The Disproportionate Costs of Uncertainty:Evidence from Dodd Frank and Small Banks

Working Paper

Raymond Kim †

A. Gary Anderson School of ManagementUniversity of California at Riverside‡

January 16, 2018

Abstract

I examine the disproportionate impact of regulatory uncertainty on smaller banks.Baker, Bloom, and Davis (2016) indicates that firm level policy exposures matters foreconomic outcomes. I find that cost of regulatory uncertainty has a greater impact onsmaller banks with fewer resources and limited access to capital markets. Uncertaintystems from delayed Dodd Frank rulings on the cost of bank hedging activities. Usingfirm fixed effects with interactions, I find that during regulatory uncertainty, smallerbanks with an increase of 1% in compliance costs leave 18% more balance sheet riskexposures and provide 8.1% fewer corporate clients with hedging services. These find-ings support the need for tailored regulations that reduce the cost of uncertainty forsmaller banks.

JEL classification:

Keywords : Regulatory uncertainty, banking regulation, hedging, risk management, interestrate risk, derivatives, mortgage backed securities, Dodd Frank

†Advisor: Dr. Jean Helwege, Area Coordinator for Finance, University of California at Riverside. I wouldlike to thank Dr. Jean Helwege for her guidance and mentorship. I would also like to thank Peter Chungand Scott Hein for their helpful comments.

‡University of California at Riverside, 900 University Ave. Riverside, CA 92521email:raymondn@ucr.edu

1. Introduction

During the global financial crisis, concerns about policy uncertainty intensified as central

banks and corporate firms noted that uncertainties about policies were contributing to steep

economic declines and tepid recoveries. Measures of economic policy uncertainty (EPU)

Baker et al. (2016) and monetary policy uncertainty (MPU) Husted, Rogers, and Sun (2017)

have measured policy uncertainty that foreshadow declines in business investment and ac-

tivity. In banking, Gissler, Oldfather, and Ruffino (2016) looks at uncertainty of mortgage

regulation and it’s negative effects on bank lending at the firm level while Bordo, Duca,

and Koch (2016) also looks at EPU and it’s aggregate effect lending. In a highly regulated

industry like banking, academic literature also highlights costs of funding uncertainty and

crisis regulations. 1.

This paper measures the costs of policy uncertainty by looking at the bank’s hedging ac-

tivities. Given that banks are repositories for interest rate risk Gorton and Rosen (1995),

the function of bank hedging activity is a central risk management function. As far as I

know, this is the first paper to look at the effects of regulatory on hedging activity and its

disproportionate impact on smaller banks. Using a unique dataset from the Conference of

State Board Supervisors and quarterly Call Report data from the Federal Reserve, I ex-

amine a period of regulatory uncertainty regarding the undecided risk weightings of bank

balance sheet assets, which impacted smaller banks that were well capitalized during the

global financial crisis. 2. This period of uncertainty starts from the passage of The Dodd-

Frank Wall Street Reform and Consumer Protection Act in 2010 until February 2015 when

risk weightings for regulatory capital were finally released by the Federal Insurance Deposit

Corporation (FDIC) and the Federal Reserve FDIC (2015) 3. Dodd Frank’s regulations were

mostly centered around regulating complicated large too-big-to-fail (TBTF) banks, which

drew out the process of final rulings on various regulatory measures such as risk weightings

for interest rate derivatives. This unusually long period of regulatory uncertainty in a highly

regulated industry provides a unique environment to study the costs of uncertainty on a

bank’s hedging activities.

1 Ritz and Walther (2015) points to bank level differences in lending due to costs of funding uncertaintywhile Banerjee and Mio (2017) and Cyree (2016) research the actual effects of crises regulation on banks.

2”[[The impact] of changing risk weight calculations [on assets] is surprising to many that have been andremain well capitalized through the most recent economic difficulties.” in a letter by the American Banker’sAssociation to the Office of the Comptroller of the Currency and the Board of Governors of the FederalReserve System on October 25, 2012

3 [”Proposed changes starting March 31, 2015] will include an increased number of risk-weight cate-gories to which on-balance sheet assets, derivatives, off-balance sheet items, and other items subject to riskweighting would be allocated.” in a news release by the FDIC on February 20, 2015 FDIC (2015)

1

Using a unique dataset of surveyed compliance costs collected by the Conference of State

Board Supervisors, Dahl, Meyer, and Neely (2016a) as found that smaller banks have been

disproportionately burdened by regulation intended for ”Too Big to Fail” (TBTF) banks.

Dahl et al. (2016a) finds that the smallest banks faced Dodd Frank compliance costs of 8.7%

of noninterest expenses vs. 2.9% for medium sized banks. Given that Dodd Frank was

enacted to curb excessive risk taking by TBTF banks, a natural question emerges: What

are the costs to non-TBTF banks regarding one size fits all regulatory measures? Does

uncertainty surrounding pending regulations affect non-TBTF banks? As smaller banks

prefer less risk Hein, Koch, and MacDonald (2005) and more conservative capital structures

Purnanandam (2007), has their ability to hedge interest rate risk been burdened? Examining

this question requires an understanding of the period of uncertainty that non-TBTF banks

experienced.

The Dodd-Frank Wall Street Reform and Consumer Protection Act was passed in 2010 as

a response to the financial crisis of 2008. Over 22,000 pages long and intended to decrease

risks to the US financial system, many smaller banks have been snared in an increasingly

complex set of rules designed for larger banks.

”it’s important to look for ways to relieve regulatory burden on community

banks and smaller institutions to tailor regulation so that it’s appropriate for the

systemic risk profile of the particular institutions”

- Federal Reserve Chairman Janet Yellen

Press Conference

December 14, 2016

Janet Yellen, the Federal Reserve Chairman made these remarks regarding the regulatory

burden that smaller banks have faced throughout the implementation of Dodd Frank. This

naturally leads to the question, did uncertainty regarding Dodd Frank’s regulations restrict

smaller banks in unintended ways, such as restricting their ability to reduce balance sheet

risk so that they can make more loans and grow their lending assets? In particular, this paper

focuses on the beginning of uncertainty for banks in Dodd Frank’s Subpart D, Section 34

which outlines future changes for risk weighting of interest rate derivatives. By eliminating

the 50% risk weight cap on interest rate derivatives, Dodd Frank waved a broad stroke

across all types of derivatives throwing smaller banks into uncertainty regarding the new

weightings of their hedging portfolios. Elimination of the risk weight cap on interest rate

derivatives opened up uncertainty as to what these new risk weight caps would actually be.

2

This uncertainty would last from the passage of Dodd Frank in 2010 until February 20, 2015

when the Federal Deposit Insurance Corporation (FDIC) released a last round of changes

before the first Dodd Frank regulatory risk weightings were to be used in March 31, 2015

FDIC (2015). This long period of regulatory uncertainty would have a particularly negative

cost for banks with limited resources for regulatory compliance. The literature shows that

high costs of compliance for smaller banks translate into having fewer staff to efficiently

process regulatory costs Dahl et al. (2016a), and regulatory uncertainty would tax smaller

banks into simply reducing the use of interest rate derivatives. This type of behavior would

fall in line with Boyson, Helwege, and Jindra (2014) and the selling of assets by banks in

order to meeting regulatory requirements.

Such reduction in hedging would also be broadly consistent with literature on banking regu-

lation and how uncertainty has negative effects on bank lending (Gissler et al., 2016; Bordo

et al., 2016), contributing a banking perspective to the growing literature on effects of eco-

nomic policy uncertainty (Baker et al., 2016; Brogaard and Detzel, 2015; Bonaime, Gulen,

and Ion, 2016; Koijen, Philipson, and Uhlig, 2016; Giavazzi and McMahon, 2012) on the

economy. In order to extend the literature to include the costs of regulatory uncertainty on

bank hedging, the next section will outline the use of interest rate derivatives in banking.

2. Interest Rate Derivatives in Bank Hedging

Literature in this growing field focused on the positive relationship between derivative use

and loan growth (Brewer III, Minton, and Moser (2000))(Landier, Sraer, and Thesmar (2013)

since loans expose banks to macroeconomic cash flow risks and interest rate derivatives help

to mitigate these cash flow risks. So as banks increase their loans held to maturity, they are

exposed to more interest rate risk, which increases the benefit of mitigating interest rate risk

by using interest rate derivatives. Lending policies become less sensitive to macroeconomic

shocks with the use of interest rate derivatives Purnanandam (2007) and subsequently lend

more than non users. However, use of derivatives for portfolio hedging has differed between

TBTF banks and smaller banks. While non-TBTF banks use interest rate derivatives to

hedge, allowing a bank to address other balance sheet risks more relevant to it’s simpler

banking model, TBTF banks additionally use CDS to hedge credit risk for bonds held as

investment and as credit insurance to sell bonds to buyersShan, Tang, and Yan (2014).

As banks face cash flow risk from mismatched maturities, repricing risk, bankruptcy risk

Smith and Stulz (1985), financing risks Froot, Scharfstein, and Stein (1993), interest rate

3

risk, monetary policy risks, and market risks, derivative use has become common practice

to hedge exposures to balance sheet risk. Larger banks use interest rate derivatives not only

for hedging4, but also dealer intermediation5 , and speculation6

The use of interest rate derivative in banks can be likened to a the next layer of risk manage-

ment after the fixed-rate deposit insurance system, such as the Federal Deposit Insurance

Corporation (FDIC). By insuring a bank’s liabilities, banks can borrow at below market

rates from depositors, and invest in riskier loans at higher interest rates. As Merton (1977)

noted, insured deposits act as a “put option” on bank assets with a strike price equal to

the maturity value of its debt. The diversification of financial services over the past several

decades were argued to contribute to the stability of the financial system as Demsetz and

Strahan (1995) looked at data from 1980-1993 and found that large bank holding compa-

nies (BHC) were indeed more diversified in their lending portfolio than smaller BHC’s. It

was during this time that banks started to increase their use of interest rate derivatives

for risk management purposes. Brewer III et al. (2000) noted how since 1985, commercial

banks started to use interest rate derivative products. By using derivatives to address the

credit risk and interest rate risk of their loans, banks that used these derivatives experienced

greater growth in their lending portfolios than banks that didn’t. The use of derivatives

allowed banks to focus on their comparative advantage in credit monitoring Diamond (1984)

So how does this increasing use of derivatives for hedging help grow a bank’s lending port-

folio? Much of the academic literature revolves around a bank using derivatives to manage

interest rate exposures, monetary policy risks, and external financing risks. Diamond (1984)

model suggests that derivatives lead to a reduction in loan costs that provide incentives

for banks to lend more. Brewer III et al. (2000) show that empirically there is a positive

relationship between interest rate derivatives and commercial and industrial loans. Banks

also use derivatives to hedge themselves against costly external financing and the cost of

financial distress. Banks that use derivatives are more immune to policy shocks in lending

volumes, while non-derivative users experience significantly declining volumes Purnanandam

(2007). The use of derivatives in risk management allows banks to weather storms better,

and strong risk management at banks has been successful in mitigating “tail risk”, leading

to lower non-performing loans and better operating performance during the financial crisis

4 Brewer, Minton and Moser (2000) found that banks that use interest rate derivatives experience greatergrowth in their loan portfolios than banks that did not use them.

5Begenau, Piazzesi and Schneider, 2015 find that banks use pay-fixed positions in swaps to insure againstsurprise interest rate increases. Hentschel and Kothari (2001) and Chernenko and Faulkender (2011) alsoshow this empirically, while Jermann and Yue (2012) use a theoretical framework to study why non-financialfirms need pay-fixed swaps

6Gorton and Rosen 1995 find that agents claim that speculative risk taking was unintentional.

4

Ellul and Yerramilli (2013). Greater derivative use and a greater fraction of income from

non-banking activities have become an accepted practice of effective bank risk management.

With adequate bank capitalization and derivative skills, loans and interest rate derivatives

can be seen as complimentary assets that move together on a bank’s balance sheet. However,

this relationship can become one of substitution when a regulatory uncertainty lowers the

marginal risk weighted return of a derivative, so that the re-calibration of a t+1 equilibrium

will have a substitution effect between the two assets. Boyson et al. (2014) outlines the

process of how a bank can sell assets in order to meet regulatory requirements such as

capital adequacy ratios. As the cost of uncertainty literature has shown, anticipation of such

change can also have economic costs for banks as well.

3. Data

We obtained the data for this study from three main sources: 1) Call Reports from the Federal

Reserve for quarterly financial data of all US insured commercial banks, 2) Compliance cost

data from the Conference of State Board Supervisors and 3) Chicago Board of Exchange for

swap interest rate data, covering the period from 2010 Q4 to 2017 Q3. 2010 Q4 was chosen

because it is the first full reporting quarter after Dodd Frank was signed into federal law by

President Barack Obama on July 21, 2010. Implementation of new risk weight and capital

adequacy ratios was implemented January 1, 2015 for non-TBTF banks. Final details of risk

weightings for off balance sheet items were not released until February 20, 2015, mandating

2015 Q1 as the first reporting quarter where regulatory certainty was established.

TBTF banks were dropped from the data in order to study the effects of regulatory uncer-

tainty on firms that were not subject to additional regulatory requirements such as stress

tests. For the logistical regression in Table 3, all non TBTF bank observations were included

to test whether the amount of loan assets were an effective proxy for interest rate risk and

whether there is predictive power for a bank hedges it’s interest rate risk. For the sum-

mary statistics and all other regressions, banks with non-missing values for mortgages and

interest rate derivatives were used. Consolidated and domestic bank data were merged and

duplicates were eliminated. Due to various changes to reporting requirements such as the

implementation of Dodd Frank, some call report variables are not consistent over time. Con-

sistent time series were formed by looking at the Call Report forms and matching variables

as they change from quarter to quarter. All bank quarter observations reflect banks with

5

active mortgage lending and derivatives divisions. The total number of bank quarter obser-

vations are 16,930 from 2010 Q4 to 2017 Q3. Summary statistics show that banks increased

in asset size and all ratios after regulatory certainty was established. Concerns that increase

in assets, lending and hedging are endogenous to an economic recovery are addressed by the

use of two way and three way interaction variables between compliance burden, regulatory

certainty, and covariates in Table 5.

In Table 4, pooled OLS where compliance cost ratios were regressed on balance sheet co-

variates demonstrating that compliance cost is highly negatively related to bank size. Data

on compliance costs were derived from Dahl et al. (2016a) and Call Reports. Covariates

were log scaled and bank observations were pooled across 2010 Q4 - 2017 Q3. Tier 1 Ratio

was obtained from Schedule RC-R of the Call Reports. Off balance sheet data on bank’s

use of interest rate derivatives were obtained from the Schedule RC-L of the quarterly Call

Reports. Interest rate derivatives are defined as total gross notional amounts of interest

rate derivative contracts held by either trading or non trading purposes. In examining the

off balance sheet data, about 80-90% of derivative use consists of interest rate derivatives.

Trading and non trading interest rate derivatives were combined for the variable ”Interest

Rate Derivatives”.

Bank loan data was taken from the Schedule RC-P and Schedule RC-C of the quarterly

Call Reports. While Schedule RC has overall loan data, it does not break out residential

mortgages, which is listed in more detail in the Schedule RC-C and RC-P. The Schedule

RC-P reports a bank’s residential mortgage activities while the Schedule RC-C reports a

bank’s general loans and leases activity. Residential mortgage loans held to maturity is not

directly given in the Call Reports, so it must be calculated. Total residential mortgage loans

are calculated by combining mortgages secured by first and junior liens in the Schedule RC-

C. This figure includes both loans held to maturity and loans held for sale or trading. In

order to subtract out loans held for sale or trading, the Schedule RC-P was used to calculate

this figure. By using the amount of residential mortgages held to maturity, we can isolate

bank risk management in regards to macro and interest rate risk in its long term balance

sheet. Residential mortgages did not have it’s risk weighting and capital treatment of 50%

affected in Dodd Frank regulation. The first differences of this variable measures the change

in willingness of banks to hold loans on it’s long term balance sheet. This variable can be

shown in simple form:

Loans Held to Maturityresidential = Loanstotal − Loansheldforsale

6

Table 3 also shows that residential mortgages are significant estimators of interest rate deriva-

tive usage by banks using both logit regression. This supports our use of residential mort-

gages held for investment as a proxy for interest rate risk. Loans Held to Maturity represent

a loan that is of higher credit quality and lower risk than a loan held for sale. n order to

isolate the risk management function of derivatives to balance sheet risk, control factors

were used. Banks mainly use interest rate derivatives for three reasons: hedging Brewer III

et al. (2000), dealer intermediation Begenau, Piazzesi, and Schneider (2015) Hentschel and

Kothari (2001) Chernenko and Faulkender (2012)Jermann and Yue (2013), and speculation

Gorton and Rosen (1995). In order to control for dealer activity and speculation,Chicago

Board of Options Exchange (CBOT) data consisting of the 10-Year Interest Rate Swaps

were used.

Interest rate swaps are a core derivative product that banks use for dealer activities as well

as hedging their balance sheet interest rate risk. Gorton and Rosen (1995) show that interest

rate risk is non diversifiable and banks are repositories of interest rate risk. This interest

rate risk is hedged using swaps ( Gorton and Rosen (1995)) and in larger banks, swaps used

for hedging are difficult to separate from speculation. Recent interest rate swap models are

outlined in Begenau et al. (2015) and Vuillemey (2015). For the purposes of this paper,

interest rate swaps are in demand by bank clients who have floating interest rate liabilities.

When interest rates fall, companies seek to lock in lower fixed interest rate payments and

takes a fixed payer swap position with a bank. This is why we expect a banks interest rate

derivative assets to increase when interest rates fall and volatility rises. Companies want to

lock in lower interest rates and also hedge the volatility of their interest rate exposures.

4. Empirical Model and Results

To test the impact of regulatory uncertainty on bank core operations, I designed two empirical

tests using panel data from 2010 Q4 to 2017 Q3.

4.1. Logistic Regression of Loans as a Predictor of Bank Hedging

The first tests whether loans held on a bank balance sheet is a significant predictor of banks

hedging with interest rate derivatives. Significant results will suggest that mortgage loans

are a proxy for bank interest rate risk. The greater interest rate risk a bank has, the more

7

likely they will use interest rate derivatives to hedge that balance sheet risk. Using a logistic

regression design, I use a cross section of banks in the latest available quarter (2017 Q3)

that files quarterly financial reports with the Federal Reserve. Table 3 tests the following

empirical model:

ln[ pi1− pi

]= α + β1(Loan Assets)i + β2

Tier 1 CapitaliRiskWeightedAssetsi

+ εi

Where pi is the probability that bank i is a user of interest rate derivatives are defined as

Interest Rate Derivativesi = Non Trading IRDi + Trading IRDi

For independent variables I use quarterly holding of loan assets on the balance sheet to see

whether amount of loans is a contemporaneous predictor of hedging use by a bank. For the

purposes of this logit regression, loan assets are scaled in units of a $1 billion.

Loan Assets ∈[Residential Mortgages, Commerical & Industrial Loans

]Tier 1 Ratio is used as a control for financial health of a bank. Tier 1 ratio is a measure

of a bank’s financial strength and ability to absorb unexpected losses. Logistic regression

results show that both residential mortgages and consumer and industrial loan assets are

economically and statistically significant predictors of whether a bank is a hedger of loans or

not. In regression (2) the coefficient for Residential Mortgages is 0.85 and has a t-statistic

of 6.62, representing an increase in log odds of .85 for every $1B of increase in mortgage

assets held on the balance sheet. .85 log odds is a 134% increase in odds of a bank being in

a user of derivative hedging. The Tier 1 Ratio is also significant shows a coefficient of -.10,

representing a decrease in log odds of -.10 for every 1% increase in a bank’s Tier 1 Ratio.

-.10 log odds is a 9.6% decrease in odds of a bank being a derivative user.

In regression (4) results are similar for C&I loans where the coefficient of .69 is significant

with a t-statistic of 5.11, representing an increase of .69 log odds for every $1B increase in

C&I loans. This translates to a 99% increase in odds of a bank being a user of derivative

hedging for every $1B increase in commerical and industrial loans. The Tier 1 Ratio also

shows negative significance with a coefficient of .09, representing a decrease in log odds of

-.09 for every 1% increase in a bank’s Tier 1 Ratio. -.09 log odds is a 8.7% decrease in odds of

a bank being a derivative user. The results of Table 3 demonstrates the significance of loans,

especially residential mortgages, as a proxy for interest rate risk. For the next empirical

model, I will use residential mortgages as a proxy for interest rate risk in understanding the

8

costs of hedging in the context of regulatory uncertainty.

4.2. Firm Fixed Effects Panel Data Regression with Three Way Interactions

This section will deal with the main empirical model which tests for the costs of regulatory

uncertainty for banks that have varying degrees of regulatory burden. There are two main

hypothesis being tested. The first is that the costs of uncertainty is higher for banks with

fewer resources and higher compliance costs. The second hypothesis is that a bank under

constraint of uncertainty will prioritize resources to core operations. The results of the main

empirical model is shown in Table 5. The specifications of the empirical model with White

standard errors is as follows:

Interest Rate DerivativesitAssetsit

= αit +Compliance Costit

Assetsit∗ Regulatory Certainty∗[Loans Held to Maturityit

Assetsit+

Loans Held for SaleitAssetsit

+Loans SolditAssetsit

+ Swap Rate]

+ εit

The ”Hedging” dependent variable represents interest rates derivatives not used for trading

scaled by total assets. ”Trading” likewise represents interest rates derivatives used for trading

scaled by assets. ”Total” represents the combined amount of Hedging and Trading interest

rates derivatives scaled by assets. The trading dependent variable should not show significant

relationships with the covariates and interactions for hedging risk because by definition in

the Call Reports, trading interest rate derivatives are for speculation and trading. Loans

Held to Maturity represent the safest loans on a bank’s balance sheet while Loans Held for

Sale and Loans Sold have greater interest rate risk and require hedging. Swap rates are 10

year interest rate swaps. When swap rates fall, corporate clients demand interest rate swaps

in order to replace their floating interest rate payments with low fixed interest rates, hedging

their interest rate liabilities. Compliance burden is calculated using a unique dataset from

Dahl et al. (2016a) and is high for firms with high compliance costs. From Table 4 that

compliance burden has a highly negative relationship with bank size. Regulatory certainty

is a dummy variable with a value of 1 and represents clarity on Dodd Frank’s risk weightings

for derivatives, which were finally implemented in 2015 Q1.

Hedging various classes of mortgages on the balance sheet represent stabilization of core lend-

ing operations of a commercial bank. Selling fixed interest rate swaps to corporate clients

9

represent a business line that hedges corporate client’s interest rate risk instead of a bank’s

own interest rate risk. If the cost of uncertainty is higher for banks with higher compliance

burden then there should be a positive interaction between Regulatory certainty and Com-

pliance Burden. Banks with high compliance burden will see a disproportionately larger

increase in use of hedging after risk weightings of interest rate derivatives are finalized for

reporting in 2015 Q1. Table 5 confirms this hypothesis as the two way interaction coefficient

between Regulatory Certainty and Compliance Burden is positive and economically and sta-

tistically significant with a value of 17.8 and a t-statistic of 2.53. Table 5 also shows that

even in times of uncertainty, banks regardless of compliance cost will hedge their portfolio

for residential mortgages held for sale, as the coefficient for Loans Held for Sale (LHS) is .758

and has a t-statistic of 3.04. The lack of significance for the three way interactions between

Certainty * Compliance * LHS/LS shows that banks prioritize resources to hedging their

portfolio regardless of whether there is regulatory uncertainty or not. The results shown in

the ”Hedging” panel data regression generally holds for the ”Total” panel data regression,

showing that banks derivative use revolve around hedging activity and not trading.

Conclusion

The costs of uncertainty has been highlighted as aggregate declines in investment, output,

and employment Baker et al. (2016). This paper contributes to the findings of Gissler

et al. (2016) and Bordo et al. (2016) regarding the costs of regulatory uncertainty on bank

hedging practices and the disproportionate impact on smaller firms with higher compliance

costs, fewer resources, and limited ability to raise capital in the equity markets. Evidence

suggests that the tradeoff for firms dealing with uncertainty is to divert resources to core

operations and to pursue fewer growth opportunities until the costs of uncertainty have

subsided. Banks as repositories of interest rate risk Gorton and Rosen (1995) and have

acute needs for risk management practices to hedge loan portfolio risks, which comprise core

operations. Banks with higher compliance costs disproportionately decreased hedging when

specifics about Dodd Frank’s regulations were not known, and disproportionately increased

its use after risk weightings were known and implemented. Greater detail in banking data,

especially in core operations and off balance sheet items, as well as an unusually long period

of uncertainty, allows for further research into the cross sectional disparities of the cost of

uncertainty.

The results of this paper suggest that Dodd Frank’s long period of uncertainty in assigning

10

risk weightings to hedging products have disproportionately harmed banks that were largely

not responsible for the global financial crisis and were well capitalized throughout the reces-

sion. From a policy standpoint, these findings support the benefits of tailored policies that

reduce the costs of uncertainty for ”good” banks, while also addressing systematic issues

arising from ”TBTF” banks.

11

Appendix

Table 1: Summary Statistics for All Banks During UncertaintyData is from Call Reports (2010 Q4 - 2014 Q4). Ratios are scaled by total assets and are rep-resented as percentages. Regulatory uncertainty centers around the risk weighting of assets forbank assets and liabilities which were unknown before Dodd Frank was fully implemented in 2015 Q1.

Statistic N Mean Median Pctl(25) Pctl(75) St. Dev.

Total Assets ($M) 10,045 2,766.76 1,048.85 496.75 2,413.26 6,887.99Interest Rates Derivative Ratio 10,045 9.03 2.68 0.88 8.79 18.60Hedging Rates Ratio 10,045 7.62 1.90 0.51 6.64 17.27Trading Rates Ratio 10,045 1.42 0.00 0.00 0.00 6.94Fixed Swaps Ratio 10,045 0.73 0.00 0.00 0.00 2.69Residential Inventory Ratio 10,045 15.93 13.05 8.39 19.84 11.23Residential Held Ratio 10,045 2.55 0.45 0.14 1.54 6.34Residential Sold Ratio 10,045 10.47 2.60 1.03 7.38 25.72Compliance Cost Ratio 10,045 0.10 0.07 0.04 0.11 0.14Tier 1 Capital ($M) 10,045 257,660.90 97,258 45,624 226,598 615,228.50Risk Weighted Assets ($M) 10,045 1,890,294.00 715,925 351,735 1,677,713 4,914,440.00Tier 1 Ratio 10,045 0.14 0.13 0.12 0.15 0.05

Table 2: Summary Statistics for All Banks During certaintyData is from Call Reports (2015 Q1 - 2017 Q3). Ratios are scaled by total assets and arerepresented as percentages. Regulatory certainty is after the risk weighting of assets forbank assets and liabilities were known when Dodd Frank was fully implemented in 2015 Q1.

Statistic N Mean Median Pctl(25) Pctl(75) St. Dev.

Total Assets ($M) 6,885 4,873.42 1,471.60 725.92 3,872.48 13,037.24Interest Rates Derivative Ratio 6,885 10.40 3.59 1.05 11.08 20.02Hedging Rates Ratio 6,885 8.33 2.36 0.67 7.88 18.11Trading Rates Ratio 6,885 2.08 0.00 0.00 0.00 8.94Fixed Swaps Ratio 6,885 1.09 0.00 0.00 0.37 2.88Residential Inventory Ratio 6,885 17.16 13.93 9.10 21.16 12.39Residential Held Ratio 6,885 2.30 0.31 0.10 1.20 6.48Residential Sold Ratio 6,885 8.63 1.69 0.67 5.28 23.56Compliance Cost Ratio 6,885 0.09 0.06 0.03 0.09 0.24Tier 1 Capital ($M) 6,885 457,279.50 143,141 70,929 375,931 1,178,064.00Risk Weighted Assets ($M) 6,885 3,688,787.00 1,100,667 534,141 2,968,841 10,232,543.00Tier 1 Ratio 6,885 0.14 0.13 0.11 0.14 0.04

12

Table 3: Logistic Regression: Using Loans as a Predictor of Bank Hedging UsageUsing a cross section of Call Report data for 2017 Q3, logistic regression results find that residentialmortgages held for investment is a significant indicator of whether or not a bank hedges their portfoliousing interest rate derivatives. Interest rate derivatives are reported on Schedule RC-L of the Call Re-ports as Derivatives and Off Balance Sheet Items. Residential mortgages and Commercial and IndustrialLoans are reported on Schedule RC of the Call Reports as Balance Sheet Items. Residential mortgagesare defined as loans and leases that a reporting bank has the intent and ability to hold both for the fore-seeable future or until maturity or payoff and for the immediate future before sale, representing interestrate risk on a bank’s balance sheet. This table also shows that Commercial and Industrial loans are alsoa significant indicator in a bank being a user of interest rate derivatives. Tier 1 Ratio is reported on theSchedule RC-R of the Call Reports a Regulatory Capital item. These results control for a bank’s Tier 1Ratio defined as Tier 1 Capital

Risk Weighted Assetswhich is a proxy for a bank’s financial health. Residential mortgages

and C&I loans represent an interest rate risk on a bank’s balance sheet that require hedging. Resultsare similar for bank observations in other quarters.

Dependent variable:

Bank as User of Interest Rate Hedging

(1) (2) (3) (4)

Residential Mortgages 0.96∗∗∗ 0.85∗∗∗

(7.23) (6.62)

Tier 1 Ratio −0.10∗∗∗ −0.09∗∗∗

(−11.73) (−11.00)

CI Loans 1.00∗∗∗ 0.69∗∗∗

(6.25) (5.11)

Constant −0.16∗∗∗ 1.50∗∗∗ −0.10∗∗ 1.44∗∗∗

(−3.52) (10.68) (−2.30) (10.41)

Observations 2,681 2,681 2,681 2,681Log Likelihood -1,786.31 -1,659.04 -1,795.43 -1,681.02Akaike Inf. Crit. 3,576.63 3,324.08 3,594.86 3,368.03

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

13

Figure 1. Use of Hedging Across Various Banks by Size and TypeScatterplot chart of the log of total assets on the x-axis with the log of

(Interest Rate Derivatives

Residential Mortgages

)on the y-axis

for 16,930 bank quarter observations from 2010 Q4-2017 Q3. Scaling was used for ease of exposition.

14

Table 4: Pooled OLS on Compliance CostCompliance data is obtained from a unique dataset in Dahl et al. (2016a). The compliance cost ratio isregressed against log scaled assets, log scaled total residential mortgages, log scaled interest rate deriva-tives, and Tier 1 Ratio. Results show that assets size is highly negatively significant with compliancecost ratio. The results of this pooled OLS show that firms with higher compliance costs are smallerbanks, even when controlling for balance sheet loans, off balance sheet derivatives, and financial health.

Dependent variable:

Compliance Costs/Assets

(1) (2) (3) (4)

log(Total Assets) −4.79∗∗∗ −7.89∗∗∗ −7.52∗∗∗ −8.84∗∗∗

(−46.35) (−39.14) (−37.00) (−43.29)

log(Residential Mortgages) 3.49∗∗∗ 3.17∗∗∗ 2.27∗∗∗

(17.86) (16.14) (11.66)

Tier 1 Ratio 35.50∗∗∗ 40.48∗∗∗

(11.55) (13.45)

log(Interest Rate Derivatives) 2.13∗∗∗

(28.07)

Observations 16,863 16,863 16,863 16,863R2 0.11 0.13 0.14 0.17Adjusted R2 0.11 0.13 0.14 0.17Residual Std. Error 17.24 17.08 17.01 16.63F Statistic 2,148.33∗∗∗ 1,253.92∗∗∗ 886.99∗∗∗ 893.30∗∗∗

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

15

Table 5: Firm Fixed Effects Panel Data Regression with Three Way InteractionsUsing Panel Data from Call Reports (2010 Q4-2017 Q3) and compliance costs from the Conference of State Bank Supervisors I use the followingempirical model:

Interest Rate Derivativesit

Assetsit= αit +

Compliance CostitAssetsit

∗ Regulatory Certainty∗[Loans Held to MaturityitAssetsit

+Loans Held for Saleit

Assetsit

+Loans Soldit

Assetsit+ Swap Rate

]+ εit

Loans Held to Maturity represent the safest loans on a bank’s balance sheet while Loans Held for Sale and Loans Sold have greater interestrate risk and require hedging. Swap rates are 10 year interest rate swaps. When swap rates are low, demand for swaps goes up as corporateclients hedge their floating interest rate liabilities with fixed rate swaps. Compliance burden is a continuous variable and is high for firms withhigh compliance costs. Regulatory certainty is a dummy variable with a value of 1 when Dodd Frank’s risk weightings for derivatives wereimplemented in 2015 Q1. The three way interaction of Certainty * Compliance * Swap Rate shows that banks with high compliance costsbenefit from regulatory certainty and generate more hedging business with corporate clients when interest rates fall. The two way interactionbetween Certainty * Compliance shows us that banks with high compliance costs hedge more during times of regulatory certainty. Derivativesused for trading by definition are not used for hedging purposes. Hedging + Trading = Total Interest Rate Derivatives. T-statistics are inparenthesis and are estimated using White standard errors.

Interest Rate Derivatives/Total Assets:Total Hedging Trading

Compliance Burden 52.145*** 54.535*** -2.39(4.24) (4.45) (-0.84)

Regulatory Certainty * Compliance 22.119*** 17.871** 4.248(3.51) (2.53) (0.92)

Loans Held to Maturity (LHM) -0.066 -0.077 0.011(-0.76) (-0.75) (0.22)

Loans Held for Sale (LHS) 0.907*** 0.758** 0.15(3.91) (3.04) (1.68)

Loans Sold (LS) 0.192** 0.19** 0.002(3.08) (2.92) (0.12)

Swap Rate -0.007* -0.004 -0.003*(-2.3) (-1.43) (-2.47)

Compliance * LHM -49.73 -51.219 1.489(-1.94) (-1.83) (0.12)

Certainty * Compliance * LHM -11.803 2.504 -14.307(-0.51) (0.08) (-0.85)

Compliance * LHS -70.922 -74.353 3.43(-1.38) (-1.5) (0.27)

Certainty * Compliance * LHS 61.548 69.768 -8.22(1.19) (1.39) (-1.22)

Compliance * LS 4.212 4.838 -0.625(0.66) (0.7) (-0.23)

Certainty * Compliance * LS -5.777 -6.434 0.657(-1.29) (-1.36) (0.61)

Compliance * Swap Rate -17.575*** -18.336*** 0.761(-4.78) (-4.9) (0.73)

Certainty * Compliance * Swap Rate -9.068** -8.135** -0.933(-2.74) (-2.6) (-0.98)

R2 0.47 0.49 0.01F-Statistic 21.55 15.75 1.66Observations 16,930 16,930 16,930Groups (Bank Fixed Effects) 1,040 1,040 1,040Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

16

Figure 2. Time Series Chart of Ratio of all Banks (Non-TBTF) from 2010Q4-2017Q3Time Series chart of 1

n

∑ni=1

Residential MortgagesitTotal Assetsit

, 1n

∑ni=1

Interest Rate DerivativesitTotal Assetsit

, 1n

∑ni=1

Interest Rate DerivativesitResidential Mortgagesit

.

The large spike in 2014 Q3 in 1n

∑ni=1

Interest Rate DerivativesitTotal Assetsit

occured after first draft of the final riskweightings on interest rate derivatives was released in June 27, 2014 by the FDIC, Federal Reserve,and Office of the Comptroller of the Currency. Comments were accepted until August 2014, and afterthousands of responses by FDIC financial institutions, a final draft was subsequently delayed until 2015Q1.

17

References

Adrian, T., Liang, N., ???? Monetary policy, financial conditions, and financial stability. CEPR Discus-sion Paper No. DP11394 .

Adrian, T., Shin, H. S., 2008. Financial intermediaries, financial stability, and monetary policy. FRB ofNew York staff report .

Adrian, T., Shin, H. S., 2010. The changing nature of financial intermediation and the financial crisis of2007–2009. Economics 2.

Alessandri, P., Nelson, B. D., 2015. Simple banking: profitability and the yield curve. Journal of Money,Credit and Banking 47, 143–175.

Allen, K. D., Davidson, T. R., Hein, S. E., Whitledge, M. D., 2016a. Dodd-frank’s federal depositinsurance reform. Community Banking Conference of 2016, Federal Reserve Bank of St. Louis .

Allen, K. D., Hein, S. E., Whitledge, M. D., 2016b. The evolution of the federal reserve’s term auctionfacility and community bank utilization .

Baker, S. R., Bloom, N., Davis, S. J., 2016. Measuring economic policy uncertainty. The QuarterlyJournal of Economics 131, 1593–1636.

Banerjee, R. N., Mio, H., 2017. The impact of liquidity regulation on banks. Journal of FinancialIntermediation .

Barth, J. R., Caprio, G., Levine, R., 2004. Bank regulation and supervision: what works best? Journalof Financial intermediation 13, 205–248.

Begenau, J., Piazzesi, M., Schneider, M., 2015. Banks’ risk exposures. Tech. rep., National Bureau ofEconomic Research.

Berger, A. N., Black, L. K., Bouwman, C. H., Dlugosz, J., 2015. The federal reserve’s discount windowand taf programs:’pushing on a string?’ .

Bernanke, B. S., Gertler, M., Gilchrist, S., 1999. The financial accelerator in a quantitative businesscycle framework. Handbook of macroeconomics 1, 1341–1393.

Berndt, A., Gupta, A., 2009. Moral hazard and adverse selection in the originate-to-distribute model ofbank credit. Journal of Monetary Economics 56, 725–743.

Blanchard, O., Dell’Ariccia, G., Mauro, P., 2010. Rethinking macroeconomic policy. Journal of Money,Credit and Banking 42, 199–215.

Bonaime, A. A., Gulen, H., Ion, M., 2016. Does policy uncertainty affect mergers and acquisitions? .

Bordo, M. D., Duca, J. V., Koch, C., 2016. Economic policy uncertainty and the credit channel: Aggre-gate and bank level us evidence over several decades. Journal of Financial Stability 26, 90–106.

Borio, C., Zhu, H., 2012. Capital regulation, risk-taking and monetary policy: a missing link in thetransmission mechanism? Journal of Financial Stability 8, 236–251.

18

Boyd, J. H., De Nicolo, G., 2005. The theory of bank risk taking and competition revisited. The Journalof finance 60, 1329–1343.

Boyson, N., Helwege, J., Jindra, J., 2014. Crises, liquidity shocks, and fire sales at commercial banks.Financial Management 43, 857–884.

Brewer III, E., Minton, B. A., Moser, J. T., 2000. Interest-rate derivatives and bank lending. Journal ofBanking & Finance 24, 353–379.

Brogaard, J., Detzel, A., 2015. The asset-pricing implications of government economic policy uncertainty.Management Science 61, 3–18.

Bruno, V., Shin, H. S., 2015. Capital flows and the risk-taking channel of monetary policy. Journal ofMonetary Economics 71, 119–132.

Chernenko, S., Faulkender, M., 2012. The two sides of derivatives usage: Hedging and speculating withinterest rate swaps. Journal of Financial and Quantitative Analysis 46, 1727–1754.

Chodorow-Reich, G., 2014. The employment effects of credit market disruptions: Firm-level evidencefrom the 2008–9 financial crisis. The Quarterly Journal of Economics 129, 1–59.

Cornett, M. M., McNutt, J. J., Strahan, P. E., Tehranian, H., 2011. Liquidity risk management andcredit supply in the financial crisis. Journal of Financial Economics 101, 297–312.

Cyree, K. B., 2016. The effects of regulatory compliance for small banks around crisis-based regulation.Journal of Financial Research 39, 215–246.

Cyree, K. B., Griffiths, M. D., Winters, D. B., 2013. Federal reserve financial crisis lending programsand bank stock returns. Journal of Banking & Finance 37, 3819–3829.

Dahl, D., Meyer, A., Neely, M., 2016a. Bank size, compliance costs and compliance performance incommunity banking .

Dahl, D., Meyer, A. P., Neely, M. C., 2016b. Scale Matters: Community Banks and Compliance Costs.The Regional Economist .

De Nicolo, G., Gamba, A., Lucchetta, M., 2014. Microprudential regulation in a dynamic model ofbanking. Review of Financial Studies 27, 2097–2138.

Dell’Ariccia, G., Laeven, L., Suarez, G. A., 2016. Bank leverage and monetary policy’s risk-takingchannel: Evidence from the united states. The Journal of Finance .

Demirguc-Kunt, A., Detragiache, E., 2002. Does deposit insurance increase banking system stability?an empirical investigation. Journal of monetary economics 49, 1373–1406.

Demsetz, R. S., Strahan, P. E., 1995. Historical patterns and recent changes in the relationship betweenbank holding company size and risk. Economic Policy Review 1.

Diamond, D. W., 1984. Financial intermediation and delegated monitoring. The Review of EconomicStudies 51, 393–414.

19

Duffee, G. R., Zhou, C., 2001. Credit derivatives in banking: Useful tools for managing risk? Journalof Monetary Economics 48, 25–54.

Duffie, D., Scheicher, M., Vuillemey, G., 2015. Central clearing and collateral demand. Journal of Fi-nancial Economics 116, 237–256.

Ellul, A., Yerramilli, V., 2013. Stronger risk controls, lower risk: Evidence from us bank holding com-panies. The Journal of Finance 68, 1757–1803.

Esty, B. C., 1998. The impact of contingent liability on commercial bank risk taking. Journal of FinancialEconomics 47, 189–218.

Fahlenbrach, R., Prilmeier, R., Stulz, R. M., 2012. This time is the same: Using bank performancein 1998 to explain bank performance during the recent financial crisis. The Journal of Finance 67,2139–2185.

Farhi, E., Tirole, J., 2012. Collective moral hazard, maturity mismatch, and systemic bailouts. TheAmerican Economic Review 102, 60–93.

FDIC, FFIEC, F. R., 2015. Summary of changes to call report schedule rc-r, part ii, and schedule rc-l effective march 31, 2015. https://www.fdic.gov/news/news/financial/2015/fil15008a.html February20.

Fishburn, P. C., Burr Porter, R., 1976. Optimal portfolios with one safe and one risky asset: Effects ofchanges in rate of return and risk. Management science 22, 1064–1073.

Froot, K. A., Scharfstein, D. S., Stein, J. C., 1993. Risk management: Coordinating corporate investmentand financing policies. the Journal of Finance 48, 1629–1658.

Gerali, A., Neri, S., Sessa, L., Signoretti, F. M., 2010. Credit and banking in a dsge model of the euroarea. Journal of Money, Credit and Banking 42, 107–141.

Giavazzi, F., McMahon, M., 2012. The household effects of government spending. In: Fiscal policy afterthe financial crisis , University of Chicago Press, pp. 103–141.

Gissler, S., Oldfather, J., Ruffino, D., 2016. Lending on hold: Regulatory uncertainty and bank lendingstandards. Journal of Monetary Economics 81, 89–101.

Gorton, G., Rosen, R., 1995. Banks and derivatives. In: NBER Macroeconomics Annual 1995, Volume10 , MIT Press, pp. 299–349.

Hein, S. E., Koch, T. W., MacDonald, S. S., 2005. On the uniqueness of community banks. EconomicReview-Federal Reserve Bank of Atlanta 90, 15.

Hentschel, L., Kothari, S. P., 2001. Are corporations reducing or taking risks with derivatives? Journalof Financial and Quantitative Analysis 36, 93–118.

Husted, L. F., Rogers, J. H., Sun, B., 2017. Monetary policy uncertainty .

Jermann, U. J., Yue, V. Z., 2013. Interest rate swaps and corporate default .

20

Jimenez, G., Ongena, S., Peydro, J.-L., Saurina, J., 2014. Hazardous times for monetary policy: Whatdo twenty-three million bank loans say about the effects of monetary policy on credit risk-taking?Econometrica 82, 463–505.

Keeley, M. C., 1990. Deposit insurance, risk, and market power in banking. The American EconomicReview pp. 1183–1200.

Kim, R., 2017. Too big to lend? the ”invisible hand” separating commercial and investment banking.Working Paper .

Koijen, R. S., Philipson, T. J., Uhlig, H., 2016. Financial health economics. Econometrica 84, 195–242.

Krishnamurthy, A., Vissing-Jorgensen, A., 2011. The effects of quantitative easing on interest rates:channels and implications for policy. Tech. rep., National Bureau of Economic Research.

Laeven, L., Levine, R., 2009. Bank governance, regulation and risk taking. Journal of Financial Eco-nomics 93, 259–275.

Landier, A., Sraer, D., Thesmar, D., 2013. Banks’ exposure to interest rate risk and the transmission ofmonetary policy. Tech. rep., National Bureau of Economic Research.

Li, S., Marinc, M., 2014. The use of financial derivatives and risks of us bank holding companies.International Review of Financial Analysis 35, 46–71.

Lins, K. V., Servaes, H., Tamayo, A., 2007. Does derivative accounting affect risk management. Inter-national survey evidence .

Maddaloni, A., Peydro, J.-L., 2011. Bank risk-taking, securitization, supervision, and low interest rates:Evidence from the euro-area and the us lending standards. Review of Financial Studies 24, 2121–2165.

Merton, R. C., 1977. An analytic derivation of the cost of deposit insurance and loan guarantees anapplication of modern option pricing theory. Journal of Banking & Finance 1, 3–11.

Minton, B. A., Stulz, R., Williamson, R., 2009. How much do banks use credit derivatives to hedgeloans? Journal of Financial Services Research 35, 1–31.

Paligorova, T., Santos, J. A., 2016. Monetary policy and bank risk-taking: Evidence from the corporateloan market. Journal of Financial Intermediation .

Parlour, C. A., Winton, A., 2013. Laying off credit risk: Loan sales versus credit default swaps. Journalof Financial Economics 107, 25–45.

Purnanandam, A., 2007. Interest rate derivatives at commercial banks: An empirical investigation.Journal of Monetary Economics 54, 1769–1808.

Purnanandam, A., 2011. Originate-to-distribute model and the subprime mortgage crisis. Review ofFinancial Studies 24, 1881–1915.

Rajan, R. G., 2006. Has finance made the world riskier? European Financial Management 12, 499–533.

Ritz, R. A., Walther, A., 2015. How do banks respond to increased funding uncertainty? Journal ofFinancial Intermediation 24, 386–410.

21

Seelig, S. A., 1974. Rising interest rates and cost push inflation. The Journal of Finance 29, 1049–1061.

Sette, E., Gobbi, G., 2015. Relationship lending during a financial crisis. Journal of the EuropeanEconomic Association 13, 453–481.

Shan, S. C., Tang, D. Y., Yan, H., 2014. The effects of credit default swaps on bank capital, lending,risk and return. Tech. rep.

Shao, Y., Yeager, T. J., ???? The effects of credit derivatives on us bank risk and return, capital andlending structure .

Shen, X. S., Hartarska, V., 2013. Financial derivatives at community banks .

Shleifer, A., Vishny, R. W., 2010. Unstable banking. Journal of financial economics 97, 306–318.

Sinkey, J. F., Carter, D. A., 2001. Evidence on the financial characteristics of banks that do and do notuse derivatives. The Quarterly Review of Economics and Finance 40, 431–449.

Smith, C. W., Stulz, R. M., 1985. The determinants of firms’ hedging policies. Journal of financial andquantitative analysis 20, 391–405.

Taylor, J. B., 2009. The financial crisis and the policy responses: An empirical analysis of what wentwrong. Tech. rep., National Bureau of Economic Research.

Vuillemey, G., 2015. Derivatives and interest rate risk management by commercial banks. Available atSSRN 2644681 .

Wheelock, D. C., 1992. Regulation and bank failures: New evidence from the agricultural collapse ofthe 1920s. The Journal of Economic History 52, 806–825.

22