Two articles will be attached below in order to understand the question given.

The main effects of Regulation Q Restrictions and technological changes on the scale economies for commercial banks. In the article, "An Empirical Investigation of Changes in Scale Economies for the Commercial Banking Firm", the author's final conclusion is that "In contrast to years prior to 1982, statistically significant scale economies were found in the later years for banks in branching and unit states"?

Given the current changes to regulatory and supervisory focus pointed by Yellen's article (2015), do such changes increase cost structure for smaller or larger banks? Why or Wy not?

Article 1: An Empirical Investigation of Changes in Scale Economies for the Commercial Banking Firm:

An Empirical Investigation of Changes in Scale Economics for the Commercial Banking Firm, 1979- 1986 THE 1980s REPRESENT n PERIOD of technological and reg- ulatory change perhaps unparalleled in American banking history. There have been tremendous advances in technology from the computer and communications indus tries. Regulatory restrictions regarding the types of services that banks and other depository institutions may offer have changed, blurring the traditional distinctions between nancial rms. The Regulation Q restrictions on explicit interest payments on some types of deposits were phased out. In addition, banks face growing com- petition from other rms seeking to provide bank-like services. Geographic re- strictions, which previously limited both the number of competitors banks faced and the organizational structure which banks used, have changed markedly. These tech- nological and regulatory changes are likely to affect the cost structure of commercial banks. This study examines data from 1979 to 1986 in order to update previously published results, and to determine the direction and magnitude of the possible shifts in the structure of bank costs. In particular, the focus here is on economics of scale for smaller and medium-sized banks. This paper is divided into the following sections. A brief summary of the current literature is given, followed by a description of the model used in this study, the data on which the estimations were conducted, and the analytical results. The author thanks Frederick W. Bell. T. Randolph Beard. Steven B. Caudill. Robert 1... Court, Douglas Evano', harry Frieder. Roger Garrison, loan G. Haworth, James E. Long, Stephen M. Miller, seminar participants at Auburn University and the Federal Reserve Bank of Dallas, an anonymous referee, and the editor for helpful comments on earlier drafts on this paper. Allen Berger also provided helpful guidance at the start of this project. Computer support from Economic Research Services, lnc., is gratefully acknowledged. _ 'DhNIEL M: GROPPER irTss-st'stanr professor of economics at Auburn University. Journal ofMortey. Credit, and Banking, Vol. 23, No. 4 (November 1991} Copyright 0 199] by The Ohio State University Press LITERATURE SUMMARY While the topic of cost structure of rms in the nancial services industry has generated a substantial literature. some recent studies report results that stand at odds with earlier ndings. The conclusions of some of the relatively recent studies may be of somewhat limited usefulness for current public policy guidance because they analyze data from the 19'21'Cls,1 although these are several studies which analyze data from the early 1980s.2 There are certain aspects of cost structure where a consensus appears to have been reached, and others where considerable variation exists. State regulations on bank structure, in the form of branching restrictions, appear to matter. There still exists some question in the literature as to whether scale economies exist beyond some small level of output. While most of these studies conclude that any scale economies are exhausted at relatively low levels of output, Hunter and Timme (1986) and Lawrence and Shay (1986) found economies of scale across a broad range of bank sizes. In a recent study by Noulas, Ray, and Miller (1990) economies of scale were found for banks up to $3 billion in total assets. These contrasting results may be due to the di'erent data sets used in the empirical estimations, alternative model specications, and some model misspecications.3 In almost all of the multiproduct cost studies referred to above, some version of the transcendental logarithmic cost function was used to model bank costs. The translog cost model is presented briey below. THE MODEL Production in the multiproduct rm can be represented by the transformation function: f1(Y., 1'2, . . . Ym,-X.,X2, . . .X") = 0 where Y), are the m outputs and XIf are the n inputs. If this function is strictly convex in inputs Xi, duality insures that there exists a unique corresponding cost function: C =f2(}'l,Y2, . . . Ym; P.,P2. . . .Pn) where P,- are the prices paid for the X; inputs, and Y} are the m outputs. This cost function is homogeneous of degree one and concave in factor prices, and non- decreasing in both factor prices and output quantities. These restrictions insure that there exists a unique correspondence between the cost function and the underlying l'I'l'tuese include Benston, Hanweck, and Hum rey (1982). Clark (1984), Gilligan and Smirlock (I934), Gilligan. Smirlock, and Marshall (I982, I 84), Murray and White (1983), and Nelson (1985). 2For a recent survey of this literature. see Clark (1988]. 3See Zardkoohj, Rangan, and Kolari (1986) on these misspecications. 720 : MONEY, CREDIT, AND BANKING production function.4 The cost function describes the production process as com- pletely as does the transformation function, and thus cost function estimates can be used to make inferences about production characteristics. For empirical estimation, a flexible functional form that imposes minimal re- strictions consistent with a dual relation between production and cost is desired. The translog cost function is essentially a second-order expansion in input prices and output quantities, and thus it provides a second-order approximation to an arbitrary cost function. Translog cost curves are not restricted to the monotonically increase ing or decreasing shapes imposed by the Cobb-Douglas or CES specifications. 6 This functional form has been applied to the study of financial firms by several researchers. The general form of the translog cost function is shown below: In C = do + E, a; In Y; + >, B; In P, + # E; Ex O;k In Y; In Yk + 1 E, Eh 8,, In P, In P, + E; E; Ty; In Y; In P, . (1) The Y, represent the n outputs, and the P, represent the m input prices. Linear homogeneity in input prices is insured by the restrictions E, B, = 1, E, 8; = 0, and E, TH; = 0. This cost function can be estimated alone, or the factor share equations can be derived using Shephard's lemma, and the system estimated simultaneously. The factor share equations are as shown below: a In C/a In P, = S, = B, + 2, 8jn In Pn + E, Ty In Y, (2) where S, is the share of expenditures on the jth input in total cost. Because of the restriction of linear homogeneity in input prices, the factor share equations must sum to one. To avoid singularity problems, one of the share equa- tions must be excluded in the estimation process. The parameter estimates are invariant with respect to which equation is excluded from the estimated system. 8 The overall scale economies (OSE) realized when all outputs are increased by a common factor can be obtained by differentiating the cost function with respect to all Y; OSE = E, a; + E; Ojk In Yx + E; T; In P, . (3) 4For an elaboration on this point, see Diewert (1973). For a discussion of the translog functional form see Christensen, Jorgensen, and Lau (1973), Denny and Pinto (1978), or Diewert (1973). The restrictions imposed by the Cobb-Douglas or CES forms can be tested by restricting certain parameters in the translog model. Thus constant elasticity of substitution can be a testable, rather than a maintained hypothesis. Several studies have conducted tests of this and other restrictions. See Clark (1984) and Lawrence (1989). "These include Benston, Hanweck, and Humphrey (1982), Berger, Hanweck, and Humphrey (1987), Flannery (1983), Gilligan, Smirlock, and Marshall (1982, 1984), Gilligan and Smirlock (1984), Hancock (1985), Lawrence and Shay (1986), Murray and White (1983), and Mullineaux (1978). 8It is assumed that banks operate in competitive input markets, and that they minimize cost for a given level of output.The cost measure OSE will be less than one if the bank is experiencing increasing returns to scale, since costs will rise proportionately less then output. An OSE value equal to one indicates constant returns to scale, and a value greater than one indicates decreasing returns to scale? The issue of the existence of economies of scale for the banking rm is of concern in determining appropriate regulatory policy toward bank acquisitions and mergers and in predicting future industry structure. If economies of scale exist in banking, then least cost provision of banking services would be accomplished by having relatively fewer, but larger banks. Competitive market forces will contribute to further consolidation of rms in the banking industry, as larger, more cost clftcient rms drive smaller, less efcient rms from the market. THE DATA Data from the Functional Cost Analysis program of the Federal Reserve System for the years 197986 were used to estimate the cost function parameters. This program is administered to participating nancial institutions on a voluntary basis by their regional Federal Reserve bank, and coordinated nationally by the Federal Reserve Bank of New York. This data set has been used widely in previous research, and contains detailed information on the number of accounts of various types that the institution has, as well as their dollar volume. In addition, information on the inputs used in the' bank production process, such as the number of officers and employees, their average salaries and benets, and the number and type of ofces the institution has is reported in the FCA program. A functional allocation of some costs across the institution's activities is made, although this allocation by the reporting institution's employees is not used in this study. The advantages of using FCA data are that detailed information on both inputs and outputs is provided in a standardized format and its use enables comparison with previously published re- search; disadvantages include nonrandom sampling of the nation's nancial institu- tions and the omission of the largest institutions. Another drawback is that the data made available to the public has some information masked to protect the conden- tiality of the participating institutions. For this reason, holding company affiliation is not revealed, the state in which the institution is located is unknown, and institu- tions cannot be tracked from one year to the next. However, detailed information on the various inputs used and the size and number of deposit, loan, and trust accounts is not available from other sources, such as Call Report data. The vast majority of the roughly fourteen thousand commercial banks in the United States over the time period of this study had total assets comparable in size to the banks in the FCA program. Although the very largest banks are not represented, banks from less than $10 million in total assets to over $2 billion are in the FCA data set. While the FCA data should not be used to draw conclusions about the nation's largest banks, patterns found in the FCA data may well pr0vide insights about 9111c variance of OSE is calculated as follows: For (OSE) = 2.- Var (of) + 2 2.- 2,- Cov (tinny). 722 : MONEY, CR-IT, AND BANKING trends affecting the smaller and medium-sized banks which make up over 90 percent of the rms in the U.S. banking industry. Since participation in the FCA program is voluntary, it seems plausible to think that the FCA banks may be more cost con- scious, and possibly more eicient, than nonparticipating banks. In a study whose purpose is to investigate possible changes in the cost structure of banking, the nonrandomness of FCA data is not necessarily undesirable. If it is believed that the FCA banks are likely to be more cost efficient than other banks, it may be possible to detect cost function shifts earlier among FCA banks than among other banks, as the FCA banks adapt more rapidly to regulatory and technological changes. In this study, the banking rm is viewed as a nancial intermediary producing various types of loans and investments using labor, capital, and funds as inputs. Some cost studies have excluded the interest costs of funds and focused on operating costs alone. Others have included interest costs, and they are included here follow- ing the intermediation approach outlined in Mester (1987a) which contains a discus- sion of both approaches. Total costs are the sum of labor. capital, and interest costs. Consistent with the intennediation approach, the dollar volume of accounts is the desired output measure for this study. The output categories of investments, total loans, and trust accounts are used.lo For the cost function estimation, the prices of the inputs labor, capital, and funds are used. The sum of wages, salaries, and benefits divided by the total number of employees and officers provides an approximation to an annual price of labor. Similarly, the total interest payments divided by the quantity of funds on which interest was paid gives an approximate price of funds. The price of capital is a composite weighted average price of physical and nancial capital following the general procedure used by Hancock (1985). It includes occupancy costs divided by the book value of building and equipment for physical capital. The cost of nancial capital includes dividends paid on common and preferred stock, and interest on capital notes and debentures.\" Each of the above measures give estimates of average, rather than marginal prices. If the marginal dilfers greatly from the average, this would be a problem. However, to the extent that individual nancial institutions are price takers in the market for labor, funds, and capital, the marginal and average prices paid for additional units of input would be similar, so that expansion of output could occur for each rm individually without signicantly raising the prices of the necessary inputs. If all institutions were to simultaneously attempt to expand output, the same conclusion might not hold, as pecuniary externalities could be important. Previous research has found that state branching regulations have an important impact on the rm's cost function. Since this nding was supported in preliminary l0For zero levels of output, which occurred for some banks with respect to the trust account output, .01 was substituted for zero. Additional estimates were made where Smaller constants were substituted, without materially affecting the empirical results reported in Tables I and 2. \"Data limitations prevent the inclusion of changes in stock prices which are an important component of the retum to investors, and thus may affect the costs of raising capital for banks. An alternative capital price specication following Mester (1987b) was tried in preliminary analyses, but that specication did not change the general pattern of increasing economies of scale reported in 1hbles l and 2. DANIEL M. GROPPER : 723 analyses for this study, separate equations are estimated for banks in unit and branching states. The number of total offices (full and limited service) the bank has is included in the model for banks in branching states, consistent with previous research which has found this to be an important element affecting cost. RESULTS The raw, unedited data from the Functional Cost Analysis data set contain several problems. For example, several banks reported having zero offices, some reported interest payments in excess of total account balances, and others reported negative account balances. The raw data included information on 4,390 banks, which was reduced to 4,277 after screening for the above problems. This results in a sample of from 514 to 735 banks in each year. Estimates were obtained using Zellner's (1962) Seemingly Unrelated Regressions (SUR) technique on a system of equations con- sisting of the total cost equation and two of the three cost share equations. Estimates of overall economies of scale are constructed from the parameter estimates. 12 Esti- mates of economies of scope are not calculated in this paper. '3 In calculating the scale estimates, the input prices are held constant at the sample means, and the output levels are allowed to vary. In order to examine economies of scale for banks of different sizes, scale estimates were conducted over a range of asset size catego- ries. The OSE measures were calculated at the output means for each size category. Estimation of a separate model for each of these years allows the comparison of the behavior of the cost function across the time period. The results of the SUR estimates of overall scale economies for banks in branch banking states are shown in Table 1. Some general patterns can be observed. Scale economies generally diminish as banks get larger within each yearly sample. Statis- tically significant diseconomies of scale are found for banks at the larger end of the TABLE 1 OVERALL ECONOMIES OF SCALE BY YEAR: BRANCH BANKING STATES Total Asset 1979 1980 1981 1982 983 1984 1985 1986 less than $50 .9504 9340 9223 9433 9270 9348 9151 .9270 (4.30) (5.65) (6.09) (3.32) (3.59) (3.22) (4.64) (4.74) $50-100 .9689 9613 9602 9685 .9300 9436 9316 .9370 (3.34) (4.23) (4.01) (2.33) (4.59) (3.70) (4.83) (4.96) $100-200 .9824 .9814 9904 9899 .9345 .9509 .9462 9446 (2.07) (2.15) (1.05) (0.87) (4.94) (3.73) (4.17) (4.62) $200-300 .9922 1.0007 1.0151 1.0073 .9392 .9577 .9573 .9508 (0.87) (-0.07) (-1.46) (-0.59) (4.40) (3.11) (3.31) (4.11) $300-500 1.0046 1.0136 1.0353 1.0234 .9430 .9628 .9668 .9576 (-0.46) (-1.29) (-3.23) (-1.71) (3.51) (2.32) (2.26) (3.30) over $500 1.0253 1.0431 1.0819 1.0541 9492 .9728 9871 .9747 (-1.87) (-2.96) (-5.22) (-2.75) (2.28) (1.17) (0.68) (1.43) NOTE: T-statistics in parentheses. 7-tests were constructed to test the hypothesis that OSE = 1. Total assets are in millions of dollars. OSE values less than ] indicate economies of scale; values greater than I indicate diseconomies of scale.1'24 .' MONEY. CREDIT, AND BANKING TABLE 2 Ovens LL Economies or SCALE av Yeas: UNIT BANKING STATES Total Neel: I979 I980 I981 I982 198} I984 1985 [986 less than $50 .96 l6 .9759 .9681 .8904 .8849 .9398 .9736 .8504 (3.04) (1.01) (1.70) (5 37} (5.06) (2.16) (0.87) (5.00) 850 l00 .9806 .9853 .9854 .9183 .9066 .9633 .9894 .9005 { I .94) ( I .48) (0.98} (5.62) {5.48) (2.01) (0.52) (4. 8?) 3100200 .9925 .9925 .9928 .945l .9252 .9806 .9948 .9392 {0.74) (0.73) (0.52) (3.82) (4.65) (I .02) (0.26) (2.2?) 3200-300 1.0015 1.0013 1.0006 .97l8 .9431 .9332 HI)\" .9692 (0.12) (-0.10) (-0.04) (1.56] (3.20] (0.51) ("0.07) (0.33) 3300500 I.0| I3 [.0000 1.0134 .9890 .95l6 1.00\" [.0106 .9875 (-0.79) (\"0.00) [ 0.65) (0 .60) (2.28) (-0.05) ( #033) (0.29) over $500 1.0246 [.0086 1.0Q30 1.003] .9803 I .0004 1.0111 1.0288 (-1.35) (\"0.39) (0.87) (0.1 I) (0.69) (-0.50) (-0.25) (-0.51) let; Tsu'listics in sea. T-tests m crawled. to test the hypothesis 00110.55 = I. Total assets are in millions ofdollm. USE values less Ilum l i ' at: economies of scale; values gleam than I indicate diseoonomies ofscale. sample size in the early years, but not after I982. Statistically signicant economies of scale are found for the smaller size categories (less than $100 million in total assets) of banks in all years. The range aver which signicant economies of scale are found increases up to $500 million in total assets in 198386, and extends over $500 million in 1983. The overall pattern appears to be one of increasing economies of scale over the entire time period. The scale economies estimates for unit banking states are shown in Table 2. The general pattern is similar to that for branching states, although the pattern is not as strong. Statistically signicant economies of scale are found for banks with less than $50 million in total assets in 1979, 1982, 1983, 1984, and 1986. The extent of scale economies increases in most of the later years, with statistically signicant economies of scale found for banks through the $100$200 million size range in 1982, 1983. and 1986. While the general pattern of increasing economies of scale is found for banks in both unit and branching states, the pattern appears stronger for the branching state banks. One possible explanation for this result is the dierence in the size of the branching and unit state samples. The branch state sample contains roughly seventy to two hundred more banks than the unit state sample in each year. The larger number of observations should result in cost function parameter estimates that are close to the \"true\" underlying functional relationships. There also was a greater proportional sample size decline in the unit state sample which makes any shift in the underlying cost function relationships less likely to be detected by statistical investigation. Another possibility is that the branching restrictions effectively con- strain efcient bank structure, and this constraint prevents larger banks from exploit- l2Parameter estimates are available from the author upon request. \"The translog functional form is not well suited to the calculation of economies of scope. Scope economies were calculated from a set of hybrid o'anslog parameter estimates. As a referee has pointed out, the scope estimates are sensitive to the choice of it in the Box-Cox metric used in the hybrid musing MI. Rather than fully investigate this issue, this paper focuses on scale economies. DANIEL M. GROPPER : 725 ing scale economies which could be realized from operating a branching organization. POOLED MODEL RESULTS AND COST SHIFTS An alternative to using the results from the yearly regressions to study possible shifts in production technology across the 1979-1986 time period is to pool the annual data, and compare the results across pools. While pooling these data raises a number of issues, it was done to provide a comparison of results with the annual analyses. 14 In the pooled data for banks in both unit and branching states, the asset size levels through which significant economies of scale were found and those where disec- onomies of scale were realized both increased over the time periods. This general pattern is the same as that exhibited in the annual regressions, although, as ex- pected, the pattern from the pooled data appears to be somewhat stronger. SHIFTS ACROSS PERIODS Two methods were used to examine the possibility of structural change in the cost function across time periods. The first was to construct a joint test that all param- eters in each model were the same across time periods. The hypothesis that all parameters are the same in each period is rejected at the 1 percent level in each case, indicating structural changes in the cost function. The second method was to focus on the estimated scale economies measures, and test whether those were the same across periods. For the respective sample mean banks in both branch and unit states, there were statistically significant increases in economies of scale between the 1979-82 and 1983-86 time periods. This is also true between the 1979-81 and 1984-86 time periods. In general, the overall trend again appears to be one of increasing economies of scale. SUMMARY AND CONCLUSIONS The behavior of the cost function for a sample of commercial banks over the years 1979-1986 has been investigated. For the years prior to 1982, little evidence to 14These issues include but are not limited to the stability of the sample size across years, the actual banks in the sample in each year, the geographic distribution of the sample in each year, and the effects of inflation. The pooled analysis was done with two pooling periods: 1979-82 and 1983-86. Another analysis was done with three pooling periods: 1979-81, 1982-83, and 1984-86. The pooled analyses were conducted after adjusting all output quantities, expenditures and prices to constant 1982 prices, and repeated without making any adjustments. These analyses are available on request from the author. There are only minor differences in the OSE estimates from the adjusted and unadjusted analyses. The primary point, that there are statistically significant increases in economies of scale over the 1979-1986 time period, is unaffected by adjusting the regression variables for the effects of inflation. This conclusion is also true for the annual analyses.726 : MONEY, CREDIT, AND BANKING suggest that economies of scale exist beyond small levels of output was found. This result is consistent with most of the empirical literature on bank costs. In contrast to the earlier years and previous literature, statistically significant scale economies were found in the later years for banks in branching and unit states. The degree of scale economies also increased over the 1979-86 time period. These results suggest that the effect of recent regulatory and technological changes may have been to give larger banks a cost advantage over that which existed in previous years. These results also indicate that there may be increased cost pressure for smaller banks to become larger, either through mergers and acquisitions or though internal growth. This may lead to further consolidation pressures within the industry, and reductions in the overall number of banking firms. Virtually all of the previous cost studies concluded with the caution that the results they had found might not apply as the deregulation process continued, and further technological developments occurred. The results found in the present study indicate that those authors were correct. LITERATURE CITED Benston, George J., Gerald A. Hanweck, and David Humphrey. "Scale Economies in Bank- ing: A Restructuring and Reassessment." Journal of Money, Credit, and Banking 14 (November 1982), 435-56. Berger, Allen N., Gerald A. Hanweck, and David Humphrey. "Competitive Viability in Banking: Scale, Scope, and Product Mix Economies." Journal of Monetary Economics 20 (1987), 501-20. Christensen, Laurits, Dale Jorgensen, and Lawrence Lau. "Transcendental Logarithmic Pro- duction Frontiers." Review of Economics and Statistics 55 (February 1973), 28-45. Clark, Jeffrey. "Estimation of Economies of Scale in Banking Using a Generalized Func- tional Form." Journal of Money, Credit, and Banking 16 (February 1984), 53-68. "Economies of Scale and Scope at Depository Financial Institutions: A Review of the Literature." Federal Reserve Bank of Kansas City Economic Review (Sep- tember/October 1988), 16-33. Denny, M., and C. Pinto. "An Aggregate Model with Multi-Product Technologies." In Production Economics: A Dual Approach to Theory and Applications, edited by Fuss and Mcfadden. Amsterdam: North-Holland, 1978. Diewert, W. E. "Functional Forms for Profit and Transformation Functions." Journal of Economic Theory 6 (June 1973), 284-316. Flannery, Mark J. "Correspondent Services and Cost Economies in Commercial Banking." Journal of Banking and Finance 7 (March 1983), 83-99. Gilligan, Thomas, and Michael Smirlock. "An Empirical Study of Joint Production and Scale Economies in Commercial Banking." Journal of Banking and Finance 8 (March 1984), 67-77. Gilligan, Thomas, Michael Smirlock, and W. Marshall. "Multiproduct Cost Structures in Commercial Banking." Proceedings of a Conference on Bank Structure and Competition. Federal Reserve Bank of Chicago, 1982. "Scale and Scope Economies in the Multi-Product Banking Firm." Journal of Monetary Economics 13 (July 1984), 393-405. Hancock, Diana. "The Financial Firm: Production with Monetary and Nonmonetary Goods." Journal of Political Economy 93 (October 1985), 859-79.Let me begin by thanking the organizers for inviting me to participate in this important dialogue on the role of finance in society. The financial sector is vital to the economy. A well-functioning financial sector promotes job creation, innovation, and inclusive economic growth. But when the incentives facing financial firms are distorted, these firms may act in ways that can harm society. Appropriate regulation, coupled with vigilant supervision, is essential to address these issues. Unfortunately, in the years preceding the financial crisis, all too many firms took on risks they could neither measure nor manage. Leverage, interconnectedness, and maturity and liquidity transformation escalated to dangerous levels across the financial system. The result was the most severe financial crisis and economic downturn since the Great Depression. Almost 9 million Americans lost their jobs, roughly twice as many lost their homes, and all too many households ended up underwater on their mortgages and overburdened with debt. To be sure, some individuals and families borrowed unwisely, but too often financial institutions encouraged the behavior that resulted in such excessive debt. In my remarks today I will discuss some important reasons why the incentives facing financial institutions were distorted and the steps that regulators are taking to realign those incentives. The Important Role of the Financial Sector Before discussing the incentives that contributed to the buildup of risk at financial institutions, I would like to highlight the important contributions that the financial sector makes to the economy and society. First and foremost, financial institutions channel society's scarce savings to productive investments, thereby promoting business formationand job creation. Access to capital is important for all firms, but it is particularly vital for startups and young firms, which often lack a sufficient stream of earnings to increase employment and internally finance capital spending. Indeed, research shows that more highly developed financial systems disproportionately benefit entrepreneurship.' The financial sector also helps households save for retirement, purchase homes and cars, and weather unexpected developments. Many financial innovations, such as the increased availability of low-cost mutual funds, have improved opportunities for households to participate in asset markets and diversify their holdings. Expanded credit access has helped households maintain living standards when suffering job loss, illness, or other unexpected contingencies. Technological innovations have increased the ease and convenience with which individuals make and receive payments. The contribution of the financial sector to household risk management and business investment, as well as the significant contribution of financial-sector development to economic growth, has been documented in many studies. Such research Recent reviews have highlighted potential costs of a distorted financial sector, but such reviews also emphasize the range of both theoretical and empirical work that has documented the many ways in which the financial sector can support economic efficiency; see, for example, Greenwood and Scharfstein (2013) and Zingales (2015). Guiso, Sapienza, and Zingales (2004) discuss evidence that financial development supports entrepreneurship, and Fort and others (2013) examine the importance of financing for business formation and young firms. Greenwood and Scharfstein (2013) discuss how an important fraction of growth in the U.S. financial sector reflects the greater demand of households for asset management services and credit. Malkiel (2013) considers similar issues and reviews how improved access to low-cost investment options has benefited households. Krueger and Perri (2006) analyze how an increase in access to credit contributed to households' ability to smooth spending despite substantial income volatility. Changes in payment technologies have been rapid, and an area of particular interest is the fast growth of mobile payment and financial service technologies. The Federal Reserve has conducted several surveys to understand these developments, and the most recent results are discussed in Board of Governors (2015). A substantial body of research finds that financial development supports economic growth, including Goldsmith (1969), King and Levine (1993), Rajan and Zingales (1998), and Levine (2005). It is noteworthy that this research emphasizes differences across countries, and that the United States is among the most financially developed countries in the world.shows that, across countries and over time, financial development, up to a point, has disproportionately benefited the poor and served to alleviate economic inequality. Distorted Incentives in the Financial Sector Despite these benefits, as we have seen, actions by financial institutions have the potential to inflict harm on society. Instead of promoting financial security through prudent mortgage underwriting, the financial sector prior to the crisis facilitated a bubble in the housing market and too often encouraged households to take on mortgages they neither understood nor could afford. Recent research has raised important questions about the benefits and costs of the rapid growth of the financial services industry in the United States over the past 40 years." A combination of responses to distorted incentives by players throughout the financial system created an environment conducive to a crisis. Excessive leverage placed institutions at great risk of insolvency in the event that severe, albeit low-probability, problems materialized. Overreliance on fragile short-term funding by many institutions left the system vulnerable to runs. And excessive risk-taking increased the probability that severe problems would, in fact, materialize. Moreover, regulators--and the structure of the regulatory system itself--did not keep up with changes in the financial sector and were insufficiently attuned to systemic risks. Once concerns began to develop about escalating losses at large firms, insufficient liquidity and capital interacted in an adverse " Beck, Demirguc-Kunt, and Levine (2007) show that financial development reduces poverty and inequality in a study examining evidence across countries. Zingales (2015) raises a number of questions regarding ways in which distortions in the financial sector may contribute to "rent seeking" activity that may promote inefficiency. Philippon and Reshef (2012) examine trends in compensation in the financial sector and the contribution of such trends to the increase in income inequality in the United States in recent decades. Philippon and Reshef (2013) and Cecchetti and Kharroubi (2012) revisit the links between financial development and economic growth, focusing particularly on these relationships around periods of rapid growth in the financial sector or among economies with a large financial sector.feedback loop. Funding pressures contributed to \"re sales\" of nancial assets and losses, reducing capital levels and heightening liquidity pressures--culminating in the near collapse of the nancial system in late 2008. Capital and liquidity Several factors encouraged excessive leverage, including market perceptions that some institutions were \"too big to fail.\"3 Financial institutions also had an incentive to engage in regulatory arbitrage, moving assets to undercapitalized off-balance-sheet vehicles. The complexity of the largest banking organizations also may have impeded market discipline. In addition, nancial intermediation outside of the traditional banking sector grew rapidly in the years up to 2007, leaving gaps in the regulatory umbrella. And conicts in the incentives facing managers, shareholders, and creditors may have induced banks to increase leverage.9 To strengthen banks' resilience, the Federal Reserve and the other banking agencies have substantially increased capital requirements. Regulatory minimums for capital relative to risk-weighted assets are signicantly higher, and capital requirements now focus on the highest-quality capital, such as common equity. In addition to risk- based standards, bank holding companies and depositories face a leverage ratio requirement. Also, signicantly higher capital standards-both risk-weighted and leverage ratios--are being applied to the most systemically important banking organizations. Such surcharges are appropriate because of the substantial harm that the B For a review of many factors that may have contributed to leverage in the nancial sector and a discussion of how, in some cases, these factors reect distortions that imply leverage was excessive, see Admati and others (2013a). 9 The notion that \"agency problems\"--that is, conicts in the interests of managers and various stakeholders in rms--may contribute to excessive debt has a long history, most notably following the notion of \"debt overhangs\" from Myers (1977). Hanson, Kashyap, and Stein (2011) emphasize the potential importance of this issue for the nancial sector. Admati and others (2013b) present a related mechanism. failure of a systemic institution would inict on the nancial system and the economy. Higher capital standards provide large, complex institutions with an incentive to reduce their systemic footprint. We are also employing annual stress tests to gauge large institutions' ability to weather a very severe downturn and distress of counterparties and, importantly, continue lending to households and businesses. Firms that do not meet these standards face restrictions on dividends and share buybacks. As a result of these changes, for the largest banks, Tier 1 common equity--the highest-quality form of capital--has more than doubled since the nancial crisis. New liquidity regulations will also improve incentives in the nancial system. Prior to the crisis, institutions' incentives to rely on short-term borrowing to fund investments in riskier or less liquid instruments were distorted in two important ways. First, many investors were willing to accept a very low interest rate on short-term liabilities of nancial institutions or on securitizations without demanding adequate compensation for severe-but-unlikely risks, such as a temporary loss of market liquidity. Perhaps these rms expected government support or simply considered illiquidity a very remote possibility. Second, institutions' attempts to shift their holdings once concerns about credit or liquidity risk arose created a re-sale dynamic that amplied declines in market values, causing unanticipated spillovers onto other institutions and across markets.\" Recently implemented regulations aim to strengthen liquidity. For example, a new liquidity coverage ratio requires internationally active banking organizations to hold sufficient high-quality liquid assets to meet their projected net cash outflows during a 30- day stress period. A new process--the Comprehensive Liquidity Analysis and Review-- sets supervisory expectations for liquidity-risk management and evaluates institutions' practices against these benchmarks. A proposal for a net stable funding ratio would require better liquidity management at horizons beyond that covered by the liquidity coverage ratio. A proposed capital surcharge for the largest firms would discourage overreliance on short-term wholesale funding. Also, the Securities and Exchange Commission has adopted changes in regulations that may help avoid future runs on prime money market mutual funds (that is, money funds that invest primarily in corporate debt securities). And reforms in the triparty repo market have reduced risks associated with intraday exposures. Large, complex institutions and too big to fail In the aftermath of the crisis, the Congress tasked the banking regulators with challenging and changing the perception that any financial institution is too big to fail by ensuring that even very large banking organizations can be resolved without harming financial stability. Steps are under way to achieve this objective. In particular, banking organizations are required to prepare "living wills"--plans for their rapid and orderly resolution in the event of insolvency. Regulators are considering requiring that bank holding companies have sufficient total loss-absorbing capacity, including long-term debt, to enable them to be wound down without government support. In addition, the Federal Deposit Insurance Corporation has designed a strategy that it could deploy(known as Single Point of Entry) to resolve a systemically important institution in an orderly manner. The crisis also revealed that risk management at large, complex financial institutions was insufficient to handle the risks that some firms had taken. Compensation systems all too frequently failed to appropriately account for longer-term risks undertaken by employees. And lax controls in some cases contributed to unethical and illegal behavior by banking organizations and their employees. The Federal Reserve has made improving risk management and internal controls a top priority. For example, the Comprehensive Capital Analysis and Review, which includes the stress tests that I mentioned, also involves an evaluation to ensure firms have a sound process in place for measuring and monitoring the risks they are taking and for matching their capital levels to those risks. Also, supervisors from the Fed and other agencies have pressed firms to improve their internal controls and to make their boards of directors more directly responsible for compensation decisions and employee conduct. Changes to Regulatory and Supervisory Focus As I noted, the financial crisis revealed weaknesses in our nation's system for supervising and regulating the financial industry. Prior to the crisis, regulatory agencies, including the Federal Reserve, focused on the safety and soundness of individual firms-- as required by their legislative mandate at the time--rather than the stability of the financial system as a whole. Our regulatory system did not provide any supervisory watchdog with responsibility for identifying and addressing risks associated with activities and institutions that were outside the regulatory perimeter. The rapid growth of the "shadow" nonbank financial sector left significant gaps in regulation.In response, the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010 (Dodd-Frank Act) expanded the mandate and authority of the Federal Reserve to allow it to consider risks to financial stability in supervising financial firms under its charge. Within the Federal Reserve System, we have reorganized our supervision of the most systemically important institutions to emphasize what we call a "horizontal perspective," which examines institutions as a group and in comparative terms, focusing on their interaction with the broader financial system. We also created a new office within the Fed to identify emerging risks to stability in the broader financial system--both the bank and nonbank financial sectors--and to develop policies to mitigate systemic risk. The Dodd-Frank Act created the interagency Financial Stability Oversight Council, chaired by the Treasury Secretary, and the Federal Reserve is a member. It is charged with identifying systemically important financial institutions and systemically risky activities that are not subject to consolidated supervision and designating those institutions and activities for appropriate supervision. And it is charged with encouraging greater information sharing and policy coordination across financial regulatory agencies. Where We Stand My topic is broad, and my time is short. Let me end with three thoughts. First, I believe that we and other supervisory agencies have made significant progress in addressing incentive problems within the financial sector, especially within the banking sector. Second, policymakers, including those of us at the Federal Reserve, remain watchful for areas in need of further action or in which the steps taken to date need to be adjusted. And, third, engagement with the broader public is crucial to ensuring that any future steps move our financial system closer to where it should be. Active debate anddiscussion of these issues at this conference and in other forums is important to improve our understanding of the challenges that remain