A better retirement through quantitative investing

September 22, 2014

Today I finally get around to doing something I mentioned way back in my first post on quantitative investing. Quant portfolios can help an investor handily beat the market over time and often with even lower risk. If that is the case then how could they impact safe withdrawal rates (SWRs) in retirement? Lets find out.

I’ve sort of already shown the effect of quant or automatic investing systems on SWRs in my posts on the various flavor of the IVY portfolios. For example, you can more than double your SWR with certain versions of the IVY portfolios. Beyond the basic buy and hold versions, the IVY portfolios are just different versions of quant systems in which whole asset classes and ETFs are used instead of individual stocks. In the IVY systems the quant metrics just happen to be trend following and momentum for the whole asset class. But I warned that the attractive SWRs from the IVY portfolios were too good to be true because the historical data did not encompass the worst period to retire in history, 1966, which defines the worst case SWRs. I want to extend that analysis of SWRs to quantitative systems using individual stocks I’ve discussed here before, namely the trending value portfolio and the combined consumer staples/utilities portfolio. These portfolios exhibit very high risk adjusted returns. Also, maybe the best part, is that we have historical data that encompasses 1966, the worst case retirement period which allows a very valid comparison to traditional retirement portfolios. OK, lets get to it.

First thing to be done is to re-run the standard stock, bond retirement portfolio (SP500) data for the period studied which is 1964 to 2013. Then do the same for the trending value (TV) portfolio and the consumer staples/utilities portfolio (XLP/XLU). Diversified portfolios are constructed with the US 10 year note. Below is a comparison of the SWRs for different stock bond allocations for the three portfolios.

SWRs for SP500 TV XLPXLU Portfolios Sep 2014

Although it’s not my purpose here, note that when you remove the great depression from the data set that even for the SP500 portfolios SWRs rise along with allocation to stocks. This is not the case when you include the great depression where the SWR falls off after its peak at 70% stocks 30% bonds. Probably a whole other topic for another day where I’ll muse about institutional learning, the rise of fiat monetary systems and the like. But for now lets just say there is not much incremental benefit for SWRs beyond 60/40. At 60/40 look at the increase in SWRs from both the quantitative portfolios; 6.39% SWR for the XLP/XLU portfolio and 7% SWR for the TV portfolio. Pretty darn good. But what about the two other retirement portfolio metrics I like to look at; maximum drawdown and average end wealth? I present those two tables below.

Max DD and AEW for SP500 TV XLPXLU Portfolios Sep 2014

MaxDD in the top table is the maximum drawdown, peak portfolio value to trough portfolio value, on an annual basis. It is the same as the worst year if there is only one down year. In the case of multiple down years it represents the compounding of those bad years after retirement withdrawals which is what matters most. $AEW is the Average Ending Wealth, or ending portfolio value, of all the 30 year retirement periods in the analysis. This is the value of the portfolio you could potentially bequeath to your family and/or charities. The quant portfolios deliver on these metrics as well. Those high SWRs in the first table come with lower drawdowns and more wealth on average across the board. Again great news across the board.

In the MaxDD and $AEW table I highlighted the 30% stocks, 70% bond rows for a good reason. The more and more I talk to retired investors the more I think that SWR is not the parameter that most retirees should try to maximize. Losses or drawdowns are what gives retired investors the most fits and more importantly is what most often derails their investment plans which leads to very poor investment results. Keeping that in mind I highlighted in the MaxDD table a level that I think most investors can tolerate, less than 10% peak to trough loss on a yearly basis. That level is at a 30% stock 70% bond allocation for the various portfolios. With the quant portfolios even at that conservative allocation level you can support an SWR of over 5% and at the same time end your retirement with an average wealth that is almost as great as a 100% stock allocation if you were just using the SP500 for your equity investments. This is an example of what Larry Swedore calls ‘low-beta, high tilt’ portfolios where in you use the bond allocation to control beta or risk (i.e. drawdowns) to a tolerable level and use a ’tilt’ to factors that increase return such as small cap, value, and momentum. This is just one example. You can choose your own optimal allocation based on your preferences.

In summary, quantitative portfolios can significantly improve the important metrics of retirement portfolios; SWRs, maximum drawdown or loss, and average ending wealth. Or at a minimum they can vastly improve the odds of a successful retirement even at very conservative SWRs or equity allocations.

The flip side of a successful retirement: spending

September 17, 2014

Most of the ink spilled in talking about retirement is limited to the investment side of the equation. How much do you need for retirement? How much can you withdraw from your portfolio in retirement? How should I invest my retirement assets? And obviously, all these questions are critical. But just as important and not talked about as often is the amount of spending in retirement. I’ll touch on some of my thoughts on this topic and my personal experience.

A few weeks ago I was reminded of an old saying, ‘The easiest way to double your money is to take it out of your pocket, fold it in half, and put it back in your pocket’. Another one I like goes something like, ‘a dollar not spent is a 100% guaranteed return on your money’. These are somewhat old school thoughts. It reminds me of how my grandparents funded their retirement with basically a 100% cash allocation and tight control on spending. Just getting them into bank CDs was a major multi year effort. However, there is a ton of wisdom in some old school thoughts. And spending in retirement is just as important as investing in retirement. The three aspects of spending in retirement are the absolute level of spending, the yearly change in spending required to maintain the same lifestyle, and the amount of flexibility in the spending level. Here I want to focus on the absolute level of spending.

Your spending level in retirement can be a lot less than what most retirement resources tell you. The typical recommendation is to plan for 85% to 100% of your pre-retirement spending. Obviously, this is highly individual but I think its important to point out that you can have an amazing retirement by spending a lot less than you did pre-retirement. At least that has been my experience. Nina and I spend approximately half of what we did before we left our careers, almost nine years after the fact. Below is our level of spending by year as a percentage of our pre-retirement spending. 2005 was our last year of pre-retirement spending so that is the base year.

Change from pre retirement spending levels sep 2014


As you can see it fluctuates quite a bit but the general trend has been down. And our quality of life is infinitely higher. Not working in the traditional sense opens up a host of possibilities to spend less. For example, we spend less in “rent” (RV camping fees plus fuel) than we spent on association fees for our condo in San Francisco. And don’t even ask me about property taxes. Working has an implied cost that is commonly overlooked (clothes, commuting, etc…). When you remove the costs and constraints of working it can translate into a needing lot less to retire than if you planned for 100% of pre-retirement spending. Some times less is truly more. This is a huge topic and other people cover it a lot better than I can give it time. A good resource and community on spending less can be found at Mr Money Mustache. How would your life decisions change if you could spend half of what you spend now?

Once you have the absolute level of spending down then the task moves to managing the change in spending every year and the surprises that inevitably come down the line to throw your plans into the air. Anyone notice 2009 in the table above? That was not really a surprise but in late 2008 we decided to move back from Hong Kong to the US. The surprise was the sticker shock from the cost of the move. What better time to be flexible, right? I took a job in the US and negotiated a decent relocation allowance that paid for our move back. I stayed in the job for just less than year and during that time we lived the high life and started planning our RV adventure. The increase in spending from 2008 to 2009 was pretty much covered by my salary (I still withdrew from my retirement assets that year but it was a small percentage). In 2010 we moved into the RV and the rest is history so to speak. What I think is key here is that you have more control and flexibility than you think. If there is another big surprise down the road whether it be from a big spending increase or a big portfolio loss I have the flexibility to cut spending and/or increase income.

In summary, you don’t need to spend as much in retirement as the common wisdom suggests. This also means you don’t need as much to retire as well. Controlling spending is as big a part of successful retirement as the investment side of the equation. You have more control and flexibility than is commonly thought.

Understanding modern retirement calculators

August 31, 2014

In today’s post I want to explain and demonstrate how modern retirement calculators work. There are two basic ways to calculate how much you can safely withdraw from your portfolio in retirement; looking backwards using historical data for past retirees and looking forward using possible future investment returns. On this blog I’ve pretty much only described the historical safe withdrawal rate (SWR) approach. That’s where the 4% SWR number comes from that you hear talked about all the time. Modern retirement calculators do not use this approach. They use a forward looking approach called Monte-Carlo analysis to come up with their SWRs. Lets see how these calculators work.

What’s wrong with using historical data to come up with SWRs? The problem is history only represents one outcome of all the possible future paths from that previous point. If you were back in 1966 looking forward to a 30 year retirement the actual outcome of your 30 year retirement period, which happened to become the worst in history, was only one of many possibilities. Modern retirement calculators eliminate this drawback by looking at many many possible future retirement outcomes. That’s the fundamental difference. The rest is calculation details. How do they do this? They assume that future investment returns follow some probability distribution, since the future is fundamentally indeterminate and probabilistic. So, for example, they assume that future investment returns are normally distributed (they also use some really fancy distributions like multi-variate log normal but the results are similar), i.e. follow a bell curve. They then randomly generate a time series of investment returns and use those investment returns to calculate what the SWR would be in each of those random series of returns. Since they can’t do this for the infinite variety of possible returns they usually stop at a sufficiently large number that generates statistically robust results. The generally accepted large enough number for these retirement calculators is 10,000 trials.

Now, here is the catch with these retirement calculators. They still involve some kind of bias or human forecasting under this seemingly unbiased random robust scientific method. For those assumed normal distributions  you still need to pick a mean and a standard deviation to generate those random investment returns. And you still need to forecast inflation as well. You can’t model all possible potential future outcomes. Confused yet?

It’s really pretty simple, it’s just the calculations that get intense. I put together a simple retirement calculator in excel to illustrate how this is done. In one excel file I generated 1,000 random series of 30 year investment returns (note: every time you open, save, or make any changes to this file the random numbers change, unless you turn off automatic calc in options). I generated stock and bond returns independently. I then copied these results to another excel file where I did the work of calculating the SWR for those 1,000 random simulations. In this first pass I use the historical mean and standard deviations for US stocks and US 10 year government bonds from 1929 through 2013. I also used historical inflation. And this leads to the next important observation regarding retirement calculators. All of them define a successful SWR to one that has an 80% to 85% success rate. In a 10,000 trial simulation that means 1,500 to 2,000 of the possible time series of investment returns result in failure, i.e. running out of money. It’s just that those cases are lower probability events. Yet still possible. Using an 85% success rate my simple model yields and SWR of 3.92% for a 50/50 stock bond allocation. Pretty close to the historical SWRs for such a simple model.

But none of the modern retirement calculators I’ve seen use historical returns in their calculators. They all use a return forecast that is significantly lower than historical returns have been. I think the simplest and best retirement calculator is the one from Vanguard. I like it because there is no need for a ton of inputs on the user’s part. It just generates the SWR based on the retirement period and one of three asset allocations. All the gory details are behinds the scenes. Their future return assumptions are updated once a year and explained in the Vanguard Capital Markets Model. It’s well worth the read. Basically, they use diversified portfolio returns of 3-5% real going forward, less than returns have been in the past. For a 30 yr retirement period and a moderate asset allocation their model yields and SWR of 3.8%. I used these lower future expected returns in my simple model and that yielded an SWR of 3.30%. You can access those files here and here.

At the end of all this fancy analysis we have a tool that tells us that future SWRs will be lower than the past because we are forecasting future returns to be lower than historical returns. And specifically lower than the 30 year retirement period that started in 1966 which is what gave us the worst case historical SWR to begin with. Seems kind of obvious when you put it that way. So, basically this all relies on a forecast of the future which are notoriously inaccurate. But its the best tool we have. From today’s valuation levels and interest rates it is possible the SWRs will be slightly to a good deal lower than 4%. But they may not be.

That pretty much sums up how modern retirement calculators work. Now you can at least use these tools with an understanding of what’s under the hood. You can even generate your own models if you want.

Note: the T Rowe Price retirement calculator is another very popular one. I find the added inputs and complexity unnecessary.

The impact of realistic spending on your retirement

May 28, 2014

In a post last week I showed that for the majority of retirees the standard recommended spending model in retirement does not match reality. Spending  in retirement actually grows by less than inflation over time. In this post I want to show you how using a more realistic sending model impacts how much you can withdraw from your portfolio in retirement.

If we go back and use the same model we used to calculate SWRs, we can adjust annual spending to see the impact of using a more realistic spending model on SWRs. For this post I simply changed the annual spending adjustment in retirement to grow by 1% less than inflation during the 30 year retirement periods (-1% per year real spending). This is actually pretty conservative compared to the actual retiree data.

Second, I took at look at combining a realistic spending model with a flexible spending model I’ve discussed in the past. The Floor Ceiling method basically allows spending adjustments in retirement based on the performance of the portfolio. For this post I used a floor ceiling model with unlimited upside and a max floor of -10% to inflation adjusted spending.

The results are shown in the table below for 30 year retirement periods, a 70-30 SP500 10yr UST bond stock portfolio with data going back to 1929.

SWRs for base vs realistic vs fcm model may 2014

The standard retirement model says your SWR is 4.39% for this portfolio. By simply using a more realistic spending model the SWR for the portfolio goes up to 4.90% an increase of almost 12%. The FCM model with its flexible spending says your SWR is 4.84% for this portfolio. By combining both approaches the SWR for the portfolio goes up to 5.39%, an increase of over 22%! That’s a raise of 22% per year in retirement simply by being more realistic and flexible. Very powerful stuff.

For those still building for retirement the implications are perhaps even more impactful. The percentage increases in SWR are equivalent to a reduction in the amount you need to fund your retirement. If you can spend 22% more with these models in retirement it also means you would need 22% less to retire. That could translate into several years of work or no work in this case.


Spending Realities in Retirement

May 20, 2014

It’s usually worthwhile to question conventional wisdom. At the minimum you usually learn where that wisdom came from. Often you learn the conventional wisdom is not the best approach. This happens to apply rather well to the topic of retirement spending. Lets take a look at the conventional wisdom regarding retirement spending and see why it may not be the best approach.

All the standard retirement models tell retirees to adjust spending every year for inflation right? This how the SWR (Safe Withdrawal Rate) model works. You pick a percentage of the portfolio to withdraw in year 1 and then adjust that spending every year for inflation. Simple enough. The theory being that at the minimum we want to maintain the same ‘standard of living’, i.e. real spending, from year to year. The biggest question to ask is this model actually how retiree spending evolves over time?

Anectodally, this model has always felt a bit off to me. For one, I meet a lot of retirees and speak to them quite a bit about finances and spending. Many I’ve met have held spending constant for years while maintaining their quality of life. In my personal situation, on average spending has either stayed the same or declined every year. What gives?

Fortunately, some recent research has shed some light on this topic. Michael Kitces has a great post covering this research which I highly recommend. I’ll just highlight a few points from the post. Lets first look at how retirees actually spend. Here is the data:

Annual Real Change in Consumption for Retirees May 20 2014

The orange line in the graph above is spending adjusted for inflation, real spending. This is the line all retirees fall on according to the recommended retirement models. The reality is somewhat different. For the majority of retirees real, inflation adjusted spending declines in retirement. Even towards the end of life. The declines are on the order of 1% a year, growing to 2% a year, then back to 1% a year in the later years but still never positive. Definitely a different reality than the models portray. An important point: this is real spending. Actual spending may still actually increase. For example, inflation is 3%, spending goes up 2%, real spending is then down 1%. Now lets look at how you can use this in your retirement.

I think the best way to use this research is to take a look at your retirement in stages. Michael also talks about this in his post but basically the idea is to divide retirement into three spending stages. Like pictured here;

three stage retirement model may 2014


Each of the three stages can be modeled with a different change in retirement spending. For first stage, say 60 to 70, -1% per year in real spending could be used (or conservatively the standard model’s o% per year), for the second stage, 70 to 85, -2% per year, and for the third stage back to -1% per year. This can be tailored to be fit your specific circumstances which is also much better than the standard retirement model.

OK, so far all we’ve done is match the retirement model to a better representation of retirees’ actual experiences. Now for the punchline. Being that realistic scenario is less spending than the model means one of several things. One, the standard SWRs are too conservative. Two, retirees need less money to retire to being with. Three, worst case, the SWRs from the standard retirement model plus a more realistic spending outlook means a much higher probability of success in retirement. I’ll take a look at some of these implications in later posts.

Standing on the edge: diversified portfolios & the modern retiree

May 16, 2014

What’s the hardest part of the retirement decision? For many its taking the leap of faith that financially what’s worked in the past will continue to work in the future. And often the start of every single year, during those early years of retirement, brings the anxiety right back again. I’m in my ninth year of withdrawals and while the anxiety has lessened by orders of magnitude, it is still there. Someday I may recount the year by year story, maybe after year 10. In today’s post I want to look back a bit in order to look forward. We’ll go back to the beginning of the year 2000 and the year 2008 and look at the decisions facing those retirees. Then we’ll look at how those portfolio decisions performed through the end of 2013.

For every retiree since 1996 the retirement decision matrix, SWR decisions vs stock allocation, has looked exactly the same. The table below shows what those choices are:

SWRs by allocation May 2014

Why have the historical SWRs not changed since 1996? Because, as I show in this post, the SWR is solely determined by the worst case retiree in history which happens to be the 1966 retiree. That 1966 was the worst case period was not know until that 30 year retirement period was over in 1996. Lets assume that the 2000 and the 2008 retiree chose an SWR of 4.39%. They then were faced with several investment portfolios offering different levels of diversification and risk management. They used data similar to what I presented in this post. I’ll use the various diversified IVY portfolios, their timing relatives, and the Permanent Portfolio. How have the 2000 and the 2008 retiree fared if they chose one of those portfolios? Here is the data.

Diviersified portfolios and the modern retiree

First, lets look at the year 2000 retiree, 14 years later. Everyone of these portfolio choices would have worked out just fine compared to the worst case 1966 retiree. In every case the more diversified portfolios did even better and the timing/trend following portfolios did even better. Only the the retiree that chose the standard 70% US stock, 30% US Bond portfolio, ended with less money than when they started with. Several of the portfolio choices would also have given the 2000 retiree the opportunity to adjust their withdrawal rates higher due to the great performance. Diversification works and it paid off in spades. All as expected.

Now, lets turn to the 2008 retiree. Hmm, notice something interesting here. For the last 6 years, diversification has basically failed to give a portfolio much benefit. The bull market in US stocks has been quite powerful. Only the two most aggressive portfolios, that combine timing and momentum (GTAA 3 and GTAA 6) beat the standard 70/30 portfolio substantially. What should we make of this? Time to abandon diversification all together? I hope you don’t think so. The last 6 years of US market out performance does not trump the long term historical data. Nor should it. If anything I think it suggests the diversified portfolios are set to outperform going forward. And so far, in 2014, that is proving itself out. But hey, that’s forecasting and who knows.

There is something more important here to take away. While it may be intellectually easy to say, stick with diversification, emotionally, while looking at your portfolio statements, it is a different story for most people. Any strategy you choose will have periods where it under performs other strategies. And those periods of under performance can be significant. This uncertainty can make taking the leap into retirement quite scary. How do you deal with this?

This is a concept called base rates which I’ve discussed in my quant strategy posts but the concept applies here as well. How best to deal with base rates less than 100%, i.e. that your portfolio choice will under perform for various periods often a lot longer than you feel comfortable with? The best way is not to go with one strategy all or nothing. A part of your retirement portfolio could be a standard 70/30 stock bond portfolio, another part could be more broadly diversified IVY B&H 13 for example, etc.. This way during any given period part of your portfolio is outperforming. This I think is the best way to deal with the emotional factors involved in retirement investing.

In summary, the 2000 and the 2008 retiree while both doing just fine with any of the various retirement portfolios discussed here have had quite a different experience. While for one diversification has been a boon, for the other it has been mostly a meh. This is quite the exhibit of the uncertainty in retirement. It is important I think to let the long term record speak for itself but at the same time choose a few of the strategies to make the emotional aspect of retirement investing easier to deal with and thereby increase the chances of success.


Revisiting the worst times to retire in history (2013 update)

May 12, 2014

Today I wanted to do another yearly update post. This time I’ll refresh the worst times to retire in history with data from 2013 which adds another year in the retirement story of the 5 ‘unluckiest’ retirees in history. See here for last year’s update.

As I said in last year’s update:

The worst time to retire since 1929 turns out not to be the Great Depression, as most people would believe. In fact, the worst time to retire in history was 1966, followed by 1965, then by the Great Depression year of 1929. The SWR (safe withdrawal rate) of 4.39% for a 70% stock 30% bond portfolio is solely determined by the retirement results of the 1966 retiree. Of course, we care about the present not the past. How are the year 2000 and the year 2008 retiree doing compared to those who retired during these worst of times? The table below (click for a larger image) shows the year by year progress for 5 retirees; the 1929 retiree, the 1965 retiree, the 1966 retiree, the 2000 retiree, and the year 2008 retiree. It uses a starting portfolio value of $1M, with a 70% stock 30% bond portfolio, which yielded a max SWR of 4.39%.

Here is the updated table with 2013 data.

Worst times to retire in history 2013 update CWR vs years

After 14 years, the year 2000 retiree is doing pretty well compared to history. It may not seem that way. The year 2000 retiree’s portfolio is down to $787K and the $61K yearly withdrawal is now 7.7% of the portfolio. CWR is defined here. But that is way better than the 1929, 1965, or 1966 retiree, all of whose portfolios survived a 30 year retirement. The year 2008 retiree is doing even better. After 6 years, despite the financial crisis, portfolio value is back to where it started and so is the current withdrawal rate. This is a much better situation than any of the other 4 example retirees.

In short, both our ‘unlucky’ modern retirees, 2000 and 2008, are doing just fine. With respect to history they are in a better position than the three worst retirement periods in history (1929, 1965, 1966). The 4% SWR rule looks to be just fine for yet another year.

P.S. I only discuss the worst retirement periods in history and the worst current periods, 2000 and 2008. This means that any other starting year in retirement has better results than that shown here. For example, the 2001 retiree would be better off than the 2000 retiree.



Get every new post delivered to your Inbox.

Join 1,651 other followers

%d bloggers like this: