There's a common idea - for example here, here and to a lesser extent here - that quantitative easing involves the Bank of England giving money to banks.

This is not quite right. QE does not hand cash only to banks but rather to anyone who wants to sell gilts to the Bank.

In fact, banks have not greatly participated in QE. With considerable cussing, I've got the data out of the Bank*. These show that during the first round of QE, from March 2009 to January 2010, the Bank bought ''198.3bn of gilts. This more than offset net issuance of ''177.4bn by the government. There was, therefore, a net ''20.9bn money injection - underfunding, to use the 1980s term.

But it was not banks who got this cash, at least not directly. They actually bought ''6.5bn of gilts. It was the non-bank private sector - institutional investors mainly - who sold gilts and got the Bank's money. They sold ''33.8bn of gilts. (I'm ignoring foreigners here, so the numbers don't add up).

For the latest phase of QE, we have data for the fourth quarter. These show that the Bank bought ''51bn of gilts, and that banks sold ''10.5bn whilst the non-bank private sector sold ''29.4bn - almost three times as much.

These numbers show that QE does not mainly work (insofar as it does) through banks. Instead it bypasses them. The Bank hopes that institutional investors will use some of their cash to buy corporate bonds and equities, which in turn will reduce their yields and so encourage firms to borrow in capital markets. Bank lending has little to do with it.

This policy did have some success. Non-financial firms' capital issues did pick up in 2009, partly thanks to those lower borrowing costs.

This is not to say that QE will be very powerful. It won't be. Nor is it to deny that QE is a welfare state for bankers. It is. Insofar as it raises price of corporate bonds, it increases the value of some banks' assets. And insofar as it encourages more issuance, it improves investment banks' corporate finance business.

My point is simply that we should not judge the efficacy of QE by what happens to bank lending.

* It's in tables A3.2, B1.4 and D2.3.1 of Bankstats. Code identifiers are: B4CA for bank purchases (VQIF and TBMH for bank and building society buying respectively before 2010) and VTMB for the other domestic private sector.