Earlier this week, the president of the Federal Reserve Bank of Minneapolis, Neel Kashkari, took strong exception to the conclusions of a new FedFin paper in which we found that capital regulation may well have a direct and adverse impact on credit availability. Calling our approach “nonsense,” Mr. Kashkari argued forcefully that the biggest banks could come up with the multiples of current capital he craves with nary a dimple on the placid sea of credit availability. Nonsense to this and then some, say I, and here’s why.

First, Mr. Kashkari uses in his defense the same misleading statistic Janet Yellen deployed last month in defense of the FRB’s capital rules. Under strong criticism from Republicans, Ms. Yellen pulled out one finding from an NFIB study which found that only four percent of small businesses think there is a credit-availability challenge. True, but one cannot then draw from this the conclusion she did that 96 percent of small businesses are comfy with their credit – our paper goes through the data in detail and also points to a finding in the same survey showing that more small businesses think credit is hard to get than are satisfied with it. So, on point one, Mr. Kashkari is simply wrong unless he has other sources to muster in his defense (and we’ve seen none that do).

Mr. Kashkari also says that borrowing costs are at a “near record low.” I’ve no clue what that means, but I would guess he is referring to low funding costs. Given ultra-low rates, this aspect of borrowing costs is indeed low, but so what? The money’s not moving for loans with low return on capital. Capital is of course another borrowing cost and it’s anything but at a record low. Not only are capital requirements for banks at unprecedented highs, but the cost of capital – supposed to go down as capital goes up – is also stubbornly high.

On to point three: Mr. Kashkari also asserts that credit data show huge amounts of loans sloshing around. First, this isn’t right when one compares credit to GDP, at least based on the BIS data we point to in our paper. Secondly and far more importantly, it doesn’t tell you anything about the impact of capital regulation on credit availability even if all of the gross data on all the U.S. lending are solid. This is for two reasons detailed in our paper: 1) Non-banks make a lot of loans, in large part because capital pushes out banks – thus, lending can stay the same or even go up as non-banks expand but the impact of capital rules on banks is nonetheless a critical consideration; and 2) gross credit data don’t tell you diddly about the types of credit critical for economic growth and full employment.

To look at the impact of capital on the ability of banks to make loans, it’s essential to look not at gross data, but rather at how many loans banks can make based on their balance sheets. To do this, we looked at the FDIC data on insured depositories to calculate (the FDIC doesn’t) what percentage of loans are on bank balance sheets over time compared to other assets. Using FDIC data on loans and leases (which lump into them unearned income) and overall banking assets, we find that loans were 61 percent of bank assets in 2006 but have fallen to 55.5 percent in 2016, a nine percent drop. Going deeper than this gross data for 2016 shows an even more interesting change: the smallest banks (less than $100 million) are right about where the industry was a decade ago (holding 59 percent of assets as loans), with this percentage rising to 70 percent for banks between $1 billion and $10 billion (the class of banks exempt from the liquidity rules and others with significant credit impact also addressed in our recent work). Then the number drops precipitously with clear correlation to regulatory requirements – the 2016 percentage for loans hits 48 percent for banks with assets over $250 billion.

FDIC data do not permit one to break out banks over the $50 billion threshold at which capital rules begin to bite or at the GSIB level where the costly surcharge kicks in. However, the significant differences in lending based on size and the correlation between this and the increasing rigor of the capital rules at the least suggest that other factors – e.g., market and technological changes that are largely uniform across the industry – play less of a role in diminishing bank lending than the new capital rules. Causation is a lot more complex than correlation, but these data are at the least suggestive and show significant credit challenges due to the huge role giant banks play in the U.S. lending market.

I’ll give Mr. Kashkari one point: the job of CEO is, as he says, to “boost stock prices and dividends, not protect taxpayers.” Well, yes – CEOs actually run private companies, not government agencies.

As I noted last week, banks that are forced to hold uneconomic amounts of capital may be safe, but they’re dead ducks from a long-term franchise value perspective. Investors are no more altruistic than Mr. Kashkari’s CEOs – they put their money into banks to get it back with a market rate of return. If they don’t, they find another place to invest and the bank from which they withdraw suffers a loss in market capitalization and a slow slide thereafter into oblivion.

We must thus decide if we want loans to come from regulated banks or if we are content to regulate banks so rigorously that credit availability depends on unregulated companies. We can’t have a bank-dependent system of financial intermediation, capital that makes banks both less profitable and (hopefully) bullet-proof, and investors that keep their money in these same low-return banks. Wanting it won’t make it so – investors know their way to the door.