Allocation waterfall model

[TwoByFour]

I am 63, retired and living off my assets. The way I manage them is a little different than the norm, I believe. Generating cash flow is interesting. Here is what I do:

• My goal is to maximize capital that will be passed on to heirs.
• I do not AA between equities and bonds in the traditional sense of AA. My fixed-income bucket is really just a cash buffer for when the market is off.
• My fixed-income bucket is not a percentage, it is a fixed amount. I set it to be 5 years worth of withdrawals. At this time, that comes out to about 18% of assets.
• Within the equities bucket, I have three risk levels - high, medium, low. I have set low-water and high water marks for each risk level. When the high water mark is exceeded, I always sell. When the fund is below the low-water mark I will never sell. Low-water is determined by minimum growth (basically keeping pace with inflation). High water is set by the 10-year growth average for the fund.
• The high-risk funds are mid-caps (IJH) and a specialty fund; medium risk is large-cap (IVV). Lowest risk is VWINX. When I reallocate, I skim off any gains above the low-water mark from the higher risk funds and push those into the lower risk funds. If my cash-bucket is low, I replenish it from the lowest risk fund (VWINX).
• My cash bucket I am still struggling with (see the other thread I posted on fixed income choices) what it should be. Right now it is almost pure cash but I want it to at least keep pace with inflation.
• About 50% of the equities are in a roll-over IRA.

This strategy pretty much ignores bonds. Short and medium-term variability is not a concern for me. What is a concern are events like the crash of 2000 and 2008 so I have the cash buffer to let me sit pat for those events and not touch equities. The "waterfall" allocation model lets me control risk in the same way reallocation does with bonds, but rather than bonds I use low risk equity (or blended) funds. I know, not the same thing. But frankly in this low interest environment I cannot stomach buying bonds.



Anyway, comments are welcome.

 

[Totoro]

A few questions:
* Isn't VWINX 60% bonds?
* Have you seen a strong correlation between IJH and IVV?

With regards to the 5-year cash / fixed income buffer: depends on how you accessible you want it. You can set up a CD Ladder, that should keep up (barely) with inflation. But then it won't be available for emergency expenses.

 

[kaudrey]

Totoro is correct - you aren't ignoring bonds. WVINX is 61% bonds as of 4/30/2014. So, depending on how large a percentage that is of your portfolio, you could be significantly invested in bonds already.

So you should see what your equity/bond allocation really is, and then maybe ask a different question?

 

[TwoByFour]

I view VWINX as a black box. The chart for it looks just like a low risk equity - it predominately moves with the equities market. It does not seem to significantly move with interest rates. It does not protect principal. So I view it as an equity as far as risk and performance characteristics. It matters little what the managers are actually doing, i.e. bonds vs. equities, within the fund. I just look at how it performs.

So, I am not against bonds per se. The VWINX managers are clearly being more active than simple reallocation. My point is, I don't want to play the equity-bond reallocation game because I can never do better than them anyway.

 

[Totoro]

I looked up VWINX: it has a policy of staying within 60% - 65% bonds.

In other words, you have decided on the equity-bond allocation issue, not someone else, it's just not that visible.

By setting the initial allocation to your three funds there is an overall equities vs. bond balance allocation.

Your waterfall model then adds a rebalancing component to it: You sell high. and never buy (low or high).

It's not that far from what I plan on doing for the coming years, but in reverse. Never sell (low or high), only buy (DCA in). That's mostly because my allocation is too light on equities for my taste right now (40%).

 

[Independent]

Presumably, the "fixed amount" in the fixed-income bucket is a maximum. You plan to spend that bucket down "when the market is off".

Have you back-tested your plan for periods of poor market performance? How low would your cash go in a 2007-2012 event? How about 1965-1981?

 

[TwoByFour]

Your waterfall model then adds a rebalancing component to it: You sell high. and never buy (low or high).

When I sell the higher risk funds, I buy the lower risk funds with the proceeds. That is the waterfall. But I will never buy more of the highest risk fund since it is at the top of the heap. So if my overall NW goes up, the overall risk is going down since only lower risk funds will ever be bought.

The allocation now is about 30% high risk, 50% medium risk, 20% lowest risk. Perhaps I should make that 20, 30, 50.

 

[TwoByFour]

Presumably, the "fixed amount" in the fixed-income bucket is a maximum. You plan to spend that bucket down "when the market is off".

Have you back-tested your plan for periods of poor market performance? How low would your cash go in a 2007-2012 event? How about 1965-1981?

My idea is that as I bleed off gains and distributions from equities, those are the source of my withdrawals. But if the market is bearish, there are no gains so I won't generate income off of it. That is when I hit the cash.

Good idea on the back test. I have not done that but I will.

 

[rayvt]

This sounds like a very convoluted way of taking withdrawals from the various categories in keeping with your desired asset allocation.

Logically, there is no difference between:
a) taking money from the cash bucket and then having money cascade (waterfall) down from the higher risk buckets, and
b) taking money proportionally from all the buckets.
Whichever way you do it, the end result is the same after you've done all the money movement for the year.

I created a spreadsheet to backtest the standard "cash bucket" method and found that all the complexity just created a mirage. In just about every historical period, the ending balance was greater if you just used a fixed asset allocation and kept it re-balanced. The withdrawals are the same, just the final balance is different.

It is here: https://www.dropbox.com/s/xf4ma5blug27aws/SPY_Withdraw_by_CashBucket_rules.xls

It's set up for SPY & T-bills, but you could tweak it to use any other funds you like. I doubt the overall results will change.