Kuantile
What Is Value at Risk (VaR)? A Plain-Language Guide
You check your portfolio's value every day — but do you know how much you could plausibly lose tomorrow? In institutional finance the answer to that question is Value at Risk, or VaR. From bank treasury desks to pension funds, professionals define their risk limits with this metric; Kuantile runs the same calculation for your portfolio.
What VaR says in one sentence
"1-day VaR at 99% confidence = 12,000 TRY" means: under normal market conditions, on 99 out of 100 days your daily loss will not exceed 12,000 TRY. Read in reverse: roughly one day in a hundred, it may. VaR is not a forecast — it is a boundary drawn from the historical distribution of returns.
The historical simulation method
There are several ways to compute VaR; Kuantile uses the most transparent one, historical simulation:
- Take each asset's historical daily returns, in TRY terms (currency effect included).
- Apply today's portfolio weights to those returns — producing thousands of "what your portfolio would have done" days.
- Sort the results; for 99% confidence, the boundary of the worst 1% is your VaR.
The method's virtue is that it assumes no distribution: the 2018 lira shock and the 2020 Covid crash enter the calculation exactly as violently as they happened.
95% or 99%?
The 95% level draws a calmer line — expected to be breached about once in 20 days; 99% about once in 100. A short-term trader can manage daily swings with 95%; a long-term saver should look at 99% to see tail risk. Kuantile lets you switch between the two with one click.
What VaR does not tell you
VaR is a threshold; it says nothing about what lies beyond it. On that 1-in-100 bad day the loss may sit just past the line or far deeper. That is why VaR should be read together with stress tests, which show what your portfolio would actually endure in history's known disaster windows.
A practical habit
Experienced investors track VaR as a ratio and a trend rather than an absolute number: what share of the portfolio is at risk, and is it growing? Kuantile's weekly reports state automatically whether your VaR rose or fell versus the start of the period — add a volatile asset and you will see its risk bill in the next report.
FHS: volatility-aware VaR
Historical simulation has one weakness: if it measures a turbulent day using data from a calm year, it can understate risk. That is why Kuantile also uses a second method — filtered historical simulation (FHS). Here past returns are rescaled to today's market volatility; RiskMetrics' exponentially weighted volatility model (EWMA, λ=0.94) gives more weight to recent days. The result is a VaR tuned to "today's temperature": the figure rises automatically when markets are stressed and falls when they calm down. The headline VaR in the app is this FHS value; the plain historical VaR sits beside it as a reference.
Backtesting your VaR
You only find out whether a risk model works by backtesting it. The logic is simple: if the 99% VaR is correct, actual daily loss should exceed that line only about 1% of the time. Kuantile counts these "breaches" in your portfolio's historical returns and evaluates the model's calibration with Basel traffic-light logic (green/amber/red). Too many breaches means the model underestimates risk; zero breaches means it is overly conservative. This transparency answers the question "why should I trust this number?"
Beyond the tail: Expected Shortfall
VaR tells you the boundary, but not the average loss beyond it. Expected Shortfall (ES) fills that gap: "in the worst 1% of days, how much do I lose on average?" It is usually meaningfully larger than VaR, and regulators (Basel) have shifted toward ES over VaR in recent years. Kuantile shows both figures side by side; the size of the ES/VaR ratio reveals how "fat" your portfolio's tail is — that is, how exposed you are to rare but severe losses.
Compute your portfolio's VaR →