資產價格、違約風險與貨幣政策

Abstract

本論文包含五個章節。第一章介紹本論文架構。第二章主要探討在小型開放經濟體下，資本流動的動態調整與國內土地價格、產出的波動。本篇採以結合Iacoviello (2005)與Aoki et al. (2007)的模型，發展一個一般均衡的理論。模型的特色為，國外資本流動會透過擔保限制式與資產價格影響經濟體，而家計單位與企業兩者都受限制於擔保限制式，且國外借款人的擔保資產會比國內借款人的受限較大。研究結果發現相對於封閉經濟體，小型開放經濟體在面對外生衝擊時，會引起較大的總體經濟變數，如：消費、產出與資產價格的波動。再者，金融深化於經濟體面對各種衝擊時，在決定總體經濟變數動態過程中扮演著重要的角色。與封閉經濟體相比較，小型開放經濟體在金融體系發展程度增加時，不只引起立即且對於衝擊有持續性的反應。最後，我們根據Schmitt-Grohe and Uribe (2004)衡量家計單位與企業預期終身效用的福利分析，結果顯示央行的貨幣政策在小型開放經濟體下對於資產價格變動需做正向反應。 第三章主要探討擔保品對於違約機率的邊際效果，我們採用傾向分數分配方法，控制了樣本選擇偏誤問題後，來衡量有使用擔保品借款與沒有使用擔保品借款的違約機率之差異。本文使用1995-2008年台灣借貸交易追蹤資料，經實證結果發現，控制了樣本選擇偏誤後，有使用擔保品借款比沒有使用擔保品借款的違約機率顯著地高了2.8%- 3.9%，此結果比過去實證文章使用Logit模型的結果來得小。另外，我們將樣本根據房屋價格漲跌週期分成兩個時期：1995-2001年房屋價格下跌期及2002-2008年房屋價格上漲期來做進一步探討。在房屋價格下跌期，有無使用擔保品借款違約機率並沒有顯著差異；但在房屋價格上漲期，有使用擔保品借款比沒有使用擔保品借款的違約機率顯著高了4.1%-5.7%，此結果也比過去實證文章使用Logit模型的結果來得小。此外，我們資料也說明了，在房屋價格上漲期，銀行放款標準較低，而房屋價格上漲加上放款條件的寬鬆造成了違約風險較高的後果。 第四章以台灣為例子，主要探討在小型開放經濟體下，央行雙率政策（利率政策與外匯干預政策）的反應函數。對於小型開放經濟體，央行的匯率政策對於影響經濟體的重要性並不亞於利率政策，而匯率政策的執行通常透過外匯干預。雙率政策的反應函數從理論模型裡央行極小化重要經濟變數的波動推導而來。當面對匯率波動時，央行的利率政策和外匯干預政策與匯率之間會有內生關係，因此，本文採用Kourtellos et al. （2007）的內生門檻模型，探討央行雙率政策的執行是否在匯率升值或貶值期間而產生不對稱之干預行為，其中內生門檻變數為匯率變動。實證結果顯示，央行的利率政策僅對預期通膨反應，且門檻效果並不存在。另一方面，我們發現，外匯干預政策對於產出、匯率與預期通膨確實有不對稱干預之傾向；且在台幣強力升值時，央行干預之幅度較大。這結果支持了央行為了維持出口競爭力的優勢而進行外匯干預。第五章為結論。

The dissertation consists of three essays. Chapter 1 indicates the introduction. In chapter 2, we study the dynamics of capital flows and the fluctuations of domestic land prices and outputs when the economy is open to foreign capital. Following Iacoviello (2005) and Aoki et al. (2007), we build a general equilibrium model in which foreign capital affects the economy through interactions between collateral constraints and asset prices. Both households and entrepreneurs are subject to collateral constraints, and collateralizable assets for international borrowing are more restricted than domestic borrowing. We find that opening to foreign capital brings about larger fluctuations in economic aggregates, such as consumption, output, and asset prices. Furthermore, financial deepening plays a significant role in determining the dynamics of economic aggregates upon various shock impacts. Compared to a closed economy, an increase in the financial institution development raises not only the immediate impact but also the persistence of an exogenous shock. We also conduct welfare analysis based on Schmitt-Grohe and Uribe (2004) by evaluating the expected lifetime utility functions of households and entrepreneurs up to the second order approximation. The result suggests that monetary policy should strongly respond to changes in asset prices in this environment. In chapter 3, we examine the marginal effect of collateral on the probability of default. Linking a panel of Taiwan's loan transactions with individual firms' financial information from 1995-2008, this study uses propensity score matching to measure the difference in default probability between collateralized and non-collateralized loans to control for the sample selection problem. After controlling for sample selection bias, the difference in the marginal effect of collateral on default probability for the two groups ranges from 2.8% to 3.9%. Though statistically significant, the difference is substantially lower than a conventional logit model implies. After splitting the sample into two sub-periods according to the house price cycle ('bust,' 1995-2001 and 'boom,' 2002-2008), the former shows no difference in the marginal effect of collateral on default probability between collateralized and non-collateralized loans. However, the latter sub-period shows a significantly larger difference between the two groups, ranging from 4.1% to 5.7%. The data indicate that lending standards were lower during the boom phase of the house price cycle, and imply that lower lending standards during housing-market booms may inevitably pave the way for a higher credit risk. In chapter 4, for a small open economy, the exchange rate policy of the central bank is no less important than the interest rate policy in affecting the economy, and the policy is usually conducted through foreign exchange intervention. In this chapter, we investigate the central bank's reaction functions of both of the interest rate and foreign exchange intervention for a small open economy, using Taiwan as an example. The reaction functions are derived from the model where the central bank minimizes fluctuations in key macro variables. When facing the fluctuations of the exchange rate, central bank's foreign exchange intervention would have feedback effects on the exchange rate itself. Therefore, the functions are then estimated by a two-regime TAR with endogenous threshold model based on Kourtellos et al. (2007) where the regimes are determined by the speed and the direction of change of the exchange rate. Results show that the interest rate policy mainly reacts to expected inflation and there is no evidence of threshold effects. On the other hand, we find strong evidence of asymmetric responses of intervention to changes in output, exchange rate, and inflation and the responses are shaped in favor of promoting exports. Chapter 5 concludes the findings in the previous chapters.