Andrew Kuhn September 3, 2019

Why You Might Want to Stop Measuring Your Portfolio’s Performance Against the S&P; 500

By Geoff Gannon

December 8, 2017



Someone emailed me this question about tracking portfolio performance:

“All investors are comparing their portfolio performance with the S&P 500 or DAX (depends were they live). I have asked a value investor why he compared the S&P 500 performance with his portfolio performance…for me as a value investor it makes no sense. A value investor holds individual assets with each of them having a different risk…it’s like comparing apples and oranges.

The value investor told me that…Warren Buffett compares his performance with the S&P 500. But I believe he did it, because other investors…expect it or ask for such a comparison.

How do you measure your portfolio success? Do you calculate your average entry P/E and compare it with S&P500 or Dow P/E to show how much you (over)paid for your assets? Or do you avoid such a comparison and calculate only the NAV of your portfolio?”

I don’t discuss my portfolio performance on this blog.

And I think it’s generally a good idea not to track your portfolio performance versus a benchmark.

It’s certainly a bad idea to monitor your performance versus the S&P 500 on something as short as a year-by-year basis.


Well, simply monitoring something affects behavior. So, while you might think “what’s the harm in weighing myself twice a day – that’s not the same thing as going on a diet” – in reality, weighing yourself twice a day is a lot like going on a diet. If you really wanted to make decisions about how much to eat, how much to exercise, etc. completely independent of your weight – there’s only one way to do that: never weigh yourself. Once you weigh yourself, your decisions about eating and exercising and such will no longer be independent of your weight.

Knowing how much the S&P 500 has returned this quarter, this year, this decade, etc. is a curse. You aren’t investing in the S&P 500. So, tethering your expectations to the S&P 500 – both on the upside and on the downside – isn’t helpful. The incorrect assumption here is that the S&P 500 is a useful gauge of opportunity cost. It’s not.

Let me give you an example using my own performance. Because of when the 2008 financial crisis hit, we can conveniently break my investing career into two parts: 1999-2007 and 2009-2017.

Does knowing what the S&P 500 did from 1999-2007 and 2009-2017 help me or hurt me?

It hurts me. A lot.

Because – as a value investor – the opportunities for me to make money were actually very similar in 1999-2007 and 2009-2017. In both periods, I outperformed the S&P. However, my outperformance in 2009-2017 was small while my outperformance in 1999-2007 was big. In absolute terms, my annual returns were fairly similar for the period from 1999-2007 and 2009-2017. It is the performance of the S&P 500 that changed.

Many value investors have a goal to outperform the S&P 500. But, is this a useful goal?

I don’t think so. Let’s look at the 1999-2007 period to see why it’s not a good goal. The S&P 500 returned very little from 1999-2007. However, a value investor – like me – who was looking to own only 3-5 specific stocks at a time could have made 15% a year. It’s this number – something like 15% a year – that’s the rabbit you should have been chasing in 1999-2007, not the 5% or so the S&P 500 did.

Let’s assume the S&P 500 returned about 5% a year from 1999 through 2007. And then let’s assume that some value investor returned 8% a year during the same time period. This value investor is very proud of himself.

Should he be?

No. He blew it.

If you were an individual investor who made 8% a year from 1999-2007, you missed a lot of opportunities you should have taken advantage of. It was possible to make 15% a year without taking a ton of risk. And, so, you left something like 7% a year on the table over a period of 8 full years.

That kind of underperformance versus your potential rate of compounding has – when it occurs over a period as long as 8 full years – a significant (really permanent) influence on your lifetime investment outcome. If you had been making 15% a year from 1999-2007 you’d have $1.72 for every $1 you’d have if you were making 8% a year from 1999-2007. And yet, because you outperformed the S&P 500 – you chalked this period up as a win for you.

When we look at the period 2009 through 2017, we see the opposite problem. We especially see it in the last couple years. But, let’s look at the full period. Let’s say that from 2009 through 2017 you again made 15% a year.

This time you’d say you just about tied the S&P 500. You certainly didn’t beat it. And so, if you did 15% a year from 2009 through today, you’d say you don’t deserve any accolades at all.

Is that right?

Well, it depends on how you achieved it. The argument that people make about why you should track your results against the S&P 500 is that the S&P 500 is always a viable alternative. So, if you didn’t do much better than 15% a year from 2009 through 2017 – you didn’t add value.

Your effort was wasted.

There’s a logical problem here. As an investor, you don’t control outcomes. You only control process. Any benchmarking against historical performance is limited in the sense that it starts at one exact beginning point and finishes at one exact end point.

Basically, we are assuming clairvoyance on your part. I don’t just mean that we’re assuming you knew everything that could be known in 2009 and reasoned everything out correctly from there. I’m not just saying you were omniscient and infallible. I’m saying you could actually foresee the future free from uncertainty.

That’s no way to measure results.

I’m now going to take you on a very involved philosophical detour. But, it’s an important side trip to take if we’re going to understand that what we want to judge is always our process not our outcome even though the most readily available tool for measuring our process (indirectly) is reasoning backwards from our outcome.

Just remember this: we have no control over outcomes. We control only process. So, going forward, our goal is to fix our process – not our outcome.

In the long run: outcomes will follow process.


The Past is Only Fixed in Retrospect

People intuitively understand that the future is not fixed. However, they don’t always make the (correct) logical leap that the present we are now living was never a 100% certainty at any point in the past. In other words, even when you are able to correctly predict the future – you should never make a 100% bet on that future. Retrospectively, this means that while the particular past that got us to today’s present is mostly informative of what we should have done in the past – it is never fully informative. The moment we are in now is somewhere near the average of moments that might have been. It is not – all by itself – ever a perfectly precise measure of what the likely future was at any point in the past.

This is very important when considering something like the performance of the S&P 500 over the past year. Let’s say the S&P 500 returned 20% in 2017. How correct was a bet made in January 1st of this year that the S&P 500 would be 20% higher on December 31st? I’d say such a bet was a bad bet even though the outcome was positive. What I mean is this: rolling one die and betting 1-to-1 odds that it will come up 3 is a bad bet even if it does come up 3. A bet that the S&P 500 might return 10% this year could be an okay bet. But, a bet that the S&P 500 would return 20% a year – which turned out to be the truth – is still a bad bet. This isn’t obvious because you lived only one 2017. But, if you could live 1,000 2017s in a row one-after-another in some sort of set of parallel experiences – it’s not likely that the central tendency of those 1,000 different last years would work out anything like the last year we all just lived.


Why It’s Better Not to Know EXACTLY What the Dow Really Earned Last Year

About 11 years ago now, I took a look at returns in the Dow based on a smudged history approach. Instead of assuming that the EPS reported for the Dow in any one year was the inevitably “correct” EPS for that year (as if you could have foreseen that it always had to work out that way), I assumed that the best way to think of the Dow’s earnings in any one year was to look at the 15 years preceding it and then draw 15 lines – moving at 6% a year – forward in time till they reached the current year. You then – metaphorically – took your finger and smudged those 15 endpoints.

That’s the normal earnings for the Dow.

It’s the central tendency suggested by 15 points from the past rather than the actual observed point we’re at now (the present). As it turns out, using the central tendency suggested by 15 past points works much better than relying on one present-day point.


Never Measure One Point In Time

This sounds strange, counterintuitive, and overly complex. But, it’s actually a much more logical way to look at the historical level of anything. If I was shooting a pistol at a target and sometimes hit the left shoulder and sometimes the right and then maybe the head and once or twice the abdomen and then finally – with my very last bullet – I hit the target right in the heart…

Should we award you any points for guessing that I’d hit the heart with that last shot? In what way does it even make sense to say I hit the heart? I mean, I also hit both shoulders and the head and the gut.

So, what would constitute a good guess on your part?

If you correctly guessed the central tendency of my shots to cluster in some particular part of the target over 5 attempts, 50 attempts, 500 attempts, or 5,000 attempts – that would be something worth giving you credit for.

So, history is certain and precise. But, the certainty and precision with which we can measure the past is not necessarily useful. Over very long periods of time – for example, from 1999 all the way through 2017 – a comparison of your results versus the S&P 500 might tell you something.


What’s Easiest to Measure vs. What Matters Most

But, even then: does it tell you what you care most about?

Should the average person really be aiming to get a better performance than the S&P 500?

In terms of building wealth, it’s the long-term rate of compounding that matters.

I once broke this down as follows…

If you string together back-to-back-to-back 15-year periods of 5% annual returns: you aren’t going to achieve any of your long-term financial goals.

If you string together back-to-back-to-back 15-year periods of 10% annual returns: you may achieve most of your long-term financial goals if you make enough, are frugal enough, etc.

And finally…

If you string together back-to-back-to-back 15-year periods of 15% annual returns: you will achieve all your long-term financial goals.

Now, I’ve brushed over the issue of inflation there. But – aside from differences in the rate of inflation over your investment lifetime – what I’ve said is true.

What matters is getting a good absolute rate of compounding over 40-50 years. Getting a good relative rate of compounding over 4-5 years isn’t important.

Because some financial cycles – like interest rates, P/E ratios, bull markets, etc. – are so long: I suggest measuring your annual returns over 15-year intervals.

And – because you can’t eat relative returns – I suggest you think in terms of absolute returns.

If, over the last 15 years, you’ve done 5% a year: that’s not good enough.

If, over the last 15 years, you’ve done 10% a year: that’s fine.

And if, over the last 15 years, you’ve done 15% a year: that is good enough.

As an individual investor, your goal should be to try to do 2 things:

1)      Avoid doing anything that might get you returns as low as 5% a year for as long as 15 years

2)      Seize any opportunities that come along that seem likely to get you returns as high as 15% a year for as long as 15 years

When I say that: I’m talking about individual stock picks, strategies you can adopt, process improvements – everything. If you think it can get you to 15% a year over 15 years – go chase it down.

Okay. So far I’ve said:

1.       Measure your absolute returns

2.       Think in terms of 15-year intervals

Is that all you can measure?

No. You can measure the performance of individual stock picks.


Measure Your Individual Forced Outcomes

In fact, you will find that several of the picks you make will resolve themselves permanently in less than 15 years. Some stocks you pick will suffer a permanent impairment of intrinsic value – they may even end up in bankruptcy. Others will be acquired at a price higher than you paid for them.

For example, I bought a stock called IMS Health in 2009. That company went private. It’s public again. But, my investment was permanently resolved – by that private equity buyout – in the sense that I had no choice of whether or not to take a profit. I was forced to take a profit.

Likewise, in 2010, I bought a stock called Bancinsurance. About 9 months later, that stock was taken private. Again, I had no choice of whether or not to take a profit. I was forced to take a profit.

Before the financial crisis, Warren Buffett invested in some Irish bank stocks. They ended up being basically worthless. So, Buffett didn’t choose to take a loss in that stock. He was forced to take a loss.

You can always measure your performance in stock picks that permanently resolve themselves irrespective of your actions. You can look at the annual returns in those stocks.

For example, my return in Bancinsurance was about 40% in about a year. Do I spend time thinking: what was the return in the S&P 500 during the period when I held Bancinsurance stock?


I just think that any time you can find something that makes you about 40% a year – that’s a win. Now, there was obviously a chance that Bancinsurance could have ended up much, much worse. I couldn’t have foreseen a 100% probability of a 40% profit ahead of time. But, when you look back at what I knew when I made my decision to invest – you could imagine that, based on the probabilities of the situation, a return of 20% a year was a good guess. At the time I made the investment, the most likely outcome seemed to be a return of no less than 3% but no more than 60% in perhaps less than a year. The important thing was this: it seemed a good enough bet in the sense it was likely to return 15% or better annualized.

And, as an individual investor, that’s what you’re looking for. You’re looking for absolute returns. You’re looking to avoid anything – like an S&P 500 index fund today – that seems more likely to return 5% a year over the next 15 years rather than 10% a year over the next 15 years. And you’re looking to jump on any opportunities that seem more likely to return 15% a year rather than 10% a year.

So, for individual investors, I think benchmarking your results against an index is bad for two reasons: 1) The time period you care about is long-term (1-year results, 3-year results, even 5-year results aren’t going to be informative much of the time) and 2) The returns you care about are absolute returns not relative returns.

If you’re picking stocks for yourself and only yourself – your long-term, absolute rate of compounding is all that matters.


Why Individual Investors Have It Easy

Who should use benchmarking then?


Professionals are looking to attract and retain clients. Clients care about relative returns. Should clients care about relative returns? That’s a question for another day. But, they do. And they often care about short-term results. For example, I just read a blog post that said Bill Ackman’s results were “mixed” because:

At Pershing Square he significantly underperformed from 2015 to 2017 and outperformed from 2004 to 2014.”

Now, the period 2004-2014 is a lot longer than 2015-2017. And, more importantly, this blog post showed results from 2004-2016. Ackman’s cumulative results from 2004-2016 were a lot better than the S&P 500 (his benchmark). And yet, the general feeling is that Ackman’s track record is mixed. When people say “mixed” they mean his long-term results are good but his short-term results are bad. For an individual investor, that’s not a mixed track record. That’s simply a good track record. But, for a professional – that kind of track record means clients are going to pull their money.

So, professionals track performance versus a benchmark, because: 1) Clients care about short-term performance and 2) Clients care about relative performance.

There is one other reason why benchmarking makes sense for most professionals and yet doesn’t make sense for many individual investors.

This reason doesn’t apply to Bill Ackman (a concentrated investor). But it does apply to most money managers (diversified investors).

I try to own 3-5 stocks at a time. There are 500 stocks in the S&P 500. And there are more like 5,000 stocks that are investable for me. So, at any moment in time – I only need to say “yes” to something like 1% to 0.1% of all potential investments. Professionals who are diversified may – depending on how big the pile of assets they manage is – need to say “yes” to more like 1% to 10% of all stocks they learn about.

Let’s take an extreme case. Imagine someone is managing $10 billion in a diversified way. This manager will often need to stick to S&P 500 type stocks. And he will often need to hold 30-50 stocks. In reality, holding 50 stocks instead of 20 stocks isn’t going to add much diversification in terms of the resulting volatility versus the benchmark. But, the convention among professional money managers is often to hold up to 50 stocks for the purpose of diversification whether or not this can be shown to make a meaningful difference in the ups and downs of the portfolio or not.

A money manager who holds 30-50 stocks of S&P 500 sized companies needs to say “yes” about 6% to 10% of the time. This is 6 to 100 times more frequently than I have to say “yes”. As a result, the performance of such a money manager’s portfolio – whether he wants this to be the outcome or not – is going to be a lot closer to his benchmark from year-to-year.

Now, this is a position number / position size issue (that is, lack of selectivity issue) not a professional versus individual investor issue. A professional money manager who runs $10 million by allocating that portfolio to just 3-5 stocks would be in the same boat at the individual investor rather than the worst case professional (a big, diversified fund) I showed here.

In reality, a professional money manager who has to invest $10 million or maybe even $100 million and is willing to keep that in as few as 3-5 stocks is in the same position as an individual investor except for the fact that he has an unstable source of funding (his clients) who may pull money the second his short-term relative performance weakens.

So, benchmarking makes more sense for people who:

1.       Rely on other people’s money

2.       Own a lot of different stocks and

3.       Manage a lot of money

Benchmarking makes less sense for people who:

1.       Invest only their own money

2.       Own very few stocks and

3.       Manage a small amount of money

Most readers of this blog fall more into the second group (as I certainly do) than the first group.

Therefore, benchmarking is something you are probably better off eliminating from your investing process.

Personally, I don’t care what the S&P 500 does. I only care what I do. So, I do my best to ignore all benchmarks.