Thank you! The math works great. The hard part is having the knowledge & confidence to overrule your emotions and stay the course.
About 15 years ago I read that most recessions have been 18-24 months from start to finish, or 6-8 quarters from peak to trough. (Of course there have been notable exceptions.) Four years' cash is more than enough for a recession, although you could put some of that cash into bonds or else carry a higher asset allocation of stocks.
We decided that two years' expenses in cash (8% of our investment asset allocation) seemed like a prudent risk. We put one year of spending money in a money-market account and the other year in a ladder of three-year CDs. That way we boosted the yield of holding cash (without losing liquidity), and our only risk was an early-redemption penalty if we had to break into a CD.
Here's a post on the nitty-gritty details of that plan:
Over the last couple years, we've moved on from that tactic and we've been drawing down the cash buffer to barely a year. One reason is that our income has risen faster than inflation-- both in our portfolio's dividends and from our rental property. Another reason is that our spending has actually dropped during the last few years (that retirement spending smile phenomenon). A third reason is that we definitely reverted to variable spending during the recessions, and of course that's difficult for the 4% SWR computers to model. I'm confident that variable spending more than makes up for "failures" in the 4% SWR.
When my spouse's military pension starts in 2022, our cash flow will turn positive and we'll never face sequence-of-returns risk again. Unless I buy a submarine.
Someone with cash flow from rental properties would also be much more protected from sequence-of-return risks, as long as their tenants kept paying the rent despite the recession.
Someone who's living strictly off their assets (no military pension) could rightfully worry that the 4% SWR study only covers 30 years. If you're FI in your 40s then you might have to make your assets last for 60 years. In other words, you'd worry about four consecutive 10-year periods of sequence-of-returns risk before you'd reach the final 20 years of your life and feel comfortable that the 4% SWR has a 100% success rate.
However that's a shortfall of the SWR computer simulation.
I'd suggest that someone in that 60-year scenario has Social Security (or the international equivalent) starting in their 60s. (SS is not part of the 4% SWR.) I also think that holding a high-equity asset allocation means that a portfolio will grow faster than inflation and will eventually grow out of the sequence-of-returns risk. Even if there's a nasty recession in the first decade, you'd have two years' cash and you'd cut spending or work part-time to give the portfolio some room for recovery. Over that first 30-40 years of FI, and despite withdrawals, the portfolio would eventually grow above a value that would give a 100% success rate and be bullet-proof to further sequence-of-return risk.
What it also means is that most of us work too long to accumulate too much net worth to ensure that we're maximizing our success rate at the 4% SWR. The reality is that we overshoot the mark, and in a few cases waaaay overshoot it.