Volatility

1. Volatility

What is Risk?

Risk in finance refers to the uncertainty about future outcomes, particularly the potential for losing money on investments.
It’s essentially the uncertainty about future returns. Investors worry about how much their actual returns might differ from what they expect.

Measuring Uncertainty

Uncertainty might seem hard to quantify, but in finance, we often model it using probability distributions.
Think of the log return (which measures the continuous rate of return over a period) as a random variable, which can take different values with various probabilities.

Key Statistical Concepts

Mean (Expected Value):
- This is the average value you’d expect if you observed the log return over many periods.
- Mathematically, if $R$ is the random variable for log returns, the mean $μ$ is:

\mu = \mathbb{E}[R] = \frac{1}{n} \sum_{i=1}^{n} R_i

- Here, $\mathbb{E}[R]$ is the expected value, and $R_i$ are individual observations of log returns. 2. **Standard Deviation:** - This measures the spread of the distribution, indicating how much the returns deviate from the mean on average. - For a set of log returns, the standard deviation $\sigma$ is given by: $$ \sigma = \sqrt{\frac{1}{n} \sum_{i=1}^{n} (R_i - \mu)^2} $$ - This $\sigma$ is what we refer to as **volatility** in finance. #### **Volatility: A Measure of Risk** - **Volatility** is simply the standard deviation of the log returns. It represents the variability or spread of returns around the expected value. - **Interpreting Volatility:** - If returns are normally distributed: - **Mean (Expected Return):** Say 5% per year. - **Volatility:** Say 6% per year. - **Range of Likely Returns:** About 95% of the time, the returns will fall within $\mu \pm 2\sigma$ (i.e., between -7% and 17%). #### **Importance of Volatility** 1. **Risk Assessment:** - Investors use volatility to understand how much their returns might fluctuate. Higher volatility means higher risk. 2. **Portfolio Management:** - Traders use volatility to determine the size of positions in a portfolio. Instruments with higher volatility might require smaller position sizes to manage risk. 3. **Alpha Strategies:** - In trading strategies, volatility can help decide when to buy or sell certain stocks. 4. **Options Pricing:** - Volatility is a crucial factor in determining the price of options, as it affects the likelihood of the option being profitable. 5. **Volatility Trading:** - Some instruments, like those tracking the volatility of the S&P 500 (e.g., VIX), allow traders to speculate directly on volatility. ### Summary Volatility is a key metric in finance that captures the variability of returns, offering insights into the risk and potential range of outcomes for investments. It's a versatile tool used by traders, portfolio managers, and investors to make informed decisions in the financial markets. ## 2. Historical Volatility ### Estimating Volatility: A Step-by-Step Guide #### **Historical Volatility** To estimate the future volatility of a stock, one common approach is to calculate **historical volatility**. This involves using past price data to measure how much the stock's returns have varied over time. Here's how you can do it: #### **Step 1: Calculate Log Returns** - **Log Return Formula:** - The log return between two consecutive prices $P_t$ and $P_{t-1}$ is calculated as: $$ R_t = \ln\left(\frac{P_t}{P_{t-1}}\right) $$ - Here, $R_t$ is the log return at time $t$. - **Example:** - If you have price data for $n + 1$ days, you will calculate $n$ log returns. The first price point doesn't have a preceding value, so the first log return is undefined. #### **Step 2: Calculate Standard Deviation of Log Returns** - **Volatility Formula:** - Volatility $\sigma$ is simply the standard deviation of the log returns $R_t$. The formula is: $$ \sigma = \sqrt{\frac{1}{n} \sum_{t=1}^{n} \left(R_t - \mu\right)^2} $$ - Where $\mu$ is the mean of the log returns. - **Interpretation:** - This calculated $\sigma$ gives you the **daily volatility** if your log returns were based on daily prices. #### **Example Calculation:** - Suppose you perform this calculation on daily data and find $\sigma = 0.025$. This is your estimate of the **daily volatility**. #### **Step 3: Understanding Time Frequency and Volatility** - **Impact of Time Frequency:** - If you use weekly or monthly prices instead of daily prices, the calculated volatility will be different. This is because price changes tend to be larger over longer time intervals (e.g., a week vs. a day). - For example: - **Daily Volatility:** Typically ranges from 0.006 to 0.03. - **Weekly Volatility:** Typically ranges from 0.01 to 0.07. #### **Step 4: Annualizing Volatility** - **Why Annualize?** - To make volatility comparable across different datasets (e.g., daily vs. weekly), it's standard practice to annualize the volatility. - **Annualized Volatility Formula:** - If $\sigma_{\text{daily}}$ is the daily volatility, the annualized volatility $\sigma_{\text{annual}}$ can be calculated as: $$ \sigma_{\text{annual}} = \sigma_{\text{daily}} \times \sqrt{T} $$ - Here, $T$ is the number of trading days in a year, typically taken as 252. - Similarly, if you have weekly volatility $\sigma_{\text{weekly}}$, you would multiply by $\sqrt{52}$ to annualize it. ### Summary - **Historical Volatility** is calculated as the standard deviation of log returns based on past price data. - The **time frequency** of your data (daily, weekly, monthly) affects the calculated volatility. - **Annualized Volatility** is a standardized way to express volatility over a year, allowing for comparison across different datasets. In the next lesson, we’ll dive deeper into how to calculate and interpret annualized volatility. ## 3. Rolling Windows ### 1. **Volatility Measurement** - **Volatility** is typically measured as the standard deviation of returns. It gives us an idea of how much prices deviate from the average, indicating the level of risk or uncertainty in the market. ### 2. **Market Dynamics** - Markets are dynamic, with periods of high and low volatility. For instance, the 2008 financial crisis saw extreme market fluctuations, whereas 2016-2017 experienced relatively low volatility. ### 3. **Rolling Window Approach** - **Rolling Window**: This is a technique used to calculate volatility over time. To estimate today's volatility, you calculate the standard deviation of returns from a set period up to yesterday. - **Window Length**: The length of the window is crucial. A longer window captures more historical data, leading to a smoother but less responsive volatility estimate. A shorter window is more responsive to recent changes but can be too volatile. ### 4. **Choosing the Window Length** - **Application-Dependent**: The ideal window length depends on your strategy. - **Long-term Strategies**: If your strategy involves holding assets for a long period, a longer window is suitable, as it smooths out short-term fluctuations. - **Short-term Strategies**: For short-term trading, a shorter window is preferred to capture quick changes in market conditions. ### 5. **Multiple Window Sizes** - To gain deeper insights, you can use multiple window sizes. For example: - **Short-term Window**: Captures immediate market reactions. - **Medium-term Window**: Provides a balance between responsiveness and stability. - **Long-term Window**: Offers a broader view of market trends. ### 6. **Mathematical Perspective** - Suppose $R_t$ represents the log return at time $t$. The volatility $\sigma_t$ over a rolling window of length $n$ can be calculated as:

\sigma_t = \sqrt{\frac{1}{n-1} \sum_{i=0}^{n-1} (R_{t-i} - \bar{R})^2}

where $\bar{R}$ is the mean of the returns over the window. By adjusting the window size, you can control the balance between responsiveness to new data and stability of the volatility estimate, depending on the nature of your investment strategy. ## 4. Exponentially Weighted Moving Average To estimate today's volatility while giving more importance to recent data, you can use an **Exponentially Weighted Moving Average (EWMA)**. This method allows you to weigh recent log returns more heavily compared to older ones, which is often more reflective of the current market conditions. Let's break down the process of calculating volatility using EWMA. ### Step 1: Understanding Historical Volatility Suppose you're using a window of `n` days to estimate today's volatility, denoted by $\sigma_t$. The standard formula for historical volatility involves calculating the variance, $\sigma_t^2$, as the average of the squared log returns over the past `n` days:

\sigma_t^2 = \frac{1}{n} \sum_{i=1}^{n} (r_{t-i} - \mu)

Here: - $r_{t-i}$ is the log return on day $t-i$, - $\mu$ is the mean log return over the period. ### Step 2: Simplifications 1. **Assume the Mean Log Return ($\mu$) is Zero**: For short intervals like daily log returns, the mean is typically small compared to the standard deviation. Thus, we simplify:

\sigma_t^2 = \frac{1}{n} \sum_{i=1}^{n} r_{t-i}

2. **Approximate $n-1$ by $n$**: For large `n`, this change has little impact on the result, allowing us to simplify further:

\sigma_t^2 \approx \frac{1}{n} \sum_{i=1}^{n} r_{t-i}

This is a straightforward arithmetic average, where each squared log return is weighted equally. ### Step 3: Introducing the Exponential Weighting To give more weight to recent observations, we introduce a parameter $\lambda$ (where \$0 < \lambda < 1$). The weight for the log return $r_{t-i}^2$ is then $\lambda^i$. The weighted sum becomes:

\sigma_t^2 = \frac{\sum_{i=0}^{n-1} \lambda^i r_{t-i}^2}{\sum_{i=0}^{n-1} \lambda^i}

Where: - $r_t^2$ is the squared log return for yesterday, - $r_{t-1}^2$ is the squared log return for the day before yesterday, and so on. ### Step 4: Calculate the Weights The weights $\lambda^i$ decrease exponentially as you go back in time. For example: - The weight for yesterday's log return: $\lambda^0 = 1$ - The weight for the day before: $\lambda^1$ - The weight for the day before that: $\lambda^2$ ### Step 5: Normalizing the Weights To ensure that the weighted sum represents a proper average, divide by the sum of the weights:

\text{Sum of weights} = \sum_{i=0}^{n-1} \lambda^i = \frac{1 - \lambda^n}{1 - \lambda}

T h u s, t h e E W M A v a r ian cees t ima t e i s :

\sigma_t^2 = \frac{\sum_{i=0}^{n-1} \lambda^i r_{t-i}^2}{\frac{1 - \lambda^n}{1 - \lambda}}

### Step 6: Estimate Volatility Finally, to get the volatility, take the square root of the variance:

\sigma_t = \sqrt{\frac{\sum_{i=0}^{n-1} \lambda^i r_{t-i}^2}{\frac{1 - \lambda^n}{1 - \lambda}}}

### Summary - **Exponentially Weighted Moving Average (EWMA)** assigns more importance to recent data by weighting log returns exponentially. - **Parameter $\lambda$** determines how quickly the weights decrease (a smaller $\lambda$ means faster decay). - This approach gives you a volatility estimate that adapts more quickly to recent market conditions compared to traditional methods. This method is particularly useful in finance when you need to account for the changing volatility in asset prices. ## 5. Forecasting Volatility Forecasting market volatility is a crucial aspect of trading, and while predicting prices can be challenging, volatility tends to show a more consistent pattern. This consistency allows traders to anticipate market conditions and make informed decisions. Let's dive into the concept of modeling volatility using **ARCH (Autoregressive Conditionally Heteroscedastic)** and **GARCH (Generalized ARCH)** models. ### 1. Understanding Volatility and "Sticky" Behavior Volatility refers to the degree of variation in asset prices over time. In financial markets, volatility is often "sticky," meaning that high volatility periods tend to be followed by more high volatility, and low volatility periods are often followed by low volatility. This makes volatility somewhat predictable, unlike prices, which can be more random. ### 2. Autoregressive Models: The Basis of ARCH - **Autoregressive**: In an autoregressive model, the current value of a variable is related to its past values. For example, today's volatility could be influenced by the volatility of the previous days. - **Heteroscedasticity**: This means that the variability (or variance) of a variable isn't constant but changes over time. In financial markets, some periods are more volatile than others. - **Conditional**: This refers to the idea that the variance (volatility) at a given time is dependent on past information, like previous log returns. ### 3. ARCH Model: Predicting Variance Based on Past Returns The **ARCH model** is used to model and predict the variance (volatility) of a time series based on past squared returns. Here's how it works: #### **ARCH(1) Model**: For a log return $r_t$ at time $t$, the variance $\sigma_t^2$ is given by:

\sigma_t^2 = \alpha_0 + \alpha_1 r_{t-1}

- $\alpha_0$: Baseline variance, representing the minimum level of variance that is always present. - $\alpha_1$: Weight that determines how much of yesterday’s squared return $r_{t-1}^2$ affects today's variance. This model assumes that today's variance depends only on the squared log return of the previous day. #### **ARCH(m) Model**: This is an extension where the variance depends on the last $m$ squared log returns:

\sigma_t^2 = \alpha_0 + \alpha_1 r_{t-1}^2 + \alpha_2 r_{t-2}^2 + \ldots + \alpha_m r_{t-m}

Here, $\alpha_1, \alpha_2, \ldots, \alpha_m$ are parameters that weight the influence of each past squared return. ### 4. GARCH Model: Incorporating Past Variances The **GARCH (Generalized ARCH)** model extends the ARCH model by including terms for past variances. The idea is that today's variance can depend not only on past squared returns but also on past variances. #### **GARCH(1,1) Model**: In its simplest form, the GARCH(1,1) model is given by:

\sigma_t^2 = \alpha_0 + \alpha_1 r_{t-1}^2 + \beta_1 \sigma_{t-1}

- $\alpha_0$: Baseline variance. - $\alpha_1$: Weight on the squared return from the previous day. - $\beta_1$: Weight on the previous day's variance. This model captures the idea that today's variance is influenced both by yesterday's squared return and by yesterday's variance. The GARCH model is often more realistic than the ARCH model because it accounts for the "persistence" of volatility. #### **GARCH(m,n) Model**: This is a more general version where the variance depends on $m$ past squared returns and $n$ past variances:

\sigma_t^2 = \alpha_0 + \sum_{i=1}^{m} \alpha_i r_{t-i}^2 + \sum_{j=1}^{n} \beta_j \sigma_{t-j}

- $\alpha_i$: Parameters for the past squared returns. - $\beta_j$: Parameters for the past variances. ### 5. Practical Applications: Using ARCH/GARCH in Trading Once you've built and calibrated an ARCH or GARCH model, you can use it to predict future volatility and make trading decisions. Here are some strategies: - **Volatility-Based Signals**: Generate buy or sell signals based on predicted volatility levels. For example, you might decide to enter a trade if predicted volatility exceeds a certain threshold. - **Strategy Activation**: Some trading strategies perform better in high volatility conditions. You can use your volatility predictions to activate these strategies only when high volatility is expected. - **Risk Management**: Adjust your position sizes based on predicted volatility. Higher volatility may lead to smaller positions to manage risk, while lower volatility could allow for larger positions. ### Summary - **ARCH** models are used to predict future variance (volatility) based on past squared returns. - **GARCH** models extend ARCH by including past variances in the prediction, capturing the persistence of volatility. - These models are valuable tools in finance, providing insights into market behavior and helping traders make more informed decisions based on expected volatility. ## 6. Markets Volatility Volatility in financial markets is a complex phenomenon that isn't solely driven by new information reaching the market, as you might intuitively think. While news can cause price fluctuations, research suggests that volatility is significantly influenced by the trading process itself. Let's break down this concept and explore the implications for trading strategies and the role of volatility indices like the VIX. ### 1. The Source of Volatility: Trading vs. Information #### **The Information Hypothesis** A natural assumption is that volatility arises when new information causes traders to reassess the value of a stock. For example, if a company releases a surprising earnings report, the stock price may spike or plummet as investors adjust their expectations. This kind of volatility is driven by external news and events. #### **The Trading Hypothesis** However, studies have shown that the connection between new information and volatility is not as strong as one might expect. Research comparing the variance of stock price returns between consecutive trading days and across weekends has found that weekend variance isn't as high as the information hypothesis would suggest. For example, you might expect the weekend variance to be three times the midweek variance because of the longer time period, but in reality, it's often only about 1.5 times higher. This suggests that much of the volatility is actually generated by trading activity itself. #### **Volatility and Volume Correlation** Volatility is frequently correlated with trading volume--when more shares are traded, volatility tends to increase. This correlation implies that the act of trading, rather than just new information, plays a key role in driving price fluctuations. In times of high trading volume, prices can be pushed up due to increased demand, but in the absence of new information, these price changes might not reflect a real change in the company's value. This leads to the possibility of prices reverting back to their mean, which is the basis for certain trading strategies. ### 2. Trading Strategies Based on Volatility Given the nature of volatility, traders can adapt their strategies depending on the current market conditions: #### **Mean Reversion Strategies** - **High Volatility**: When volatility is high, prices often deviate significantly from their average (or "mean"). Since these deviations are not necessarily driven by new information, they may be temporary. **Mean reversion strategies** capitalize on this by betting that prices will eventually return to their average level. This strategy is particularly effective during volatile periods when prices are more likely to oscillate around a mean. #### **Momentum Strategies** - **Low Volatility**: In periods of low volatility, prices tend to trend more smoothly. **Momentum strategies** are based on the idea that stocks that are moving in one direction (up or down) are likely to continue moving in that direction. These strategies work better in low-volatility environments where trends are less likely to be disrupted by sudden, large price movements. ### 3. Market Volatility and the VIX Index Volatility often behaves asymmetrically with respect to market movements: - **High Volatility in Down Markets**: Volatility tends to spike when the market is falling, as fear and uncertainty drive rapid selling. - **Low Volatility in Up Markets**: When the market is rising, volatility is usually lower, as optimism and steady buying prevail. Because of this behavior, volatility is often seen as a "gauge of fear" in the market. #### **The VIX Index** - The **VIX Index**, also known as the "fear gauge," measures the 30-day expected volatility of the S&P 500 Index. It is calculated based on the prices of S&P 500 Index options and represents the market's expectation of future volatility. - When the S&P 500 dips, the VIX often spikes, reflecting increased market uncertainty and risk. For example, during the 2008 financial crisis, the VIX reached a record high of 80, indicating extreme volatility as markets were in turmoil. #### **Trading the VIX** - Just as with other financial instruments, the VIX itself has become a tradable asset. You can trade VIX futures and options, essentially making bets on the future volatility of the S&P 500. - There are also other volatility indices published by the Chicago Board Options Exchange (CBOE), covering various assets like commodities, currencies, and even specific stocks. - Interestingly, there's even a **VVIX** (Volatility of Volatility Index), which measures the volatility of the VIX itself, adding another layer of complexity for traders who specialize in volatility. ### Summary Volatility in the stock market isn't solely driven by new information but is also significantly influenced by trading activity. This insight leads to different trading strategies: - **Mean reversion strategies** thrive in high volatility environments, betting on price corrections. - **Momentum strategies** work better in low volatility periods, capitalizing on ongoing trends. The VIX Index serves as a key measure of market sentiment, especially fear, and can be traded like any other financial asset, offering opportunities for those who specialize in volatility trading. ## 7. Using Volatility For Equity Trading Volatility is a key metric in finance, representing the degree of variability in asset returns over time. It's versatile and can be applied in numerous trading strategies, portfolio management decisions, and risk assessments. Let's explore some of the ways volatility can be utilized in trading, along with a deeper dive into specific strategies and practical applications. ### 1. **Classifying Stocks by Volatility** One of the simplest uses of volatility is to classify stocks into high or low volatility categories. This classification can guide trading strategies: - **Low Volatility Stocks**: - **Mean Reversion Strategy**: In this strategy, traders take a short position when an asset's price rises significantly, betting that the price will revert to its mean. Low volatility stocks, with less uncertainty around their fair value, are more likely to revert after a large deviation. - **Performance Mystery**: Interestingly, low volatility stocks often outperform high volatility stocks, which contradicts the traditional risk-return relationship. This could be because investors are drawn to more "exciting" high volatility stocks, overlooking the steady, reliable returns of low volatility ones. - **ETFs for Low Volatility**: Some exchange-traded funds (ETFs) focus specifically on low volatility stocks, optimizing portfolios to include these steady performers. ### 2. **Normalizing Metrics Using Volatility** Volatility can also be used to normalize other financial metrics, making comparisons across different assets more meaningful: - **Momentum Signal Normalization**: - Suppose you have a momentum signal based on a stock's returns over the past year. If both Netflix and Walmart have a 30% return, their performance might seem identical. However, because Netflix is more volatile, its 30% return might not be as significant as Walmart's. - To account for this, you can normalize the momentum signal by dividing it by the stock's volatility. This adjustment helps compare stocks more accurately, ensuring an "apples to apples" comparison. This method is broadly applicable across many signals and metrics, improving the consistency and reliability of analyses. ### 3. **Determining Position Sizes Based on Volatility** Volatility plays a crucial role in determining the size of trading positions: - **Volatility-Based Position Sizing**: - Traders often reduce position sizes in more volatile markets to minimize potential losses. A common approach involves a formula like this:

\text{Position Size} = \frac{R}{\sigma \times M \times \text{Last Close}}

ByteCrypt

Explorer

Volatility

1. Volatility

What is Risk?

Measuring Uncertainty

Key Statistical Concepts

Graph View

Backlinks