It’s come to our attention that the precious metals investing community has been rendered a little “worried” by a sudden and sizable accumulation of inventory in the iShares physical Trust, the SLV for short. (H/T Kid Dynamite)
According to ZeroHedge, 572 tonnes were added to the trust in just one day. And while that does not represent a record for the fund, it is “the biggest one day addition of physical silver to SLV in ordinary course operations”. Or so, at least, ZeroHedge says (though we haven’t double checked the numbers ourselves at this point).
You can see historical accumulations (source not named) here.
So the question is, is this sinister or not? And more specifically, does it, as Zerohedge suggests, imply..:
“…in kind basket creation taking place. Whether this was due to arbitrage, or simply the need to create inventory we don’t know.”
Well, no, we certainly don’t think this is sinister. At least no more sinister than the usual issue we have with ETFs, which is the manner in which they overly smooth flow, mask price discovery and manufacture arbitrage spread for a handful of officially nominated Authorised Participants (the only market participants allowed to redeem or create at source. So a bit like Primary Dealers at the Fed).
Indeed, after years of watching this area, it is our conclusion that the simplest way to understand the impact that ETFs have on the market is to think of them as giant shock-absorbers and flow deflectors.
The funds themselves are in effect nothing more than giant depositories, where inventory (or flow) can be temporarily parked in the “encumbered” off market arena until its potential price impact becomes desirable on the public markets. The intermediaries who manage these flows make money from arbing the small spread opportunities that become available from subtle changes in the curve structure, their own cost of funding and the flow they have available to them, largely by diverting flows in and out of the ETFs versus the public market when it pays them to do so.
It’s nothing more than the old contango/backwardation trade applied to equities.
Except unlike with contango or backwardation trades, the off-market inventory (or short) ends up sending the wrong signal to passive investors, much more obviously.
Whereas it’s taken for granted in commodity markets that a large accumulation of commodity inventory is a bearish signal, the opposite is true of ETF markets. If assets under management rise exponentially this is usually interpreted as a bullish underlying market.
Yet in reality we’d argue it all comes down to yield.
When you think about it you realise that the vast majority of ETFs are negative yielding (due to the fact they don’t pay out dividends and charge a management fee). In a way they are the equity market equivalent of cash. They’re more liquid and more easily exchangeable precisely because they have more of a “hot potato” effect, since sitting in an ETF is an opportunity loss compared to sitting in the underlying. The only reason to hold an ETF is because it offers market access and exposure to the appreciation of the underlying.
When you see assets accumulating you’re actually seeing accumulations of negative yielding cash deposits which have a negative yield curve. The underlying, however, can have a very real positive yield if it’s hedged with futures into a contango curve.
ETF = price exposure risk, negative yield. Underlying = hedged position, yield.
This, we believe, is why ETF flows can appear so counterintuitive. When price action is bearish, ETF assets under management tend to rise because flows which would only depress the market further are redirected into the off-market ETF stash — a la contango trade — implicitly funded by passive investors and hedged via attractively-priced derivative positions offering “carry” to the intermediaries. When price action is bullish, redemptions tend to be the norm, as the exact opposite situation becomes true.
In short the money making flows are about making synthetic yields, not flat price returns.
Either way, APs don’t create or redeem unless there’s a spread to be made. And if an “in kind” creation request comes their way without the correct market conditions being in place to offer them carry, you can be sure an AP would charge the client in question an equally compensatory spread for the privilege of taking on the inventory — especially if it’s against the general flow of the market (since borrowing or buying the underlying would not be cheap enough to make it worthwhile).
A large creation of the sort described by Zerohedge in the SLV is much more likely in our opinion to be the result of an AP — after having offset all his flows — being left with an overhanging sell order, which he knows the market won’t take without adjusting downwards.
Rather than putting that order through the public market, the AP channels the inventory into the ETF depository via a creation. Since the market is in decline, the act of not dumping further into the market stabilises the physical price (for now). In other words, supply in the parallel ETF market is expanded in a bid to counteract the oversupply in the physical market.
This has an important impact on the derivative positions.
How does the AP make money? Just like any player who puts on a contango trade.
Imagine a global traffic system. During equilibrium all the traffic is perfectly distributed throughout all the cities in the country. Every city accommodates exactly the amount of cars it can handle and is completely traffic-free. All traffic tolls are priced the same.
But imagine what happens if a popular concert is held in New York and the city begins to attract more traffic than it can handle. In order to manage the flow, toll prices might have to adjust upwards.
Those who really don’t need to be in New York would leave to take advantage of lower toll costs elsewhere. Staying in New York is now an opportunity cost. Those who really need to be there would, however, be prepared to pay the higher prices.
And that (in a very over simplified way) is how the market works.
So what’s an ETF?
It’s basically a parallel road system, which runs on top of the existing infrastructure.
Every time the city gets overcrowded, gatekeepers come in and offer the opportunity to drive the parallel network for a premium price. And since it tracks the original network almost identically, that makes sense because it’s literally almost as good.
The only problem is that how much it costs to come in and out of the network is determined entirely by the gatekeepers, who take their cue for prices from the perception of traffic on the main network.
If the roads are chock full, the gatekeepers see an opportunity to charge ever higher access fees for the network. Furthermore, if people see that the traffic is getting lighter on the main network, the gatekeepers have the power to charge a fee to allow them back in.
Either way the gatekeepers profit. It’s just that sometimes they’re earning on access into the parallel system, sometimes they’re earning on access out.
All of which means, if you’re going against the traffic flow it can be extremely cost effective for you.
How does this all relate to SLV?
Well, in the current context of the market the question is why would anyone set themselves up on the parallel network when everyone is already leaving the city. (And remember, you have to pay to get out at this point.)
The tempting answer, of course, is to suggest that someone is taking the opposite view and anticipating that the price into the parallel network will soon go up because the flows will reverse. (Even though this is obviously a risky position to take, and one which yields nothing.)
The more cynical view, however, is that this is nothing more than a diversion tactic by the gatekeepers designed to fool people in the parallel network into staying a bit longer. And this is because the gatekeepers fear those in the parallel network and in the main network will soon leave the city, and take away their funding source.
Since they can’t control the outflow from the main network, all they can hope to do is achieve the maximum fee they can once the exit happens. And that may mean getting someone to flow into the parallel network at a small cost (or for no fee), to insure either that when the stampede happens it’s much bigger than anyone expected and thus justifies a higher exit fee, or the counter flow really does fool people into coming back.
Manufacturing arbitrage with ETFs - FT Alphaville