TIP$TER's transparent, realistic equity return models
The Problem The Solution: TIP$TER Case Studies Sample Reports Download TIP$TER TIP$TER User Guide Support
Preferred model: exploratory simulation
"In finance, exploratory simulation is generally the most useful.... The use of historical data provides as realistic a model of real-world behavior as can be achieved."
David Nawrocki, Ph.D., "Finance and Monte Carlo Simulation," Journal of Financial Planning, Nov. 2001.
TIP$TER's default mode of simulation is exploratory simulation. In this mode, TIP$TER simulates a financial plan and the equity portion of a planned portfolio against over 1600 rolling intervals of a looped set (spanning Jan. 1871 to Dec. 2011) of real S&P 500 return data that is scaled to match the expected long-term annualized real return that you expect the market to yield.
TIP$TER tests your financial plan against every available interval in this looped data set. Assume that you make a 50-year financial plan. The first simulation trial would simulate your financial plan as if it started in January 1871, and ended in December 1920. The second simulation trial would simulate your financial plan as if it started in February 1871, and ended in January 1921. The 1201st simulation trial would simulate your financial plan as if it started in January 1971, continued through Dec. 2011 and, because the data set is looped picked up at Jan. 1871 before finally terminating in Dec.. 1879. The last simulation trial would start with Dec. 2011, pick up at Jan. 1871, and terminate in Nov. 1920.
You might notice that this approach is similar to the one used by FIRECalc®*. But there are some key differences, described here. The most significant difference is that TIP$TER, unlike FIRECalc®*, scales the historical data set to match your forward-looking expected return.
Historical return set scaled to match your forward-looking return expectations
TIP$TER is not the only financial planner that performs exploratory simulation. Money Guide Pro®* does it. J&L Financial Planner does it. And FIRECalc®* does it.
But the way TIP$TER does exploratory simulation is unique and more fundamentally sound than its peers. TIP$TER unlike Money Guide Pro®* (annual subscription cost to financial advisors: $1295), unlike J&L Financial Planner* (annual subscription cost to financial advisors: $140), and unlike FIRECalc®* (free) scales the historical data set to match your forward-looking expected return.
By performing exploratory simulation of a "mean-adjusted" historical data set, TIP$TER avoids the flaws associated with almost every other financial planning simulator. In the "Problem" section, we noted that most financial planning simulators fall into one of two flawed categories: (1) Monte Carlo simulators that oversimplify return patterns, treating returns as if they are independently and randomly distributed; and (2) historical/exploratory simulators (e.g., FIRECalc®) that merely backtest a planned portfolio against raw historical nominal or real return data.
As noted in that section, the "random walk" assumption modeled by traditional Monte Carlo simulators is fundamentally unsound. It grossly exaggerates long-term cumulative outcomes.
Ordinary historical backtesting the kind performed by FIRECalc®*, for example is too optimistic: for equities to sustain their historical performance (6.4% annualized real return from 1871-2008), the economy would have to grow at an unprecedented pace: an annual inflation-adjusted rate of at least 4% per year.
By using a historical return data set, TIP$TER avoids the oversimplification and "random walk" assumption of conventional Monte Carlo simulation. By scaling the returns of the historical data set to match your forward-looking return expectations, TIP$TER avoids the chief flaw of typical historical backtesters.
Also supports several types of Monte Carlo simulation
Although TIP$TER defaults to the exploratory simulation option, TIP$TER is ultimately agnostic about your simulation preferences. Indeed, TIP$TER accommodates several simulation options, including random sampling of returns from:
a stationary normal distribution
(i.e., conventional Monte Carlo)
a stationary lognormal distribution (what most "advanced" Monte Carlo simulators do)
a stationary double lognormal distribution (fatter-tailed and more leptokurtic)
a non-stationary normal distribution (to incorporate mean reverting behavior)
a non-stationary lognormal distribution; and
a non-stationary double lognormal distribution
What, exactly, is the point of providing these other models? So you can compare them, and study the differences, and perhaps gain a deeper understanding of why the model matters.
user manual provides detailed information on each of these simulation
Next: Provides meaningful measures of simulated portfolio performance
Previous: Encourages sound fundamental thinking about future returns
J&L Planner is a trademark of J&L Software, LLC, which which Prospercuity claims no sponsorship, connection, affiliation, or association.
Money Guide Pro is a trademark of PIETech, with which Prospercuity claims no sponsorship, connection, affiliation, or association.