Central bank equity buying: some evidence

Over the past few months there seems to have been a growing debate, here and elsewhere, as to the impact of central banks upon equity markets.   While few would dispute the notion that many monetary authorities would like to see equities rise, and indeed put in place macro policies that are intended to effect that outcome, the degree and influence of CBs' direct participation has been a source of disagreement.

In an effort to put to rest the tiresome trope that "central banks are in buying" seemingly every time that equities rally, Macro Man decided to do a little digging.  Fortunately, the ETF and J-REIT purchases by the BOJ are a matter of official policy, and thus there is an official dataset of the historical date and size of BOJ participation in equity markets.  Macro Man decided to parse the data to see what he could find.

If the conspiracists are correct, what should we expect to see?  In other words, what verifiable predictions can this "CBs directly drive equities higher" model of the world make?  Two obvious ones come to mind:

* We should see better index performance on buying days than non buying days

* We should see a larger volume on buying days than non buying days.

So what does the data tell us?

A couple of pretty strong patterns emerge.   The first is that yes, volume does tend to be higher on BOJ purchase days, by roughly 10%.   Macro Man did not run "value traded" numbers, but this strikes him as a volume boost that is likely to be greater than that accounted for by direct BOJ purchases.

The second is that the Nikkei fares much, much, much worse, on a consistent basis, on BOJ buying days than it does on non-buying days.  In fact, the average daily return on buying days is negative, both overall and every single year since the program was inaugurated in 2011.

Now, there could be a simple explanation for this: namely, that the BOJ has their monthly purchase bogey, and on a high frequency basis choose to time their buys on days when the market sells off out of the gate.  This seems pretty reasonable to Macro Man; after all, Japan was the birthplace of the PKO (price-keeping operation.)

However, whatever buying that the BOJ is doing, it is not sufficient to propel the market back to gains on the day they do it.  Now, adherents to the "CBs control all equity price action" theory might respond that "yeah, of course, they buy when the market's down, and the impact carries over to the next day."

OK fine, but that's contrary to the original claims, and also suggests that the original theory is wrong, to wit:  those days that the stock market rises are precisely those days when the CBs are not in buying.

You can't have it both ways, fellas, by claiming that equities are up because central banks are buying, and then say that they only buy on days when the market goes down.  Having it both ways isn't a theory, and it certainly isn't science; it's some sort of cultish financial religion.