Do big banks have lower operating costs?
* But such limits could have adverse effects if they were to undercut the economies of scale associated with large banking firms.
* Reasoning that scale economies may be achieved in part through lower operating costs, the authors of this study examine the relationship between BHC size and non-interest expense.
* Their analysis, which considers these costs at a finer level of detail than in past studies, reveals a robust negative relationship between BHC size and scaled non-interest expenses, including employee compensation, information technology, and corporate overhead costs.
* The results suggest that limits on BHC size may, in fact, increase the cost of providing banking services--a drawback that must be weighed against the potential financial stability benefits of limiting firm size.
The largest U.S. banking firms have grown significantly over time, their expansion driven by a combination of merger activity and organic growth. In 1991, the four largest U.S. bank holding companies (BHCs) held combined assets equivalent to 9 percent of gross domestic product (GDP). Today, the four largest firms' assets represent 50 percent of GDP, and six BHCs control assets exceeding 4 percent of GDP. Despite recent financial reforms, there is still widespread concern that large banking firms remain "too big to fail"--that is, policymakers would be reluctant to permit the failure of one or more of the largest firms because of fears about contagion or damage to the broader economy (see, for example, Bernanke ).
A growing number of market observers advocate shrinking the size of the largest banking firms in order to limit the problem of too-big-to-fail. The most direct approach would be to simply impose a firm cap on the size of assets or liabilities; for example, Johnson and Kwak (2010) propose a size limit of 4 percent of nominal GDP. An alternative would be to impose levies or progressively higher capital requirements on large banking firms to encourage them to shed assets.
Would such policies impose any real costs on the economy? A number of recent academic papers suggest that the answer may be "yes" because of the presence of economies of scale in banking. Scale economies imply that the cost of producing an additional unit of output (for example, a loan) falls as the quantity of production increases. A number of papers find evidence of scale economies even among the largest banking firms (Hughes and Mester 2013; Wheelock and Wilson 2012; Feng and Serletis 2010). Taken at face value, this research implies that the introduction of limits on bank size would impose deadweight economic costs by increasing the cost of providing banking services.
We contribute to this line of research by studying the relationship between size and components of non-interest expense (NIE), with the goal of shedding light on the sources of scale economies in banking. NIE includes a wide variety of operating costs incurred by banking firms: examples include employee compensation and benefits, information technology, legal fees, consulting, postage and stationery, directors' fees, and expenses associated with buildings and other fixed assets. Our hypothesis is that lower operating costs may be a source of scale economies for large BHCs, because large firms can spread overhead such as information technology, accounting, advertising, and management over a larger asset or revenue base.
Our analysis therefore tests for an inverse relationship between BHC size and scaled measures of different components of NIE.
One novel contribution of this paper is to make use of detailed non-interest expense information provided by U.S. banking firms in the memoranda of their quarterly regulatory FR Y-9C filings. The Y-9C reports contain detailed consolidated financial statements and other data for U.S. BHCs (see Section 3 for details). Since 2001, about 35 percent of total non-interest expense is classified in the Y-9C as part of a broad "other non-interest expense" category. For the period 2008 to 2012, we disaggregate this line item into nine author-defined categories, using memoranda information from Schedule HI of the Y-9C.
In part, this involved manually classifying about 5,500 individual "write-in" text fields reported by individual BHCs. To our knowledge, ours is the first paper to make use of these data.
We start by estimating the relationship between bank holding company size (measured by the natural logarithm of total assets) and total non-interest expense scaled by net operating revenue, assets, or risk-weighted assets. We find a statistically and economically significant negative relationship between BHC size and these NIE ratios, robust to the expense measure or set of controls used. Quantitatively, a 10 percent increase in assets is associated with a 0.3 to 0.6 percent decline in non-interest expense scaled by income or assets, depending on the specification. In dollar terms, our estimates imply that for a BHC of mean size, an additional $1 billion in assets reduces non-interest expense by $1 million to $2 million per year, relative to a base case in which operating cost ratios are unrelated to size. (1)
These results hold across the size distribution of banking firms, and over different parts of our sample period. We find no evidence that these lower operating costs flatten out above some particular size threshold. The point estimate of the slope of the relationship steepens, if anything, although the statistical uncertainty associated with the estimate becomes larger owing to the small sample.
The relationship between size and the NIE ratio is negative for each of the three main components of non-interest expense reported in BHC regulatory filings: employee compensation, premises and fixed asset expenses, and other non-interest expense. Using our novel by-hand classification of other NIE into nine subcomponents, however, we find significant variation in the size-expense relationship among the subcomponents. The inverse relationship between size and expense is particularly pronounced for corporate overhead (for example, accounting, printing, and postage); information technology (IT) and data processing; legal fees; other financial services; and directors' fees and other compensation. In contrast, large BHCs spend proportionately more on consulting and advisory services than do smaller firms, relative to revenue or assets. Large BHCs also incur proportionately higher expenses relating to amortization and impairment of goodwill and other intangible assets.
Overall, our results are consistent with the presence of scale economies in banking, as found in recent academic literature (for example, Wheelock and Wilson ; Hughes and Mester ; Feng and Serletis ) and industry research (Clearing House Association 2011). In particular, our findings suggest that these scale economies stem in part from an operating cost advantage of large BHCs in areas such as employee compensation, information technology, and corporate overhead expenses.
We emphasize that a number of caveats apply to our results. First, our estimates represent reduced-form statistical correlations; caution should be exercised in drawing a causal interpretation from them. Although our regressions control for a wide range of BHC characteristics, firm size may still be correlated with omitted variables that are also associated with lower expenses, such as the quality of management. This caveat also seems to apply more generally to the existing literature on scale economies in banking.
Second, our results may also reflect factors other than scale economies. One possibility, closely related to scale economies but conceptually distinct, is that large firms operate closer to their production frontier on average; that is, they have greater X-efficiency (see Section 2 for a discussion). (2) Another possibility is that large banking firms have greater bargaining power vis-a-vis their suppliers and employees. If cost differences are due only to bargaining power effects, then limiting the size of the largest BHCs would not necessarily generate deadweight economic costs, although it might instead reallocate rents to employees or suppliers. An additional possibility is that our results are influenced by too-big-to-fail subsidies for large BHCs. Our prior is that such subsidies would be more likely to be manifested as a lower cost of funds for large firms, or a more leveraged capital structure, than as lower operating costs. However, it is still possible that a too-big-to-fail banking firm could respond by reducing expenditures on functions such as information technology or risk management; these would show up as part of non-interest expense.
These caveats aside, our results and those of related research suggest that imposing size limits on banking firms is unlikely to be a free lunch. For example, taking our estimates at face value, a back-of-the-envelope calculation implies that limiting BHC size to no more than 4 percent of GDP would increase total industry non-interest expense by $2 billion to $4 billion per quarter. (3) Limiting the size of banking firms could still be an appropriate policy goal, but only if the benefits of doing so exceeded the attendant reductions in scale efficiencies.
A second contribution of this article is to present new evidence on other determinants of BHC operating costs. In particular, we find that proxies for organizational complexity (for example, the number of distinct legal entities controlled by the BHC), as well as measures of the diversity of business activities, are robustly correlated with higher expense ratios. This result appears consistent with prior research on the diversification discount in banking (for example, Goetz, Laeven, and Levine ). A third contribution is to present new stylized facts about the composition of non-interest expense, based on our data collection efforts. For example, we document the large share of NIE that is composed of corporate overhead, investment technology and data processing, consulting and advisory services, and legal expenses.
The remainder of the article proceeds as follows: Section 2 presents background and reviews the literature on economies of scale in banking. Section 3 describes the data, discusses our method for classifying other non-interest expense, and presents descriptive statistics. Section 4 presents multivariate analysis of the relationship between size and non-interest expense ratios. Section 5 studies components of non-interest expense. Section 6 summarizes our findings.
2. BACKGROUND AND LITERATURE
Our analysis is closely related to academic literature on scale economies and organizational efficiency in banking. In a microeconomic production model, the cost function traces out the relationship between output and the minimum total cost required to produce that output, for a given set of input prices. A firm exhibits economies of scale if minimum cost increases less than proportionately with output--for example, if the firm could double its output by less than doubling its costs, holding input prices fixed.
A large literature empirically estimates the cost function for banks and/or BHCs, and tests for the presence of scale economies by measuring whether the elasticity of total costs with respect to output is greater than, equal to, or less than unity (indicating diseconomies of scale, constant returns to scale, or economies of scale, respectively).
The earliest studies of scale economies in banking (for example, Benston ), estimated during an era when U.S. banking organizations were on average much smaller than today, found evidence of modest economies of scale. Subsequent research, using more flexible cost functions, found that these scale economies were limited to small banks (for example, Benston, Hanweck, and Humphrey  and Peristiani ; see also Berger and Humphrey  for a survey).
More recent research, however, has found evidence of scale economies even among the class of large banks and bank holding companies. Examples include Wheelock and Wilson (2012), Hughes and Mester (2013), Feng and Serletis (2010), and Hughes et al. (2001). This departure from earlier findings reflects greater statistical power, attributable to the use of larger datasets with many more observations for large banking firms, as well as the evolution of empirical techniques. For example, Wheelock and Wilson (2012) estimate a nonparametric cost function rather than the typical parametric translog function estimated in earlier literature, while Hughes and Mester (2013) and Hughes et al. (2001) endogenize bank risk and capital structure decisions. The difference in time periods may also play a role (for example, the greater use of information technology may have changed the extent to which scale economies are present).
The theoretical derivation of the cost function assumes that the bank maximizes profits, or equivalently, minimizes costs for any given level of output. A related body of literature on bank efficiency, however, finds evidence of surprisingly large cost differences between otherwise similar banks. These differences are viewed as evidence of X-inefficiencies, that is, firms operating inside their production possibilities frontier because of agency conflicts, management problems, or other inefficiencies (DeYoung 1998; Berger, Hunter, and Timme 1993; Berger and Humphrey 1991).
Rather than analyzing total scale economies or X-efficiency, this paper instead presents disaggregated evidence on the relationship between firm size and detailed components of non-interest expense. We have in mind the idea that operational and technological efficiencies related to size are likely to show up in the data in the form of lower operating costs in areas such as information technology and corporate overhead (for example, accounting and human resources) because large BHCs are able to spread the fixed component of these costs over a broader revenue or asset base. Our goal is to shed additional light on the mechanisms driving differences in efficiency between small and large firms. We note that our empirical finding that large BHCs have lower average operating costs could be driven by the presence of scale economies in the production of banking services, higher average X-efficiency for large firms, or both. For some categories of NIE, it could also be possible that lower costs for larger banking firms not only reflect technological efficiencies, but also greater bargaining power relative to suppliers, customers, or employees.
Our analysis is related to recent research by the Clearing House (2011) that uses proprietary management information systems data from a number of large banks to estimate product-specific scale curves in seven areas: online bill payment, debit cards, credit cards, wire transfers, automated clearing house, check processing, and trade processing. The Clearing House finds that in each of these areas, unit costs are decreasing in production volume, a conclusion that suggests the presence of fixed costs or other technological benefits of size. The economies of scale associated with these seven services are estimated to total $10 billion to $25 billion per year.
Although our approach is similar in some respects to the analysis by the Clearing House, we make use of data from audited regulatory filings, rather than internal management information system data, and study components that together sum up to total non-interest expense, rather than just a subset of NIE (the seven items studied by the Clearing House together cover only 7 to 10 percent of NIE). We also study the entire cross-section of BHCs, while the Clearing House sample consists of only six firms.
Our approach is related to the literature on banking mergers that uses accounting variables to estimate the effects of mergers on operating performance. Kwan and Wilcox (2002) find evidence that bank mergers reduced operating costs, although more so for the early 1990s than the late 1980s. Cornett, McNutt, and Tehranian (2006) examine different measures of efficiency improvements for large mergers, and find evidence for cost-efficiency improvements in addition to other revenue improvements. Hannan and Pilloff (2006) show that cost-efficient banks tend to acquire relatively inefficient targets. Using German banking data, Niepmann (2013) finds a negative correlation between size and scaled operating costs--a result consistent with our findings for U.S. firms.
Davies and Tracey (2014) argue that standard estimates of scale economies for large banks are influenced by too-big-to-fail (TBTF) subsidies, and that scale economies are no longer present after controlling for TBTF factors. Hughes and Mester (2013) dispute this conclusion, arguing that the cost function used by Davies and Tracey is misspecified. One potential advantage of our focus on non-interest expense is that operating costs (for example, information technology, printing, postage, and advertising) may be relatively more likely to reflect technological features of the firm's production process than any distortions due to TBTF. Instead, TBTF seems most likely to affect the firm's funding costs and capital structure. It seems difficult, however, to rule out the possibility that TBTF subsidies may affect our results or those of previous literature.
3. DATA AND DESCRIPTIVE STATISTICS
Our analysis is based on quarterly FR Y-9C regulatory data filed by U.S. bank holding companies. The Y-9C filings include detailed balance sheet and income data, as well as information about loan performance, derivatives, off-balance-sheet activities, and other aspects of BHC operations. Data are reported on a consolidated basis, incorporating both bank and nonbank subsidiaries controlled by the BHC (see Avraham, Selvaggi, and Vickery  for more details). Our analysis considers only "top-tier" BHCs--that is, the ultimate parent U.S. entity. Our sample includes top-tier U.S. BHCs with a foreign parent, although it excludes "stand-alone" commercial banks that are not owned by a BHC, and BHCs that are too small to file the Y-9C (the Y-9C reporting threshold varies over time, but is currently $500 million). Our sample excludes investment banks, thrifts, and other types of financial institutions, unless those firms are owned by a commercial BHC.
Noninterest expense is reported in the consolidated Y-9C income statement (Schedule HI), broken down into five categories. Note that non-interest expense does not include loan losses due to defaults, trading losses, gains and losses on owned securities, or taxes; these are recorded in other parts of the income statement. (4) Our analysis focuses on non-interest expense because it is the most likely area in which firms would realize operating cost advantages from size.
We compute several normalized measures of non-interest expense. The first measure, widely used by practitioners and industry analysts, is the "efficiency ratio," defined as the ratio of non-interest expense to "net operating revenue" the sum of net interest income and non-interest income:
Efficiency ratio = non-interest expense/net interest income + non-interest income
A higher efficiency ratio indicates higher expenses, or equivalently, lower efficiency. Effectively, this ratio measures the operating cost incurred to earn each dollar of revenue. Efficiency ratios vary widely across BHCs, as we document below, but typical values range from 50 to 80 percent. Efficiency ratios are sometimes computed excluding certain noncash items from non-interest expense, such as amortization of intangible assets. We refer to such measures as "cash" efficiency ratios.
One limitation of the efficiency ratio is that it is sensitive to quarter-to-quarter movements in net operating revenue. For example, ratios spiked for many BHCs during the financial crisis, because of trading losses and other non-interest losses. (In rare cases, the efficiency ratio even flips sign, because the sum of net interest and non-interest income is negative.) To provide an alternative normalization that is less sensitive to these concerns, we also present results based on scaling non-interest expense by total assets or risk-weighted assets (RWA), rather than net operating revenue:
Expense asset ratio = non-interest expense/total assets (or risk-weighted assets)
These normalizations can be computed for total non-interest expense, or for NIE subcomponents such as compensation.
3.1 Descriptive Statistics
Table 1 presents descriptive statistics for non-interest expense over the period from first-quarter 2001 to fourth-quarter 2012. We selected this period to take advantage of additional detail on non-interest income expense that was added to the Y-9C in 2001, thereby allowing us to separate non-interest income (which we use as a control) into components such as investment banking fees, income from insurance fees, deposit fees, and servicing fees. Note that the sample period for our regression analysis in Section 4 begins in first-quarter 2002 because we incorporate lagged income variables from the previous four quarters. A total of 2,810 BHCs are present in the sample for at least one quarter.
Panel A of the table reports summary statistics for four normalized measures of non-interest expense: the efficiency ratio, the cash efficiency ratio (which excludes goodwill impairment and amortization from non-interest expense), non-interest expense scaled by total assets, and non-interest expense scaled by RWA. The industry efficiency ratio averages 66.3 percent over 2001-12, although it is somewhat higher (71.7 percent) in 2012. The standard efficiency ratio and the cash efficiency ratio differ little on average, reflecting the fact that goodwill impairment and amortization expense generally represent a small total of total non-interest expense.
The distribution of the expense ratios is skewed to the right. For example, the difference between the 5th percentile of the efficiency ratio and its median is 19.5 percent, significantly smaller than the difference of 28.0 percent between the median and the 95th percentile value. Furthermore, the right tail includes some extremely high values (for example, the 99.5th percentile is 198.4 percent), likely driven by one-time spikes in revenue. To reduce the influence of outliers, our regression analysis winsorizes the top and bottom 0.5 percent of observations for each non-interest expense ratio (all data below and above the bottom and top 0.5th percentiles, respectively, are set equal to the 0.5th and 99.5th percentiles).
We examine the components of non-interest expense in Panel B of the table, based on the five non-interest expense categories reported on Schedule HI. (5)
* Compensation (49.4 percent of industry total over the sample time period, reported on FR Y-9C as "salaries and employee benefits"). This category includes wages and salaries, bonus compensation, contributions to social security, retirement plans, health insurance, employee dining rooms, and other components of employee compensation.
* Premises and fixed assets (11.6 percent of total, reported on Y-9C as "expenses of premises and fixed assets net of rental income") includes depreciation, lease payments, repairs, insurance and taxes on premises, equipment, furniture, and fixtures. It excludes mortgage interest on corporate real estate.
* Goodwill impairment (1.8 percent of total, reported on Y-9C as "goodwill impairment losses") represents losses incurred when goodwill exceeds implied fair value and is revalued downwards. This item is reported separately from "other non-interest expense" from 2002 onwards.
* Amortization expense (1.9 percent of total, reported on Y-9C as "amortization expense and impairment losses for other intangible assets") includes amortization of goodwill and other intangible assets owned by the BHC, as well as impairment losses for intangible assets other than goodwill. This item is also available from 2002 onwards.
* Other (35.0 percent of total) includes a broad range of other operating costs, such as telecommunication and information technology costs, legal fees, deposit insurance, advertising, printing, postage, and so on. Additional information on these expenses is provided in the memoranda to Schedule HI, as we explain in detail below.
Chart 1 plots the time series evolution of the four normalized measures of total industry NIE. Each expense measure declined between 2001 and mid-2007, a period when the revenues and assets of the banking system grew rapidly. For example, the industry efficiency ratio fell from 65.4 percent in quarter-one 2001 to 58.8 percent in quarter-two 2007, while the expense asset ratio declined from 0.96 percent to 0.72 percent over the same period. This downward trend was reversed during the 2007-09 financial crisis. Since the efficiency ratio is mechanically inversely related to net operating revenue, the reversal for that NIE measure is perhaps unsurprising. However, the expense asset ratio also increased, whether normalized by total assets or risk-weighted assets. In recent years non-interest expense ratios have stabilized at levels higher than those prevailing prior to the onset of the crisis. The rise in the efficiency ratio in part simply reflects the decline in net operating revenue and measures of profitability for the banking industry, owing to compression of net interest margins and lower non-interest income.
Appendix B also plots the evolution of the relative shares of the five non-interest expense subcategories. (6) Goodwill impairment expenses are almost entirely concentrated in 2008, with negligible levels for this expense category before and after 2008. Other non-interest expense makes up a progressively larger fraction of total NIE over the past five years. (In 2012, this category represented 39.9 percent of total NIE, a share similar to that held by compensation expenses).
As a first look at the relationship between firm size and normalized non-interest expense, the main focus of this paper, we present scatter plots of BHC size and the efficiency ratio (Chart 2). The plots are based on year-to-date 2012 expense data and assets as of the end of 2012. A striking feature of the chart is the variability in non-interest expense across firms, particularly among smaller BHCs. This finding is also borne out in our multivariate analysis in Section 4. The variability points to the importance of adding controls for those observable differences in BHCs' activities that are associated with different types of expenses. These controls are described in Section 3.3.
3.2 Classifying Other Noninterest Expense
The category "other NIE" represents more than one-third of industry non-interest expenses since 2001. To shed light on these costs, we examine data from the memoranda to Schedule HI. Since 2008, Schedule HI has allowed BHCs to
classify other NIE into eleven standardized subcategories; (7) in addition, space is provided for BHCs to report additional "write-in" expense items that were not captured by the standardized fields. For the eleven standardized subcategories, BHCs are instructed to record items for amounts greater than $25,000 that also exceed 3 percent of total other non-interest expense. Write-in items bear the additional requirement that the expense item exceed 10 percent of total other non-interest expense. Since 2008, amounts in the eleven standardized categories have made up 38 percent of total other non-interest expense, while the write-in fields have constituted another 28 percent of other NIE. The remaining 34 percent of other non-interest expense is not reported in the Schedule HI memoranda, presumably because it does not meet the reporting thresholds described above.
It is particularly challenging to classify and analyze items recorded in the write-in expense fields, because these amounts are reported using nonstandardized language by each BHC. For example, non-interest expenses related to foreclosures and to properties that are "other real estate owned" (8) are variously written in as "reo," "ore," "R.E.O," "oreo," "foreclose," and so on, as well as various misspelled text strings such as "oero" and "forclosuer" (sic). Overall, more than 30,000 text strings are written in by the BHCs in our sample between 2008 and 2012. Approximately 5,500 of these strings are unique. Individual BHCs often tend to use the same text field from one quarter to the next when referring to a given data item, a practice that reduces the total number of fields to be classified.
We classify each unique text string into broad categories, proceeding in two steps. First, we classify each string into one of ninety subcategories, such as "card rewards," "custodian fees," "affordable/low-income housing," "servicing," "dues/ memberships/subscriptions," and "lockbox fee." We chose these subcategories by grouping together apparently similar items, employing our institutional knowledge where possible, as well as internet searches and our best judgment. A list of these subcategories, along with the percentage of non-missing values, is presented in Appendix B to this paper. This classification was in part done by hand, and in part via Stata code that conducted Boolean searches for keywords within each text string. The subcategories include four separate "miscellaneous/other" categories, one for text strings that are well-defined but do not fit into any obvious category (for example, "cattle feed," "livestock," and "image processing"), one for items that we did not understand (for example, "tops expense"), one for items that are vague or otherwise unclassifiable (for example, "sundry loss"), and one for text strings that combine multiple items with values listed.
Since most of the subcategories are fairly sparsely populated, as documented in Appendix B, we then aggregate them into nine categories that are better suited to statistical analysis. We also assign each of the eleven standardized memoranda items to one of the same nine author-defined categories. By doing this, we are able to classify 66.2 percent of other non-interest expense into the nine high-level categories, which are listed below:
* Corporate overhead (18.6 percent of other NIE). This category, which is intended to measure general corporate expenses, includes four standardized Y-9C items: "accounting and auditing," "printing, stationery, and supplies," "postage," and "advertising and marketing." It also includes write-in expenses related to corporate overhead costs, such as travel, business development, recruitment, professional memberships and subscriptions, and charitable contributions.
* Information technology and data processing (12.6 percent of other NIE). This category covers the standardized Y-9C item "data processing expenses," as well as write-in expenses related to information technology, software, and internet banking.
* Consulting and advisory (11.1 percent of other NIE). This category is the standardized Y-9C item "consulting and advisory expenses." It does not include any write-in expenses.
* Legal (6.7 percent of other NIE). This category includes the standardized Y-9C item "legal fees and expenses," as well as write-in line items related to "litigation," "settlement," "records retention," "legal reserve," and similar items. (9)
* Retail banking (6.4 percent of other NIE). This category is intended to reflect operating costs related to lending and deposit-taking. It includes the standardized NIE category "ATM and interchange expenses," as well as write-in items related to loans, retail banking, or credit cards (for example, costs related to real estate owned properties, credit reports, credit card rewards, branch closing costs, lockbox fees, check fraud, and so on).
* Federal Deposit Insurance Corporation (FDIC) assessments and other government-related expenses (5.8 percent of other NIE). This category includes the standardized Y-9C item "FDIC deposit insurance assessments" and write-in expenses related to the Community Reinvestment Act, compliance with regulation, and other items. In practice, deposit insurance fees make up the bulk of these expenses.
* Other financial services (3.0 percent of other NIE). This category embraces written-in expense items for financial activities other than traditional lending and depository services--in particular, asset management, insurance, and miscellaneous derivatives- and trading-related expenses.
* Directors' fees and other compensation (0.3 percent of other NIE). This category includes the standardized Y-9C category "directors' fees," as well as write-in fields related to director compensation or other compensation costs.
* Miscellaneous (1.8 percent of other NIE). The final category reflects the four types of miscellaneous categories described above--that is, items that cannot be easily classified or are not understood by us based on the content of the write-in field.
In a small minority of cases, the write-in field content suggests an expense item that may have been classified as other NIE by mistake (for example, costs related to employee compensation). We did not attempt to reclassify these expenses, given the limited context and detail in the write-in fields.
Descriptive statistics for these nine author-defined categories of other NIE are presented in Panel B of Table 2. Note that the individual percentiles and standard deviations reported in the table are based on annual expenses, rather than quarterly values. We adopt this approach because of the significant number of zero values reported for even these nine aggregated categories. Our analysis of the other NIE subcategories is based on these year-end cumulative expenses.
The variation across BHCs in the relative size of different components of other NIE is striking. For example, the category "other financial services," which includes non-interest expense related to insurance and other nonbanking financial services, has a median value of zero, but at the 99.5th percentile, it is 15.9 percent of total other non-interest expense. This varied distribution of expenses is consistent with the dispersion in products and services offered by BHCs.
Operating costs are likely to vary significantly across BHCs engaged in different business activities. While the decision to enter different businesses is endogenous, and may be related to size, we are primarily interested in understanding how size is related to operating expenses on an apples-to-apples basis. For this reason, our regression analysis controls for a variety of BHC characteristics reported in the FR Y-9C. Summary statistics for these controls are presented in Table 3. In order to show how these controls are related to bank size, we also present industry averages for the following size cohorts: largest 1 percent, 95 to 99 percent, 75 to 95 percent, 50 to 75 percent, and smallest 50 percent. (10) Differences in BHC characteristics by size are clear from differences in sample means within the cohorts. However, there is substantial variation in business models apparent within size cohorts as well.
The controls in Table 3 are grouped into six categories, as follows:
* Asset shares. Our asset composition control variables measure the fraction of balance sheet assets held in various types of loans and other assets (for example, trading assets, securities, cash, and fixed assets). As shown in Table 3, small firms hold a higher fraction of total assets in the form of loans, while trading assets are a significantly higher share of total assets for the largest BHCs than for any other group.
* Risk. We control for two additional measures of asset risk: risk-weighted assets as a percentage of total assets, and nonperforming loans (NPLs) as a percentage of total loans. The relationship between firm size and risk is non-monotonic for both risk measures, although we note that the largest firms have significantly higher nonperforming loan ratios than other BHCs.
* Revenue composition. Revenue composition refers to the percentage of net operating revenue (the sum of interest and non-interest income) that is earned from different sources:
(i) interest income, (ii) trading income, and (iii) five different components of non-interest nontrading income. Since these components can be volatile, in the regressions we include these variables in the form of a four-quarter rolling average lagged value. (The standard deviation reported in the table is based on this four-quarter rolling average.) It is notable that large BHCs earn a significantly higher percentage of revenue from non-interest income.
* Funding structure. In some specifications, we include two controls for funding structure, the ratio of deposits to assets, and a dummy for whether the BHC is a publicly traded company (firms with foreign parents are coded as private, regardless of whether their ultimate parent is public). Large firms fund less of their assets with deposits, on average.
* Business concentration. Research in organizational economics has found that diversified firms tend to be less efficient and less profitable than focused firms. In studies that are most relevant to our analysis, Goetz, Laeven, and Levine (2013) find that geographically diversified commercial banks have lower valuations, while Laeven and Levine (2007) find a diversification discount (based on the firm's activity mix) in an international cross-section of banks. In the spirit of these studies, we include Herfindahl-Hirschman Index (HHI)-style measures of asset and income concentration, computed as the sum of squared asset weights and income weights, respectively, based on the categories presented in Table 3. Higher values of these measures indicate greater concentration. As the table shows, large firms have more diversified assets and activities (lower HHI), reflecting their greater reliance on financial activities outside of traditional lending and deposit taking.
* Organizational complexity. Organizationally complex firms may also have higher operating costs, because of various internal inefficiencies (for example, duplication of efforts across different subsidiaries or divisions within the same firm). It is important to attempt to disentangle the effects of size and structure, given that large firms are likely to be organizationally complex. Our analysis includes three measures of organizational structure, the log number of subsidiaries (following Avraham, Selvaggi, and Vickery ), the percentage of subsidiaries domiciled overseas, and
a dummy for whether the BHC has a foreign parent. As shown by the sample means across size cohorts, large firms have more complex organizational structures than small firms on each of these dimensions. The differences are striking: the largest BHCs (those in the top 1 percent of the size distribution) have 962 subsidiaries on average, 22.7 percent of which are domiciled overseas. BHCs below the sample median in size, however, have only 4 subsidiaries on average, and only 4.8 percent of these subsidiaries are domiciled outside the United States.
In this section, we study the relationship between BHC size and measures of total non-interest expense scaled by revenue or assets, examining how this relationship is affected by controlling for differences in firms' business models and by the normalization of non-interest expense used. Our analysis progressively adds controls for a wide range of measures of the composition of BHC assets and sources of income, on the presumption that some types of assets or activities are likely to be more complex and time-consuming to manage than others. For example, a BHC with a large portfolio of other real estate owned assets will likely incur significant property maintenance and management expenses associated with these assets, compared with an otherwise similar banking firm that has liquidated such properties in return for cash, government securities, or other simple assets. Similarly, a portfolio of consumer loans is likely to have different screening and monitoring costs than a portfolio of commercial loans. Including these controls seems particularly important given that asset composition varies significantly by firm size, as documented in Section 3.
4.1 Total Noninterest Expense
Table 4 presents ordinary least squares estimates of the relationship between the efficiency ratio and BHC size measured by the log of total assets. We find a statistically and economically significant inverse relationship between size and the efficiency ratio in each regression specification. That is, non-interest expenses per dollar of net operating revenue are lower for large BHCs.
The first column of results controls only for time-series variation in the efficiency ratio, through the inclusion of quarter fixed effects. Each subsequent regression specification successively adds more explanatory variables associated with differences in BHCs' business activities. We begin with simple controls for the composition of BHC assets and add more detailed measures of the risk of those assets, the composition of revenue, funding structure, business concentration, organizational complexity, and geography.
Looking across the models, we see that the inclusion of additional controls tends to steepen the inverse relationship between BHC size and the efficiency ratio. Including controls for BHC asset composition (for example, the percentage of assets in fixed assets, residential real estate loans, trading assets, and so on) increases the magnitude of the coefficient on bank size by 54 percent (from -1.32 in specification 1 to -1.96 in specification 3), and increases the explanatory power of the model by 13 percentage points. Controlling for the percentage of income generated by different activities (for example, trading, investment banking, and deposit service charges) shifts the coefficient to -2.63 (specification 6). The inclusion of controls for organizational complexity further steepens the association between BHC size and the efficiency ratio; the coefficient increases in magnitude from -2.98 (specification 8) to -4.13 (specification 9).
For the model including all controls but excluding firm fixed effects (specification 10), the coefficient on size of -4.151 implies that a 10 percent increase in size is associated with a 42 basis point decrease in the efficiency ratio, equivalent to 0.6 percent of the sample average efficiency ratio. In dollar terms, the coefficient implies that for a BHC at the mean of the data ($9.1 billion in assets), an increase in size of $1 billion is associated with a reduction in operating expenses of $437,000 per quarter, relative to a counterfactual in which the efficiency ratio is not associated with size. The corresponding calculation for the smaller coefficient from column 2 implies a reduction in operating expenses of $199,000 per quarter.
The final specification in Table 4 includes BHC fixed effects, and thus examines only changes in size within bank holding companies. This within-firm analysis includes both size changes from organic growth and size changes from mergers. While still statistically significant, this coefficient is somewhat smaller in magnitude than that of specification 10 (-2.47 compared with -4.15). There is some evidence that non-interest expenses after mergers are inflated by one-time merger related costs (Kwan and Wilcox 2002), which may account for this difference. The standard error of the size coefficient estimate from specification 11 is much larger than in the other specifications; in other words, the coefficients are estimated with lower power, owing to the smaller residual variation in the efficiency ratio not absorbed or accounted for by the fixed effects and other controls.
As expected, observable differences among BHCs explain a significant fraction of the variation in non-interest expenses. Simple asset controls alone more than double the adjusted [R.sup.2] of the initial specification. However, even the fixed effects specification in column 11 has an [R.sup.2] of only 54.9 percent, implying a large amount of residual variation in operating costs. Furthermore, the inclusion of BHC fixed effects nearly doubles the [R.sup.2] relative to specification 10, a result suggestive of large persistent differences in operating costs across observably similar firms. This finding seems consistent with prior literature on X-inefficiency, which shows that many banking firms operate significantly inside the efficient production frontier (see, for example, Berger, Hunter, and Timme ). It is worth noting that BHC size alone explains only a very small fraction (less than 1 percent) of the total variation in non-interest expense in the data, as illustrated graphically in Chart 2.
In sum, Table 4 provides consistent evidence that large BHCs have lower operating costs as measured by the efficiency ratio, although the strength of the relationship is sensitive to the set of controls used. Instead of taking a strong stance on the "appropriate" set of controls, throughout the paper we present results for specifications using controls from columns 1, 2, and 10 from Table 4. A comparison of the results across these specifications enables the reader to observe how the relationship between non-interest expenses and size is influenced by the inclusion or exclusion of controls for the mix of BHC assets and business activities.
Although our main focus is on the relationship between operating costs and firm size, estimates for several of the controls included in Table 4 are also of independent interest. In particular, BHC organizational complexity, measured by the log number of subsidiaries, is associated with higher non-interest expense ratios. BHCs with a foreign parent also have higher expenses. Proxies for greater organizational focus are associated with lower non-interest expense: BHCs that have more concentrated asset portfolios and more concentrated sources of non-interest income have lower expenses, all else equal, although the marginal explanatory power of additional concentration is relatively low. Each of these relationships is robust to the inclusion of BHC fixed effects (column 11). Although not shown in Table 4, these relationships are also robust to specification changes that allow for a more flexible linkage between size and the efficiency ratio. This finding suggests that our results are not likely to be driven only by the largest BHCs.
Caution should be exercised in applying a causal interpretation to these associations, given that we do not have a convincing econometric instrument for organizational complexity or focus. But taken at face value, each of these estimates implies that complex, diversified firms have higher operating expenses than focused or organizationally simple firms, consistent with the conclusions of prior literature on the diversification discount in banking (Goetz, Laeven, and Levine 2013; Laeven and Levine 2007).
4.2 Other Functional Forms
The specifications so far assume a log-linear relationship between BHC size and the efficiency ratio. Next we allow for a more flexible functional form by estimating fractional polynomial specifications that permit the data to determine the shape of the relationship between size and the NIE ratio. An alternative to regular polynomials, fractional polynomials provide flexible parameterization for continuous variables. We use the Stata function fracpoly to determine an optimal polynomial specification (optimal polynomial) and also estimate a specification with exponents ranging from -2 to 2--that is, log assets raised to the -2, -1, 0, 1, and 2 power (flex polynomial). These best-fit polynomials are shown in Chart 3 along with the ordinary least squares line of best fit.
Overall, the log-linear functional form assumed in Table 4 appears to be a good approximation, although we note that, based on point estimates, the point-estimated relationship between log assets and the efficiency ratio is somewhat concave at the tails. Specifically, the relationship between BHC size and the NIE ratio is relatively flat among small BHCs (those with assets below $150 million), while the relationship is steeper among the largest BHCs (those with assets above $750 billion). For the vast range of asset sizes, the relationship between log size and efficiency ratio is close to linear, and the 95 percent confidence interval of the alternative forms is very similar. Thus, we use a log-linear specification for the remainder of the analysis.
In addition to investigating flexible polynomial specifications, we separate the sample into different size cohorts, re-sorted in each quarter, and estimate separate specifications for each cohort. This approach allows the relationship between NIE and control variables, as well as size, to vary by BHC size class. (In the fractional polynomial approach, the coefficients on explanatory variables other than size are unrelated to size.) Each column of Table 5 represents specifications 1, 2, and 10 of Table 4 estimated on a subset of the BHCs sorted by size in each year. The first column replicates the results on the entire sample, for comparison. Without including controls for BHC asset mix, it appears that much of the coefficient on size is driven by BHCs below the median asset size (column 6). As additional controls are included, economies of scale become apparent in many of the size cohorts. In the specification including all controls, the estimated coefficient on size is negative in all cohorts and statistically significant. As suggested by the flexible polynomial specifications, the point estimate coefficient on size is largest in the top 1 percent of the sample.
What do these findings imply for the policy debate around size limits for the largest BHCs? We find no evidence that the inverse relationship between size and operating costs disappears above any particular size threshold; indeed, our point estimates suggest that, if anything, the relationship is steeper among the largest firms. This result is consistent with scale economies from sources other than bargaining power to the extent that we believe that differences in bargaining power may be small within the top 1 percent of BHCs. The statistical precision of our estimates is limited, however, given the small number of observations for the largest BHCs.
4.3 Alternative Measures of Operating Costs
The efficiency ratio may be distorted in periods when net operating income is temporarily low. (11) Next, we test the sensitivity of our results to other normalizations of non-interest expense: the expense asset ratio discussed in Section 3 (NIE / total assets), NIE / risk-weighted assets, and a "cash" efficiency ratio, which excludes noncash expenses such as goodwill amortization in the numerator. We do this because noncash expenses are often associated with one-time costs relating to mergers and acquisitions that are not likely to persist, and may be associated with size. We also estimate a specification using the log of non-interest expense as an alternative measure of operating costs.
As before, for each normalization of NIE, we re-estimate specifications with the set of right-hand-side variables from columns 1, 2, and 10 of Table 4 and present the coefficient on asset size. Results are presented in Table 6. Regardless of the normalization used, the coefficient on size is negative and statistically significant once BHC controls are included. In the specification including all controls, the estimated coefficient on size is approximately 7 to 10 percent of the average expense ratio.
For the specifications using the log of non-interest expense as the dependent variable, the coefficient on log assets can be directly interpreted as the elasticity of operating costs with respect to size. In line with our other results, this elasticity is less than unity--in other words, a 10 percent change in BHC size is associated with a less than 10 percent change in NIE operating costs, a finding consistent with the presence of scale economies in operating costs. For the specification including all controls, the operating cost elasticity is 0.899, much smaller than one, although it is significantly closer to one for the specification just including asset controls (0.979). Both these estimates are statistically significantly smaller than unity.
5. DECOMPOSITION OF NONINTEREST EXPENSE
This section examines the relationship between BHC size and components of non-interest expense. First, we consider the five major components of non-interest expense reported in the Y-9C income statement. Probing more deeply, we then analyze the nine subcomponents of "other non-interest expense," using our manual classification of these expenses as described in Section 3.
One goal of this disaggregated analysis is to shed additional light on the sources of the lower operating costs enjoyed by large BHCs. Although these lower costs could be due to scale economies or other efficiency benefits of size, they could also reflect implicit government guarantees for large BHCs, or the greater bargaining power of these firms. For example, large banks may endogenously select riskier activities, but invest less in risk management because of implicit insurance associated with being "too big to fail" Alternatively, large banks may simply take advantage of greater bargaining power to reduce expenses. These different explanations have very different normative welfare implications. Efficiency benefits of size imply that limiting size would impose deadweight economic costs, while explanations relating to bargaining power and TBTF primarily relate to the allocation of economic rents. Although the breakdown of expenses in the Y-9C does not allow us to fully disentangle these different explanations, we are able to draw some suggestive conclusions.
5.1 Major Components of Noninterest Expense
We begin by studying the five expense categories reported on Schedule HI: compensation (49.4 percent of non-interest expense), premises and fixed assets expense (11.6 percent), goodwill impairment (1.8 percent), amortization (1.9 percent), and other (35.0 percent). Results are presented in Table 7. As before, we normalize each expense by net operating revenue, and for parsimony, focus on the coefficient on log assets for specifications 1, 2, and 10 from Table 4.
Each of the three largest categories of non-interest expense declines as a percentage of net revenue as size increases, all else equal, with or without the inclusion of controls for BHC characteristics. The final column of the table presents the estimated coefficient scaled by the mean of the dependent variable in question (that is, an elasticity of the component efficiency ratio with respect to firm size). Focusing on the specifications including these controls (either for asset composition alone, or for all controls), we find that the inverse relationship between BHC size and scaled non-interest expense is steepest for compensation, followed by other non-interest expense, based on this calculated elasticity. For the specifications including all controls, a 10 percent increase in size is associated with a 0.735 percent decline in compensation scaled by net operating revenue and a 0.683 percent decline in the corresponding ratio for other non-interest expense. The result for employee compensation is perhaps surprising, given that large BHCs have more employees in highly compensated roles such as investment banking and trading. However, the higher productivity and additional revenue earned by these employees (the denominator of the efficiency ratio) appears to offset this higher compensation.
Expenses related to premises and fixed assets may represent a category of operating costs for which scale efficiencies are lower (for example, building lease costs are roughly proportionate to the size of the leased space, at least within a specific geographic area). Given this, it is perhaps unsurprising that estimated economies of scale are smaller for premises and fixed assets expense: for this category, our point estimate implies that a 10 percent increase in size is associated with a 0.478 percent decline in expenses scaled by operating revenue.
Significantly, expenses related to the impairment and amortization of goodwill and other intangible assets are actually proportionately higher for large firms--a fact that distinguishes these expenses from the other categories. We estimate a positive, statistically significant (in most specifications) coefficient on these expenses. The likely key reason for this finding is that large BHCs often have grown by way of acquisitions, which will sometimes result in goodwill when the acquisition purchase price exceeds the tangible book value of assets purchased. Consequently, these firms report higher expenses related to the amortization or impairment of these assets. Although the positive slope for these two expense categories is economically significant, the two categories together make up only a relatively small proportion (3.7 percent) of total industry NIE.
5.2 Subcomponents of Other Noninterest Expense
In this section, we examine the nine subcomponents of "other NIE" identified in section 3.2. (Recall that these categories reflect both standardized memoranda items reported on the Y-9C since 2008 and "write-in" text strings classified by us.) Previous work estimating scale curves for these disaggregated categories has been based on case studies or has had limited sample size (for example, Clearing House Association ).
Overall, we find evidence that scaled expense falls with size for most, but not all, components of other non-interest expense, especially after including controls for BHC asset and income composition. When controls for the composition of assets and income sources are included in the specification, large BHCs exhibit lower expenses in categories in which a fixed cost can be spread across an expanded scale of operations, such as corporate overhead, information technology, and data processing.
The lower part of Table 7 presents results for the other NIE components, listed in descending order of size. Corporate overhead is the largest component of other non-interest expense, and a component for which we estimate significant scale efficiencies (a high estimated coefficient on size relative to mean level of expense). Corporate overhead includes expenses such as accounting and auditing, advertising and marketing, treasury expenses, travel and business development, charitable donations, insurance, and utilities. These expenses appear to have significant operational leverage; the estimated coefficient on size is -0.33, approximately 7 percent of the mean level of corporate overhead expenses.
Similar scale economies are observed for expenses associated with information technology and data processing, with an estimated coefficient on size that is -6.6 percent of mean IT expense. This finding is consistent with the view that spreading overhead expenses associated with technology may be one source of cost advantage for large banking firms.
In contrast to these two categories, we find that expenses associated with consulting and advisory services are proportionately higher for large BHCs. Prior to adding controls for BHC characteristics, our estimates show that the coefficient on size and consulting expenses is positive and statistically significant. This coefficient remains significant when asset composition controls are included, although once all controls are included, the coefficient is positive but no longer statistically significant. This suggests that consulting and advisory services may be related to non-interest income, rather than to the composition of BHC assets. Despite recent publicity surrounding large BHCs' legal issues and large-dollar-value settlements, over the 2008-12 period, legal expenses also increase less than proportionately with BHC size, particularly in the specification including the full set of controls (specification 10 from Table 4). This expense category includes both legal fees and retainers paid for legal services performed, as well as expenses associated with legal settlements and reserves, to the extent we can identify these expenses from the write-in text fields. Some part of this finding may reflect the fact that small banks may lack internal legal departments, for which expenses would be recorded as part of compensation, and thus have higher external legal fees.
The assignment of write-in fields to retail banking requires perhaps the most judgment on our part. This category includes collection expenses, credit reports, mortgage-related expenses such as appraisal and title fees, branch expenses, checks, lockboxes, and robbery, among many others. After including asset composition controls, the estimated coefficient remains negative although not statistically significant. This result may reflect the wide variation in the types of retail banking businesses that are not well captured by our BHC characteristics. Alternatively, economies of scale may be limited or not present for branch banking (at least among the set of expenses classified into this category), since many costs only scale until the next branch is opened.
Similarly, we find a negative but statistically insignificant relationship between size and normalized FDIC assessments and other government-related expenses after including the full complement of BHC characteristics. The majority of the expenses in this line item are due to deposit insurance, and thus it would be surprising to uncover economies of scale once we control for the amount of deposit financing. This coefficient would likely shrink further if our regression specification included a control for the fraction of insured deposits, rather than total deposits.
The category "other financial services" represents the sum of expenses associated with BHCs' non-banking businesses, such as asset management, trust and custody services, and insurance. Given likely differences in the non-interest expenses of these businesses, it is not surprising that the estimated coefficient changes sign from positive to negative once we control for the composition of BHCs' assets and non-interest income. Banking firms that earn a high percentage of income from fee income should naturally have higher expenses. But holding all else equal and controlling for income composition, we find that larger BHCs have lower scaled expenses in this category: we estimate a coefficient of 7.4 percent of the mean value. This result is consistent with cost economies of scale in noncompensation expenses associated with businesses such as insurance and asset management.
The component of other non-interest expense for which scale economies are largest in percentage terms is directors' fees and other compensation. For this category, the coefficient on size is almost three times as large as the sample mean.
This makes intuitive sense; even though directors of large BHCs have higher compensation, board size does not increase dramatically with firm size. This coefficient is negative and significant regardless of the set of controls used.
Miscellaneous expenses include items as varied as expenditures for cattle feed and reducing gold to market. It also includes nonspecific write-in text fields such as "miscellaneous expense," "miscellaneous fee," and "other expense." Regardless of the controls for bank businesses used, we do not see economies of scale in these varied expenses, although some economies may exist in the residual category "other expenses," which includes all non-interest expenses not otherwise classified.
We find a robust inverse relationship between the size of bank holding companies and scaled measures of operating costs. Quantitatively, a 10 percent increase in assets is associated with a 0.3 to 0.6 percent decline in non-interest expense scaled by income or assets, depending on the specification. In dollar terms, our estimates imply that for a BHC of mean size in our sample, an additional $1 billion in assets reduces non-interest expense by $1 million to $2 million per year, relative to a base case where operating cost ratios are unrelated to size. This inverse relationship is robust to various changes in model specification, although the magnitude of the relationship is sensitive to the set of controls used.
Unpacking our results, we find that while size is associated with lower scaled operating costs for most components of non-interest expense, the largest contributions in dollar terms come from employee compensation, premises and fixed assets, corporate overhead, and information technology and data processing. While not a large component of total non-interest expense, directors' fees and other compensation account for the largest proportionate savings, presumably a reflection of the fact that corporate boards do not expand with firm size, even if their members are better paid on average.
Our results likely reflect a combination of three factors: First, large BHCs benefit from "operational leverage" or economies of scale, whereby they effectively spread costs over a higher revenue or asset base. Second, "X-efficiency"--a factor closely related to operational leverage--may be higher for large BHCs; that is, these firms may operate closer to the production frontier on average. Third, large BHCs may have greater bargaining power than smaller firms with suppliers or employees. We are not able to pin down with confidence the relative contribution of these three factors. We emphasize, however, that the inverse relationship between BHC size and scaled measures of NIE is not limited to particular components of expense or particular segments of the BHC size distribution.
Consistent with recent research that identifies the presence of scale economies in banking, our results suggest that imposing size limits on banking firms would be likely to involve real economic costs. Although the limitations of our econometric methodology must be borne in mind, a back-of-the-envelope calculation applied to our estimates implies that limiting BHC size to be no larger than 4 percent of GDP would increase total non-interest expense by $2 billion to $4 billion per quarter. These costs should be weighed against the potential benefits of size limits as policymakers address the "too-big-to-fail" problem.
APPENDIX A: VARIABLE DEFINITIONS Income Statement Variables Variable Definition Y-9C Mnemonic Construction/Variable Source Net interest income bhck4074 [1981:Q2-present] Noninterest income bhck4079 [1981:Q2-present] Trading revenue Includes the net gain bhcka220 or loss from trading [1996:Q1-present] cash instruments and off-balance-sheet derivative contracts (including commodity contracts) that has been recognized during the calendar year-to-date Fiduciary income Includes income from bhck4070 + bhckb494 fiduciary activities, [2001:Q1-2002:Q4], fees and commissions bhck4070 + bhckc386 + from annuity sales, bhckc387 [2003:Q1 - underwriting income 2006:Q4], bhck4070 + from insurance and bhckc887 + bhckc385 + reinsurance bhckc387 activities, and income [2007:Q1-present] from other insurance activities Investment banking Includes venture bhck b491 + bhckb490 income capital revenue, fees [2001:Q1-2006:Q4], and commissions from bhckb491 + bhckc886 securities brokerage, + bhckc888 and investment [2007:Q1-present] banking, advisory, and underwriting fees and commissions Service charges Service charges on bhck4884 on deposits deposit accounts in [1981:Q2-present] domestic offices Net servicing fees Includes income from bhckb492 servicing real estate [2001:Q1-present] mortgages, credit cards, and other financial assets held by others Other income Total noninterest Derived income not accounted for in the five categories listed above Net operating revenue Net interest income bhck4074 + bhck4079 plus noninterest [1981:Q2-present] income Noninterest expense bhck4093 [1981:Q2-present] Compensation Salaries and employee bhck4135 benefits [1981:Q2-present] Premises and fixed bhck4217 assets [1981:Q2-present] Amortization Amortization expense bhckc232 expense and impairment losses [2002:Q1-present] for other intangible assets Goodwill impairment Goodwill impairment bhckc216 losses [2002:Q1-present] Other Total noninterest Derived expense not accounted for in the four categories listed above Data processing Eleven standardized bhckc017 expenses other noninterest [2002:Q1-present] Advertising and expense items reported bhck0497 marketing in Schedule HI: [2002:Q1-present] expenses Memoranda of the FR Y- Directors' fees 9C beginning either in bhck4136 2002 or in 2008. BHC [2002:Q1-present] Printing, filers only report bhckc018 stationery, amounts greater than [2002:Q1-present] and supplies $25,000 that exceed 3 Postage percent of total other bhck8403 noninterest expense [2002:Q1-present] Legal fees and bhck4141 expenses [2002:Q1-present] FDIC deposit bhck4146 insurance [2002:Q1-present] assessment Accounting and bhckf556 auditing [2008:Q1-present] expenses Consulting and bhckf557 advisory [2008:Q1-present] expenses ATM and bhckf558 interchange [2008:Q1-present] expenses Telecommunications bhckf559 expenses [2008:Q1-present] TEXT8565 Description of the bhck8565 "write-in" components [1994:Q1-present] TEXT8566 of other noninterest bhck8566 expense. BHCs only [1994:Q1-present] TEXT8567 report amounts that bhck8567 exceed 10 percent of [1994:Q1-present] total other noninterest expense Consolidated Balance Sheet Variables Variable Definition Y-9C Mnemonic Construction/Variable Source Total assets bhck2170 [1991:Q1-present] Total loans bhck2122 [1991:Q1-present] Residential real The sum of 1) all bhdm1797 + bhdm5367 estate loans other loans secured by + bhdm5368 one-to-four-family [1991:Q1-present] residential properties: secured by first liens; 2) all other loans secured by one-to-four-family residential properties: secured by junior liens; 3) revolving, open-end loans secured by one- to-four-family residential properties and extended under lines of credit Commercial real The sum of 1) one-to- bhdm1415 + bhdm1460 + estate loans four-family bhdm1480 [1990:Q3- residential 2006:Q4], bhckf158 + construction loans; 2) bhckf159 + bhdm1460 + other construction bhckf160 + bhckf161 loans and all land [2007:Q1-present] development and other land loans; 3) real estate loans secured by multifamily (five or more) residential properties; 4) loans secured by owner- occupied nonfarm nonresidential properties; 5) loans secured by other nonfarm nonresidential properties Credit card loans Loans to individuals bhck2008 [1991:Q1- for household, family, 2000:Q4], bhckb538 and other personal [2001:Q1-present] expenditures (that is, consumer loans). Includes purchased paper: credit cards Other consumer loans The sum of 1) loans to bhck2011 [1991:Q1- individuals for 2000:Q4], bhck2011 + household, family, and bhckb539 [2001:Q1- other personal 2010:Q4], bhckb539 + expenditures--that is, bhckk137 + bhckk207 consumer loans [2011:Q1-present] (includes purchased paper): other revolving credit plans; 2) automobile loans to individuals for household, family, and other personal expenditures--that is, consumer loans (includes purchased paper); 3) other consumer loans to individuals, for household, family, and other personal expenditures (includes single payment, installment, and all student loans) All other loans Total loans minus the derived sum of residential real estate loans, commercial real estate loans, credit card loans, and other consumer loans Cash and balances due The sum of 1) non- bhck0081 + bhck0395 + from depository interest-bearing bhck0397 [1991:Q1- institutions balances and currency present] and coin; 2) interest- bearing balances in U.S. offices; 3) interest-bearing balances in foreign offices, edge and agreement subsidiaries, and international banking facilities Trading assets Assets held in trading bhck2146 [1981:Q2- accounts include but 1994:Q4], bhck3545 are not limited to [1995:Q1-present] U.S. Treasury securities; U.S. government agency and corporation obligations; securities issued by states and political subdivisions in the United States; other bonds, notes, and debentures; certificates of deposit; commercial paper; and bankers acceptances. Assets held in trading accounts also include the amount of revaluation gains from the "marking to market" of interest rate, foreign exchange rate, and other off-balance-sheet commodity and equity contracts held for trading purposes Federal funds and The sum of 1) bhck1350 [1981:Q2- repurchase agreements outstanding amount of 1988:Q1][1997:Q1 - federal funds 2001:Q4], bhck0276 + sold--that is, im- bhck0277 [1988:Q2- mediately available 1996:Q4], bhdmb987 + funds lent (in bhckb989 [2002:Q1- domestic offices) present] under agreements or contracts that have an original maturity of one business day or roll over under a continuing contract, excluding such funds lent in the form of securities purchased under agreements to resell and over-night lending for commercial and industrial purposes; 2) securities resale agreements, regardless of maturity, if the agreement requires the bank to resell the identical security purchased or a security that meets the definition of substantially the same in the case of a dollar roll, and purchases of participations in pools of securities, regardless of maturity Investment securities Held-to-maturity bhck0390 [1981:Q2- securities (at 1993:Q4], bhck1754 + amortized cost) plus bhck1773 [1994:Q1- available for sale present] securities (at fair value) Other real estate owned The book value (not to bhck2150 exceed fair value), [1981:Q2-1990:Q2] less accumulated [2001:Q1-present], depreciation, if any, bhck2744 + bhck2745 of all real estate [1990:Q3-2000:Q4] other than bank premises actually owned by the bank and its consolidated subsidiaries. Premises and fixed bhck2145 assets Investments in Includes the amount of bhck2130-bhck3656 unconsolidated the bank holding [1981:Q2-2009:Q1], subsidiaries and company's investments bhck2130 associated companies in subsidiaries that [2009:Q2-present] have not been consolidated; associated companies; and corporate joint ventures, unincorporated joint ventures, general partnerships, and limited partnerships over which the bank exercises significant influence (collectively referred to as "investees"). Also includes loans and advances to investees and holdings of their bonds, notes, and debentures Investments in real The book value of bhck3656 estate ventures direct and indirect [1981:Q2-present] investments in real estate ventures Intangible and other Other identifiable bhck3165 + bhck2160 + assets intangible assets plus bhck2155 [1985:Q2- other assets 1991:Q4], bhck3164 + bhck5506 + bhck5507 + bhck2160 + bhck2155 [1992:Q1-1998:Q4], bhck0426 + bhck2160 + bhck2155 [2001:Q1- 2005:Q4], bhck0426 + bhck2160 [2006:Q1- present] Nonperforming loans The sum of 1) total bhck5525-bhck3506 + loans and leasing bhck5526-bhck3507 financing receivables [1990:Q3-present] that are ninety days or more past due and still accruing; 2) total loans and leasing financing receivables in nonaccrual status. Risk-weighted assets BHC risk-weighted bhcka223 assets net of all [1996:Q1-present] deductions Total deposits 1) Non-interest- bhdm6631 + bhdm6636 + bearing deposits 2) bhfn6631 + bhfn6636 total interest- [1981:Q2-present] bearing deposits in foreign and domestic offices Other Characteristics and Organizational Structure Variables Variable Definition Y-9C Mnemonic Construction/Variable Source Public Dummy=1 if firm has Federal Reserve Bank PERMCO, Dummy=0 of New York. 2013. otherwise CRSP-FRB Link Number of Total number of NIC Top Holder Table: subsidiaries offspring entities top holder variable whose relationship to rssd9003 the bank holding company is regulated, that is, governed by applicable banking statutes, which are either federal or state banking laws Foreign subsidiaries Total number of NIC Country Name offspring entities Directory: domestic that are not domiciled indicator rssd9101 in the United States Foreign parent Dummy=1 if the highest NIC Board Derived entity in the Items Table: foreign organization is not family ID rssd9360 domiciled in the United States, Dummy=0 otherwise Source: Board of Governors of the Federal Reserve System, Microdata Reference Manual. Note: BHC is bank holding company; FDIC is Federal Deposit Insurance Corporation; CRSP is Center for Research in Securities Prices; NIC is National Information Center.
Note to Readers:
Appendix B, "Additional Materials" is available as a separate file at http://www.newyorkfed.org/research/ epr/2014/1412kovn_appendixB.pdf.
Avraham, D., P Selvaggi, and J. Vickery. 2012. "A Structural View of U.S. Bank Holding Companies." Federal Reserve Bank of New York ECONOMIC POLICY REVIEW 18, no. 2: 65-81.
Benston, G. 1972. "Economies of Scale of Financial Institutions." JOURNAL OF MONEY, CREDIT, AND BANKING 4 (May): 312-41.
Benston, G., G. A. Hanweck, and D. B. Humphrey. 1982. "Scale Economies in Banking: A Restructuring and Reassessment" JOURNAL OF MONEY, CREDIT, AND BANKING 14 (November): 435-56.
Berger, A., and D. Humphrey. 1991. "The Dominance of Inefficiencies over Scale and Product Mix Economies in Banking." JOURNAL OF MONETARY ECONOMICS 28, no. 1 (August): 117-48.
--. 1994. "Bank Scale Economies, Mergers, Concentration, and Efficiency: The U.S. Experience." Wharton Financial Institutions Center Working Paper no. 94-25.
Berger, A., W. Hunter, and S. Timme. 1993. "The Efficiency of Financial Institutions: A Review and Preview of Research Past, Present, and Future." JOURNAL OF BANKING AND FINANCE 17, no. 2-3 (April): 221-49.
Bernanke, B. 2013. SEMIANNUAL MONETARY POLICY REPORT TO THE CONGRESS. Testimony before the Committee on Financial Services, U.S. House of Representatives, Washington, D.C., July 17, and before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate, Washington, D.C., July 18.
Clearing House Association, L.L.C. 2011. Understanding the Economics of Large Banks. TCH Research Study, November.
Copeland, A. 2012. "Evolution and Heterogeneity among Larger Bank Holding Companies: 1994- 2010." Federal Reserve Bank of New York Economic Policy Review 18, no. 2: 83-93.
Cornett, M., J. McNutt, and H. Tehranian. 2006. "Performance Changes around Bank Mergers: Revenue Enhancements versus Cost Reductions." JOURNAL OF MONEY, CREDIT, AND BANKING 38, no. 4 (June): 1013-50.
Davies, R., and B. Tracey. 2014. "Too Big to Be Efficient? The Impact of Implicit Subsidies on Estimates of Scale Economies for Banks." JOURNAL OF MONEY, CREDIT, AND BANKING 46, no. s1 (February): 219-53.
DeYoung, R. 1998. "Management Quality and X-Inefficiency in National Banks." JOURNAL OF FINANCIAL SERVICES RESEARCH 13, no. 1 (February): 5-22.
Feng, G., and A. Serletis. 2010. "Efficiency, Technical Change, and Returns to Scale in Large U.S. Banks: Panel Data Evidence from an Output Distance Function Satisfying Theoretical Regularity" JOURNAL OF BANKING AND FINANCE 34, no. 1 (January): 127-38.
Goetz, M., L. Laeven, and R. Levine. 2013. "Identifying the Valuation Effects and Agency Costs of Corporate Diversification: Evidence from the Geographic Diversification of U.S. Banks." REVIEW OF FINANCIAL STUDIES. 26, no. 7: 1787-1823.
Hannan, T., and S. Pilloff. 2006. "Acquisition Targets and Motives in the Banking Industry." Board of Governors of the Federal Reserve System FINANCE AND ECONOMICS DISCUSSION SERIES, no. 2006-40.
Hughes J., and L. Mester. 2013. "Who Said Large Banks Don't Experience Scale Economies? Evidence from a Risk-Return-Driven Cost Function." JOURNAL OF FINANCIAL INTERMEDIATION 22, no. 4 (October): 559-85.
Hughes, J. P., W Lang, L. J. Mester, and C. G. Moon. 2001. "Are Scale Economies in Banking Elusive or Illusive? Evidence Obtained by Incorporating Capital Structure and Risk-Taking into Models of Bank Production." JOURNAL OF BANKING AND FINANCE 25, no. 12 (December): 2169-208.
Johnson, S., and J. Kwak. 2010. THIRTEEN BANKERS: THE WALL STREET TAKEOVER AND THE NEXT FINANCIAL MELTDOWN. New York: Pantheon.
Kwan, S., and J. Wilcox. 2002. "Hidden Cost Reductions in Bank Mergers. Accounting for More Productive Banks." In J. Kensigner, ed., RESEARCH IN FINANCE 19, 109-24. Bingley, U.K.: Emerald Group Publishing Limited.
Laeven, L., and R. Levine. 2007. "Is There a Diversification Discount in Financial Conglomerates?" JOURNAL OF FINANCIAL ECONOMICS 85, no. 2 (August): 331-67.
Niepmann, F. 2013. "Banking across Borders with Heterogeneous Banks." Federal Reserve Bank of NEW YORK STAFF REPORTS, no. 609, April.
Peristiani, S. 1997. "Do Mergers Improve the X-Efficiency and Scale Efficiency of U.S. Banks? Evidence from the 1980s." JOURNAL OF MONEY, CREDIT, AND BANKING 29, no. 3 (August): 326-37.
Wheelock, D. C., and P. W. Wilson. 2012. "Do Large Banks Have Lower Costs? New Estimates of Returns to Scale for U.S. Banks." JOURNAL OF MONEY, CREDIT, AND BANKING 44, no. 1 (February): 171-99.
Anna Kovner is a research officer, James Vickery a senior economist, and Lily Zhou a former senior research analyst at the Federal Reserve Bank of New York.
The authors thank Peter Olson for outstanding research assistance and Gara Afonso, Jan Groen, Joseph Hughes, Donald Morgan, an anonymous referee, and workshop participants at the Federal Reserve Bank of New York for helpful comments and suggestions. The views expressed in this article are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System.
(1) For details of this calculation, see Appendix B, available as a separate file at http://www.newyorkfed.org/research/epr/2014/1412kovn_appendixB.pdf. The appendix was omitted from the main document because of space constraints.
(2) Our analysis does not attempt to separate the effects of X-efficiency from those of scale economies. We note, however, that Hughes et al. (2001) and Hughes and Mester (2013) find that estimated scale economies are larger for more efficient banks than for less efficient ones, controlling for size.
(3) Details of this calculation are presented in Appendix B, http://www .newyorkfed.org/research/epr/2014/1412kovn_appendixB.pdf.
(4) BHC net income in Schedule HI is calculated as follows: net income = net interest income + noninterest income - noninterest expense - provision for loan and lease losses + realized securities gains (losses) - taxes + extraordinary items and other adjustments - net income attributable to noncontrolling interests. See Copeland (2012) for descriptive information on how the main components of BHC income have evolved over time.
(5) A detailed definition of these five variables can be found in the Federal Reserve Microdata Reference Manual data dictionary, available at http://www .federalreserve.gov/apps/mdrm/data-dictionary.
(6) Appendix B is available at http://www.newyorkfed.org/research/ epr/2014/1412kovn_appendixB.pdf.
(7) The eleven standardized memoranda categories are (a) data processing expenses, (b) advertising and marketing expenses, (c) directors' fees, (d) printing, stationery, and supplies, (e) postage, (f) legal fees and expenses, (g) FDIC insurance assessments, (h) accounting and auditing expenses, (i) consulting and advisory expenses, (j) automated teller machine (ATM) and interchange expenses and (k) telecommunications expenses. See FR Y-9C Schedule HI Memorandum Item 7.
(8) "Other real estate owned" refers to real estate owned by a bank as a result of the foreclosure of a mortgage loan.
(9) The standardized "legal fees and expenses" other NIE category includes fees and retainers paid for legal services obtained, but excludes legal settlements and legal expenses associated with owned real estate. Legal settlements and legal reserves established against expected future settlements are recorded in the write-in text fields, if separately reported.
(10) To compute the industry average for the asset and income ratios, we sum the numerator and denominator of the ratio across all firms in the size cohort, and then take the ratio of the two sums. In contrast, the mean and standard deviation reported in the first two columns represent the unweighted mean and standard deviation of the individual observations in the sample. Of course, the mean of the individual observations may differ substantially from the industry mean if the ratio in question is correlated with firm size.
(11) During the 2007-08 financial crisis, trading losses and other losses brought net operating income close to zero for several large BHCs.
TABLE 1 Noninterest Expense Summary Statistics Industry Individual Observations Full 2012 p0.5 p5 p25 Sample Panel A: Efficiency Measures, in Percent: 2001-12 Efficiency ratio 66.32 71.68 29.07 46.31 58.26 Cash efficiency ratio 63.29 70.39 28.69 45.81 57.72 Expense-to-asset ratio 0.82 0.82 0.25 0.45 0.63 Expense-to-RWA ratio 1.22 1.35 0.35 0.61 0.87 Panel B: Components of Noninterest Expense, as a Percentage of Total: 2001-12 Compensation 49.36 48.68 18.08 40.45 50.31 Premises and fixed assets 11.63 9.64 2.79 7.78 11.47 Goodwill impairment 1.75 0.02 0.00 0.00 0.00 Amortization expense 1.93 1.78 -0.03 0.00 0.00 Other 34.95 39.88 10.02 20.93 26.22 Individual Observations p50 p75 p95 p99.5 Panel A: Efficiency Measures, in Percent: 2001-12 Efficiency ratio 65.77 74.44 93.71 198.40 Cash efficiency ratio 65.17 73.72 92.07 168.11 Expense-to-asset ratio 0.75 0.88 1.25 3.95 Expense-to-RWA ratio 1.05 1.28 1.89 6.02 Panel B: Components of Noninterest Expense, as a Percentage of Total: 2001-12 Compensation 54.67 58.58 64.59 74.30 Premises and fixed assets 13.67 16.01 20.16 26.53 Goodwill impairment 0.00 0.00 0.00 16.28 Amortization expense 0.00 0.97 3.57 9.03 Other 30.04 34.71 45.82 69.29 Individual Observations Mean Standard Deviation Panel A: Efficiency Measures, in Percent: 2001-12 Efficiency ratio 68.10 18.69 Cash efficiency ratio 67.05 16.64 Expense-to-asset ratio 0.80 0.37 Expense-to-RWA ratio 1.15 0.58 Panel B: Components of Noninterest Expense, as a Percentage of Total: 2001-12 Compensation 53.96 13.54 Premises and fixed assets 13.84 5.45 Goodwill impairment 0.29 5.03 Amortization expense 0.76 1.72 Other 31.11 16.15 Source: Board of Governors of the Federal Reserve System, Consolidated Financial Statements of Bank Holding Companies (FR Y-9C data). Notes: The table reports summary statistics for 2,810 unique bank holding companies from 2001:Q1 to 2012:Q4, a total of 58,217 firm- quarter observations. The column labeled "industry" reports the average industry efficiency ratio, calculated by summing across all bank holding companies each quarter, taking the ratio, and then taking the time-series mean, either over the 2001:Q1--2012:Q4 sample period or over calendar year 2012. The denotation "p" refers to percentiles of individual observations (for example, "p50" is the median). Variables are defined in Appendix A. RWA is risk-weighted assets. TABLE 2 Components of Other Noninterest Expense Panel A: FR Y-9C Classification of Other Noninterest Expense: 2008-12 Percentage of Total Other Noninterest Category Expense, Industry In Y-9C 37.99 Text classified 28.21 Unclassified 33.80 Total 100.00 Panel B: Components of Other Noninterest Expense, as a Percentage of Total Other Noninterest Expense: 2008-12 Individual Observations Component (Author-Defined) Industry p0.5 p5 p25 p50 Corporate overhead 18.63 0.00 2.43 10.29 16.26 Information technology and 12.63 0.00 0.64 8.21 13.84 data processing Consulting and advisory 11.07 0.00 0.00 0.00 2.31 Legal 6.68 0.00 0.00 0.00 3.53 Retail banking 6.35 0.00 0.00 0.00 6.41 FDIC assessments and other 5.81 0.00 0.00 6.80 11.53 government Other financial services 3.01 0.00 0.00 0.00 0.00 Directors' fees and other 0.25 0.00 0.00 0.00 0.00 compensation Miscellaneous 1.76 0.00 0.00 0.00 0.00 Total classified 66.20 4.02 35.11 55.83 66.87 Unclassified 33.80 Individual Observations Standard Component (Author-Defined) p75 p95 p99.5 Mean Deviation Corporate overhead 22.70 34.58 50.95 17.07 10.07 Information technology and 19.81 29.91 45.01 14.54 8.69 data processing Consulting and advisory 5.78 12.73 29.97 3.74 5.23 Legal 6.19 12.43 24.71 4.16 4.71 Retail banking 13.48 29.64 55.24 9.24 10.55 FDIC assessments and other 16.95 25.54 37.34 12.26 7.58 government Other financial services 0.00 4.00 15.85 0.56 2.72 Directors' fees and other 3.45 6.99 14.60 1.91 2.85 compensation Miscellaneous 0.00 5.75 24.91 0.84 3.98 Total classified 75.05 85.72 95.35 64.32 15.73 Unclassified Source: Board of Governors of the Federal Reserve System, Consolidated Financial Statements of Bank Holding Companies (FR Y-9C data). Notes: The table reports summary statistics for 2,810 unique bank holding companies from 2008 to 2012. Annual data are as of year-end, for a total of 4,999 firm-year observations. Panel A summarizes information on the following types of noninterest expense: (i) FR Y-9C line items: eleven standardized other noninterest expense items reported in FR Y-9C Schedule HI: Memoranda, (ii) text classified: other noninterest expense items reported in Schedule HI: Memoranda as text fields, and (iii) unclassified: other noninterest expense items not classified in Schedule HI (for example, because the amounts do not exceed the reporting threshold). Panel B includes summary statistics for the nine author-defined other noninterest expense categories, which are constructed from the FR Y-9C line items and the text fields. These data are described in Section 3.2. FDIC is Federal Deposit Insurance Corporation. TABLE 3 Summary Statistics for Control Variables Industry, by Size Cohort Top 1% 95-99% 75-95% Asset shares (percentage of total assets) Total loans 42.08 59.58 64.65 Residential real estate loans 14.94 16.63 16.55 Commercial real estate loans 4.26 15.65 28.12 Commercial and industrial loans 8.64 12.54 11.20 Credit card loans 3.53 2.33 0.59 Other consumer loans 4.68 6.11 4.19 All other loans 6.03 6.32 4.00 Trading assets 15.52 1.45 0.24 Federal funds and repurchase agreements 13.67 2.20 1.24 Cash 5.49 5.76 4.41 Investment securities 12.65 20.60 22.94 Other real estate owned 0.11 0.12 0.31 Fixed assets 0.70 1.24 1.62 Investments in unconsolidated 0.33 0.18 0.09 subsidiaries Investments in real estate ventures 0.08 0.05 0.02 Intangible and other assets 8.02 6.77 3.89 Risk Risk-weighted assets (percentage of 63.85 75.08 71.72 total assets) Nonperforming loans (percentage of 2.94 1.85 2.05 total loans) Revenue composition (percentage of net operating revenue) Interest income 50.61 51.56 65.08 Trading income 7.38 1.58 0.28 Noninterest nontrading income 45.38 46.85 34.65 Fiduciary income 7.86 9.63 4.54 Investment banking fees 12.96 7.32 8.60 Service charges on deposits 5.43 6.53 7.40 Net servicing fees 3.48 1.52 0.65 Other income 15.55 21.85 13.45 Funding structure Deposits/assets (percent) 43.67 62.76 74.85 Publicly traded (percentage of sample) 76.85 79.16 60.18 Business Concentration HHI assets 0.25 0.41 0.48 HHI income 0.53 0.56 0.59 Organizational complexity Number of subsidiaries 962.25 68.78 10.76 Percentage of subsidiaries foreign 22.71 14.46 3.88 BHC is foreign-owned (percentage of 23.15 18.06 3.28 sample) Sample statistics: Regression sample (2002-12) N 604 2,405 12,197 Average number of firms 14 56 282 Average asset size (millions of 599,180 42,761 3,153 dollars) Industry, by Size Cohort Bottom 50-75% 50% Industry Asset shares (percentage of total assets) Total loans 67.84 67.57 48.39 Residential real estate loans 17.32 18.08 15.53 Commercial real estate loans 31.47 29.77 9.48 Commercial and industrial loans 10.25 9.94 9.65 Credit card loans 0.26 0.17 2.93 Other consumer loans 3.72 3.87 4.89 All other loans 4.83 5.73 5.91 Trading assets 0.04 0.04 10.89 Federal funds and repurchase agreements 1.61 2.07 9.95 Cash 4.65 4.91 5.43 Investment securities 20.56 20.46 15.34 Other real estate owned 0.42 0.49 0.14 Fixed assets 1.92 2.02 0.93 Investments in unconsolidated 0.12 0.07 0.27 subsidiaries Investments in real estate ventures 0.03 0.02 0.07 Intangible and other assets 3.19 2.97 7.24 Risk Risk-weighted assets (percentage of 72.95 71.82 67.04 total assets) Nonperforming loans (percentage of 1.83 1.95 2.51 total loans) Revenue composition (percentage of net operating revenue) Interest income 73.25 77.26 53.01 Trading income 0.08 0.09 5.44 Noninterest nontrading income 26.68 22.66 43.90 Fiduciary income 3.96 2.64 7.83 Investment banking fees 1.38 0.83 10.73 Service charges on deposits 7.84 7.79 5.93 Net servicing fees 0.47 0.52 2.69 Other income 13.03 10.88 16.66 Funding structure Deposits/assets (percent) 79.58 81.17 51.49 Publicly traded (percentage of sample) 30.81 12.69 30.02 Business Concentration HHI assets 0.51 0.51 0.29 HHI income 0.64 0.67 0.53 Organizational complexity Number of subsidiaries 6.22 4.07 18.29 Percentage of subsidiaries foreign 4.54 4.83 16.15 BHC is foreign-owned (percentage of 0.39 0.62 2.02 sample) Sample statistics: Regression sample (2002-12) N 15,181 27,830 58,217 Average number of firms 352 705 1,410 Average asset size (millions of 838 424 9,065 dollars) Individual Observations Standard Mean Deviation Asset shares (percentage of total assets) Total loans 66.44 13.36 Residential real estate loans 17.78 10.62 Commercial real estate loans 28.27 15.02 Commercial and industrial loans 10.42 6.84 Credit card loans 0.32 2.93 Other consumer loans 4.25 5.14 All other loans 5.40 7.83 Trading assets 0.20 1.75 Federal funds and repurchase agreements 2.14 3.93 Cash 4.64 4.01 Investment securities 21.35 12.38 Other real estate owned 0.36 0.89 Fixed assets 1.90 1.05 Investments in unconsolidated 0.09 1.38 subsidiaries Investments in real estate ventures 0.02 0.94 Intangible and other assets 3.19 2.11 Risk Risk-weighted assets (percentage of 71.68 11.89 total assets) Nonperforming loans (percentage of 1.65 2.65 total loans) Revenue composition (percentage of net operating revenue) Interest income 77.62 12.54 Trading income 0.19 1.14 Noninterest nontrading income 22.26 12.30 Fiduciary income 2.84 4.97 Investment banking fees 0.99 2.83 Service charges on deposits 7.87 4.56 Net servicing fees 0.60 1.58 Other income 9.77 9.32 Funding structure Deposits/assets (percent) 79.21 10.42 Publicly traded (percentage of sample) 27.75 44.78 Business Concentration HHI assets 0.52 0.13 HHI income 0.69 0.17 Organizational complexity Number of subsidiaries 15.75 139.99 Percentage of subsidiaries foreign 0.75 5.18 BHC is foreign-owned (percentage of 1.78 13.24 sample) Sample statistics: Regression sample (2002-12) N Average number of firms Average asset size (millions of dollars) Source: Board of Governors of the Federal Reserve System, Consolidated Financial Statements of Bank Holding Companies (FR Y-9C data). Notes: The table reports summary statistics for 2,810 unique bank holding companies from 2001:Q1 to 2012:Q4, a total of 58,217 firm-quarter observations. The first six columns are industry ratios (computed by first summing numerator and denominator across all firms in the relevant size class), or are statistics weighted by firm size, except for the two indicator variables "publicly traded" and "BHC is foreign-owned." Size cohorts are recalculated in each quarter. The last two columns are unweighted statistics across all firms. Note that the sample period for the regression analysis begins in 2002:Q1, not 2001:Q1, because specifications include lagged income variables from the previous four quarters. See Appendix A for variable definitions. HHI is Herfindahl-Hirschman Index; BHC is bank holding company. TABLE 4 BHC Size and the Efficiency Ratio Specification (1) (2) (3) Log assets -1.320 *** -1.892 *** -1.962 *** (0.235) (0.228) (0.226) Asset shares (percentage of total assets) Total loans -50.105 *** (7.446) Residential real -41.250 *** estate loans (7.850) Commercial real estate -55.329 *** loans (7.452) Commercial and -41.365 *** industrial loans (8.235) Credit card loans -70.539 *** (11.455) Other consumer loans -63.106 *** (8.749) All other loans -69.382 *** (8.442) Trading assets -2.154 -2.418 (18.177) (18.105) Federal funds and -20.466 * -18.125 repurchase agreements (9.526) (9.598) Investment securities -44.233 *** -46.246 *** (7.538) (7.420) Other real estate owned 511.223 *** 516.118 *** (59.960) (58.233) Fixed assets 195.591 *** 195.896 *** (31.754) (31.448) Investments in -74.519 *** -64.972 *** unconsolidated (13.295) (16.201) subsidiaries Investments in real -72.295 *** -64.503 *** estate ventures (15.963) (16.499) Intangible and other 92.308 *** 90.825 *** assets (18.720) (17.868) Revenue composition (percentage of net operating revenue) Trading income Noninterest nontrading income Fiduciary income Investment banking fees Service charges on deposits Net servicing fees Other noninterest income Funding structure Deposits/assets (percent) Public [1=yes] Business concentration HHI assets HHI income Organizational complexity Log number of subsidiaries Percentage of subsidiaries that are foreign Foreign-owned [1=yes] Constant 101.061 *** 143.904 *** 146.053 *** (3.377) (8.397) (8.432) Time fixed effects Yes Yes Yes State fixed effects Firm fixed effects [R.sup.2] 0.080 0.195 0.207 N 58,217 58,217 58,217 Specification (4) (5) (6) Log assets -2.044 *** -2.509 *** -2.631 *** (0.246) (0.239) (0.240) Asset shares (percentage of total assets) Total loans Residential real -42.777 *** -28.889 ** -30.446 *** estate loans (8.211) (8.877) (8.367) Commercial real estate -63.050 *** -46.223 *** -46.866 *** loans (9.352) (10.172) (9.729) Commercial and -43.923 *** -30.428 ** -32.324 ** industrial loans (10.014) (10.676) (10.189) Credit card loans -84.648 *** -79.998 *** -81.301 *** (10.068) (11.430) (10.945) Other consumer loans -67.709 *** -54.509 *** -53.905 *** (9.973) (10.805) (10.353) All other loans -74.193 *** -59.828 *** -61.058 *** (9.793) (10.711) (10.216) Trading assets -1.657 -3.909 -12.428 (17.966) (17.525) (16.434) Federal funds and -22.468 * -17.305 -19.636 * repurchase agreements (9.278) (9.253) (9.194) Investment securities -47.976 *** -35.704 *** -36.532 *** (7.135) (7.792) (7.487) Other real estate owned 218.441 *** 224.027 *** 227.645 *** (50.156) (52.325) (51.683) Fixed assets 213.179 *** 182.093 *** 190.166 *** (30.379) (29.035) (29.664) Investments in -56.758 *** -69.983 *** -75.613 *** unconsolidated (13.768) (12.469) (11.868) subsidiaries Investments in real -54.043 *** -66.178 *** -71.900 *** estate ventures (15.216) (14.355) (13.837) Intangible and other 55.478 ** 34.231 31.273 assets (21.111) (20.543) (19.928) Revenue composition (percentage of net operating revenue) Trading income 49.008 45.614 (26.304) (25.079) Noninterest nontrading 19.746 *** income (3.151) Fiduciary income 30.172 *** (4.570) Investment banking fees 37.832 ** (12.036) Service charges 13.020 * on deposits (6.356) Net servicing fees -1.060 (16.367) Other noninterest income 21.814 *** (3.837) Funding structure Deposits/assets (percent) Public [1=yes] Business concentration HHI assets HHI income Organizational complexity Log number of subsidiaries Percentage of subsidiaries that are foreign Foreign-owned [1=yes] Constant 144.782 *** 136.250 *** 138.941 *** (8.075) (8.276) (8.036) Time fixed effects Yes Yes Yes State fixed effects Firm fixed effects [R.sup.2] 0.247 0.258 0.261 N 58,217 58,217 58,217 Specification (7) (8) (9) Log assets -2.886 *** -2.983 *** -4.131 *** (0.271) (0.273) (0.334) Asset shares (percentage of total assets) Total loans Residential real -31.136 *** -23.170 * -21.549 * estate loans (8.579) (9.415) (8.859) Commercial real estate -47.723 *** -38.003 *** -36.868 *** loans (9.922) (10.596) (9.990) Commercial and -32.581 ** -24.657 * -25.291 * industrial loans (10.276) (10.748) (10.249) Credit card loans -80.567 *** -69.742 *** -66.710 *** (10.950) (12.164) (11.620) Other consumer loans -54.258 *** -45.243 *** -45.078 *** (10.466) (11.060) (10.654) All other loans -60.776 *** -52.092 *** -51.257 *** (10.442) (10.901) (10.321) Trading assets -10.508 -5.105 -3.128 (16.871) (18.359) (17.552) Federal funds and -18.727 * -18.063 -16.537 repurchase agreements (9.378) (9.220) (8.875) Investment securities -36.918 *** -35.623 *** -32.975 *** (7.660) (7.625) (7.248) Other real estate owned 228.260 *** 224.115 *** 223.890 *** (51.743) (52.201) (51.959) Fixed assets 197.031 *** 187.538 *** 189.759 *** (29.974) (29.496) (28.939) Investments in -74.270 *** -75.580 *** -86.452 *** unconsolidated (12.733) (13.733) (13.632) subsidiaries Investments in real -70.348 *** -29.115 -42.377 * estate ventures (14.470) (19.204) (19.251) Intangible and other 26.103 23.238 19.702 assets (20.804) (20.893) (20.411) Revenue composition (percentage of net operating revenue) Trading income 47.794 44.346 30.746 (25.203) (26.351) (25.803) Noninterest nontrading income Fiduciary income 29.695 *** 25.165 *** 27.327 *** (4.580) (4.793) (4.822) Investment banking fees 37.510 ** 33.487 ** 29.794 ** (12.140) (11.527) (11.075) Service charges 13.072 * 3.950 5.965 on deposits (6.284) (6.448) (6.294) Net servicing fees 1.707 -5.177 -1.426 (16.477) (16.582) (16.699) Other noninterest income 21.688 *** 20.629 *** 20.181 *** (3.919) (3.751) (3.716) Funding structure Deposits/assets (percent) -0.497 -2.194 -0.643 (3.119) (3.075) (2.980) Public [1=yes] 1.474 * 1.314 * 1.787 ** (0.606) (0.608) (0.621) Business concentration HHI assets -10.565 * -9.828 * (4.220) (4.093) HHI income -8.101 ** -7.205 * (3.023) (2.934) Organizational complexity Log number of subsidiaries 1.883 *** (0.395) Percentage of subsidiaries -3.813 that are foreign (5.341) Foreign-owned [1=yes] 14.895*** (2.481) Constant 142.911 *** 152.872 *** 161.137 *** (9.438) (9.380) (9.324) Time fixed effects Yes Yes Yes State fixed effects Firm fixed effects [R.sup.2] 0.262 0.264 0.271 N 58,217 58,217 58,217 Specification (10) (11) Log assets -4.151 *** -2.471 * (0.326) (1.156) Asset shares (percentage of total assets) Total loans Residential real -22.379 * -31.408 ** estate loans (8.910) (10.472) Commercial real estate -31.123 ** -45.328 *** loans (9.894) (10.340) Commercial and -15.721 -43.188 *** industrial loans (10.201) (10.512) Credit card loans -59.817 *** -36.635 (10.812) (19.167) Other consumer loans -34.291 ** -37.861 *** (10.619) (11.343) All other loans -41.791 *** -60.073 *** (10.233) (13.084) Trading assets -1.641 -9.133 (18.084) (33.833) Federal funds and -15.062 -16.323 * repurchase agreements (8.654) (7.514) Investment securities -29.990 *** -28.246 *** (7.193) (6.448) Other real estate owned 248.885 *** 264.291 *** (51.125) (54.925) Fixed assets 223.443 *** 289.553 *** (29.775) (36.789) Investments in -81.657 *** 7.582 unconsolidated (13.386) (42.429) subsidiaries Investments in real -36.690 * 58.462 estate ventures (17.849) (50.434) Intangible and other 16.813 0.999 assets (20.255) (21.117) Revenue composition (percentage of net operating revenue) Trading income 46.602 35.616 (25.765) (43.903) Noninterest nontrading income Fiduciary income 24.057 *** 34.635 *** (4.718) (8.471) Investment banking fees 35.915 *** 46.586 ** (9.925) (14.453) Service charges 14.567 * 49.324 *** on deposits (7.094) (10.446) Net servicing fees 14.615 -9.113 (14.275) (15.153) Other noninterest income 21.462 *** 0.801 (3.730) (3.656) Funding structure Deposits/assets (percent) -1.061 4.577 (2.903) (3.770) Public [1=yes] 1.418 * -0.704 (0.626) (1.705) Business concentration HHI assets -10.531 ** -10.581 * (3.907) (5.091) HHI income -8.681 ** -8.903 ** (2.902) (3.447) Organizational complexity Log number of subsidiaries 1.771 *** 1.404 ** (0.396) (0.534) Percentage of subsidiaries -5.668 2.694 that are foreign (5.139) (8.515) Foreign-owned [1=yes] 13.512 *** 15.046 ** (2.436) (5.529) Constant 157.186 *** 122.139 *** (9.372) (19.637) Time fixed effects Yes Yes State fixed effects Yes Firm fixed effects Yes [R.sup.2] 0.296 0.549 N 58,217 58,217 Source: Authors' calculations. Notes: The table presents an analysis of the relationship between size, measured by log of total assets, and efficiency ratio, defined as total noninterest expense normalized by net operating revenue. All explanatory variables are lagged by one quarter. Revenue composition variables are the rolling average for the absolute value of the income share over net operating revenue. HHI (Herfindahl-Hirschman Index) assets is the sum of squared asset shares, by asset type, and HHI income is the sum of squared four-quarter rolling average income shares, by income type. See Appendix A for further detail on controls included in the models. Models are estimated with robust standard errors and two-way clustering by firm and quarter. Standard errors are in parentheses. *** p<0.01 ** p<0.05 * p<0.1 TABLE 5 Coefficient on Log Assets, by Size Cohort (1) (2) (3) (4) All Top 1% 95-99% 75-95% Table 4, -1.320 *** 1.860 1.273 -1.790 ** Specification (1) (0.235) (1.647) (1.164) (0.687) Table 4, -1.892 *** -2.864 -0.379 -1.888 ** Specification (2) (0.228) (2.020) (1.278) (0.674) Table 4, -4.151 *** -8.018 * -5.138 *** -4.132 *** Specification (10) (0.326) (3.931) (1.442) (0.696) N 58,217 604 2,405 12,197 (5) (6) 50-75% Bottom 50% Controls Table 4, -0.768 -6.140 *** Time fixed effects Specification (1) (1.509) (1.633) Table 4, -1.914 -3.195 * Asset shares Specification (2) (1.352) (1.334) Table 4, -4.238 *** -5.055 *** All controls Specification (10) (1.204) (1.311) N 15,181 27,830 Source: Authors' calculations. Notes: The table presents an analysis of the relationship between size, measured by the log of total assets (lagged by one quarter), and efficiency ratio, defined as total noninterest expense as a percentage of net operating revenue. Each row represents the coefficient on size for specifications (1), (2), and (10) of Table 4, estimated on a subset of bank holding companies sorted by size in each quarter. Specification (1) includes time fixed effects. Specification (2) includes time fixed effects as well as controls for the percentage of assets in each broad category (asset shares). Specification (10) includes the controls from specification (2) as well as controls for types of loans, revenue composition, funding structure, business concentration, organizational complexity, and headquarters state fixed effects. Robust standard errors reported in parentheses are clustered by bank holding company and quarter. *** p<0.01 ** p<0.05 * p<0.1 TABLE 6 Alternative Measures of Operating Costs Noninterest Expense/ Risk-Weighted Assets Table 4, Specification: (1) (2) (10) Log assets 0.007 -0.044 *** -0.115 *** (0.010) (0.011) (0.013) Asset share controls Yes Yes All controls Yes [R.sup.2] 0.016 0.231 0.487 N 58,217 58,217 58,217 Noninterest Expense/ Assets Table 4, Specification: (1) (2) (10) Log assets 0.003 -0.018 ** -0.083 *** (0.006) (0.007) (0.009) Asset share controls Yes Yes All controls Yes [R.sup.2] 0.007 0.171 0.430 N 58,217 58,217 58,217 Cash Noninterest Expense/ Net Revenue (Cash Efficiency Ratio) Table 4, Specification: (1) (2) (10) Log assets -1.686 *** -2.239 *** -4.339 *** (0.231) (0.217) (0.303) Asset share controls Yes Yes All controls Yes [R.sup.2] 0.078 0.208 0.325 N 58,217 58,217 58,217 Log Noninterest Expense Table 4, Specification: (1) (2) (10) Log assets 0.993 *** 0.979 *** 0.899 *** (0.007) (0.007) (0.008) Asset share controls Yes Yes All controls Yes [R.sup.2] 0.935 0.949 0.968 N 58,192 58,192 58,192 Source: Authors' calculations. Note: The table presents an analysis of the relationship between size, measured by the log of total assets (lagged by one quarter), and different measures of efficiency. The dependent variables in the first three specifications are cash efficiency ratio, defined as total noninterest expense less goodwill impairment and amortization expense over net operating revenue; in the next three specifications, NIE-assets ratio, defined as total noninterest expense (NIE) over total assets; and in the final three specifications, NIE-RWA ratio, defined as total noninterest expense over total risk-weighted assets (RWA). For each alternative measure of efficiency ratio, specifications (1), (2) and (10) of Table 4 are presented. Specification (1) includes controls for quarter fixed effects. Specification (2) includes the controls from specification (1) as well as controls for the percentage of assets in each broad category. Specification (10) includes the controls from specification (2) as well as controls for types of loans, revenue composition, funding structure, business concentration, organizational complexity, and headquarters state fixed effects. Models are estimated with robust standard errors and two-way clustering by firm and quarter. *** p<0.01 ** p<0.05 * p<0.1 TABLE 7 Bank Holding Company Size and the Efficiency Ratio, by Component of Noninterest Expense Table 4 Log Standard Significance Specification Assets Error Level Total noninterest 1 -1.320 (0.235) *** expense 2 -1.892 (0.228) *** 10 -4.151 (0.326) *** Components of noninterest expense Compensation 1 -1.135 (0.126) *** 2 -1.472 (0.133) *** 10 -2.385 (0.175) *** Premises and 1 -0.265 (0.045) *** fixed assets 2 -0.103 (0.048) * 10 -0.365 (0.073) *** Other 1 -0.283 (0.127) * 2 -0.658 (0.125) *** 10 -1.585 (0.167) *** Amortization 1 0.181 (0.016) *** expense 2 0.164 (0.018) *** 10 0.159 (0.024) *** Goodwill 1 0.044 (0.015) ** impairment 2 0.042 (0.014) ** 10 0.017 (0.011) Components of other noninterest expense Corporate 1 -0.002 (0.073) overhead 2 -0.212 (0.063) *** 10 -0.334 (0.074) *** Information 1 -0.106 (0.044) * technology and 2 -0.150 (0.054) ** data processing 10 -0.213 (0.068) ** Consulting and 1 0.285 (0.047) *** advisory 2 0.208 (0.053) *** 10 0.053 (0.054) Legal 1 0.006 (0.035) 2 -0.022 (0.034) 10 -0.118 (0.045) ** Retail banking 1 -0.225 (0.058) *** 2 -0.068 (0.087) 10 -0.205 (0.118) FDIC assessments 1 -0.249 (0.048) *** and other 2 -0.103 (0.042) * government 10 -0.036 (0.068) Other financial 1 0.038 (0.019) * services 2 -0.022 (0.011) 10 -0.058 (0.017) *** Directors' fees 1 -0.142 (0.012) *** and other 2 -0.182 (0.015) *** compensation 10 -0.190 (0.019) *** Miscellaneous 1 0.026 (0.014) 2 0.017 (0.017) 10 -0.004 (0.022) Unclassified 1 -0.129 (0.115) other noninterest 2 -0.063 (0.102) expenses 10 -0.289 (0.134) * Adjusted Mean Coefficient [R.sup.2] (Percent) Controls /Mean (Percent) Total noninterest 0.080 Time FE -1.99 expense 0.195 66.32 Asset shares -2.85 0.296 All -6.26 Components of noninterest expense Compensation 0.048 Time FE -3.50 0.103 32.44 Asset shares -4.54 0.242 All -7.35 Premises and 0.025 Time FE -3.47 fixed assets 0.135 7.64 Asset shares -1.35 0.257 All -4.78 Other 0.111 Time FE -1.22 0.256 23.20 Asset shares -2.84 0.354 All -6.83 Amortization 0.077 Time FE 14.00 expense 0.106 1.29 Asset shares 12.68 0.163 All 12.29 Goodwill 0.031 Time FE 3.01 impairment 0.032 1.46 Asset shares 2.88 0.039 All 1.16 Components of other noninterest expense Corporate 0.018 Time FE -0.04 overhead 0.074 4.77 Asset shares -4.45 0.212 All -7.00 Information 0.006 Time FE -3.28 technology and 0.023 3.23 Asset shares -4.64 data processing 0.139 All -6.59 Consulting and 0.069 Time FE 9.92 advisory 0.097 2.87 Asset shares 7.24 0.210 All 1.84 Legal 0.008 Time FE 0.33 0.141 1.79 Asset shares -1.23 0.263 All -6.57 Retail banking 0.017 Time FE -13.59 0.108 1.66 Asset shares -4.11 0.208 All -12.38 FDIC assessments 0.242 Time FE -16.51 and other 0.393 1.51 Asset shares -6.83 government 0.536 All -2.39 Other financial 0.009 Time FE 4.86 services 0.146 0.78 Asset shares -2.81 0.211 All -7.42 Directors' fees 0.095 Time FE -221.31 and other 0.139 0.06 Asset shares -283.65 compensation 0.259 All -296.12 Miscellaneous 0.002 Time FE 5.62 0.010 0.46 Asset shares 3.68 0.042 All -0.87 Unclassified 0.004 Time FE -1.48 other noninterest 0.147 8.72 Asset shares -0.72 expenses 0.229 All -3.32 Source: Authors' calculations. Notes: The table presents an analysis of the relationship between size, measured by the log of total assets (lagged by one quarter), and the components of noninterest expense normalized by net operating revenue. The first nineteen rows present the specifications for NIE and its large components: compensation, premises and fixed assets, other, amortization expense, and goodwill impairment. The remaining rows present three specifications each for the nine subcomponents of other, as well as for unclassified expense, the total other noninterest expense less the nine constructed components of other noninterest expense. All noninterest expense components are normalized by net operating revenue. Each row presents specifications (1), (2), and (10) of Table 4 for each main component of noninterest expense. Specification (1) includes time fixed effects. Specification (2) includes time fixed effects as well as controls for the percentage of assets in each broad category (asset shares). Specification (10) includes the controls from specification (2) as well as controls for types of loans, revenue composition, funding structure, business concentration, organizational complexity, and headquarters state fixed effects. See Appendix A for further detail. The sample mean for each component is presented, and the final column is the estimated coefficient on size normalized by the sample mean for the NIE component. Robust standard errors reported in parentheses are clustered by BHC and quarter. FE is fixed effects; FDIC is Federal Deposit Insurance Corporation. *** p<0.01 ** p<0.05 * p<0.1
|Printer friendly Cite/link Email Feedback|
|Comment:||Do big banks have lower operating costs?|
|Author:||Kovner, Anna; Vickery, James; Zhou, Lily|
|Publication:||Federal Reserve Bank of New York Economic Policy Review|
|Date:||Dec 1, 2014|
|Previous Article:||The financial market effect of FOMC minutes.|
|Next Article:||Evidence from the bond market on banks' "too-big-to-fail" subsidy.|