notes-investing4


some industry specific KPIs (many of these are from businessweek's annual 'facts and figures'):

notes from Businessweek 2014 Year Ahead Figures and Facts

http://www.bloomberg.com/visual-data/industries/rank/name:market-share

key industry metrics:

apparel design: Revenue by segment - retail ($M) Inventory growth to sales growth Gross margin change YOY (basis point) Marketing ratio (%)

Advertising & Marketing Sales - one-year growth Operating margin (%) Ebitda ($M) Ebitda margin (%)

Aerospace & Defense Operating margin (%) Sales - one-year growth Total backlog ($M) Orders received ($M)

Airlines Revenue passenger miles Miles (M) Sales - one-year growth Yield per passenger mile Ebitda margin

Apparel Design Revenue by segment - retail ($M) Inventory growth to sales growth Gross margin change YOY (basis point) Marketing ratio (%)

Asset Management Assets under management ($M) Net inflows/outflows ($M) Pretax margin Net revenue - one-year growth

Auto Parts Supply chain revenue (%) Trailing 12-month operating income ($M) Trailing 12-month sales growth (% ) Operating margin (%)

Automobile OEM Vehicles sold WW Operating margin (%) Number of employees SG&A rate (%)

Banking Tier 1 capital ratio Nonperforming loans to total loans Coverage ratio Return on equity (%)

Beverages Total beverage volume produced Hectoliters (M) Brewer revenue per hectoliter Free cash flow margin BEst DPS YOY growth

Biotech Price-to-sales ratio Diluted P/E from cont ops Gross margin R&D expenditure to net sales

Cable & Satellite Total cable subscriber adds Units (K) Internet/data subscriber additions Units (K) Total cable subscribers Units (M) Internet/data subscribers Units (M)

Casinos Adjusted Ebitda ($M) Casino - number of tables Casino - number of slots Gaming revenue ($M)

Chemicals Chemicals revenue growth YOY (%) Ebit margin (%) (R&D+SG&A) / sales (%) Return on invested capital (%)

Coal Mining Ebitda margin (%) Return on equity (%) Coal revenue growth YOY (%) Return on assets (%)

Construction Materials Sales - one-year growth Ebit margin (%) Return on invested capital (%) Return on equity (%)

Containers & Packaging Sales - one-year growth Ebitda margin (%) Return on assets (%) Return on equity (%)

Credit & Debit Return on equity (%) Managed card loans ($M) Pretax margin Net revenue - one-year growth

Department Stores Same-store sales Inventory growth to sales growth Basis point change in gross margin YOY Sales per retail square footage

Electrical Equipment Five-year average sales growth (%) Operating margin (%) Free cash flow ($M) Organic growth rate

Engineering & Construction Five-year average sales growth (%) Operating margin (%) Free cash flow ($M) Cash flow to net income

Express & Courier Services Average daily package volume growth (%) Operating margin (%) Return on assets (%) Return on equity (%)

Food Mfg Return on invested capital (%) Sales per employee Free cash flow margin BEst DPS YOY growth

Food Retailers Same-store sales Retail - number of locations (end) Average retail square footage sq feet (K) Operating margin (%)

Forest & Paper Products Sales - one-year growth Ebit margin (%) Return on invested capital (%) Return on equity (%)

Computer Hardware Sales - one-year growth Operating margin (%) Gross margin Free cash flow ($M)

Home & Office Products Ebitda margin (%) Return on equity (%) Sales - one-year growth Return on assets (%)

Homebuilders (Single Family) Number of homes contracted Average order price Number of selling communities Homebuilder gross margin

Household Products Mfg Organic growth rate Gross margin Operating margin (%) Free cash flow yield

Internet Media Industry revenue ($M) Free cash flow ($M) Total cost as % of revenue (ex SBC) Advertising revenue ($M)

Investment Banking Return on equity (%) Pretax margin Est Basel III Tier 1 CE ratio fully phased in Global banking & markets revenue ($M)

IT Services Total employees Sales - one-year growth Gross margin Ebitda to revenue

Life Insurance Net premiums earned - life ($M) Benefit ratio (life) Operating return on equity Investment income (losses) ($M)

Local Media Operating margin (%) Sales - one-year growth Ebitda ($M) Ebitda - one-year growth

Lodging Ebitda ($M) Sales - one-year growth Operating margin (%) Number of hotel rooms

Machinery Machinery Five-year average sales growth (%) Machinery Operating margin (%) Machinery Normalized return on equity (%) Machinery Cash flow to net income

Managed Care Commercial risk enrollment Total medicare enrollment Medicaid enrollment Managed care-medical care ratio

Mass Merchants Inventory to sales spread SG&A rate (%) Same-store sales Basis point change in gross margin YOY

Medical Equipment Gross margin Operating margin (%) Return on equity (%) Return on assets (%)

Metals & Mining Ebitda margin (%) Return on equity (%) Metals & mining revenue growth YOY (%) Return on assets (%)

Oil & Gas Services Operating margin (%) Net debt to Ebitda EV to BEst Ebitda Return on equity (%)

Pharma Cost of goods sold ($M) R&D as a % sales R&D expense ($M) Free cash flow yield

Property & Casualty Net premiums earned (non-life) ($M) Combined ratio (non-life) Operating return on equity Investment income (losses) ($M)

Rail Freight Transport Revenue per carload growth (%) Carloads growth (%) Operating ratio (%) Gross ton miles per employee (K)

Real Estate Operations Sales - one-year growth Ebitda ($M) Ebitda to revenue Net debt to Ebitda

Oil Refining & Marketing Refining capacity BBLs/day (K) Refining utilization (%) Return on equity (%) Normalized return on capital employed

REIT Occupancy rate FFO per share Debt to enterprise value Dividend indicated yield - gross

Renewable Energy Sales - one-year growth Operating margin (%) R&D expenditures ($M) Free cash flow ($M)

Restaurants Free cash flow ($M) Same-store sales Retail - number of locations (end) Operating margin (%)

Retail Discretionary Same-store sales Inventory growth to sales growth Basis point change in gross margin YOY Working capital ($M

Semiconductors Devices Gross margin Trailing 12M Cash-adjusted P/E ratio Return on invested capital (%) Sales - one-year growth

Software License revenue ($M) License revenue growth Revenue growth @ constant currency Software - deferred revenue ($M)

Steel Producers Ebitda margin (%) Return on equity (%) Sales - one-year growth Return on assets (%)

Telecom Carriers Wireless revenue ($M) Total wireless subscribers Units (K) Wireline revenue ($M) Ebitda margin (%)

Tobacco Cigarette sales volume Units (M) Actual sales per employee Free cash flow margin BEst DPS YOY growth

Utilities Electric customers - total Generating capacity (MW) Total electric sales - retail kWh (M) Return on equity (%)

---

ppl who make a list of wealth managers:

five star wealth managers

--

money and gender and myers-briggs:

http://www.selbygroup.com/docs/MoneySexPersonality.pdf

--

how much can you safely withdraw from retirement savings each year? "The recommendation is 4 percent the first year, increasing that amount for inflation each year after that" -- http://www.selbygroup.com/docs/MoneySexPersonality.pdf

--

" What You Can Do

The good news is that regardless of y our personality type or gender, you can dramatically improve your fi nancial situation fairly easily.

Attend at least one seminar on financial pl anning for retirement. A 2004 study by Annamaria Lusardi, professor of economics at Dartmouth, found that those who had attended retirement seminars had a 20 percent increase in net worth. Interestingly, those with the least amount of money and education seem to get the most from the events.

Do “a lot” of retirement planning. In a separate study, Lusardi teamed up with Olivia- Mitchell, professor of insurance and risk ma nagement at the University of Pennsylvania and executive director of the Pension Research Council, to analyze how much retirement planning people ha d done. Those who said they had done “a lot” of planning had a net worth of $200,000 co mpared to $84,000 for those who said they had done “hardly at all.”

Learn to calculate simple compound interest. Lusardi and Mitchell also found that those who grasped compound inte rest had a median net wort h of $309,000 compared to $116,000 for those who missed the question.

Make a plan. Virtually all financial gurus re commend making a plan, re viewing it at least annually, and adjusting it accordingly.

Do not count on your home to protect you. According to research by Lusardi and Mitchell, the typical older boomer (ages 51 to 56) household has nearly 50 percent of its $110,000 net worth in home equity. This is frightening because, despite reverse mortgages and home equity lines, it is difficult to live off your house. If housing prices reverted just to 2002 levels, the older boom er household would lose 13 percent-of the total value in the home.

Set a budget. John Ameriks, a senior investm ent analyst at Vanguar d, looked at the relationship between budgets and net worth. Not surprisingly, he found that budgeting is closely linked to greater wealth. It seems too obvious to mention, but how much money you make matters less than how much you spend. "

-- http://www.selbygroup.com/docs/MoneySexPersonality.pdf

---

"open" hedge fund/trader education

http://www.sanglucci.com/

http://www.businessweek.com/articles/2014-05-08/hedge-fund-sang-lucci-tries-to-boost-traders-brains-and-returns

--- may 24 2014

intersection of

market cap by avail supply, market cap by total supply, volume

1 Bitcoin $ 6,628,556,816 $ 517.52 12,808,225 BTC $ 30,880,677 -1.29 % 2 Litecoin $ 316,233,117 $ 11.05 28,623,804 LTC $ 6,960,072 -3.76 % 3 Darkcoin $ 55,315,680 $ 12.80 4,320,135 DRK $ 6,069,032 +2.76 % 4 Peercoin $ 48,249,347 $ 2.25 21,421,953 PPC $ 513,949 -0.80 % 5 Nxt $ 36,144,942 $ 0.036145 999,997,096 NXT* $ 125,019 -3.01 % 6 Dogecoin $ 33,121,249 $ 0.000419 79,104,946,278 DOGE $ 738,608 -1.48 % 7 Ripple $ 25,270,967 $ 0.003232 7,817,888,647 XRP* $ 1,392,174 -11.33 % 8 Namecoin $ 21,414,859 $ 2.42 8,837,732 NMC $ 466,403 +3.82 % 9 Mastercoin $ 13,209,076 $ 23.46 563,162 MSC* $ 27,396 +24.71 % 10 BlackCoin? $ 9,792,863 $ 0.13 74,541,407 BC* $ 1,586,396 +43.21 % 11 Counterparty $ 7,711,790 $ 2.91 2,649,268 XCP* $ 79,579 +7.19 % 12 BitShares?-PTS $ 7,151,230 $ 4.37 1,637,991 PTS $ 57,429 +1.46 % 13 MaidSafeCoin? $ 6,991,998 $ 0.01545 452,552,412 MAID* $ 15,182 +32.79 %

1 Bitcoin $ 6,628,556,816 $ 517.52 12,808,225 BTC $ 30,880,677 -1.29 % 2 Ripple $ 323,245,384 $ 0.003232 99,999,990,162 XRP* $ 1,392,174 -11.33 % 3 Litecoin $ 316,233,117 $ 11.05 28,623,804 LTC $ 6,960,072 -3.76 % 4 Darkcoin $ 55,315,680 $ 12.80 4,320,135 DRK $ 6,069,032 +2.76 % 5 Peercoin $ 48,249,347 $ 2.25 21,421,953 PPC $ 513,949 -0.80 % 6 Nxt $ 36,144,942 $ 0.036145 999,997,096 NXT* $ 125,019 -3.01 % 7 Dogecoin $ 33,121,249 $ 0.000419 79,104,946,278 DOGE $ 738,608 -1.48 % 8 Namecoin $ 21,414,859 $ 2.42 8,837,732 NMC $ 466,403 +3.82 % 9 CRTCoin $ 18,597,709 $ 1.85 10,039,066 CRT $ 8,495 +0.93 % 10 Mastercoin $ 14,529,979 $ 23.46 619,478 MSC* $ 27,396 +24.71 % 11 BlackCoin? $ 9,792,863 $ 0.13 74,541,407 BC* $ 1,586,396 +43.21 % 12 Counterparty $ 7,711,790 $ 2.91 2,649,268 XCP* $ 79,579 +7.19 % 13 Isracoin

1 Bitcoin $ 6,628,556,816 $ 517.52 12,808,225 BTC $ 30,880,677 -1.29 % 2 Litecoin $ 316,233,117 $ 11.05 28,623,804 LTC $ 6,960,072 -3.76 % 3 Darkcoin $ 55,315,680 $ 12.80 4,320,135 DRK $ 6,069,032 +2.76 % 10 BlackCoin? $ 9,792,863 $ 0.13 74,541,407 BC* $ 1,586,396 +43.21 % 7 Ripple $ 25,270,967 $ 0.003232 7,817,888,647 XRP* $ 1,392,174 -11.33 % 29 Libertycoin $ 1,053,328 $ 0.06381 16,507,234 XLB* $ 1,326,619 +118.73 % 6 Dogecoin $ 33,121,249 $ 0.000419 79,104,946,278 DOGE $ 738,608 -1.48 % 25 Billioncoin $ 1,420,078 $ 0.000182 7,820,646,324 BIL $ 539,757 +0.86 % 4 Peercoin $ 48,249,347 $ 2.25 21,421,953 PPC $ 513,949 -0.80 % 8 Namecoin $ 21,414,859 $ 2.42 8,837,732 NMC $ 466,403 +3.82 % 19 Monero $ 2,600,184 $ 2.85 911,701 MRO $ 255,300 +11.14 % 42 X11Coin $ 433,663 $ 0.094274 4,600,040 XC $ 164,332 ? 33 Cinni $ 605,514 $ 0.040335 15,011,964 CINNI* $ 150,001 -2.53 %

isection =

btc ltc darkcoin peercoin dogecoin ripple namecoin blackcoin

note: if not volume, then we also get the interesting (to me) nxt, mastercoin, counterparty. nxt and counterparty have significant volume, mastercoin not so much if not market cap, , then we also get the interesting (to me) monero, which has almost insigificant market cap but not quite

http://www.reddit.com/r/CryptoCurrency/comments/2643st/wired_darkcoin_the_shadowy_cousin_of_bitcoin_is/

   darkcoin open source, 9% premine

out of the biggies, the most interesting ones to me are:

btc darkcoin namecoin nxt

but i also like: mastercoin monero bytecoin (why is BCN not on this list? https://news.ycombinator.com/item?id=7765455 claims its market cap should be around $10mil, which would make it a contender)

now what will drive adoption? i'm thinking awesome tech and dev community. so todo: see which coins have an active dev community

btw etherium and zerocash sound awesome

---

http://managed-futures-blog.attaincapital.com/2014/07/17/alternative-investments-why-do-we-care/?utm_source=hs_email&utm_medium=email&utm_content=13513588&_hsenc=p2ANqtz-8pviPWK6L3qiXe9yBmmGoWSVhuNnOmTygIoYmGZqJlF_cJ0vkk82u0wNcegEccsG9IPbozM8wQPbKocpGSx15gIoGeNA&_hsmi=13513588

---

http://vixandmore.blogspot.com/2013/02/beyond-splv-expanding-universe-of-low.html

---

http://vixandmore.blogspot.com/2012/05/best-post-of-year-on-exchange-traded.html

---

etfdb.com/etf-education/101-etf-tips-tricks-every-financial-advisor-should-know/

---

http://vixandmore.blogspot.com/2012/05/cheating-with-partial-hedges.html

---

http://vixandmore.blogspot.com/2012/12/volatility-during-crises.html

---

http://vixandmore.blogspot.com/2012_05_01_archive.html

---

http://vixandmore.blogspot.com/2012/06/performance-of-volatility-hedged-etps.html

---

http://vixandmore.blogspot.com/2012/05/cheating-with-partial-hedges.html

i've heard that historically, in fact it makes sense to hedge. i wonder if it would make sense to be more hedged for small falls or for big ones? To have a linearly increasing or decreasing number of put options at various falls? To finance distant puts with distant calls?

---

http://seekingalpha.com/article/300307-book-review-buy-and-hedge-the-5-iron-rules-for-investing-over-the-long-term

" And so, the authors argue, the solution is to define risk. Ideally, every investment should be hedged. Alternatively, the portfolio as a whole can be hedged. The authors outline a range of strategies, from married puts, collars, and ITM options to vertical and diagonal spreads, to accomplish this goal.

Since for the most part the authors view options as hedging vehicles, not speculative instruments, the focus is on risk management. They boil risk management down to four metrics, all of which should be used in analyzing a portfolio: capital at risk, volatility, implied leverage, and correlation. "

---

i think what i described above is a partial collar: http://www.theoptionsguide.com/the-collar-strategy.aspx

one proposal: buy 1000 units of an underlying stock that you think will rise (recall that 1 option counts against 100 units of stock). Buy a put with a strike $10 below. Buy another put with a strike $20 below. Buy another put with a strike $30 below. Etc, until you get a total of 10 puts. So, at $100 below, you are fully insured, above that, you are only partially insured.

One annoyance with this strategy is that you are now a net insurance purchaser. Someone else is profiting off of selling you insurance. To make some of this back, you can sell calls at strike prices above the current price, with a similar graded pattern. If you want to capture more of the upside, sell calls at the exp(-log(return)) of the prices that you bought the puts, e.g. if you buy a put at 50% of the current price, then sell a call at 200% of the current price (rather than 150%).

The reason to do all of this is only if a high proportion of your total savings are tied up in things correlated to one stock (e.g. a market index). In this case, the reasoning behind the Kelly criterion shows that a 50% downturn is worse for you than a 50% upturn (if everything were invested in one stock, then you care about the log of the expected return, rather than the expected return).

The reason i have a graded pattern of partial insurance is the same logic; you care more about large falls than small falls.

In many cases, it would be better to invest in a diversified portfolio of uncorrelated things (and rebalance between them frequently). But this has disadvantages, because often one asset class and its correlates (such as stock) is expected to outperform, and so you'd rather invest a lot in that.

---

counterintuitive observations about investing:

---

http://finance.yahoo.com/news/7-truths-investors-simply-cannot-accept-195752739.html

---

http://strubelim.com/dumb-investment-of-the-week-junk-bond-funds/ looks at market returns from 1983 to 2012.

in the image http://strubelim.com/wp-content/uploads/2014/03/03-17-14-return-table-300x64.jpg they show four portfolios, and for each one, the average annual return, the standard deviations, and the max yearly drawdown:

Junk bonds (Barclays High Yield Bond Index), S&P 500, T-bonds, 50/50 S&P 500 and T-bonds, 75/25 S&P 500 and T-bonds: average annual returns: 9.27%, 10.06%, 8.13%, 9.61%, 9.98% standard deviations: 15.74%, 17.24%, 9.66%, 9.73%, 13.07% max yearly drawdowns: -26.16%, -37%, -11.12%, -8.45%, -22.75%

from this we can calculate returns/standard_deviation (related to the Sharpe ratio, but here's i'll assume that the risk-free return is zero) and returns/variance (a measure related to the Kelly criterion if Gaussian probability distributions are assumed, see On Automated Trading by Max Dama ( http://www.decal.org/file/2945 ), section 6.1, Optimizing Kelly, or see http://epchan.blogspot.com/2006/10/how-much-leverage-should-you-use.html or see The Kelly Criterion in Blackjack Sports Betting, and the Stock Market, by Edward O. Thorp ( http://www.edwardothorp.com/sitebuildercontent/sitebuilderfiles/KellyCriterion2007.pdf ), section 7.1 Wall street: the biggest game: Continuous approximation, page 23):

in Python: returns = array([ 0.0927, 0.1006, 0.0813, 0.0961, 0.0998]); standard_deviations = array([ 0.1574, 0.1724, 0.0966, 0.0973, 0.1307]); print returns/standard_deviations; print returns/(standard_deviations2)

Junk bonds (Barclays High Yield Bond Index), S&P 500, T-bonds, 50/50 S&P 500 and T-bonds, 75/25 S&P 500 and T-bonds: returns/standard_deviations: 0.58894536 0.58352668 0.84161491 0.98766701 0.76358072 returns/standard_deviations^2: 3.74171132 3.38472553 8.71236964 10.15074007 5.84223963

for both of these measures, the rank ordering from best to worse is: 50/50, all T-bonds, 75/25, all junk bonds, all equity

So the standard advice to go 60/40 equity/bonds is on the mark. This isn't to say that a portfolio with a few bonds other than T-bonds in there might not do even better.

note that various sources (eg http://www.edwardothorp.com/sitebuildercontent/sitebuilderfiles/KellyCriterion2007.pdf , http://content.time.com/time/business/article/0,8599,1982327-2,00.html (Ian Ayres and Barry Nalebuff), http://ddnum.com/articles/leveragedETFs.php) claim that the ideal is about 2x leverage (NOTE: leveraged ETFs probably aren't equivalent to this because they have daily rebalancing costs eg http://seekingalpha.com/article/1677722-drilling-down-on-volatility-decay )

what are the comparable numbers for angel investing?

http://www.keiretsuforum-midatlantic.com/sites/default/files/docs/Angel-Investing-Worth-the-Effort-by-Geoff-Roach.pdf claims that from 2000-2006, angel investing returns in its dataset of 120 investments of 100 companies were:

returns = array([.2038, .2132, .2824, .2620, .3246, .1455 ,.2013])

calculating the geometric mean:

exp(mean(log(1 + array(returns)))) - 1

we get 0.23201052508116793 (note: in this case this is very close to the arithmetic mean, 0.23325714285714286).

They note that the Sharpe ratio was 3.39. They don't report the standard deviation, but if they used a mean of 23.20% and a risk-free rate of zero, the standard deviation would have been .2320/3.39 = 0.06843657817109144, that is, a standard deviation of 6.84%.

(this is way too low; that's lower than the S&P 500's standard deviation)

Then they recompute with real estate deals removed:

returns = array([-.0201, .1523, .0680, .0482, .1572, .0852, .3238]) exp(mean(log(1 + array(returns)))) - 1

Now the (geometric) mean is 11.19% ( 0.11191580150697611 )

They say the Sharpe is .78. So, if they used a mean of 11.19% and a risk-free rate of zero, the standard deviation would have been .1119/.78 = 0.14346153846153845 or about 14.35%.

Imo this is still too low, as this is about the same as the S&P 500's standard deviation (however, an optimist might say that a portfolio of many angel investments is less correlated than a portfolio of many S&P 500 companies, and so might actually experience a lower standard deviation). I conclude that the study's size (and probably, timescale) was probably too short to capture enough large losses.

But if it were correct, then mean/std^2 would be 5.4369973190348535

Some tentative conclusions, though:

other data sets:

guesses: http://talkfast.org/2010/04/28/angel-investing-simulation-part-2/ http://possibleinsight.com/2010/05/11/simulating-angel-investment-kevins-remix/

AIPP data assembled by Wiltbank http://sites.kauffman.org/aipp/ http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1027911 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1028592 http://possibleinsight.com/2013/01/14/how-many-angel-investments/

i downloaded the AIPP data and used libreoffice's 'Descriptive statistics' (Statistics->Descriptive statistics->Descriptive statistics) on the 'basemultiple' column, which (i think) for each investment tells you if the total cash out was e.g 10x of the total cash invested, although for many rows this column was empty (the descriptive stats seemed to ignore the empty ones).

the stats are:

Mean 11.8858071278826 Standard Error 3.95114329178981 Median 0.83 Mode 0 Standard Deviation 86.2942720575701 Sample Variance Kurtosis 150.622194252997 Skewness 11.6669023488701 Range 1332.8 Minimum 0 Maximum 1332.8 Sum 5669.53 Count 477

however, these investments were held for various numbers of years.

so here's what i did. In the spreadsheet, column AY was basemultiple, and column AZ was 'yearsheld'. So i inserted two new columns, BA and BB. In BA2 i put the formula

AY2^(1/AZ2)

and i copied that to the rest of the BA column. This gave the answer i wanted (annualized multiple) in BA except that it put "#VALUE!" in cells in which AY and AZ were blank. So in BB2 i put

iferror(BA2,)

and copied that to the rest of the BB column. Then I did descriptive stats on the BB column as well.

The results were:

Mean 1.28209885463984 Standard Error 0.11057155593939 Median 1 Mode 0 Standard Deviation 2.33512885433649 Sample Variance 5.45282676635483 Kurtosis 107.951601307366 Skewness 8.57879250943967 Range 35 Minimum 0 Maximum 35 Sum 571.81608916937 Count 446

Instead of multiples, you might want stats on the returns. So i changed BA2 to =(AY2^(1/AZ2)) - 1 and copied to the rest of BA, then did descriptive stats on BB again. The results:

Mean 0.28209885463984 Standard Error 0.11057155593939 Median 0 Mode -1 Standard Deviation 2.33512885433649 Sample Variance 5.45282676635483 Kurtosis 107.951601307366 Skewness 8.57879250943967 Range 35 Minimum -1 Maximum 34 Sum 125.81608916937 Count 446

Note that the standard deviation appears to be the unbiased estimator, not the sample std, that is, like std(data, ddof=1) in Python (not that it'll make much difference since we have 446 data points, i think).

So from this we can calculate a Sharpe-like measure, mean/std, and a gaussian-Kelly-like measure, mean/(std^2):

mean/std = 0.28209885463984/2.33512885433649 = 0.12080654740570895 mean/(std^2) = 0.28209885463984/(2.33512885433649^2) = 0.05173442449711635

comparing this to other portfolios above, where the mean/std ranged from .58 to .98 and the mean/std^2 ranged from 3.3 to 10.2, we see that angel investing is, by itself, probably not a very good strategy. Perhaps it would be good as a small part of a large portfolio, however, if it turns out to be uncorrelated to the rest of the portfolio. Note that one interpretation of the Kelly computation is the recommended leverage to apply if your entire portfolio were to be composed of this; so if you have access to only a single angel investment, your leverage should be 0.05, that is, you should put only 5% of your portfolio into it and hold the other 95% as cash. Interestingly, this coheres with http://talkfast.org/2010/04/28/angel-investing-simulation-part-2/ 's suggestion that you need at least about 20 companies in an angel portfolio (although note that http://possibleinsight.com/2013/01/14/how-many-angel-investments/ , who also worked off of the AIPP dataset, suggests many more, at least about 70) (note also that people generally think that investing using Kelly is overly optimistic because you tend to underestimate the standard deviation due to the heavy-tailedness of the actual distribution; also the mean/var^2 approximation is wrong and, i'm guessing, overly optimistic; people tend to recommend the heuristic 'half-Kelly', that is, divide Kelly by half, giving about 2.5% in this case). Of course, this isn't exactly what the Kelly criterion is saying; if you had access to multiple angel investments, then to the extent that they are uncorrelated, they would decrease the standard deviation of the portfolio, after which the half-Kelly criterion would recommend allocating more than 2.5%

The keiretsu forum study is not directly comparable because the numbers reported there were the rate of return and Sharpe ratio for an imaginary aggregate portfolio of all investments made by all investors. So the standard deviation for that is comparable to that of an index like the S&P 500, where the standard deviation of the AIPP data is comparable to that of an individual stock. So although the Kelly leverage for the AIPP data is 0.05 and for the keiretsu forum data is 5.44, this is not as different as it sounds because that is like saying if you can only invest in one stock, put 5% of your capital into it, but if you can invest in an index fund, borrow 5x your capital for it (or, if you use half Kelly, 2.5% and 2.5x; note that some people recommend leveraging equity index funds to 2x and some even say leverage to 3x, so this isn't crazy).

Note if you are being pessimistic, you may want to take the keiretsu forum study's mean return and the AIPP standard deviation when computing mean/std and mean/std^2 for individual angel investments, yielding:

mean/std = 0.11191580150697611/2.33512885433649 = 0.047927034646992175 mean/(std^2) = 0.11191580150697611/(2.33512885433649^2) = 0.020524364022991053

So in this case, the half-Kelly is about 1%.

So the annual returns for angel investments are on the order of somewhere between 11% to 28%, the mean/std is on the order of somewhere between .05 to .78, and the mean/std^2 between 0.02 to 5.44.

What size makes a "portfolio"? The sample portfolio had 100 investments; other people recommend a minimum of 12-70; the reciprocal of 2.5%, 40, seems reasonable to me.

So, if i had access to a single angel investment, i should invest about 1%-2.5% of my capital. If i had access to a portfolio of 40 or so such investments, i should invest about as much in that portfolio as i'd invest in public company equities (so, a 60/40 equity/bond portfolio turns into a 30/30/40 angel/public equity/bond portfolio; and note that ideally this whole thing would be 2x leveraged), possibly a bit more or less depending on if the true numbers are better or worse than these.

this is in the range of https://www.equitynet.com/blog/wp-content/uploads/2011/09/Survey-Angel-Investing-in-the-US.pdf , which says that angels invest about 20% of their personal wealth into a portfolio of about 6-7 ventures (about 3% of their personal wealth per venture)

The bottom line: each angel investment should get about 1.5% of my portfolio, up to about 30% of the total

(so, my entire portfolio could be approximately:

35% bonds (mostly treasuries and a few investment grade bonds) up to 25% angel investments (each one 1.5%; if i can't find enough, add the rest to the public equites slice) 25% public equities and a few junk bonds (probably just a few index fund ETFs) 10% other alternative investments 5% cash

)

https://www.equitynet.com/blog/wp-content/uploads/2011/09/Survey-Angel-Investing-in-the-US.pdf

http://gadoci.me/is-angel-investing-worth-it

http://www.frbatlanta.org/documents/pubs/wp/wp1014.pdf

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1027911


random QA on the Kelly criterion

http://quant.stackexchange.com/questions/1374/optimality-of-kelly-criterion-in-non-normal-environment


Good and bad properties of the Kelly criterion

http://www.edwardothorp.com/sitebuildercontent/sitebuilderfiles/Good_Bad_Paper.pdf

---

" Michael James said...

    The Kelly criterion is based on the assumption that you can borrow at the risk-free rate. In your example, if you borrow at 4% over the risk-free rate, then the optimal leverage works out to be 6.25 (half-Kelly in this case).
    See the following post where I discuss the effect of higher borrowing rates: link
    However, I don't know where you can find an investment that has a 12% return with a standard deviation of only 8%. If the standard deviation is 20% (roughly matching long-term volatility of the S&P 500), then the Sharpe ratio drops to 0.4, and the Kelly criterion drops to f=2. Toss in the assumption of borrowing at 4% above the risk-free rate, and the optimal leverage point drops to f=1 (i.e., no borrowing at all).
    It's amazing how the case for leverage evaporates as you repair unrealistic assumptions.
    Tuesday, February 19, 2008 at 9:27:00 AM EST 

Ernie Chan said...

    Dear Michael,
    You are right that if your funding cost increases, the Kelly leverage will be reduced. However, your assumption of 4% over risk-free rate is far too onerous. For e.g., if you trade through Interactive Brokers (popular choice for many independent traders and small hedge funds), the funding rate is only 1% or less above risk-free rate.
    You are also correct in pointing out that if we own S&P500 index, we will not get returns of 12% vs. standard deviation of 8% (a Sharpe ratio of 1.5). However, many market-neutral strategies that quantitative traders employ have even better returns vs. risk ratio than what I used as an example here. In fact, Sharpe ratios (after cost) of over 2 are quite common. So I don't believe these are unrealistic assumptions -- and certainly high leverage is very appropriate for such strategies.
    Ernie
    Tuesday, February 19, 2008 at 10:20:00 AM EST Michael James said...
    Dear Ernie,
    I can't say that I know much about brokers other than the one I use. I can borrow only a modest amount from my broker in the form of margin. After that, I woud need to get a loan elsewhere at higher rates. It is unlikely that I could borrow even 3 times the size of my portfolio at any interest rate.
    Of course, hedge funds are very likely to have better access to borrowed funds than I have. However, given the number of hedge funds that go under, I'm guessing that many of them assess the volatility of their strategies incorrectly.
    Suppose that a hedge fund has a strategy with expected excess return 20% and standard deviation of 10% (Sharpe Ratio 2). Borrowing at 1% excess leads to f=19. If such a fund had $10 million in assets and sought to borrow $180 million at 1% excess, the potential lender would be insane (in my opinion) to agree. The strategy seems too good to be true. The expected leveraged geometric return is 181.5%. But this drops to -44% if the standard deviation is actually 15% instead of 10%.
    If crazy lenders exist out there then I can't blame hedge funds for trying their strategies. It amounts to a huge free roll. If things work out, then everybody gets rich. If it all goes south, then the lender loses a pile of money.
    Michael
    Tuesday, February 19, 2008 at 11:07:00 AM EST 

" -- http://epchan.blogspot.com/2006/10/how-much-leverage-should-you-use.html?showComment=1203431220000#c3163263245499769623

" Anonymous said...

    Another interesting observation: If you believe that you don't have any special portfolio management skills and for the invested asset mu = rf + rp*sigma then the optimal leverage under GBM etc. works out to f=rp/sigma, i.e. the market price of risk divided by your investment's volatility.
    Tuesday, July 3, 2007 at 3:50:00 PM EDT " -- http://epchan.blogspot.com/2006/10/how-much-leverage-should-you-use.html?showComment=1183492200000#c1815882118160577040 (he later explains that GBM is short for geometric brownian motion

 Anonymous said...
    Dear Ernie,
    In example 6.2 in your book, you calculate that the compounded return on SPY using the recommended 2.528 Kelly leverage is 13.14%. Do you mean that the return to the investor on the full leveraged amount of $252,800 (using $100,00 in capital), would be 0.1314*252,800 = $33,218 (33% of capital)? Or, do you mean you would get 0.1314*100,000 = $13,140 back using the full $252,800 (using $100,000 in capital)? If you would only get $13,140, this seems very low considering how much money has been borrowed and the 9.8% return without borrowing.
    Using the formula in the appendix for example 6.2, I got g = 0.04 + 2.528*(0.1123-0.04) + (.1691*2.528)^2/2 = 31.4%. But what initial investment does gain value this apply to ($100,000 or $252,80?)? Does this mean using $100,000 in capital, and 2.52 leverage I would make $31,400? Essentially, how much money would I have at the end of the year if I had $100,000 in capital, and used the Kelly leverage to borrow up to $252,800 to buy SPY in your example?
    Thanks very much for your great book and for your help.
    -E
    Sunday, April 3, 2011 at 7:28:00 PM EDT Ernie Chan said...
    E,
    In example 6.2, if you have a $100K capital, and buy SPY with a leverage of 2.528, you should expect to earn $13,140 at the end of the first year.
    It is not much higher than the $9,800 that you would earn with no leverage because of the low Sharpe ratio of SPY.
    Ernie
    Wednesday, April 6, 2011 at 9:08:00 AM EDT 

 Anonymous said...
    Hello Ernie,
    What do you recommend in the case of $100k capital trading futures contracts across 3 models (to figure out the # of contracts / model)? One could calculate f for each using e.g. half Kelly and then round down to the get the # contracts / model, however, this does not take into consideration the portfolio risk as a whole...
    Grateful for your thoughts and / or links on where to research this.
    best regards
    Ed
    PS presumably the formula values of m & r could be either annualized or summed (if longer or shorter period than 1 year) so long as they are both treated equally?
    Tuesday, February 28, 2012 at 10:33:00 AM EST Ernie Chan said...
    Hi Ed,
    I am not sure why you think that portfolio risk is not taken into account using Kelly formula. As you can see from example 6.3 of my book, covariance between each futures are used to generate allocation across the 3 contracts.
    Means and variances can be of any time frame: Kelly formula is time-scale invariant, unlike the Sharpe ratio.
    Ernie
    Tuesday, February 28, 2012 at 11:04:00 AM EST 

-- http://epchan.blogspot.com/2006/10/how-much-leverage-should-you-use.html


toread:

https://en.wikipedia.org/wiki/Constant_proportion_portfolio_insurance


http://managed-futures-blog.attaincapital.com/2014/10/28/pursuing-portfolio-perfection/

http://mebfaber.com/2013/07/31/asset-allocation-strategies-2/

---

ok i'm not sure if this is right, but i'm trying to figure out when to fix a car vs. buy a new one.

Let's assume you pay up front for every car (no interest charges), and that the only costs of a car that vary with age is fixing stuff, and that you will spend your whole life buying one (new) car after another, always the same type of car and always driven the same amount and way, so that on average the expected cost of fixing stuff is the same for each car depending only on age, and that there is no opportunity cost to spending money buying a new car.

So, over the course of your life, the average annual cost of owning a working car is:

(cost of buying a new car - money you get from selling the old car) / (years in between buying new cars) + (average cost of fixing stuff per year)

If you have already owner your car for x years, then let's assume you are only comparing owning each car for x years and then getting a new one, vs. owning each car for x+1 years and then getting a new one. So this is:

(cost of buying a new car - money you get from selling the old car) / (x) + (average cost of fixing stuff per year if you keep it x years) vs (cost of buying a new car - money you get from selling the old car) / (x+1) + (average cost of fixing stuff per year if you keep it x+1 years)

first let's look at the part: (cost of buying a new car - money you get from selling the old car) / (x) - (cost of buying a new car) / (x+1)

the derivative of k/x is -k/x^2, so we can use that ((cost of buying a new car)/x^2) as an approximation of this magnitude.

In the scenario where you just took your old car to the shop and was told it would cost $Y to fix it, then you might assume that at a minimum the average fixit cost per year would increase Y/x should you choose to keep this car another year (in fact, it will probably go up more, because as cars get older they need more fixing; but there's a small chance that this fixit cost was a fluke accident and that the car won't need this much fixing again for longer than the amount of time you've owned the car so far, in which case this would be an overestimate; but, usually you can assume that if you double the holding period of the car, the fixing cost will at least double).

So, under the above assumptions, you should buy a new car whenever

(cost of a new car - money you get from selling the old car) / x^2 < (cost of this repair)/x

simplifying:

(cost of a new car - money you get from selling the old car) / x < (cost of this repair)

or:

(cost of a new car - money you get from selling the old car) / (how long you've owned this car) < (cost of this repair)

and, most likely, you should buy a new one sooner, as the amount of repairs per year in the future will most likely be more than the amount of repairs per year in the past, and also there are non-monetary costs of the inconvenience of future repairs, the uncertainty of future breakage, and the safety hazard of driving an old car that is more likely to break. To include these, you might want to multiply "cost of this repair" by a constant fudge factor, such as 1.5. (on the other hand, there is also the inconvenience of shopping for a new car to consider).

(cost of a new car - money you get from selling the old car) / (how long you've owned this car) < 1.5*(cost of this repair)

or, in other words, at any given time you have a threshold. This threshold is the most you're willing to pay for a repair before you decide to sell the car instead of fixing it. This threshold is:

(cost of a new car - money you get from selling the old car) / (1.5 * how long you've owned this car)

the results of this formula are not terribly different from just saying "sell the car when maintanence outstripts depreciation", but are easier to calculate.

note that many car rental companies only keep their cars for a short time:

" The average age of a Hertz rental vehicle nearly doubled from 10 months in 2006 to 18 months in 2012, The Wall Street Journal reported last year. "

presumably they are doing a calculation similar to this one, although theirs will be somewhat different since we can assume their cars will get a lot more wear.

---

currently, Jan 2015, CapitalOne?'s 360 Saving Account rate is better than many 1-year CDs: https://home.capitalone360.com/online-savings-account

currently, Ally is offering 1%

---

http://managed-futures-blog.attaincapital.com/2015/02/13/what-you-dont-know-cant-hurt-you/

---

risk management: different kinds. Some of them:

---

more on drawdown recovery time: http://managed-futures-blog.attaincapital.com/2013/10/08/the-2-important-drawdown-measurements-how-deep-how-long/

---

http://managed-futures-blog.attaincapital.com/2015/04/01/sharpe-ratio-the-black-sheep-of-risk-analysis/

notes on one that they mention; the 'ulcer index'; it's the RMS of drawdown

---

ok i have a suggestion for another metric which may be 'better' than the 'ulcer index'; instead of doing the RMS of drawdown, do the exp(-(RMS of (log return-from-beginning-ratio, but with positive returns set to zero))).

Eg if you had an investment which you held for 4 periods, and at the end of the first, second, and forth period it was worth 75% of what you bought it for, and at the end of the second period, 25%, then:

a = array([.25, .25, .75, .25])

exp(-sqrt(sum(log(array([min(x,1) for x in a]))2)/len(a)))

0.29845016832915766

(higher is better, eg higher is less drawdown; range is from 0 (an investment that has, at some point in time, gone to zero) to 1 (an investment that has never gone down, only up))

using the log accentuates large drawdowns more than the linear RMS (the ulcer index), which would give you only sqrt(sum(array([min(x,1) for x in a])2)/len(a)) == 0.4330127018922193

another example:

a = array([.25, .25, .75, .25, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 ,1 ,1])

sqrt(sum(a2)/len(a))

0.91515026088615636

exp(-sqrt(sum(log(a)2)/len(a)))

0.58231173485472842

compare also to standard deviation and variance and their length-normalized variants:

a = array([.25, .25, .75, .25]) std(a) == 0.21650635094610965 var(a) == 0.046875 std(a)/sqrt(len(a)) == 0.10825317547305482 var(a)/len(a) == 0.01171875

a = array([.25, .25, .75, .25, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 ,1 ,1]) std(a) == 0.26809513236909022 var(a) == 0.071874999999999994 std(a)/sqrt(len(a)) == 0.059947894041408996 var(a)/len(a) == 0.0035937499999999997

and the same for log returns:

a = array([.25, .25, .75, .25]) std(log(a))

0.47571307544817298

exp(std(log(a)))

1.6091612418463055

var(log(a))

0.22630293015235911

exp(var(log(a)))

1.2539554686695085

std(log(a))/sqrt(len(log(a)))

0.23785653772408649

exp(std(log(a))/sqrt(len(log(a))))

1.2685271939719327

var(log(a))/len(log(a))

0.056575732538089778

exp(var(log(a))/len(log(a)))

1.0582067524488608

a = array([.25, .25, .75, .25, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 ,1 ,1]) std(log(a))

0.4929300182371098

exp(std(log(a)))

1.6371059498198504

var(log(a))== 0.24298000287923741 exp(var(log(a)))

1.2750431266721411

std(log(a))/sqrt(len(log(a)))

0.11022250289283884

exp(std(log(a))/sqrt(len(log(a))))

1.1165264731928997

var(log(a))/len(log(a))

0.01214900014396187

exp(var(log(a))/len(log(a)))

1.0122230990179069

a = array([.25, .25, .75, .25]) b = array([.25, .25, .75, .25, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 ,1 ,1])

std(log(a)) / std(log(b)) exp(std(log(a))) / exp(std(log(b))) var(log(a)) / var(log(b)) exp(var(log(a))) / exp(var(log(b))) std(log(a))/sqrt(len(log(a))) / std(log(b))/sqrt(len(log(b))) exp(std(log(a))/sqrt(len(log(a)))) / exp(std(log(b))/sqrt(len(log(b)))) var(log(a))/len(log(a)) / var(log(b))/len(log(b)) exp(var(log(a))/len(log(a))) / exp(var(log(b))/len(log(b)))

In [85]: std(log(a)) / std(log(b)) Out[85]: 0.96507223712909462

In [86]: exp(std(log(a))) / exp(std(log(b))) Out[86]: 0.98293042183578894

In [87]: var(log(a)) / var(log(b)) Out[87]: 0.93136442287735544

In [88]: exp(var(log(a))) / exp(var(log(b))) Out[88]: 0.98346121981170043

In [89]: std(log(a))/sqrt(len(log(a))) / std(log(b))/sqrt(len(log(b))) Out[89]: 0.1078983562709226

In [90]: exp(std(log(a))/sqrt(len(log(a)))) / exp(std(log(b))/sqrt(len(log(b)))) Out[90]: 1.1361371399859073

In [91]: var(log(a))/len(log(a)) / var(log(b))/len(log(b)) Out[91]: 0.011642055285966943

In [92]: exp(var(log(a))/len(log(a))) / exp(var(log(b))/len(log(b))) Out[92]: 1.0454283778700257

if you are thinking of making a 'performance ratio' out of this by putting returns on top, and then comparing that performance ratio to the Sharpe using the above examples, these might be of interest to you:

0.29845016832915766/0.58231173485472842

0.5125264535560377

0.58231173485472842/0.29845016832915766

1.9511188018915866

0.21650635094610965/0.26809513236909022

0.8075728530872481

0.046875/0.071874999999999994

0.6521739130434783

0.10825317547305482/0.059947894041408996

1.8057877962865378

0.01171875/0.0035937499999999997

3.2608695652173916

you could make it symmetric, instead of only counting downside (downdowns) (as Sharpe is symmetric while Sortino is not) by not setting the upside returns to zero at the beginning (the array([min(x,1) for x in a]) part); but this isnt a good idea, because investments that appreciate over time would be penalized.

in order to create a performance ratio, you'd might just multiply this by the mean of the log of the return-from-beginning-ratios:

mean(log(returnRatios))*exp(-sqrt(sum(log(array([min(x,1) for x in returnFromBeginningRatios]))2)/len(returnFromBeginningRatios)))

(recall that the second factor here, the risk metric, ranges from 0 (bad) to 1 (good))

or, you might leave the risk metric in (the negation of) log space and divide by it:

mean(log(returnRatios))/sqrt(sum(log(array([min(x,1) for x in returnFromBeginningRatios]))2)/len(returnFromBeginningRatios))

note: to calculate returnFromBeginningRatios as used above, you'd really have to calculate them beginning with each starting point in the dataset, and include all of those points in the vector returnFromBeginningRatios

(which suggests that all time points are weighted equally; but really, you might want to weight by marketcap at that time, right?)

---

so, to summarize the above, we now have another risk metric to add to our pile. My favorites performance metrics so far:

in what situations would someone care about drawdowns, as opposed to std? If std is symmetric and if returns are i.i.d. then std should suffice. You care about drawdowns when there are temporal dependencies between returns in subsequent periods. Two examples might be "momentum", when something keeps going down over multiple periods, or "anti-momentum", which something that went down in one period tends to bounce back in the next period. Note however that if you have the same set of returns, if you scramble their locations in time, you get the same overall return, even though grouping them together produces a greater drawdown. Note also that markets tend to go down sharply and then rise slowly.

---

the principal that income is good, maintenance expenses (eg owning a home) is bad

increasing wealth via earning more vs via spending less

---

http://www.quora.com/What-should-everyone-know-about-economics

---

http://www.businessinsider.com/how-to-invest-100-1000-or-10000-2014-1

---

http://lesswrong.com/lw/yv/markets_are_antiinductive/

---

https://www.cxoadvisory.com/what-investing-approaches-work-best/

https://www.cxoadvisory.com/

i was impressed by their article on https://www.cxoadvisory.com/3265/sentiment-indicators/mark-hulbert/ (not recommending that, it's on a specialized topic, it just makes me think cxoadvisory might be good)

---

QMN

" The big daddy of market neutral ETPs (and the only one not presently foundering) is, oddly enough, an atypical product. The IQ Hedge Market Neutral Tracker ETF (NYSE: QMN), unlike the others, isn’t built on a long/short framework. Instead, QMN takes a fund-of-funds tack. Holding more than a dozen ETPs, QMN’s managers achieve beta neutralization with fixed income and currency exposures. More than 70 percent of the fund’s assets, in fact, are given over to non-equity products. Global stock exposure comes from ETPs tracking the Russell 2000 index and the MSCI EAFE and Emerging Market indices.

This setup may explain QMN’s relatively high correlations to equity and fixed income. Against the S&P 500 for example, QMN’s R-squared coefficient is .47, versus an average of zero for the other ETPs (1.00 is 100 percent correlated). It also dampens the fund’s volatility. With an annualized standard deviation of just 3.2 percent, QMN’s price variability is roughly half that of competing products.

QMN doesn’t suffer the cash drag of short sales but, as a trade-off, absorbs the passed-through costs of its acquired ETP portfolio. Still, with a total expense of 91 basis points (.91 percent), QMN is among the cheapest market neutral products. "

http://wealthmanagement.com/etfs/perils-market-neutral-funds

--- toread mb: http://www.investingdaily.com/10977/best-stock-screening-tools-on-the-web/ http://www.lagunabeachbikini.com/index.php/2012/11/08/screenings-stocks-or-etfs-by-sharpe-ratio/

this can screen by sharpe:

http://funds.us.reuters.com/US/screener/screener.asp

(but is it total return?.. YES, b/c 3-yr AGG price chart is negative but total return is positive, and the Sharpe shown is positive) (but i think it may be only 3-year Sharpe! at least it also has alpha, beta, R^2..)

Yahoo finance also shows a 3-yr, 5-yr and 10-yr Sharpe, and it seems to be total return (AGG is positive, and the 3-yr number is the same as what reuters reports; i bet they are all using the same datasource from morningstar or something): http://finance.yahoo.com/q/rk?s=AGG+Risk


todo look into schwab's research more. it doesnt appear to have Sharpe, tho (!)

optionsexpress's stock screen does not have Sharpe

this guy said Zecco used to have one that could screen for Sharpe, Zecco merged with TradeKing?: http://www.lagunabeachbikini.com/index.php/2012/11/09/screening-etfs-by-sharpe-ratio/ https://www.tradeking.com/trading/tools/screeners

todomb look at these high-Sharpe stocks: http://www.lagunabeachbikini.com/index.php/2012/11/09/screening-etfs-by-sharpe-ratio/


some current high Sharpe pics:

GSY: http://funds.us.reuters.com/US/etfs/overview.asp?symbol=GSY

Short/Intermediate Corporate YTD Return (Cumulative) 0.30% Classification Ultra-Short Oblig

" The Fund seeks maximum current income. The Fund uses a low duration strategy to seek to outperform the Barclays Capital 1-3 Month U.S. Treasury Bill Index in addition to providing returns in excess of those available in U.S. Treasury bills, government repurchase agreements and money market funds. "

Guggenheim Invest S&P 500 Eql Wght Health Care ETF (RYH)

http://funds.us.reuters.com/US/etfs/overview.asp?symbol=MINT.K

MINT ("maximum current income, consistent with preservation of capital and daily liquidity. The Fund invests under normal circumstances at least 65% of its total assets in a diversified portfolio of Fixed Income Instruments of varying maturities.") -- the so-called 'money market' ETF, i think

in order (as of Sep 3 2015; presumably the info in reuters isnt quite up to date, tho): GSY Enhanced Short Duration MINT money market RSY health care, equal weight FLRN.K floating rate bonds XLV health care; VHT, XHS, IYH, IXJ, FXH, RXL, IHF, PJP, CURE.K, PPH, KNOW.K insider sentiment FLOT.K floating rate bonds FPX US IPO PPA aerospace IHE health care PWB largecap growth; SCHG.K IHI health care ITA aerospace RPX large cap growth (by now we're at Sharpe 2.15) QQXT.O NASDAQ ex-tech FXG Consumer Staples BSCF.K Guggenheim bullet 2015 IYC consumer services XAR aerospace RYF financials, equal eweight VOE Mid-Cap Value FTC lg-cap growth TTFS.K mid-cap core VCR Consumer Discretionary Index DBEF.K EAFE 100% Hedged to USD (we're at Sharpe 2.04)

simplifying by consolidating similars: GSY Enhanced Short Duration; FLRN.K, MINT, FLOT.K RSY health care; XLV, VHT, XHS, IYH, IXJ, FXH, RXL, IHF, PJP, CURE.K, PPH, IHE, IHI KNOW.K insider sentiment FPX US IPO PPA aerospace; ITA, XAR PWB largecap growth; SCHG.K, RPX, QQXT.O, FTC FXG Consumer; IYC, VCR BSCF.K Guggenheim bullet 2015 RYF financials VOE Mid-Cap Value TTFS.K mid-cap core DBEF.K EAFE 100% Hedged to USD (we're at Sharpe 2.04)

ok now rescreen by requiring Lipper Consistent Return 10-year rating 4 or 5 (in the top 40% of Lipper's idea of consistent/risk-adjusted returns over 10 years). NOTE: this removes all the bonds, for some reason!:

PJP health care; IBB.O (bio) IYC consumer; VCR, XLY VO mid-cap ONEQ.O NASDAQ index RSP S&P 500/lg cap; IWB, VV, IVV, SPY VFH financials JKG midcap; IWR IUSG.K growth; IWF, SPYG.K, IVW, VTI total mkt; IWV, ITOT.K

ok now rescreen by dropping the Lipper and checking all of the category boxes, then unchecking them as we see 2 things from that category (unless they seem like outliers eg if they arent prototypical of the category theme) (i didnt always write down the second one)

GSY ultra-short obligation; MINT RYH sector health/biotech; IXJ Global Health/biotech; RXL leveraged health care ('equity leveraged funds') FLRN.K Short Inv Grade debt (they mean short duration, not shorting it); FLOT.K KNOW.K insider sentiment (midcap value; outlier?) FPX US IPO (multicap growth) PPA aerospace (industrials funds) PWB large-cap growth QQXT.O nasdaq ex-tech (multicap growth) FXG consumer goods; IYC consumer services BSCF.K Guggenhm Bullet 2015 ETF (Short-Intermediate Investment Grade Debt; outlier?) FXO financial services; RYF PXLG.K lg cap core; RWL; DBEF.K (international lg cap core) TTFS.K mid cap core; VO VOE mid cap value (this is the last Sharpe >=2 one!) DBJP.K Japanese dollar hedged; DXJ EDEN.K Denmark (European region) RSP, multicap core; PKW "Buyback Achievers"; FWDD.K, SPHQ.K, SCHB.K, VTI, VTHR.O, IWV, VOO S&P 500; SPY EEH Lg Cap US Sector Momentm VOT mid-cap growth RPV multicap value; IWS (midcap value); JKI; PRV; VTV lg cap value; VONV.O YCS (leveraged short yen) SCPB.K Short Inv Grade; FLTR.K EIRL.K ireland CWB convertibles; (Mixed Equity) (we're down to Sharpe 1.8 now) ALFA.K AlphaClone? Alt Alpha ETF (Long/Short Equity) BSJF.K Guggenhm Blt 2015 HY ETF (High Yield Funds); ANGL.K (fallen angels) VSPY.K Direxion S&P 500 Volatility Response Shares (Mixed Equity) EZM small-cap core; VB PEY High Yield Equity Dividend Achvrs; DTD FDN Dow Jones Internet Index Fund (Science & Technology); RYT (scitech); FONE.O (telecom) PXSG.K smallcap growth; VBK SPFF.K flexible income (Global X SuperIncome? Preferred north america); PFF; PowerShares? Financial Preferred Portfolio (Global Financial Services); PGX PSP Global Listed Private Equity Portfolio (Global Financial Services) DFE Europe SmallCap? Dividend Fund (International Equity Income) AOA (Mixed-Asset Target Allocation Aggressive Growth ) AOR Target Risk Growth Index (Mixed-Asset Target Allocation Growth; Mixed Equity) SDYL.K (leveraged dividends) BSCG.K Guggenheim BulletShares? 2016 EWK Belgium URTH.K MSCI World (International Large-Cap Growth) HEDJ.K WisdomTree? Europe Hedged Equity VBR small-cap value; IJJ; IVOV.K TOK Global Large-Cap Core ZIV VIX inverse medium term ('dedicated short bias'?!?) (we're down a little below Sharpe 1.6 now) IXG sector Global financials CYB Yuan currency, and money market rates in China (carry trade?) IXN sector global tech FAN sector global alternative energy; GEX (Global Natural Resources Funds); CLN.O (Natural Resources) VT global multicap core ACWV.K All Country World Minimum Vol (Global Multi-Cap Growth) SCZ International Small/Mid-Cap Core; MDD DVYL.K (leveraged dividend aristocrats) EIS Israel capped (Intl Sm/Mid-Cap Value) GIVE.K absolute return (we're down to Sharpe 1.4 now) DEF defensive equity (https://www.sabrient.com/Defender-1.html) BKLN (Loan Participation) EFAV.K EAFE Minimum Volatility (International Multi-Cap Growth) CUT timber (Basic Materials Funds); WOOD.O; XLB; VAW BWV BuyWrite?; PBP also buywrite (Long/Short Equity) IVOP.K VIX inverse MNA merger arb (Alternative Event Driven Funds ) DSUM.K Dim Sum Bond Index CSLS.K Credit Suisse Tremont Long/Short Equity Hedge Fund Index (Alt Long/Short Eq) IDHQ.K QSG Developed International Opportunities (International Small/Mid-Cap Growth; HQ firstworld) IGN sector north america telecom IOO global lg cap value CNY Renminbi currency PYZ basic materials IFEU.O international real estate XIV.O VIX inverse; SVXY GULF.O wisdom tree middle east dividends PDN RAFI Developed Mkts ex-US (International Small/Mid-Cap Core); FDT (Developed Markets Ex-US AlphaDEX?) ROOF.K US Real estate; KBWY.K AGA (leveraged short sector agriculture; ADZ; CMD (commodity)) CMBS.K US mortgage FNI Chindia (we're down to 1.1 Sharpe now) PGJ China HDG Hedge Repl ('Long/Short Equity') MATH.K tactical ('Flexible Portfolio') BSCI.K Guggenhm Bullet 2018 PXF RAFI Developed Mkts ex-US (International Multi-Cap Value) ; DWM ("companies in developed markets outside of the U.S. and Canada that pay regular cash dividends") EFV EAFE value; International Large-Cap Value IGF global infrastructure; GII

now let's look up other categories i'm interested in that i didnt see too much of:

BOND.K Core Plus Bonds CJNK.K Corporate Debt Funds BBB-Rated VPL Pacific region; FPA Pacific Ex-Japan; ADRA.O SCHO.K short US treasury; VGSH.O; TUZ; SHY BSCJ.K Guggenhm Bullet 2019 QAI Sharpe 0.87 ITR Intermediate Investment Grade Debt; CIU MES Glf States Id BUNL.K German Bond Futures BSCK.K Guggenhm Bullet 2020 JGBL.K Japan bond WPS intnl developed real estate; UJB High Yield VMBS.O US Mortgage (we're at .8 Sharpe now) RWO global real estate; FFR SMIN.K indian SMMU.K short muni INKM.K income; flexible portfolio COBO.K USD Covered Bond GMF emerging asia VCIT.O Intermediate Investment Grade Debt RALS RAFI Long/Short (.7 Sharpe) AGZ agency .61 Sharpe MLPN.K Energy MLP PCEF.K PowerShares? CEF Income Composite Portfolio; S-Network Composite Closed-End Fund Index "The Fund is a "fund of funds," as it invests its assets in the common shares of funds included in the Index rather than in individual securities. The Index currently includes closed-end funds that invest in taxable investment grade fixed-income securities, taxable high yield fixed-income securities and others utilize an equity option writing (selling) strategy" (Mix Tgt All Mod) IYLD Morningstar Multi-Asset Income CORP Core Plus Bond Funds; CBND (Intermediate Investment Grade Debt) VQT (Sharpe .52) AGG CRED (BBB), BND, LQD (BBB); BIV ~.5 Sharpe FEMS.K emerging small cap alphadex EEMA.O emerging asia EMCD.K emerging debt; CEMB.K .47 Sharpe EEMS.K emerging CHEP.K emerging markets neutral .4 Sharpe EMCB.O emerging debt; CEMB.K GYLD global high yield .3 Sharpe GTAA tactical .3 Sharpe PCY em sovereign debt .3 Sharpe EMB em bond EEMV Emerging Markets Minimum Volatility .28 Sharpe EWX em small cap CLY 10+ Corporate Debt Funds BBB-Rated BLV Corporate Debt Funds A-Rated long-term GMM em; BIK (BRIC), PIE (em momentum), SCHE.K, VWO, EMFT.K (largest 50) TIP .16 Sharpe EZA africa; .13 Sharpe BKF (BRIC) BICK.O DGS GREK.K FEM MOM EEM .06 Sharpe EEHB.K WDTI.K wisdomtree Managed Futures (-0.01 Sharpe) EMAG.K Latin America EM Agg Bd EELV.K PowerShares? S&P EM LV PERM.K Glbl X Permanent ETF (-0.08 Sharpe) ADRE.O Em Mkts 50 ADR WIP Intl Govt Inflation-Protected Bond EEB BRIC ADR LTPZ (-0.16 Sharpe) CSMA.K merge arb TIP -0.27 Sharpe

i think we're missing commodities b/c they crashed..

note: can also check 10-yr Sharpes at eg http://performance.morningstar.com/funds/cef/ratings-risk.action?t=IWM which has 15 year Sharpes too for some. These ratings differ from the above for many. Eg 10-yr TIP is .42 in Morningstar, 3 yr Sharpe is -.27 above, and -.3 in morningstar.

here they are with the longer Sharpes (those that go back to 10 yrs or longer only; only the guys i listed above in the lefthand col; in descending order of Sharpe):

SYM          timeperiod              Sharpe       recent-Sharpe       (3 yrs)
AGG                10.0                0.88                 0.5       
ONEQ                10.0                0.52                 1.5      
VCR                10.0                0.51                1.59       
JKG                10.0                0.48                1.54       
VO                10.0                0.47                1.61        
VTI                10.0                0.46                1.46       
RSP                10.0                0.46                1.55       
PWB                10.0                0.46                1.68       
XLV                15.0                0.46                2.04       
TIP                10.0                0.43                -0.3       
VBR                10.0                 0.4                 1.3       
IXN                10.0                0.39                 1.1       
VTV                10.0                0.39                1.42       
IYC                15.0                0.37                1.61       
PGJ                10.0                0.36                0.71       
EZA                10.0                0.32                -0.0       
PXSG                10.0                0.28                1.23      
EWK                15.0                0.27                1.25       
EEM                10.0                0.26               -0.15       
IOO                10.0                0.26                0.84       
ADRE                10.0                0.24               -0.25      
VPL                10.0                0.22                0.62       
EFV                10.0                0.19                0.66       
IGN                10.0                0.17                 0.9       
PEY                10.0                0.16                1.51       
VFH                10.0                0.11                1.49       
IXG                10.0                0.09                1.06       
IUSG                15.0                0.09                1.47      

Charts that report 'total return'

https://www.bogleheads.org/forum/viewtopic.php?t=89613 todomb reply to this and cite my findings below (stockcharts, etfreplay) after i'm sure they're good

schwab's 'growth of 10,000 investment' chart

example that shows big growth in total return but not in price, over 3 years from Oct 2012 to Aug 2015: AGG (price from 111 to 109, total return factor 1.05)

DOES NOT show total return (at time of writing): Yahoo finance

DOES show total return (at time of writing): http://www.etfreplay.com/charts.aspx http://funds.us.reuters.com/US/etfs/overview.asp?symbol=AGG (only on OVERVIEW tab, not on CHARTS tab) http://stockcharts.com/h-perf/ui

---

todo find more high Sharpe pics, eg via http://funds.us.reuters.com/US/screener/screener.asp todo look for more Sharpe stock screeners, there's gotta be a few

---

http://blogs.cfainstitute.org/investor/2012/09/11/chasing-warren-buffett-alpha/

---

btw if you ever find some historical stock certificate lying around and want to know about it, here's a company that apparently does that:

http://www.scripophily.com/

---

https://www.google.com/search?q=otcbb+index&oe=utf-8&oq=otcbb+index&gs_l=mobile-heirloom-serp.3..0.1723.8207.0.8427.25.18.1.3.3.1.367.3484.3j6j4j4.17.0....0...1c.1.34.mobile-heirloom-serp..11.14.1764.BJR5vwuURLE

---

http://www.investopedia.com/ask/answers/05/nanomicrocapindex.asp

---

what is this?: http://www.bloomberg.com/quote/BSDONX:IND

---

https://www.google.com/search?q=otc+%22sharpe+ratio%22+portfolio&oe=utf-8&oq=otc+%22sharpe+ratio%22+portfolio&gs_l=mobile-heirloom-serp.3...18035.20827.0.20955.15.10.0.0.0.3.487.1462.4j2j2j0j1.9.0....0...1c.1.34.mobile-heirloom-serp..13.2.455.0SZC1aau87E

https://www.google.com/search?q=pink+sheet+%22sharpe+ratio%22+portfolio&oe=utf-8&oq=pink+sheet+%22sharpe+ratio%22+portfolio&gs_l=mobile-heirloom-serp.3...20115.23284.0.23630.10.10.0.0.0.0.338.1652.1j7j1j1.10.0....0...1c.1.34.mobile-heirloom-serp..7.3.459.BPG50iwCy4g

https://www.google.com/search?q=pink+sheet+%22sharpe+ratio%22+portfolio&oe=utf-8&oq=pink+sheet+%22sharpe+ratio%22+portfolio&gs_l=mobile-heirloom-serp.3...20115.23284.0.23630.10.10.0.0.0.0.338.1652.1j7j1j1.10.0....0...1c.1.34.mobile-heirloom-serp..7.3.459.BPG50iwCy4g

https://www.google.com/search?q=how+to+buy+broad+nanocap+protfolio&ie=utf-8&oe=utf-8

https://www.google.com/search?q=nanocap+portfolio&oe=utf-8&oq=nanocap+portfolio&gs_l=mobile-heirloom-serp.3...89875.95967.2.96285.38.18.0.1.1.0.927.9139.1j6j1j1j6-9.18.0....0...1c.1.34.mobile-heirloom-serp..39.4.1087.ubWxV8hkWx0

---

http://www.efalken.com/papers/Taleb2.html

---

" Sharpe Ratios reported by hedge fund indices are commonly calculated without regard to their monthly serial correlation . This practice often has the effect of understating a return stream's annual standard deviation and thus overestimating a particular strategy’s risk - adjusted return . The "square root of time" formu la as an estimator for the annual standard deviation of hedge fund return s is imprec ise when serial correlation exists in the data ... How is the Sharpe R atio for the Credit Suisse Broad Hedge Fund Index currently calculated? Credit Suisse divides the index's annual geometric mean return less the Risk Free Rate ( Credit Suisse uses the annualized rolling 90 day T - bill rate as an estimate for the Risk Free R ate ) by the index's annualized monthly standard deviation :

(Annual Geometric Mean Return - Annualized rolling 90 day T - bill Rate) / Standard Deviation of Monthly Returns * SQRT (12)

...

The distinction is important , particul arly because you can easily calculate the standard deviation of the i ndex's 2 0 years of annual returns (1994 - 2013) and discover that the actual measured standard deviation of annual returns is sign ificantly higher (10.92%) than the annualized monthly st andard deviation calculated above by Credit Suisse (7.28%) . Taking it one step further, i f you were to insert the standard deviation of annual returns (10.92%) in to the equation above in pl ace of the annualized mont hly standard deviation (7.28 %), you wou ld calculate a Sharpe Ratio of 0.53 as demonstrated below:

(Annual Geometric Mean Return - Annualized rolling 90 day T - bill Rate) / Standard Deviation of Annual Returns

(8.65% - 2.82%) 10.92 %

5.82% 10.92 %

0.53

Using the annual ized monthly standard deviation , as opposed to the annual standard deviation , results in a Sharpe Ratio that is overstated by 50% .

...

Over the years, the financial community has become so accustomed to the term "annualized standard deviation" that it has i mproperly come to equat e it with "annual standard deviation" . The reason we use "annualized standard deviation " is to estimate the annual standard deviation.

...

in the event that monthly da ta does not exhibit serial correlation, a calculated annual standard deviation is just as suitable an estimator of the population annual standard deviation as a formula that uses an annualized monthly standard deviation . In the event that serial correlation is present in the monthly data, it is preferable to use t he annual data to calculate an annual standard deviation. It does not matte r how many years of data are available (as long as it is 2 or greater) because "24 months " (or 36, 60, 240, e tc.) of data results in a comparable annual standard error as 2 years (or 3, 5, 10, etc.).

...

Statistician Gerald van Belle developed a simple formula (which can be used for any regression with serial correlation when used to calculate a t - s tatist ic) which we have adapted for annualizing monthly standard deviations that accounts for serial correlation, when serial correlation is present 2 . The steps are as follows:

1. Calculate the serial correlation ("v") of the monthly return stream using a one mont h lag period ;

2. Calculate the monthly standard deviation of the return stream ;

3. Multiply the monthly standard deviation of the return stream by the square root of time (SQRT (12)) ; and

4. D ivide the result calculated in Step 3 by SQRT((1 - v)/ ( 1+v)) where "v" is the measured serial correlation calculated in Step 1

...

Table 2: Sharpe Ratios of Hedge Fund Strategies using three different calculation techniques:

A. Using annualized monthly standard deviation

(Ann ual Geometric Mean Return - Annualized rolling 90 day T - bill Rate) / Standard Deviation of Monthly Returns * SQRT (12)

B. Using the van Belle adjusted annualized monthly standard deviation

(Annual Geometric Mean Return - Annualized rolling 90 day T - bill Rate) / ( Standard Deviation of Monthly Returns * SQRT (12) ) / SQRT((1 - v)/(1+v))

C. Using annual standard deviation

(Annual Geometric Mean Return - Annualized rolling 90 day T - bill Rate) / Standard Deviation of Annual Returns "

-- http://www.msrinvestments.com/Sharpe%20Ratios%20Reported%20by%20Hedge%20Fund%20Indices%20Underestimate%20Annual%20Standard%20Deviation.pdf

http://www.msrinvestments.com/Sharpe%20Ratios%20Reported%20by%20Hedge%20Fund%20Indices%20Underestimate%20Annual%20Standard%20Deviation.pdf also provides a corrected table of hedge fund category returns on PDF page 5

---

http://blogs.cfainstitute.org/investor/2012/09/11/chasing-warren-buffett-alpha/

" From November 1976 to the end of 2011, Warren Buffett delivered an average annual return of 19% in excess of the Treasury bill rate, as measured by shares of his publicly traded conglomerate, Berkshire Hathaway (BRK.A, BRK.B), versus a 6.1% average excess return for the stock market. In addition, Berkshire’s Sharpe ratio — a measure of return per unit of risk — is higher than all U.S. stocks that have been traded for more than 30 years from 1926 to 2011, as well as all U.S. mutual funds in existence for more than three decades.

So how does he do it?

If a newly published paper is any guide, the answer is pretty straightforward. According to “Buffett’s Alpha http://www.cfainstitute.org/learning/products/publications/contributed/Pages/buffett_s_alpha.aspx ,” authored by AQR Capital Management‘s Andrea Frazzini, David Kabiller, CFA, and Lasse Pedersen, who also teaches finance at the NYU Stern School of Business, Buffett buys low-risk, cheap, and high-quality stocks; he employs modest leverage to magnify returns; and he sticks to his investment discipline even during rough periods in the markets that would force investors with less conviction or capital “into a fire sale or a career shift,” as the authors put it.

Previous researchers analyzing Buffett’s returns using conventional size, value, and momentum factors haven’t been able to adequately explain his outperformance, the authors say, leaving admirers to conclude that Buffett’s magic is pure alpha. So they extend the analysis by testing Buffett’s impressive returns — as measured by Berkshire’s stock — against two factors that better reflect his folksy investing wisdom: One called “Betting Against Beta,” which represents safe, low-beta stocks, and another called “Quality Minus Junk,” which represents the stocks of high-quality companies that are profitable, growing, and paying dividends.

The results? “Controlling for these factors,” the authors write, “drives the alpha of Berkshire’s public stock portfolio down to a statistically insignificant annualized 0.1%, meaning that these factors almost completely explain the performance of Buffett’s public portfolio.” The factors also explain “a large part” of Berkshire’s overall stock return, the authors add, as well as Berkshire’s private portfolio, insofar as their alphas also become statistically insignificant.

As one commentator put it, “It’s some evidence that Buffett is doing what he says he’s doing.” But the takeaway is more nuanced. Buffett is in fact best known as a value investor par excellence, yet the authors’ findings suggest that his focus on safe, quality stocks “may in fact be at least as important” as his value bent in accounting for his consistent outperformance.

One of the most interesting aspects of the paper is its analysis of Buffett’s use of leverage. The authors deconstruct Berkshire’s balance sheet and find that on average the conglomerate is levered 1.6 to 1, which they describe as “non-trivial” and say at least partly explains why the volatility of its stock is high relative to the market — 24.9% versus 15.8% — despite investing in many relatively stable businesses. Still, they note that leverage alone does not account for Buffett’s stellar returns: Applying the same 1.6-to-1 leverage to the market yields an average excess return that is still nine percentage points below Buffett’s over the time span studied by the authors.

Of course, cheap financing doesn’t hurt. The authors note that Buffett benefited from Berkshire’s AAA rating from 1989 to 2009 and that he reaps the benefit of its cheap insurance float, which checks in at an estimated average annual cost of 2.2% — more than 3 percentage points below the average Treasury bill rate. The authors find that 36% of Berkshire’s liabilities, on average, consist of insurance float.

The paper also tackles a provocative question: Can Warren Buffett be reverse engineered? The authors take a stab at it by constructing a hypothetical “Buffett-style strategy” that is similarly leveraged and tracks Buffett’s market exposure and stock-selection themes — and find that it “performs comparably” to the real Berkshire Hathaway. (In fact, it outperforms, but the authors caution that the simulated strategy does not account for transaction and other costs, and also benefits from hindsight. The main takeaway, they assert, is the high covariation between the actual and simulated Buffett strategies.)

So why don’t investors just mirror Buffett’s trades? Blame hubris: As detailed in a separate academic paper published a few years ago, between 1980 and 2006 an investor could have achieved investment results similar to Buffett’s simply by following his trades as disclosed in public filings — yet the market seemed to underreact to such disclosures. The authors of the paper surmise that analysts and fund managers overestimate their own stock-picking skill or the “precision of their independent private information” and underweight the value of public disclosures, even Buffett’s.

There is at least one more very practical reason why investing like Buffett is harder than it may seem: Thanks to current U.S. Securities & Exchange Commission disclosure rules, he doesn’t always have to tell us what he owns." -- http://blogs.cfainstitute.org/investor/2012/09/11/chasing-warren-buffett-alpha/

http://blogs.cfainstitute.org/

---

https://www.google.com/search?client=ubuntu&channel=fs&q=stock+screener+sharpe&ie=utf-8&oe=utf-8

---

margin: a brokerage account with 'margin' means the broker will lend you money. The amount of margin you can get is generally a multiple of the account value (often 2), adjusted by considerations like how risky your broker thinks your investments are. If the net value of your account gets too low, the brokers will generally try to liquidate the account before the balance goes negative, but if they screw up they SOMETIMES DO try to get you to pay them the negative balance! ( http://www.forexcrunch.com/some-brokers-still-going-after-negative-balances/ ). The broker charges you interest on the loan, as you might expect; but it's typically less interest than, say, a credit card or a mortgage (you can't just wire all of your money, including the borrowed money, out of the account to do something else with it, because the amount of margin they'll give you is a multiple of your account value).

i've heard of 3 ways that trading with an account with 'margin':

---

interesting but exposed to stock market crashes:

hvpw http://www.etftrends.com/2013/03/generating-income-with-the-new-put-write-etf/

http://stockcharts.com/h-perf/ui

huh, this related index did better than the SPY: http://www.cboe.com/delayedquote/advchart.aspx?ticker=PUT

---

https://orchardplatform.com/blog/2014521diversification-of-online-lending-portfolios-prosper/

https://orchardplatform.com/blog/bitcoin-powered-lending-with-btcjam-and-bitbond/

https://www.bitbond.com/

https://btcjam.com/

pdf page 18 has a list of best and worst options ETF performers: http://www.etf.com/docs/magazine/1/ETF_Report_July2014.pdf

---

https://www.aqr.com/library

---

financial markets have (at least) three roles:

note that options, because they are offered at different strikes and with different expirations, allow society to query the mind of Mr. Market for an entire PDF of some value at some future time (as opposed to just the price of the underlying, which is sort of like the just getting to know a single number (the number being something like the expectation of the discounted future value of some asset)).

---

what Sharpe ratio is 'good'?

the statistics of Sharpes:

analysis: potential cutoffs mentioned above for a good Sharpe range from .2 (SPY) to 2.99 (best performing hedge fund category in Lo's table). The geometric mean of that is .77. Throwing out the high and lowest numbers from the above (.2 from SPY and 2.99 for the highest category in Lo's hedge fund table), the range is .4 to 2; the geometric mean is .9. Throwing out the highest 2 numbers mentioned above (.4 from predicting the future without leverage in a 40-60 portfolio, 2 from the cutoff for 'what marketing dept?' in Muller's interview article), we get .5 to 1.4; geometric mean of that is .84. This is also the Sharpe of the Vanguard 500 Index as calculated in the Lo article.

So one could say:

Perhaps a more robust statement is:

---

td ameritrade has an ETF screener with Sharpe but it seems to only be the trailing 3-year one (same as reuters)

morningstar has Sharpe ratings up to 15-years but their ETF screener doesn't offer it as a criterion.

eg

http://performance.morningstar.com/funds/cef/ratings-risk.action?t=AGG

http://funds.ft.com/us/Screener/PreScree has Sharpe but it also seems to be 3-year

https://research.tradeking.com/research/etf-screener.asp has Sharpe but it also seems to be 3-year; but at least its UI is nicer than the Reuters one (results display the full name of the symbol;

---

when you look at some stock screeners, sometimes you see 'Z score'. This is not the usual stats metric, instead this is short for 'Altman Z-Score', which is a formula intended to serve as a red flag for companies that may be going bankrupt (when the Z score is lower than about 2, or maybe 1 or 3). It's a formula:

(Working Capital / Total Assets)*1.2 + (Retained Earnings / Total Assets)*1.4 + ( Earnings Before Interest and Taxes / Total Assets)*3.3 + (Market Value of Equity / Book Value of Total Liabilities)*0.6 + (Sales/ Total Assets)*1

http://www.stockopedia.com/screens/altman-z-score-screen-4/ http://www.stockopedia.com/content/the-altman-z-score-is-it-possible-to-predict-corporate-bankruptcy-using-a-formula-55725/

(as a thought experiment, what would Altman's Z-score look like for a startup on day 1? Well, a startup would probably be capitalized by its founder(s) upon incorporation, probably at least partially in the form of cash (working capital), and would have no earnings, sales, or liabilities. So the first factor would be close to 1, the second and third would be close to 0, the fourth would be close to infinity, and the fifth would be close to 0. Because the fourth term has liabilities in the denominator, Altman's Z cannot go below 3 until Book Value of Total Liabilities > .2 * Market Value of Equity; that is, a necessary condition for Altman's Z score to wave a red flag is for Book Value of Total Liabilities to exceed 20% of market cap; so even a brand-new startup "doesn't have to worry" (about THIS metric at least!) until that is happening. What if our brand-new startup starts selling a lot, increasing total liabilities, but without making earnings yet; what could prevent Altman's Z from going low? Then the working capital and sales factors would have to compensate; the last factor could stand on its own if sales were 3x assets (the first factor cannot stand on its own because working capital can't exceed total assets, i think))


another screener/tool site, doesnt have Sharpe afaict: https://www.macroaxis.com/

probably a good idea to plug your portfolio into that and see what it says, at least

---

get the 15-year Sharpes for each of the above ETFs

---

interesting theory for the market crash in Aug 2015:

" Risk-parity strategies are designed for low volatility and generally allocate more toward bonds than equities. Often, though, clients want additional volatility, so a manager applies leverage through borrowing.

Chintan Kotecha, an equity-linked analyst at Bank of America, said that these types of portfolios have been known to add what is known as a "volatility control" component, a mechanism that responds to changes in market volatility by leveraging or deleveraging.

In the last several days, the equity market fell sharply and investors did not rush to buy up safe-haven bonds, causing the bond market to decline modestly as well.

The lack of bond buying caused a big spike in volatility in these funds - more than investors wanted - so they responded with a dramatic level of selling, particularly in equities. "

see also http://www.brandes.com/docs/default-source/brandes-institute/the-risks-of-risk-parity

table on the last page of that PDF: " Traditional Portfolio Allocation (%) Contrib. to Returns (%) Contrib. to Volatility (%) U.S. Equity 37 46 57 Non-U.S. Equity 15 15 25 Fixed Income 37 26 3 Real Estate 8 11 12 Commodities 3 2 2 Source: FactSet?, as of 12/31/2013 Risk Parity Portfolio Allocation (%) Contrib. to Returns (%) Contrib. to Volatility (%) U.S. Equity 10 15 19 Non-U.S. Equity 9 11 20 Fixed Income 58 46 15 Real Estate 11 17 27 Commodities 12 10 19 "

---

random argument that P/E/CAPE ratios has been getting higher in the long term:

http://m.fool.com/investing/general/2014/03/10/is-the-stock-market-overvalued-it-all-depends-on-h

note: the CAPE ratio is currently LOWER than when that article was written:

http://www.multpl.com/shiller-pe/

http://www.multpl.com

here's the bear case based on valuations, fed tightening:

http://www.businessinsider.com/stock-market-crash-2015-7

note: the Buffet ratio is high but not so much that Buffet is worried: "Warren Buffett has gone on record saying that he believes the best way to check the broader market’s valuation is to look at total stock market capitalization against GNP.

In a speech given in 2001, he stated:

    If the percentage relationship falls to the 70% or 80% area, buying stocks is likely to work very well for you. If the ratio approaches 200% — as it did in 1999 and a part of 2000 — you are playing with fire.

We can see that total stock market cap/GDP (works out to be quite similar) is just over 125%.

Certainly not cheap. Although, the ratio was 133% back when Buffett gave that speech in 2001 and he stated:

    I would expect now to see long-term returns run somewhat higher, in the neighborhood of 7% after costs.

“Somewhat higher” is in reference to a similar speech he made back in 1999, when the ratio was much higher and Buffett thought long-term returns would thus be lower. So this ratio apparently doesn’t indicate the need for panic. Furthermore, Buffett just spoke on this subject quite recently when asked about it at the Berkshire Hathaway Inc. (BRK.B) annual shareholders meeting and stated that he only views stocks as currently expensive if rates start to rise. Otherwise, they’re not particularly cheap or particularly expensive." -- http://www.dividendmantra.com/2015/06/waiting-for-a-correction-when-many-stocks-have-already-corrected/


global or international P/E ratios: http://www.starcapital.de/research/stockmarketvaluation

---

hmm seems like even ex-US international ETFs that have done relatively poorly over the last three years declined recently in synch with SPX, as has energy:

http://stockcharts.com/h-perf/ui?s=$SPX&compare=EMAG,TIP,EZA,EWX,SMIN&id=p03875156107 http://stockcharts.com/h-perf/ui?s=$SPX&compare=TIP,EZA,EWX,SMIN&id=p81269214377 http://stockcharts.com/h-perf/ui?s=$SPX&compare=GMM,EZA,GREK,SMIN&id=p28898120770 http://stockcharts.com/h-perf/ui?s=$SPX&compare=GMM,EZA,MLPN&id=p70408936037 http://stockcharts.com/h-perf/ui?s=$SPX&compare=GMM,EEB,MLPN,USCI&id=p49818554250

this seems to add support to the narrative that the entire world's financial markets have become coupled (global beta)

global beta is expected in any case, but a further concern right now is that emerging economies may have become dependent upon US Fed stimulus:

"Many emerging market economies are concerned that a Fed rate rise would trigger large outflows of capital from emerging economies into dollar-denominated assets, creating market turmoil that would hurt growth." -- http://www.reuters.com/article/2015/09/05/us-g20-imf-fed-idUSKCN0R50PE20150905

---

 "

Roughly 95% of all exchange-traded products are “passive,” meaning they are designed to replicate the benchmark of an index. Actively-managed ETFs have gathered momentum over the years; there are now about 114 active exchange-traded products, which, combined, manage over $16 billion in assets. Be sure to check out our Actively-Managed ETF Portfolio.

Below is a table of the five most popular actively-managed funds:

Ticker ETF (MINT A+)

Enhanced Short Maturity Strategy Fund (BOND B+)

Total Return Exchange-Traded Fund (EMLP B+)

North American Energy Infrastructure Fund (FTSM n/a)

Enhanced Short Maturity ETF (SRLN B+)

SPDR Blackstone GSO Senior Loan ETF "

-- http://etfdb.com/etf-industry/a-visual-guide-to-the-etf-universe/

---