Thinking

Why is my VWAP Off?

February 26, 2016 Ofir Gefen
Volume Weighted Average Price (VWAP) is a popular benchmark in Asia Pacific and in many cases traders will use a VWAP trading strategy when aiming to achieve a VWAP price outcome. So why does this strategy often miss the benchmark?

In this research we use a quantitative analytical approach to isolate the components that lead to this slippage against the benchmark and propose ways to improve trading outcome with the goal of reducing tracking error from VWAP.

KEY POINTS

  • Only about 25% of Asian stocks display highly stable volume profiles.
  • For the other 75%+ of Asian stocks with moderate or unstable volume profiles, sticking rigidly to a volume profile increases the chances of a price outcome that is “off” the desired VWAP benchmark.
  • In particular, adherence to a strict volume profile can drive an active trading strategy to cross the spread more often to “keep up”, causing greater price slippage.
  • A smarter VWAP trading strategy therefore dynamically balances price outcome with volume schedule in a more flexible way. Typically this will take more passive liquidity to help minimise slippage against a VWAP benchmark.
  • Some stocks with highly unstable volume profiles will not be well suited for a VWAP trading strategy at all – alternatives should be considered.

THE ANALYSIS

When examining the slippage of the traded actual price (actual) to the market/ benchmark price (VWAP), we can break it down to two components:

  • Deviation of our actual volume profile from today’s market volume profile (Market Price * Market Volume) – (Market Price * Executed Volume)
  • Deviation of our actual price achieved (per bin or per day) from the VWAP benchmark price for that bin/day (Market Price * Executed Volume) – (Execution Price * Execution Volume)

Pricing deviation stems to some extent from pricing volatility as well as the trader’s effort to “keep to a profile”. (Here it is worth noting that volume is not a deterministic variable in itself, and thus has its own volatility and distribution.)

Tracking VWAP is therefore exposed to volatility in both price and volume and particularly the correlation between these two random variables, i.e., that volume may spike at some price levels.

PUTTING THE THEORY INTO PRACTICE

Using this analysis we can break down any VWAP trade to its basic components and determine which is the critical contributor in the overall slippage of a trade from its VWAP benchmark.

Gefen_ImpactOfVolume_2016_HTML

Sticking rigidly to expected volume profile leads to significant price slippage versus benchmark.

In this example for a Skyworth Digital (751 HK) sell order:

On October 20 shortly after the Opening Auction, price and volume spiked.

  • The order’s actual traded volume tracked the historical volume profile closely, but these expected volume profiles deviated from the actual market volume profile significantly between Bin 1 and Bin 7.
  • As a result, price slippage for the order relative to the market VWAP opened up and did not recover for the rest of the day.
  • Why? Because by sticking strictly to the expected volume schedule the trade missed volume at “good” prices and although the volume deviation was not large, the resulting price deviation was.

THE 80-20 ISSUE

We can use this methodology to analyze a large sample of VWAP strategy trades in order to create a more systematic approach to improving slippage of a VWAP trading strategy against a VWAP benchmark. The chart below breaks down the VWAP performance slippage between volume deviation and price deviation by month.

Overall we can see that pricing deviation is responsible for 80% of the slippage, while profile deviation for only 20%.

Gefen_VWAPCostDecomp_2016_HTML

The reason why becomes clear when we analyse the data further based on whether each fill was passive (earned the spread) versus aggressive (crossed the spread). The VWAP strategy is looking to minimize volume deviation by keeping to profile, and as a result “pays up” to catch up with volume by crossing the spread, thus increasing the pricing deviation.

Gefen_VWAPCostDecompby-Fills_2016_HTML

As the proportion of passive fills reduces, the size of the pricing deviation increases. In particular as the proportion of passive fills falls below 60% (i.e., at least 40% of fills are from crossing the spread) pricing deviation grows exponentially.

DYNAMIC VWAP – IMPROVES TRADING STRATEGY TO HELP ACHIEVE A GOAL OF VWAP

Our data-based analysis demonstrates that on average approximately 80% of the slippage against a VWAP benchmark is caused by price deviation.

This typically stems from “paying” for liquidity when crossing the spread due to rigid adherence to the expected volume profile.

ITG is adjusting its Asia Pacific VWAP Algorithm to dynamically balance price and profile deviation, helping traders achieve their intended goal of a VWAP benchmark.

Additionally, our analysis identifies Asia Pacific stocks with less stable volume profiles. These may not be well suited to a volume-based trading strategy at all and other alternatives should be considered. ITG’s electronic execution team will pro-actively alert clients to orders which fit these unstable profiles when a VWAP strategy is selected for them, and discuss alternative trading strategies accordingly.

  • Ofir Gefen

    Managing Director, Head of Electronic Brokerage, Asia Pacific

    Mr. Gefen is a Managing Director and Head of Electronic Brokerage in Asia Pacific.

    Mr. Gefen oversees ITG’s electronic and block trading, as well as algorithmic and dark pool product development in Asia Pacific. Mr. Gefen has been with ITG for over 15 years and in this time has developed experience across all regions through roles in product development and execution services.