To unpath’d waters, undream’d shores

Shakespeare A Winters Tale, Act 4, Scene 4

This exploration began four years ago, with an initial objective of building a passive income of $58,000 per annum by July 2021. Since that time, goals have evolved and changed enabling bringing forward the achievement of this initial goal.

Each year in early January I spend time reviewing my investment goals and how I plan to reach them.

This post explains the findings from my annual review, details my updated portfolio goal and assumptions, and discusses how I will approach my financial independence journey through 2021 and beyond.

The aim is to have a clear written record of the objectives, approaches and reasoning underlying the plan, to serve as a reference point through the year. The process also enables the updating of plans and assumptions for changes in circumstances, thinking, as well as data and evidence.

Ports of call and the future of the voyage

In the middle of December 2020 I passed the single portfolio objective I set myself in early 2020, six months earlier than targeted.

This target of $2,180,000 was based around an assumption of the portfolio producing a real annual income of $87,000 in 2020 dollars. This level was chosen because it reflected the equivalent of adult Australia full-time ordinary earnings, and was close to my then estimated spending of $89,000 per annum.

Passing this target forced close and careful thought about the meaning and significance of the target. For the journey so far, the target has been a way to measure progress towards the goal. It has not served as a countdown clock to immediate early retirement.

The value of the process of setting the target has been in large part as a forcing mechanism to clarify what type of post-financial independence is envisaged.

In this respect, the target goal of adult ordinary full-time earnings represents an objective external measure of a continuing capacity to access and participate in the range of goods, services and daily experiences of those in the community. That it sits close to my actual spending also provides an assurance that post-early retirement life would not see a material diminishment in or limitation of lifestyle.

Setting a revised portfolio goal

To recognise these same factors, I have revised my existing goal and formed a new Portfolio Objective.

The new objective is to reach a portfolio of $2,585,000 by 31 July 2022. This would produce a real annual income of about $90,500 (in 2021 dollars).

This revised target is an increase of just over $400,000 on my previous portfolio objective. The key reasons for this change are two fold:

• Updating the target for movements in inflation and earnings – simply moving the target to nominal 2021 dollars and updating average ordinary earnings increases the target by a small amount
• Switch to a clear ‘safe withdrawal rate’ target approach – changing from previous ‘average portfolio returns’ based approach to adopting an explicit safe withdrawal rate. This represents a more material adjustment.

The passive income target for this objective is updated with 2020 data and expected inflation. It seeks to reflect in nominal 2021 dollars the approximate equivalent of average Australian full-time ordinary earnings. It is also close to my current estimated average annual spending between 2013-2020 of $85,200.

This income goal is designed to reflect a ‘business as usual’ lifestyle, rather than a ‘leanFIRE’ approach. This is personal choice. As described above, as I considered my goals for financial independence, this level is closer to the level of expenditure at which I think I will be truly indifferent to working or not in terms of lifestyle.

Movement to a safe withdrawal rate

This year I have changed my approach to setting the portfolio goal to explicitly incorporate a safe withdrawal rate in the calculation.

In previous years I have estimated the portfolio goal by dividing the passive income target by a projected average real portfolio return (for example, 3.99 per cent in the case of 2020). The real return assumption used in this past calculation was based on the total return arising from the portfolio allocation. These return assumptions are discussed further below in the second half of this entry which reviews allocation plans and portfolio design.

Such an approach was always a useful rough approximation of a safe withdrawal rate, and past annual reviews recognised it as an imperfect but serviceable proxy measure for use on the journey of accumulation.

The weakness of this approach is that it fails to account for sequence of returns risk. This is the potential for early negative returns affecting the portfolio in the first years of drawdown to lead to unsustainably high portfolio withdrawals, depleting the portfolio prematurely.

At this stage of the journey it is important to move beyond approximations and start to target explicitly a portfolio goal that better takes into account a reasonable estimate of a safe withdrawal rate.

I have chosen a target safe withdrawal rate of 3.5 per cent. This figure underpins the new portfolio goal (i.e. $90,500/0.035 = $2,585,000)

Reasoning for target safe withdrawal rate

A safe withdrawal rate is a personal decision. The choice is informed by risk tolerances, personal factors and views about the value of historical data in forward-looking decision of high personal consequence.

The basis for previous reviews has been the finding that based on Australian equity data over the period 1900-2011, a safe withdrawal rate of 4.0 per cent has had a 88 per cent success rate over a 40 year time horizon. This rate lifts to 97 per cent at 3.0 per cent.

A further study by Professor Wade Pfau, an eminent academic commentator in this area, has indicated that the safe maximum withdrawal rate for Australia using 1900-1979 equity market data was 3.68 per cent.

These findings might suggest a safe withdrawal rate around 3.7-4.0 per cent, however, several factors have led to a more cautious rate being adopted.

These factors are:

• Survivorship bias – Due to strong equity market performance Australia has had one of the highest safe withdrawal rates in the world. Effectively projecting continued higher than globally average equity market returns indefinitely into the future may not be realistic. Many other highly developed countries with unique advantages have experienced lower safe withdrawal rates. There are no particular reasons to suppose that the outperformance of Australia’s equity markets over the past century will reliably persist (see Pfau An International Perspective on Safe Withdrawal Rates from Retirement Savings: The Demise of the 4 Percent Rule? 2010).
• High equity market valuations – Safe withdrawal rates should account for relatively high equity market valuations at present, which can justify a lowering, for example, of US safe withdrawal rates to 3.25 per cent.
• Impact of historically low bond yields – There is some academic evidence that a safe withdrawal rate of 4.0 per cent is simply too high considering the extremely low bond yields currently available. By some estimates, a 40 year drawdown horizon in these conditions implies a safe withdrawal rate of between 1.6-2.2 per cent, assuming a tolerance of failure of between 5-20 per cent (see Finke, Pfau and Blanchet The 4 Percent Rule is Not Safe in a Low-Yield World 2013).
• Portfolio management fees not taken into account – Academic and other estimates of safe withdrawal rates routinely assume away real world investment management fees. For the investment portfolio at present, these average around 0.23 per cent per annum.
• Statistical portfolio ‘success’ can look a lot like failure – Many statistical ‘successes’ with a higher safe withdrawal rate would actually be close run exercises. This is because success is frequently defined in such studies as having at least $1 in the retirement portfolio at the end of the drawdown. Suffice to say, this does not match many people’s intuitive desire for a margin of safety and security late in life’s journey. In short, as this analysis shows, ‘terminal values’ and the experienced path of the drawdown matters.

These factors all argue for a safe withdrawal figure lower than 4 per cent. In moving to set the final figure, I have re-read key parts of Early Retirement Now’s brilliant US-focused safe withdrawal series as well as some works of AussieHIFIRE and Ordinary Dollar who have also produced excellent shorter and simpler analyses of Australian returns.

A final consideration in the other direction has been evidence on the potential value of flexibility in withdrawal rates.

The flexibility to not draw on one year of distributions can make a significant difference in the performance and success rate of a reasonable safe withdrawal rate. Similarly, this (pdf) Canadian Vanguard study shows that a dynamic flexibility in spending can lift a safe withdrawal rate for a high equity allocation portfolio across 40 years from 3.5 per cent to 4.3 per cent.

On balance, taking all these considerations into account, a target value of 3.5 per cent appears a reasonable estimate, and to match the personal degree of assurance I am comfortable with.

Resetting the timeframe of the portfolio goal

Setting a timeframe for the achievement of the portfolio goal, given past targets and the unexpected path of progress to date, seems a triumph of hope over experience.

Nonetheless, an indicative timeline can help focus the mind on planning and offers a perspective on progress.

As with previous years, to reset the target timeframe for how long it may take to achieve this revised objective, I used three alternative measures. These were past average monthly portfolio gains, an estimate taking into account just contributions, and a simple trend-line forward projection of progress based on monthly data points over the past three years.

Each of these are highly imperfect approximations, and cannot fully account for the volatility which will be encountered in any projection into the medium term. They also cannot account for deviations in distributions, extra savings from past changes to the levels of emergency funds, or a dozen other complicating factors.

Nonetheless, a rough average of the approaches suggests that reaching the target by 31 July 2022 – or around 19 months away – is feasible based on past progress.

As with previous targets, mathematically, as the portfolio grows in size the actual timeline for meeting the target is increasingly driven more by market fluctuations over the next 18 months than raw saving and investment efforts. This can be seen quite clearly from the chart below of volatility on the journey to date.

As is quite evident from the above, a number of other outcomes are also entirely possible.

Using pessimistic assumptions of no net portfolio growth at all over the next 18 months or so, the new portfolio target would not be reached until the end of 2022.

Further significant market falls as occurred in March 2020 could easily delay the target by more than a year, and falls in any part of the portfolio could easily take the portfolio below the previous target level.

Tracing the path to financial independence

Broadly, I will continue to measure my progress in simple percentage terms against the revised total portfolio goal needed to deliver the passive income target, as described above.

Similarly, after expanding the reporting of progress in recent years to recognise that I have some significant superannuation assets that sit outside of the financial independence portfolio, I will continue to assess progress using two key benchmarks:

• a ‘Portfolio’ benchmark – this the primary measure based on the total value of assets in the financial independence focused investment portfolio alone; and
• an ‘All Assets’ benchmark – a secondary ‘All Assets’ measure which takes into account portfolio assets, but also the additional value of superannuation assets.

This latter ‘All Assets’ approach recognises the incompleteness of a too narrow portfolio-only view, which ignores a substantial set of financial assets that are large enough to have significant implications for the achievement of financial independence, even if they ultimately have some potential accessibility restrictions and some legislative risk.

Due to these risk and restriction factors, my first preference is to continue to target financial independence through my private investment portfolio alone, with superannuation providing an additional ‘margin of safety’ and buffer.

Recognising this, I will continue to just report a total ‘All Assets’ measure, rather than detail or write about my superannuation arrangements (which in any case are simple enough, being almost exclusively in a low cost index fund).

I will also to report against a slightly revised set of benchmarks, beyond just my formal investment objectives.

Previously, I have reported against two additional measures.

My average annual Credit card purchases (a ‘credit card FI’ benchmark) was one, and the second (Total expenses) was previously determined using an aggregated rough estimate of total current annual expenditure ($89,000).

From this month I will discontinue the credit credit card measure, as it has been fully and decisively met some time ago. I will also update the ‘Total expenditure‘ metric with an adjusted target figure, which will now represent the average total annual expenditure since 2013.

Current this is $85,200, but I will recompute this figure based on the full post-2013 monthly data at each portfolio update. In this way, the total expenditure measure will become a moving average, responding to either lifts or falls in expenditure. This means that if my average expenses rise over time, actual impact on progress will be better represented than the use of a static ‘once in time’ estimate that can become out of date.

I will report these progress percentages in future monthly updates. An example of this reporting, using the portfolio position on 1 January of this year as inputs, is set out below.

Measure	Portfolio	All Assets
Portfolio objective – $2,585,000 (or $90,500 pa)	87.6%	105.4%
Total average expenses – $85,200 pa	93.4%	112.0%

Trimming the sails – approaching flexibility

The resetting of the portfolio goal to a higher value, beyond that recently achieved, flows from a fundamental consideration about the nature of seeking financial independence.

This new target is not adopted as a binary measure, below which represents failure, and above which represents immediate or welcome exit from the workforce. Rather, at this point it represents a useful, salient measuring stick of a more subtle set of plausible combinations of degrees of financial freedom and potential future lifestyles.

The passing of the previous objective also meant careful re-examination of what was being measured. The evidence and studies above indicate clearly that while a 4 per cent return and safe withdrawal rate was not inherently unreasonable, it did imply and contain risks which I was personally not willing to bear at this time.

The progress of the past year, and its volatility has also highlighted the low probability that any finely adjusted projections and plans at this stage of the journey will play out. One set of specific circumstances – now unknown and unknowable – will prevail over the next crucial 5-10 year period, not simply an ‘average’.

Taking into account the best evidence on safe withdrawal rates, however, allows greater confidence that the portfolio will find itself robust to the task assigned to it than previous approaches.

The updating of the income estimate will ensure the maintenance of a real standard of living over time, which is based on an external measure. The inclusion of an actual total expenditure measure ensures longer-term changes in expenditure patterns are reflected in measures of progress over time.

These changes allow the confidence to know when the FI objective has been reached, in a manner consistent with both my personal risk tolerances, and the best available quantitative evidence. This evidence, is, however, always incomplete and backward-looking.

The prospect of this upfront confidence is more attractive than a false sense of security while taking short-cuts or entertaining unexamined risks. Likewise, ignoring available evidence now where adverse and irreversible risks and consequences may arise in future decades would be short-sighted.

Checking the stops – conditions of voyage

What is noticeable from the progress update above is that on an ‘All Assets’ basis both the revised portfolio goal, and the goal of financial independence based on total average expenses, are still met.

This highlights that the focus of this phase of the journey may ultimately turn out to be less about whether a particular portfolio figure is reached than previous phases. Rather, it may focus more on whether a number of broader conditions informing a decision to change the nature of my work in the future are met.

After some consideration, these conditions are:

1. Portfolio target – in particular, is the portfolio target met or exceeded, such that applying a 3.5 per cent safe withdrawal rate can be expected to be sustainable over a forty year period? The logic for this target and withdrawal rate is detailed above.
2. Minimum equity target – has a minimum equity holding of $1.93 million been achieved (this representing the target allocation of 75 per cent of the total revised portfolio goal), to ensure that the long-term core of the portfolio which is assumed to generate income and capital growth is able to serve its function? If the equity holdings are below this level, any decision to retire would be at appreciably higher risk
3. Cash reserve – is a capital reserve in place that is equivalent to at least one year of total normal expenditure, to minimise or avoid required portfolio withdrawals in future equity market drawdowns?

Currently, none of these three conditions are met. Each being met will be important consideration in any future early retirement decision.

So from here my approach will be to focus on the portfolio target (condition #1) and also meeting conditions #2 and #3 – seeking with all efforts the achievement of the target. Yet this will be done recognising that all past experience has demonstrated that volatility in the market, or changes in my career are likely to disrupt the specific portfolio goal and projected timeline.

The progress of the FI portfolio, and growth in passive income over the past four years gives me confidence that I will be able to adapt to a broad range of future events – which are both probable and not able to be forecast.

Reviewing the investment plan and assumptions

One has to pay attention to the different tides and to trim one’s sails according to the wind.

Confusion de confusiones, Joseph de la Vega (1688)

Each year I review my investment policy to ensure it stays relevant to available data and evidence, and will serve my ultimate investment goal of financial independence.

The summary is based on my past reviews and decisions. It is designed to serve as a reference for the chain of reasoning underpinning portfolio design and allocation decisions through the year – so it remains unchanged where my conclusions have not changed.

In 2019 I increased the equity allocation in the portfolio to 75 per cent. This remains the long-term target.

The most significant change following the review this year is to move towards a long-term target of an equal allocation of global and Australian equities, which is illustrated and discussed below.

Specific asset allocation targets

Based on the review and considerations below the portfolio allocation targets are as follows:

• 75 per cent equity-based investments, comprising:
  • 37.5 per cent international shares
  • 37.5 per cent Australian shares
• 15 per cent bonds and fixed interest holdings
  • 10 per cent Australian bonds and fixed interest
  • 5 per cent international bonds and fixed interest
• 10 per cent gold and commodity securities and Bitcoin
  • 7.5 per cent physical gold holdings and securities
  • 2.5 per cent Bitcoin

Reasons for allocation targets and assumed asset returns

Equity returns, safe withdrawal rates and international diversification

The equity component of the portfolio provides the fundamental engine of returns in the portfolio, with the best chance of outperforming other asset classes, and maximising after inflation returns.

Calculation of the overall portfolio target previously involved setting an assumed return for each asset class, however, this has been replaced this year with the use of the safe withdrawal rate.

To anchor forward expectations, however, I still find the estimation of assumed returns a question of interest. In terms of long-term real equity returns, last year I adopted a more conservative estimate of 4.9 per cent, based on the geometric mean of Australian equity returns over risk-free assets from 1883 to 2018. This value has been retained this year.

For my specific purposes, the geometric mean appears to continue to be more appropriate than the previously used average of the arithmetic and geometric mean, and its lower value also reflects lower expectations going forward.

For global equities the equivalent real return estimate is 5.2 per cent, a long-term historical figure sourced from the 2020 Global Investment Returns Study.

The split between Australian and international equities is designed to maximise total returns and minimise portfolio volatility, while taking advantage of the tax-advantaged nature of Australian franked dividends.

The equities sub-targets above have changed.

Previously, I sought to achieve a target 60/40 split between Australian and foreign equities, which this academic survey published in 2013 estimates to be optimal for most Australian investors (see Klement, Greenrod and O’Neill Optimal Domestic Equity Allocations for Australian Investors and the Role of Franking Credits published in the Journal of Wealth Management and also discussed previously here).

A key finding of the study is that Australian equity exposures at higher rates significantly increase portfolio volatility, and maximum potential losses.

The specific optimal 60/40 split suggested by the study is, importantly, a product of the historical data and the characteristics of volatility in past markets. As such, it retains value in application only to the extent that fundamental relationships between Australian and global equities have not changed since 2013. I have previously noted that as time continues, that assumption becomes more questionable, meaning that a 60/40 split provides no certainty of continuing to be optimal going forward.

This year I have changed the target allocation to a 50/50 split of Australian and foreign equities. This approach is adopted to:

• Assume a neutral position on relative equity market performance – Over the long-term in conceptual terms, an equal allocation reduces the portfolio risks of any future Australian equities underperformance compared to global equity indices.
• Recognise the after-tax benefits of franking credits – The remaining 50 per cent weighting to Australian equities still sits close to the optimal balance for reduced portfolio variance and maximisation of franking credit benefits suggested in the Klement et al paper discussed above
• Reduce portfolio variance – A 50/50 per cent weighting is supported as a minimum variance (i.e. lowest volatility) portfolio allocation by this 2012 Vanguard study (pdf).
• Recognise that pure ‘market capitalisation’ weighting is not required – Due to correlations between Australian and global markets, and the specific benefits provided by the current franking credit regime, it is not optimal to hold the Australian market at the low weighting (2-3 per cent) which would arise from a pure market capitalisation weighting approach.

The Vanguard study (pdf) mentioned above provides an excellent structured framework for individual investors thinking about these issues (see in particular Figures 4 and 11).

Having set this target, I do not intend to realise capital gains in seeking to immediately meet it, or even to target meeting it in the short-term by exclusively redirecting new investments and distributions to global shares. Rather, it sits as a longer-term goal to guide, but not fully determine, the pattern of future investments.

The reason for this is that exclusively switching to purchases of global shares across these particular relative market valuation and currency regimes would effectively embed a set of risks into the portfolio which are better diversified by regular investments over time. In this way, time itself can be employed as a diversifying factor, in a similar way as dollar cost-averaging.

This means that overall, the target for the achievement of this 50/50 allocation is longer-term, around 5-7 years.

Bonds and fixed interest

Bonds and fixed interest play a role in diversification, reducing overall portfolio volatility.

The assumed return of 2.0 per cent for these assets is in line with long term global averages measured since 1900, sourced from the 2020 Global Investment Returns Study and based on data from the Dimson, Marsh and Staunton book Triumph of the Optimists – 101 Years of Global Investment Returns. It is certainly difficult to see bonds producing returns close to these historical results in the near-term, following a multi-decade fall in interest rates that have acted as a tail-wind for returns.

A separate review of bond holdings in the portfolio and the relevant investment literature has reinforced the value of a small bond holding, but caused a slight adjustment in the target allocation from a simple equal weighting of Australian and foreign bonds, to a position that reflects the greater diversification benefits of international bonds.

Property

I have no formal property allocation, excepting my small exploratory investments in fractional residential real estate through BrickX. In the current market environment my assessment is Australian property is likely to enjoy low yields and returns for a considerable period, and not offer much diversification benefit over Australian equities or other asset classes.

Gold and Bitcoin

As described previously, the role of gold and Bitcoin in the portfolio are primarily as non-correlated financial instruments for diversification, and as an insurance against extreme capital market events or conditions.

No real return is assumed for either asset, and I plan to only rebalance by purchasing low cost gold index ETFs if the overall alternatives asset class falls well below its 10 per cent allocation.

Bitcoin has been included as an asset in the portfolio following the unexpected growth in value of a small exploratory investment (representing around 0.5 per cent of 2015 portfolio value) to a sizeable component of current overall portfolio value. It is still unclear what role Bitcoin may or may not play in investment markets, or as a store of value, and as such it remains a high risk and volatile component of the portfolio.

Overall long-term portfolio return estimate

Taking into account the above asset allocation and return assumptions, the overall portfolio return is estimated on a weighted average basis at 4.09 per cent, up from 3.99 per cent last year.

This is equal to a nominal return of 6.59 per cent based on an assumption of inflation being in the middle of the Reserve Bank’s target band over the long-term.

Though providing some guidance around expectations, as highlighted above this estimate no longer plays a determining role in setting the level of the portfolio goal. Rather, this goal is now calculated by reference to the safe withdrawal rate assumption.

With this annual review and trimming of the sails completed, I look forward to where unpathed waters shall lead the portfolio and journey.