AC: I’d like to start by asking a little about your role and what you’re focused on.
AM: Anything we do here at Vanguard we want to make sure we have an emphasis on minimising the cost to our shareholders. And we take that approach regardless of asset class– whether it is equities or fixed income products or FX, we look at our orders holistically and make sure we’re managing the risk and also the cost of execution appropriately.
AC: How do you go about doing that?
AM: The most important thing we’ve done to manage the FX markets is open a trading desk in each of our trading locations—Melbourne Australia, London and here in Pennsylvania with full-time FX traders in each. With our global team in place, we look at the orders coming in on an aggregate basis and at the benchmark and based on those two factors we decide which time zone is the most appropriate to trade those orders in. We have the ability to trade at optimal liquidity windows in the market.
AC: You must work with a number of banks and brokers.
AM: Yes, we do. It’s imperative that we have as many tools at our disposal so that we can make sure we can see a trade and decide the best method of execution. For example, it could be a request for a stream to a number of banks, or it could be a request going out to an individual bank that we might feel is active in that currency.
Lastly, it could be a request going out through an agency model and using an FX algorithm to source liquidity from many different venues. By using this agency model, we take a large order and cut it into smaller pieces in order to reduce our execution costs.
AC: Are there any important or dominant trends in the agency model?
AM: Just to take a step back for a moment, I think one of the advantages Vanguard has is that we manage assets all over the globe and we hedge many products to multiple currencies. We have Australian products that hedge back to the Australian dollar, we have UK products that hedge back to the pound, we have euro products that hedge back to the euro and we have dollar products that hedge back to the dollar. So as a global team we seek to bring all of those different flows under one centralised desk. From there, we take a look at what exposure we have on a net global basis and our orders so that we’re only going to the market with as little a position as possible in order to minimise our impact and footprint.
Once we see what our net exposure is, we decide the best way to execute and in what time zone given the different parameters of the particular order. If you can imagine a situation where we have many different funds trading a euro position, we can bring all those positions together and net as much as possible so that we do the right thing for our shareholders by not going to the market with the entire position.
Based on our structure, we have the ability to trade pieces of an execution algo—for example, here in the US we do so in order to take advantage of liquidity. If we don’t find sufficient liquidity or if spreads are higher than they normally are, we can pass the order to our Australian desk once liquidity starts to wind down in the US. As Singapore, Tokyo and Hong Kong come online, that Australian desk is now ready to take advantage of any liquidity and tighter spreads that are available in those markets.
AC: You spoke about netting as much as possible. Could you describe how this works?
AM: Our portfolio managers all over the globe send in orders to our centralised FX desk with their benchmarks and their sensitivity to that benchmark. Then, our traders aggregate all those positions together taking the parameters outlined by the portfolio managers into consideration. Once we have those orders all aggregated net, we decide how exactly we’re going to implement the execution strategies on those orders. Netting is a huge component of what we do here as it provides a significant savings to our shareholders.
AC: Is this a growing trend for larger funds?
AM: I cannot speak to what the rest of the industry is doing, but as mentioned, it’s a huge cost saving for our clients and the reason why we have a centralised FX desk to help offset counterflow.
AC: One of the issues we’re looking at in this edition is the classic buy versus build question? What’s Vanguard’s view on that, given that your size means you have scope to do either?
AM: Most often, we utilise bank-generated algos. We believe these algos give us the best access to the different venues that exist. We work very closely with our bank partners to make sure the algorithms are working in our best interests. We use multiple TCA methods to evaluate our executions.
AC: What would those conversations lead to?
AM: We look at a number of factors in assessing the best venues. Last look, for example, is a factor we consider or we might question the venue if there looks to be a large amount of mean reversion happening on that particular venue.
AC: How do you go about adding that venue back?
AM: It’s just an ongoing dialogue with our algo partners. Obviously, there are a lot of new venues that pop up all the time so we really rely on our partners to help us vet these venues and then we evaluate execution reports.
AC: How do you choose which providers to use?
AM: We look at the algo providers as a total package of their offering. A great deal of it has to do with what venues they route to, how their internal liquidity looks, and what their transaction cost analysis looks like. As these algorithms advance in the FX space constantly, they start to appear very similar, almost like a commodity. As such, assessing the specific details of what the total package provides is important.
AC: You mentioned TCA. A number of buy-side participants have talked about how there is no universal standard in the FX world. Does that present an issue to evaluate one set of algos versus another
AM: It can, but I think this goes back to the fact that end users need to set proper expectations with the bank they are working with. We ensure that our algo providers understand the parameters we want to evaluate our executions on. We have strict criteria that we want to look for when we get our TCA reports.
AC: There seems to have been an uptick in volatility in recent months. What does that mean for you?
AM: Volatility has been on the rise from where it was the last couple of years, at historically low levels. That being said, I think that does play into the hands of using algorithms for executing orders. An increase in FX bid ask spreads can increase the savings we realize by using algorithms. Trading smaller clips of a large order, being passive and providing liquidity to the market as opposed to paying the full spread can result in large savings.
Ultimately it is going to reduce your overall transaction cost if you can deal with the opportunity cost of that price moving against you.
AC: Are there any indicators you focus on in terms of how you deal with market impact and slippage?
AM: One of the things we absolutely look at is an estimated risk transfer price at the beginning of our order. If we can beat that risk transfer price more often than not, that takes into account opportunity cost and impact cost. And that’s what we hold ourselves accountable for on this desk. We’ve saved significant amounts to that estimated price. Additionally, we want to look at what the market mid was at the time we started executing. This is a really difficult price to achieve because it assumes a zero cost of trading.
AC: You’re essentially becoming a price maker rather than a price taker.
AM: Exactly. Which we’ve seen in our performance. Just absolutely great numbers, which boils down to significant savings for our shareholders.
AC: Can you talk about the trends in the share of risk transfer versus algo executed trades?
AM: I would say at this point we are predominantly algo executed. There are situations where we’ll use a risk transfer price, typically on our smaller trades, where it doesn’t necessarily make sense to use an algo due to the fact that there is cost to using algos and spreads are relatively tight in a lot in small size. But where we see there’s an advantage to using an algo is the large block trade and the more illiquid currencies.
AC: Are you noticing any interesting developments with providers?
AM: One algo feature I expect to see a lot more of is the matching utilities that many of the different banks are trying to come up with. Essentially, matching off flow between firms. They are a bit like dark pools in the equity market algos. These FX matching engines are designed to match up buyers and sellers at mid-pricing without any risk.
AC: For large block orders, the equities market buy side has Liquidnet. Are there any similar development in FX?
AM: I think that’s probably the next stage, especially as there’s a lot of flow centered around different benchmarks. A lot of flow is coming together at exact points in time and is really ideal for matching utilities. What’s unique about the FX market is that a lot of the flow gravitates toward benchmark times, which actually makes it easier to find offsetting flow.
The market structure for FX will lend itself well to a matching utility. One of the issues with the FX market is that there seems to be an arms race to see who can come up with a matching utility first. You really don’t need nine matching engines, you really need one predominant matching engine so that all the flow gravitates to that one point to maximise your ability to match and offset against someone opposite you.
AC: On a personal note, what attracted you to the FX market?
AM: My background is actually in equities. I was an equity portfolio manager and an equity trader. I primarily used algorithms to execute our equity orders. And when we formed the FX desk, I moved over to run this desk and we use a lot of similar approaches that I used in the equities space and applied them to the FX market. As the FX market has gradually shifted more toward the agency model, it’s actually been a great fit for me.
AC: Finally, were there any things that surprised you in adapting to FX?
AM: Not really. I traded FX over the last 10 years. So I witnessed the gradual transformation of the FX market as it shifted towards the agency model.