قدرت نفوذ، اندازه ترازنامه و منابع مالی عمده فروشی

Positive co-movements in bank leverage and assets are associated with leverage procyclicality. As wholesale funding allows banks to quickly adjust leverage, banks with wholesale funding are expected to exhibit higher leverage procyclicality. Using Canadian data, we analyze (i) if leverage procyclicality exists and its dependence on wholesale funding, (ii) market factors associated with this procyclicality, and (iii) if banking-sector leverage procyclicality forecasts market volatility. The findings suggest that procyclicality exists and that its degree positively depends on use of wholesale funding. Furthermore, funding-market liquidity matters for this procyclicality. Finally, banking-sector leverage procyclicality can forecast volatility in the equity market.

In the aftermath of the recent financial crisis, the high levels of leverage among financial institutions has widely been identified as one of the major causes of the crisis. This has focused attention on both how financial institutions manage their leverage ratios (defined as assets divided by equity) and on potential regulatory actions required to prevent the build up of excessive levels of leverage in the financial sector. Given the nature and severity of the recent financial crisis, leverage has quickly become one of the focal points of both academic research and policy-oriented discussions related to financial stability. While the slow build up of leverage over several years among financial institutions is an important issue, some studies also identify the importance of higher frequency movements of leverage. In this regard, one major question has been raised: Does leverage positively co-move with assets, and if so, what are its implications for market volatility? Adrian and Shin (2010) study this relationship between financial institution leverage and assets in the United States and find evidence that such a correlation exists. One channel in which this positive correlation can be observed is when a financial institution actively manages its balance sheet with respect to changes in the value of equity. For example, when the value of equity increases due to a rise in prices of some marked-to-market assets, a financial institution’s leverage ratio decreases. If the financial institution actively manages its balance sheet, it can raise non-equity liabilities and lever up. In this process, the newly-raised liabilities are invested into new assets, leading to a positive relationship between changes in leverage and balance-sheet size. Furthermore, as prices of assets tend to increase during booms and decrease during busts, leverage becomes procyclical to economic activity in addition to balance-sheet size. This paper highlights the interaction of leverage procyclicality with the use of wholesale funding. The degree of procyclicality varies across different types of financial institutions and with respect to changes in macroeconomic and market environments. Financial institutions that use wholesale funding (e.g., institutional deposits, repos, commercial paper and banker’s acceptances) display high degrees of procyclicality as these market-based funds are readily available at short notice for quick adjustments to leverage. However, the crisis disrupted short-term wholesale funding markets, revealing the high funding-liquidity risks associated with these funds. With reduced access to wholesale funding, financial institutions lost the ability to adjust leverage easily and quickly, which dampened the degree of procyclicality. Specifically, we have three main objectives. First, we show that leverage of Canadian financial institutions is procyclical (i.e. positive correlations between leverage and balance-sheet size) and that the degree of procyclicality depends on the usage of wholesale funding. Second, we identify macroeconomic and market variables that are important for the degree of procyclicality. Third, we study if banking-sector leverage procyclicality can forecast aggregate volatility in the equity market. The empirical strategy chosen to achieve the first two objectives is a two-step method, similar to the approach outlined by Kashyap and Stein (2000) in their work on the bank lending channel of monetary policy. The first step cross-sectionally estimates the degree of leverage procyclicality based on monthly bank-level balance sheet data for all federally chartered deposit taking institutions in Canada over the period 1994–2009. 1 The analysis for the first objective is derived from the outcome of this step. Then, the second step determines if and how the degree of procyclicality changes over time following macroeconomic and market-wide changes. The results from this step are used for the discussion of the second objective. For the third objective, we construct volatility measures from the Toronto Stock Exchange Broad Index to gauge aggregate market volatility. We regress these volatility measures on banking-sector leverage procyclicality. With respect to the first objective, we find strong procyclicality of leverage. In addition, we find significantly higher degrees of procyclicality among financial institutions that use more wholesale funding over those that use less. This confirms the findings by Adrian and Shin (2010) that leverage among US investment banks, who mainly rely on market-based wholesale funding to fund their investment activities, is strongly procyclical. They do not find such leverage procyclicality for commercial banks, which rely less on wholesale funding.2 These results consistently prevail through various robustness checks and model extensions. Secondly, we find that degrees of procyclicality change with liquidity in short-term wholesale funding markets, where funding-market liquidity is measured as changes in the trading volume of repos and the volume of outstanding commercial paper and banker’s acceptances. Specifically, for wholesale funding users, we find that procyclicality is high when the liquidity of these markets is also high. Hence, when these markets become illiquid, wholesale funding users lose the ability to quickly adjust leverage, leading to weaker procyclicality of leverage. This result is consistent with Brunnermeier and Pedersen (2009) who provide a theory linking market liquidity (i.e., the ease with which an asset is traded) with funding liquidity (i.e., the ease with which funds are obtained) through margin requirements for financial intermediaries. Since margin requirements for raising funds (e.g., haircuts on collateral and discounts on bank debts) can increase during downturns, available funds for investment decrease, reducing market liquidity. Such market and funding illiquidity would show up as weaker procyclicality of leverage, as the financial institution’s ability to adjust leverage and investment declines. We observe weaker procyclicality with illiquid market conditions only for those financial institutions that rely on short-term wholesale funding markets. Finally, we find that lagged banking-sector leverage procyclicality forecasts equity market volatility. While this effect is positive and significant during the pre-crisis period, it is insignificant during the crisis period. We interpret this result as the ability for banking-sector leverage procyclicality to forecast overall market volatility during pre-crisis periods. There are, however, multiple other factors that would have contributed to movements in market volatility during the crisis (such as various global and domestic government/central bank interventions), leaving the estimate insignificant. Our paper is related to a few different strands in the existing literature. Regarding wholesale funding of banks, Huang and Ratnovski (2010) analyze a model with a tradeoff between using wholesale funding vs. retail deposits. On one hand, wholesale funding improves efficiency as uninsured wholesale financiers monitor banks. On the other hand, the monitoring incentives of the financiers depend on the available information set, which could lead to inefficient liquidations. This study is similar to ours in spirit, since it also evaluates the decisions and the riskiness of banks under different funding structures (retail deposits vs. wholesale funding). Our study is also related to the literature on the regulation of bank leverage, since banks in Canada face regulatory leverage limits. Blum (2008) provides a theoretical motivation for leverage limits in a world where a supervisor knows that different types of banks (safe and risky) exist, but without knowing the actual risk types of each bank. In such a setting, self-reporting and assessment of risks (in a manner similar to Basel II) is not optimal, since risky banks have an incentive to understate their risks. Blum (2008) shows that having a simple leverage ratio cap along with capital requirements based on banks’ internal risk assessments can result in truthful revelations of banks’ risk levels. Geanakoplos (2010) theoretically analyzes adverse effects of leverage fluctuations in an environment where leverage is determined in equilibrium together with interest rates. The paper shows how leverage cycles damage the economy and argues for regulations to control them. Bordeleau et al. (2009) discuss the historical evolution of regulatory leverage limits in Canada and analyze how large Canadian banks manage leverage with respect to these limits. They find that some large banks maintain a buffer between their leverage and the regulatory limit, implying some flexibility to adjust leverage. Finally, The Committee on the Global Financial System (2009) provides some international policy discussions regarding leverage procyclicality. The rest of the study is as follows: Section 2 presents some basic balance sheet arithmetic to explain the link between asset growth and leverage growth. Section 3 provides a brief discussion of the Canadian banking sector. Section 4 presents the data. Section 5 explains the empirical methodology and Section 6 describes the results for the first two objectives. Section 7 analyzes the relationship between banking-sector leverage procyclicality and market volatility. Section 8 discusses our robustness checks. Finally, Section 9 concludes.

We study the extent of procyclicality of leverage in the Canadian banking sector. The study is motivated by the theory developed by Brunnermeier and Pedersen (2009) and empirically studied in Adrian and Shin (2010) that a link exists between funding liquidity and market liquidity through financial institutions’ balance-sheet management. Our analysis utilizes a variation of the two-step empirical estimation method first proposed by Kashyap and Stein (2000). We use monthly balance sheet data covering 15 years and establish that leverage is highly procyclical among Canadian financial institutions. The degree of procyclicality is higher among banks that are more dependent on wholesale funding, e.g., leverage rises as assets increase. Furthermore, the gap in the degree of procyclicality between high wholesale funding users and the rest of the banking sector has grown larger during the 2000s. We then investigate macroeconomic and market-wide variables associated with leverage procyclicality and its divergence between different wholesale funding groups. The result suggests that leverage becomes more procyclical during times of increased liquidity in repo, BA and CP markets. Finally, we argue that banking-sector leverage procyclicality is important for aggregate economy by providing empirical evidence that banking-sector leverage procyclicality forecasts aggregate market volatility in the equity market. Since procyclicality of leverage could lead to aggregate volatility, current leverage regulations may not adequately address potential consequences of market and funding liquidity risks. Other regulations, such as those being discussed in the Basel Committee on Banking Supervision, that enforce counter-cyclical capital holdings and directly restrict banks’ balance-sheet liquidity-risk management have the potential to address this issue. However, potential costs of such regulations need to be taken into account.