Short answer: Because they destroy the banking system.
For a longer answer let’s start with the following chart (via allhambrapartners),
These are Italian banks’ total holdings of Italian sovereign debt – EUR 725 billion, at last count.
The reason this is so huge (it looks similar all across the Euro periphery) is QE, or in the words of Jeffrey Snider,
In July 2012, the total assets reported by the combined Italian banking system were €3.54 trillion. In April 2016, the latest data available, total assets were €3.48 trillion. Worse for Draghi, of those assets, Italian banks have nearly doubled their holdings of government securities. In other words, the ECB through his promise did not actually encourage monetary expansion into credit expansion but rather front-running into government debt at the expense of total lending. The reason for that is obvious; with Draghi’s promise in their pocket, Italian banks as banks all over Europe found an outlet for “risk free” profit in buying government bonds at huge discounts. It was not an economic consideration. The mechanism for making good on Draghi’s promise was, in essence, a (deeply) negative factor on the further expected monetary transmission!
Remember: from a bank’s perspective buying the sovereign is attractive, as it is not required to hold any capital against it, and does not even have to look at clients or financial markets for funding. Thus you get crowding out of the relatively less attractive private sector loans – the opposite of “stimulus” (and inflation).
Now, you can imagine how bad the underlying situation in Italy must be when, despite this huge riskless profit opportunity, profitability over the last few years was terrible, as evidenced ever declining stock prices and rumors of new capital needs.
But it gets worse: by following their Keynesian logic, where the only thing that matters is interest rates (and “confidence”), the ECB has started to impose negative rates and stepped up QE purchases of corporate bonds thereby driving down yields across the curve further hoping to finally revive lending to the private sector – if it hasn’t worked, try more of it.
However, whereas QE and the associated reduction of interest rates at least subsidized individual banks, if not the “real” economy, courtesy of riskless profits, negative or minuscule interest rates on sovereign bonds have the potential to destroy the banking system.
Let’s assume Italian banks have earned a net interest margin (NIM) of 3 percent on the sovereign loan book above – say, they bought at an average yield of 4 percent and funded at an average yield of 1 percent. Further, assume they intend to hold the loans to maturity and booked them correspondingly, thereby shielding them form mark to market volatility. A NIM of 3 percent on 725 billion translates into a “riskless” annual operating income of about 22 billion Euros for the entire Italian banking system.
This is big, just compare to the mere 40 billion, as estimated by Prof. Zingales (see here), in capital that is supposedly needed to fix Italian banks (again) and already causing another political stir in the EU.
The problem: those bonds will have to be refinanced at much lower yields in the near future resulting in a collapse in (riskless) income.
Worse, the ECB is now actively competing for Italian (and all other European) banks’ best costumers by buying the bonds of large corporates, hence driving down the margins on the (shrinking) private sector loan book as well.
How on earth, I ask myself, are they going to compensate for the shrinking operating income?
The ECB apparently thinks that starving the banks of income will finally force them to expand their private loan books and thus “stimulate” the economy.
Fine, unfortunately this is almost certainly not going to work.
Here is why:
Assume that the NIM on the average SME or private client in Italy is currently 4 percent (in Austria it is much lower, but let’s be generous). Now, deduct risk costs of, say, 0.5 percent to arrive at a risk adjusted NIM of 3.5 percent. If you agree with my assumptions, it means that banks need to originate about 630 billion in private sector loans (mainly SME, as the margin on the large corporates will me much lower, courtesy of corporate bond QE) – about 1/3 of Italian GDP – to compensate for the missing revenue on the “riskless” government bonds.
This, now doubt, would be inflationary.
But there are technical problems. Not only do banks need to find enough solvent borrowers in what is a market with an aging population and huge structural issues, they also need capital for that – about 60 billion Euros, if you use a 10 percent capital requirement. After all, the low hanging fruit in the form of no-capital government bonds is no more.
(Note that this is a best case: if the Italian banking market would offer that attractive margins, there would almost certainly be heavy competition from German and other yield starved European banks cannibalizing these margins driving up the required loan volume.)
To sum it up: Italian banks need 40 billion in new capital just to deal with their NPLs and about 60 billion of new capital (under very generous assumptions) to be in a position to increase their loan book to offset the decline in operating income from lower rates on government bonds to maintain their meager profitability.
Not going to happen.
What is going to happen, instead, is that the earnings capacity of European banks will be more and more impaired with each passing day of low rates. When the business model is impaired, bank managements either respond by doing dumb things (remember those CDO² that ended up with German Landesbanks?), or by shrinking their business.
Shrinking bank balance sheets are deflationary by definition.
(Disclosure: no position in any bank)