September 7, 2026

9 min read

Highest Sharpe and Sortino Ratio ETFs, by Asset Class

Which ETFs earned the most return per unit of risk, in equity, fixed income and specialty. Each ratio is printed with its standard error, because most of the ordering sits inside it.

Sharpe and Sortino measure the same thing differently

Both divide the return earned above cash by a measure of variability, so a fund that returned 12% with wild swings can be compared with one that returned 7% quietly.

Sharpe = (return − risk-free rate) ÷ volatility
Sortino = (return − risk-free rate) ÷ downside deviation

Ratio	Definition	What it counts as risk
Sharpe	Excess return ÷ total volatility	Counts every deviation as risk, upside and downside alike. The default when returns are roughly symmetric.
Sortino	Excess return ÷ downside deviation	Counts only returns below zero. Rewards a fund whose volatility is mostly upside, so it separates from Sharpe only when returns are skewed.

Sharpe treats a 5% gain and a 5% loss as equally risky. Sortino counts only the loss, which in principle suits an investor who does not mind upside volatility. In practice the two rank this universe almost identically, and the tables report how closely.

A Sharpe ratio comes with an error bar

A ratio computed from a year of data is not the fund's Sharpe ratio. It is one sample from a distribution, and the sample is small. Andrew Lo derived the standard error in The Statistics of the Sharpe Ratio (Financial Analysts Journal, 2002):

SE(S) = √( (1 + S²/2) ÷ T )
T in years, for an annualised ratio S

A fund showing a Sharpe of 1.0 over twelve months carries a standard error of about 1.22. The two-standard-error range runs from below −1.4 to above 3.4. At five years the same reading carries about 0.55, and over twenty years about 0.27. Lengthening the window is the only fix, which is why these tables rank on five years and print the error beside it.

Two things make the true uncertainty larger still. Returns are not independent, which the derivation assumes. And ranking many funds and reporting the winner is a multiple comparison: with enough candidates, the top of any table is partly a record of which fund was luckiest. Bailey and López de Prado formalise the correction in The Deflated Sharpe Ratio (Journal of Portfolio Management, 2014). The count of funds indistinguishable from each leader is reported below for the same reason.

Our ETF universe is curated, not comprehensive

Roughly four thousand ETFs trade in the United States. These tables rank a curated subset, the same one the Quantlake alert system monitors, and the curation is what makes the ranking readable. Screener output produces three problems at once: the same exposure appears under several tickers, funds of funds compete against the funds they hold, and cash-like instruments sit at the top.

The last is the worst. A fund holding three-month Treasury bills is close to the risk-free asset itself. Its excess return over cash is near zero and so is its volatility, and the ratio of two small numbers is unstable and frequently large. It wins the table without having delivered anything.

The three groups exist so the comparison is like for like. A Sharpe ratio measures return per unit of risk, but how much risk gets paid is a property of the asset class rather than of the fund. Through a rate-tightening window every bond fund can post a negative ratio while every equity fund posts a positive one, and a merged table would then be sorting by asset class rather than by fund. That answers a question nobody asked. The allocation decision comes first; these tables address the one that follows it, which is that given a decision to hold bond exposure, some funds delivered it better than others.

So: one fund per exposure, no funds of funds, no cash proxies, no volatility or inverse products. The composition is reported with the tables and refreshed with them, so the list can be audited rather than trusted. Quantlake also concentrates on passive and rules-based funds, which is what makes a five-year comparison mean anything. An index fund's record is a property of the exposure it holds; an active fund's is partly a property of a manager who may have changed mid-window.

One year measures momentum, five years measures character

The two windows are not competing estimates of the same quantity. Over five years a Sharpe ratio is a classifier: it describes the risk-adjusted character an exposure has held through more than one market state.

Over one year it is a momentum reading. It says what has been working recently on a risk-adjusted basis, which is the same quantity a systematic investor tracks when scaling positions by volatility. How well that ranking persists, and how well it forecasts, are measured below. They are very different numbers, and the gap between them is the difference between describing the present and predicting the future.

Reading the tables without trusting the order

For a systematic investor the useful output is not the first row. Three readings follow from the tables without any need to trust the ordering.

Whether an asset class is being paid at all. A group where every five-year Sharpe is negative did not beat cash over that window, whichever fund is on top. That is an allocation input, not a selection one.

Whether a leader is stable or lucky. The count of funds within one standard error of the leader is printed with every table. Where most of the group sits inside that band, the ranking carries no information and cost, liquidity and tax treatment are what still differ.

Whether the regime is turning. The charts show the same funds through time, on the same measure the table ranks them by. A group crossing zero together is changing state, and that shows months before a five-year column moves.

The league tables

96 funds across 11 issuers, of which 77 are passive and 15 rules-based. Each window is measured against the Treasury bill return over that same window rather than one rate applied to all of history; cash has paid 3.8% over the trailing year. This block refreshes monthly.

Equity

The group is evenly split: two dividend, two factor and two sector funds, plus one country and one structural themes fund. Over the past twelve months, Taiwan rose 95.4%, Value Factor gained 74.4%, and Energy climbed 51.5%.

XLE leads 59 equity funds on 5y Sharpe at 0.85. The standard error on that reading is 0.52, and 30 of the 59 funds sit inside it. The order of the rows is not the finding. The level is.

SPY is not in the table. It ranks 10th of 59 on 5y Sharpe at 0.51. A ratio rewards return per unit of risk, not return, so the funds most widely held sit mid-table whenever something more volatile has been paid for its volatility.

Equity: top 8 by 5y Sharpe

59 funds · ranked on 5y · as of September 4, 2026

ETF	Since	1Y	3Y	5Y ± SE	Life	Sortino 5Y	Return 1Y	Return 5Y p.a.
XLE Energy	Dec 1998	2.20	0.54	0.85 ± 0.52	0.24	1.18	+51.5%	+25.8%
IVLU Intl Value	Jul 2015	1.88	1.35	0.75 ± 0.51	0.38	1.08	+32.9%	+16.3%
VLUE Value Factor	Apr 2013	3.44	1.59	0.73 ± 0.50	0.65	1.07	+74.4%	+17.5%
EWT Taiwan	Jun 2000	2.93	1.52	0.70 ± 0.50	0.20	1.01	+95.4%	+20.7%
IDV Intl High Dividend	Jun 2007	2.16	1.62	0.64 ± 0.49	0.16	0.91	+31.6%	+13.9%
XLK Technology	Dec 1998	1.50	0.99	0.60 ± 0.49	0.32	0.86	+43.4%	+19.5%
VYM High Dividend	Nov 2006	1.66	1.09	0.58 ± 0.48	0.43	0.84	+20.7%	+12.0%
IYW US Technology	May 2000	1.44	1.11	0.56 ± 0.48	0.28	0.80	+38.9%	+18.8%

Ranking the same funds by Sortino gives a rank correlation of +1.00 and the same five funds in the top five. Sortino separates from Sharpe only where returns are skewed, and this group is not skewed enough for the choice of denominator to change the answer.

The same funds, on the same measure the table ranks them by, through time. XLE's five-year Sharpe alone travels 1.50 across the decade. A figure quoted as though it were a property of the fund is a property of the window it was measured over.

Fixed income

Credit accounts for five of the eight fixed income funds, and six of the eight did not beat cash over the past twelve months. Over that period EM Sovereign Bonds rose 4.9% and beat cash, High Yield Bonds rose 4.3% and beat cash, while High Yield rose 3.9% but did not beat cash.

HYG leads 17 fixed income funds on 5y Sharpe at −0.05. The standard error on that reading is 0.45, and 5 of the 17 funds sit inside it. The order of the rows is not the finding. The level is.

AGG is not in the table. It ranks 11th of 17 on 5y Sharpe at −0.70. A ratio rewards return per unit of risk, not return, so the funds most widely held sit mid-table whenever something more volatile has been paid for its volatility.

Fixed income: top 8 by 5y Sharpe

17 funds · ranked on 5y · as of September 4, 2026

ETF	Since	1Y	3Y	5Y ± SE	Life	Sortino 5Y	Return 1Y	Return 5Y p.a.
HYG High Yield	Apr 2007	−0.01	0.67	−0.05 ± 0.45	0.31	−0.07	+3.9%	+3.6%
JNK High Yield Bonds	Dec 2007	0.12	0.67	−0.06 ± 0.45	0.30	−0.08	+4.3%	+3.5%
EMB EM Sovereign Bonds	Dec 2007	0.18	0.60	−0.26 ± 0.45	0.29	−0.37	+4.9%	+1.4%
VTIP Short TIPS	Oct 2012	−1.16	0.25	−0.29 ± 0.46	0.16	−0.43	+2.2%	+3.2%
PFF Preferred Stock	Mar 2007	−0.40	0.15	−0.31 ± 0.46	0.12	−0.43	+0.9%	+0.8%
VCIT Corporate Credit	Nov 2009	−0.82	0.19	−0.50 ± 0.47	0.46	−0.71	+0.6%	+0.7%
MBB Mortgage-Backed	Mar 2007	−0.44	−0.03	−0.55 ± 0.48	0.22	−0.78	+2.0%	+0.2%
VCSH Short Corporate	Nov 2009	−0.85	0.29	−0.55 ± 0.48	0.47	−0.82	+2.3%	+2.4%

Ranking the same funds by Sortino gives a rank correlation of +0.99 and the same five funds in the top five. Sortino separates from Sharpe only where returns are skewed, and this group is not skewed enough for the choice of denominator to change the answer.

The same funds, on the same measure the table ranks them by, through time. VCSH's five-year Sharpe alone travels 2.19 across the decade. A figure quoted as though it were a property of the fund is a property of the window it was measured over.

Specialty

Commodities account for four of the eight specialty leaders, with two structural themes, one alternatives fund and one hard assets fund. Over the past twelve months, Oil rose 96.0%, Semiconductors rose 94.4%, and Gold Miners rose 51.0%.

SMH leads 20 specialty funds on 5y Sharpe at 0.81. The standard error on that reading is 0.52, and 11 of the 20 funds sit inside it. The order of the rows is not the finding. The level is.

Specialty: top 8 by 5y Sharpe

20 funds · ranked on 5y · as of September 4, 2026

ETF	Since	1Y	3Y	5Y ± SE	Life	Sortino 5Y	Return 1Y	Return 5Y p.a.
SMH Semiconductors	May 2000	2.30	1.35	0.81 ± 0.52	0.31	1.17	+94.4%	+33.8%
GLD Gold	Nov 2004	0.65	1.24	0.81 ± 0.51	0.48	1.13	+23.0%	+19.2%
GDX Gold Miners	May 2006	0.92	1.19	0.61 ± 0.49	0.10	0.87	+51.0%	+27.0%
FTLS Long/Short Equity	Sep 2014	1.26	0.96	0.60 ± 0.49	0.63	0.86	+14.3%	+10.2%
PAVE Infrastructure	Mar 2017	0.75	0.76	0.54 ± 0.48	0.53	0.78	+19.4%	+15.8%
USO Oil	Apr 2006	1.91	0.51	0.54 ± 0.48	−0.22	0.75	+96.0%	+24.0%
DBA Agriculture Basket	Jan 2007	0.56	0.69	0.51 ± 0.48	0.00	0.73	+10.1%	+11.0%
PDBC Broad Commodities	Nov 2014	2.31	0.51	0.48 ± 0.47	0.13	0.67	+50.8%	+13.3%

The same funds, on the same measure the table ranks them by, through time. DBA's five-year Sharpe alone travels 2.08 across the decade. A figure quoted as though it were a property of the fund is a property of the window it was measured over.

What the one-year column is worth

Does the one-year ranking persist, or forecast?

Rank correlation across the universe · 298 monthly observations

Horizon	Ranking still holds	Predicts the returns that follow
6 months later	0.51	0.06
12 months later	0.13	0.09

Six months on, the one-year Sharpe ordering still correlates 0.51 with itself. Its correlation with the returns that actually followed is 0.06. The first number is why a one-year reading is a usable momentum measure. The second is why it is not a forecast.

Ratios use close-to-close volatility over each window, against the Treasury bill return over that same window rather than a single rate applied to all of history. Standard errors follow Lo (2002) and assume independent returns, so they are a floor on the uncertainty. As of September 4, 2026.

Frequently asked questions

Which ETF has the highest Sharpe ratio?

It depends entirely on the window, which is the honest answer rather than an evasive one. The tables on this page rank on five years and also show one year, three years and full history, and the leader is frequently a different fund in each column. A ranking quoted without its window and its sample size is not a measurement.

What is a good Sharpe ratio?

Below 1 is sub-par, 1 to 2 is good, above 2 is strong but rare over a full cycle. Those bands apply to a ratio measured over a long window. Over one year the standard error alone is larger than the gap between the bands, so a one-year reading of 2.0 is not evidence of a strong fund.

Is Sortino better than Sharpe?

Only when returns are skewed. Sortino counts just the downside as risk, so it rewards a fund whose swings are mostly upward. On this universe the two produce nearly the same ordering, and the live tables report the rank correlation so the claim can be checked rather than assumed.

How long a track record do you need to rank funds?

Longer than most rankings use. The standard error of an annualised Sharpe falls with the square root of the number of years, so halving the uncertainty takes four times the history. One year is not enough to separate funds; five begins to, and a full history separates them better still.

Does a high Sharpe ratio mean an ETF will keep performing?

No, and this page measures the gap rather than asserting it. Across the universe the one-year Sharpe ranking holds together reasonably well six months later, so it describes a state that persists. Its correlation with the returns that actually follow is close to zero. High recent risk-adjusted performance is information about the present, not about the future.

Can I use the Sharpe ratio to pick ETFs?

As one input, and not as the deciding one. On these tables a large share of each group sits within one standard error of the leader, so the ordering cannot support a choice on its own. Where the ranking is that tight, cost, liquidity, tracking difference and tax treatment are the variables that still differ meaningfully between funds.

Why isn't the biggest or best-known ETF at the top?

Because a Sharpe ratio rewards return per unit of risk, not return. A fund can return more than another and still rank below it by carrying more volatility to get there, which is routinely the case for concentrated exposures against broad ones. Where each group's benchmark ranks is printed with its table and refreshes monthly, so the answer is the current one rather than a claim about any particular fund.

How can a fund make money and still show a negative Sharpe ratio?

Because the ratio measures return above cash, not return. A fund that gained 3% in a window when Treasury bills paid 4% has a negative excess return and therefore a negative ratio, despite having made money. This happens across whole asset classes when short rates are high, and it is a statement about the rate environment over that window rather than about the funds.

How often should these rankings be reviewed?

The five-year tables move slowly and are worth a quarterly look. The one-year column and the rolling charts change monthly and are the part worth monitoring, because that is where a change of regime shows up first.

Romain Gandon

CEO, Quantlake