The Federal Home Loan Bank (FHLBank) System extends secured loans (i.e., advances) to its member institutions to provide liquidity in the housing finance ecosystem. To understand how much of this borrowed liquidity translates into more home lending, researchers at the Urban Institute and the Government Accountability Office (GAO) conducted separate analyses in 2025.
Although the studies covered different time periods and used different methods, both reached a similar conclusion: An increase in advances is associated with an increase in both overall lending and residential real estate lending.
In this article, we compare the two studies, focusing on residential real estate lending, mostly home mortgage lending, which connects most directly to the FHLBanks’ mission of providing liquidity to the housing finance system.
Differences in data, methods, and results
The two studies aren’t directly comparable because they explore different aspects of the same idea. Our study follows the same bank over time, asking whether its lending increases in quarters when its advances rise. GAO’s study mostly compares multiple banks in the same quarter, asking whether banks with more advances outstanding also have more loans outstanding.
Here’s a simplified example. Say Bank A had $50 million in advances last quarter and $60 million this quarter. Our approach asks, How much did Bank A’s residential lending change? It’s a before-and-after comparison of the same bank.
GAO’s approach asks, Does Bank A, which has $60 million in advances outstanding, also have more residential loans outstanding than Bank B, a similar bank with $50 million in advances outstanding? This approach compares two banks at one point in time, not one bank before and after.
Both approaches are reasonable ways to study the relationship between advances and lending, but they answer different questions. Their results therefore describe related but distinct relationships.
Sources: Jung Hyun Choi, Jun Zhu, Laurie Goodman, John Walsh, Katie Visalli, and Bryson Berry, The Value of the FHLBank System to Promote Housing and Community Development Lending (Urban Institute, 2026); and GAO, Federal Home Loan Banks: Role During Financial Stress and Members’ Borrowing Trends and Outcomes (GAO, 2025).
Note: FHLBanks = Federal Home Loan Banks; GAO = Government Accountability Office.
The studies also differ in terms of result type. Our estimate is in dollars: $100 more in advances is linked to about $18 more in lending. GAO’s first two estimates are percentages, which cannot be compared directly with our dollar-based estimate. GAO’s third specification, however, can be expressed approximately in dollars.
GAO found that a bank whose advances are 1 percentage point higher as a share of assets has about 0.9 percent more residential loans. Between 2015 and 2024, single-family loans outstanding averaged about 20 percent of FHLBank members’ total assets. All residential loans outstanding, including multifamily loans, averaged about 32 percent of FHLBank members’ total assets.
Applying 0.9 percent to those balances implies that $100 in advances is linked to $18 to $29 more in residential lending. Our within-bank estimate is at the low end of that range.
In addition to the baseline model, our report also tested an instrumental variable model for the 2010–24 period to address the possibility that advances and lending are jointly influenced by an unobserved factor that is not controlled for in the regression model. This approach produced a substantially larger estimate, consistent with the assumption that the baseline $18 figure is conservative rather than overstated.
Despite the differences between our approach and GAO’s, both studies find a statistically significant, positive relationship between advances and residential lending, with resulting dollar estimates of a similar magnitude. Reaching the same basic conclusion through different approaches strengthens the overall evidence.
Accounting for differences in bank size
Whether the relationship holds across various bank sizes matters because large banks account for most advance dollars, while small and midsize banks make up most of the FHLBanks’ membership and have fewer alternative funding sources.
In its analysis, GAO separates banks with more than $10 billion in assets from smaller banks. It finds a larger relationship between advances and lending among banks with more assets, but the result is not statistically significant, meaning the pattern could be attributable to chance. Large banks account for less than 3 percent of GAO’s observations, making it harder to detect a reliable pattern. This evidence should be viewed as suggestive rather than conclusive.
We accounted for differences in bank size by taking changes in advances and lending as a share of each bank’s total assets, a common approach in finance research. We then divided our sample into the same groups and found $37 per $100 of advances for large banks versus $17 for small and midsize banks, with both estimates statistically significant. As in the GAO study, however, the large-bank sample is small, so this result should be interpreted with caution.
Quantifying the aggregate lending increase from advance borrowing
Because our estimate produces a dollar-for-dollar relationship, we can quantify the aggregate residential lending associated with increases in advances.
Between 2002 and 2024, we find that member banks averaged advance balance increases of $198 billion a year. Applying our baseline estimate, this translates to about $35 billion a year in residential lending associated with increases in advances, roughly $811 billion over the full period.
We believe this estimate is conservative for several reasons:
- Bank data do not capture the total amount drawn and repaid during the quarter. The advance balance increases are the sum of positive quarterly increases across member banks, which are calculated at the end of each quarter. A bank that started with $100 million in advances, borrowed another $80 million, and repaid $50 million before quarter-end would show an increase of just $30 million, even though it obtained $80 million in new funding. Advances drawn and fully repaid within the quarter do not appear in our calculation, so it likely understates total borrowing activity.
- The relationship between advances and lending is stronger when advances go up. We found that a $100 increase in advances is linked to about $26 more in lending, while a $100 decrease is linked to only $6 less in lending. Our baseline $18 estimate averages both directions. But we want to capture the effect of borrowing itself, so blending in the weaker decline effect makes $18 a conservative estimate. The $18 figure is also lower than the estimate from the instrumental variable approach, which is used to isolate the causal impact of advances on lending.
- Our estimate captures only the direct increase in lending associated with advances. Additional mortgage or construction lending may generate broader economic activity through home purchases, construction work, and related spending, but these additional effects are not included in our estimate.
This exercise measures lending statistically associated with advances, but it does not tell us whether that lending would have occurred without the FHLBank System, as we cannot observe how banks would have borrowed and lent in that alternate world. We therefore rely on the conservative $18 estimate, meaning the cumulative $811 billion figure likely understates the aggregate value associated with advances.
Because GAO’s primary models examine the relationship between lending and advance balances in percentage terms, rather than dollar changes, their coefficients do not map onto the quarterly increases summed here.
Areas for future research
Despite examining different periods and using different methods, our study and GAO’s lead to the same conclusion: FHLBank advances are positively associated with residential real estate lending. Together, the two studies fill a gap in the empirical evidence. Policymakers, regulators, and the FHLBanks should consider this evidence when assessing the role and value of advances in supporting the FHLBank System’s housing finance mission.
Beyond providing advances, the FHLBanks directly support affordable housing and community development through targeted mission programs. In separate research, we examined the reach and economic contributions of these programs over the past decade. Given the nation’s critical housing needs, the FHLBanks should continue strengthening and expanding these programs to address those needs more effectively.
We have also conducted an analysis of credit unions and found a similar relationship between advances and lending, but neither study examined insurance companies. Insurance companies have been FHLBank members since 1932 and have grown significantly within the system. Because insurers do not file the call reports used in our study and in GAO’s, analyzing them would require different data and outcomes that reflect their roles as portfolio lenders and investors in mortgage-backed securities. We leave this question for future research.
