Any debate on the impact of central banks’ shrinking balance sheets must start with some humility. This has never happened before and no one really knows what will happen or what the unintended consequences will be! But our job is to come up with a well-argued opinion. Here’s mine:
I think a good way to estimate the impact of any 'difficult-to-quantify' factor, such as QE, is to work backwards from what we know. For instance, how much of the 260% rally in the S&P 500 since its lows in March 2009 can be explained with well-understood, standard factors?
You can see how these add up in the chart below. Whatever is left is a good starting point for estimating the effect of QE.
The largest contribution to the bull market and also the easiest to quantify is earnings. US corporate earnings per share have grown by 130%, about half of the rise in share prices. Of course, this has come from various sources, but the biggest contributions have been operational: revenue growth and operating margin expansion. Share buybacks have contributed about 1.5% per annum by reducing the share count.
A natural push-back here is that earnings are inflated by low interest expenses, which in turn have been driven by QE. But a look at US corporate income statements shows that this has not been the case. Interest expenses have not fallen since QE has started. Rather, it seems that the effect from the decline in yields has been roughly offset by that from rising debt.
Multiple expansion – the increase in price-to-earnings ratios (PEs) – is technically responsible for the remainder of the equity bull market. And a re-rating is arguably where the effect of QE should show up most directly (interestingly we find little correlation between the size of central bank balance sheets and PEs). But of course not all PE expansion can be attributed to central banks’ asset purchases. It is quite normal for PEs to go up in a bull market; this has happened for many decades before the advent of QE. Going all the way back to 1946 the average bull market has seen PE expansion of 70%, so we use that figure in the chart below.
This leaves 60% of the 260% rally unaccounted for. Let’s take a conservative approach and attribute all of this to the decline in bond yields and its effect in lowering the discount rate for equity cashflows. This then leads us to the debate over how much of the decline in bond yields can be attributed to QE. There is a large range of estimates from various researchers, but they find that QE has depressed bond yields by 50-100bps and those estimates typically include the effect of policy guidance. That leaves us with 40-50% of bull market returns explained by non-QE bond yield declines (eg. lower growth and inflation) and the remaining 10-20% of the explained by central bank quantitative easing.
Pulling all of the above together, the S&P 500 is up 260% since March 2009, around 130% comes from earnings, 70% from normal bull market PE expansion, 40%-50% from a non-QE related decline in bond yields and about 10%-20% from QE. With the above mentioned humility in mind this analysis suggests the complete withdrawal of QE could take 4% to 8% off the S&P 500.