# How can I slowly decay a percentage over time (not linearly)?

I'm not a statistician, but I am incredibly interested in personal finance, budgeting, and investment management.

I've been building a large spreadsheet of my personal expenses, savings, and investments. I have a large table that predicts investment returns year-over-year, up until age 65.

Currently the table models 10% return every year until age 65, but in reality, investments will become much more conservative over time (more CDs, bonds, etc.), so returns will decrease.

I could just create a linear formula starting at age 26 (current age), whereby every year until 65 the estimated rate of return decreases, until it eventually plateaus around 1-2% at age 65.

However, I think it would be a longer, decaying downward slope instead of a linear drop in returns.

Anyone have ideas as to how to model this? Something like exponential decay or smoothing?

Here's an image of my spreadsheet so you can see what I'm currently doing.

Another dimension in the space of QF is the Markowitzian Theory of Portfolio Optimization, which focuses on e.g. quarterly rebalancing of a portfolio for maximizing returns, $\mu$, and minimizing volatilty, $\sigma$ -- see the efficient frontier. By far, the best portfolio optimization for personal use I have used is Smartfolio -- but this is mostly for rebalancing assets and risk projection (VaR). Last is FinAnalytica, if you want to step up to commercial grade.