Better definition of investment risk

[OverThinkMuch]

All measures of investment risk I've seen rely on volatility. If the market drops suddenly, that's considered more risky than the absence of a market drop. Yahoo shows R squared, standard deviation, Sharpe ratio, and Treynor ratio. For example, the S&P 500:
https://finance.yahoo.com/quote/SPY/risk?p=SPY

In theory, if an investment makes +50% more with +100% more volatility, it's "more risky". But this risk, as I understand it, is about selling after a market drop, and missing out on the gains in a recovery. In both the dot-com crash and 2008 crisis, many investors sold while I kept my investments. So does that mean investors can be sorted by risk tolerance? Do many investors gain more risk tolerance over time?

Looking at an R squared or Sharpe ratio, despite investing for decades I don't know what it means to me. It seems like the use of volatility to measure risk seems rather incomplete. I wish there was a better definition that included both investor and the investment.

 

[njhowie]

Looking at an R squared or Sharpe ratio, despite investing for decades I don't know what it means to me. It seems like the use of volatility to measure risk seems rather incomplete. I wish there was a better definition that included both investor and the investment.

Basically looking at your personal volatility/returns versus the market or other index. Are you getting more stable returns relative to the market or "your" index? Or are your returns less stable (more volatile)?

If you are with Fidelity, on the Performance tab they give a nice chart and table (click the Show Details button) on the bottom for Risk/Return using R squared and Sharpe Ratio. Study it a little and play with the indexes it uses for comparison - it will all make sense pretty quickly.

 

[Lawrencewendall]

If you are with Fidelity, on the Performance tab they give a nice chart and table (click the Show Details button) on the bottom for Risk/Return using R squared and Sharpe Ratio. Study it a little and play with the indexes it uses for comparison - it will all make sense pretty quickly.

I just discovered this and it shows I would have lower risk and higher returns with just S&P index. Not sure what time-frame they are looking at. May cause me to rethink my strategy.

 

[njhowie]

I just discovered this and it shows I would have lower risk and higher returns with just S&P index. Not sure what time-frame they are looking at. May cause me to rethink my strategy.

At the top left of the graph, and right below the bolded word Timeframe there is a dropdown where the choices are 3 Year and 5 year (for me). If you've had your account with Fidelity for longer or shorter it probably gives more or fewer selections.

 

[Out-to-Lunch]

All measures of investment risk I've seen rely on volatility. If the market drops suddenly, that's considered more risky than the absence of a market drop. Yahoo shows R squared, standard deviation, Sharpe ratio, and Treynor ratio. For example, the S&P 500:
https://finance.yahoo.com/quote/SPY/risk?p=SPY

In theory, if an investment makes +50% more with +100% more volatility, it's "more risky". But this risk, as I understand it, is about selling after a market drop, and missing out on the gains in a recovery. In both the dot-com crash and 2008 crisis, many investors sold while I kept my investments. So does that mean investors can be sorted by risk tolerance? Do many investors gain more risk tolerance over time?

Looking at an R squared or Sharpe ratio, despite investing for decades I don't know what it means to me. It seems like the use of volatility to measure risk seems rather incomplete. I wish there was a better definition that included both investor and the investment.

Coincidentally, a contemporaneous thread about Portfoliovisualizer references another measure he terms "the Ulcer Index": https://portfoliocharts.com/2017/11/01/the-ulcer-index-is-a-helpful-way-to-quantify-portfolio-pain/. I do not know if it is a good measure or not, but may be worth a look to you.

 

[Lsbcal]

All measures of investment risk I've seen rely on volatility. If the market drops suddenly, that's considered more risky than the absence of a market drop. Yahoo shows R squared, standard deviation, Sharpe ratio, and Treynor ratio. For example, the S&P 500:
https://finance.yahoo.com/quote/SPY/risk?p=SPY

In theory, if an investment makes +50% more with +100% more volatility, it's "more risky". But this risk, as I understand it, is about selling after a market drop, and missing out on the gains in a recovery. In both the dot-com crash and 2008 crisis, many investors sold while I kept my investments. So does that mean investors can be sorted by risk tolerance? Do many investors gain more risk tolerance over time?

Looking at an R squared or Sharpe ratio, despite investing for decades I don't know what it means to me. It seems like the use of volatility to measure risk seems rather incomplete. I wish there was a better definition that included both investor and the investment.

My comment addresses the red above. I don't really think about risk in terms of a math definition. To me it's about having a nice pile of money to not have to worry about future spending needs. I did not sell in 2008-2009 but I do not expect the next really bad decline to be similar. It could be much much worse. Or not.

So I have a Plan B which acknowledges my own personal preferences. It involves selling some or even all equities. Will depend on the factors existing at that time point ... which I hope never happens.

Such risk will be different depending a lot on your net worth. An investor with 10 million who spends maybe 150k a year will probably react differently then an investor with 1 million and who spends 100k a year. For the wealthier investor loosing 60% would be ugly but livable. Not so for the other guy.

 

[chassis]

All measures of investment risk I've seen rely on volatility. If the market drops suddenly, that's considered more risky than the absence of a market drop. Yahoo shows R squared, standard deviation, Sharpe ratio, and Treynor ratio. For example, the S&P 500:
https://finance.yahoo.com/quote/SPY/risk?p=SPY

In theory, if an investment makes +50% more with +100% more volatility, it's "more risky". But this risk, as I understand it, is about selling after a market drop, and missing out on the gains in a recovery. In both the dot-com crash and 2008 crisis, many investors sold while I kept my investments. So does that mean investors can be sorted by risk tolerance? Do many investors gain more risk tolerance over time?

Looking at an R squared or Sharpe ratio, despite investing for decades I don't know what it means to me. It seems like the use of volatility to measure risk seems rather incomplete. I wish there was a better definition that included both investor and the investment.

A better definition exists. An investor needs to learn how to assess and choose investments. This leads the investor to holding individual names and generally away from indexing.

Risk metrics:
- all of the aforementioned ones in the original post
- currency risk
- country risk
- counterparty risk
- technology risk (ask the buggy whip manufacturers about this)

More can be added. In the end the investor needs to study and understand each of these. Each risk element is present in each investment, to a greater or lesser extent. Then the investor needs to gird the lions and make choices to build and manage the portfolio.

 

[OverThinkMuch]

Most of my assets are at Vanguard, which doesn't have much in the way of tools. Another example is my Roth IRA, which didn't lose much then suddenly shot upwards. Anything very profitable is considered volatile and therefore risky by most defintions.

I couldn't find Tesla's standard deviation on Yahoo or Morningstar, but I think it illustrates the point. It has a 5 year standard deviation of 57%. Did it ever lose 57% in the past 5 years? No - the volatility is based on upwards moves - and that, I claim is not risk.
Tesla Inc (NASDAQ: TSLA) Stock Report

 

[OverThinkMuch]

A better definition exists. An investor needs to learn how to assess and choose investments. This leads the investor to holding individual names and generally away from indexing.

Risk metrics:
- all of the aforementioned ones in the original post
- currency risk
- country risk
- counterparty risk
- technology risk (ask the buggy whip manufacturers about this)

I am not talking about the risks listed in a prospectus. The places which list currency and country risk do not mention Sharpe ratio as a risk. I'm referring to a general metric for measuring risk, not the specific ones you list here.

 

[OverThinkMuch]

Basically looking at your personal volatility/returns versus the market or other index. Are you getting more stable returns relative to the market or "your" index? Or are your returns less stable (more volatile)?

I think you're restating the conventional definition, and I'm disagreeing with that definition. If my investments are much more volatile, but I'm not selling, where is the risk? Did I lose something when my investments dropped and recovered?

I think there's a level of risk I'm ignoring. Take an example of an investor who checks their investments in Jan 2020, then again in Jan 2021. If they ignore the entire year, they never see a pandemic fueled drop of -35% from peak to bottom. They might login or get an annual report and just say "oh, 18%, that's a good year." Where is the risk in that scenario?

 

[Closet_Gamer]

I think you're restating the conventional definition, and I'm disagreeing with that definition. If my investments are much more volatile, but I'm not selling, where is the risk? Did I lose something when my investments dropped and recovered?

I think there's a level of risk I'm ignoring. Take an example of an investor who checks their investments in Jan 2020, then again in Jan 2021. If they ignore the entire year, they never see a pandemic fueled drop of -35% from peak to bottom. They might login or get an annual report and just say "oh, 18%, that's a good year." Where is the risk in that scenario?

I think risk is situational and therefore defies a single specific number to sum it up.

To your example, someone with a 3+ year investment horizon faces very little risk from equity volatility.

That same person who is intending to use that money to pay for a college education in 6 months faces substantial risk of not being able to meet a financial commitment due to volatility.

To someone with a well diversified, well funded portfolio the risk of loss from buying Carnival Cruise Line bonds in June of 2020 may have been quite modest in exchange for upside. To someone with a more modest portfolio, the absolute risk of default would have dramatically increased the level of risk.

I have my portfolio intentionally broken into pieces and I manage the AA in each piece based on timing and injury if something is missed.

Kids college funds being dispersed over the next 5-7 years are in bonds.

Money for a beach house I want to buy in a year is conservatively invested but chasing a little yield. Very little risk of injury.

Retirements funds at about 80/20 stocks bonds. Won't be touching those for 15 years.

 

[chassis]

I am not talking about the risks listed in a prospectus. The places which list currency and country risk do not mention Sharpe ratio as a risk. I'm referring to a general metric for measuring risk, not the specific ones you list here.

Are you looking for a universal risk metric? I think you are at risk of overthinking it.

 

[OverThinkMuch]

Are you looking for a universal risk metric? I think you are at risk of overthinking it.

I'm looking for a replacement for Sharpe ratio and standard deviation. Typically I see investment risk equals volatility.

Some people can't tolerate 30% equities, and some ride through crashes with 100% equities. Maybe investors need Sharpe ratio tolerance, to adapt it to them?

 

[big-papa]

Coincidentally, a contemporaneous thread about Portfoliovisualizer references another measure he terms "the Ulcer Index": https://portfoliocharts.com/2017/11/01/the-ulcer-index-is-a-helpful-way-to-quantify-portfolio-pain/. I do not know if it is a good measure or not, but may be worth a look to you.

Ulcer Index has become my go-to risk measure these days. The portfoliocharts site is a very good place to start - it also has a link to a website from the person who came up with it which, in turn, has a link to some excel code. It measures the drawdowns from the previous peak and then calculates a root mean square of all of the drawdowns across whatever timeframe you're interested in. Best not done with just annual data - for example, if you did that, you'd miss much of the the fun in March/April of 2020.

Anyway, it's as good as anything I've seen if you just want to roll up things into a single number. I prefer to look at how things would have actually unfolded in time. Besides, for me "risk" means the inability to actually meet my goals. And that could mean a lot of different things for different people who are in different situations.

 

[chassis]

I'm looking for a replacement for Sharpe ratio and standard deviation. Typically I see investment risk equals volatility.

Some people can't tolerate 30% equities, and some ride through crashes with 100% equities. Maybe investors need Sharpe ratio tolerance, to adapt it to them?

Investors need to learn how to assess a business (stock) and either buy it, or don't buy it. Learning how to assess a business is multivariate calculus and cannot be sufficiently boiled down into one metric/KPI. Volatility (standard deviation) and Sharpe ratio are good, and I use them. They are however not sufficient.

 

[OverThinkMuch]

Ulcer Index has become my go-to risk measure these days ...
Besides, for me "risk" means the inability to actually meet my goals. And that could mean a lot of different things for different people who are in different situations.

But do you see how those two things don't align? If the stock drops 10% and recovers, that adds to the Ulcer index, right? But how does that hurt your goals?

That problem is what I'm trying to capture - and maybe people relying on their portfolio in retirement have a different view than those accumulating.

 

[all4j]

But do you see how those two things don't align? If the stock drops 10% and recovers, that adds to the Ulcer index, right? But how does that hurt your goals?

That problem is what I'm trying to capture - and maybe people relying on their portfolio in retirement have a different view than those accumulating.

The more I read, the less I understand what "the problem" is. So several thoughts that might prompt more questions than answers. But 1st, congratulations on thinking about risk. After several years of central bank intervention trying to 'eliminate' risk, too many think they understand investing & too few think about risk.

As illustration of what is confusing to me, I'll pick on your example. In this case, I understand the risk under discussion is about one asset in a portfolio. The portfolio should have a goal & each asset in the portfolio will have one. So, this one asset fell 10% then 'recovered' (presumably got back to starting point?). How does that "hurt your goal"? So, what if my goal for that asset was to provide stability & protect my portfolio from a margin call. But, that temporary dip did trigger a margin call....& so it hurt my goal.

I don't think it has to do with accumulation vs retirement. Rather, understanding these data points & how they are tools to use. It does help to have a definition of risk -- where the thread started. When I think about risk, I'm thinking about meeting my future goals. Any measurement of past performance won't magically tell me future results. Metrics such as sharpe ratio are really used to compare 2 assets/portfolios for a given time period. Taken in isolation a sharpe ratio for 1 stock won't tell me how it will perform.

If risk can be distilled to a single number that everyone can use, is it really risk?

 

[big-papa]

But do you see how those two things don't align? If the stock drops 10% and recovers, that adds to the Ulcer index, right? But how does that hurt your goals?

That problem is what I'm trying to capture - and maybe people relying on their portfolio in retirement have a different view than those accumulating.

That's why I included the other part in my post: "Anyway, it's as good as anything I've seen if you just want to roll up things into a single number."

I find just about all single number metrics to be less than adequate, which is why I prefer to see how things unfolded over time.

 

[chassis]

Use the "financial goals" and "risk parity" models on portfoliovisualizer. There are a reasonably large number of parameters that are calculated based on a list of tickers the user enters. It's an excellent tool imo.

Max drawdown is a good one. So are beta, r-squared and alpha. These are portfolio descriptors. Some address risk, some address return. You may have your eyebrows raised when you see the safe withdrawal rate and perpetual withdrawal rate metrics.

I also like total portfolio correlation, which means the aggregate correlation of all portfolio holdings with respect to each other.

 

[moontower]

Absolutely we become more risk tolerant over time because we see the opportunity that calamities like 2000 and 2008 and 2020 provide to buy low and sell high much much later.

Richard Thaler has shown most "retail" money does the opposite of what's wise: they sell when stocks are going down and buy when they're going up (usually around 52 week highs) - should do the opposite.

To me one useful non-quant-ish definition of investing is gambling with the odds in your favor over time. Research and discipline tilts the odds in your favor, esp over long periods.

 

[target2019]

I'm looking for a replacement for Sharpe ratio and standard deviation. Typically I see investment risk equals volatility.

Some people can't tolerate 30% equities, and some ride through crashes with 100% equities. Maybe investors need Sharpe ratio tolerance, to adapt it to them?

If the trove of indicators on Yahoo charts do nothing for me (https://www.yahoo.com/now/complete-guide-trend-following-indicators-100425674.html), I'd have to go to school to learn the language. For example, this book goes deep into the math of various risk measurement approaches:
https://www.value-at-risk.net/

In a recent Morningstar interview Bengen mentioned that he uses a Risk Management Service. I think advisers use something like that on a subscription basis. Perhaps there is a software or spreadsheet for poor investors such as myself?

Maybe there is a mega-thread on Risk Management at Bogleheads.

What you are looking for has probably been invented, it's just that we don't know what to call it.

 

[chassis]

Absolutely we become more risk tolerant over time because we see the opportunity that calamities like 2000 and 2008 and 2020 provide to buy low and sell high much much later.

Richard Thaler has shown most "retail" money does the opposite of what's wise: they sell when stocks are going down and buy when they're going up (usually around 52 week highs) - should do the opposite.

To me one useful non-quant-ish definition of investing is gambling with the odds in your favor over time. Research and discipline tilts the odds in your favor, esp over long periods.

Great post.

 

[OverThinkMuch]

Absolutely we become more risk tolerant over time because we see the opportunity that calamities like 2000 and 2008 and 2020 provide to buy low and sell high much much later.

Richard Thaler has shown most "retail" money does the opposite of what's wise: they sell when stocks are going down and buy when they're going up (usually around 52 week highs) - should do the opposite.

To me one useful non-quant-ish definition of investing is gambling with the odds in your favor over time. Research and discipline tilts the odds in your favor, esp over long periods.

Richard Thaler is considered a founder of behavioral finance. If he comes up with a definition of risk, that's probably exactly what I'd want. Then again, economists want to help the most people, and need to make compromises for outliers... and I'm probably an outlier, hoping the formula can include me.

Maybe people need to be categorized by how they behaved during past corrections and crashes. So a risk profile for someone who ignores big crashes and keeps holding would be different than someone who panic sells after a 15% drop. I imagine it will be hard to put that into a risk formula, but it's probably relevant.

But I also kept indexing in my first major crash. I read the years of data in "A Random Walk Down Wall Street", and it make a strong impression. So experience can also come from reading, which could make this even harder to quantify. What is the risk profile of someone with no crash experience behind them, one who reads a few books and one who doesn't?

 

[OverThinkMuch]

If the trove of indicators on Yahoo charts do nothing for me (https://www.yahoo.com/now/complete-guide-trend-following-indicators-100425674.html), I'd have to go to school to learn the language. For example, this book goes deep into the math of various risk measurement approaches:
https://www.value-at-risk.net/

In a recent Morningstar interview Bengen mentioned that he uses a Risk Management Service. I think advisers use something like that on a subscription basis. Perhaps there is a software or spreadsheet for poor investors such as myself?

Maybe there is a mega-thread on Risk Management at Bogleheads.

What you are looking for has probably been invented, it's just that we don't know what to call it.

Trend following is an investment approach, not risk assessment. Things like "trend following" and SMA are forms of technical analysis, so I don't think those are related to devising a measure of risk.

I've heard of risk parity, but I don't think it's relevant for most investors (or me). An example might be that stocks are 10x more volatile than bonds, so you invest 9% stocks and 91% bonds... the 9% stocks will be as volatile as 91% bonds. My plan to hold mostly stocks is a rejection of risk parity.

As to a better definition of risk, look at Morningstar and Yahoo Finance pages for the S&P 500 ETF "SPY". Both have a "risk" tab, and both list the same things: alpha, beta, R squared, standard deviation, Sharpe ratio. They seem to be in agreement on how risk is measured, but none of those captures the lack of risk when someone doesn't sell in a crash.
https://www.morningstar.com/etfs/arcx/spy/risk
https://finance.yahoo.com/quote/SPY/risk?p=SPY

 

[target2019]

In addition to value-at-risk link, you can pick through the spreadsheets at https://www.financialwisdomforum.org/gummy-stuff-tutorials/

Risk is mentioned in several titles and perhaps you will find something there to define exactly how this risk of yours could be modeled. Since you're not satisfied with the common measures we all know about, then you have a significant research task ahead of you. As I mentioned there are Risk Management Services who provide more complicated tools to advisors. And since your measurement is theoretical, it could be that it has been studied and analyzed at some university or think tank.