6.1. Some General Observations
Given the concern of the bank regulatory authorities since the early 1980s with regard to brokered deposits, it is important to consider the impacts of such deposits on bank performance, failures, and failure costs. To begin, we use data for the top 100 banks (by ratio of brokered deposits-to-total deposits) from Table 1
to examine some fairly simple and suggestive relationships among variables. In particular, we examine the relationship between the ratio of brokered deposits-to-total deposits and: (1) the number of branches a bank operates; (2) a bank’s efficiency ratio; and (3) its capital-to-asset ratio.
shows a significantly negative relationship between the number of branches and the ratio of brokered deposits-to-total deposits. Banks, of course, can secure deposits through branches or brokers, or some combination of the two sources of funds; and they will incur either the costs of operating their branches or the fees of acquiring brokered deposits, or both costs if they are securing funds from both sources. It is important to acknowledge that the business models of some banks make it less costly to rely on brokers than to operate a network of branches. This may help explain the finding that the higher a bank’s brokered deposits ratio, the fewer branches it operates.
looks at the relationship between a bank’s efficiency ratio and its ratio of brokered deposits-to-total deposits, and shows a significantly negative relationship between the two, consistent with a view that banks with higher brokered deposit ratios operate more efficiently than those with lower ratios. This finding is consistent with the previous finding insofar as banks with fewer branches are most likely to incur lower non-interest expenses.
examines the relationship between a bank’s capital-to-asset ratio and its ratio of brokered deposits-to-total deposits. It shows a significantly positive relationship, indicating that the higher is the brokered deposits ratio, the higher is the capital asset ratio. This finding, coupled with the previous two findings, suggests that on average that greater use of brokered deposits is associated with higher capital ratios and better efficiency ratios for the top 100 banks.
These findings also indicate that brokered deposits may be an important source of funds for some banks, depending on their business models, and do not pose the types of problems of concern to regulators. Indeed, brokered deposits may enable some banks to operate more safely and soundly.
However, it is important to consider more rigorous studies of the impacts of brokered deposits on bank performance, bank failures, and bank failure costs, which we do next.
Before we examine the impacts of brokered deposits on bank performance, bank failures, and bank failure costs, it is important to clarify that banks hold two types of deposits: brokered deposits and core deposits. Through 2010 (when the definition underwent a change; we address this shortly), the Federal Financial Institutions Examination Council (FFIEC) included as “core accounts” all demand and savings deposits, including money market deposits, NOW, and ATS accounts, other savings deposits, and time deposits in amounts under $
100,000 (FDIC 2011, p. 115
Regulatory authorities do not, of course, treat core deposits and brokered deposits equally; they have historically perceived, and categorized, core deposits as stable, less costly funds obtained from local customers who maintain relationships with the institution. Meanwhile, they perceive brokered deposits to be volatile because they draw customers more broadly, mainly in search of yield.
However, this view is not necessarily supported if one looks at the characteristics of both types of deposits. Core deposits typically have few or no restrictions on early withdrawals—which makes the banks more vulnerable to “runs” during periods of uncertainty. Brokered deposits, on the other hand, do not permit early withdrawals unless the depositor dies or is declared incompetent by a court of law, making it impossible for these depositors to flee. However, even considering these factors, the volatility of any deposits should depend ultimately on whether the bank itself is well capitalized and whether the rates it offers on its various deposits are competitive in the marketplace.
Despite these historical perceptions of stability versus volatility, the FDIC nonetheless stated in 2011 that “examiners do not necessarily view the presence of any certain source of funding as inherently bad,” and adds that “there should be no particular stigma attached to the acceptance of brokered deposits per se and the proper use of such deposits should not be discouraged” (FDIC 2011, p. 32
). However, brokered deposits are certainly not treated as core deposits. Worse, this treatment is not justified by any consensus based upon a thorough quantitative analysis.
looks at core and brokered deposits in terms of their relative roles in funding, respectively, total deposits and bank assets. Looking first at core deposits, we see that they constitute a significantly larger portion of total deposits than they do of total assets. Over the period 1992−2018, core deposits ranged from 67% to 83% of total deposits. The percentage declined over the first half of the period to its lowest value of 67% in 2006, and then increased during the second half to end at 78% in 2018. The pattern is similar when we examine the role of core deposits in funding total assets, but they are on average 22% lower. This means that all non-core sources of funds, and not just brokered deposits, are quite important in supporting the assets of banks.
Turning to brokered deposits, Figure 12
shows that their percentage of both total deposits and total assets generally increased over the period, beginning in 1992 with values of 2% and 1%, respectively, and ending in 2018 with values of 8% and 6%. Of note, the core deposit ratio generally decreased leading up to and during the financial crisis, while the brokered deposit ratio generally increased over the same period, as many banks turned to not only brokered deposits but non-deposit liabilities as sources of funds. However, toward the end of the period, the brokered deposit ratios were tending to level off, even as the core deposit ratios were still increasing.
Since non-core sources of funds constitute a relatively significant portion of total assets, a comparison of these funds with brokered deposits is helpful. Figure 13
shows the percentage of total assets funded by core deposits, insured brokered deposits, non-insured brokered deposits, and equity capital and other liabilities. The figure shows that brokered deposits, including both insured and non-insured deposits, fund a relatively small portion of total assets, as noted in Figure 12
Of interest here, insured brokered deposits were included in core deposits through 2010. However, on 31 March 2011, a new definition of core deposits included time deposits up to $
250,000, but excluded brokered deposits under $
250,000. At the time, those insured brokered deposits accounted for 81% of total brokered deposits. The FDIC study behind the decision cites a number of research papers on core and brokered deposits, but it did not rely on comparable data or variables, making the reason for the change, and the evidence supporting it, unclear (FDIC 2011
). In the study, some of the research papers cited that find core deposits to be beneficial include insured brokered deposits in core deposits. Other cited papers do not distinguish between insured and non-insured brokered deposits. In addition, the study states that one deposit characteristic that can contribute to potential problems is being uninsured. Furthermore, it does not appear from the empirical work that insured and non-insured brokered deposits are included as separate variables, or that the regressions separately control for all non-core deposit variables.
6.2. Some Research Studies
Over the years, of course, many studies have focused on the causes of bank failures, the costs of resolving bank failures, and bank instability. Here, we examine and review a number of these studies to see if there is any consensus in the role brokered deposits play in bank performance, bank failures, and bank failure costs, and why. It seems appropriate to focus first on two studies undertaken by bank regulatory authorities themselves. The first, by the FDIC in 2011, is titled “Study on Core Deposits and Brokered Deposits” (FDIC 2011
), while the second was released the same year by the Office of Inspector General (OIG) of the Board of Governors of the Federal Reserve System and titled “Summary Analysis of Failed Bank Reviews” (Office of Inspector General 2011
The FDIC study identifies three of the most important potential problems deposits can pose. The first, referred to as “rapid, risky growth,” occurs if a bank acquires deposits too easily and thus has more funds than it can prudently invest (e.g., if a bank pays a higher rate on its deposits than it earns on its loans, it will ultimately fail). The second, “deposit volatility,” is the greater likelihood that a depositor will withdraw funds for higher rates elsewhere when the bank is under stress, resulting in the greater risk that the bank will encounter liquidity problems. It should be noted that the FDIC stated that “[t]he net effect of brokered deposits on liquidity is, therefore, uncertain and variable for different types of institutions and in different regions” (see US Department of the Treasury 1991, pp. IV−8
). The third problem, “lower franchise value,” occurs when potential buyers of failed banks find that some kinds of deposits—those with low relative costs, those that offer a continuing customer relationship, and those likely to remain at the bank after acquisition—more attractive than others and demand discounts on the “more volatile” brokered deposits.
The FDIC relies on five factors to determine whether brokered deposits create any or all of these three potential problems: (1) deposit accounts that pay high interest rates (which are likely to exhibit all three of the problems)21
; (2) many forms of brokered deposits (which can be acquired quickly and in bulk); (3) deposits that are not based on a customer relationship (which again are likely to present all three problems)22
; (4) uninsured deposits (which can exacerbate liquidity problems at a weakened bank); and (5) the duration of a deposit (which can present or mitigate the problem of a deposit leaving a bank for higher rates or when the bank is under stress).23
Based on these five characteristics, the FDIC concludes that reciprocal deposits should be considered brokered deposits. Sweep deposits from affiliates fall within the purview of the primary purpose exception and therefore should not be considered as brokered deposits, although sweep deposits from non-affiliates should be considered as brokered deposits. Referrals of deposits from affiliates and their agents should be considered as brokered deposits. Listing service deposits have not yet been identified as a potential problem on account of insufficient data.
However, all high-rate deposits pose a problem. Furthermore, the FDIC recommends that Congress not amend or repeal the brokered deposit statute, since “increasing levels of brokered deposits are correlated with a higher probability of failure and higher losses to the FDIC in the event of failure… [and] …increasing levels of brokered deposits are associated with lower core deposit ratios, more rapid growth, and riskier underwriting standards, each of which is correlated with a higher probability of failure” (FDIC 2011, p. 59
Furthermore, the FDIC provided an analysis that included banks and thrifts that failed between 1 January 1988, and 8 April 2011. It found a strong, statistically significant link between the use of brokered deposits and asset growth rates, as well as with higher future rates of noncurrent and nonperforming loans. It also found that bank average growth rates are higher the larger the share of bank assets funded with brokered deposits, but it acknowledged that the relationship is likely to be the result of a complex series of choices made by bank management that drive both a bank’s growth rate and its use of brokered deposits. This means that “[t] he underlying structural choice models are undoubtedly much more complex than the models estimated in this analysis” (FDIC 2011, p. 82
). It should be noted that the asset growth rate is not included in all the regression models, nor is it interacted with brokered deposits, and in some cases the statistical significances of brokered deposits are mixed.
Of significance, the FDIC study lacks consistency in its findings. In some cases, there is a statistically positive relationship, showing that brokered deposits increase the probability of a bank failure; in others there is no statistically positive relationship. The results are therefore mixed, and their use to support a differential regulatory treatment of brokered deposits is questionable.
Perhaps even more important, while the study provides some information about correlations, it provides no information about causation as it relates to the impact of brokered deposits on bank performance, bank failures, and bank failure costs. This is a major weakness and must be addressed because causation may in fact not derive from brokered deposits, but from the opposite direction—in that troubled institutions can turn to these deposits late in the game and as a last-ditch effort to grow out of their problems by investing the funds in excessively risky assets. When they then fail, it might seem reasonable on the surface to point to brokered deposits as the cause, but that would be an error. The real underlying cause would be that these troubled institutions were allowed to take in more funds and invest them in the risky assets, whether the sources of those funds were brokered deposits or some other sources, including high-rate non-brokered deposits. The FDIC study leaves this question unanswered.
Another flaw of the study is that it provides no direct information about the relationship between brokered deposits and the terrible trio of problems: “rapid, risky growth,” “deposit volatility,” and “lower franchise value.” Even more confusing, a few years after the study was released, the FDIC stated that “[b]rokered deposits can be a suitable funding source when properly managed as part of an overall, prudent funding strategy. However, some banks have used brokered deposits to fund unsound or rapid expansion of loan and investment portfolios, which has contributed to weakened financial and liquidity positions over successive economic cycles. The overuse of brokered deposits and the improper management of brokered deposits by problem institutions have contributed to bank failures and losses to the Deposit Insurance Fund” (FDIC 2016b
In other words, the problem again is not brokered deposits per se, but troubled institutions and their attempts to fund unsound or rapid expansion of loans and investments, as well as their overuse of brokered deposits. It would seem to follow naturally that regulations (regulators) should, therefore, be directed (focused) on these factors, rather than on brokered deposits. After all, other sources of funding can contribute to unsound or rapid expansion of excessively risky loans and investments.
The 2011 OIG study examines selected failed state member banks for the period December 2008−February 2011. It cites four common elements in the failures of the 20 institutions that displayed “unusual circumstances,” namely: (1) management making poor decisions as they pursued aggressive growth objectives and made strategic choices; (2) rapid loan portfolio growth that exceeded the bank’s risk management capabilities and/or internal controls; (3) asset concentrations that were tied to commercial real estate (CRE) or construction, land, and land development (CLD) loans, thus increasing the bank’s vulnerability to changes in the marketplace and compounding the risks inherent in individual loans; and (4) management failure to raise sufficient capital to cushion mounting losses (Office of Inspector General 2011, p. 8
The OIG also stated that “[i]n … supplemental research and analysis comparing failed banks to those that withstood the financial crisis, we found that lower commercial real estate and CLD concentration levels, strong capital positions, and minimal dependence on non-core funding were key differentiating characteristics. Our research also revealed a correlation between high CLD concentration levels and the likelihood of failure during the recent financial crisis” (Office of Inspector General 2011, p. 2
This seems to suggest that the OIG did not consider brokered deposits to be an important factor in the bank failures. Indeed, even when mentioning non-core funding as a factor, the OIG wrote that “[f]unding … can be very sensitive to changes in interest rates, … [which includes] brokered deposits, certificates of deposit greater than $
100,000, federal funds purchased, and borrowed money” (Office of Inspector General 2011, p. 52
). In short, brokered deposits, which accounted for only 10% of non-core funding during the period of bank failures studied by the OIG, played a minor role, if any. The obvious question again is: Why treat brokered deposits differently than other non-core funding?
Several other studies have also considered the role of brokered deposits in bank failures, failure costs, and banking instability. Rather than discuss each in detail, we summarize the more important findings of empirical studies in Table 2
. Of the studies, nineteen focus on the relationship between brokered deposits and the likelihood of bank failure, while the other four studies focus on whether brokered deposits increase bank failure costs.
Of the nineteen studies that examine the relationship between brokered deposits and the likelihood of a bank failure, eight find a significantly positive relationship between brokered deposits and bank failures, five find no such relationship, and six find mixed results. In the case of the remaining four empirical studies that examine the relationship between brokered deposits and bank failure costs, two find no relationship between these two variables, while one study actually finds that an increase in brokered deposits is associated with a decrease in bank failure costs. Two find mixed results in that some relationships were significantly positive, while others were not significant.
Interestingly, a study finds in one case that core deposits were statistically and positively related to the cost of resolving failed banks, while in another case that such deposits were not significantly related to the failure costs of banks (FDIC 2011, p. 104
). This is a study by the FDIC.
The bottom line here is that most of the empirical studies, those focusing on either bank failures or bank failure costs, do not provide justification for the current regulatory treatment of brokered deposits. In this regard, the Department of the Treasury in its report titled “Modernizing the Financial System: Recommendations for Safer, More Competitive Banks,” stated that “studies of depository institution failures have not found a consistent, statistically significant relationship between brokered deposits and either the probability or cost of failure” (US Department of the Treasury 1991, pp. IV−4
Most important, the studies do not consider different types of brokered deposits or control for all non-core sources of funding used by banks in the empirical work. They also do not generally take into account the underwriting standards used when loans are made or the extent of fraud involved in bank failures, among other limitations.
Moreover, none of the studies provide direct evidence that brokered deposits are a causal factor with respect to bank failures, failure costs, or banking instability. For example, the FDIC study states that “[b]rokered deposits are correlated with behaviors that increase the risk of failure” (FDIC 2011, p. 47
). However, the correlations that do emerge are totally consistent with the view of Rossi
(2010, p. 22
), who states that “a picture emerges supporting the view that brokered deposits do not drive asset growth, risk-taking or insolvency. … Instead, it was shown that greater risk-taking could promote increased usage of brokered deposits when faced with a constraint on retail deposits.”
The bottom line, more generally, may be best found in statements made by the FDIC in an older but timely study titled “Deposit Insurance for the Nineties: Meeting the Challenge.”
According to the study, proposals regarding limits on brokered deposits “ignore FDIC examination experience, which suggests that supervision can, in general, effectively discriminate between sound and unsound use of brokered funds. … Moreover, recently proposed changes in reporting requirements should enhance examiners’ ability to detect brokered-deposit abuses early. Supervisors will get clear signals that closer scrutiny is warranted. These signals take the form of increases in offering rates and the growth of brokered-funds purchased. Once in the bank, supervisors can evaluate the quality of lending in the usual manner. This indicates that brokerage of funds is not a special problem, but part of the more general incentive problem in deposit insurance” (FDIC 1989, pp. 95−96