In the last post Existential Doc posed the question what to do? I went through how I would analyze it but a more formal presentation might be useful.
What would happen to survivability if the stock market dropped in half permanently? How does AA affect this? Lets start with the good old bogglehead triple fund. The triple is a 80% stocks 20% bonds fund of US stocks Global stocks and total bonds and if you used it as your retirement portfolio for 30 years and a 4% WR the probabilities would look like this:
82% survival for 30 years. I’m going to treat this as an 80/20 portfolio socks to bonds. Let’s say stocks fall in half. A 1M portfolio now becomes 200K in bonds and 400K in stocks so the portfolio is worth 600K and the AA is 67/34. Lets monte carlo that 67/34 portfolio using the same ratios and WR on the 600K principal
43% survive at 40K/yr WR. If you knock the WR down to 28K we are at 82% survival once again. So to survive a 50% equity hair cut using an 80/20 bogglehead at equal levels of survival and a 1M start you’re 40K/yr WR must drop to 28K/yr or a 30% cut. That’s almost 2500/mo. What about a 60/40 Stocks v bonds at 4% WR and 30 year horizon?
Just by changing the AA to 60 /40 and getting rid of foreign survivability goes to 96%. What happens if we loose 50% of the equities? on a 1M portfolio the bonds are 400K and the stocks are 300K, so you immediately start with an extra 100K in the bank compared to the BH3 example. The AA of that 700K is 43/57 stocks v bonds so lets MC that sucker 700K, 40K WR and a 43/57 AA
so a 40K WR on 700K drops survival to 70% What if we re-balance back to 60/40?
74% It helps a little, so let’s keep 60/40 and try reducing WR to 30K/yr
We see a 94% success at 30K/yr nearly the same as the original 96%.
So what does all this mean? It’s how you plan. What do you do if you loose 1/3 of your money? What do you do before you loose your money? We saw what happened to a typical 80/20 portfolio not on the efficient frontier. WE saw what happened to a safer 60/40 portfolio on the efficient frontier, and we saw the survival rates. To get the bogglehead 3 to 92% survival required dropping the WR from 40K/yr to 19K/yr after the disastrous loss.
The other thing that matters is when SS kicks in. If you are close to SS when the disaster hits you are largely immunized. If you are far away from SS because you retired early…
6 Replies to “How to Survive Demographics”
That is a good analysis and confirms that I want to move towards 60/40 and keep working at least part time as long as I can stand it. I think planning to retire with a paid off residence and the ability to drop WR helps too.
But when to decrease equity allocation? As you say, maybe not today… but soon? I’ll keep watching the Fed.
“Moving some to bonds and shortening the window of retirement, thereby reducing risk and decreasing WR (since your pile at retirement is bigger) is the way to go but maybe not yet quite today.”
Simply look at returns for either AA on the efficient frontier and calculate the percent differences. A BH3 has an expected return of 9.02 and risk of 11.95 A 60/40 has a return of 9.18 and a risk of 9.02. virtually the same return for 25% less risk (1-(9.02/11.95) = .25 or 25%. What you will find is the change in risk is often significantly greater than the loss in return so just decide what you can live with and buy that. In my case I already “sold high” and locked in my gains, so my”FED trigger” is playing defense on my already locked in gains. A little different strategy than your case. Changing to a 60/40 effectively locks in your gains. When it comes to taking profits the better is the enemy of the good. Walking 1 foot from the edge of a cliff is probably good enough when taking in the view. Walking to 3 inches from the edge does not improve the view only increases the risk.
Great analysis yet again Gasem and convincing argument to do 60/40 with just 2 holdings.
Watching the video that started this all was quite a sobering event. Very detailed explanation of the retirement crises we are facing with the boomer generation about to leave the workforce and what that impact will have on the economy.
You know the storm clouds are gathering. SS getting cut, geezers retiring, pension funds like teacher unions running out of money and everyone expecting the “government” to bail them out. The problem is we is them. There is no other them to soak. The whole system is way over levered IMHO. Over leverage is what happened in 2008
Do you think the difference in results between your and big ERNs results are only the methods ie monte carlo vs historical perspectives. Big ERN series on SWR supports a much more aggressive AA of 80/20. I read the guest blog you did on his site and he seemed to down play the differences.
It is in our approaches and optimization. Karsten likes historic analysis because it allows him to add custom parameters to the analysis like the Sharpe’s score. I don’t have a way to do that in Monte Carlo, but Monte Carlo has the ability to front load failure. Karsten and I also tend to ask different questions. His question is what does a normative retirement for a 45 year old look like. My question is what does a bullet proof portfolio look like? Age is a difference as well. I’m 67, so 20 years closer to death than he is, my appetite for risk is different and our portfolio sizes are likely different (though I don’t know what he owns). So it’s not just based on the approach of the calculator, but also the length of retirement, WR, when SS kicks in, taxes in retirement, side gigs, and how leveraged the principal needs to be to reach the end game.
Once you reach 20 years left it’s pretty hard to run out of money. Not so if you have 50 years to travel. I have great respect for Karsten’s method. If the economy stays relatively constant on a decades long growth curve his method will be pretty accurate IMHO. But the market follows a dual credit cycle. One curve oscillates about every 10 years and the other oscillates every 70-100 years aka once in a life time. The crash of 1929 is almost 90 years old, Oct 28 is its birthday. Karsten does deal with that in his WR analysis, so does Monte Carlo. Karsten uses glide path. Monte Carlo uses Gaussian statistics.
So there are multiple factors involved. Hope this hits close to what you were asking