# rmoorman/gwern.net forked from briansniffen/gwern.net

Fetching contributors…
Cannot retrieve contributors at this time
254 lines (155 sloc) 40 KB
 --- description: Cookie prices & tax filings as evidence of corporate inefficiency ... I was reading a new blog, [ribbonfarm.com](http://www.ribbonfarm.com/) by the interesting & polymathic Venkatesh Rao, because [one post](http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-or-the-office-according-to-the-office/) on the true organization of corporations had been linked approvingly [on LessWrong](http://lesswrong.com/r/all/lw/1pq/rationality_quotes_february_2010/1jtf); he presents a lifecycle of corporations by way of describing how a corporation inevitably degenerates into operating. They are constantly being born and then ripped apart by hungry entrepreneurs (a nicer name for Venkatesh's 'sociopaths'), and this creative destruction keeps things relatively efficient and cuts down on indefinite [rent-seeking](!Wikipedia) and other such problems. Very few indeed survive more than a century.[^darwin] [^darwin]: [Mike Darwin](!Wikipedia), ["The Armories of the Latter Day Laputas, Part 5"](http://chronopause.com/index.php/2011/07/11/the-armories-of-the-latter-day-laputas-part-5/): > "In 2009, the Japanese business analysis and survey firm Tokyo Shoko Research (a combination of Dunn and Bradstreet and TRW in Japan), conducted an examination of the founding dates of the 1,975,620 enterprises in their database.They found 21,666 companies which have existed for over 100 years. The Bank of Korea conducted a similar evaluation of their database and found that there are 3,146 firms founded over 200 years ago in Japan, 837 in Germany, 222 in the Netherlands and 196 in France. There are 7 companies in Japan over 1,000 years old; 89.4% of the companies with over 100 years of history are for profit businesses. However, a closer examination of the history of these long-surviving enterprises reveals that many underwent takeovers, buyouts and essentially a complete restructuring of mission and the nature of the business the firms were engaged in – often more than once in their history. Thus, the chances of a business entity (excluding religious and academic institutions) surviving for >100 years is 1.096%." Physicist [Geoffrey West](!Wikipedia) offers [the fruits of his own research](https://www.nytimes.com/2010/12/19/magazine/19Urban_West-t.html) into corporate longevity and the life cycle: > "But it turns out that cities and companies differ in a very fundamental regard: cities almost never die, while companies are extremely ephemeral. As West notes, Hurricane Katrina couldn’t wipe out New Orleans, and a nuclear bomb did not erase Hiroshima from the map. In contrast, where are Pan Am and Enron today? The modern corporation has an average life span of 40 to 50 years. This raises the obvious question: Why are corporations so fleeting? After buying data on more than 23,000 publicly traded companies, Bettencourt and West discovered that corporate productivity, unlike urban productivity, was entirely sublinear. As the number of employees grows, the amount of profit per employee shrinks. West gets giddy when he shows me the linear regression charts. “Look at this bloody plot,” he says. “It’s ridiculous how well the points line up.” The graph reflects the bleak reality of corporate growth, in which efficiencies of scale are almost always outweighed by the burdens of bureaucracy. “When a company starts out, it’s all about the new idea,” West says. “And then, if the company gets lucky, the idea takes off. Everybody is happy and rich. But then management starts worrying about the bottom line, and so all these people are hired to keep track of the paper clips. This is the beginning of the end.” The danger, West says, is that the inevitable decline in profit per employee makes large companies increasingly vulnerable to market volatility. Since the company now has to support an expensive staff — overhead costs increase with size — even a minor disturbance can lead to significant losses. As West puts it, “Companies are killed by their need to keep on getting bigger.”" But of course, some corporations last a very long time. Millennia is [not unheard of](!Wikipedia "List of oldest companies"), and the Catholic Church resembles modern corporations in many respects. It's interesting looking at their areas of expertise - food, construction, and religion seem to predominate. # Barriers to entry All areas with what one might call 'unearned' advantages; if a temple tells you that Buddha endorses that temple and worshiping elsewhere might condemn you to the [hell](!Wikipedia "Naraka (Buddhism)") or rebirth as a [hungry ghost](!Wikipedia "Preta"), they probably didn't earn that endorse on the free market with really good homiletics and especially effective philanthropy. If a construction company has been building the local Shinto temples for the past couple centuries, they'll go on getting business regardless of whether they're doing an efficient job. The [candies](!Wikipedia "Mochi") produced near [Ise](!Wikipedia "Ise Grand Shrine") will go on selling well despite transgressions like selling [expired frozen goods](http://www.nytimes.com/2007/10/31/world/asia/31iht-31japan.8123604.html) as fresh and forging the documentation. When King Henry VIII expelled the Catholic Church and seized [monastery assets](!Wikipedia "Dissolution of the Monasteries"), he seized something like $\frac{1}{4}$ of the entirety of England^[I also found estimates which went as high as $\frac{1}{3}$!]; one has to wonder if the monasteries were *that* good at agriculture & business. Modern corporations have seamy underbellies even if they seem utterly unobjectionable^[Why is IKEA, normally considered a paragon of virtue with its commitment to renewable energy, such a [bizarre](!Wikipedia "IKEA#Corporate structure") [legal labyrinth](http://www.economist.com/node/18229400)? What moral high ground was forfeited when IKEA decided to help finance Communist regimes by shifting manufacturing to cheaper Poland by 1965? Much of _[The Omnivore's Dilemma](!Wikipedia)_ explores the deep compromises made by American organic farming in order to scale to large sales volumes. The images of IKEA and organic farming remain largely intact & untarnished despite the considerable coverage both have received over the years.]. So if one sees an old corporation, one is justified in asking how it has survived so long - does it have some sort of unique stable niche it alone can fill, or an unfair advantage? It might be a simple case of luck. With millions of corporations over the years, surely a few will simply happen to survive a long time. But it seems to me that if this were the case, one would expect to see such corporations shifting niches and surviving in competitive areas for long periods; but instead, if one looks at examples like the *former* oldest corporation in the world, [Kongo Gumi](!Wikipedia), one sees a story which runs more like this: they lived in a protected niche for an extremely long time and then they died shortly after venturing into more competitive areas; and indeed, Kongō Gumi was founded in 578 AD as a *temple* construction company and [went bankrupt](http://www.businessweek.com/smallbiz/content/apr2007/sb20070416_589621.htm) after it tried to expand into general real estate in the middle of a [bubble](!Wikipedia "Japanese asset price bubble").^[The [real estate bubble](!Wikipedia) aspect makes the tale of Kongō Gumi even sadder; a 1400-year old corporation has seen many bubbles come and go on the local and international scene, and so really ought to know better.] Gumi illustrates another failure mode: a single leader can draw on the accumulated capital (financial & social) of an ancient institution. How did Gumi, a relatively small construction company in a shrinking business, have enough credit to hang itself? Perhaps because it was such an ancient institution. The folly of one generation can be fueled by the prudence of many generations. Charities are usually [corporations](!Wikipedia "501(c)"), incidentally, with curious spending patterns.[^fivepercent] [^fivepercent]: Tyler Cowen, _Good and Plenty: The Creative Successes of American Arts Funding_: > Since 1969, tax-exempt nonprofit institutions must spend a certain percentage of their income each year. At first the clause required either 6 percent or the rate of return on assets, whichever was greater. In 1981 the Internal Revenue Service revised the rule and put it at 5 percent.... > > This little-known rule is a significant part of American arts policy. On average, most foundations spend 5 to 10 percent of their assets, with many foundations clustered right at the 5 percent level. The legal constraint thus appears to bind in many cases. Indeed when the legal requirement changed from 6 to 5 percent in 1981, the aggregate grant rate fell from 8 to 6 percent over the following decade. Before 1969 many foundations were little more than methods of tax evasion and storehouses for private wealth. They were explicitly marketed to donors in those terms, and 70 percent of business-linked foundations had no programs for charitable giving whatsoever. The Ford Foundation, a leading arts donor, had been set up in 1937 but did not actively make grants until the 1950s...Note that from 1981 to 1999, 84.5 percent of the increase in foundation assets (which rose almost threefold over that period) came from new donations and the creation of new foundations, rather than from returns on existing foundation assets [Mehrling, Perry. 1999. “Spending Policies for Foundations: The Case for Increased Grants Payout.” Economics Department, Barnard College] That is all background for the main question. # Cookie prices On a boardwalk in San Francisco, we passed some [Girl Scouts](!Wikipedia) selling the [famous cookies](!Wikipedia "Girl Scout cookies"). We bought 1 box of thin mint [Girl Scout cookie](!Wikipedia)s. It cost $4. (I asked around, and they're going for$4 in San Diego too.) $4 is a lot, really, compared to some things. I could get a nice meal of smoked trout for twice that, or buy a bunch of used books, or host my website for a few months. All of those would last longer too - a box of thin mints could disappear in a few minutes. As well, they're pretty good cookies, but they aren't many times greater than similar chocolate cookies. If they weren't from the Girl Scouts, a _bona fide_ charity giving money to which makes a good person, I wouldn't have spent the hundreds of dollars that I have.^[$20 would be a small Girl Scouts cookie order for me, and I have bought for many years.] And it wasn't always a solid $4. It used to be a nice$2.50, easy to explain: 8 boxes for a Jackson, 4 for a Hamilton, 2 for a Lincoln. And before that it was $2. And that wasn't that long ago either; I was selling cookies (with my sisters) at those prices in the '90s. ## Cookie prices and inflation Cookies aren't computers, of course, but this doubling exceeds inflation. In [1992](http://community.seattletimes.nwsource.com/archive/?date=19930513&slug=1701067), boxes were$2-3; in 2006 or 2008, you could [find boxes](http://web.archive.org/web/20110718074845/http://www.dchieftain.com/news/78448-03-09-08.html) for $3.50; in 2011 they are$4^[There are scattered claims of boxes selling for [$4.50](http://authenticorganizations.com/harquail/2009/08/03/wal-mart-knocks-off-the-girl-scouts/#comment-1214). Even if currently untrue or just isolated incidents,$4.50 is inevitable.]. Total inflation from 1992-2011 was [57%](http://data.bls.gov/cgi-bin/cpicalc.pl?cost1=100&year1=1992&year2=2011) (presumably core inflation was lower), but the price increased 100%. From 2006-2011, annual US inflation was close to 1% over that period and the net inflation was [9.25%](http://data.bls.gov/cgi-bin/cpicalc.pl?cost1=100.00&year1=2006&year2=2011); there was a 14% price increase. It is true that food prices are [highly volatile](http://www.cbsnews.com/stories/2011/03/16/business/main20043737.shtml), but those prices are already included in the inflation figures and large national bakers likely hedge their commodity supplies with futures contracts like airliners hedge their oil supplies. If we look at wheat prices in particular, the [UIll. farmdoc](http://www.farmdoc.illinois.edu/manage/uspricehistory/us_price_history.html) dataset puts 1992 wheat at $3.38 a bushel and 2006 wheat at$4.04; [USDA statistics](http://usda.mannlib.cornell.edu/usda/nass/AgriPric//2010s/2011/AgriPric-05-31-2011.pdf) put 2010 prices in a range from $4.16 a bushel (June) to$6.45 a bushel (December), and [1992 prices](http://usda.mannlib.cornell.edu/usda/nass/AgriPric//1990s/1992/AgriPric-12-31-1992.pdf) at $3.34 (December). The 2011 volatility aside, can wheat prices explain the price increase from 1992-2010? Taking the 2 highest numbers ($3.38 and $6.45) produces an increase of around 90%, so wheat's price outpaced the overall 57% inflation but cannot explain the entire doubling (how much of the cost of manufacture could all the ingredients represent? And then how much of that would be wheat?). In fact, even if all the ingredients experienced a similar 90% increase, this increase would explain no more than half the doubling, since the total cost of manufacture - in 2011 - represents only around$1 of the price, so the ingredients' additional cost could be no more than 90¢. Perhaps the Girl Scout leadership is to blame. # Judging charities Presumably we would 'blame' a leadership if it was failing to accomplish its publicly claimed goals and serving itself instead (the fancy economics term for this is the [principal-agent problem](!Wikipedia)). A charity focused on character-building and recreation is a bit harder to judge than one trying to stamp out polio; the latter can just point to falling death-rates.^[Indeed, 'saving lives' is an easy enough metric that there is a 'meta-charity', [Givewell](!Wikipedia), which ranks charities by efficiency; it is very sobering to reflect that the price of a life is just $1000 or$2000. If one's ethics say that failing to stop a murderer is as morally bad as murdering the victim yourself, then that leads to an unpalatable conclusion about every $2000 one spends on luxuries like vacations or restaurants.] There are a number of criteria we could try to judge Girl Scouts on: 1. Is the administrative overhead growing or decreasing? Computerization and general technological developments means that the necessary bureaucracy ought to be more efficient than, say, in the 1950s. 2. Is the enrollment of Girl Scouts as a fraction of the young female population increasing or decreasing? If Girl Scouts is shrinking, then the good the organization can do is necessarily also shrinking. It's hard for Girl Scouts to help a lot of young girls learn self-discipline or character if they're all fleeing. 3. Are camps being more or less utilized, or worse, being closed? Camping and the outdoors are one of the key aspects of any Scouting organization and in a sense, what makes them more than a club for arts & crafts^[My family is a Scouting family and though I was not a Girl Scout, I went on my fair share of Girl Scout camping trips and did more than my fair share of their arts & crafts. I was left with a higher opinion of the former than the latter, and a sympathy for [the girls](!Wikipedia "Yeaw v. Boy Scouts of America") [who would](http://www.bsa-discrimination.org/html/yeaw-media.html) rather join Boy Scouts.]. If they are going unused, then it'd be like a library no longer lending books - it may still be doing *something* useful, but one wonders what exactly. How can we judge Criterion 1? Easily, it turns out. Girl Scouts is a charity, charities are nonprofit corporations, corporations keep records of spending, and often make them public; as an IRS-recognized charity, Girl Scouts is obligated to make [Form 990](!Wikipedia) financial reports public, and some Googling then turns up [GuideStar](!Wikipedia) as a source for the reports, at which point it's easy to download their PDF and finally answer the question - what does Girl Scouts spend its money on? (That charity filings are publicly available seems to be generally unknown; people routinely ask questions about charities [that could](http://lesswrong.com/lw/1mc/normal_cryonics/1hmp) [be answered](http://lesswrong.com/lw/2l8/existential_risk_and_public_relations/2ywt) by looking at the Form 990s.) Criterion #2 is harder, but total enrollment is public information, as is US census data about how many young girl there are. Criterion #3 is very difficult to research; it seems unlikely that camp attendance or closing data is easily found online, and it is possible that not even Girl Scouts National (GSUSA) knows the answer. We'll punt on #3, and see how far we can get on #1 and #2 without looking at the tax filings. Here are some facts to consider. Girl Scouts was founded in 1912 and is a century old. The GSUSA leadership is apparently unelected. And the [current era](!Wikipedia "Executive pay#Trends in executive compensation") is well-known as one of stratospheric executive compensation, which would be as easy to justify in charity as anywhere else ("shouldn't our managers get a fair living wage? You don't want the Girl Scouts crippled by headhunters poaching it.") Particularly troubling are the cookie sales themselves. At first glance, they seem almost designed to concentrate control & money, rather than decentralize it: - The council probably control the cookie sales through standard methods like copyright or trademark or [trade secrets](!Wikipedia) - prices and availability reportedly are now [set nationally](http://blog.cleveland.com/lifestyles/2008/01/some_girl_scout_cookies_change.html) for each area. This is a very recent development. - Cookies have been official fundraisers and sold since the 1920s; they were commercialized in 1934, and commercial bakers used nationally in 1936 - Competing would be hard because so much of the value is in the brand (and brands are so prized by manufacturers because they are a license to print money). - Economies of scale discourage alternatives; there are 2 official bakers in 2011, compared to [14](http://www.girlscouts.org/program/gs_cookies/cookie_history/1960s.asp) bakers in 1961. - There are few fundraising alternatives: Girl Scouts sometimes sell calendars or magazine subscriptions or nuts or popcorn, but the sales of those are peanuts and most only sell cookies. - Returns of cookies are forbidden # Cookie *costs* How much are the cookies worth? Let's do some [guessing](!Wikipedia "Fermi problem"), then look it up. A Girl Scout troop can run around 10-20 girls. Most of them will sell cookies. A reasonable guess from my own cookie-selling experience is that the average number of boxes sold is ~80.[^sales] Each box is$4, so each girl sells $320, and a troop$3,200-6,400. (If this sounds high, a large council might have 4 or 5 girls who sell over 1000 boxes each, which is >$16,000 in sales right there. The all-time record is at least [17,238 boxes](http://www.foxnews.com/story/0,2933,355758,00.html).) With around 2.4 million girl members, that works out to around 184 million boxes sold annually; in [1992](http://community.seattletimes.nwsource.com/archive/?date=19930513&slug=1701067), 175 million were sold, and in [2007](http://www.girlscouts.org/news/news_releases/2007/2007_cookie_report.pdf), around 200 million were sold - so our guess was off only by ~8%. [^sales]: "Last year, Girl Scouts in the old Lake Erie Council, which served Cuyahoga, Lake and Geauga counties, sold more than 986,000. Each girl peddles an average of 115 to 130 boxes." quoted in [_The Plain Dealer_](!Wikipedia "The Plain Dealer (newspaper)"), ["Some Girl Scout cookies change their names, but the flavor's the same"](http://blog.cleveland.com/lifestyles/2008/01/some_girl_scout_cookies_change.html). 200 million boxes is, of course, \$800 million in sales. Quite a bit. How much of that would be profit? Hard to say. A box of Thin Mints is around 7oz of cookies. (Or less. Aside from the price increases, the cookie count has been reduced, so the previous figures about inflation understate the true price increases.) A consumer can buy 5 pounds of flour for \$4 currently, or 80 cents for a pound, or 40 cents for half a pound (8oz) of flour. Chocolate commodity prices aren't so easy to figure; but 36 1.55 oz Hershey Milk Chocolate costs \$21.30 in 2011 (they have also cost [\$33.62](http://www.amazon.com/Hershey-Milk-Chocolate-ounces-Boxes/dp/B001HXN7YU/) in 2011 or [\$24.15](http://web.archive.org/web/20090304221125/http://www.amazon.com/Hersheys-Milk-Chocolate-1-55-Ounce-Bars/dp/B000IXWCQO) in 2009), which is 2.62 ounces per dollar ($\frac{36 \times 1.55}{21.3}$), or 38 cents an ounce. This is roughly 3 times as much as the bulk cocoa itself, but sugar & milk aren't free[^cocoa]. The chocolate layer on Thin Mints is very thin^[It's reasonable to focus on Thin Mints, as they are the best-seller, making up a [quarter](http://www.girlscouts.org/program/gs_cookies/cookie_faqs.asp#bestselling) of sales.], so 1 ounce may be an overestimate. Let us generously add the 8oz of flour to the 1oz of chocolate; our cost per box is 78 cents. Let's throw in 22 cents more to cover the flavoring and cooking oil and baking process, giving us $1 a box. A cross-check on this$1 manufacturing cost estimate is looking at Walmart's [knockoff](http://authenticorganizations.com/harquail/2009/08/03/wal-mart-knocks-off-the-girl-scouts/) ['Great Value: Fudge Mint Cookies'](http://www.walmart.com/ip/Great-Value-Fudge-Mint-Cookies-10-oz/11997740), which are 10oz for [$2.49](http://fuzziebutter.com/?p=326)^[Walmart refuses to post prices and most reviews take an almost perverse pleasure in omitting the exact price; so I'm going on this blogger's offhand mention.]; 3oz represents a 42% increase, so we scale our Girl Scout cost estimate to$1.42 to compensate, and we see Walmart has only $1 of margin. So our estimate can't be *too* erroneous - obviously it couldn't possibly cost >=$2 to make the cookies because then Walmart would be unable profitably to offer 42% more for $2.49. [^cocoa]: ["Global cocoa prices: 2010-2011 forecast"](http://blog.euromonitor.com/2010/08/global-cocoa-prices-20102011-forecast.html), _[Euromonitor International](!Wikipedia)_: > "Prices for September-delivered cocoa peaked at £2,465 a tonne on 19 July, the highest level for a second-month contract in 32 years."$4270/tonne is $4.27 per kilogram or$1.94 a (Imperial) pound or 12 cents an ounce. And the other $3 are apparently pure profit. How is *that* distributed? The official Girl Scout website [says](http://www.girlscouts.org/program/gs_cookies/cookie_faqs.asp#money_where) > "With every purchase, approximately 70% of the proceeds stays in the local Girl Scout council to provide a portion of the resources needed to support Girl Scouting in that area, including the portion that goes directly to the group selling the cookies. The balance goes directly to the baker to pay for the cookies." So the baker is supposed to get 30% of the$4 a box. The baker gets $1.2 per box? What? That makes no sense - we deliberately used highly conservative assumptions (consumer prices for flour & chocolate rather than wholesale, 2oz of additional material, an arbitrary 22% fudge factor, etc.). And we're supposed to believe that our estimate was *still* low by 20%? Something stinks here. But surely everything is kosher - I mean, GSUSA is specifically not getting a cut. # Revenue streams (other than cookies) But how are they being funded, anyway? Their FAQ [also says](http://www.girlscouts.org/program/gs_cookies/cookie_faqs.asp#revenue_GSUSA): > "Girl Scouts of the USA is paid a royalty for use of the licensed trademarks by its licensed vendors based on gross annual sales volume. Girl Scout councils do not provide any portion of their cookie revenue to Girl Scouts of the USA. No other revenue from cookie sales goes to Girl Scouts of the USA. Girl Scouts of the USA provides contractual services and approves all educational materials developed by the bakers, as well as providing coordination and training for national media, safety standards, leadership programs and sale guidelines." That's sensible. Let's move on. At this point an economist is slapping his head somewhere. Money is [*fungible*](!Wikipedia "Fungibility"). What matters is the whole system, not arbitrary labels and divisions. If you give Kim Jong-Il$50 worth of rice, you just freed up $50 for his nuclear weapons program. If the 2 bakers are paying GSUSA a licensing fee, *that fee is coming from somewhere*. [TANSTAAFL](!Wikipedia "There ain't no such thing as a free lunch"). It's coming from the cost of the cookies! The license fee must be built into the cost structure. If the councils & troops are charged$1.20 for a box and the baker then pays 20 cents to GSUSA, you could change the arrangement to the baker having no license fee, charging $1 a box, and then GSUSA taking 20 cents. Everyone makes the same amount of money in the end, but suddenly no longer are '70% of proceeds staying in the local Girl Scout council'. There may be innocent reasons for these wording choices, but the more cynical know what they think: GSUSA is hiding its rake by blaming it on the 2 scapegoat bakers. Fascinatingly, if the$1 figure is a ceiling, GSUSA could be getting a figure comparable to the individual troops, since the troops only get around 10% or 40 cents; if the true cost of baking is <80 cents and the licensing makes up the other >40 cents, GSUSA would be getting *more*! The other 60% goes to the local council, which apparently covers all the little rewards and incentive programs for the girls.[^bonuses] [^bonuses]: Figures can vary a lot. If you look at the following quote, it seems to imply that $0.47+1.33=1.80$, leaving $1.70. Subtract$1.2 for baking, and they need 50 cents per box to buy advertising, handle the non-existent returns, and offer incentives to the girls? > "Cookies cost $3.50 per box. > > Of this amount, Bruney said, 40-55 cents, for an average of 47 cents, goes directly to the troop. The council receives another$1.33 per box to subsidize events and camps. > > The rest of the proceeds pay the bakers, buy advertising and incentives for the girls, and take care of any debt from people not paying for their cookies, Bruney said." ["Girl Scout cookies take on new shape"](http://web.archive.org/web/20090119214130/http://www.dchieftain.com/news/78448-03-09-08.html), _El Defensor-Chieftain_ Are there any stats breaking things down further? The _Seattle Times_ [mentions](http://community.seattletimes.nwsource.com/archive/?date=19930513&slug=1701067) that > 'Girl Scouts USA doesn't share in the local councils' cookie proceeds, though it does get a 2 percent royalty from the commercial bakers' net sales.' Net sales presumably means the full $800 million, or to put it another way, 2% of each$4, which would be 8 cents, which brings the bakers' $1.20 down to$1.12. That's more reasonable. $16 million is a lot but it wasn't the hundreds of millions we feared. One Girl Scouts leader [comments](http://news.ycombinator.com/item?id=2479833) that the bakers get$0.90 of each $4.00 (22.5%), which is smaller than our conservative estimate of 20 cents more. # Staying honest: predicting *before* the experiment So, before we peek at the tax filings, what predictions would we make? It's important to predict before we look at more data; there are too many [cognitive biases](!Wikipedia) which let you decide that the results [confirm your theory or beliefs](http://lesswrong.com/lw/im/hindsight_devalues_science/) (even if you were [just](http://lesswrong.com/lw/ip/fake_explanations/) [lied to](http://lesswrong.com/lw/if/your_strength_as_a_rationalist/)). If your theory can account for any result, then it's a pretty worthless theory. (If your predictions don't eliminate outcomes as possibilities, then you wind up [holding weird beliefs](http://lesswrong.com/lw/ih/absence_of_evidence_is_evidence_of_absence/) like WWII sabotage confirms the existence of Japanese saboteurs and the *absence* of sabotage *also* confirms the existence of said saboteurs, or that [confessing to witchcraft](http://lesswrong.com/lw/ii/conservation_of_expected_evidence/) confirmed you as a witch and *not* confessing also confirmed you. Which is obviously daft.) ## Predictions Well, one would expect a high salary for the GS CEO (>$300k seems like a sum I would start to wonder about), one would expect substantial benefits^[$50,000 of health benefits is worth more than$50,000 to any self-serving employee, because it is taxed lightly, may keep her salary from being taxed in a higher bracket, and is harder for the public to object to.], a large number of employees in general (empire building); a major red flag would be any employees [shifting employment](!Wikipedia "Revolving door (politics)") between Girl Scouts and the 2 licensed bakers (and any [shared directors](!Wikipedia "Interlocking directorate") basically end our inquiry right there). This essay for me has been something of an intellectual exercise. Without knowing the facts of the matter, going on general knowledge and calculations, I would try to make a case for corruption in GSUSA. The case so far may or may not be convincing depending on how much trust one puts in the 2% figure. The IRS tax filings would reveal important data like employee salaries, which trump anything I've said so far. Will my suspicions be vindicated? Or will I be revealed as a cynic or paranoiac, and the _Seattle Times_ and others right in pointing to the councils as hogging all the money? ## The data The standard source for nonprofit filings is [GuideStar](!Wikipedia)'s entry on the [national organization GSUSA](http://www2.guidestar.org/organizations/13-1624016/girl-scouts-united-states-america.aspx) and more specifically, its [Form 990](!Wikipedia). ### 2009 When I first wrote this, the [2009](http://www.guidestar.org/FinDocuments//2009/131/624/2009-131624016-06105256-9.pdf) was the latest filing available, so we'll start with it. What interesting things do I note on the first read? - the headquarters are on [Fifth Avenue](!Wikipedia) in NYC - what is a frugal charity doing on a street whose Wikipedia entry describes how it is the most expensive street in the world? - the 'governing body' has just 30 voting members - so it actually is elective (in a very narrow sense) - there are *539* employees - it received $6m in donations & grants ($1.5m government) and paid out $4.1m in similar (all to various local councils, it seems) - it received$37.7m in 'program service revenue' The later section (page 10) is very confusing to me. I'm not sure how to read it. It may imply that $32m came from membership dues and the other$5.1m from 'meeting and learning events'. - investments are an odd item; in 2008 they earned $2.15m, but are recorded are -$3.9m in 2009! Did they get burned badly in the housing bubble? More interestingly, the detailed breakout section includes an item for 'royalties' which must be the cookie royalties! The royalties are $8,122,461, much less than the$16m figure. They also earned $2.8m from 'dividends, interest, other similar amounts'. There's one set of items I have no idea to interpret. 'Gross amount from sales of assets other than inventory' lists '(i) Securities: 53,116,009' and then 'Less cost or other basis and sales expenses' lists '(i) Securities: 59,853,735' for a 'Gain or (loss)' of -6,737,726. Another curious entry is 'Gross sales of inventory, less returns and allowances' ($37.9m), 'Less cost of goods sold' ($16.6m) for a 'Net income or (loss) from sales of inventory' of$21.2m. It's possible that this is some sort of cookie reselling scheme, but more likely it's just GSUSA selling Girl Scout merchandise like uniforms at a healthy 57% markup. - the 'Salaries, other compensation, employee benefits' expense item constitutes $37.3m; and GSUSA spent$34.6m the previous year Here is another area where the details (page 11) begin to trip me up. 'Compensation of current officers, directors, trustees, and key employees' amounts to only $2.4m! 'Other salaries and wages' is a total of$24m. The remaining $9m seems to go to$2.7m in pension contributions, $5.7m in 'Other employee benefits', and$2.2m in payroll taxes. - All 'Other expenses' make up $44.5m Some of them make one wonder.$489k in investment management fees seems somewhat reasonable if $14m is being held in 'savings and temporary cash investments',$93m in 'publicly traded securities' & $22m in 'other securities', as does$913k in legal fees; but what did they bury in the $11.8m laconically listed as 'Other'? Did they really have to spend$6.1m on 'occupancy' and $3.7m on 'travel'. It is surprising how little they pay in insurance ($305k) but then again, the insurance often cited as one of the necessary overhead expenditures probably is more a problem for the councils which actually own the camps & facilities. - the form, in case you missed it, spells out the implication: GSUSA was $17m in the red in 2009, and$7m in the red in 2008. How on earth does that work? Where is the money coming from? Am I misunderstanding something? Did they really choose to spend an extra $3m on salary & benefits even as they fell off a financial cliff? - the summary is particularly stark: Total liabilities 2008,$19.35m; 2009, $40.8m. Its net assets in 2009 amount to$118.9m. If it continues to deficit spend at $20m a year, then it'd be down to$100m in 2010, and $80m in 2011, and bankrupt around 2015 or so. I had expected to find a large salary/benefits item like$38m, but I had not expected to find GSUSA in such parlous state. - the checklist does have one interesting item, #28, which asks about former employees etc. having a business relationship; this was all marked no, so if there's a revolving door, it's using some loophole or dodge, which makes it less likely. Phew! One bit of good news. - part 7, section A covers the juicy salary information - but omits the details! All it reports is that 70 individuals are included, who collectively earned $3,793,979 in salary & benefits from GSUSA 'and related organizations', and at least one earned >$150,000. Fortunately, on page 17 (part VII, section Aaa), we find that the CEO gets $435,352 and$63,058, for a total of $498,410. The VP makes$431,758, the chief of staff $310,630, the executive vice president makes$311,618, the senior vice president makes $248,979, the 'VP & general manager' makes$297,407, and a passel of 7 other vice presidents of various sorts make similar sums in the 300-190k range. On page 42, we read of still more payments. The CEO has 82k in 'other compensation', 48k in 'deferred compensation', and 16k in 'nontaxable benefits' (healthcare?). These may or may not subsume her $63k from 'other organizations' previously mentioned, but if we add it all together ($352649 + 82703 + 47553 + 15505 + 63058$), she got$561,468. Nor is she unique - the other executives are all getting their other/deferred/nontaxables as well. - the 'Independent Contractors' section is more detailed. 2 'Technical Services' firms earned $2.3m; 1 legal firm earned$0.9m; and 2 marketing firms earned $1.2m. This doesn't seem so bad, but a final item mentions that there are 45 other independent contractors each paid >$100,000! (If they averaged $250k each, then the 45 were paid$11.25m, for total payments to independent contractors of ~$15.6m.) - One random item I was struck by was$192k spent on direct lobbying (of a legislative body), and a total of $866,500 in direct lobbying 2005-2009. One doesn't expect the Girl Scouts to have much lobbying to do. - One suggested cost for cookie money is scholarships, but the endowment section ($112m) mentions spending $1.2m on 'grants or scholarships', so I doubt any cookie money is going directly there. It's par for the indebted course that the GSUSA endowment declined to$107m. ### 2010 The [2010 Form 990](http://www.guidestar.org/FinDocuments//2010/131/624/2010-131624016-07162b63-9.pdf) is likewise available from Guidestar. - The first page is as striking as before: - voting members is down 2 to 28 - employees is down 21 to 518 - but compensation is still up, from $37.3m to$40.1m - Total revenue: $80.7m. Total expenses: \$85.5 - $4.8m in the red is much better than 2009's$17m - Revenue seems mostly similar (roughly same amount of grants & dues) - but what on earth is GSUSA doing to earn $85k for 'software maintenance'? - Expenses is pretty similar - 'compensation of current officers, directors, trustees etc.' has fallen from$2.4m to $1.9m, despite the previously noted increase in employee compensation - 'other salaries and wages' is$23.5m - less than 2009? - the 2009 report listed just $1.8m for 'Other', but the 2010 lists$12m! What on earth? - Travel fell $0.8m, so perhaps they have been cutting back - The 'total net assets or fund balances' by the end of the year totaled$112m, down from $118.9m in 2009; this is more than I would have guessed from a$4m deficit. - 2010 lobbying was substantially higher - $307k vs$192k - pg 21 covers the GSUSA endowment with convenient current/prior year comparisons. - No prizes for guessing that the balance, contributions, & grants all fell - however, investment income soared to a gain of $9.5m from a loss of$1.4m - 'other expenditures for facilities and programs' increased by a solid $1.1m - Salary information is very interesting (page 39): - the CEO compensation for Cloninger now totals$621k, up from $500k - a remarkable$121k salary increase - next is Corsello (CFO/senior VP), who now makes $500k - a jump of ~$70k ## Assessment of data & predictions So, what's the upshot here? I'm far from an expert on these sorts of financial matters - an interested layman, no more. To me, the picture the filing paints is of a bureaucratic self-serving organization paying ever more in salary, benefits, and pension to a gaggle of executives (half a million or more to its CEO) even as its revenues & endowment shrink and it sinks deeper in debt. But it apparently is *not* doing this on the back of the cookie prices, which seem to be relatively unimportant; the $8m in royalties is less than$\frac{1}{3}$of their membership fees and less than$\frac{1}{2}\$ of their merchandise profit. The price increases outpacing inflation *may* be due to a desperate or greedy GSUSA, but they seem to be running at a loss only in the last few years, long after much of the price increase, and increases could only net them a little more revenue. Honesty compels me to say that the culprits are almost certainly the local councils, who reap something like 60% of the revenue and can recapture the other 10% from the troops through the methods described in the previously linked articles. On the other hand, GSUSA does seem to be in trouble. So perhaps I was right after all? If I was, it is an equivocal sort of right, and one I can't really claim as a victory. Introspecting, I think I may have fallen prey to a sort of [conjunction fallacy](!Wikipedia) which one might call a 'storytelling bias' - I had a nifty story about a reliable gusher of cookie revenue corrupting a self-selected elite, and I ignored that the elite could become sclerotic and self-serving by many mechanisms, not just the one that captivated my mind at the moment. # See also - [Charity is not about helping]() -(examining Folding@home with an uncharitable eye) # External links Looking at financial filings for charities can be interesting from the points of view of sustainability and efficiency. Here are some links where people attempt lay analysis similar to the foregoing: - [ALCOR](!Wikipedia): [Mike Darwin](!Wikipedia), ["The Armories of the Latter Day Laputas, Part 5"](http://chronopause.com/index.php/2011/07/11/the-armories-of-the-latter-day-laputas-part-5/) - [JSTOR](!Wikipedia) 1. [myself](http://lists.wikimedia.org/pipermail/wikien-l/2011-July/109234.html) 2. [Andrew Gray](http://www.generalist.org.uk/blog/2011/jstor-where-does-your-money-go/) - [SIAI](!Wikipedia): ["SIAI - An Examination"](http://lesswrong.com/lw/5il/siai_an_examination/), Brandon Reinhart - [Edge Foundation, Inc](!Wikipedia): myself, ["Case Study: Reading Edge's financial filings"](http://lesswrong.com/r/discussion/lw/7gy/case_study_reading_edges_financial_filings/) - [Lifeboat Foundation](http://lifeboat.com): [myself](http://lesswrong.com/lw/7gy/case_study_reading_edges_financial_filings/4sbg) (short)