t - 1 } + \mathbf { \Phi } \mathbf { X } _ { i , t - 1 } + \varepsilon _ { i , t } \tag{1}$
the subscript -i represents an equally weighted portfolio of peer firms based on the TNIC industries (firm B in the model)

The dependent variable, $I_{it}$, is the ratio of capital expenditure in that year scaled by lagged fixed assets (property, plant, and equipment).

The explanatory variable $Q_{it-1}$ is the Tobin’s Q of firm i in year t-1 as defined above and $Q_{-it-1}$ , is the Tobin’s Q of firm i’s peers, computed as the average Q of all firms belonging to the same TNIC industry as firm i in year t - 1, excluding firm i.

The vectors $\mathbf { X }$ include control variables known to correlate with investment decisions.

用標準差對解釋變量做了標準化，於是係數 $\eta$ 和 $\beta$ 分別是公司投資和自己股價和同行股價的協方差。同時，這樣標準化後，係數的經濟意義直接就是效果的經濟顯著性，而非僅僅是統計顯著性

We scale all independent variables by their standard deviation. Hence, coefficients $\eta$ 和 $\beta$ are the empirical counterparts of $\operatorname { cov } \left( I , p _ { \mathrm { B } 1 } ^ { \mathrm { s } } \right)$ and $\operatorname { cov } \left( I , p _ { \mathrm { A } 1 } ^ { \mathrm { s } } \right)$(see the discus- sion in Section 2.4). Another advantage of this scaling is that the magnitude of the estimated coefficients is directly informative about the economic significance of the effects.

具體估計分兩步：

We use estimates of these coefficients to test the main implications of the model. We proceed in two steps.

首先，估計模型(1)，顯示 $\eta$ 和 $\beta$ 都是統計顯著的。這是“學習性同行效應“成立的必要條件，但不是充分條件

First, in Sections 4.1 and 4.2, we establish that estimates for $\eta$ and $\beta$ are both statistically significant in our sample. This is indeed a necessary condition for the “learning from peers” channel to play a role (see Corollary 3). This condition is not sufficient, however.

$Q_i$ and $Q_{-i}$有兩層含義：(i) 都是管理層關於投資機會私有信息的代理變量 (ii) 都是管理層從股價中學習信息的決定因素。因此， $\eta$ 和 $\beta$ 顯著可能並非是由於學習效應，而僅僅是因爲私有信息的作用

Indeed, as highlighted by the model, $Q_i$ and $Q_{-i}$ play a dual role in Eq. (1): they are both (i) proxies for unobserved managers’ private information about their investment opportunities and (ii) determinants of investment if managers learn information from stock prices. Thus, a significant association between investment and Tobin’s Q might exist even if managers do not learn from stock prices, simply because Tobin’s Qs act as a proxy for unobserved managers’ signals (the correlated information channel).

爲了解決上述的Concern，我們檢驗了錨定於 “learning from peers” 場景下的模型顯著性。即檢驗價格的信息含量，管理層信息，公司股價相關性是否影響 $\eta$ 和 $\beta$ 的顯著性。

Hence, in a second step (Section 4.3), we test the predictions of the model about $\eta$ and $\beta$ that are specific to the “learning from peers” scenario. That is, we test whether proxies for the model’s parameters (price informativeness, managerial information, and the correlation in demand shocks between firms) affect $\eta$ and $\beta$ as Table 1 predicts when managers learn from their peers, with a special focus on the five predictions specific to this scenario.

具體來說，檢驗投資者和同行股價之間的相關性是否受到

• 知情交易(informed trading)
• 排除股價的管理層信息($\gamma$)
• 公司股票與其同行股票之間需求量的相關性($c(\rho)$)

的影響(引入交乘項)：
$\begin{aligned} I _ { i , t } = & \alpha _ { i } + \delta _ { t } + \eta _ { 0 } Q _ { - i , t - 1 } + \eta _ { 1 } \left[ Q _ { - i , t - 1 } \times \phi _ { i , t - 1 } \right] + \beta _ { 0 } Q _ { i , t - 1 } \\ & + \beta _ { 1 } \left[ Q _ { i , t - 1 } \times \phi _ { i , t - 1 } \right] + \cdots \end{aligned}$
Specifically, we study how the covariation between investment and own’s and peers’ stock prices varies with measures of (i) informed trading ($\pi_A$ and $\pi_B$), (ii) managerial information other than stock prices($\gamma$), and (iii) the correlation in demand shocks between a firm and its peers ($c(\rho)$).