The total cost of owning an ETF: a guide on expense ratio and trading costs
Endowus Insights

CPF is for your housing, and so much more.

find out more
.

The total cost of owning an ETF: a guide on expense ratio and trading costs

Updated
4
Sep 2026
published
4
Sep 2026
The total cost of owning an ETF: a guide on expense ratio and trading costs

Number of Pax
Charity List
Select your preferred charity/charities
    This event is only for Accredited Investors (AI) in Singapore. Please verify that you are an AI.
    • An ETF’s expense ratio captures only the cost of holding the fund. Trading costs — the bid-ask spread, premiums or discounts to net asset value (NAV), and the market impact of large orders — are paid at the point of execution and do not feature in a fee table.
    • “Slippage” is defined as the gap between the price an investor expects when they put an order, and the price at which the order ends up filling. It tends to grow with order size and shrink with the liquidity of an ETF and its underlying holdings.
    • The execution channel matters: institutions often trade ETFs through request-for-quote (RFQ) channels, where market makers compete to price an entire order at one firm price, while retail orders typically fill against the visible exchange order book. Unlisted funds, by contrast, deal once a day at NAV, with no spread or slippage at the point of dealing.

    Exchange-traded funds (ETFs) typically appear attractive cost-wise for retail investors. Their headline fees have fallen steadily for two decades, and a total expense ratio (TER) below 0.10% is now routine for the largest index trackers. That visible number, typically printed on all factsheets, has become the default way investors compare products.

    Yet the expense ratio measures only the cost of holding an ETF. It says nothing about the cost of trading one. Every purchase and sale carries execution costs — the bid-ask spread, any premium or discount to fair value, and the market impact of the order itself — that are paid at the moment of the trade. For investors who trade frequently, or in size, these costs may be substantial. 

    Far from recommending to either buy or not buy an ETF, this article simply examines the components of an ETF’s total cost of ownership, explains how slippage arises when orders meet the order book, compares on-exchange execution with request-for-quote (RFQ) trading, and sets out how unlisted funds price differently.

    What does the expense ratio measure?

    TER accrues daily inside the fund’s net asset value (NAV). It compensates the manager, the custodian, and the administrator, and it applies whether an investor trades once a decade or once a week. Trading costs are unrelated to it: brokerage commissions, the bid-ask spread, and any gap between execution price and fair value are all borne by the investor at the point of transaction.

    The two cost layers do not move together. State Street Investment Management, analysing the 100 largest US equity ETFs as of February 2026, found no correlation between expense ratios and bid-ask spreads — some low-fee funds trade with wide spreads, and some higher-fee funds trade very tightly. 

    Cost Component Institutional Trading Vehicle (e.g., SPY) Low-Cost Core Index Fund (e.g., SPYM)
    Primary Design Intent Institutional liquidity & active trading Retail / Advisory buy-and-hold core
    Expense Ratio (Holding) Higher (e.g., 0.09%) Lowest possible (e.g., 0.02%)
    Bid-Ask Spread (Trading) Extremely tight (e.g., ~$0.01 or <0.01%) Can be wider during market volatility
    Secondary Market Liquidity Massive; high daily trading volume Lower; relies heavily on Authorized Participants
    Ideal Use Case Horizon Short-term traders / Hedgers (Trading cost savings outweigh higher holding fees) Long-term investors (Low annual holding costs outweigh entry/exit spreads)

    Disclaimer: Figures shown are estimates and for illustrative purposes only.

    Which layer dominates depends on behaviour. The total cost of ETF ownership separates holding costs (expense ratio, tracking difference) from trading costs (spread, premium/discount volatility, and market impact). For a buy-and-hold investor, holding costs compound over years and dominate. For an investor trading monthly, the trading layer may matter more, because it is paid on every round trip.

    The bid-ask spread is the toll paid on every trade

    An ETF quote always has two prices: the bid, at which buyers stand ready to purchase, and the ask, at which sellers stand ready to sell. An investor buys at the ask and sells at the bid, so the spread is the compensation for market makers, regardless of whether there is a commission attached. 

    Spreads are not fixed. They widen - for example - when the ETF’s underlying holdings are hard to price or thinly traded, and more broadly tend to move in either direction according to market conditions. Dimensional’s ETF trading guidance notes that spreads tend to be wider near the market open, when some underlying holdings have not yet begun trading, and during volatile periods, when market makers’ hedging costs rise. The same ETF can be cheaper to trade (i.e. have a lower bid-ask spread) on a calm Tuesday afternoon and more expensive to trade in the first minutes of a stressed session.

    An ETF’s market price is not its fair value

    Unlike a unit trust, an ETF does not transact at NAV. Its price is set continuously by supply and demand on the exchange, and it can drift above the value of the underlying holdings (a premium) or below it (a discount). Arbitrage normally keeps the gap small: authorised participants can create new ETF units by delivering the underlying basket, or redeem units for the basket, and this mechanism pulls the market price back towards fair value.

    The mechanism is powerful but not instantaneous. In stressed or fast-moving markets, premiums and discounts can widen meaningfully before arbitrage closes them. An investor who buys at a premium and sells at a discount pays that round trip on top of the spread, which can be an additional stealth cost. 

    Slippage grows with order size

    Slippage is the difference between the price an investor expects when placing an order and the average price at which the order actually fills. For small orders in liquid ETFs, it is usually negligible. For larger orders, it becomes a function of the order book itself.

    The exchange displays a finite number of shares at each price level. A large buy order consumes the shares offered at the best price, then fills at the next price up, then the next — pushing the average execution price progressively above the quote the investor first saw, and potentially above the fair value implied by the fund’s underlying holdings. A large sell order works the same way in reverse, eating through the standing bids at lower and lower prices. Traders call this “walking the book”, and it is the dominant form of slippage for size.

    Order fill Shares bought Execution price Cost
    1st fill — at the quoted offer 200 US$50.00 US$10,000
    2nd fill — next price level 500 US$50.05 US$25,025
    3rd fill 200 US$50.10 US$10,020
    4th fill 100 US$50.20 US$5,020
    Total order 1,000 US$50.065 (average) US$50,065
    Slippage vs quoted price +US$0.065 per share (+0.13%) +US$65

    Illustrative example only. Prices, share quantities, and order book depth are hypothetical and do not reflect any actual security, exchange, or trade. The investor sees a quoted offer of US$50.00, but only 200 shares are available at that price; the remainder of the 1,000-share order fills at progressively higher price levels, lifting the average execution price to US$50.065.

    Displayed liquidity, however, understates what is actually available. Because market makers can create or redeem ETF units against the underlying basket, they can absorb orders far larger than the screen suggests. ETF liquidity ultimately reflects the liquidity of the underlying holdings, not the ETF’s own trading volume. The practical question is how an investor reaches that deeper liquidity. 

    This is where the execution channel matters.

    On-exchange and RFQ execution can produce different prices for the same ETF

    On-exchange execution means an order fills against the visible public order book - the investor takes the prices on screen, level by level, as explained thoroughly above. In an RFQ execution, the investor sends the details of the intended trade simultaneously to several market makers, each of which responds with a firm price for the entire order. The investor picks the best quote, and the whole trade executes at that single price.

    The appeal for large trades is threefold. 

    1. Competition among market makers may tighten the quoted price. 
    2. A single firm price removes the walking-the-book problem, because there are no successive price levels to consume. 
    3. Market makers can quote sizes well beyond displayed volume, because they can hedge against — or create and redeem against — the underlying basket. 
    RFQ response Firm price for all 1,000 shares Total cost
    Market maker A US$50.04 US$50,040
    Market maker B — best quote US$50.03 US$50,030
    Market maker C US$50.06 US$50,060
    Market maker D US$50.05 US$50,050
    Trade executed with Market maker B US$50.03 — single price, no averaging US$50,030
    Slippage vs quoted price of US$50.00 +US$0.03 per share (+0.06%) +US$30
    (Potential) saving vs walking the book (US$50.065 average) −US$0.035 per share −US$35

    Illustrative example only. Prices and quotes are hypothetical and do not reflect any actual security, market maker, or trade. The investor requests quotes for the full 1,000-share order; each market maker responds with a firm price for the entire size, and the trade executes in one fill at the best quote. Comparison row refers to the walking-the-book example, in which the same order filled at an average of US$50.065 on-exchange.

    RFQ platforms such as Tradeweb and Bloomberg have made this channel standard practice for institutional ETF trading, and average trade sizes on these venues run to multiples of what the public order book displays at the touch.

    Whether RFQ is automatically cheaper remains contested, and there may not be a definitive answer. For most retail investors in Singapore, the RFQ channel is not directly accessible - it is institutional infrastructure. 

    But where a platform can route client orders through the RFQ channel rather than the public book, a retail investor gains the competitive quoting and single firm price that institutional desks rely on, without needing the counterparty relationships or infrastructure to reach those venues directly. For larger allocations - for example, when client money is pooled - that access may make a meaningful difference to the execution outcome. 

    Unlisted funds price differently: one NAV, once a day

    There is a second structural route to fair-value execution, and it predates the ETF: the unlisted fund. Unit trusts deal on a forward-pricing basis, so every subscription and redemption received before the dealing cut-off transacts at the next NAV calculated after the order arrives. In the United States, forward pricing has been a regulatory requirement for mutual funds under SEC Rule 22c-1 since 1968; the same dealing convention applies to unit trusts offered in Singapore.

    The consequences follow from the structure. There is no bid-ask spread, because there is no continuous two-sided market. There is no premium or discount, because the transaction price is the NAV by definition. And there is no walking the book, because a S$1,000 order and a S$1 million order both deal at the same forward NAV. Execution costs do not vanish — the fund incurs transaction costs when it invests inflows, and these are shared across the fund — but the point-of-trade costs an ETF investor bears individually are absent at the moment of dealing.

    The trade-off is intraday flexibility. A fund investor cannot act on a mid-morning price, use a limit order, or exit in the afternoon of a volatile day. Forward pricing also means the exact transaction price is unknown when the order is placed. 

    Investment implications

    The practical conclusion is that “low cost” is a property of an investor’s whole arrangement — product, trading behaviour, and execution channel together. An infrequent trader in a liquid, tightly quoted ETF may find the expense ratio a fair summary of cost. A frequent trader, or one dealing in size or in less liquid exposures, may pay more at the point of execution than in annual fees.

    In our view, if cost is the largest determinant on whether to hold an ETF or a unit trust, it is worth looking at the investing behavior. But broadly speaking, both ETFs and unit trusts can be used to build a diversified portfolio. 

    Investors trading ETFs directly can reduce execution costs with limit orders rather than market orders, by avoiding the open and periods of high volatility, and by preferring funds whose underlying holdings are liquid. Our advisers can help you assess which specific product fits your investing pattern and goals.

    On the one hand, ETFs offer intraday liquidity, transparent pricing, and decreasing headline fees. On the other hand, investors should look at trading costs on top of holding costs. 

    Frequently asked questions

    What is slippage in ETF trading?

    Slippage is the difference between the price an investor expects when placing an order and the average price at which the order actually executes. It arises because prices move between order placement and execution, and because large orders consume successive levels of the order book, filling at progressively worse prices.

    Is the expense ratio the total cost of owning an ETF?

    No. The expense ratio covers the cost of holding the fund. The total cost of ownership also includes trading costs — brokerage commissions, the bid-ask spread, any premium or discount to NAV at the time of the trade, and market impact — which are paid at the point of execution.

    What is the difference between on-exchange and RFQ execution?

    On-exchange execution fills an order against the visible public order book at the prices displayed. RFQ (request-for-quote) execution asks several market makers to compete on a single firm price for the entire order. RFQ is used mainly by institutions trading in size; retail orders in Singapore typically route to the exchange, though platforms can offer RFQs by aggregative clients’ orders. 

    Do unit trusts have bid-ask spreads?

    No. Unit trusts deal at a single forward-priced NAV, calculated after the dealing cut-off. All investors transact at the same NAV regardless of order size, so there is no spread, premium, or discount at the point of dealing. Fund-level transaction costs still exist but are incurred inside the fund and shared across all investors.

    How can investors in Singapore reduce ETF trading costs?

    Practical steps include using limit orders instead of market orders, avoiding trading at the market open or during volatile sessions when spreads widen, checking an ETF’s typical spread and premium/discount history before trading, and sizing orders with the ETF’s liquidity in mind. 

    Disclaimers
    +
    .

    Do Crazy Rich Asians only invest in real estate?

    Singapore property investment - pros and cons - property vs stocks: which is a better investment?
    .

    Change is the only constant in life: Are index funds truly "passive"?

    Change is the only constant in life: Are index funds truly "passive"?
    .

    The real difference between unit trusts and ETFs

    Shocked man in blue in front of blue background
    The total cost of owning an ETF: a guide on expense ratio and trading costs

    Table of Contents

      find out more
      Check out the top-tier funds included under CPFIS
      find out more
      find out how

      Grow your cash with yields up to

      2.4%

      *
      No lock-ups. No investment limits. No fuss.
      *Not guaranteed. Net yields calculated as of 31 Jul 2026.
      find out how

      Still have questions?

      We're here to help — drop us a message to get instant support.
      connect with us