By Vincent Deluard, CFA, Vice President, Global Macro Strategy at the BD Division of INTL FCStone Financial Inc. 

Index funds were originally launched as an answer against the excesses of active management with the explicit goal of minimizing trading expenses. 40 years after the launch of First Index Investment, it is somewhat ironic that the most traded security in the world is an index fund, and that ETFs account for about a third of all trading volume on U.S. exchanges.  J. Bogle, the godfather index investing, recently criticized the evolution of the ETF industry:  providers have turned a simple low-cost vehicle into a fee-generating scheme, with thousands of niche ETFs and a push towards the oxymoronic concept of “smart beta”.

He also argued that excessive turnover “takes a toll on investor wealth”. This claim surprised me. I always viewed ETFs’ heavy volume as a positive thing: high liquidity allows investors to complete large trades of thousands of stocks and bonds all day long with minimal market impact.

The “race to zero” has greatly reduced explicit trading costs: SPY (Spiders S&P 500 Index) has traded at a penny-wide spread for more than 3,100 days and retail brokers offer hundreds of commission-free ETFs.

However, J. Bogle argues that speculation has hidden costs beyond fees. When investors rush to trade ETFs after every piece of news (or Presidential tweet), they confuse the noise for the signal. They buy high and they sell low, and underperform cooler-headed investors.

This effect can be quantified because providers disclose the daily flows in and out of ETFs. This sequence of cash flows can be turned into an internal rate of return – a precise measure of the actual return achieved by investors’ collective trading decisions.

The example below shows the assets, NAV, and price of a fictional ETF over 5 days. The bottom line shows cash flows from the collective perspective of ETF investors. On day 0, their position was worth $1,000 which is recorded as an outflow (the cost of initiating the position). ETF outflows are recorded as positive cash flows (money leaving the fund going to investors) and ETF inflows are recorded as negative cash flows (money leaving investors for the fund). The terminal value on day 4 is a positive cash flow. Just as we would with any bond[1], we can compute the internal rate of return of this sequence of cash flows. In this case, investors achieved a return of 4.24%[2], which is higher than the buy-and-hold return of 3%.

ETFChart2 (1)

How did investors realize a 4.24% return on a fund that only gained 3%? By trading smartly: investors bought on Day 1 and Day 2 when the fund was cheap, and they sold on day 3 and 4 when it was expensive.

I replicated this calculation for the 1,000 largest ETFs in the world by comparing each fund’s flow[3]-weighted return to its “normal” time-weighted return. The difference represents the active return achieved by ETF investors’ collective trading decisions.

I measured ETFs’ active trading returns from November 8, 2016 to March 16, 2017 – an interesting period because it covers the market’s reaction to the unexpected outcome of the U.S. presidential election.

Flow-weighted returns underperformed time-weighted returns by a median of 0.23% over the period[4]. The active trading return was negative for 617 funds and positive for 417 funds. A 20-basis point difference over four months is probably insignificant but the asset-weighted active trading return was more negative at 1.15%. This owes to the fact that active trading returns were more negative for the heavily-traded ETFs. On the surface seems consistent with the Bogle hypothesis of “over-trading”. But the reality is more complex than that.

 

 

 

 

 

First, it is important to keep in mind that some trading is a necessity: managers need to invest new money coming in, they need to sell to meet redemptions, they need to adjust portfolios as clients’ conditions change, and they need to pay taxes. The alternative of “buying-and-holding forever” presented by the passive school is not realistic beyond the ivory towers of academia.

If some market participants consume liquidity, others need to provide it, and they need to earn a rate of return for providing this service. ETF investors paid a median 20 basis point for the convenience of liquidity[5] over a four-month period. It does not seem excessive.

A more relevant question for real-world investors is whether trading allows investors to achieve their goals, and at what cost. Our analysis of flow-weighted returns shows that most of the return slippage came from a handful of large sector ETFs. Specifically, ETF investors scrambled to adjust to the violent rebound of cyclical sectors that followed the election.  The S&P 500 Financials Index soared by about 11% in just three days, and much of the inflows into the sector came after that spike. Similarly, defensive sector ETFs were mostly sold after their initial post-election plunge. ETF investors bought relatively high and sold relatively low.

Yet, we must consider the alternative, which would have been to do nothing and not adjust sector allocation following the election. I created a fictional portfolio that passively held the main sector ETFs based on their pre-election market cap. This portfolio would have started with a market value of $92.8 billion on November 8, and would have grown to $102.7 billion by November 16, for a total gain of $9.9 billion.

In the real world, investors added a net $18.8 billion to the main sector ETFs over the period, and their assets grew to $121.7 billion, for a net gain of $10.0 billion. This is about $100 million more than if they had done nothing.

Does this contradict our earlier observation that their timing was generally poor? Not necessarily: their poor timing was more than offset by two positive factors. First, investors were net buyers, a beneficial decision in a rapidly-rising market. Second, they made the right sector picks. The re-allocation from bond proxies into cyclicals may have been made at the wrong time (i.e. investors did not capture the rebound in full), but it was still better than the alternative of not trading at all.

In conclusion, critics of active managers are right to point that trading is never free, and that on aggregate, ETF investors tend to buy high and sell low. We view this slippage as the cost of consuming liquidity. This cost needs to be measured against the benefits of achieving investors’ goals. In the case of sector ETFs following the Trump election, achieving a more appropriate sector mix was worth losing a few basis points on poorly-timed trades.

NOTE: All underlying ETF market data in this article was sourced from Bloomberg.

 

[1] For bonds, this would be the yield-to-maturity: the discount rate that ensures that net cash flows have a present value of zero.

[2] This return can be approximated by dividing the net gain of $49 (1,249+100-200-100-1,000) by average invested assets of $1,160

[3] Flow is measured as the daily change in shares times volume-weighted average price (VWAP)

[4] This does not include explicit trading costs such as commission and fees

[5] This slippage is captured by non-ETF investors, who collect a premium for satisfying ETFs’ demand for shares when prices are relatively high.