Perhaps the single most important lesson learned from the financial crisis is that liquidity transformation — that is, funding illiquid assets with short-term investments — is the tinder that feeds a financial conflagration. The crisis saw that effect in non-banks (especially money market funds and securities firms like Lehman Brothers), banks, and government-sponsored entities.

In response, post-crisis bank regulation has significantly (and intentionally) reduced the liquidity transformation performed by banks. Banks, particularly large ones, are subject to strict regulation of the relative liquidity and maturities of their assets and liabilities. For their part, prime money market mutual funds are now required to have a floating NAV (Net Asset Value), to ensure that they and their investors internalize any funding mismatch; the results have been dramatic, as investors have reacted by moving hundreds of billions of dollars of investments elsewhere.

Unfortunately, though, the same rules that have caused banks and money market funds to decrease their liquidity transformation have also caused a corresponding increase in the liquidity risk borne by the Federal Home Loan Bank System, a government-sponsored enterprise. As documented in three recent Federal Reserve blog posts (here, here and here), and as discussed further below, FHLBs have responded by growing rapidly in recent years, funding longer-term loans with shorter-term borrowings. The irony here is profound: liquidity transformation at banks has been limited in large part to prevent them from having to rely on discount window lending under stress, because the discount window has come to be viewed by some as a bailout; as a result, though, liquidity transformation has shifted to the Federal Home Loan Banks, GSEs implicitly backed by the taxpayer, every day. This seems like a bad trade.

Background

There's money to be made by funding illiquid assets with money-like liabilities. Individuals and businesses strongly desire investments that they can draw at a moment's notice, so they will accept a low return on them. And borrowers prefer loans with long terms that won't be called unexpectedly, so they will pay a higher rate for them. Matching those investors and borrowers and managing the resultant interest rate and maturity risks has throughout our history been the core business of banking. To reduce the attendant run risk, the government has established deposit insurance and a lender of last resort (i.e., the discount window). As an unfortunate consequence, short-term creditors and depositors have a reduced incentive to price for risk when lending to banks, resulting in moral hazard. It is largely to control that moral hazard that banks are regulated.

Liquidity regulation makes balance sheet liquidity more valuable to banks. Banks will accept lower returns on highly liquid assets and pay more for stable funding, moreover; they will demand higher return on illiquid assets. What Harvard Professor and Former Fed governor Jeremy Stein calls "the money premium" will go up.

Of course, to the extent that non-banks perform a similar maturity transformation role, they run the same risks. If one prevents banks from engaging in liquidity transformation, others have a profit incentive to run risks that banks no longer can.

History

While many think of the global financial crisis as a residential real estate bubble popping, its darkest, most systemic moments were ultimately a run on the shadow banking system. Deposit-like investment at prime money funds funded illiquid, risky subprime loans through a chain of intermediation passing through private mortgage-backed securities (MBS) backed by the mortgages, asset-backed securities backed by the MBS, and asset-backed commercial paper backed by the asset-backed securities that was bought by the prime money funds. Prime money funds were able to perpetuate the illusion of offering risk-free deposits by always fixing the value of their investments at par regardless of fluctuations in the value of the underlying assets. When one large money fund with substantial investments in Lehman paper "broke the buck" — that is, had to value its shares below par — hundreds of billions of dollars were pulled out of all similar funds, and the real crisis began.

In reaction — although it was a delayed reaction — the SEC has forced prime money funds to abandon their fixed NAV ("net asset value") in favor of one that "floats" (i.e., changes as the value of the fund's portfolio changes). A floating NAV removes the illusion that investment in prime money funds are deposits, forcing the investor to recognize that the value behind the shares fluctuates.

In parallel, standalone broker-dealers obviously fared disastrously in the global financial crisis, as their reliance on short-term wholesale funding of longer-term assets represented another form of unstable maturity transformation. Post-crisis, all major broker-dealers are now affiliated with banks, and subject to the same liquidity regulation as banks — most notably, the Liquidity Coverage Ratio (LCR) and the Federal Reserve's Comprehensive Liquidity Assessment and Review (CLAR). (Short-term-wholesale funding is also discouraged by the U.S. GSIB surcharge.)

As Night Follows Day

Liquidity transformation abhors a vacuum, so what has been gained by money fund reform and bank liquidity regulation appears to have been largely lost as liquidity transformation has shifted to the Federal Home Loan Banks. Liquidity regulation encourages banks to borrow at longer terms and invest in high quality liquid assets, defined by liquidity rules to include (subject to some limitations) GSE debt — Fannie and Freddie, but less noticed, FHLBs. Money fund reform has resulted in a massive shift out of prime money funds, whose assets were primarily loans to banks, to government money funds, which can only hold short-term government and agency debt (including FHLB debt) or reverse repos collateralized by such debt. Meanwhile, under the LCR, a bank's liquid asset requirement is based on a projection of net cash outflows under stress, and advances from FHLBs mostly mature outside the LCR's 30-day window and therefore do not count as outflows. Those that do come due are assumed to rollover at a 75 percent rate when they mature, while loans from other private entities are assumed to not roll over at all.

In sum, the Federal Home Loan Banks have accommodated the resulting shifts in demand and supply by lending at term to banks funded with short-term securities sold to banks and money funds, expanding advances to banks by 80 percent over the past five years and incurring a substantial liquidity and maturity mismatch in their portfolios. As noted in the recent Fed blog posts, while FHLBs have not noticeably shortened the maturity of their assets over the past decade, the maturity of their liabilities has shortened materially, with the share of debt maturing in less than one year rising from 40 percent before the crisis to nearly 80 percent currently.

FHLBs have been able to assume this role because even though their debt caveats that it is not guaranteed by the U.S. government, it is perceived by the market, and therefore trades, as being guaranteed. A recent American Banker column states, without irony:

How much rollover risk is borne by the home loan banks — and ultimately by the taxpayers who backstop them — as a result of the increased maturity mismatch on their balance sheets is unclear at this stage. The odds that the FHLBs will face a liquidity shock are quite low. It would require a loss of confidence among the investors who buy their obligations, and this is unlikely except in the extreme case that calls FHLBs' implicit public guarantee called into question.

Thus, it is precisely because the FHLBs are perceived as too big to fail that they have been able to conduct the liquidity transformation squeezed out of banks in an effort to prevent banks from being "too big to fail."¹

This state of affairs appears ripe for FSOC review. Certainly, more careful supervision of the FHLBs appears warranted. Furthermore, the Fed could take actions to increase the supply of, or reduce unnecessary demand for, the high quality liquid assets (HQLA) that satisfy liquidity requirements. In particular, we have recently proposed that the Fed could count required reserves — the percentage of customer deposits that banks must in turn deposit at the Fed — as HQLA. The Fed would be able to do so if it changed the regulation governing required reserves to stipulate that banks can use them to help meet customer outflows. Relatedly, the Fed could take steps to ensure that banks are willing and able to use their HQLA when there is a stress event, reducing demand by banks for HQLA over and above their LCR requirements. Lastly, and in the longer term, as described in a TCH working paper, the Fed could consider recognizing the liquidity value of collateral pledged to the discount window when evaluating each banks' LCR. Such recognition can be accomplished in compliance with the LCR by allowing banks to borrow against the collateral but hold the proceeds of the loans at the Fed until needed. Such an arrangement would also have the advantage of charging banks for the implicit liquidity support they currently receive for free.

Disclaimer: The views expressed in this post are those of the author(s) and do not necessarily reflect the position of The Clearing House or its membership, and are not intended to be, and should not be construed as, legal advice of any kind.

¹ As one of the Fed blog posts cited above puts it, "Given the short tenors of FHLB debt and the fact that the debt is primarily held by a wide range of cash investors, mainly government money market funds, the lynchpin for the more benign scenario seems to be continued confidence among money market participants in the implicit government guarantee for FHLB debt."