Direct Answer
Negative Volume Index (NVI) is a cumulative indicator that only recalculates on days when volume decreases from the prior day, adding that day's percent price change (as a share of the running total) whenever the update condition is met and holding flat otherwise. It's the mirror-image companion to the Positive Volume Index (PVI), which does the same thing on rising-volume days instead. The traditional interpretation, NVI above its own long-term moving average read as bullish, is a dated heuristic from Dysart and Fosback's original research, not a proven predictive signal.
Key Takeaways
- NVI only updates when today's volume is lower than yesterday's; on every other day it carries forward unchanged.
- The update, when it happens, adds today's percent price change to the running total, not the raw price change and not the volume figure itself.
- NVI and PVI are mirror images: every session updates exactly one of the two lines, never both, since a given day's volume is either higher, lower, or equal to the prior day's.
- The traditional signal compares NVI with its own long-term moving average (commonly a roughly one-year lookback, mirroring the convention used for PVI).
- That signal comes from Dysart's original observations and Fosback's later research, decades-old work, not a proven or independently re-validated predictive relationship.
What Is the Negative Volume Index?
Negative Volume Index (NVI) is a cumulative indicator built from close-to-close percent price change, but it is selective about which days it counts. It only recalculates on a day when volume is lower than the previous day's volume. On any day where volume is flat or higher, the indicator simply carries its prior value forward unchanged.
The idea behind isolating lower-volume days is associated with research by Paul Dysart and later Norman Fosback, on the premise that price moves on quieter, lower-volume days might reflect a different kind of participation than moves on heavy-volume days, sometimes described loosely as "smart money" activity, since large crowds chasing a move tend to show up as higher volume. NVI is the counterpart to the Positive Volume Index (PVI), which applies the identical rule to rising-volume days instead of falling-volume days.
Key takeaways: NVI is a selective cumulative indicator, it only updates on lower-volume days and holds flat otherwise. The update adds that day's percent price change to the running total, not the raw price or volume value. NVI and PVI are mirror images built from the same close-to-close logic, just triggered by opposite volume conditions. The traditional signal compares NVI with its own long-term moving average, and that signal is a dated heuristic from decades-old research, not a proven predictive tool.
The Formula
Rule: if Volumetoday is less than Volumeyesterday, then:
NVItoday = NVIyesterday + NVIyesterday × ((Closetoday − Closeyesterday) ÷ Closeyesterday)
Otherwise, NVItoday = NVIyesterday, the line is unchanged.
| Condition | What happens | Interpretation note |
|---|---|---|
| Volumetoday < Volumeyesterday | NVI adds yesterday's NVI value multiplied by today's percent price change | This is the only condition under which the line moves, the calculation deliberately ignores what happened on higher- or equal-volume days. |
| Volumetoday ≥ Volumeyesterday | NVI carries forward unchanged | A flat reading on this kind of day says nothing by itself about that day's price move. It means the day simply didn't meet the update condition. |
| Starting value | Arbitrary (commonly set to a round number such as 1,000) | Because the indicator is cumulative and multiplicative, only its shape and trend over time carry information, the absolute number is not comparable across instruments. |
Note the multiplicative structure: each update multiplies the prior NVI value by (1 + percent change), so the effect of a given percent move compounds relative to the current level of the line rather than adding a fixed amount. This is the same style of calculation used by On-Balance Volume's volume-accumulation logic, though OBV adds raw volume rather than a percent price change.
Worked Example
Hypothetical example, for education only.
Start with an arbitrary baseline of NVI = 1,000 on Day 1, when the stock closes at $50.00.
Day 2: volume is lower than Day 1's volume, and the close rises to $51.50 (+3.00% from Day 1). Because the volume condition is met:
NVIDay 2 = 1,000 + 1,000 × (1.50 ÷ 50.00) = 1,000 + 1,000 × 0.03 = 1,030
Day 3: volume rises above Day 2's volume, so the update condition is not met. NVIDay 3 stays at 1,030, regardless of what price did that day.
Day 4: volume falls again, below Day 3's volume, and the close drops to $50.47 (−2.00% from Day 3's $51.50). The volume condition is met again:
NVIDay 4 = 1,030 + 1,030 × (−1.03 ÷ 51.50) = 1,030 + 1,030 × (−0.02) = 1,009.40
Across these four days, NVI only changed on the two days where volume declined from the prior session, it ignored Day 3's price action entirely because Day 3's volume didn't meet the rule. This is the mechanical trade-off at the heart of NVI: it deliberately discards information from higher-volume days to isolate what happened on quieter ones.
How NVI Is Commonly Used
NVI versus its own long-term moving average
The traditional interpretation, associated with research by Paul Dysart and later Norman Fosback, compares the NVI line with a long-term moving average of itself, commonly a roughly one-year lookback, mirroring the convention used for PVI. When NVI sits above that moving average, the reading is traditionally described as bullish; below it, more cautious or neutral. This is a dated heuristic from decades-old research, not a proven or independently re-validated predictive signal, and it should be treated with that caveat rather than as a mechanical buy/sell rule.
Trend and slope
Beyond the moving-average comparison, traders sometimes read the general slope of NVI on its own, a rising line suggests that lower-volume sessions have skewed toward gains; a falling line suggests they've skewed toward losses. As with the moving-average signal, this describes historical behavior on a subset of days, not a forecast of what happens next.
Pairing with PVI
Because NVI and PVI are calculated from the same close-to-close logic on opposite volume conditions, some traders look at both lines together rather than NVI in isolation, for example, comparing whether gains are concentrating on lower-volume or higher-volume days. Reading the two together doesn't resolve the underlying limitation that both are built from a fairly narrow definition of "volume decreased" or "volume increased" versus the single prior day.
Common Moving-Average Settings
The traditional NVI signal is read against its own long-term moving average, commonly a roughly one-year lookback, by analogy to the convention used for PVI. Different charting platforms and analysts have used somewhat different exact lengths for that lookback, and the underlying NVI calculation itself has no adjustable period, every day is either included (volume down) or excluded (volume flat or up) by a fixed rule. The only real parameter choice is the length of the moving average laid on top of it, and there's no universal, independently proven optimal length for that; traders should verify the exact period their charting platform uses before comparing readings.
Limitations
- The traditional signal is decades old and not independently re-validated. Dysart's original observations and Fosback's later research describe a historical relationship in a specific dataset from an earlier era of market structure, not a guarantee that the same relationship holds in current markets.
- The volume comparison is strict and single-day. "Lower than yesterday" depends only on the immediately prior session, so a single noisy volume reading can flip whether a given day updates the line at all.
- Raw NVI values aren't comparable across instruments. Because the calculation is cumulative and multiplicative from an arbitrary starting value, only the line's shape and trend carry meaning, not its absolute level.
- It discards information from higher-volume days entirely. A large price move on a heavy-volume day contributes nothing to NVI, even though that day may be highly informative on its own.
- Data source matters. Consolidated-tape versus single-venue volume reporting can change which days meet the "volume decreased" condition, which can shift the resulting line between providers.
Common Mistakes
- Treating "NVI above its moving average" as an automatic buy signal. It's a historical correlation from older research, not a mechanical trading rule with a proven edge.
- Comparing raw NVI levels across different stocks. The starting value is arbitrary and the calculation compounds over time, so absolute levels aren't meaningfully comparable between tickers.
- Ignoring PVI. Looking at NVI alone means only ever seeing what happened on lower-volume days, half of the same underlying logic is simply never examined.
- Assuming lower-volume days always represent informed activity. The "smart money" framing is a loose historical interpretation, not something the calculation itself can verify.
- Using an undocumented moving-average length and comparing results to research based on a different lookback. Different lengths produce different signals, and the difference isn't disclosed unless it's checked directly.
An Indicator That Only Watches the Quiet Days
The Negative Volume Index updates only on days when volume fell from the prior session, on the premise that informed participants act when the crowd is inactive. That premise is a piece of market folklore from an era of different market structure, and it is worth treating as a hypothesis rather than an established mechanism.
If you use it, use it on long horizons and on broad instruments. The indicator was developed and discussed in the context of index-level analysis over months, and applying it to individual securities on short timeframes asks it to do something the original reasoning does not support.
The mistake is reading it as a volume indicator in the ordinary sense. It ignores the price change on every high-volume day, which means it discards the sessions where the most trading occurred, and that is a deliberate feature rather than an oversight.
The underlying premise is also harder to defend in current markets than when it was proposed. Algorithmic execution deliberately spreads large orders across quiet periods, and the distinction between informed and uninformed participation is not reliably captured by the day's volume total.
NVI FAQs
What does NVI stand for?
Negative Volume Index. It's a cumulative indicator that only updates on days when volume falls from the prior day, tracking the percent price change on those lower-volume sessions specifically.
Is NVI a bullish or bearish indicator?
Neither by itself, NVI is just a running total. The traditional interpretation reads NVI above its own long-term moving average as bullish, but that reading comes from decades-old research and is not a proven predictive signal.
What is the difference between NVI and PVI?
They're mirror images built from the same close-to-close percent change. PVI only updates on days when volume rises from the prior day; NVI only updates on days when volume falls. Every session updates exactly one of the two lines, never both.
What is the standard NVI moving average setting?
NVI is traditionally compared to its own long-term moving average, commonly a roughly one-year lookback by analogy to the convention used for PVI. It isn't a rule enforced by the calculation itself, and traders should verify the exact period their charting platform uses before comparing readings.
Is the NVI-above-its-moving-average signal reliable?
It's a dated heuristic from Dysart and Fosback's original research, not a proven predictive signal. Historical correlation observed decades ago in one dataset does not guarantee the same relationship holds in current markets.
Does NVI work for cryptocurrency, or is it stock-specific?
The calculation itself works on any instrument with reported daily volume and closing price. The original research behind the traditional interpretation was done on equity market data, so applying the same signal to crypto's very different volume and volatility patterns hasn't been separately validated.
Why does the index only update on lower-volume days?
The construction rests on an idea that informed participants trade when volume is quiet while crowd activity dominates heavy-volume days, so restricting updates to lower-volume sessions is intended to isolate the former. Whether that separation holds is not established, and the premise dates from an era with a very different market structure. The mechanism is a hypothesis embedded in a formula rather than a measured effect.
What happens to the index during a long stretch of rising volume?
It stops updating entirely, holding its last value until a lower-volume session occurs. A chart therefore shows flat segments that represent absence of qualifying days rather than absence of price movement. Readers unaware of this can interpret a flat line as stability when it is simply a gap in the measurement.
Is this index useful on instruments where volume is fragmented across venues?
Its entire mechanism depends on classifying each session as higher or lower volume than the previous one, so an unreliable volume series makes that classification arbitrary. On instruments where reported volume varies by data source, the same chart can produce different index values depending on the feed. This limits it to instruments with a consolidated, dependable volume record.
References
- CMT Association: Technical Analysis Body of Knowledge and Research
- CFA Institute Research and Policy Center: Investment Research
- TA-Lib: Technical Analysis Function Library Documentation
- The NVI/PVI calculation and the moving-average comparison discussed on this page are associated with research by Paul Dysart and later Norman Fosback, a dated, debated heuristic from early volume-index research.
- SEC Investor.gov: Introduction to Investing