| 摘要: | 流動性創造（liquidity creation，以下簡稱LC）是銀行其中一項最重要的功能，透過將短期、流動性高的負債，轉換為長期、流動性低的資產，可向經濟體系的其它非銀行部門提供流動性。由於資本與流動性為確保銀行穩健度之關鍵因素，故此二者除在巴塞爾資本協定三（Basel III）中受到高度管制，它們之間的關係在2008年金融危機後的文獻中也受到廣泛討論。
然而，目前的研究仍未對哪些影響資本與LC的因素達成一致的共識，因此也無法建立出一個兩者之間合理的連結，此議題提供本文研究動機。我們提出：在多數文獻中被忽略的資本適足性可能為影響資本與LC交互關係的一個重要因素。由於資本與LC實際上是密切相關的，故本文將從兩個相反方向來剖析此一議題。
本文第一章探討資本適足性是否會影響由資本向LC的連結。透過考慮符合巴塞爾資本協定三架構之總資本適足率（total capital adequacy ratio，以下簡稱CAR）以及槓桿比率（leverage ratio，以下簡稱LR）的嶄新方式，定義出我們的「資本適足性」以及兩類銀行。我們提出：弱資本適足（weak capital adequate）銀行的風險意識較為不足，為符合文獻上之「金融脆弱性／排擠假說」（financial fragility/crowding-out hypothesis）；強資本適足（strong capital adequate）銀行的風險意識則較為審慎，為符合文獻上之「風險吸收假說」（risk-absorption hypothesis）。前類銀行會傾向創造更多流動性、且具有負向的資本與LC 關係；後者則恰好相反。
第二章檢驗資本適足性如何影響LC向資本變動的連結。結合CAR以及LR兩個比率，我們定義：若某銀行的兩個比率均滿足，則視為資本充足（sufficient-capital，以下簡稱SC）銀行；若某銀行的兩個比率有任一者不滿足，則歸類為兩類資本不足（deficient-capital，以下簡稱DC）銀行。我們提出：SC銀行符合「資本緩衝成本效果」（buffer cost effect），且在創造超額流動性之後會降低其資本；DC銀行則符合「流動性不足風險效果」（illiquidity risk effect），且在創造超額流動性後將提高其資本。
利用1996至2016年36個經濟合作暨發展組織（Organization for Economic Cooperation and Development，以下簡稱OECD）國的18,247家商業銀行為樣本，本文證實資本適足性確實會影響資本與LC之間的交互關係。對於由資本向LC的連結而言，我們發現：弱（強）資本適足銀行的風險意識較不足（充足）、會創造較多（較少）流動性、且顯示負向（正向）的資本與LC關係。對於由LC向資本變動的連結，我們發現：SC銀行比較在乎由較高資本緩衝所帶來的成本，也因此在創造超額流動性之後會降低其資本；DC銀行則將避免流動性不足的風險視為其優先目標，因此在創造超額流動性之後會提高其資本。此外，為顯著提高其資本比率，SC與DC銀行會分別增加第二類（Tier 2）與第一類（Tier 1）資本，以因應超額的流動性創造。
本文的發現不僅證實資本適足性會透過影響銀行的風險意識與資本決策，進而影響資本與LC的交互關係，也同時呼應了巴塞爾資本協定三的精神。藉由引進與傳統CAR互補的簡單LR，巴塞爾資本協定三可降低諸如流動性錯配（liquidity mismatch）等超額風險行為，以有效確保銀行的穩健度以及金融穩定。
Liquidity creation (LC) is one of the most important functions of banks. Converting their short-term, liquid liabilities into long-term, illiquid assets, banks can provide liquidity for non-bank sectors in an economy. Since capital and liquidity are critical factors that ensure the soundness of banks, they are highly regulated by Basel III and their interrelationship has been widely discussed in literature after the 2008 financial crisis.
However, the existing studies have not reached a consensus on the factors that influence capital and LC and hence have failed to develop a reasonable connection between them. This motivates this dissertation. We propose that capital adequacy, which has been ignored in much literature, may be a pivotal factor that affects the interrelationship between capital and LC. As the capital and LC are closely related in practice, we plan to analyze their relationship from two opposite directions.
The first chapter of this dissertation investigates whether capital adequacy affects the link from capital to LC. We define our capital adequacy and two types of banks in a novel way by considering the total capital adequacy ratio (CAR) and the leverage ratio (LR) that follow the framework in Basel III. We propose that weak capital adequate banks are less aware of risk, which adheres to the financial fragility/crowding-out hypothesis, while strong capital adequate banks are prudently aware of risk, which follows the risk-absorption hypothesis. The former tends to create more liquidity and has a negative correlation between capital and LC, whereas the reverse is true for the latter.
The second chapter examines how capital adequacy affects the link from LC to the
change of capital. Combining CAR and LR, we define a sufficient-capital (SC) bank that is well-capitalized in both capital ratios, and two types of deficient-capital (DC) banks that are not well-capitalized in either CAR or LR. We propose that SC banks follow the buffer cost effect and will decrease their capital after conducting excessive LC (ELC), whereas DC banks adhere to illiquidity risk effect and will increase their capital after conducting ELC.
The data contain 18,247 commercial banks from 36 Organization for Economic Cooperation and Development (OECD) countries spanning 1996 to 2016. This dissertation robustly corroborates that capital adequacy is indeed affecting the interrelationship between capital and LC. For the link from capital to LC, weak (strong) capital adequate banks are found to be less (more) aware of risk, create more (less) liquidity, and show a negative (positive) correlation between capital and LC. For the link from LC to the change of capital, SC banks care more about the costs arising from higher capital buffers and thus decrease their capital after conducting ELC. Conversely, DC banks regard the prevention of illiquidity risk as their priorities and hence increase their capital after conducting ELC. Furthermore, to significantly increase their capital ratios, SC and DC banks raise Tier 2 and Tier 1 capital, respectively, in response to the ELC.
Our findings not only confirm that capital adequacy affects the interrelationship between capital and LC through the impact on banks’ risk awareness and capital decisions, but also echo the spirit of Basel III. Introducing a simple LR that complements the traditional CAR, Basel III can effectively ensure the soundness of banks and financial stability that thereby reduces excessive risk activities such as liquidity mismatches. |
