Summary: Nasdaq computers store stock prices as a 32-bit number representing the number of 1/100'ths of a penny, which means the highest dollar amount it can store is $429,496.7296 (2 ^ 32 / 10000).
It probably would be for quite some time. The article lists the next highest stock is $5,100. This is all due to stubbornness on the part of Warren Buffet, for better or worse.
Well, this isn’t really that big of a problem. Just let the price overflow. That should make it easier for them to just split the stock next time they get there ;)
With all the weird structures of stock lending and synthetics going around, I believe it's been extremely valuable for Buffet to keep it as it is for all these years.
There are no exchange traded options on BRK.A, lending shares to be shorted gives you interest on the total $ value, not the number of shares. Neither of these are the reason for not splitting.
Creating a second class of share is a very very bad solution: It splits exchange liquidity. It also complicates all the investor relations correspondence, adding lawyer fees.
A huge fraction of publicly traded companies have preferred stock in addition to their standard ticker... They just have different terms (eg pricing vs underlying, dividends, etc).
Because the owner of the other class of share believes themselves to own part of the company, but they don't. It's confusing. What you really own is some sort of bond from the company that guarantees they will pay you an amount equivalent to a dividend or something, not a share of the company.
> Because the owner of the other class of share believes themselves to own part of the company, but they don't.
Shares, whether in one class or many, are packages of claims against the company, not really ownership of the company in the simple sense. That’s fairly fundamental to the corporate form as distinct from, say, partnerships.
This should be easy for Nasdaq to solve: just force companies to split their shares before they reach certain limits. Markets can create rules that listed companies need to follow, and this seems to be a useful one.
Berkshire Hathaway is NYSE listed, so Nasdaq do not set those rules for them. If a primary market created rules to mitigate systems issues, they would be harshly judged for it, and likely lose business to competitors.
Exchanges can create rules, but there's a limit to how far they can push it:
If Nasdaq makes too many rules that investors or companies dislike, they'll list respectively trade elsewhere.
Developments in the rest of the market make stock splitting less of an imperative than in the past. Eg Robinhood and many other retail brokerages allow you to own partial shares now.
If you follow the logic of these recent developments, individual shares might become a thing of the past, and we will just directly trade fractions of whole companies.
(On your computer, rational numbers are easy to implement: just represent them as a pair of arbitrary-length integers.
Operations will be a lot slower than on floats or ints, but the exchange doesn't need to do too many operations per second. And market makers and hedge funds etc can use whatever approximation scheme they want in their internal systems.)
The same way you can own part of a house, part of a business, or even part of a bottle of wine. Contracts can easily dictate who owns how much of a thing, they wouldn't have to literally rip a paper stock in two.
I'm not convinced. Can you go to the Berkshire shareholder meeting and vote for your 0.0001 shares?
I find it much more likely that it's actually your broker who truly owns the share and you have a contract with them. This is not at all like a house which can actually have two true owners.
That doesn't mean there's anything wrong with such an arrangement, of course.
> Cede technically owns substantially all of the publicly issued stock in the United States.[2] Thus, investors do not themselves hold direct property rights in stock, but rather have contractual rights that are part of a chain of contractual rights involving Cede.[3]
Modifying the last link in that chain isn't such a big deal.
> I still wonder what might happen if the broker went bust and owed X different people various fractions of a single Berkshire share.
Legal liabilities ultimately are generally settled in dollars, though if an entity “goes bust” they tend to be written off or resolved at a small fraction of their theoretical worth based on the liquidation value of the remaining assets.
If you "own" one share, or 100, there are a couple of layers of other people who "truly own" the share, and you have a contract with them. You can look up Cede & Co and the DTCC.
Your arrangement with your broker for fractional shares is similar, though not identical. They're not just selling you a derivative product that they may optionally hedge with BRK/A.
> I find it much more likely that it's actually your broker who truly owns the share and you have a contract with them. This is not at all like a house which can actually have two true owners.
Well, a share of stock can have two true owners, it's just that public stocks don't.
The legal ownership structure isn't hard to find out; it is actually mandated by law. All publicly traded stocks are owned by the company set up for that purpose. Doesn't matter whether you're the contractual owner of a tenth of a share or the "sole owner" of a full share; the share is owned by DTCC.
>Can you go to the Berkshire shareholder meeting and vote for your 0.0001 shares
You could form a contract with your co-share owners describing how you allocate that benefit, just like would be necessary with a house with co-owners.
I have no idea if that happens, but it's certainly just as possible.
A share is already just part of the company, a fraction of the company. Or more specifically, a fraction of the outstanding shares.
Suppose you buy one share of a company that has a thousand shares outstanding. You now own one thousandth of the company, or 1/1000 of the company.
Suppose instead you buy one thousandth of a share. Now you own a thousandth of a thousandth, 1/1000/1000. Or you can call it a millionth of the company, 1/1000000. It's the same number either way you write it.
The point is that a share is already a fraction, and it's no great leap to have a smaller fraction than that.
Ah, thank you for the clarification, that is a nuance I completely missed.
A follow-up question since you know more about this than I do: if some number of people each buy a fraction of a share, who gets to vote that share, nobody? Does it become effectively a non-voting share?
A fractional share is a new financial instrument operated by the seller of the fractional share.
If you take the example of Robinhood, here's the relevant section[1].
> "I understand that fractional shares within My Account (i) are unrecognized, unmarketable, and illiquid outside the Robinhood platform".
Robinhood remain the owner of the share and retain the voting rights. Robinhood's terms describe how they distribute dividends and voting "rights", but this is a commercial agreement between Robinhood and the fractional share buyers, and doesn't invoke the legislation surrounding share ownership.
When it comes to fractional shares as a consumer-facing product, I imagine it would be effectively a non-voting share, yes.
However the vote could be exercised by whatever entity actually owns the whole share. That would be entirely a decision for the people who chose to divide up the resource. Presumably BH will only accept a single binary vote per share, so it would be up to the owners of the fractional share to agree (or not).
I used to think the same. Recently, I was surprised to learn that many US retail brokerage firms now offer this service.
If you Google (or DuckDuck) "fractional share ownership", you will find a bunch of hits with a list of brokerages and an explanation about how it works.
I assume these products are "captive audience" -- you must buy or sell your fractions with the same broker. Thus, they cannot be transferred to other brokers. That said, I also assume it a very price competitive product so the margins would be thin and prices fair.
More like a deliberate plan to concentrate voting control.
You can perfectly well buy Berkshire Hathaway without plonking down a half million dollars--there are class B shares. It's just class B shares don't have anything like the voting power that class A shares have. Class A shares can be divided up into class B shares but you can't combine class B shares to make a class A share. The result is the owners of the class A shares have voting power far beyond their percentage of the company.
Warren Buffet has no need for a market in class A shares to even exist. The common man should buy class B.
To be precise Lindt & Sprüngli (LISN) is currently trading for CHF 90’900 roughly USD 99’485. And as a fun fact they have quite a famous natural dividend a suitcase full of chocolate :)
I found a couple by going through the OTC daily lists and looking at dividend payers. I’d double take seeing somecompany paying a $500 or $1,000 dividend, look it up and they would have some crazy stock price.
But note: they’d also only trade once a month or so.
First, as other posters have mentioned, there are b shares that trade in the hundreds of dollars - BRK.B. Yes, their voting rights are disproportionately lower, but you can still get a stake in the ownership without having $400,000 in your bank account.
I own B shares. It doesn’t bother me that I have fewer voting rights because I trust my fellow Berkshire owners. So there’s little barrier to entry in merely owning a small stake in Berkshire’s economic output.
Second, if you want to argue there’s a barrier to entry for voting purposes, that also doesn’t make sense: you need to have a ton of ownership in any stock to make a difference in ownership as an outside investor. It doesn’t matter if the share price is $1 million or $10 if I need $50 million in share value to make a dent in voting.
Third, Buffett has stated that he doesn’t want to split shares because he wants to encourage long-term owners. There’s a lack of liquidity in A shares; I’ve heard usually only about 1,200 trade a day. He, and I’m sure many other Berkshire shareholders, want their fellow shareholders to think like long-term owners of a private business, especially for shareholders with a lot of voting influence. A high share price and its corresponding low liquidity encourages that.
It's not a matter of trusting one's fellow Berkshire owners. It's a matter of keeping power in the hands of those he trusts to follow his strategy. I trust his running of the company far more than I trust my fellow shareholders (I also own some BRK.B.)
He created the BRK.B share class years ago and has split that once to keep the price manageable. It stands at 1/1500 of a class A share today. The B shares do have reduced voting rights relative to the A shares on a proportional basis.
Lol. Some brokers allow you to vote on fractional shares, and technically you don't own the share, but why would they care if you do or not, it's about the voting rights.
Generally, yes. Most companies will go through splits long before that, turning every share outstanding into two or more shares (usually with half the value) on a predetermined date. Having a very high stock price is actually kind of inconvenient, as it makes it hard for random day traders to afford.
To be clear: "random day traders" are retail with small balance. Institutional players absolutely do not care about stock price. It is notional (or delta) that matters.
Samseong in Korea also had (has?) a huge stock price that used to cause issues.
There's nothing strange about it and you're referencing things that aren't exactly relevant to the issue. The SEC has rules about lit market quotes which exist to give integrity to the price-time priority. No exchange is allowed to publicly quote fractional pennies so as to give someone a way of "jumping" the queue so to speak, but there are plenty of ways to transact at fractional pennies so long as one isn't quoting it on the lit market.
For example one can submit a mid-point peg order which can be filled at a fractional penny:
There are also auctions that take place on a daily basis, such as the opening auction and closing auction, where trades may execute at a fractional price.
Furthermore NASDAQ also operates as a reporting facility, and as such reports trades made through ADFs such as FINRA, dark pools, off-exchange block trades, and inside quotes which are quite often executed as fractional pennies.
Anyways, all this to say that the situation is a lot more detailed that you make it seem from your two references. It's not just a simple matter of SEC says orders have to be to the penny, so NASDAQ must be doing something really bizarre here.
It looks like the smallest increment before decimalization was 1/16 of a dollar, or $0.0625, which could be represented exactly in 4 decimal places. So, by storing the value of the stock to this precision, both old and new prices could be stored in this format.
In that case they can store quarters of a cent instead of hundredths of a cent, and still be able to store $0.0625, while increasing their range by a factor of 25.
I'm trying to remember how we used to store it -- decimalization was 20 years ago this year!
I think we had a discriminant for the denominator, so it wasn't a rational type, but we could also easily represent both fractional and decimal prices. We could handle fractions smaller than 1/16, though I don't know if any equities traded at such prices. What you suggest would also be reasonable, but would have had to be rewritten when decimalization happened.
Frequently if the bid is say 40.10 and the ask is 40.09, the mark is shown as 40.095. Not sure if that's why they have hundredths, or just to support those legacy cases of odd fractions.
Orders are generally quoted in cents, but trades execute off-book at fractional cent prices on a regular basis (e.g. when a HFT firm is providing price improvement).
At a previous job (over a decade ago), I worked on software that stored prices in centipennies. I'm pretty sure it's no longer in production. At least, I hope it isn't.
But these things sometimes surprise me. A few years ago I paid a visit to my coworkers at the job I had in high school in the mid nineties. I had written some sort of email processing program in C; I don't even remember what it did any more. I was told they are still using it, 25 years later. Awesome or scary, I don't know which.
Honestly, I always found precedence rules horribly ambiguous. I just use brackets for clarity and forget about it. Or is this considered 'bad practice'?
I've never found it to be an issue with math. Unary minus > powers > multiplication/division > addition/subtraction.
I have seen differences in the precedence of logical operators between languages (and > or but what happens when you have both math and logic in the same statement isn't entirely consistent) and always use parenthesis in those cases--although these days if Resharper says it's redundant I let it remove them.
I don't consider it bad practice to insert parentheses, especially with exponentiation because many people are less familiar with its precedence rules.
I don't understand why this was down voted. It is an interesting comment!
I always assume the floor for equity is the first tick above zero, e.g., 1 US cent or 1 Euro cent or 1 Japanese yen. Honestly, I don't know if any traditional stock exchanges allow zero price or less for shares. If anyone knows an example, please post about it!
Another reason why this comment matters: Does anyone remember when oil futures went negative? (April 2020) I am sure more than a few computer systems were unprepared for that scenario. I remember when short-end Japanese rates were zero and negative after the 2008 Global Financial Crisis. It was a real monkey wrench in the machine!
There's a reason futures can go negative, that is when the contracts get close to expiry, the holder becomes legally liable to receive the oil. Since the speculators have no way to actually do this (and that storage was at capacity anyway), it makes sense to sell to shift the liability to someone else. There is not such analog for shares that I can see.
You wrote: <<the holder becomes legally liable to receive the oil>> Very good point! We should be clear for other readers: For many commodities futures, there are two types: physical and cash settle. (From my experience, liquidity can vary wildly!) Example: Physical oil futures require you to deliver (short)/receive (long) oil at expiry. Cash (settle) futures only require an exchange of cash at expiry.
Probably downvoted because 2^32-1 is the same as (2^32)-1, it's 2^(32-1) you're confused with. Although downvoting just for that feels a bit pedantic to me.
The the intention of both comments was clear... To discuss signed vs unsigned integer.
They could also be applying an arbitrary offset to support some range of negative value. This whole tangent seems a bit pedantic though. :-) They probably should have used uint64!
Or perhaps just move to arbitrary-length integers. (Eg what Python gives you by default.) Or even rational numbers.
Both are a bit slower than using ints built into the machine, but exchanges don't have to be particularly fast. (In contrast to HFT participants. But they just need to be faster than their competitors.)
My suspicion is that the field is a fixed size for serialization inside a database table or something. Truly variable length fields are trickier to store and then subsequently iterate over non-sequentially.
When you say rational numbers, do you literally mean a fraction, or are you talking about a floating point representation? I'm not sure what a fraction would do over fixed point in this usage, and floating point is best avoided for anything finance related due to the fact that they're approximate.
They probably regret not just using 64-bit from the start, but on the other hand somebody likely saved a measurable amount of money on storage over the years. Potentially not worth the cost to figure out a workaround now, though.
Perhaps they could implement a rule where if the field equals MAX_UINT32, then the value must be looked up in an auxiliary database table. Depending on how much crufty code there is assuming 32-bit though, it may be their equivalent of y2k.
I meant a genuine fraction. So you have arbitrary precision.
Yes, variable length fields are a bit annoying. But my argument is that the exchange itself doesn't need to be that fast. And the market makers just need to be faster than each other.
(For anyone else, exchanges just need to be fast enough.)
Yes, just using a 64 bit integers is probably the easiest here.
Floats would be worse than my suggestion to use variable length rational numbers, for exactly the reasons you know.
> When pressed on the issue, [Warren Buffet] has told shareholders that a lower price would bring unsophisticated short-term investors into the stock.
With respect to Buffet who has talent and is not a total fraud he gets a large fraction of his outsized returned in trades not available to you and me. Off market placements for example. Do you think your returns would have improved if you could buy Goldman Sachs at a 20% discount to market? Yeah me too.
Marketing and promotion is one way he gets those offers you and I don't. (No not the only way, he has billions to invest too, but his name helps). Having the highest share price for a single unit of stock is something else to talk about and color a story if you're reporting it. He's very good at this promotion. His annual letter was a previous generation's version of Elon Musk's tweeting, hiring onion writers, flame throwers and so on that gets Elon so talked about.
It's a skill, a talent. Finding the opportunities and exploiting them in a way that doesn't significantly blow back at you. Buffet executes it brilliantly. Something to consider in your own business. Is there something else you can exploit like that which would make for a paragraph in the story of your company?
You can go to the meeting with BRK.B. Buffett prefers long-term holders, and the feeling was a higher share price makes things a little less liquid. With fractional share ownership it's kind of a wash now.
Even BRK.B would be 50x more its current price but it had to be split when they bought Burlington Northern with stock to accomodate BNSF shareholders.
The major difference is voting power. Economic interest is pro rata, but voting is much different (not that it matters in the current environment). That, and BRK issues direct tender offers to BRK.A holders.
Voting power is in my understanding the real reason. Dollar for dollar a shares have seven times the voting power of a b share. Following his death, the foundations he has bequeathed his wealth to need to sell his shares over the course of ten years. Ordinarily this would result in institutional investors having a majority of the voting power. To avoid this the a shares will be converted to b shares in order to reduce the voting power. A shares are by and large held by investors who have been with Berkshire for decades. Hopefully they will continue the long term focus of Berkshire. Once their progeny start inheriting the shares I think all bets are off. I worry activist investors will convince institutional investors to break Berkshire apart.
Many of the individual businesses within Berkshire are able to act with a long term horizon that I think is uncommon within the stock market as a whole. Profits and float are able to be reinvested within Berkshire in a much more tax efficient way than separate companies.
One example is BNSF. Most railroads are on warpath to increase operating margin as much as possible. Cutting routes, laying off staff, and neglecting customer satisfaction are the norm. Imagine being laid off after your company has had its most profitable year ever. These actions are being dictated by institutional investors. BNSF is the only major railroad not taking on precision railroading. It's possible that this will be a failure on BNSF's part, but Berkshire is willing to take the bet and act differently from everyone else on the idea that this will be better long term. Institutional investors are not interested in that.
I think you get tickets with BRK.B too. A friend of mine sells the tickets he gets on ebay and says that's a bigger dividend than the stock will ever pay :)
Pricing for auto insurance is all fully public information, that states' approve. There is no "discretion". You can look it up yourself. Search for "SERFF" and your state.
Also, auto insurance has the lowest profit margins of any type of insurance.
Unfortunately, in some states, even after reading and cross-referencing dozens of obtuse rules and tables from a given company's rules and rates filing, you may eventually find that the final premium relies on the output of some manner of GLM/tree model that you'll never have access to.
True, you might need to review those rate/rule filings across several states to tease out the more obscure aspects...or any large actuarial firm can do it for you.
All of those things are for sure taken into account, but the key input for pricing is:
- type / year / color of car (gives price and your attitude)
- how many km driven / year (no risk if it stays always on your drive!)
- where is it parked
A small anecdote: I did a project for an insurer once. When we discussed the extreme detail the actuarials wanted for each policy, the client remarked half jokingly that the color + engine size was almost enough for the actuarial model.
(Disclosure: I work for a car insurance company. I am not an actuary and have no special insights into how premiums are calculated.)
Gender famously matters too - a 18/M is (statistically) likely to be a worse driver than a 18/F. Rates reflect this. Having once been a 18/M, I have no trouble believing this.
That said, it's very difficult to compare across states. Car insurance is regulated at the state level, so inevitably different states (or territories, or DC, etc.) have made different choices over time. You can do a dollars-to-dollars comparison between any two states but there's a good chance the elements of the policies and liability rules aren't the same.
I suspect there's also some adjustment to premiums done for how much it costs to repair things where you live. Medical care, building repair, and car repair are all things that likely come up a lot on claims. If you're some place where these are twice national averages, you might have higher premiums than you would in a place where these cost half the national average. In addition to the chance of a loss varying by location.
You should probably include your full credit history as well, since in most US states (i.e. where legally allowed) that has a very large impact on your premium.
You would need coverage amounts and driving histories. Even then, it’s impossible to do comparisons without having all the data for losses in various locations and personal history of claims.
Car insurance is extremely regulated and transparent. If you think you’re getting ripped off, go to a different website and shop around.
That is only relevant if the person buys coverage for the car. Legally, you only have to buy liability insurance, so you can compensate any other parties you injure or damage.
If you borrow money for the car, the lender might require coverage for the car itself. I pay for $500k bodily injury and $100k property damage liability only coverage for $40 per month per vehicle for 10k miles per year.
I had Geico on the east coast at a similar price ($50), but they were more expensive on the west coast so I switched to Amica.
With less than 1 you're not a shareholder, only your broker is. All you have is a contract with your broker about how they split a part of the profits of their share in Berkshire Hathaway.
Not really. A small shareholding is still a shareholding. You can still go. If you read the meeting instructions they ask for a statement showing you have a holding in street name. I've been a couple of times flying from London to Omaha. It was kind of fun.
Also you are technically and legally a shareholder even if your broker holds the shares on your behalf as is usually the case these days, and attending any shareholders meeting is a legal right whatever the size of the shareholding.
Although just now they are online due to covid but I'm sure the physical meetings will return.
If you do hold 0.001 shares with a broker and you want to go you should request with them to be sent the annual report, also a legal right, and that'll have the invite card. Or if not you can phone Berkshire and ask for one.
A small shareholding counts, but that still means owning at least 1 share. Anything less means you don't own the share. Fractional shares are a different construction (compared to the normal case where the broker holds your real non-fractional shares for you) between you and some brokers that offer it, which result in you not really owning the share and thus not going to the shareholder meeting.
Joke's on those that think having 0.9 share at Robinhood is the same as owning a real share. Try voting or going to a shareholder meeting with your fractional share and see how it goes...
"a lower price would bring unsophisticated short-term investors into the stock"
Haven't "sophisticated" investors been conned, suckered, and fleeced as long as stocks existed and even before?
From participation in the South Sea Bubble (the original "bubble") to the Tulip Craze to Bernie Madoff to Enron to cryptocurrencies and beyond.
Not to mention that deep pockets don't necessarily equal sophistication (even when they spend enormous sums to hire professionals to manage their money), and the fatter your pockets the juicier a mark you're going to make for people whose lives are dedicated to funneling money from your pocket to theirs.
I think it's just a matter of terminology/framing -- large sophisticated, moneyed market participants is a "financial crisis" and smaller players getting conned/suckered/fleeced is "enabling gambling". On one hand I don't care for the moralizing and condescension but on the other hand, Berkshire has more business/market knowledge in their trashcans than I have in my head.
That said, Berkshire has always been on the side of the little guy in a sense, even if they're a bit condescending about it -- Warren Buffett has been very vocal about investing in index funds (even putting a tiny bit of his money where his mouth is and challenging fund managers to beat it) this whole time and it's been a good strategy over a very long time frame.
Now, one of the things that no one tells you about the stock market is that a LOT of players make money by front-running and middle-manning the whales (index funds, pension funds, etc), so you'd maybe think it was a trap to just make the whales bigger, but in general the strategy of just investing in broad index funds has been good regardless for individual investors.
Another thing that's been made pretty clear in the previous year is that governments will bail out businesses first in crises. One of the really crazy things that Warren mentioned was just how wide "spreads" (difference between pristine debt aka bonds and debt from struggling companies aka junk bonds) got in march of 2020 -- near/surpassing 2008. He joked that Berkshire couldn't even have floated debt in that market, which is why the Fed doing what it did was necessary. The usual arguments of businesses as "job creators" is maybe valid in some sense, but capitalism without repercussions/negative feedback is what we seem to be trying to drive hard to. Companies that are not prudent with their cash and record profits for the last 10 years of extremely loose fiscal policy should be punished. We expect individuals to save for a rainy day, why don't we expect companies to do so?
I remember it was a discussion topic during my graduate degree in finance. According to theory, the story about unsophisticated investors is stupid since you could basically start an ETF with only BRK stuff in it and sell it at a lower price.
You can skip ETF management fees and just look up their latest SEC filing. Here is a pie I made in M1 (which you can invest with the click of a button if you have an account) which is similar to their public holdings but with tech stocks (Apple and SNOW) removed: https://m1.finance/6rAREGn34Ai2
Fund shareholders typically don’t show up to the shareholder meetings of the fund’s underlying holdings and don’t vote because the fund itself is the shareholder, rather than the fund’s shareholders. (If that makes sense...) So it did in a way keep the unsophisticated investors from exercising too much control but it did bring fund managers into the mix. Which is arguably worse. Hence BRK.B.
Surprising that they are they rushing to fix it at the last minute when Berkshire Hathaway's stock price has been steadily climbing towards that limit for decades. For about half the time that Nasdaq even existed, it's been more than 20% of the way there ($70,000 in 1998). Surely they've touched their software sometime between then and now and had the opportunity to notice the problem.
I remember a friend complaining that he had to work over the weekend when they hit 10,000. That 5th digit screwed up all their column formatting. Technical debit is ever present, unfortunately.
Oh no! I just remembered I have a time-limited trial feature in my own software that won't work past 2022 because when I implemented it, I assumed I'd either rewrite it or be out of business by then. Now I have some more empathy.
1. Banks and other stakeholders subscribe to binary formats of price data. Making this change means forcing everyone to update. And those subscribers are paying customers who will likely be against this change.
2. These are old codebases with likely many uses of 32 bit prices, and if written in C/C++, conversion from int is particularly hard because those languages like to automatically cast to int without a warning (so you can’t rely on compiler errors to tell you what you have left to fix).
3. These are performance critical systems. Adding an extra 32 bits of all zeros (or, all zeros 99% of the time) means you slow everything down for one special case.
I’m not saying it cannot be done, but I can see being very pissed at any company who makes you do tons of work for their marketing trick. (Surely a 200k stock is just as out of reach from the average person?)
Back in the mid 2000s I was working on a stock ticker display for an embedded device. I got into regular arguments with people that couldn't be convinced that a stock price would ever exceed 3 digits.
BRK.A was written in black Sharpie on a post-it above my desk, I would just point to it. I think it was $90,000 back then.
From what I can tell, BRK has not been remarkable for at least 10 years now. Or at least since the Goldman deal where Buffett famously got a discount. In fact, I’m pretty sure Buffett lost out on a ton of growth because he was anti tech for a while, since he liked to claim he did not invest in businesses he did not understand.
I doubt the insurance companies are driving BRK’s price growth anyway. The biggest contributor is probably his stake in Apple.
Any broad market index fund ETF probably did just as well, if not better.
If you choose to _sell_ at $1 you will likely get a warning because it will fill instantly at the current market rate despite being a limit order. However you can choose to buy at $1 and there's no reason for it not to be accepted into the order book. In fact you often see this via Nasdaq TotalView during low liquitiy events such as right as the premarket opens where someone is bidding $1 or $0.01 for shares of a blue chip company.
Market makers, which are required to maintain a quote, have been known to put in a bid at $1, knowing it will never actually execute. Until the price drops below a dollar, which it turns out does occasionally happen. Oops.
I don't know about refuse, but if it did go through it would definitely raise a bunch of red flags from the anti-money laundering checks, IME these are taken very seriously. Accounts involved would probably get locked pretty quickly.
A yes, the famous S500K problem. Like the even more famous Y2K problem people (a small number in this case) will rush to fix it in time, and then the consensus will be that the whole thing was overblown.
Making limiting simplifications is one of the important legs of engineering (the other two being making tradeoffs and focusing on accomplishing the goal).
Sorry, it was a bit of a joke: Y2K was Year 2000 and my made up name S500K was Stockprice over around $500K. I suppose I should have written "famous".
Those of us over a certain age remember how much work went into patching the Y2K issue, and how it changed the software industry, and yet was pooh-poohed by the general press in the final year of the last millennium (2000) when nothing material went wrong.
My point was the article was a little bit alarmist, which it almost had to be to spin out a brief article about what is otherwise an amusing piece of trivia that would fit in a sentence: "Berkshire Hathaway's price has almost reached the point where it can't be represented in NASDAQ's computer, but they are patching their code and expect everything to work fine by the time the stock reaches that point, if ever."
I'd be glad to chuckle in passing upon encountering that sentence, but was annoyed at the padding used by the poor reporter to make an article of it. Given the state of the world it's hard to blame anyone (reporter, editor, Bloomberg management...) for this phenomenon.
There were things like the supermarket whose inventory computer was tossing a bunch of perfectly good product until someone noticed that the "expired" goods they were disposing of weren't expired at all.
When you are designing a very high throughput system you do often end up needing to trim bits from a transaction in order to fit an integral number into a page/packet/buffer or even a cache line.
Other kinds of code may not have these kinds of requirements in which case I agree that these kinds of limitations can be short sighted.
But that kind of decision is what I was talking about with respect to engineering.
The higher share price serves as a mechanism to stabilize the share prices as well. I have a few friends who disdain at the idea of selling even a single share because they don't want to bother converting it to B shares and/or don't want to trigger a tax on the capital gain.
Well, when they bought it, the share price was sub $100k. Some even lower. The B shares didn't exist before either (pre 1996), so some of these are legacy shares.
There's a great Acquired podcast episode about Warren Buffet and his obsession with compounding returns. After listening to that episode, it totally makes sense why he'd be opposed to a stock split.
I highly recommend it for those curious about Buffet
"[S]ignals the power of compounding" -- only to unsophisticated (frequently, retail) investors. Institutional investors always have historical data that is split-, and even dividend-, adjusted. If I had to guess, refusing to split is like a form of virtue signalling, nothing more. When I was a kid, IBM usually had a high stock price, over 100 USD. I never understood why; nor did my father, a stock broker.
"Unsophisticated" investors see the big number and the price history (which is much simpler without splits) and see the power of compounding plainly. There's not much more to it than that.
A common misunderstanding is that if a bunch of investors pool their stocks, they’ll earn more compounding interest on the larger pooled principal than they would individually, or conversely that a stock should not be split or it will earn less interest.
The continuously compounding exponential return formula is
P(t) = P0 e^rt
Note that only time enters the argument of the exponent, not the principal. So the returns are invariant to divisions of the principal (ignoring human behavior effects).
I like the pricing system. The shares were $16 when Buffett started buying, $420k now, haven't paid significant dividends, done splits or anything so the return to investors has been 420k/16 = 26250x. With most stocks it's really hard to do a similar calculation - you'd have to look at 60 years of history to figure the payments, splits and so on. That's basically why Buffett does it - he likes clarity and openness.
As an aside a lot of my bank / brokerages basically give a number out of range error if you try going back more than five years or so. Which is a pain for this sort of stuff - Buffett started buying into Berkshire in 1962.
Also the whole digits thing gets much worse in crypto where you have have 10^24 shares of some coin worth $10^-23 per coin or similar. I was trying to write some software the other day and the only thing I found was to store the numbers as strings in the database and convert them to something like Python long integers which can have unlimited digits.
Most market data vendors (including yahoo finance, google finance etc) account for splits (by dividing the historical price) and dividends (by simulating reinvestment). You don't have to think about those events unless you're doing something special in response to them.
It would be cool if you could store floating point (or double precision floating point) values in some sort of pseudo-type where you want to ensure that you have at least this much precision, and any time you insert a new row into the db table, if any of the individual values exceed the configured db max precision, as specified in the initial migration, or explicit table creation, it just automatically creates a migrated to bump the precision, (in Django you have 2 required fields for a decimal, it's like "max digits" and "precision" iirc)
EDIT: like an expandable floating point type, where as you increasing the needed max_digits, you just have this implicit migration that occurs. Idk this is super half/quarter-baked but just going off script a bit
Creating a migration to bump the precision would rebuild the entire table in most SQL databases. That's not something you want happening at random intervals.
It just seems silly that I'm building some PoC, and 3 years later all the assumptions I applied initially become something I need to actively design around, or just refactor my data schema. How many layers of abstraction would I need to add to create a kind of more flexible types that just know to expand if I add more significant figures, or contract (on a row basis) depending on the entry.
Like I was building a django app 2 weeks ago, and pulled in django-money to deal with currencies without reinventing the wheel, and then with another field on a model (meant to simulate a crypto asset), I had to arbitrarily decide what level of precision and what the "max digits" were for a class of potential instances of this model. I get that this might be overoptimization, but really - this is silly. By specifically asserting a max length for some field, aren't you wasting space? If not, then my entire point is moot.
Not super quantitative, but just thinking out loud a bit.
Assuming you mean like fixed-point arithmetic? Well, if you know you want to support all of 1 BTC's satoshis but also want to support Danish krone's, you could be wasting a lot of space by just delegating to the currently largest precision-requiring currency (or whatever, depending on your use case)
Unless I'm not understanding what you mean by fixed point?
For accounting, you'd want fixed point. I guess in this case it's decimal and not binary. Aside from that, there are arbitrary-precision ints. Python ints are. PHP promotes overflowing ints to floats, but this was probably a bad idea. We're also at the point that you can afford 64-bit ints.
Or we could follow Shakespeare's lead and not worry about spelling too much. Plenty of named also get anglicized into slightly different or multiple spellings. Are those Chinese people called Wang or Wong?
Not just his own name, nobody at that time could spell English consistently. Which might have had something to do with not even being able to read or write.
People's names are fundamentally oral and spelling is just an encoding to communicate them in writing. Being worried about spelling correctness it like being worried about the choice of binary encoding of the spelling. My name is only correctly encoded in binary as ASCII, not some bastardized similar-looking unicode codepoints, and certainly not that archaic IBM encoding! Future generations had better respect my preference! /s
It's the same character in both dialects. I don't think a particular romanization counts as somebody's language when plenty of native speakers can't even read or write it. If you want to respect spellings, we should write all Chinese names in Chinese characters, but we don't because we don't respect them enough to actually bother with effort.
Buffett sees the stock price as a feature, not a bug (IIRC he's also said that you should only buy BRK.A if it's at a significant discount compared to BRK.B)
Ahh, another 32 bit overflow problem. I still think I'll be making big bucks after my retirement working on the Y2038 epoch overflow problem [1], just as COBOL programmers were reportedly in high demand for Y2K.
I confess to having committed this sin. The order tracking at my employer suddenly started crashing during load and this happened when I was on the other side of the world with only e-mail access.
The culprit turned out to be a Y2K38 bug on my part triggered by a salesman fat-fingering an order into the 2060s. The only Y2K bug I had wasn't worth fixing--one list of reports was "date"-sorted (dates in the filenames) and making the display wider was more of a downside than dealing with the mis-sort at rollover.
The flip side of this problem is keeping track of exchange rates when currencies go through hyper-inflation. For example the exchange rate between Venezuelan Bolivar to USD right now is 0.000000.
I never quite got why Berkshire doesn't do splits - the only explanation I can come up with is flexing and keeping investors to be a small, exclusive, contained club.
In fact, splitting has become surprisingly rare in recent years
This is what I'd consider if something like this came up at work, and there was too much crufty old code in the 32-bit systems to get management to accept updating the whole thing to 64-bit.
Change it so that what is listed on Nasdaq as BRK.A is not actually BRK.A but rather 1/2 BRK.A. So when Nasdaq says BRK.A is $X it is really $2X, and when they say N shares traded that day it was really N/2 shares. Only allow transactions that involve an even number of the 1/2 BRK.A shares.
Hopefully the code to handle all that would just need to be in the interfaces between the 32-bit systems and the rest of the world, and should be a lot less work than changing the whole thing to 64-bit.
Outside of forcing a stocksplit, what do people anticipate an architectural solution would be? Is it simply having to rewrite a bunch of code to deal with 64bit numbers?
IIRC this is intentional, I seem to recall Buffet responding to this question that he intends Berkshire to be a "hold long" stock, so for that purpose it makes sense to have the retail investors pushed to class B shares (equivalent but with less voting rights, and stock splits have been done for class B shares) so that the voting shareholders are predominantly larger investors who generally have a longer term perspective than retail.
What the hell is the point of shares without voting rights? At that point all it is sans regular dividends is a speculation vehicle, which Buffett has long been against...
In theory, economic rights. If someday a dividend is paid, a stock buyback occurs, the company is liquidated, etc., all shareholders of BRK.B will be paid the same as those of BRK.A (accounting for the 1500 to 1 exchange rate).
In fact it seems like voting is actually not that important, since if you didn’t trust Buffett to run the company for you, you would be investing in something else.
In practice the rise of stuff like cryptocurrencies and NFTs shows that this theory is only partially true, and perhaps there is some nonzero lift on the price due to people just buying it because they think the company is cool or whatever.
It will drop when that happens, but it would probably not make sense to make that bet.
If you can construct a portfolio that exposes you to only that specific event (and not something like the further increases in value until that point -- seems difficult but maybe possible?) you might learn the probability of it happening at various points in time, through its price!
Maybe you'll find that portfolio seriously underpriced (another comment gave some potential reasons why it would be) in which case there's money to be made.
> If you can construct a portfolio that exposes you to only that specific event (and not something like the further increases in value until that point -- seems difficult but maybe possible?)
You'd do that by going long in all stocks that Berkshire owns, and short Berkshire itself. The problem with that is that Berkshire also has a bunch of non-listed subsidiaries you can't buy.
I think so. And I find it hard to believe that is already priced in, because only Buffett fans are interested in buying BRK.A now, and fans are frequently not rational about their fandom.
You don’t need to buy BRK.A to benefit from Buffett’s financial acumen. You only need to buy BRK.A if you want to say you own BRK.A.
Also, the harder it is to transact a stock (and a price over $400,000 makes it hard), the less efficient price discovery is.
He recently publicly announced his successor. No issues. No one blinked. One of the best long term companies to own shares in. Investing to build wealth is a long game. All of my "bets" are looking out 20-30 years. Short term some years I have not out performed the market, but over the lifetime since I started I have (15 years so far..)
Warren wouldn't characterize a long-term investment as speculating, even with no dividends and no voting rights. Speculation is when you buy something for its short-term price movements, whereas investing is buying a share of ownership in a company based on the intrinsic cash-generating ability of the company. Even with no dividends the profits don't vanish in thin air, they can get re-invested or improve price through share buy-backs
I've always felt this was a weak point for him. He goes over what he thinks makes a good investment for a normal person and why he thinks Berkshire is one of them a lot but glosses over this point any time I've heard him talk about it.
I think the idea is that he doesn't want to return cash while there are still good investments and made the B shares available for those who need liquidity.
They issued a 2nd class of share, BRK.B, which is around $280/share right now and equates to 1/1500th of an BRK.A share. Berkshire A shares are convertible to B shares, but the converse isn’t possible.
Yes, but Berkshire’s stock transfer agent won’t handle the conversion from B to A. You sell them on the market and are taxed accordingly.
Their stock transfer agent will give you B shares in exchange for an A share, which is not a taxable event. It’s a mechanism to encourage arbitrageurs to keep the prices of the share classes in lockstep.
Berkshire has two classes of stock A and B. B is targeted towards retail investors (trading ~$280.50 per share ATM). They don't want to split A shares because they believe that the high price attracts like minded investors focused on long term profits.
BRK is an unusual case where they offer both A and B, wherein B is retail-accessible. BRK.A mostly exists for bragging rights at the shareholder meetings / getting closer to a celebrity a lot of people deify.
For other businesses, stock splits can increase investment/price since it makes that stock more accessible to retail investors and more importantly their huge pile of retirement savings. For example automatic investment plans that can buy whole shares of a company each month (via salary redirected savings) are much more popular than sitting on cash for up to months to buy a single share (and that is harder to automate/brokers don't support it).
Fractional Shares should solve this issue, but many 401Ks simply don't support them, different brokers have a patchwork of what can/cannot be done with fractional shares (e.g. market orders only), and there's also questions about voting power/stock splits/broker transfers/legal rights/etc.
If fractional shares were offered not at the broker level but at the exchange level with identical [proportional] power/legal rights/access/etc to regular shares none of this would matter, but they don't at all. So we've got a system where share price is both meaningless mathematically but also somehow an impediment to an entire investment class.
Right, my point was poorly articulated. I know why most companies split, I'm just not sure what incentives BRK to do this now. Do you think their cap is negatively effected by lack of retail access? They are a weird case.
Kind of concerning that the systems running a major financial exchange are struggling with problems solved by game devs already.
What's newsworthy here isn't the technical problem, but the institutional problem that allowed 32bit overflow to even approach affecting a production system like this.
in a game engine you have several ms to process events from the user/network
if you're handling depth from US exchanges you're lucky if you have a microsecond
for maximum decoding speed: most major exchanges represent data using fixed point with fixed message offsets (CME being the main exception, but then they have fractional prices)
the 4dp was likely picked due to some past requirement (e.g. fractional prices)
doing a major upgrade of a protocol like this is not easy when you have to co-ordinate the upgrade with all of your members (all of which have their own software/appliances, and some likely pass that 4dp through their systems too)
Thanks for knowledgeable response. It does shed some light on the issue.
It's interesting that decoding speed is the major bottleneck here.
However it is still surprising that this issue wasn't foreseen and that the burden of responsibility seems to fall on the berkshire and not the system architects.
Remember that a lot of action on the stock market is built on speed. If they change the storage format to be 64-bit instead of a 32-bit number that would be a non-backwards compatible data format change. So the data format specification would need a new version, all the software that interacts with it would also need to be updated, and a lot of algorithms will get a bit slower since they now have to deal with a 64-bit number or some other representation.
There are probably some elegant ways to handle this, but again it would require that timing-sensitive software will need to do extra work to see if the value is in the extended format; or they would have to process twice as many bits if all data values are updated to the new format.
Also in in the past most stocks will split for many other reasons when the per share price gets above an easily to handle number. What wasn't predicted is that a company wants to avoid a stock split for whatever reasons they have.
The current pandemic was predicted by many experts. Obama even set up a task force to prepare.
That didn't stop idiots ignoring expert predictions. One orange idiot in particular disbanded the task force to save millions of dollars, and cost the US economy billions because of the slow response to a highly predictable disaster.
It's actually fairly comparable. In both cases, it's likely to happen at some point, you don't know when but by the time it gets close to being a problem you dont have that much time to make the necessary changes.
The stock price of BRK has been grinding higher gradually for a year and was within reach months ago. That this could happen should've been on the radar for a long time and it's definitely a failure if this was a "surprise"
Munger told a story about a secretary they deeply loved and then she died. They thought she couldn't be replaced. Then they realized the new hire was better!
Except the new hires wouldn't have that first hand experience to help them understand the importance of sizing your integers appropriately. Institutional knowledge is a valuable thing.
Spot on. Firing someone for a mistake is incredibly short-sighted. After the mistake has been rectified, you can be sure there is going to be one employee in the organisation who will never ever let that problem happen again. Why do you want to get rid of that person?
Of course, if the person is making the same mistake again, then that may be reason for concern, but that's a completely different discussion
