Double the Insurance, Double the Funds?

Transcription

1 Double the Insurance, Double the Funds? Anna-Leigh Stone a,* a Brock School of Business, Samford University, 800 Lakeshore Drive, Birmingham, AL 35229, United States This Version: January 2017 Abstract The Depositors Insurance Fund and Share Insurance Fund, private organizations in Massachusetts, have provided a full insurance guarantee on all deposits held at Massachusetts state chartered savings and cooperative institutions since This paper is the first to examine the members of these organizations. I find that the member institutions hold a significant amount of deposits over the Federal Deposit Insurance Corporation (FDIC) limit when compared to the remaining 49 state chartered savings and cooperative institutions. In addition, the period surrounding the Great Recession, when the FDIC insured noninterest-bearing deposits held at participating institutions, is examined and I find that several of the member banks still participated in the program despite the additional fee they were required to pay, thus doubling their full insurance guarantee on noninterest-bearing accounts. These member institutions had more total deposits and number of accounts over the FDIC s limit, but did not have a statistically significant amount of noninterest-bearing deposits or number of noninterest-bearing accounts over the FDIC s limit. Keywords: Deposit Insurance, Depositors Insurance Fund, Share Insurance Fund, Federal Deposit Insurance Corporation, Transaction Account Guarantee Program JEL Classification: G21, G28, G01 * Corresponding author. Address: 800 Lakeshore Drive, Birmingham, AL 35229; Tel : Fax: address: alstone@samford.edu (A. Stone) The author would like to thank Matt Calnan, a senior financial analyst at the Depositors Insurance Fund for aiding in the execution of the paper. 0

2 I. Introduction During the most recent financial crisis, the Federal Deposit Insurance Corporation (FDIC) enacted the Transaction Account Guarantee (TAG) Program, which provided unlimited FDIC insurance on noninterest-bearing deposits. While nationally this is one of the few times that certain types of deposits have been guaranteed in full, deposits held at savings and cooperative institutions chartered in Massachusetts have had a full insurance guarantee on all deposits since These institutions are members of the FDIC but also members of one of two private organizations called the Depositors Insurance Fund (DIF) and the Share Insurance Fund (SIF). While the FDIC insures funds up to $250,000, the DIF and SIF ensures all deposits, regardless of category, in excess of the FDIC guarantee. One of the purposes of this paper is to study the DIF and SIF. Neither the academic literature nor the popular press have devoted much attention to the funds held by these organizations, their member institutions, or the unlimited insurance guarantee that both offer. Outside of incorporating risk varying premiums into the FDIC (Pennacchi 1987, 2003), much of the academic literature that has focused on deposit insurance has focused on the funding of the FDIC s deposit insurance fund (see Merton (1977), Marcus and Shaked (1984), Pennacchi (1999) and Kuritzkes, et al. (2005)). To the author s knowledge, this is the first paper to examine the DIF and SIF member banks. In addition, it is of interest to see how deposits changed at these banks during the financial crisis. Pennacchi (2006) stated that FDIC insurance could explain the flight to bank deposits. Thus, it is worth examining whether or not banks that offer an unlimited guarantee experienced a flight to deposits during the financial crisis. In light of discussions about potential changes to the FDIC, it is also worthwhile to examine the effectiveness of these private organizations. In an FDIC report examining available options that might be incorporated into the program, Bradley and Craig (2007) state that while Massachusetts has the only organizations that currently insure deposits in excess of the FDIC limit, there are three independent businesses that will provide additional insurance for a fee, although few banks have taken advantage of it. They also cite that the FDIC benefits from private excess insurance programs because of the reduction in 1

3 moral hazard stemming from the additional monitoring provided by these private insurers, as well as an increase in deposit coverage without a proportional increase in the size of the FDIC s Deposit Insurance Fund, risk exposure, or extra premiums assessed on banks. In a 2001 report, the FDIC recommended the need for additional analysis of private deposit insurer coverage before the possibility of further integration into the system (FDIC, 2001). Therefore, it is important to examine the DIF and SIF as examples of programs that could be applied, not just in other states, but to other classification of institutions that might bring improvements to the FDIC. The results show that DIF member institutions are larger in size, have fewer past due loans, and have a larger amount of deposits in excess of the FDIC limit than other state-chartered savings institutions located in the remaining 49 states. SIF members are smaller than DIF members and are not as liquid and have more deposits in excess of the FDIC limit than other state-chartered cooperative institutions. The finding that both DIF and SIF member institutions have more deposits over the FDIC limit suggests that depositors at these banks utilize the additional insurance guarantee that is not available to similar institutions in other states. Even prior to the financial crisis, member institutions held more in excess deposits than other state chartered institutions. The other focus of this paper is to examine the TAG program participation of these Massachusettscharted institutions. The TAG program provided unlimited FDIC insurance on noninterest-bearing deposits from October 2008 until December Over the course of the program, sixty-eight savings institutions were members of the DIF and seventy-nine cooperative institutions were members of the SIF. I first examine whether banks that maintained membership in the DIF and SIF participated in the TAG program. In order to fund the TAG program, participating banks were required to pay fees on any deposits in excess of the $250,000 limit. However, members of the DIF and SIF were already paying a premium for unlimited insurance through their respective excess insurance organizations. With the increased fee that banks were responsible for paying for participation in the TAG program, it is of interest to see how this affected participation in this program. 2

4 Surprisingly, 87% of DIF members and 91% of SIF members did not opt-out of the TAG program initially. I thus examine characteristics of state-chartered member banks that might explain participation in the program and find that these participating institutions did have more deposits in excess of the FDIC limit than the nonparticipants did. There are three reasons why member banks might have remained in the TAG program despite having full insurance on all deposits. First, if the bank believed that customers would value the additional insurance, they might have found that the benefit of staying in the program outweighed the fees. I do find that there is a positive and statistically significant relationship between growth in total deposits and participation in the program. However, growth in noninterest-bearing deposits is not a significant determinant of participation in the TAG program. Second, it could be that the banks did not want to opt out of the TAG program because they did not want the bank s name to appear on one of the lists of nonparticipants that was made publicly available. I find that 46% of DIF institutions and 56% of SIF institutions chose not to participate in the debt guarantee program (DGP), where the list was also made publicly available, but chose to participate in the TAG program. Thus, I find no evidence that participation was out of fear of negative publicity. Finally, if the bank did not have deposits in excess of the $250,000 insurance limit or had minimal deposits in excess of the limit, it could imply that the banks did not have to pay a large premium to participate in the TAG program. I find that the member banks that remained in the program had very few noninterest-bearing accounts over the limit, however, banks with more accounts in excess of the insurance limit were more likely to participate in the TAG program. This paper also adds to the growing literature on the TAG program. While the majority of the analysis has come from banks discussing the implications of the full insurance coverage, Hoskins (2012) and Schich (2008) discuss the unlimited insurance guarantees, the potential impact upon its expiration, and the implications for extending the insurance. 1 However, no empirical tests are provided in any of the 1 Goldman Sachs Research Department published US Daily: FDIC Deposit Guarantees: Another Year-End Risk on September 4, It is available at: 3

5 discussions. Shapiro and Dowson (2012) look at the extension of the unlimited insurance provisions and find that the extension would send a negative signal to the market and restrict bank competition and stability. While Boyle et al. (2015) do not directly examine the TAG program, they examine whether the onset of deposit insurance during a crisis reduced the amount of runs at institutions. They find that it does not affect deposit withdrawals in the short run. Stone (2015) examines bank participation in the TAG program and the impact on corporate cash holdings due to the full guarantee and finds that the banks that participated in the program saw a larger increase in noninterest-bearing deposits from those that opted-out of the program. The paper continues as follows: Sections II and III discuss bank participation in either the DIF or SIF and TAG program, respectively. Section IV examines the aforementioned reasons for why banks might have participated in the TAG program and Section V concludes. II. Participation in the Depositors Insurance Fund and Share Insurance Fund a. Depositors and Share Insurance Funds Both the DIF and SIF are private, industry-sponsored insurance companies that were established by the Massachusetts legislature in 1934 in response to the Great Depression. 2 Originally, the DIF was established within The Mutual Savings Central Fund, Inc., which had been created to make liquidity loans to savings institutions. Similarly, the SIF was established under The Co-operative Central Bank, which has been established to provide cash reserves to cooperative banks. Both the DIF and SIF have always been independent organizations and are not backed by the federal government or the Commonwealth of Massachusetts. The DIF insures Massachusetts-chartered savings institutions exclusively while the SIF insures Massachusetts-chartered cooperative banks. Originally, institutions that where members of the guarantees.pdf. Bank of American published Life After Full FDIC Insurance in April It is available at: 2 The DIF was created under Chapter 44 of the acts of 1932 and enacted under Chapter 43 of the acts of

6 DIF and SIF were not members of the FDIC. 3 This changed in 1956 when the DIF and SIF started insuring deposits held over the limit covered by the FDIC and was a sole excess insurer by 1986 when all banks were members of both the FDIC and their respective organization. 4 Membership in the funds are open to all Massachusetts-chartered saving and cooperative banks but both funds insure all deposits at branches affiliated with a member located out of state as well as online deposits of member institutions. Thus, insurance is granted based on the institution s headquarters and not the depositors residences. The DIF and SIF boast that no depositor has ever lost any money when insured by both the excess insurance and the FDIC. 5 Even during the 1980s when 19 DIF member banks failed, the DIF insured the $250 million in excess deposits. 6 Banks that do become members of the DIF or SIF are required to pay an annual assessment for the insurance that varies by each institution based on the amount of excess deposits held. Full excess insurance is provided on all deposit accounts at these institutions including personal and business checking, negotiable order of withdrawal (NOW) accounts, money market deposit accounts, savings accounts, certificates of deposits (CDs), and retirement deposit accounts. 7 During the 2008 financial crisis, the DIF had sixty-eight member organizations and as of October 2016, the DIF had fifty-five members. During the financial crisis, the SIF had fifty-two member organizations and as of October 2016, had fourty-eight members. Over time the number of member institutions has varied because of failure, mergers and acquisitions, and changes in charters. 8 3 This was due in part to the confusion regarding whether savings institutions were insured by the FDIC. Even though the Federal Reserve Act was amended to include a Fund for Mutuals to insure savings institutions in July 1934, there was hesitation to join and thus these institutions joined the DIF instead. Members of The Mutual Savings Central Fund were also prohibited from joining the FDIC in Section 10 of Chapter 324 in the acts of 1956 amended the original act to include FDIC membership. In 1985, the Massachusetts Commissioner of Banks required all Massachusetts savings banks to obtain federal deposit insurance. 5 See for the DIF and for the SIF While the DIF does not limit the deposits insured, member institutions may limit the size of the deposit it is willing to accept. These limits vary from bank to bank based on bank policies and funding. 8 If an institution that was a member of the DIF applies for a federal charter, it can reapply for membership in the DIF. This has happened once during this time frame when Brookline Bank switched from a federal charter back to a savings institution and became a DIF member. 5

7 b. Participation in the Programs Data on bank deposits and assets held at Massachusetts-chartered savings institutions is gathered from the FDIC Statistics on Depository Institutions (SDI) and the Federal Reserve Bank of Chicago s Commercial Bank Data. The FDIC and Chicago Federal Reserve provide information filed in call reports from all insured institutions. For the purposes of this paper, a state-chartered savings institution is one that has a bank class code provided by the FDIC of a savings bank with a state charter that is supervised by the FDIC (SB) or a Federal Reserve member state-chartered savings bank that is supervised by the Federal Reserve (SM). Additionally, each savings institution should have the Federal Reserve two-digit classification of 10, which indicates a state-chartered savings bank. A cooperative institution is one classified by the FDIC with a bank class code of a savings bank with a state charter that is supervised by the FDIC (SB) or a state savings institution supervised by the office of thrift supervision (SA). Again, each cooperative institution should have the Federal Reserve two-digit classification of 17, which indicates a cooperative bank. Table 1 presents the total assets and total deposits of state-chartered institutions that are members of the DIF and state-chartered banks that are members of the SIF, and compares them to all other statechartered savings and cooperative institutions in the remaining 49 states. DIF members are significantly larger in terms of both total assets and total deposits relative to their SIF counterparts in Massachusetts. Compared to other state-chartered savings institutions, DIF member institutions are slightly smaller in terms of the average amount of total assets and deposits. However, based on the median numbers, DIF savings institutions are larger. When looking at the cooperative banks, SIF members are significantly smaller than the average bank in the remaining 49 states. However, again based on median numbers, SIF members are slightly larger. Figure 1 plots total deposits at DIF and SIF member state-chartered savings and cooperative banks in Massachusetts from 2002 until Compared to their SIF counterparts, DIF member banks have consistently held more total deposits. Starting in 2005, DIF member banks experience an increase in total deposits that further increases during the financial crisis. SIF member banks did not experience a similar 6

8 increase in deposits, but remained relatively constant in the amount of total deposits after While Figure 1 shows that DIF member banks experienced an increase in deposits leading up to the financial crisis, did they experience a flight to deposits during the crisis? In January 2009, John Boucher, the CEO of South Shore Savings Bank, a DIF member, said his bank s experience of the flight to safety was immediate and dramatic in September and October Some of that has to do with DIF (Bielski, 2009). However, based on Figure 1 it does not appear that in the aggregate DIF member banks experienced a flight to deposits leading up to the financial crisis that is more defined than the increase that they were already experiencing. Figure 2 shows the amount of excess deposits held by the different classifications of banks. Excess deposits are defined as deposits held at the bank in excess of the FDIC insurance limit and are the type of deposits that the DIF and SIF guarantee. The excess deposit data is collected by the FDIC and reported on call reports of all institutions. However, in 2008, when the deposit insurance limit was temporarily raised to $250,000, the FDIC still reported data based on the $100,000 insurance limit. These call reports were not updated until September That is why during the fourth quarter of 2009 we see a decline in the amount and number of accounts over the limit. We see that DIF members have more deposits over the limit than SIF members. The amount and number of deposits over the limit for DIF members increases during the financial crisis, but for SIF members it seems to stay stagnant. In addition to presenting univariate results, Probit regression are run to identify traits that might be common among member banks. Control variables are included that are consistent with the literature to control for a bank s CAMELS rating. These include: the Tier 1 Leverage Ratio to measure capital adequacy, the ratio of past due loans to total loans to control for asset quality, the efficiency ratio to measure managerial effectiveness, the return on average assets (ROAA) to measure earnings, the liquidity ratio to control for liquidity, and the funding gap to control for sensitivity to interest rates. The natural log of total assets is also included to control for the size of the institution and net loan growth is included to capture changes in credit. I also control for the standard deviation of assets across the institution to capture variation in total assets. All variables are defined in the appendix. Unique to this paper, I control 7

9 for excess deposits, measured as the ratio of deposits in excess of the FDIC limit divided by total deposits, to see if member banks carry a higher percentage of deposits in excess of the FDIC limit. The dependent variable is a dummy variable, DIF member [SIF member], equal to one if the bank is a DIF [SIF] member and zero otherwise. The results are presented in Table 2. Panel A reports the results for DIF members and Panel B reports the results for SIF members. The first three columns compare the members of the DIF and SIF with other Massachusetts chartered institutions and the last 3 columns compare them to other state-chartered savings or cooperative institutions. The first and fourth columns examine the entire period from the first quarter of 2002 until the fourth quarter of In Column 4 of Panel A, we see that DIF member banks tend to have fewer past due loans, are more efficient, have a lower return on average assets, and better liquidity ratios than other state-chartered savings institutions. They are also larger and have a lower standard deviation of assets. The results also show that member institutions do not have a Tier 1 Leverage Ratio that is significantly different than nonmembers, indicating that they are not any different in terms of capital adequacy. 9 Finally, a positive and statistically significant relationship is found between excess deposits and membership, indicating that member banks do have a statistically larger amount of excess deposits than their counterparts. This could possibly indicate that members at these banks are taking advantage of the extra insurance. Columns 2 and 5 present regression results using the period from September 2009 to December 2015, when the FDIC insurance was raised to $250,000 and columns 3 and 6 present regressions using the period from January 2002 to just before October 2008 when the FDIC insurance limit was $100,000. As mentioned previously, in 2008, when the deposit insurance limit was temporarily raised to $250,000 the FDIC still reported data based on the $100,000 insurance limit and the call reports were not updated until September Thus, in order to ensure that excess deposits are accurately measured, the time period between October 2008 and September 2009 is removed. Removing the period of reporting inconsistencies 9 In addition to including the Tier 1 Leverage Ratio, the Tier 1 risk-based capital ratio and the total risk-based capital ratio are also included. All of these provide similar results to those presented here and are available upon request. 8

10 does little to change the results for DIF members. Member banks still have a statistically significant amount of excess deposits than other state chartered savings institutions. This also indicates that the increase in total deposits shown in Figure 1 might be partially driven by an increase in funds in excess of the deposit limit and not necessarily in new deposits placed at the bank. SIF member banks, shown in Panel B, have very different results. In column 4, I do not find that these banks have very many traits that are statistically different from any other state-chartered cooperative institution with the exception of the liquidity ratio. The largest difference is that SIF member banks also have more deposits in excess of the FDIC limit than their counterparts. This again shows that members might take advantage of the additional insurance and hold more in deposits at these member institutions. In column 5, the period since the financial crisis, I find that SIF members have lower capital adequacy, a lower liquidity ratio, larger net loan growth, and, again, hold more deposits over the FDIC limit. Prior to the insurance increase at the FDIC, these institutions had a smaller amount of past due loans, a higher liquidity ratio, lower net loan growth, and a smaller deviation in assets. More importantly, they still had more in excess deposits than other state-chartered cooperative institutions. Thus, again we see that members of cooperative institutions hold more in deposits than members of other state-chartered cooperative institutions suggesting members at SIF banks are utilizing the extra insurance coverage. III. Participation in the Transaction Account Guarantee Program The TAG program was part of the Temporary Liquidity Guarantee (TLG) Program passed by the FDIC in response to the financial crisis in October While the FDIC insurance limit was temporarily raised to $250,000 during this time, the TAG program provided unlimited FDIC insurance on all noninterestbearing deposits held at FDIC insured institutions from October 2008 until December The program was extended twice until June 2010 and finally December At the start of the program, the fee for insurance protection was 10 basis points on any amount over the $250,000 limit. However, during the 9

11 extensions of the TAG program the fee was increased to basis points on amounts over the $250,000 limit, depending on the riskiness of the bank. The TAG program allowed banks to opt-out of the program upon the original implementation as well as each of the two extensions. The FDIC provided the names of all institutions that chose to opt-out of all three periods of the program on their website. 10 Once a bank elected not to participate in the program, it was excluded from all the subsequent extensions. Over the life of the program, a total of 2,077 banks opted out. Out of the sixty-eight DIF member institutions, twenty-eight total institutions opted-out of the TAG program: nine during the original implementation, thirteen during the first extension, and six during the second extension. Out of the seventy-nine SIF member institutions, twenty-eight total institutions opted out: seven during the original implementation, eighteen during the first extension, and three during the final extension. This means that some institutions not only had a full insurance guarantee provided by the DIF or SIF, but also had a full insurance guarantee on noninterest-bearing accounts provided by the FDIC. At the beginning of the program, the banks that opted out were, on average, smaller as measured by total assets and total deposits than the banks that stayed in the program. During the extensions, larger banks chose to exit the program, as indicated by a shift in the average total assets and deposits. Table 3 provides summary statistics on the percentage of deposits that noninterest-bearing deposits comprised based on the member organizations and participation in the TAG program. We see that banks that remained in the program tend to have a larger percentage of deposits than those that exit during the original implementation; however, this shifts over the course of the program. Similar to the analysis of the excess insurers, a Probit regression is run to identify traits that might be common among banks that participated in the TAG program. The same bank control variables are included as in the previous model as well as a control for the amount of a bank s noninterest-bearing 10 These lists can be found at: 10

12 deposits-to-total deposits. It is of interest to see if the amount of noninterest-bearing deposits at a bank was a factor in its decision to remain in the program. 11 The dependent variable is a dummy variable, participant, equal to one if the bank was a participant in the TAG program and zero otherwise. In addition, to control for banks that exited the TAG program at the extensions the dummy variable is allowed to change across bank quarters. The sample period is also restricted from 2008Q4 until 2010Q4, as these were the only quarters that the banks had an option to opt out. The sample is also restricted to only DIF and SIF member institutions. The results are presents in Table 4. The first four columns look at participation when separately controlling for the excess deposits variable, the number of accounts over the FDIC limit, the noninterestbearing accounts over the limit, and the number of noninterest-bearing accounts over the limit. The only variable that is consistently correlated with participation is the standard deviation of assets. Banks with a smaller standard deviation of assets were more likely to participate in the TAG program. However, the total number of deposits over the FDIC limit and the number of accounts over the FDIC limit are positively correlated with TAG participation. The data on noninterest-bearing accounts over the limit is not statistically significant. The last four columns include a combination of the above-mentioned variables, but we do not see very many significant relationships other than those mentioned. IV. Why Double Insurance? It is important to note that the institutions that were members of the DIF or SIF and participated in the TAG program were paying for the additional insurance coverage provided by both organizations. Some might ask why these banks would pay for double insurance coverage. There are three primary reasons for participating in the TAG program while also maintaining insurance from the DIF and SIF: value-added to 11 In unreported regressions, I do include the amount of noninterest-bearing accounts over the limit as a percentage of total deposits. However, this data is only reliably available for nonparticipants starting in the fourth quarter of 2010, so the results are based on a much smaller sample. I do not find any evidence that the amount of noninterestbearing deposits over the limit had any effect on participation in the program during this quarter. 11

13 customers, public disclosure, and minimal fees. This section will present a discussion and analysis of these explanations. a. Added Value to Customers Massachusetts savings and cooperative banks might have paid for the double coverage because the banks believed customers might value the additional insurance. This additional benefit could manifest itself in the form of a growth in the amount of total deposits, and not just noninterest-bearing deposits, at the institution. In an FDIC report in 2001, the FDIC stated that some banks have expressed a willingness to pay an additional premium to increase FDIC insurance coverage in order to retain existing large deposits or obtain new deposits (FDIC, 2001). Thus, if the banks anticipated that deposits might follow the additional noninterest-bearing coverage, they could have remained in the program for the additional customer reassurance. It also could be that the growth in deposits might outweigh the fees of the program, thus bringing a benefit for a small cost. It was previously shown in Table 1 that the DIF member banks are on average larger and hold more deposits than the SIF member banks in Massachusetts and that while DIF members are closer to the size of other state-chartered savings institutions, SIF members are considerably smaller than other statechartered cooperative institutions. However, both classifications of banks had the additional insurance, which begs the question, do banks see a benefit to the TAG program as well as the additional excess insurance? Figure 1 also documents that while DIF member institutions did not experience a large increase in total deposits that was separate from the increasing trend these banks were already experiencing, SIF member banks did experience an increase in deposits during this time that they had not experienced prior. While Table 1 and Figure 1 only show data on total deposits I am also interested in looking at whether noninterest-bearing deposits increased. Thus, the growth in noninterest-bearing deposits for all member institutions based on participation in the TAG Program is presented in Figure 3. Keep in mind that the large cutoffs in the series are because banks were allowed to change participation at each 12

14 extension. Banks that participated in the program do not display a large increase in noninterest-bearing deposits over the course of the program. In fact, from October 2008 until December 2009, the aggregate amount of noninterest-bearing deposits at DIF members only increases by 6.7%. This represents an increase of $173 million in deposits across fifty-nine institutions. To compare, all other participating institutions in the United States experienced a 9.1% increase in aggregate noninterest-bearing accounts over this time. b. Public Disclosure In addition to the added value, banks could have felt pressure to participate in the program because the list of nonparticipants was to be published on the FDIC s website. Perhaps institutions did not want the negative impression that not participating might have in the ability to attract and retain deposits. And while the DIF and SIF members provided an unlimited guarantee despite remaining in the program, the banks might have still been seen as less attractive due to the nonparticipation. While this pressure on financial institutions is not easy to measure, I compare institutions that opted out of the TAG program with those that opted out of another TLG Program, the Debt Guarantee Program (DGP). The DGP placed an FDIC guarantee on debt issued from October 2008 until December 2009 that met certain requirements. Banks were allowed to opt-out of this program as well, but financial institutions that remained in the program were not required to issue guaranteed debt under the program nor were they assessed a fee for simply remaining in the program. The FDIC also made the list of institutions that chose not to participate publicly available on their website. 12 While the list of banks that issued debt under the program was not made publicly available until after the program had expired, institutions that issued guaranteed debt did disclose it to news outlets and in financial statements. 13 With regards to the opt-out lists for the DGP, some say that the FDIC indirectly punished banks that opted out of the program by 12 The list can be found at the same place as the TAG program opt out lists: 13 Out of all the participants in the DGP, the author only found one savings bank that issued debt backed by the FDIC. 13

15 releasing their names while not disclosing the names of the banks that choose to issue debt under the program (Black et al., 2014). Of the firms that issued guaranteed debt, Amborse, et. al. (2013) find that, on average, the stock market reacted negatively to the announcement of debt issues while the bond markets reacted positively to the news. Using the DGP opt-out list, I examine the participation of banks in both the TAG program and the DGP. It could be that banks followed similar strategies with both programs because of the public perception. Banks on the DGP opt-out list are matched based on the certificate number provided by the FDIC. Panel A of Table 5 presents the number and percentage of all DIF member institutions based on participation decision in the TAG program and DGP. Overall, 41% of DIF member institutions participated in both the DGP and the TAG program upon the creation of the program. This percentage does fall as banks opted out of the extensions of the program. Only 13% of member banks opted out of both the DGP and the TAG program and 45% of banks opted out of the DGP, but choose to participate in the TAG program. Panel B presents the number and percentages for SIF member banks. Upon the implementation of the programs, 35% of institutions participated in both while almost 9% did not participate in either program. Across DIF and SIF institutions, there does not seem to be a consensus of banks deciding to opt-out or remain in the programs. Thus, it is hard to determine if banks were fearful of the market s reaction to the opt-out list. c. Minimal Fees The final explanation for the participation in the TAG program could be fee related. The fees placed by the FDIC were on amounts in excess of the $250,000 limit. Thus, if a bank did not have deposits in excess of the $250,000 insurance limit or had minimal deposits in excess of the limit, they would not have to pay a large premium to remain in the program. Statistics on the amount of noninterest-bearing deposits held by DIF members based on participation in the program are presented in Table 6. Keep in mind that the amounts do shift after the first four 14

16 quarters due to banks withdrawing from the extensions of the program. Banks that participated initially held approximately 92% of all total noninterest-bearing accounts of member banks. However, at each extension of the TAG program more banks exited so that at the end of the program a majority of the noninterest-bearing accounts are at institutions that opted out of the program. During the first extension, thirteen institutions left the TAG program, but these institutions had almost $1 billion in noninterestbearing accounts, showing evidence of institutions with larger amounts of noninterest-bearing deposits exiting the program. The fee did increase for the last year of the program, so this exodus could be due to banks that already had a full deposit guarantee provided by the DIF, that were looking at a larger fee assessments withdrawing from the program. The number and amount of noninterest-bearing accounts in excess of the $250,000 limit are also examined. The Federal Reserve Bank of Chicago provides this information for all banks that participated in the TAG program starting in October The data is presented in Table 7. I again separate the banks based on participation in the TAG program and extensions. When looking at the participation based on DIF membership, participants had a smaller amount of deposits over the $250,000 limit as well as a smaller number of individual accounts over the limit when compared to other Massachusetts chartered banks. A participating institution, on average, had just over twenty accounts over the $250,000 limit for an average amount under $15,000,000. The average number of accounts over the limit at Other Massachusetts State Chartered institutions was around 200 for an average amount of about half a billion. The SIF participating institutions had an even smaller average amount of deposits over the limit than DIF members. Remember that nonparticipants were not required to report the amount of noninterest-bearing accounts over the limit during this time, so we cannot guarantee that they did not have any accounts over the limit as reported in Table 7. However, the participants as well as nonparticipants were required to report data the final quarter of 2010 (due to the Dodd-Frank Act), so we can guarantee that those are accurate numbers and make comparisons between participants and nonparticipants. When looking at the 15

17 participation based on DIF membership, participants had a smaller amount of deposits over the $250,000 limit as well as a smaller number of individual accounts over the limit when compared to other Massachusetts chartered banks. It is also important to note that participating institutions had a smaller amount of noninterest-bearing deposits over the limit compared to nonparticipating DIF member institutions despite there being more banks participating in the program. A participating institution, on average, had twenty accounts over the $250,000 limit for an average amount of $10,652,220. The average number of accounts over the limit at nonparticipating institutions was thirty-seven for an average amount of $26,660,110. Thus, the participating institution would pay on average between $15,978 - $26,630 to remain in the program, depending on the riskiness of the institution. Looking at the SIF members, the participating institutions had a larger average amount and number of accounts over the limit than the nonparticipants although the numbers are still smaller than DIF participants or nonparticipants. Overall, it appears that the majority of member institutions that remained in the program had fewer accounts and a smaller amount over the limit and thus might have remained in the program because of the minimal expense. In summary, it appears that the banks that remained in the TAG program and its extensions carried smaller amounts in noninterest-bearing deposits over the limit. d. Regression Results What if there was not one reason for member banks participating in the TAG program, but a combination of reasons? I empirically examine all the reasons previously discussed. I only examine the member banks in the analysis, reducing the overall sample to the 147 member institutions over the nine quarters of the TAG program. Again, a Probit regression model is used where the dependent variable is a dummy variable, participant, equal to one if the bank participated in the TAG program and zero otherwise. The goal is to predict characteristics that are common to member institutions participating in the TAG program. All the control variables from Section II are included, plus some additional variables to capture the additional reasons why the banks might have participated. 16

18 To measure customer value, the growth rate in deposits from the previous quarter is included. It is assumed that if customers valued the insurance that they increased deposits at the participating institutions and thus a positive relationship is expected. I include the growth rate in both total deposits as well as noninterest-bearing deposits to see if either had an effect on the participation. To measure the fear of being on an opt-out list, a dummy variable is included that equals one if the bank stayed in the DGP program and zero otherwise. Thus, if a bank stayed in the DGP they might have stayed in the TAG program and thus a positive correlation is expected. To measure the accounts over the limit, I include the excess deposits variable described Section II. The reasoning is that the more accounts they have above the limit, the higher the fee assessed by the FDIC, and the more likely they might be to opt out of the program, so I expect a negative relationship. 14 The results can be found in Table 8. The first three columns do not include noninterest-bearing deposits to total deposits while the last three do include the variable. I do see that member banks that had a larger growth in total deposits were more likely to remain in the program, a relationship that is statistically significant at a 5% level. However, I do not see that member banks that saw a growth in noninterest-bearing deposits were more likely to participate in the TAG program. The DGP Dummy is positive and statistically significant, suggesting banks behaved uniformly in their participation decisions in both programs. I do find that banks with more excess deposits were more likely to participate in the TAG program. This is counter to the results seen in the prior table. To summarize, it appears that participation in the TAG program was driven by the growth in deposits that banks experienced and a reluctance to exit the program due to the publicly available list. Surprisingly, banks that held more deposits over the limit were more likely to remain in the program. 14 I do run regressions including the amount of noninterest-bearing accounts over the $250,000 limit introduced in part C, but this reduces the sample extensively and removes much of the variation. However, results on the main variables of interest presented remain similar. 17

19 V. Conclusion This paper is the first to examine the private deposit insurers in Massachusetts, the Depositors Insurance Fund and the Share Insurance Fund, and the member institutions of this fund. The insurance companies are two of only a few private deposit insurance institutions in the United States and insures all deposits held at Massachusetts chartered savings and cooperative banks in excess of the FDIC limit. Examining this private insurer has implications outside of academic literature, but provides a better understanding of the private insurance guarantee if the model were to be extended to other states. It is found that banks that participate in the DIF are larger in terms of total assets and total deposits, but are very similar to other state-chartered savings institutions in terms of capital adequacy. DIF members also have fewer past due loans, are more efficient, have a lower return on average assets, and are more liquid. These institutions also seem to have more deposits over the FDIC limit than their peer savings institutions, suggesting that members take advantage of the additional insurance at these institutions. SIF members are not very different from other state-chartered cooperative institutions, with the exception that they have more deposits over the FDIC limit. I then examine a period of time where the FDIC provided unlimited insurance on noninterest-bearing deposits. Because this insurance was provided for a fee, it is of interest to examine why banks that already had an unlimited insurance guarantee on all types of deposits paid the fee to double insure just one type of deposit. It is shown that banks that participated in the TAG program had a lower return on average assets and had a greater amount of excess deposits. I then provide an extensive analysis of three reasons why member institutions might have elected to participate in the TAG program: additional value assessed by customers, public pressure to participate, and the minimal fee of the additional insurance. I find evidence that the banks did not experience a significant increase in noninterest-bearing deposits but do experience an increase in total deposits. I also find that the banks remained consistent in their choice to remain in or leave both the TAG program and DGP program. Finally, I find that banks that had a larger amount of deposits in excess of the FDIC limit were more likely to remain in the TAG program. 18

20 REFERENCES Amborse, B. W., Y. Cheng, and T. D. King, 2013, The Financial Crisis and Temporary Liquidity Guarantee Program: Their Impact on Fixed-Income Markets, The Journal of Fixed Income 23, Bielski, L., 2009, Deposit gathering in hard times, ABA Banking Journal 101, Black, J. R., S. A. Hoelscher, D. Stock, 2014, Benefits of government bank debt guarantees: Evidence from the Debt Guarantee Program, unpublished working paper. Boyle, G., R. Stover, A. Tiwana, O. Zhylyevskyy, 2015, The impact of deposit insurance on depositor behavior during a crisis: A conjoint analysis approach, Journal of Financial Intermediation 24, Bradley Ch.M., V. V. Craig, 2007, An evaluation of further possible changes to the deposit insurance system, FDIC Staff Study, 12 February. Federal Deposits Insurance Corporation (FDIC), 2001, Keeping the Promise: Recommendations for Deposit Insurance Reform Hoskins, S. M., 2012, An overview of the Transactions Account Guarantee (TAG) program and the potential impact of its expiration of extension, Congressional Research Service, R Kuritzkes, A., T. Schuermann, and S. M. Iiner, 2005, Deposit insurance and risk management of the U.S. banking system: What is the loss distribution faced by the FDIC? Journal of Financial Services Research 27, Marcus, A. J. and I. Shaked, 1984, The valuation of FDIC deposit insurance using option-pricing estimates, Journal of Money, Credit and Banking, 16, Merton, R. C., 1977, An analytic derivation of the cost of deposit insurance and loan guarantees, Journal of Banking and Finance 1, Pennacchi, G., 1987, Alternative Forms of Deposit Insurance: Pricing and Bank Incentive Issues, Journal of Banking and Finance 11, Pennacchi, G., 1999, The effects of setting deposit insurance premiums to target insurance fund reserves, Journal of Financial Services Research 16, Pennacchi, G., 2003, The Cost of Deposit Insurance for Privately Held Banks: A Market Comparable Approach, Journal of Financial Services Research 24, Pennacchi, G., 2006, Deposit insurance, bank regulation, and financial system risks, Journal of Monetary Economics 53, Schich, A., 2008, Financial crisis: Deposit insurance and related financial safety net aspects, Financial Market Trends 2, Shapiro, R. J. and D. A. Dowson, 2012, The financial hazards and risks entailed in extending unlimited federal guarantees for deposits in transaction accounts 10 October. Available at: 19

21 SSRN: or Stone, A. 2015, An Examination of Unlimited FDIC Insurance and the Effect on Corporate Cash Holdings, unpublished working paper. 20

22 2002:1 2002:3 2003:1 2003:3 2004:1 2004:3 2005:1 2005:3 2006:1 2006:3 2007:1 2007:3 2008:1 2008:3 2009:1 2009:3 2010:1 2010:3 2011:1 2011:3 2012:1 2012:3 2013:1 2013:3 2014:1 2014:3 2015:1 2015:3 Figure 1 Total Deposits of Massachusetts-Chartered Institutions by DIF and SIF Membership This figure displays the total deposits, in thousands, from January 2002 until December 2015 for Massachusettscharted savings and cooperative institutions broken down by DIF and SIF membership. $60,000,000 $50,000,000 $40,000,000 $30,000,000 $20,000,000 $10,000,000 $- DIF Members SIF Members 21

23 2002:1 2002:3 2003:1 2003:3 2004:1 2004:3 2005:1 2005:3 2006:1 2006:3 2007:1 2007:3 2008:1 2008:3 2009:1 2009:3 2010:1 2010:3 2011:1 2011:3 2012:1 2012:3 2013:1 2013:3 2014:1 2014:3 2015:1 2015:3 2002:1 2002:3 2003:1 2003:3 2004:1 2004:3 2005:1 2005:3 2006:1 2006:3 2007:1 2007:3 2008:1 2008:3 2009:1 2009:3 2010:1 2010:3 2011:1 2011:3 2012:1 2012:3 2013:1 2013:3 2014:1 2014:3 2015:1 2015:3 Figure 2 Total Deposits Over the FDIC Limit This figure displays the amount, in thousands, of total deposits and the number of deposits based on membership in the DIF, SIF, or neither. Panel A: Total Amount (in thousands) of Deposits Over the FDIC Limit $160,000,000 $140,000,000 $120,000,000 $100,000,000 $80,000,000 $60,000,000 $40,000,000 $20,000,000 $- DIF Members SIF Members Other State Chartered Institutions Panel B: Total Number of Accounts Over the FDIC Limit DIF Members SIF Members Other State Chartered Institutions 22

24 Figure 3 Total Noninterest-Bearing Deposits at DIF and SIF Members Based on Participation in the TAG Program This figure displays the amount, in thousands, of total noninterest-bearing deposits based on participation in the TAG program. $3,000,000 $2,500,000 $2,000,000 $1,500,000 $1,000,000 $500,000 $- 2008:4 2009:1 2009:2 2009:3 2009:4 2010:1 2010:2 2010:3 2010:4 SIF Members, TAG Program Participants SIF Members, TAG Program Nonparticipants DIF Members, TAG Program Participants DIF Members, TAG Program Nonparticipants 23