{"id":510,"date":"2014-11-28T09:45:49","date_gmt":"2014-11-28T09:45:49","guid":{"rendered":"http:\/\/rovdownloads.com\/blog\/?p=510"},"modified":"2014-11-28T09:45:49","modified_gmt":"2014-11-28T09:45:49","slug":"stochastic-forecasting","status":"publish","type":"post","link":"https:\/\/rovdownloads.com\/blog\/stochastic-forecasting\/","title":{"rendered":"Stochastic Forecasting"},"content":{"rendered":"

Theory<\/strong>
\nA stochastic process is nothing but a mathematically defined equation that can create a series of outcomes over time, outcomes that are not deterministic in nature; that is, an equation or process that does not follow any simple discernible rule such as price will increase X% every year or revenues will increase by this factor of X plus Y%. A stochastic process is by definition nondeterministic, and one can plug numbers into a stochastic process equation and obtain different results every time. For instance, the path of a stock price is stochastic in nature, and one cannot reliably predict the exact stock price path with any certainty. However, the price evolution over time is enveloped in a process that generates these prices. The process is fixed and predetermined, but the outcomes are not. Hence, by stochastic simulation, we create multiple pathways of prices, obtain a statistical sampling of these simulations, and make inferences on the potential pathways that the actual price may undertake given the nature and parameters of the stochastic process used to generate the time series. Four stochastic processes are included in Risk Simulator\u2019s Forecasting tool, including Geometric Brownian motion or random walk, which is the most common and prevalently used process due to its simplicity and wide-ranging applications. The other three stochastic processes are the mean-reversion process, jump-diffusion process, and a mixed process.<\/p>\n

The interesting thing about stochastic process simulation is that historical data is not necessarily required; that is, the model does not have to fit any sets of historical data. Simply compute the expected returns and the volatility of the historical data or estimate them using comparable external data, or make assumptions about these values.<\/p>\n

Procedure<\/p>\n

  • Start the module by selecting Risk Simulator | Forecasting | Stochastic Processes.<\/li>\n
  • Select the desired process, enter the required inputs, click on Update Chart a few times to make sure the process is behaving the way you expect it to, and click OK (Figure 1).<\/li>\n

    \"Untitled\"<\/p>\n

    Results Interpretation<\/strong>
    \nFigure 2 shows the results of a sample stochastic process. The chart shows a sample set of the iterations while the report explains the basics of stochastic processes. In addition, the forecast values (mean and standard deviation) for each time period is provided. Using these values, you can decide which time period is relevant to your analysis and set assumptions based on these mean and standard deviation values using the normal distribution. These assumptions can then be simulated in your own custom model.<\/p>\n

    \"snapshot\"<\/p>\n

    Notes<\/strong>
    \nBrownian Motion Random Walk Process<\/strong>
    \nThe Brownian motion random walk process takes the form of
    \nError! Objects cannot be created from editing field codes.<\/strong>
    \nfor regular options simulation, or a more generic version takes the form of
    \nError! Objects cannot be created from editing field codes.<\/strong>
    \nfor a geometric process.<\/p>\n

    For an exponential version, we simply take the exponentials, and as an example, we have
    \nError! Objects cannot be created from editing field codes.<\/strong><\/p>\n

    where we define
    \nS as the variable\u2019s previous value
    \n\u03b4S as the change in the variable\u2019s value from one step to the next
    \n\u00b5 as the annualized growth or drift rate
    \n\u03c3 as the annualized volatility<\/p>\n

    To estimate the parameters from a set of time-series data, the drift rate and volatility can be found by setting \u00b5 to be the average of the natural logarithm of the relative returns
    \nError! Objects cannot be created from editing field codes.<\/strong>
    \nwhile \u03c3 is the standard deviation of all<\/p>\n

    \"Untitled2\"<\/p>\n

    values.<\/p>\n

    Mean-Reversion Process<\/strong>
    \nThe following describes the mathematical structure of a mean-reverting process with drift:
    \nError! Objects cannot be created from editing field codes.<\/strong>
    \nTo obtain the rate of reversion and long-term rate, using the historical data points, run a regression such that
    \nError! Objects cannot be created from editing field codes.<\/strong>
    \nand we find
    \nError! Objects cannot be created from editing field codes. and Error! Objects cannot be created from
    \nediting field codes.<\/strong><\/p>\n

    where we define
    \n\u03b7 as the rate of reversion to the mean
    \nError! Objects cannot be created from editing field codes.<\/strong> as the long-term value the process reverts to
    \nY as the historical data series
    \n\u03b2 as the intercept coefficient in a regression analysis
    \n\u03b2 as the slope coefficient in a regression analysis<\/p>\n

    Jump Diffusion Process<\/strong>
    \nA jump diffusion process is similar to a random walk process but there is a probability of a jump at any point in time. The occurrences of such jumps are completely random but its probability and magnitude are governed by the process itself as introduced below.<\/p>\n

    for a jump diffusion process
    \nError! Objects cannot be created from editing field codes.<\/strong><\/p>\n

    where we define
    \n0 as the jump size of S
    \nF(\u03bb) as the inverse of the Poisson cumulative probability distribution
    \n\u03bb as the jump rate of S<\/p>\n

    The jump size can be found by computing the ratio of the postjump to the prejump levels, and the jump rate can be imputed from past historical data. The other parameters are found the same way as above.<\/p>\n

    \"Share<\/a>