onEconomy
By Jeff Huther
ABA Data Bank
The Federal Reserve’s weekly H.8 report provides aggregate data on assets and liabilities of commercial banks in the US. The data are disaggregated into three large categories: foreign, large domestic and small domestic banks. In recent months, this weekly report has received attention as the media and analysts have used it to try and track deposit flows between globally important banks and all others and more broadly, the health of the banking sector. The reality, however, is that these data, while helpful, have two important limitations: they emphasize immediacy over detail, and the size grouping, is somewhat arbitrary. “Large banks”, for purposes of the H.8 are defined as the largest 25 banks by domestic assets as reported on the most recent quarterly statements, which includes banks that are generally thought of as regional banks as well as the banks with global footprints, banks that focus on custodial services for large pension and insurance funds, and banks that specialize in consumer financing. This note covers some of the weaknesses and strengths of the H.8 report with examples and comparisons to more detailed data. While the H.8 is an important data source, it is not a proxy of bank health.
What the H.8 report can tell us
(Click to enlarge)
The large bank grouping in the weekly H.8 report combined with quarterly Call Report data can point to trends on a close-to-real-time basis if global and regional banks are expected to be responding to shocks in a similar way. Figure 1, for example, shows that the increase in large bank deposits (H.8 data) from the policy response to the pandemic in early 2020 was due to an increase in deposits for all large banks (individual bank Call Report data have been circled in blue). Aside from that single quarter though, deposit flows for at least a few banks have been in the opposite direction and varying magnitudes of the changes in the H.8 deposit data.
The Fed’s quantitative tightening, which began in June 2022, is likely to push deposits lower at banks of all sizes – quantitative tightening should look something like the reverse of the Fed’s most recent quantitative easing which flooded the banking system with deposits starting in March 2020. Recognizing when quantitative tightening outflows become significant may be an input into future monetary policy implementation. It can be important to know when aggregate changes are occurring more quickly than quarterly, and the H.8 data are better suited for this purpose.
Figure 2a. (Click to enlarge)
Figure 2b. (Click to enlarge)
The H.8 data on bank assets can also be used to show how banks coped with the massive inflow of deposits generated by the pandemic-related quantitative easing and hints at why the Fed’s tightening of interest rates since March 2022 has not led to a large contraction in bank lending. The following charts show that bank lending across major categories held steady as deposits grew. Banks maintained tight credit standards and put the initial deposit inflows into reserves (“cash assets” in Figures 2a and 2b) and, more gradually, invested in securities (largely Treasury securities and mortgage-backed securities). Banks have responded to quantitative tightening by trimming their holdings of securities rather than pulling back on lending, thus continuing to support economic growth.
Figure 3a. (Click to enlarge.)
The H.8 data can provide early warning signals of economic slowdowns. In Figures 2a and 2b, the relative stability of cash assets following the March 2023 bank turmoil-driven bump up (circled in red) suggests a cautious, but not a negative outlook by banks, a view that is reinforced by looking at the largest sub-categories of bank lending. Commercial real estate loans, for example, are considered a potential at-risk category given lower demand for some office and retail properties. Despite the headlines associated with urban commercial real estate, Figures 3a and 3b show that lenders at banks of all sizes have not pulled back from the sector, suggesting that as of summer 2023, they do not see abnormal risks for the types of properties they support.
Going forward, the H.8 data will be able to provide quick assessments of how declining deposits, from the pressures of quantitative tightening, will affect the assets side of bank balance sheets – will they build cash buffers in anticipation of additional outflow or maintain lending funded by borrowing? The answer will affect future economic activity as well as bank profitability.
What the H.8 report can’t tell us
Figure 4. (Click to enlarge.)
The large and small categorization in the H.8 partially or completely masks relevant balance sheet changes. As a result, care should be taken in drawing conclusions from comparisons between large and small banks. A recent example is the data on deposits following the collapse of Silicon Valley Bank. The H.8 data appear to corroborate the story that, at least for the week of March 15, depositors moved large balances from small banks to large ones (see Figure 4). Both large and small banks, however, lost deposits in the following week and then saw deposits grow slowly in the subsequent two weeks.
The H.8 data, however, do not distinguish between regional and global banks and are not detailed enough to determine the flow of funds around the system. In particular, flows from regional banks to the global banks do not change deposit levels at large or small banks in the H.8 since both the regional and the global banks are categorized as “large.”
Figure 5. (Click to enlarge.)
The H.8 aggregation leads to generalizations that miss important differences in trends among smaller banks. Lending as a share of current assets, for example, has grown particularly fast since the beginning of 2022 for banks with over $50 billion in assets that are not in the H.8 Large category. Loan growth at the smallest banks, while not notably different in direction, has lagged that of larger banks. (See Figure 5.) Simply looking at the H.8 data for banks not in the top 25, which are weighted towards the banks with the larger balance sheets, overstates changes for smaller institutions.
Figure 6. (Click to enlarge.)
The Call Reports can also provide additional insights that are not possible with the H.8 data. Figure 6 shows that deposits, as a share of assets, grew on average for banks of all sizes from the pandemic to 2022Q1 and then stabilized for a couple of quarters. Over the past year, however, banks not considered “large” in the H.8 have had divergent deposit growth with a notable difference between banks with less than $1 billion in assets and banks with more than $50 billion in assets. Deposits as a share of assets for the smallest banks have, on average, declined since peaking last year while deposits as a share of assets at banks with more than $50 billion in assets (but not among the top 25) have steadily increased. Notably, banks with assets between $1 billion and $50 billion have, on average, maintained steady deposit-to-asset shares over the past year.
Regardless of size, banks face similar deposit pressures in the coming year. Increased issuance of Treasury securities will continue to attract deposits and money market mutual funds, which do not benefit from any form of federal insurance, will continue to offer attractive yields relative to deposits. How these pressures play out will be evident in both the H.8 and Call Report data.
Jeff Huther is a VPfor banking and economic policy research at ABA.
Appendix: H.8 reporting details
The Federal Reserve has estimated that approximately 875 banks file weekly data on their balance sheet items. That estimate, however, appears dated (possibly from 2009) so the number of filers is likely to be smaller. The reporting form is a single page of data that correspond directly to the aggregated data provided publicly and are a small subset of the quarterly information available about individual banks. Formally, submission is voluntary but is likely to be close to 100% for larger banks. While relatively timely, the data are only available with a 9-day lag.
For policy discussions about domestic banks, the relevant H.8 data exclude the foreign banks since the distribution of foreign assets and associated liabilities is very different than that of domestic banks. Foreign banks disproportionally lend to commercial and industrial companies but not to retail customers, hold high levels of reserves, and participate heavily in financing markets. So, while foreign banks hold 13% of the reported assets, their share of bank activity ran from no participation in retail lending markets to over half of all borrowing in the fed funds market at the end of June 2023.
Depending on asset size, banks may not need to report at all or on a quarterly or annual basis. These exclusions, while not influential in determining the aggregate numbers, highlight the caution needed in interpreting how “small” bank balance sheets are evolving.
ABA Data Bank Balance sheet management
