A fund’s SRI is reviewed once a year. A portfolio’s risk, however, changes every day. This gap is at the heart of SRI use in financial advisory: understanding how it is calculated, what it does not capture, and why a portfolio’s SRI cannot be obtained by averaging the SRIs of its component funds.
01 — What is the SRI?
The SRI (Synthetic Risk Indicator) is the regulatory risk indicator for packaged investment products (PRIIPs) since 1 January 2023. It replaces the UCITS SRRI on a scale of 1 to 7, where 1 represents the lowest risk and 7 the highest. It appears in the Key Information Document (KID) that every PRIIP initiator provides to retail investors.
Its main difference from the SRRI lies in the significantly wider volatility thresholds and the inclusion of a credit risk measure (CRM) absent from the former indicator. The SRI is therefore the result of two combined measures: the market risk measure (MRM), based on historical volatility, and the credit risk measure (CRM), based on the issuer’s credit rating.
Reference texts — Regulation (EU) No 1286/2014 (PRIIPs), Delegated Regulation (EU) 2017/653 (RTS), Delegated Regulation (EU) 2021/2268 (revision applicable from 1 January 2023), ESMA Guidelines ESMA/2021/1301, AMF Position DOC-2022-08.
SRI vs SRRI: why the thresholds differ
A fund rated SRRI 3 under the old UCITS grid may end up at SRI 4 or 5 under PRIIPs — not because its risk profile has changed, but because the thresholds have been widened. The table in section 5 shows both grids side by side. Before comparing two products based on their risk indicator, it is essential to verify that they use the same reference framework.
The core problem: an annual indicator for daily risk
The SRI shown in a KID is calculated once a year based on five years of historical data. But a portfolio’s risk evolves every day, driven by markets, subscription flows and allocation decisions. This gap between the regulatory update frequency and the actual risk dynamics is the main limitation of the SRI as a continuous monitoring tool.
During periods of crisis, the gap can be stark. A fund displayed at SRI 3 on 1 January may show an annualized standard deviation corresponding to SRI 6 or 7 at the height of a market stress episode, without the KID indicator having been revised. The declared SRI does not capture implied volatility or crisis-regime correlations — which tend to converge towards 1 across asset classes.
02 — Calculating the annualized standard deviation
The MRM is based on the annualized standard deviation of the product’s returns, calculated from the series of net asset values (NAV) over the last five years. The method follows four successive steps.
Step 1: calculating discrete returns
Returns are calculated using the discrete (arithmetic) method, measuring the change in NAV relative to the start-of-period value:
rt = (NAVt − NAVt-1) / NAVt-1
- rt: discrete (arithmetic) return for period t
- NAVt: net asset value at end of period
- NAVt-1: net asset value at start of period
Step 2: calculating the series variance
σ²p = [1 / N] × Σ (rt − r̄)²
- N: total number of observations in the series
- r̄: arithmetic mean of periodic returns
Step 3: annualization
σ = √(σ²p × F)
- F: annualization factor based on data frequency
- F = 252 for daily data (trading days)
- F = 52 for weekly data (PRIIPs reference)
- F = 12 for monthly data
Step 4: assigning the MRM class
The resulting annualized standard deviation is compared against the official thresholds (section 5) to assign the MRM class from 1 to 7. This class constitutes the final SRI, unless the CRM component intervenes (section 6).
Example — Volatility calculation over 5 years
A European equity fund shows a weekly return variance of 0.000342 over 260 weeks (5 years).
Weekly volatility: √0.000342 = 0.01850, i.e. 1.850%
Annualized standard deviation: 1.850% × √52 = 1.850% × 7.211 = 13.34%
→ SRI 4 — An annualized standard deviation of 13.34% falls between 12% and 20%, corresponding to SRI class 4 under the revised PRIIPs grid.
03 — The weekly frequency: the PRIIPs reference method
The PRIIPs RTS (Annex II, point 12) adopt the weekly frequency as the reference for calculating the MRM. The NAV is recorded each week on the same day, over a rolling five-year window of 260 observations.
rt = (NAVweek t − NAVweek t-1) / NAVweek t-1
If a NAV is unavailable for the relevant week (public holiday, suspension), the NAV of the last available trading day is used. The weekly variance is annualized using the factor F = 52.
Why weekly rather than daily?
The choice is not arbitrary. Daily data introduces micro-structural noise (end-of-session effects, temporary illiquidity, bid-ask spreads) that artificially inflates measured volatility, particularly for funds investing in less liquid assets. The weekly frequency smooths these effects while retaining sufficient granularity to capture risk dynamics over five years.
In practice, a fund using daily data (F = 252) will often produce slightly higher volatility than the same series on a weekly basis (F = 52), which can result in a different SRI class. The chosen method must be documented and applied consistently from one revision to the next.
04 — Calculating the SRI with less than 5 years of history
The standard method requires 260 weeks of NAV data. When a fund does not yet have this track record, the PRIIPs RTS provide for alternative rules depending on the situation.
Using a proxy or representative benchmark
When the fund has partial history, the series is extended to 260 weeks using an index or proxy fund with a comparable strategy, investment universe and risk profile. The hybrid series (actual data + proxy data) is then treated exactly like a complete series for the standard deviation calculation.
The choice of proxy is critical: an insufficiently volatile proxy will underestimate the SRI; an overly volatile proxy will overestimate it. The RTS do not mandate a specific proxy but require the professional to justify the representativeness of the chosen substitute.
On EnvestBoard — EnvestBoard uses the fund’s identified current risk factors to complete insufficient data series, ensuring a faithful representation of the current risk profile rather than a generic proxy.
No track record at all
For a fund with no history (recently launched, fund of funds without consolidated track record), two approaches are available under the RTS:
- Similarity approach: use the complete series of a comparable fund in terms of strategy, geography and investment constraints. The SRI calculated on this series is applied to the fund, with explicit mention in the KID.
- Default SRI 6: in the absence of an identifiable and justifiable proxy, SRI 6 constitutes the regulatory prudential floor. This conservative choice protects retail investors from underestimating risk during start-up phases.
Funds with very short history (less than 1 year)
A track record of fewer than 52 weeks does not produce a representative standard deviation. The series is too short to capture a complete market cycle and results are highly dependent on the observation period. In this case, the proxy method is systematically preferable to a direct calculation on actual data alone.
⚠️ Revision at the first annual deadline — An SRI calculated on a proxy must be revised as soon as the fund has sufficient actual data to reduce or eliminate the synthetic portion of the series. Proxy substitution is not a permanent state: it decreases mechanically over time as the actual track record lengthens.
05 — SRI 1 to 7 classification thresholds
The following thresholds are taken from Annex II of Delegated Regulation (EU) 2017/653, as revised by Regulation 2021/2268 applicable from 1 January 2023.
| SRI | Min. volatility | Max. volatility | SRRI equivalent |
|---|---|---|---|
| 1 | 0% | < 0.5% | SRRI 1 |
| 2 | 0.5% | < 5% | SRRI 1 to 3 |
| 3 | 5% | < 12% | SRRI 3 to 5 |
| 4 | 12% | < 20% | SRRI 5 to 6 |
| 5 | 20% | < 30% | SRRI 6 to 7 |
| 6 | 30% | < 80% | SRRI 7 |
| 7 | 80% | No limit | SRRI 7 |
SRRI equivalents are provided for guidance only. The correspondence is not one-to-one: the same fund may be classified differently depending on the framework, due to differences in how the grids are constructed.
The most debated breakpoint: class 2
Under the SRRI, class 2 covered volatilities between 0.5% and 2%. Under the PRIIPs SRI, it extends to 5%. Concretely, short-term bond funds or dynamic euro-denominated funds that were rated SRRI 3 now appear at SRI 2, improving their displayed profile. The reverse also exists: cautious balanced funds previously at SRRI 3 may move to SRI 4 if their volatility exceeds 12%. These shifts create comparison difficulties in multi-fund reports that mix old and new KIDs.
06 — The credit risk measure (CRM)
The MRM alone is insufficient for products exposing the investor to an issuer’s default risk. A CRM component is then added to the calculation, which may raise the final SRI beyond what volatility alone would produce.
| Issuer rating | CRM assigned | Effect on final SRI |
|---|---|---|
| AAA to AA- | CRM 1 | SRI = max(MRM, 1). No effect if MRM ≥ 1. |
| A+ to BBB- | CRM 2 | SRI raised by one notch if MRM < 3. |
| BB+ to B- | CRM 3 | SRI = MRM + 1 if MRM ≤ 5, capped at 6. |
| CCC and below / unrated | CRM 4 | SRI = MRM + 2, minimum 5, capped at 7. |
For equity UCITS and the vast majority of diversified funds without exposure to a single issuer, the CRM is set at 1 and has no practical effect on the final SRI. It becomes relevant for structured bonds, capital-protected funds backed by a single counterparty, or products with a partial issuer guarantee.
07 — Common errors in reading the SRI
-
Comparing an SRI and an SRRI as if they were equivalent — A fund at SRRI 3 and a fund at SRI 3 do not present the same risk level. The grids are built on different thresholds. Any multi-fund comparison must ensure all products display the same indicator. A KID dated before 2023 shows the SRRI; one dated after 2023 shows the SRI.
-
Using the SRI as an absolute measure of risk — The SRI measures historical volatility over five years. It captures neither liquidity risk, nor concentration risk, nor counterparty risk (except via the CRM), nor tail distributions. Two funds at SRI 4 can have very different risk profiles depending on their composition.
-
Ignoring the effect of frequency changes on the class — A calculation using daily data generally produces higher volatility than the same series on a weekly basis. The difference can be sufficient to cross a class threshold. The chosen frequency must be documented and applied consistently over time.
-
Interpreting an SRI decrease as an improvement in the fund — The annual SRI revision reflects changes in historical volatility over five years. A drop in SRI may simply mean that the high-volatility period originally included in the calculation window has fallen outside the observation range. It is not necessarily a sign of improvement in the fund’s future risk profile.
-
Confusing declared SRI with actual risk during stress — A fund’s SRI is calculated once a year on five years of historical data. During a crisis, a fund at SRI 3 on 1 January may show an annualized standard deviation corresponding to SRI 6 or 7 mid-year, without its KID having been updated. The declared SRI is an annual snapshot, not a dynamic measure. Relying on it for continuous portfolio risk monitoring is like driving while looking in the rearview mirror.
-
Forgetting the annual revision — An SRI calculated in 2022 and never revised may be significantly misaligned with the fund’s actual risk profile, particularly after a period of high market volatility. The annual revision is a minimum regulatory requirement, not a ceiling for vigilance.
08 — Why calculating the portfolio SRI matters
A multi-fund portfolio’s SRI cannot be obtained by averaging the SRIs of its component funds. This is the most widespread error in practice, and it can lead to significant underestimation of the portfolio’s actual risk.
Why the arithmetic average is misleading
A portfolio composed 50% of a fund at SRI 2 and 50% of a fund at SRI 6 is not a portfolio at SRI 4. The SRI is an ordinal class built on standard deviation thresholds: the average of two classes corresponds to no actual portfolio standard deviation. To obtain the correct SRI, the annualized standard deviation of the portfolio must be recalculated from the historical performance series of the consolidated allocation and then compared against the PRIIPs thresholds.
Correlation between funds plays a decisive role. Two SRI 5 funds with high correlation will produce a portfolio whose SRI remains close to 5. Two uncorrelated SRI 5 funds can produce an SRI 4 portfolio through diversification. Conversely, funds at moderate SRI but highly correlated with each other provide no real risk reduction — something a simple average never reveals.
Crisis behavior: when the declared SRI diverges from actual risk
Under normal market conditions, a fund at SRI 3 operates within a volatility range of 5% to 12%. During an acute stress episode — liquidity crisis, systemic shock, sharp correction in a concentrated sector — realized volatility may temporarily reach levels corresponding to SRI 6 or 7, without the KID reflecting this reality.
First, correlations between asset classes tend to converge towards 1 during crises: the diversification effect that kept the portfolio’s SRI below individual SRIs disappears precisely when it would be most useful. Second, funds’ declared SRIs incorporate in their five-year window periods of calm markets that dilute recent volatility: a fund that navigated 2019–2021 smoothly shows a low SRI in 2022, even though its risk profile has changed.
The only way to detect this gap in real time is to recalculate the portfolio SRI from recent consolidated valuations, over a short rolling window, and compare it to the five-year SRI. A significant gap between these two measures is a warning signal about the relevance of the declared SRI under current market conditions.
On EnvestBoard — EnvestBoard calculates the portfolio SRI from the allocation’s consolidated valuation series, not by aggregating individual SRIs. The result is compared against PRIIPs thresholds to produce a portfolio SRI directly usable in client advisory and documentable for continuous risk profile monitoring. EnvestBoard gives you access to more than 4,000 risk factors to contextualize this SRI within the relevant peer universe.
09 — Conclusion
A fund’s SRI is a retrospective measure, calculated once a year on five years of weekly data. Useful as a first-level filter, it does not replace a dynamic analysis of actual risk. During periods of market stress, a fund’s declared SRI can be two or three classes away from the risk actually borne by the portfolio.
Three points structure a rigorous reading of the SRI in financial advisory: never aggregate individual SRIs using an arithmetic average to obtain the portfolio SRI, recalculate this indicator from the allocation’s consolidated valuations, and monitor the gap between the declared SRI and the SRI calculated over a short window to detect periods of misalignment. It is in this gap that most of the risk information the KID does not convey can be found.
