VaR, CVaR and max drawdown: how much can you lose?
What VaR, CVaR and maximum drawdown are, how they're calculated with simple examples and how to use them to see how much your portfolio could lose.
Oct 30, 2026 · 5 min read
Volatility is the most widely used risk measure, but it has a problem: it treats rises and falls the same. Nobody loses sleep because their portfolio rose 5% in a day. What worries people is losses. That's why there are measures that look only at the bad side: VaR, CVaR and maximum drawdown.
Maximum drawdown
Maximum drawdown is the worst loss from a peak to a trough during the period analyzed.
Illustrative example: your portfolio is worth 10,000, climbs to 12,000 and then drops to 8,400 before recovering. The maximum drawdown is:
(8,400 − 12,000) ÷ 12,000 = −30%
It's the most intuitive measure of all because it answers a very concrete question: how much would I have had to sit through without selling?
Some approximate historical references:
- The S&P 500 fell about 55% between 2007 and 2009, and about 34% in a few weeks in 2020.
- A broad US bond ETF fell about 18% between 2020 and 2022, its worst drop in decades.
- Many emerging markets have had drops of more than 60% measured in dollars.
Two things worth knowing:
- It depends on the period. If you analyze 3 years with no crisis, the maximum drawdown will look small. That doesn't mean something worse can't happen.
- Recovery costs more than the fall. If you lose 50%, you need to gain 100% to get back to where you started. If you lose 30%, you need 43%.
VaR: value at risk
The daily 95% VaR answers: what loss was not exceeded on 95% of days?
If your portfolio's daily 95% VaR is 1.8%, it means that on 95 out of 100 days the loss was smaller than 1.8%. On the other 5 days, it was worse.
To calculate it, the optimizer takes the portfolio's historical daily returns, sorts them from worst to best and finds the point that leaves out the worst 5% of days. It doesn't assume returns follow a "normal" curve: it uses what actually happened.
The VaR trap
VaR tells you where bad days begin, but not how bad they get. Two portfolios can have the same 1.8% VaR: in one, the worst days lose 2%; in the other, some days lose 10%. VaR can't tell them apart.
CVaR: how ugly the bad days are
The daily 95% CVaR (also called expected shortfall) is the average loss on that worst 5% of days.
Following the example:
| Portfolio A | Portfolio B | |
|---|---|---|
| Daily 95% VaR | 1.8% | 1.8% |
| Daily 95% CVaR | 2.4% | 4.1% |
Illustrative numbers. Both have the same VaR, but in B the bad days are much worse. That happens with assets prone to sharp drops: individual stocks, markets with capital controls or currencies that devalue suddenly.
By definition, CVaR is always greater than or equal to VaR. If CVaR is much higher than VaR, it's a sign of tail risk: rarely, but when things go wrong, they go very wrong.
Reading them together
Each measure looks at something different:
| Measure | Answers | Horizon |
|---|---|---|
| Volatility | How much does it move overall? | Annual |
| 95% VaR | Where does a bad day start? | Daily |
| 95% CVaR | How much do I lose, on average, on a bad day? | Daily |
| Maximum drawdown | What was the worst full losing streak? | Whole period |
A portfolio can have low volatility and still a large maximum drawdown, if it fell slowly over a long time. That's why it's worth looking at all of them.
Where to see them in QuantMundi
After calculating a portfolio in the optimizer, the metrics appear at the top of the results: expected return, volatility, Sharpe, maximum drawdown, VaR and CVaR. All are calculated in your currency, so they include the effect of the exchange rate.
An exercise:
- Load the "Latin America" example and calculate with Maximum Sharpe.
- Note the maximum drawdown, VaR and CVaR.
- Switch to Minimum variance and compare.
- Try the "Global diversified" example, which includes bonds and gold.
You'll see how a portfolio concentrated in one region has higher maximum drawdowns and CVaR than a diversified one, even if the expected return looks similar.
The limits of these measures
- They're historical. They measure what happened, not what will happen. The next crisis may be worse than any in the period analyzed.
- Daily VaR doesn't scale up easily. A 1.8% daily loss doesn't translate into a fixed monthly loss: losing streaks happen.
- They need enough history. With one year of data, the worst 5% of days is only 12 or 13 days. With more years, the measure is more reliable.
Frequently asked questions
What's the difference between VaR and CVaR? VaR shows the loss limit that wasn't exceeded on 95% of days. CVaR is the average loss on the 5% of days that did exceed it. That's why CVaR shows better how severe the bad days are.
What maximum drawdown is acceptable? It's up to you: it's the biggest loss you could watch without selling. As a reference, an all-stock global portfolio has fallen more than 50% in severe crises; adding bonds usually reduces that.
Why do banks use expected shortfall? Because it measures what happens beyond VaR. International banking regulators adopted it for market risk precisely for that advantage, which is also why it's worth checking in your own portfolio.
In short
Maximum drawdown tells you how much you'd have had to endure in the worst streak. VaR, where a bad day begins. CVaR, how bad that bad day is on average. Together, they give a much more honest picture of risk than volatility alone. Before investing, ask yourself whether you could live through the maximum drawdown without selling.
Sources and further reading
Try it with your own assets
Build a list and see which portfolio the model calculates, in your currency.
Disclaimer: Educational and simulation tool. It is not investment advice or a personalized recommendation. Past performance does not guarantee future results.