A common question about of the “4% rule” is, well, 4% of what? People like to anchor themselves to the all-time high value of their portfolio, but we can see from recent events that can be a shaky idea. I believe that you should always be prepared for stocks to fall by 50%, which means you could be taking 4% of two very different values. Folks shouldn’t act like they “lost $XX,XXX” when their stocks drop from an arbitrary all-time high, and they shouldn’t plan out the next 30 years of retirement income based on a single value either.

I’m currently reading the new book Rethinking Investing: A Very Short Guide to Very Long-Term Investing by Charles Ellis. He’s been in the industry a long time, but may be best known for his bestselling book Winning the Loser’s Game, first published in 1985.

In the book, Ellis proposed a potentially better way to set your Spending Rule in retirement. It’s not based on the most recent value of your portfolio, and definitely not the all-time high of your portfolio. Instead, he wants you to use the rolling average over the last 5-7 years. Then, you can add the 4% rule (or whatever).

In designing your own spending rule, first, average the year-end values of your assets over the prior several years (preferably more than five years) to dampen the impact of market fluctuations. Next, calculate what would be a prudent withdrawal of the averaged assets—likely 4–5%—to determine what dollar amount you can prudently withdraw from your current portfolio each year to cover some of your expenses.

This has the effect of smoothing out your annual withdrawals:

Averaging your assets over multiple years makes the funds available for your spending far more consistent and predictable. If, for example, you settle on a 5% rate of withdrawal and a six-year moving average of the year-end value of your assets, a 30% drop in the stock market would lead to only a 5% reduction in your payout that year (and much of that reduction likely would be provided by your consistent dividend income).

(Side note: This supports the idea of me tracking my consistent dividend income…)

Let’s take a look at Vanguard LifeStrategy Growth Fund (VASGX), an all-in-one fund that is diversified similarly to their Target Date Retirement Funds, but a handy benchmark since it is a constant 80% stocks/20% bonds. Here’s a Growth of $10,000 chart for the last 5 years ending 4/4:

Instead of seeing that you are about 11% off your all-time high value of about $18,300, you might appreciate that you are still above January 2024 levels, and that your 5-year rolling average of year-end values is about $15,400. If you based your 4% withdrawal rate on that value, you would be much calmer now.

I like this strategy, and I believe it should be applied even when you are still accumulating for retirement. Don’t anchor yourself the all-time high of your portfolio and make it your new “If it ever goes below this, I’ll be sad!” value. Instead, mentally track a rolling average of your net worth. I’ll look to add this concept to my portfolio updates, hopefully it’ll reduce my stress levels during volatile times.

Finally, Ellis points out another potential benefit:

Importantly, by following such a Spending Rule, you are then free to concentrate on achieving significantly higher long-term returns without the need to be overinvested in bonds. Stabilizing the investor’s income with a responsible Spending Rule frees the investment portfolio to invest more in equities and produces, over time, a higher and more rapidly rising portfolio value and income stream.

If you are a DIY investor (or professional financial planner) that is looking to geek out on the intricacies of the tax treatment for holding Treasury Inflation-Protected Securities (TIPS), check out the new paper Best Asset Location for a TIPS Ladder by Edward F. McQuarrie. I’ve been building a ladder of individual TIPS for many years, and have been extending it and filling in gaps during the recent period when long-term real rates went up to ~2.6%. Here is a chart of historical 30-year real rates (TIPS pay this much above inflation):

The paper focuses specifically on TIPS ladders, where you hold individual TIPS with staggered maturities such that when one matures each year, it creates a level, inflation-adjusted stream of annual income. The primary unique feature of this ladder is that it is guaranteed to adjust for inflation (as measured by CPI), even if it is higher than expected. Regular, nominal bonds don’t provide this protection. Of course, if inflation is lower than expected, then those nominal bonds will outperform TIPS.

The paper itself is very detailed and took a few readings to fully comprehend it all, but I definitely learned some new wrinkles. However, the overall conclusions are still useful to keep in mind if you hold TIPS. The question is, where is the preferred place to locate TIPS? In a regular taxable brokerage account? In a tax-deferred account like a pre-tax IRA or 401(k)? In a Roth IRA or 401k(k)?

Here are my takeaways, in my own words:

Individual, longer-term TIPS should be avoided if possible in a regular taxable brokerage account. This is primary due to the unique taxation of TIPS and the “phantom income” they make you pay upfront if there is inflation. You can look up “TIPS phantom income” for more details elsewhere, but the bottom line is that it’s hurts you upfront and you don’t catch up. Things only get worse at higher income tax rates, and higher inflation rates. It’s also just an extra annoyance at tax filing time.

The overall preferred location for TIPS is a Tax-Deferred Account (TDA). In other words, a pre-tax 401(K) or a Traditional pre-tax IRA where the tax is deferred but you pay taxes at ordinary income rates upon withdrawal.

It’s better to put stocks and REITs in a Roth account, so also not TIPS ideally. Roth accounts are great overall, but it’s best to take advantage of them by putting stocks and REITs inside as there is not as much added benefit for TIPS (or bonds in general).

The paper also discusses the wrinkles from state income tax and RMDs, but they don’t change the overall recommendation.

Here is a direct quote from the paper:

It follows that if the client has a more aggressive asset allocation, perhaps 2:1 stocks versus fixed income, with three accounts of roughly equal size, then stocks should first fill the Roth and then fill taxable. A TDA is always the best location among the three account types for a bond ladder, especially TIPS. Distributions are required from TDAs, and bond ladders produce distributions. Bond income is taxed as ordinary income, and distribution from TDAs are taxed as ordinary income. Characteristics of the bond asset and the TDA account are aligned.

The paper also states “The paper does not consider the best location for TIPS bonds or bond funds during the accumulation phase.” I would then add myself that if you do really want to own TIPS in a taxable account, you should consider a low-cost index ETF which is really sort of a ladder of TIPS than replenishes on its own with a roughly constant average maturity. For short-term TIPS, there is the Vanguard Short-Term Inflation-Protected Securities ETF (VTIP) with an average maturity of ~2.5 years. For a longer-term, there is the Schwab U.S. TIPS ETF (SCHP) with an average maturity of ~7 years. TIPS ETFs don’t expose you to the phantom income effect.

Again, this paper offered some additional insight for those so inclined. I hold all my individual direct TIPS in a pre-tax Solo 401(k), so I am following the advice. I am not building a strict ladder, so if I ran out of room in tax-deferred accounts, I would hold a TIPS ETF in a taxable account.

Photo by Nick Page on Unsplash

As a companion to my post on Fidelity Model ETF Portfolios, I also found Blackrock’s version of their 60/40 Model ETF portfolio.

The was prompted by the fact that Blackrock recently announced that it was adding a 1-2% allocation to Bitcoin in their model ETF portfolios.

The world’s biggest asset manager is finally allowing Bitcoin into its $150 billion model-portfolio universe.

BlackRock Inc. is adding a 1% to 2% allocation to the $48 billion iShares Bitcoin Trust ETF (ticker IBIT) in its target allocation portfolios that allow for alternatives, according to an investment outlook viewed by Bloomberg.

Of course, this coincided with the fact that last year they finally launched their own Bitcoin ETF, the iShares Bitcoin Trust ETF (ticker IBIT). That made me wonder, what exactly does Blackrock put into these model portfolio that are meant for advisors? The model portfolio below does not have the Bitcoin ETF added yet:

As with the Fidelity model portfolio, and probably all model portfolios meant for advisors, there is the appearance of technical complexity, with a lot of tiny allocations to ETFs to bump the total number involved to 18 different ETFs and cash (and possibly the new Bitcoin ETF as well). 1% to the iShares US Infrastructure ETF? 1% to iShares J.P. Morgan USD Emerging Markets Bond ETF? 1% to iShares Gold Trust?

However, what surprised me the most was hidden in their performance stats at the bottom. With a relatively low net weighted expense ratio of 0.16%, their gross overall performance (before all fees) was pretty good and hugged the benchmark indexes very closely. However, they had to disclose that their NET historical performance (what clients actually got) was a lot lower… why was it so much lower? Because their managed portfolio apparently comes with a 3% annual fee, charged quarterly!!!

Tucked deep at the bottom:

Net composite returns reflect the deduction of an annual fee of 3.00% typically deducted quarterly. Due to the compounding effect of these fees, annual net composite returns may be lower than stated gross returns less stated annual fee.

So you put your Managed Portfolio clients in a low-cost ETF portfolio, and then add a 3% annual fee on top. Wow, that’s… wow. I have trouble even believing it. I must be reading this wrong.

Another interesting note is that Vanguard’s new CEO, Salim Ramji, was the former global head of iShares and index investments at BlackRock and thus very involved in their push into model ETF portfolios and probably had a big hand in designing them. Will he adjust Vanguard’s suggested portfolios in a similar manner?

The Canada Pension Plan Fund (CPP) is one of the two major components of Canada’s public retirement income system, along with Old Age Security (OAS). The CPP mandates that all employed Canadians age 18+ to contribute a certain percentage of their earnings (with an match contributed by their employer) to the CPP, managed by a group called CPPI. (Source: Wikipedia.)

I learned all of this because the CPP has become an interesting example where we can compare an investment manager that has chosen to switch to the high-cost, “we can do better because we are smarter” philosophy: lots of highly-paid employees, lots of highly-compensated hedge fund and private equity managers, lots of fees paid, all in search of higher returns. Luckily, we can see if they succeed because they have to publish their results for all to see.

My sources are two interesting articles and the CPP 2024 Annual Report:

• Opinion: Eighteen years and $46-billion later, the CPP admits it could have earned more just by buying index funds [Globe and Mail]
• Canada Pension Plan Falls Short of Indexed Benchmark [Center for Retirement Research at Boston College]
• CPP Annual Report 2024

In their fiscal 2024, the CPP paid C$3.5 billion in fees to external investment managers. (The fees paid in 2006 were just C$36 million. As in only C$0.036 billion, 100 times less!) The pension fund itself has grown to over 2,000 employees (up from only 100 employees in 2006), and after adding all operating expenses and transaction costs, the fund’s total expenses now exceed C$5.5 billion annually.

The total assets are roughly C$630 billion. C$5.5 billion of costs on C$630 billion of assets means the fund’s annual expenses eat up 0.87% of the total assets every year. That is creeping very close to 1% of assets annually.

What have those costs bought? Not much so far. In fact, the 5-year performance lag in returns as compared to a passive benchmark portfolio is actually higher than that. The CPP chooses its own Reference Portfolio to match up with their mix of hedge funds and private investments, and it has shifted over the years going from 65% Global Equities/35% Bonds in 2015 to 85% Global Equities/15% Bonds in 2024. (Specifically, 85% MCSI World Index and 15% Canadian government bonds.)

After many pages in the CPP Annual Report explaining their very fancy system and why they believe they will outperform… here’s the one chart that shows their actual value-added. This is their own chart and language.

The CPPI says that we should be okay with this lag, partially because they are so “resilient” during market downturns. This is an often-cited reason for underperformance, but I question it on two levels.

First, with many of these illiquid investments, the values are essentially self-reported. Private REITs always have lower volatility than publicly-traded REITs because they get to report their own net asset value. What’s the value of a building or business that hasn’t actually sold on the open market? Who really knows? Sure, the numbers have to be within reason, but otherwise they are easily fudged. Second, you could have gotten lower volatility by simply holding a little less stocks and a little more bonds. That would have also been more resilient.

I’ve also read the follow-up defense pieces, but I wasn’t really swayed. It’s all the same old stuff. The benchmark wasn’t really a good benchmark (in retrospect), even though they picked the benchmark themselves. Our performance beat our arbitrarily-set target (’cause everything went up), even if it lagged the benchmark. You have to pay up for smart helpers! Don’t you understand?! “I have people skills!”

(Counterpoint: It’s not common, but it can be done with less bloat and lower fees. The Public Employees’ Retirement System of Nevada (NVPERS) is an example of a pension fund that uses low-cost index funds for all of their publicly-traded asset classes. They have two employees. Their overall fees are 0.13%, mostly because they do hold about 12% in private assets. Their trailing 1-year performance as of 9/30/24? 20% annualized. Source.)

Right now, the alarms are not ringing for the CPP because the markets are up a lot and they are generating solidly positive returns even if they lag the market by 1% or 2% annually. I will be on the lookout for future updates on the CPP to see if they can justify their high cost structure over the long run. In the meantime, perhaps Canadian taxpayers should re-read Warren Buffett’s parable warning us about expensive Helpers.

It’s been a while since Vanguard announced a big round of expense ratio cuts, and just in time for their upcoming 50th anniversary, they announced what they call the largest fee cut in Vanguard history spanning 87 funds and projected to save fundholders $350 million this year (effective 2/1/25). In this regard, I am not worried about Vanguard. They know that low costs are core to their identity.

At Vanguard, we believe our funds’ impressive long-term performance owes much to their low costs. For the 10 years ended December 31, 2024, 84% of our funds outpaced the average results of competing funds. The performance of our actively managed fixed income funds has been especially strong: 91% of our active bond funds and 100% of our money market funds have outpaced their peers’ average results.

In fact, I worry that they focus on low costs too much. Cuts are nice, but these expense ratio cuts mean probably 0.01% for the average investor, or $1 a year per $10,000 invested. I’d much rather Vanguard keep the 0.01% and spend it on maintaining a highly-trained, long-tenured staff. There is a palpable difference when talking with a typical Fidelity employee vs. Vanguard employee. I think Vanguard is heading in the right direction, but their staffing still feels much less experienced.

Here are the largest funds and ETFs with expense ratio drops:

I would point out that the ETF version of some of the mutual funds are still slightly cheaper. For example, BND is at 0.03% while VBTLX went from 0.05% to 0.04%. Again, small margins, but you can hold Vanguard ETFs easily at any brokerage and I like that optionality.

Here are a limited sample of funds that I have held in the past that were affected:

• Vanguard Total International Stock Market (VXUS) lowered to 0.05%.
• Vanguard Treasury Money Market Fund (VUSXX) lowered to 0.07%.
• Vanguard California Municipal Money Market Fund (VCTXX) lowered to 0.12%.
• Vanguard Intermediate-Term Treasury ETF (VGIT) lowered to 0.03%.
• Vanguard FTSE Emerging Markets ETF (VWO) lowered to 0.07%.
• Vanguard FTSE All-World ex-US ETF (VEU) lowered to 0.04%.

Here are the current expense ratios on the four broadest ETFs + their classic S&P 500 ETF:

• Vanguard Total US Stock Market (VTI) at 0.03%.
• Vanguard Total International Stock Market (VXUS) at 0.05%.
• Vanguard Total US Bond Market (BND) at 0.03%.
• Vanguard Total International Bond (BNDX) at 0.07%.
• Vanguard 500 Index (VOO) at 0.03%.

I find it interesting that one of the last places where the “Vanguard Effect” hasn’t shown up is in money market funds. They still have consistently the best and cheapest default cash sweep money market funds, Treasury money market funds, and Municipal money market funds. They also have some of the best muni funds in general. They are expanding their bond ETFs, but for now access to these alone is a strong reason to stay with Vanguard as a brokerage if you are DIY investor.

Vanguard recently released their most recent annual 10-year forecast as of mid-November 2024 (effectively the beginning of 2025). The beginning of the year is the time for forecasts, and that also makes it a good time to remind ourselves how badly they can be wrong and how you shouldn’t really use them for anything.

Let’s look back at how those same forecasts performed from 2011-2021, with confidence ranges within the 25th and 75th percentiles. I have some old images saved from when Vanguard gave us an update in 2021.

For US Stocks between 2011-2021, their forecast in 2011 was roughly between 6% and 12% annually for US stocks, for a median around 9%. That a wide band! The actual return? 13.4%. As of early 2025, we are still outside their confidence bands.

For Global ex-US Stocks between 2011-2021, their forecast in 2011 was roughly between 6% and 11% annually for US stocks, for a median around 8.5%. The actual return? 4.0%. As of early 2025, we are still outside their confidence bands here as well.

I’m not trying to pick on Vanguard here, but they do release these things with a certain degree of seriousness and brand authority. But honestly, I wish they wouldn’t. I mean, sooner or later they’ll be correct, but how could you possibly attribute that to skill and not luck?

I’m going to include a copy of their late 2024 10-year forecasts (close to the start of 2025) here because they usually delete the post after a couple of years. This way, we can look back again in the future. For this chart, the ranges are their median forecast with a fixed 2% range of confidence for stocks and 1% range for bonds.

Notably, the 10-year median return forecast is 3.8% for US stocks, 7.9% for Global ex-US stocks, and 4.8% for US total bond. This table includes their percentile confidence ranges.

This all reminds of me of the old joke: How can you tell economists have a sense of humor? They use decimal points.

I’ve been a Vanguard customer (ahem, owner) for 23 years now. As such, I’ve also been one of those long-time customers that has been disappointed to see their struggles with customer service for their individual retail brokerage clients. One of my big decisions in 2024 was if I would move the majority of my assets to Fidelity or Schwab. Here’s another long-winded post about my thoughts about Vanguard.

Jack Bogle made a powerful decision when he created the Vanguard ownership structure. Each of the mutual funds was its own entity, and the shareholders own the funds. In turn, the member funds own the umbrella Vanguard Group. The member funds each pay their own expenses for research, management, etc. Everything is “at-cost”. There are no outside shareholders that may call for profits to put aside or dividends to be paid out to them. In theory, this means that the goals of each individual retail investor are aligned with the Vanguard executives.

However, in practice, we are entirely passive shareholders in that we have no vote over who is CEO, who is on the Board of Director, how much each of those folks gets paid (we don’t even get to see the actual number), whether the company should prioritize customer service or growth of assets or employee benefits. As with many large non-profits, the executives at Vanguard get very large compensation packages and the target is almost always growth, growth, growth. Bigger is better; more assets means the executives can justify a larger paycheck.

When I started with Vanguard, they were much smaller and there was more “fat” in the system. Their expense ratio for the flagship S&P 500 index fund something like 0.20% annually ($20 a year for every $10,000 invested). ETFs did not exist, and mutual funds usually charged users a transaction fee unless they were on a “mutual fund supermarket” with a pay-to-play structure. In turn, this made mutual funds more expensive because they passed the costs onto the customer. Vanguard refused to pay kickbacks because that would increase the costs to shareholders, so us retail investors had to go “direct” to Vanguard to get access with no transaction fees. The cheapest option was to go direct with Vanguard, and they had a “cheap and cheerful” reputation. They weren’t the best in customer service, but phone calls were answered promptly.

Then came the exchange-traded fund (ETF). ETFs were cheaper to maintain for Vanguard (and everyone else). This drove costs even lower. ETFs could be bought and sold at any brokerage with the same transaction costs as a stock. ETFs also had inherent tax-advantages that made it much easier to avoid creating capital gains distributions. I believe a big break happened when Vanguard stopped holding the mutual fund and ETF expense ratios at the exact same level. Everyone was incentivized further to hold the ETF version.

Today, the expense ratio for the flagship S&P 500 index mutual fund is only 0.04% annually ($4 a year for every $10,000 invested). But the ETF version is only 0.03% annually ($3 a year for every $10,000 invested). There are both certainly much cheaper than 20 years ago, but today each of their ETFs also has at least two other competitors at the same low expense ratio. Vanguard probably feels forced to keep their ETF costs as low as possible, lest they hurt their “low-cost” brand.

However, since each Vanguard fund has to pay for its own expenses including customer service costs, Vanguard is now incentivized to have you hold your ETF at another brokerage. (And once you start holding Vanguard ETFs in another brokerage, technically you should be rooting for lower costs and thus Vanguard to spend less on customer service as well.) Trades are zero now everywhere. But every single customer service call still has to be paid for somehow, and from this perspective, you can begin to understand why their customer service has gone downhill. Their margins are purposefully thin and the only solutions are to either raise their expense ratios a tiny bit (slower growth and perhaps lower executive salaries) or just try to keep spending as little as possible on customer service.

Guess which one they picked? From WSJ article (gift) Vanguard’s Die-Hard Customers Have a Message for New CEO: ‘The Service Is Abysmal’:

Brokerage-account customers were also recently warned that “excessive reliance on phone associates” could lead to additional fees or account termination.

Importantly, Vanguard has limited ways to subsidize the low costs of their ETFs. Meanwhile, Fidelity still makes a ton of money upselling customers to a variety of wealth management services. Schwab earns hundreds of millions extra by quietly paying nearly zero interest on their cash sweep (they recently dropped it to 0.05% APY in December 2024), pocketing an average of 2% to 3% annually on their customer’s idle cash. Robinhood lets me trade random crypto 24/7 and promotes active trading which results in an insane amount of payment for order flow.

Is this the natural end for Vanguard? Will they just make the commodity product and let others distribute it and deal with customers?

This is why the new CEO will undoubtedly have a big focus on low-cost wealth management. This will allow them to charge customers a higher fee for increased financial advice and customer service. They would finally have something to upsell. The only other alternative is for them to raise Vanguard brokerage fees so that the retail customers pay directly for the additional services they require.

In the end, I asked myself, “If something happens to me, would I rather have my wife deal with Vanguard, Fidelity, or Schwab?” She may end up wanting to pay for extra assistance and advice. Vanguard would have the worst customer service, but perhaps they will come up with a reasonable-cost advisory system. Fidelity and Schwab would undoubtedly be happy to provide her additional financial advice as well, likely at a higher price. Fidelity has solid customer service in my opinion, but I don’t really like their wealth management options based on my past experiences helping older relatives. Schwab has a conveniently-located physical branch near us, but I have a bad taste in my mouth after their “Intelligent Portfolios” zero-interest-cash-is-good-for-you fiasco. (See CBNC article Charles Schwab to pay $187 million to settle SEC charges that it misled robo-advisor clients on fees.)

In the end, I have punted my decision and only made some smaller moves. I transferred our Vanguard IRA assets over to Robinhood as a 5-year test run (in exchange for $16,000). I already have my Solo 401k and an active Cash Management Account at Fidelity. Perhaps Vanguard will shore up their customer service to “decent enough” and use AI to create a at-cost/low-cost advisory platform for the masses. Perhaps the Fidelity model of solid customer service plus a whole bunch of both low-cost and higher-cost menu items is the best one. Perhaps I will place the highest value on a local Schwab human rep.

Image credit: Canva AI generated with prompt “HMS Vanguard in rough seas”

The most common reader question about my personal portfolio is definitely the fact that the allocation to international stocks has been a drag on performance relative to owning 100% US stocks. This is kind of a repeat topic, so I won’t dive into the full debate again, but wanted to offer some expanded and updated thoughts to my response to a comment from reader John. All numbers below are taken as of January 2025.

The divergence between the performance of US stocks and the rest of the world started around 2009, which of course coincided with most of my investing lifetime so far. 😒 Here’s a chart from the Bloomberg article Global Diversification Has Disappointed. Don’t Give Up on It (gift article for next 7 days). Worth a read.

As noted in my portfolio updates, my asset allocation floats along with total world market cap breakdown, as tracked according to the Vanguard Total World Stock ETF (VT). I remember a time when it was only 45% US and 55% Rest of the World (World ex-US). As of the end of 2024, it is now at 65% US and 35% World ex-US.

In practical terms, this means that I used to own about the same amount of Vanguard Total US Stock Market ETF (VTI) and Vanguard Total International Stock ETF (VXUS), a 1:1 ratio. But as of the end of 2024, I now own about double the amount of VTI relative to VXUS, a 2:1 ratio. So my performance isn’t exactly that of the chart above due to ongoing investments over time, but it’s still been much lower than if I used owned 100% US stocks.

I can’t change the past. The question is: Should I change my asset allocation now?

Let’s look closer. A significant chunk (not all) of the outperformance has been due to a higher P/E ratio. Below is a Yardeni chart of the P/E ratio of US stocks vs. International stocks. The gap looks like the greatest in 25 years. Can this trend continue? I don’t know, and I don’t think anyone really knows.

Are US stocks simply a better investment, forever? They might be. The US definitely offers a very business-friendly environment overall. That’s why I just let it float. If the US manages to continue this outperformance in the future, then one day my allocation might grow to 75%/25% (3:1 ratio) or even 80%/20% (4:1 ratio). My portfolio adjusts.

This is the same theory as owning all of the companies in a market-weighted S&P 500 index fund: you own all the winners, and you also own the losers, but owning the winners is good enough to pull everything up overall. If the US keeps being a huge winner, I’ll own a lot of the US. If not, I still own the entire haystack. Therefore, I plan to continue holding a chunk of international stocks according to the investable market-cap float with maybe a slight home bias.

I could sit here and lament how big my portfolio would have been if I had bet on 100% US for the last 15 years, but honestly the stock markets have been kind to me as a business owner (although I’d say at the expense of the average worker bee) even with my international stocks and bond holdings. The 10-year trailing average annualized return has been 9.33% for VT vs. 12.50% for VTI. Owning a mix of winners and losers has still worked out just fine, and I was covered in case history turned out differently. I have no complaints.

Photo by Andrew Neel on Unsplash