ساختار مالی و تاثیر ناهمگن سیاست پولی در سراسر صنایع

The two general channels by which monetary policy impacts output are the neo-classical cost of capital channel and the credit channel. This paper decomposes the output response to a change in the stance of monetary policy to each of these channels. We use a unique industry level data set that measures the financial characteristics of firms operating at the industry level through time. We bring these financial characteristics formally into the regression analysis, thus allowing for a more precise identification of the two channels. The evidence indicates that both channels are active in the Canadian economy.

In textbook applications, it is implicitly assumed that the impact of monetary policy is uniform across the entire economy—that is, each industrial sector is assumed to be impacted to the same degree. As such, the heterogeneous impact of monetary policy across an economy is ignored. Although there have been many studies that consider this heterogeneity across the regions of an economy, there are few studies which consider the heterogeneity across industries. This paper works to fill this void. To the extent there is heterogeneity in the effects of monetary policy across both industrial sectors and across regions, these differences should be considered in the formation of optimal monetary policies. Hence studies which are able to predict this heterogeneity are very useful. Studies that measure this heterogeneity typically consider the sensitivity of regional output to changes in monetary policy for the following factors: (1) the mix of interest sensitive sectors; (2) the mix of large versus small firms; (3) the ability and willingness of banks to supply loans; and (4) differences in financial structure. The first captures the neo-classical interest rate channel whereas the remaining are meant to capture the credit channel of monetary policy. It is argued, for example, that monetary policy will have a large impact in regions where there are many small firms as these firms are more reliant on bank financing. In other words, small firms are impacted more from changes in interest rates (monetary policy). Gertler and Gilchrist (1994) (henceforth GG (1994)) measure the response of firms to changes in monetary policy (for example, a movement to tight money). They use firm size to proxy for capital market access, and as such are able to use a comprehensive dataset for the manufacturing sector, although not at the firm level. The idea is that the cyclical behavior of firms is in part linked to their access to capital markets, and how that access varies over the business cycle. The strategy pursued by GG (1994) is to trace the impact of monetary policy on the time series behavior of small versus large firms. They find that in reaction to a monetary contraction, large firms borrow to finance inventories, whereas small firms shed inventories relatively quickly. In addition to this inventory effect, they show that sales and short-term debt change significantly in response to a change in interest rates, with the effects being larger for small firms. As such, small firms account for a disproportionate share of the reduction in output. They go on to suggest that further studies on cyclically sensitive sectors such as retail and wholesale trade and construction would be very useful. The current paper does exactly this: we measure the impact that monetary policy has at the industry level, which include retail and wholesale trade as well as many other industries. The stance of monetary policy is measured using the term spread, defined as the difference between the 10-year T-bond yield and the 3-month T-bill yield, as well as the overnight rate, which is the Bank of Canada's key policy rate. More importantly, however, this paper employs a unique data set on the financial conditions of firms across industries, conditions which serve to capture the balance sheet and bank lending linkages of the credit channel, as described in Bernanke and Gertler (1995). This channel is also known as the financial accelerator effect—that is, how changes in a firm's financial characteristics over the business cycle influence the effects of monetary policy. Measures of financial conditions used include measures of liquidity, inventories, coverage ratios, bank borrowings, debt equity ratios, and firm size. These data are available at the industry level and through time. We are not aware of a study that uses such precise data to capture the amplifying effect of the credit channel from a monetary policy change. The current paper has two objectives. First, we demonstrate that there is indeed heterogeneity in the impact that monetary policy has across industries, and find that manufacturing is the most interest sensitive industry. This therefore confirms the assumption made in many papers which hypothesize that the larger the share of interest sensitive sectors in a region – namely manufacturing – the larger will be the impact of monetary policy. We also document significant correlations between the magnitude of the predicted monetary policy impact at the industry level and the average values of the financial variables at the industry level. For example, it is demonstrated that the higher the average level of bank dependence in the firm's financing at the industry level the larger is the impact of monetary policy for that industry. Furthermore, the larger are firms in the industry, the smaller is the impact of monetary policy. These results hold for broad industry categories as well as within manufacturing. Second, we measure the impact of monetary policy at the industry level, and interact our indicators of monetary policy with the financial characteristics of the firms in that industry. This allows us to measure the financial accelerator effect. The results clearly show that changes in the financial characteristics of firms, at the industry level, do in fact influence the impact that monetary policy has on that industry's output. That is, as the financial characteristics of firms in the industry change over the business cycle, these changes can serve to reinforce the impacts of monetary policy or offset them. These results therefore have important implications for monetary policy, both in terms of measuring how its impact will vary across industries, as well how these effects are themselves a function of how the financial characteristics of the firms being impacted change in the aftermath of a change in monetary policy. Key insights are therefore extracted regarding the channels by which monetary policy affects firms, and ultimately output, at the industry level. One explanation often used to explain why some industries are more sensitive to contractions in monetary policy relates to the mean level of that industry's financial characteristic. For example, and as demonstrated below, it is often argued that firms with high levels of bank credit will be more adversely affected by the rise in interest rates that follows a monetary contraction. But additional evidence presented here indicates that although there is a systematic relationship between the magnitude of the impact of monetary policy and the average values of the financial characteristics at the industry level, there is no systematic relationship between the mean level of those financial characteristics and the accelerator effect of monetary policy. Rather, the financial accelerator effects are related to how balance sheets and income statements change over the business cycle, rather than the level of these financial characteristics at any point in time. This is the first study to our knowledge that has documented this. Overall the results of this analysis indicate that monetary policy within Canada impacts the economy both through the neo-classical (interest rate) cost of capital channel as well as the credit channel. That is, changes in balance sheet and income statement characteristics are systematically related to changes in output, after controlling for the change in the stance of monetary policy. The financial characteristics used include measures of liquidity, inventories, coverage ratios, bank borrowings, debt equity ratios, and firm size. As such, the evidence demonstrates clearly that the credit channel is an important one within Canada, and hence these results provide important insights that can enhance the creation and implementation of monetary policy. This paper is organized as follows. Section 2 provides a literature review. Section 3 develops the hypotheses to be tested. Section 4 provides a description of the data. Section 5 contains the empirical evidence. Section 6 concludes.

There are two general channels by which monetary policy impacts the economy: the neo-classical cost of capital channel and the credit channel. The credit channel can be broken down into the bank lending channel and the balance sheet channel. There are many studies that measure the heterogeneity in these impacts across regions. These studies relate that heterogeneity to the financial characteristics of firms and banks operating in each region, as well as the prevalence of interest sensitive industries such as manufacturing. The current study differs from these studies in several ways. First, the current study considers the heterogeneity in the impact of monetary policy across industries. Second, a unique industry-level database is used that provides information on the financial characteristics of firms operating in that industry on a time-series basis. This allows for the financial characteristics to enter directly into regressions that measure the channels by which monetary policy impacts the economy, and hence the ability to obtain more precise inferences about the importance of each of the channels by which monetary policy impacts the economy. The analysis presented here indicates that both the balance sheet and bank lending channels are very active in terms of influencing the impact that monetary policy has on the economy. The major results of the paper can be summarized as follows. First, the impact that monetary policy will have on an industry is a function of both the mean difference in the financial characteristics of firms across industries, as well as the time-series variation in the financial characteristics of firms operating in that industry. Second, although the net impact of monetary policy at the industry level is related to the average value of the industry's financial characteristics, there is no relationship between the average value of an industry's financial characteristics and the strength of the credit channel. In other words, the credit channel is related to the time series variation in balance sheets and income statements as opposed to the financial position of the industry at any point in time. Putting these two results together indicates that the mean difference in the financial characteristics of industries is systematically related to the strength of neo-classical cost of capital channel but not to the credit channel. This is the first study, to our knowledge, to document this result. Since the impact of monetary policy is found to be systematically related to the changes in the financial characteristics of firms, these results will be useful to the monetary authority and others that are anticipating what impact a change in the stance of monetary policy is likely to have on the economy.