Research

While pre-tax income in the middle and top of the distribution is relatively stable over the business cycle, income at the bottom is strongly procyclical. As a result, income inequality is countercyclical. Since bank credit supply has been shown to play an important role over the business cycle, this paper asks: Do shifts in the supply of bank credit contribute to the observed cyclical fluctuations in income inequality? Using an estimated bank credit supply shock based on Bassett et al. (2014), I find that, first, a tightening of bank lending standards leads to a rise in income inequality due to a larger decrease in income at the bottom of the distribution relative to the middle and top. Second, compared to total factor income (TFI), labor income at the bottom of the distribution contracts more following the tightening of bank lending standards. As a result, labor income inequality rises by more than TFI inequality. Third, bank credit supply shocks account for a nontrivial amount of the cyclical variation of income inequality. Depending on the specification and inequality measure, bank credit supply shocks can account for up to 42% and 33% of the volatility of labor income inequality and TFI inequality, respectively. To rationalize these observations, I build a tractable heterogeneous agent New Keynesian (THANK) model with an income-skill complementarity among households: i.e., ‘poor’ hand-to-mouth households supply unskilled labor, while ‘rich’ unconstrained households supply skilled labor. Firms require a bank loan to pay their labor wage bills in advance of production. When the representative bank reduces its supply of credit to firms, the cost of borrowing rises, leading to a fall in labor demand of both skilled and unskilled workers. However, since it is costly for firms to adjust their skilled labor, firms optimally choose to cut their unskilled workforce by more. The result is a rise of income inequality as ‘poor’ households lose a larger share of income relative to the ‘rich’ households.

There is a strong negative relationship between US bank credit and the unemployment rate since the mid 1980s. In this paper I ask: Is there a causal connection between changes in the supply of bank credit and the observed fluctuations of the unemployment rate? If so, what are the transmission mechanisms governing such a relationship? To answer these questions, I proceed down two paths. First, I use a bank credit supply measure estimated by Bassett et al. (2014) to document the effect that bank credit supply shocks have on the unemployment rate in a VAR model. The results suggest that a one percent contraction in the supply of bank credit leads to a 0.3 percentage point increase in the unemployment rate. Moreover, bank credit supply shocks account for about 30% of the volatility in the unemployment rate. Second, I rationalize these results by incorporating a banking sector into an otherwise standard DSGE model with labour search frictions and nominal rigidities. Unlike standard banking models—which restrict banks to intermediaries of preaccumulated loanable funds—this paper allows banks to finance loans through deposit (i.e., money) creation as described in McLeay et al. (2014) and Bundesbank (2017). Matched firms must obtain a bank loan to purchase their inputs before production. When banks contract credit, there is less funds for firms to purchase their capital and labour inputs and the cost of borrowing rises. This lowers the firms’ benefit of matching thereby lowering the benefits of posting a vacancy. As a result, less vacancies are posted and labour market tightness falls. In both empirical and model settings, a contraction in bank credit supply leads to a typical recession: GDP, investment and labour market tightness fall while the cost of external finance and the unemployment rate rise.

When a bank issues a loan it creates new deposits in the borrower's account of equal value to the loan. I refer to this as financing through money creation (FMC). The vast majority of business cycle literature model banks as intermediaries of pre-accumulated funds or resources. In contrast, this paper asks: Are unexpected disruptions to the level of banks' FMC a quantitatively important source of economic fluctuations?  To address this question I incorporate a novel fractional reserve-type banking sector into a standard DSGE model with nominal and financial frictions. The bank accumulates net worth and is endowed with a FMC technology that allows it to endogenously and exogenously---via a FMC shock---vary its desired reserve ratios, and therefore, varying the quantity of 'inside money' it creates ex nihilo to finance loans to non-bank firms. I use US data from 1989:Q1-2019:Q4 to estimate the model. I find that first, an unexpected reduction in the level of FMC generates a model response consistent with the financial crises of 2007-09: (i) while the level of FMC contracts, more non-FMC sources of funding are lent but the end result is a reduction in external finance, (ii) the cost of all types of external finance rises and (iii) output and investment both fall; second, FMC shocks account for the majority (55%) of the volatility in output and investment; and third, the model is able to replicate two stylized facts Adrian et al. (2013) list as obstacles for the macroeconomic literature: (i) financial institution net worth is `sticky' and (ii) financial institution leverage is procyclical. 

This paper studies the short run macroeconomic effects of a bank credit supply shock on the Canadian economy. While this has been attempted before, there remains outstanding endogeneity issues in previous attempts. To identify a credit supply shock this paper use balance-of-opinion data from the Senior Loan Officer Survey and Business Outlook Survey that is purged of macroeconomic and banking sector factors that may also affect loan demand. The estimated credit supply shock---representing changes in banks' lending standards---is included in a VAR model. I find that a tightening of bank lending standards leads to an economic contraction. However, I find that bank credit supply shocks account for a relatively small share in overall economic volatility. While the results are robust to variety of tests, I find that using a Bayesian VAR model with sign restrictions to identify bank credit supply shocks suggests that the benchmark results may be biased downward.

This paper provides a brief historical exposition of the development of banking in Canada and lists six stylized facts of the Canadian Banking system using detailed bank-level filings. The stylized facts are: first, it is dominated by five large banks; second, credit is deepening; third, net worth is ‘sticky’; fourth, leverage is procyclical; fifth, balance sheet risk is falling; and sixth, rising bank output is driven by liquid liabilities. These documented facts can help researchers make headway along two important dimensions: first, in the macro-banking modelling domain and, second, in understanding potential sources of stability of the Canadian banking sector.

This note is meant to help researchers add a banking sector that is consistent with the financing through money creation (FMC) view into a standard New Keynesian DSGE model.

This paper studies how depositor withdraw risk influence a bank's decision to extend loans when the bank finances its loans through money (i.e., deposit) creation.