Tracking Error & Tracking Difference: A Quantitative Guide to Measuring How Closely an ETF Tracks Its Index

Vietnam's onshore ETF segment closed the 28 October 2026 session with 24 HOSE-listed funds, combined estimated AUM of 22.96 trillion VND and total secondary-market turnover of just 2.25 billion VND — a turnover-to-AUM ratio of roughly 0.01%, which tells you that most of the day's price prints were struck on very thin order books. The VN-INDEX finished at 1,759.75 points (-0.12%), the VN30-INDEX at 1,886.85 points (-0.08%), and breadth across the listed ETF universe was 1 advancer, 9 decliners and 14 unchanged. Foreign investors bought 0.31 billion VND and sold 1.36 billion VND on HOSE, a net -1.05 billion VND, according to HOSE / VCI order-book data on etf.com.vn. In a session like this, the headline percentage change on a fund's ticker tells you almost nothing about replication quality — the metric that matters is the gap between the fund's return and its benchmark's return, and how volatile that gap has been.

Key Takeaways

• Tracking difference is a return gap; tracking error is its volatility. On 28 October 2026, E1VFVN30 closed at 33,870 VND (-0.38%) against a VN30-INDEX print of -0.08% — a single-session differential of -0.30 percentage points that is statistically indistinguishable from noise and must not be read as fee leakage.
• Liquidity distorts every tracking statistic. FUESSV30 printed 23,400 VND (-2.50%), 1.47% below its own intraday VWAP of 23,750 VND, on just 400 units (approximately 9.4 million VND of turnover), while FUEKIV30, FUEKIVFS and FUEMAVND recorded zero volume and closed exactly at their reference VWAP — a stale-price pattern that mechanically biases computed tracking error downward.
• Secondary-market foreign flow is not fund flow. The -1.05 billion VND net foreign sell figure is an order-book statistic from HOSE; FUEVFVND alone accounted for -0.92 billion VND of it (87.6% of the total). It says nothing about primary-market creation or redemption, which is the only mechanism that changes the 22.96 trillion VND AUM pool.

Tracking Difference vs Tracking Error: The Definitions That Drive Due Diligence

The two terms are routinely conflated in retail commentary, and the confusion produces genuinely wrong conclusions about whether a fund is doing its job. They answer different questions and are computed from different inputs.

Two Metrics, Two Different Questions

Tracking difference (TD) answers: how much return did the fund lose or gain relative to its index over a period? It is a single signed number, usually annualised. For a physically replicated, fully funded domestic equity ETF, TD should converge on the negative of the total expense ratio plus or minus frictions. Tracking error (TE) answers: how consistent was that gap? It is the annualised standard deviation of the periodic TD series. A fund can post a small TD and a large TE (consistent but noisy), or a large TD and a small TE (a stable, predictable fee drag). Institutional allocators care about TE for sizing and about TD for realised cost.

The Formulas Used on etf.com.vn

The methodology applied across our fund pages — documented in full at /en/phuong-phap-tinh — is deliberately simple and reproducible:

TD_t = (NAV_t / NAV_t-1 - 1) - (Index_t / Index_t-1 - 1), computed on NAV, not market price.

TE = stdev(TD_1..n) × √250, where 250 approximates annual HOSE trading sessions. Annualised TD over a window of n sessions is Π(1 + TD_t) - 1, then scaled by 250/n.

Three inputs must be respected. First, use NAV, not the exchange close — market price embeds a premium or discount that belongs to a separate diagnostic. Second, align the fund and index observation to the same valuation timestamp; Vietnam's NAV is struck after the close, so using intraday index prints introduces a timing mismatch that inflates measured TE. Third, never annualise a window shorter than about 60 sessions without flagging it, because the √250 scaling on a 10-day sample produces statistically meaningless figures.

Three Numbers That Are Not Tracking Error

This session illustrates the trap cleanly. E1VFVN30 closed at 33,870 VND against an intraday VWAP of 33,925 VND — a gap of -0.16% relative to VWAP. FUEVFVND closed at 32,240 VND against a VWAP of 32,345 VND, or -0.33%. FUEVN100 actually closed above its VWAP, at 24,690 VND versus 24,670 VND (+0.08%). None of these are tracking error. They are execution slippage: the difference between a fund's closing print and the volume-weighted average price at which its units actually traded during the session. Likewise, a fund trading at a premium or discount to NAV is a supply-and-demand artefact of the creation/redemption mechanism, not a replication failure. Keep four ledgers separate — TD, TE, premium/discount, and VWAP slippage — or you will systematically misdiagnose liquid funds as poorly managed and thin funds as well managed.

Vietnam ETF Session Data: 28 October 2026

The table below isolates the observable inputs that feed a tracking analysis. AUM concentration is extreme: FUEVFVND (10,443 billion VND, 45.5% of the 22.96 trillion VND universe) and E1VFVN30 (5,860 billion VND, 25.5%) together hold 71.0% of all HOSE ETF assets.

Source: HOSE / VCI order-book data on etf.com.vn, session 28/10/2026. AUM figures are platform estimates. Turnover implied by price × volume is approximately 0.71 billion VND for E1VFVN30 and 1.23 billion VND for FUEVFVND, together roughly 86% of the 2.25 billion VND market total.

Where the Volume Actually Is — and Why It Matters for Measurement

Two funds absorbed the session: E1VFVN30 traded 20,900 units and FUEVFVND 38,200 units. Everything else was residual. FUESSVFL moved 4,900 units at exactly its VWAP of 28,200 VND with zero foreign net activity. FUEKIV30, which carries 2,298 billion VND of AUM (the third-largest pool in the market), recorded zero units traded. A fund with 2.3 trillion VND of assets and no secondary-market prints is not a liquid instrument, and any TE figure derived from its exchange prices on such a day is measuring the absence of trading, not the quality of replication.

Secondary-Market Foreign Flow Is Not Creation/Redemption Flow

The market-wide foreign net of -1.05 billion VND is a HOSE order-book measurement: the difference between foreign-labelled buy value (0.31 billion VND) and sell value (1.36 billion VND) executed on the exchange. FUEVFVND's -0.92 billion VND dominates that total. This is not fund flow. An ETF's AUM changes only through primary-market creation and redemption conducted with the fund issuer by authorised participants — activity that never appears in the exchange tape. A day of heavy foreign net selling on HOSE can coexist with net creations, and vice versa. Analysts who treat exchange foreign net as a proxy for ETF inflows will misstate the direction of capital movement roughly as often as they get it right.

Decomposing the Tracking Difference Into Its Structural Drivers

Fee Drag, Cash Drag, and Dividend Timing

The dominant, predictable component of TD is the management fee. Vietnam-domiciled equity ETFs typically price in the 0.6%–0.8% annual range, which sets a structural floor on annual TD at approximately that level before any other friction. Above that floor sit three secondary drags: cash held for pending creations and redemptions (uninvested cash dilutes exposure during rising markets), dividend accrual timing (cash received but not yet reinvested sits idle between the ex-date and the reinvestment date), and the withholding treatment applied to dividends before they reach the fund. All three are estimateable from holdings disclosures and all three are controllable by the manager — which is why they are the first questions to ask an issuer.

Rebalancing Cost and Basket Concentration

Index reconstitution is where TD becomes lumpy. Semicircular rebalances force the fund to trade at the close on a known date, and the cost of that trade is a realised TD event, not a random one. Concentration amplifies the effect because the marginal position is larger relative to the fund's typical daily turnover. In the VN30 basket underlying E1VFVN30, FUESSV30, FUEKIV30 and FUEMAV30, the top seven names — FPT (12.4%), TCB (9.2%), ACB (8.6%), HPG (8.4%), VPB (7.5%), MBB (7.1%) and MWG (6.2%) — account for 59.4% of the benchmark. The VN-Diamond basket behind FUEVFVND and FUEMAVND is tighter still: FPT (15.1%), PNJ (12.8%), MWG (12.4%), GMD (10.6%), ACB (9.8%), TCB (9.1%), MBB (7.9%) and REE (6.8%) sum to 84.5%. The VNFinLead basket behind FUESSVFL and FUEKIVFS carries 74.7% in six names, led by TCB (14.8%), ACB (14.2%) and MBB (13.6%). Higher concentration implies higher expected TE for an equal-cost fund, purely because single-stock liquidity events transmit directly into the replication gap.

Building a Tracking-Error Scorecard in Practice

Window Selection and Frequency

Run three windows in parallel: 20 sessions (detects operational breaks — a botched rebalance, a cash injection, a settlement failure), 60 sessions (the shortest horizon that produces a statistically usable annualised TE), and 250 sessions (the structural read, where the fee floor becomes visible). Compare the fund against its own peer set on the comparison tool, and always show TD alongside TE. A fund with TD of -0.72% and TE of 0.35% is behaving exactly as its fee schedule predicts; a fund with TD of -0.72% and TE of 2.10% has an execution problem that the fee line does not explain.

The Stale-NAV Trap in Zero-Volume Sessions

Autocorrelation is the enemy of TE measurement. When a fund does not trade — as with FUEKIV30, FUEKIVFS and FUEMAVND on 28 October 2026, all of which closed at 0.00% and precisely at their VWAP — its recorded price series becomes a step function. Consecutive zero-change observations produce artificially small daily TD values, and the standard deviation of that series is biased downward. The result is a TE figure that flatters illiquid funds. The remedy is to compute TE on NAV (which does update, being marked from the underlying basket) rather than on market price, and to flag any window in which more than 20% of sessions had zero volume.

Cross-Checking Against the Constituent Basket

A final sanity check: recompute the index return from the published constituent weights and compare it to the official index print. If they diverge by more than a few basis points, the discrepancy is a data-vendor or corporate-action adjustment issue and will contaminate your TE series. Use the stock lookup tool for constituent-level verification and the portfolio builder to model how a basket-weighted return would have behaved. For a worked fund-level example, see the E1VFVN30 product page, the full ETF product catalogue, and the methodology primer at /en/hoc-etf.

Frequently Asked Questions

1. What is a reasonable tracking error for a Vietnam equity ETF?
For a physically replicated, fully funded domestic equity ETF with a management fee in the 0.6%–0.8% range, annualised ex-post TE computed over 250 sessions should cluster near the fee level plus rebalancing noise — meaning roughly 0.5% to 1.5% for a large-cap VN30 tracker. Readings materially above that band, sustained across 60-session windows, indicate an execution or cash-management issue rather than cost. Funds with concentrated benchmarks such as the VN-Diamond basket (84.5% in eight names) should be expected to sit at the upper end of the range, not the lower end.

2. E1VFVN30 fell 0.38% while the VN30-INDEX fell 0.08% on 28 October 2026. Is that a tracking failure?
No. A -0.30 percentage point single-session differential is well inside normal intraday timing and valuation noise, and it is measured on market price rather than NAV. Tracking error is the annualised standard deviation of the daily tracking-difference series, so one observation contributes essentially nothing to the statistic; with 250 sessions the same 0.30 pp gap recurring in a consistent direction would matter, and a one-off print would not. Note also that the exchange close at 33,870 VND sat 0.16% below the 33,925 VND intraday VWAP, which places part of the gap in the execution-slippage ledger rather than the replication ledger.

3. Why did FUEKIV30, FUEKIVFS and FUEMAVND close at exactly 0.00% with zero volume?
Because they did not trade. With no secondary-market prints, the closing reference price reverts to the prior session's level or the fund's reference VWAP — in this case FUEKIV30 at 12,990 VND, FUEKIVFS at 16,610 VND and FUEMAVND at 13,750 VND, each identical to its stated VWAP. These are stale prices, not evidence of perfect replication. Any tracking-error calculation that uses such prints unadjusted will understate risk, which is why etf.com.vn computes and displays TE on NAV-marked series and flags zero-volume sessions separately.

Go Deeper on etf.com.vn

Screen tracking statistics across all 24 HOSE-listed funds on the live market dashboard, run side-by-side TD and TE comparisons on the comparison tool, model basket-weighted outcomes in the portfolio builder, and review the full ETF product catalogue for benchmark-level concentration data before you size any allocation.

Disclaimer: This content is for informational purposes only and does not constitute personal investment advice.