Yale School of Management

Program on Financial Stability

Improving our understanding and management of systemic risk.

The FHLBs May Not be the Lenders-of-Next-to-Last Resort during the Coronavirus Crisis

March 24, 2020
: by Chase P. Ross

Access our live “financial-intervention” tracker, which is a single point-of-entry on the YPFS website to keep up with financial-crisis fighting interventions by central banks, fiscal authorities, and international organizations  here.

During the global financial crisis, the Federal Home Loan Banks (FHLBs) were the lenders of “next-to-last” resort, as banks and other FHLB members preferred to use their funding rather than the Federal Reserve’s discount window, the traditional lender of last resort (LOLR). Banks felt using the discount window might stigmatize their image with other market participants.

But it’s unclear whether the FHLBs will play that role this time. Market data suggest that banks are turning to the Fed—partly because the stigma of using the Fed is reduced, partly because of pricing.

The FHLB system is a government-sponsored enterprise operating with an implicit guarantee from the government. The FHLB system provides advances to its members—mostly depository institutions, but also insurance companies—to help finance housing-related assets. The system has a simple leverage multiple of 19.5 and has roughly $1 trillion in assets as of Q4 2019, of which $650 billion are advances.

In 2008, the FHLB system financed its LOLR activities by ramping up debt issuance. Auctioned discount notes grew from $120 billion pre-crisis to almost $300 billion in May 2008. Money-market mutual funds were important buyers of that debt. However, after the near-failure and government takeover of the other two prominent GSEs—Fannie and Freddie— money funds were less willing to buy FHLB debt. The FHLBs were “guilty by association.” The systems’ debt outstanding shrunk rapidly, and only eventually recovered after the crisis (Figure 1). 

Figure 1

In the aftermath of the Reserve Primary Fund breaking the buck in September 2008 and Treasury’s subsequent money fund guarantee, regulators sought to limit the systemic risk of money-market mutual funds. In 2016, the SEC implemented reforms that required funds to report floating net asset values (NAV) unless the fund imposed gates and fees or invested only in government securities. The gate structure allows the fund to temporarily prevent investors’ redemptions to cash in times of stress and would, in principle, limit a run from the money fund.

The effect of the reform has been a marked shift from prime funds, which invest primarily in commercial paper, toward government funds since government funds have a fixed $1 NAV without gates or fees. Money fund investors clearly prefer the fixed value and the option to run in bad times over the comparatively higher yield offered by prime funds.

In my recent working paper, I show that these post-crisis reforms made FHLBs new crucial safe asset producers. Government money funds can buy FHLB debt—as of late 2017, almost 40% of government money fund assets were FHLB debt. The FHLBs use the proceeds from the debt to make advances to banks, and banks prefer funding via FHLB advances (rather than issuing commercial paper) because post-crisis liquidity regulations are friendlier to FHLB advances.

During the rapidly unfolding coronavirus pandemic, preliminary and incomplete evidence—based on the public data available so far—suggests the FHLBs provided substantial liquidity in the initial stages of the recent funding market pressures. But it ceded that role when the Federal Reserve boosted its lending operations.

From February 26 to March 12, government money funds saw about $150 billion of inflows—a 5% increase in assets under management in just a week. They used a good chunk of these inflows to purchase FHLB debt. From January 1 to March 11, FHLB auctioned debt outstanding (with maturity less than one year, excluding overnight debt) increased about 12%, or $22 billion (Figure 2, Panel A). Government funds likely bought up the incremental issuance.

For that period, it appears the FHLBs were once again acting as a next-to-last LOLR. They used the proceeds from their increased debt issuance to help banks handle increased financing demands from the real economy. One estimate puts the magnitude of unfunded commitments of the largest banks to COVID-exposed industries via revolving credit facilities at $125 billion. Compare that to the $850 billion in cash and $2 trillion in securities at those same banks. Fed data shows that over the same time period commercial and industrial loans increased $16 billion, and loans to nondepository financial institutions also increased $16 billion.

However, beginning March 12, the Federal Reserve increased its provision of financing to the banking system via repurchase operations, and on March 16 it “encourage[d] banks to use [the] Federal Reserve discount window” after cutting the primary credit rate on March 15. Both actions had significant uptake: from March 4 to March 18, repo lending by the Fed increased from $200 billion to $450 billion, and discount window lending increased from $0 to $28 billion (Figure 2, Panels B and C).

Figure 2

Beginning March 17, short-term auctioned FHLB debt outstanding began falling quickly. FHLB debt spreads flipped from roughly -20 basis points to 30 basis points, and they’ve stopped regularly auctioning debt at the 4-week maturity.

Why? In part, the Fed’s actions siphoned some flows from the FHLBs. On March 16—after the Fed cut the discount window rate—the all-in cost to finance a highly rated MBS was 26 basis points with the discount window and 65 basis points with an FHLB advance. This is unusual; advances are typically cheaper than the discount window, as shown in (Figure 2, Panel D).

FHLBs cannot provide a large amount of liquidity now without increasing their debt issuances. In its latest quarterly filing, the system reported $5 billion in cash holdings, and it seems unlikely they’d pull back their repo lending ($50 billion in Q3 2019) to make space for more advances. The FHLBs could scale back their fed funds lending—$55 billion—to finance advances, but the action would have distributional effects. The FHLB system cannot earn interest on its account at the Federal Reserve; it instead lends in the fed funds market to foreign banks, which arbitrage the difference between interest on excess reserves and the fed funds rate. That is, if the FHLBs shifted from providing liquidity via fed funds to advances, they would be moving liquidity from foreign banks to domestic banks, despite evidence that foreign banks have stronger demand for dollar funding during the coronavirus crisis.

Are the FHLBs a systemic risk right now? Unlike Fannie and Freddie, they do not have risky investment portfolios, and their statutory super-senior lien means they are unlikely to face significant credit losses. The more worrisome problem is the amount of maturity transformation they’re doing—half of their advances are longer maturity than one year—and the average maturity of FHLB debt held by money funds is 40 days as of 2018.

The FHLBs provided a valuable lender of next-to-last resort channel in 2008—and although the Fed has worked to destigmatize the discount window in the past weeks, the FHLBs will remain important intermediators between banks and money-market mutual funds in the coming weeks. Keep an eye on the FHLBs.

See also:  The YPFS Live Financial Intervention Tracker.