Skip to main content
diversification.com
← Back to D Definitions

Discrete time models

What Is Discrete Time Models?

Discrete time models in quantitative finance represent financial processes and asset prices as evolving in distinct, measurable steps or intervals, rather than continuously over time. These models view financial phenomena at specific points in time, such as daily, weekly, or monthly intervals, making them particularly useful for analyzing data that is collected and reported periodically. Within the broader field of financial modeling, discrete time models offer a computationally tractable way to approximate complex market behaviors and are foundational for many numerical methods in areas like option pricing and risk management.

History and Origin

The concept of modeling financial phenomena in discrete steps has roots in early efforts to understand market dynamics. While continuous time models gained prominence with the development of sophisticated stochastic calculus, discrete time models often preceded or ran in parallel with them, particularly due to their computational simplicity and alignment with real-world data collection.

A pivotal moment in the history of discrete time models in finance was the introduction of the binomial option pricing model. This model, formalized by Cox, Ross, and Rubinstein in 1979, provided a simplified, yet powerful, framework for valuing options by assuming that the underlying asset price could only move to one of two possible values (up or down) over a given time step9. This approach offered a clear, intuitive method for understanding option valuation, and its discrete nature made it readily implementable even with limited computational resources. Early financial models were often constructed manually on worksheets, with the advent of spreadsheets like VisiCalc and Lotus 1-2-3 later facilitating more complex discrete calculations8.

Key Takeaways

  • Discrete time models represent financial variables and processes as changing at specific, separated points in time.
  • They are often computationally simpler and more intuitive for practical application, particularly for modeling scenarios with distinct decision points.
  • The binomial option pricing model is a prominent example of a discrete time model widely used for valuing derivative securities.
  • While they can approximate continuous processes, discrete time models might introduce inaccuracies if the time steps are too large relative to the underlying continuous dynamics.
  • Many real-world financial data are inherently discrete, making discrete time models a natural fit for analysis and simulation.

Formula and Calculation

A classic example of a discrete time model with a clear formula is the single-period binomial option pricing model. This model calculates the value of an option by considering two possible outcomes for the underlying asset's price at the end of a single time period: an upward movement or a downward movement.

The core idea is to construct a portfolio of the underlying asset and a risk-free rate bond that replicates the option's payoff at expiration. By the principle of arbitrage-free pricing, the current value of this replicating portfolio must equal the option's value.

For a call option, the current value (C_0) can be derived using the risk-neutral probability (p^*):

C0=p∗Cu+(1−p∗)CderΔtC_0 = \frac{p^* C_u + (1-p^*) C_d}{e^{r \Delta t}}

Where:

  • (C_0) = Current value of the call option
  • (C_u) = Call option value if the underlying asset price moves up
  • (C_d) = Call option value if the underlying asset price moves down
  • (r) = Risk-free rate (continuously compounded)
  • (\Delta t) = Length of the time step in years
  • (p^*) = Risk-neutral probability of an upward movement, calculated as: p∗=erΔt−du−dp^* = \frac{e^{r \Delta t} - d}{u - d} Where (u) is the upward multiplier and (d) is the downward multiplier for the asset price.

This formula demonstrates how the expected value of future payoffs, adjusted for the discount rate, determines the present value of the option in a discrete setting. The model can be extended to multiple periods by constructing a binomial tree and working backward from the option's expiration date.

Interpreting the Discrete Time Models

Interpreting discrete time models involves understanding that the underlying financial process is viewed as a series of distinct states or events occurring at specified intervals. For example, in the context of stock prices, a discrete time model might assume that the price can only change at the end of each trading day, rather than fluctuating continuously throughout the day. This interpretation aligns well with how much financial data, such as daily closing prices or quarterly earnings, is reported and analyzed.

The values generated by discrete time models, such as option prices or projected portfolio values, are typically snapshots at these discrete points. When evaluating a model's output, it is crucial to consider the granularity of the time steps. A model with monthly steps will provide a coarser view than one with daily or hourly steps. The practical application of discrete time models often involves simulating numerous paths through these discrete steps to derive probabilities or expected outcomes, which are then used for decision-making or valuation.

Hypothetical Example

Consider a simple discrete time model for valuing a European option on a stock that currently trades at $100. We'll assume a single time step of one year, a risk-free rate of 5%, and that the stock price can either go up by 20% or down by 10% over the year. The option has a strike price of $105.

  1. Calculate possible stock prices at year-end:

    • Upward movement: $100 * (1 + 0.20) = $120
    • Downward movement: $100 * (1 - 0.10) = $90

  2. Calculate option payoffs at year-end (for a call option):

    • If stock price is $120: Call payoff = max($120 - $105, 0) = $15
    • If stock price is $90: Call payoff = max($90 - $105, 0) = $0

  3. *Calculate the risk-neutral probability (p)**:

    • Here, (u = 1.20) and (d = 0.90).
    • p∗=e0.05×1−0.901.20−0.90=1.05127−0.900.30=0.151270.30≈0.5042p^* = \frac{e^{0.05 \times 1} - 0.90}{1.20 - 0.90} = \frac{1.05127 - 0.90}{0.30} = \frac{0.15127}{0.30} \approx 0.5042
    • Thus, (1 - p^* \approx 0.4958).

  4. Calculate the present value of the option:

    • Using the risk-neutral probability: C0=(0.5042×$15)+(0.4958×$0)e0.05×1=$7.5631.05127≈$7.19C_0 = \frac{(0.5042 \times \$15) + (0.4958 \times \$0)}{e^{0.05 \times 1}} = \frac{\$7.563}{1.05127} \approx \$7.19
    • According to this discrete time model, the fair value of the call option is approximately $7.19.

Practical Applications

Discrete time models are widely employed across various domains of finance due to their computational tractability and alignment with how much financial data is generated and analyzed. Their practical applications include:

  • Option and Derivative Valuation: The binomial option pricing model is a cornerstone for valuing options, especially American options which can be exercised at any time before expiry. Its step-by-step nature allows for the incorporation of early exercise decisions. Practitioners often use variations of this model due to its ability to handle complex conditions that other models may not easily apply to, particularly for longer-dated options or those with dividend payments.
  • Financial Planning and Budgeting: Companies and individuals use discrete time models to project future cash flows, revenues, and expenses on a monthly, quarterly, or annual basis. These models are essential for creating budgets, financial forecasts, and long-term strategic plans.
  • Risk Management and Stress Testing: Firms can model potential financial outcomes under various discrete scenarios (e.g., economic downturns, market shocks) to assess portfolio vulnerability and inform risk management strategies.
  • Algorithmic Trading: In high-frequency trading, while continuous-time theory may inform strategy, the actual execution and rebalancing occur at discrete intervals, often corresponding to tick-by-tick price changes. Models used here need to account for the discrete nature of market data6, 7.
  • Policy Analysis: Central banks and regulatory bodies might use discrete time models to analyze the potential impacts of policy changes on economic variables over specific periods, helping to inform monetary policy decisions5.

Limitations and Criticisms

While discrete time models offer practical advantages, they also come with limitations and criticisms:

  • Approximation of Reality: A primary criticism is that financial markets operate continuously, with prices changing moment-to-moment. Discrete time models, by definition, simplify this continuous reality into distinct steps. While these models can approximate continuous processes by making the time steps infinitesimally small, they may introduce inaccuracies, especially when state transition probabilities are high4.
  • Computational Intensity for Fine Granularity: To better approximate continuous processes, discrete time models require an increasing number of steps. This can lead to significant computational complexity, particularly for multi-period models or those with many underlying variables. For example, the worst-case runtime for the binomial option pricing model can be exponential with the number of time steps.
  • Less Mathematical Elegance: Compared to their continuous time counterparts, discrete time models can sometimes be less elegant from a purely mathematical perspective, particularly for deriving analytical solutions3. Continuous time models often allow for the application of advanced stochastic calculus, leading to more compact and general formulas.
  • Parameter Sensitivity: The accuracy of discrete time models can be sensitive to the chosen time step size and the parameters defining the asset price movements (e.g., volatility and up/down factors). Incorrect parameter calibration can lead to mispricing or inaccurate forecasts.
  • Inability to Capture Intra-Period Dynamics: By only considering changes at discrete intervals, these models may overlook important intra-period dynamics, such as rapid price fluctuations or unexpected events that occur between the defined time steps. This can be a concern for very high-frequency analyses.

Discrete Time Models vs. Continuous Time Models

The fundamental distinction between discrete time models and continuous time models lies in their perception of time. Discrete time models assume that financial variables change at specific, distinct points in time (e.g., daily, monthly). This aligns with how most financial data is observed and recorded, making them often easier to implement computationally and interpret for periodic reporting. Examples include the binomial option pricing model and most traditional financial planning spreadsheets.

In contrast, continuous time models assume that financial variables evolve smoothly and constantly, with changes occurring at every instant. These models typically employ advanced mathematical tools like stochastic differential equations and are favored in theoretical quantitative finance for their mathematical elegance and ability to provide analytical solutions, such as the Black-Scholes model for European options. While continuous time models offer a more complete theoretical description of market processes, their practical application often requires discretization for computational purposes, effectively translating them into a discrete framework for simulation or numerical solution1, 2. The choice between the two often depends on the specific problem, available data, and the desired level of mathematical rigor versus computational practicality.

FAQs

What is the main difference between discrete and continuous time in finance?

The main difference is how time is treated. In discrete time, events and changes occur at distinct, separated points (e.g., end of day). In continuous time, events and changes are assumed to happen constantly and smoothly at every infinitesimal moment.

Why are discrete time models used if markets are often considered continuous?

Discrete time models are used because real-world financial data is often collected at discrete intervals (e.g., daily stock prices, quarterly earnings). They are also computationally simpler for many problems, making them practical for valuation, simulation, and financial planning, especially for instruments like American options that allow for early exercise.

What is an example of a discrete time model?

The most well-known example in finance is the binomial option pricing model, which models the price of an underlying asset as moving up or down to one of two possible values over a given time step.

Are discrete time models less accurate than continuous time models?

Not necessarily less accurate in all contexts. While continuous time models may offer more theoretical elegance for some problems, discrete time models can provide very close approximations, especially when the time steps are made sufficiently small. For certain applications, such as valuing options with complex features, discrete time numerical methods can be more versatile or more directly applicable to observed data.

What mathematical tools are typically used with discrete time models?

Discrete time models often rely on probability theory, difference equations, and numerical methods like binomial trees, lattices, and Monte Carlo simulations. Concepts such as expected value and stochastic processes are fundamental to their construction and analysis.