BESTOGA stocks

During the course of a different discussion, the conversation strayed to BESTOGA (beer-spirits-tobacco-gaming) stocks and a model George created based on it

http://imarketsignals.com/2015/trading-the-beer-spirits-tobacco-gambling-stocks-of-the-sp500-with-the-im-bestoga-3-system/

It’s an interesting one based on the testing.

I see a learning opportunity here for the p123 community as a whole. Ill pose the question: WHY do George’s tests look as they do? (Backtest/sim results have no mean ing unless one can take a position on why the results were what they were.)

Don’t worry about his ranking system. We can do plenty just with the screen. Set the universe to SP500 and create one Screening rule:

GICS(25301010, 30201010, 30201020, 30203010)

Run a MAX backtest. It looks really good. Why?

Potential answers:

  1. The nature of the BESTOGA group. What characteristics does it have that give you reason to assume or not assume performance will be good out of sample?
  2. The decision to draw only upon the BESTOGA stocks in the S&P 500. Is it a matter of liquidity alone, or are there other substantive aspects of S&P 500 membership that influence your thoughts?

I’m not so much looking for answers as the kinds of thoughts/tweaks you might make as you get to the answers. My goal here is to get a discussion going (in the forum or within the minds of individual users) that focuses on financial, rather than statistical, approaches to strategy design. Think of it as a case exercise that’s ancillary to the on-line strategy design class.

Marc,
I think BESTOGA stocks will always work because they cater to a fundamental law in human (and animal) psychology: Quick rewards are usually preferred to long term rewards. Being addicted (tobacco, gambling, booze, sex) caters to our short term reward systems. It feels good (short term) to engage in these activities. It feels so good that after you tried it a few times, you don’t care about the long term effects, you just care about that “great” feeling short term and the flush of dopamine in your brain, which is the “reward” neurotransmitter. We are prone to be dopamine junkies.

Since this industry is trying everything to feed this vicious cycle, getting people addicted, it will always work. There are hundreds of millions out there, who are dependent on it. Therefore the tobacco industry and others will always strive.

Just using your screen GICS(25301010, 30201010, 30201020, 30203010) + Max No. Stocks + Max backtest + 1 year rebal + No Ranking …

We get to see the years it really outperformed the S&P bench (20%+ Excess) …2000-2003, 2008. So, it performs well in panic years when people are fleeing to safety. These are well known defensive sectors, so this isn’t ground breaking stuff. Not only are they consumer staples, people will give up a lot of things during hard times before they give up their booze and cigarettes. You can find many pictures of wartime soldiers with a pack of cigarettes strapped on their helmets and a random bottle in their hand and conclude the more stressful times are the more people tend to consume them. And maybe they’re a consumer staple in bad years and a consumer discretionary in good years. Cheap whiskey to forget the bad times and celebratory champagne celebrate the good times.

I think one nice thing about using this sector screen with the SP500 Universe is it seems to perform well once the markets turn more bullish as well. Good results in years 2010-2014. It’s not like people pour out of them when the coast is clear like some other hedges. Even if some do bail, it’s probably more than made up by the SP index fund wave when people start pouring money back into the market. When I use the an All Fundamental Universe with some basic liquidity and price screens, the excess results per annual year seem much more random and sporadic and doesn’t really provide a decipherable story to me.

Another benefit is these sectors don’t seem to require frequent rebalancing to draw nice returns. Even just a yearly rebalance seems to perform quite well. These could lead to some nice tax-advantaged models.

As expected, junk food is equally addictive as the products of the Bestoga group and returns are very good too.
Put this into screener: GICS(25301040)
Gives a 15.4% annualized return from 1999.

Geov - you are assuming that restaurants translate into junk food. I’d prefer to see fast food restaurants only but is not offered as an industry.
Steve

Hi Marc,

I would say that the choice of Bestoga is primarily because of point 1. These are industries with high returns and they are recession-proof (hence low drawdowns). Adult entertainment and dentistry can probably compete equally well with them, but are usually not incorporated entities.

What I am less sure is if Georg’s Bestoga model can hold up 30% OOS. It is a small universe, and too easily one can fall into the curve fitting trap.

Br, Florian

I was surprised to see that in the SP500 there are, on average, only 7-10 companies that are in these GICS. I thought there were more.
Anyway, my thought as to why it has a nice steadily growing equity curve:

  1. Reliable and consistently growing revenues.
  2. Steady and predictable operating margins
  3. Steady EPS growth for all companies in this (small) universe.
  4. Good debt management and ROx management
  5. Universe EPS growth that is higher than the SP500 EPS growth (creating excess returns over SP500)

Paradoxically, the addictive/amoral nature of these products creates a moat by holding customers to them as well as scaring off potential competitors from investing in competitive products (because of litigation concerns or just the morality of it). So these companies have some pricing power that keeps their margins up (beer, spirits and cigs have gotten steadily more expensive, even taking the tax increases into consideration). Part of their moat is their distribution network, that tends to be very old and established, giving them access to customers worldwide. Sin is universal and emerging market countries have huge and growing populations (think China and smoking). Helps grow revenue.
There will always be a market for sin. Makes us feel good (at least momentarily) and it is affordable by many. But their excess returns over the SP500 as a whole may not always be there. The Chinese may decide to stop smoking as much. So I would treat these companies as all others and not just invest in them because they have shown alpha in the past. Things change…

I think that in the future, use of hydrocarbons for energy production will fall into this same bucket. Many think it is bad, it is being hit by litigation (global warming) but it is cheap and effective at what it does. So Exxon, Total and Sinopec will always be around and profitable, no matter what. The energy density per dollar for oil is still the best (at least without a huge carbon tax).

I think there is some inefficiencies in pricing because some people and funds will not even look at these industries because they find them personally offensive.

Good point Dan.

Steve, currently there are only 5 restaurant stocks in the S&P500: CMG, DRI, MCD, SBUX, and YUM. Which one offers the best food?

HAHA I’m surprised there aren’t any higher quality restaurants in the list. I would probably exclude DRI, then would have to go back to 2000…2004…2008…2012 etc and see what restaurants were called out in the group. Plus you don’t know what the future holds so you will have to constantly have an eye on the restaurant group.

Steve

Georg, none of them would qualify to be called a “restaurant” in Italy :slight_smile:

Bingo! All of these factors are important. (By the way, this is how Warren Buffett got to become Warren Buffett.) The p123 server can’t process ideas like the addictiveness of sin, etc. To do anything with these ideas in p123, we need to do some language translation, and that list of five makes for a pretty good grammar reference.

So now, we have some interesting choices to consider:

Is Bestoga status, by itself, sufficient reason to allow us to infer that these requirements will be satisfied? (If we determine the answer is yes, and were creating a public model, something that must be sold to outsiders, we are allowed to factor the marketing aspects of a BESTOGA brand into our answer; all’s fair in love and commerce).

Are their Besoga companies that don’t fit the mold (i.e. might we enhance results by defining some rules that eliminate them)?

Are there non Bestoga companies that are at least as good (or maybe better)?

We still haven’t talked about the limitation to sp500 stocks. Anybody care to comment on that?

One can add the Restaurants to the Vice Stocks of the S&P500. The Bestoga +Restaurants = BESTOGAR.
GICS(25301010,30201010,30201020,30203010, 25301040)
The backtest return is not much less than for the Bestoga stocks only.

Now we have 12 stocks in the universe instead of only 7. All companies are in the S&P500. I think this is important because they are operating internationally and cater to consumer needs which are not much affected by economic conditions. So they have a big moat around them.

If a bestoga stock is one that depends upon an addiction to drive sales, then I would add the GICS which contains Coke (KO&COKE), Pepsi (PEP) and the like. They are not alcoholic but people are hooked on them. I would also add the candy manufacturers like Hersheys (HSY) (I know there are much, much better international companies for these kinds of items too- especially chocolate). I have been looking around my own house for suggestions and unfortunately, I came up with those as well :slight_smile:

The nice things about SP500 is that they have more stable earnings and have more financial resources to weather litigation. Plus they are able to have an international sales network. More reach.

Maybe include ADRs of SP500-like companies in Developed (not Emerging) countries who fit these GICS we have talked about. I think Developed countries use similar accounting standards. I have not had much luck when adding certain ADRs to my own custom universes in the past but it might work here.

Good idea David. Those are classified SODAPOP , GICS(30201030). There are only 5 stocks in the S&P500, and screen shows AR= 8.35%.

My gut feeling (no pun intended) is that Coke’s best days are behind them but they are very resilient and Buffet likes them…

So what have we learned from this discussion?
It appears to be advantageous to select one’s stocks from a universe that is not affected significantly by changes in the economy. The BESTOGAR stocks seem to meet that criterium. One will get better performance from models concentrating on the basics, i.e. the universe, all other factors seem to be less important.

Here is performance of a 6-stock model using the universe BESTOGAR. There is only one buy rule, a market-timing rule which ensures that stocks are only bought during up-market periods, otherwise the algorithm is identical to the one that Steve posted earlier.


Yes, financially, they tend to have more of a safety cushion – to a point.

Many small companies have much stronger balance sheets than many realize and actually, small size contributes to that. The smaller the company, the more concern the CFO has with such things as meeting payroll, paying utility bills, paying suppliers etc. leaving less time and energy for “balance-sheet management.” The large-company edge tends to come about in that they have more levers they can pull in the capital markets if push comes to shove. Also, larger companies are more likely to benefit from internal diversification (even within what appears to be a singular GICS classification, it’s possible for a company to have many different kinds of businesses with different profit characteristics. The only way to see that is through the 10-k business descriptions and trade sources.

So regardless of what how historical bets may tabulate over the course of any specifically chosen sample period, looking forward, when all probabilities are on the table, big companies have a leg-up in terms of risk mitigation. That can always be and sometimes is cancelled by company-specific matters, but pending further information, the default assumption is that bigger means less risky.

Likewise, although company-specific circumstances can always change things, the default assumption in that bigger companies have more stable profit streams. This is a function of operating leverage. The bigger you are, the more thoroughly fixed costs are covered. Conversely, the smaller the company, the larger fixed costs loom as a percent of total expense, thus equating to a presumptively more volatile profit stream, and that suggests a default more-volatile-stock assumption. Once again, this, not historical price-based data, is what future-oriented betas, standard deviation, etc. is likely to be about.

Continuing, we saw in the Value portions of the on-line seminar that higher quality equates, all else being equal, to higher ideal (warranted) PE, PS, PB, PCF, etc. This has implications for out-of-sample-oriented modeling.

First, we can choose to pay up for stability (i.e. higher probability that historical trends will persist). Many do just that.

Second, we can look for information arbitrage opportunities. Big-cap stocks are often seen as dull and offering little opportunity while smaller and super smaller stocks are favored for their high volatility high risk characteristics (i.e. the larger probability of positive outliers). We see that in the forums all the time, and we certainly see it all over the place in Smart Alpha, where designers and subscribers are hypnotized by the stupendous simulations that can be crafted among tiny stocks. And we’re not the only ones who think this way. But that can present opportunities. If big cap stocks are less appreciated than they should be, on objective grounds, well, you know . . .

This is why the easiest way to dampen the BESTOGA sims is to aim at a broader and/or smaller universe than SP500.

And by the way, SP500 and R1000 are not always equivalent. SP500, by virtue of its “cultural” status often attracts a lot of non-fundamental park-it-naively money. This, too, plays a role in volatility.

This doesn’t mean you can’t win with a small/micro BESTOGA strategy. There are some great opportunities down there. But because you’re stripping away the benefits of SP500 membership, you’ll have to work harder in screening/ranking to uncover them.

What’s the risk to an SP500 BESTOGA strategy (aside from the obvious market risks)? One ever-present one is valuation, that too many investors identify and pursue the opportunity and bid the stocks up to levels that bring volatility onto the table. I think it’s a low probability scenario since people are in love with swinging for the fences chasing bigger upside. Probably the one that bears more watching is at the company level, i.e., whether antsy action-craving managements go nuts with unwise investments and/or financing wizardry, in other words, event risk. If you expand BESTOGA to restaurants, this may become more a concern and the presence of gaming certainly puts it on the table. Complacency is also something to watch for as brands like Budweiser and Coca Cola find themselves having to fight harder than ever for shelf space and consumer affection.

Personally, I like the BESTOGA approach and think the model has a good chance of being successful. But you no matter how much anyone loves any strategy, one should always have a doom and gloom scenario in mind, since complacency is our biggest enemy. We can’t be intimidated by baggage; 100% of companies have it. But we do have to be aware of it.

Anyway, this is a sample of how I look at strategies.

Geov - one point I would like to make is that these stocks are supposed to be resilient to economic downturn. So why use market timing? While your MT rule may have worked on a backtest, it doesn’t mean it will work going forward. You may miss great opportunities being out of the market, or get whipsawed. It will also add psychological stress as one questions whether the MT rule is working in forward time.

Steve