In April, the U.S. Federal banking agencies issued a proposal to establish a new regulatory requirement for banks, the Net Stable Funding Ratio (NSFR). The regulation is purportedly intended to ensure that banks have a stable funding profile over a one-year horizon. In July, The Clearing House published a lengthy research note on the NSFR. The note provides a readable explanation of what is a complicated regulation, and identifies a series of conceptual and analytical flaws in the proposal. In this post we provide a simple demonstration of one way that the regulation is poorly and arbitrarily designed.
One component of the NSFR entails estimating the amount of “stable” funding a bank has over a one-year horizon – in other words, how sticky its deposits and other liabilities are. It does so by assigning to each category of bank liability an “available stable funding” (ASF) factor. Those factors are then multiplied by the amount of the corresponding liability and the results are added up to determine the bank’s total available stable funding. While the approach is reasonable, the various ASF factors are not consistent with each other.
The exhibit illustrates the funding situations of two banks. The pink bank is funded with $100 of non-transactional retail deposits. Under the NSFR, such deposits have an ASF factor of 90 percent, meaning that 10 percent of the deposits would be deemed to flow out of the bank over a one-year horizon (as explained in the TCH note, the conditions under which such an outflow would occur are not clearly specified by the regulation). So by 6 months, the bank would have $95 in funding left, and by 12 months, it would have only $90. The regulation concludes that the bank has available stable funding of $90.
The blue bank is funded with loans from other financial institutions (FIs). In this particular example,
1) a $6 overnight loan,
2) a $6 six-month loan, and
3) an $88 one-year loan.
The NSFR assumes that such loans are not rolled over. That is why, in the graph, liabilities of the blue bank step down to $94 when the overnight borrowing comes due and $88 when the six-month borrowing comes due.
Note that the pink bank has more stable funding than the blue bank every day of the year. Prior to the final set of revisions, the NSFR, which was then based on the bank’s funding situation at the end of the year, concluded that the pink bank had more stable funding than the blue bank ($90 versus $88).
However, in the second half of 2014, the NSFR was revised to give banks partial credit for certain liabilities that come due in the latter half of the year because the bank had at least a half year before having to potentially replace the funding.1 For one, the ASF on loans from FIs that matured between 6 and 12 months was raised from 0 to 50 percent. In this case, that would credit the blue bank an additional $3, raising its total available stable funding from $88 to $91.
While that makes some sense, the adjustment was not applied the ASF factors for other liability categories, which resulted in the ASF factors being inconsistent. In the example, the pink bank funded with deposits has a minimum of $95 in stable funding over the first half of the year, $5 above the year-end level of $90. If the bank got 50 percent credit for that additional $5, the NSFR would credit the bank with stable funding of $92.5, not $90, higher than the $91 in stable funding of the FI funded bank, as seems more reasonable.2
This example, and there are lots of similar examples, illustrates several things.
- The NSFR is an unreliable guide to a bank’s funding condition. A bank with strictly superior funding (the deposit-funded bank) would receive a lower score.
- The NSFR will establish incentives that will lead to unintended and undesirable consequences. The first bank would have an incentive to shift toward the second funding model to improve its NSFR even though that shift would lead to an inferior funding situation under the NSFR’s own assumptions.
- As banks act on those incentives, the NSFR will become an increasingly bad measure of bank condition.
With it being playoff season, a baseball analogy can help put it in simple terms. The NSFR’s treatment of ASF factors is akin to Major League Baseball deciding that teams should get triple credit for runs scored in the first three innings but only applied the change to teams whose names start with the first four letters of the alphabet. With that adjustment, the Dodgers won game 2 of the series with the Nationals and there would be no game 5 tonight.
The adjustment would result in game scores being a poor indication of which team played better. While the problem might be relatively minor to start, with only a few teams affected, the mismeasurement would worsen once teams began to change names to the Aardvarks, Badgers, Chickens, and Daredevils.
Many of the flaws in the design of the NSFR, including the flaws discussed in this note, were introduced in the final round of revisions, when rushed horse trading between the jurisdictions on the Basel Committee superseded sound design. I’m sure academics and industry stakeholders could help the banking agencies come up with a better design, but unfortunately, we’ve only gotten a chance to provide comment about aspects of the regulation that aren’t nailed down by the international agreement.
Once the banking agencies adopt the regulation, it will be very difficult to change. But, as demonstrated in the TCH research note released in July, as the U.S. financial situation becomes more normal in the future, if the NSFR is adopted, it will have an increasingly material impact, reducing banks’ willingness to make all types of loans, including those to small businesses and households.
Not only will the regulation have foreseeable unintended consequences, it is not needed: The large banks that would be subject to the NSFR are already required to conduct monthly stress tests of their liquidity situation across overnight, 30-day, 90-day, and one-year horizons. Moreover, it appears increasingly likely that banks in other jurisdictions will not be subject to key parts of the NSFR.
Thankfully, for the Nationals and fans of baseball everywhere the scoring is consistent and fair. Until the same can be said for the NSFR the U.S. banking agencies should refrain from adopting the proposal. The European Union just announced that they would not adopt the most recent round of capital regulations put out by Basel, demonstrating that the United States does not need to adopt a bad regulation simply because the Basel Committee says it has to.
1 A similar adjustment was made for assets that mature or become unencumbered within the one year horizon.
2 While this example highlights a correction to the design of the NSFR that would make the regulation a bit easier, similar adjustments to the NSFR’s assessment of the assets of the bank would make the regulation tougher, with the net effect uncertain.