Bank of England research suggests that QE of £375 billion has boosted the broad money supply by about £220 billion, or 15%, and that this in turn has lifted nominal GDP by nearly 6%. The analysis below, however, argues that the Bank has significantly overestimated QE “pass-through” – the monetary impact, relative to a no QE counterfactual, may have been as low as £80 billion, or 5%. Using the Bank’s sensitivities, this would imply a boost to nominal GDP of only 2%, split roughly equally between output and prices.
An excellent article* in the latest Bank of England Quarterly Bulletin (BEQB) considers the impact of QE1 (£200 billion between March 2009 and January 2010) and QE2 (£125 billion between October 2011 and May 2012) on the broad money supply. The researchers conclude that pass-through was similar for the two episodes, at 61% and 56% respectively, or an average 59%. Assuming that the latter figure also applied to QE3 (£50 billion between July 2012 and October 2012), the cumulative QE impact on broad money has been £222 billion.
The Bank estimates of offsets to the direct monetary impact of QE are plausible except in one important respect: they assume that commercial banks’ holdings of government securities would have been static if no QE had been conducted. Banks, however, were under strong regulatory pressure to boost their holdings of high-quality liquid assets following the financial crisis. QE, in effect, did the job for them by expanding their Bank of England reserves. In its absence, they would probably have bought many more Treasury bills and gilts. These purchases would have had a similar monetary impact to QE.
The blue line in the chart shows the ratio of banks’ high-quality sterling liquid assets – defined as physical cash, Bank of England reserves, Treasury bills and gilts – to their sterling assets; the red line separates out the contribution of T bills and gilts. Note that the rate of increase of the latter slowed sharply when the Bank embarked on QE1 and the contribution actually fell during QE2 and QE3. This supports the view that the reserves boost supplied by QE substituted for purchases of government securities that banks would otherwise have made in order to achieve their liquidity targets.
How much more would they have bought in the absence of QE? One approach is to base an estimate on the pace of purchases during the 20-month interval between QE1 and QE2, when reserves stopped increasing**. The red dashed line in the chart assumes that the rate of increase of the ratio of T bills / gilts to sterling assets during this interval*** would have applied for the whole period from March 2009, when QE1 started, to the present. This is arguably conservative, since it implies an aggregate liquidity ratio now of 7.9% compared with an actual 10.7%, assuming no change in the cash / reserves portion since February 2009.
This approach implies that, without QE, the ratio of T bills / gilts to sterling assets would now be 6.7% rather than 2.8%, consistent with additional bank buying of £144 billion. This sum needs to be subtracted from the £222 billion estimate of the broad money impact of QE derived from the Bank’s research. The suggestion is that £375 billion of QE delivered a monetary boost of only £78 billion – pass-through, in other words, was 21% rather than 59%.
The BEQB article claims that QE1 and QE2 lifted broad money by £192 billion and that this resulted in nominal GDP being about 5% higher than otherwise by mid-2012. The lower estimate here of a £78 billion impact of total QE (i.e. including QE3) suggests a nominal GDP boost of only about 2%, of which slightly less than half will reflect higher prices, based on the Bank’s research.
These results, if accepted, raise doubts about the wisdom of QE: the boost to output of little more than 1% needs to be set against the damage to the Bank’s credibility from pursuing the policy despite above-target inflation and inviting a charge of printing money to suppress government borrowing costs****. There is, however, one positive implication: unwinding QE may have a surprisingly small negative impact on the money supply and economy as banks step up their purchases of government securities to offset the drag of falling reserves on their liquidity ratios.
*Butt, Domit, McLeay and Thomas, “What can the money data tell us about the impact of QE?”, Bank of England Quarterly Bulletin, Vol. 52, No. 4.
**Reserves, in fact, fell between QE1 and QE2 as banks repaid Bank of England loans.
***Specifically, between the two dates highlighted by boxes on the red line.
****Assuming identical DMO issuance, the yield curve would probably have been modestly higher and steeper in the absence of QE because aggregate Bank of England / commercial bank purchases would have been lower and of shorter duration, owing to banks’ preference for shorter-dated securities.