History: Round lots

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
umfundi
Posts: 3361
Joined: Tue Jun 07, 2011 5:26 pm

History: Round lots

Post by umfundi » Fri Oct 21, 2011 11:54 pm

In the old days, we would save our money until we had enough to buy a round lot of shares, possibly 100. We didn't buy mutual funds, because the fees were too high. Accumulate the money, then buy individual stocks.

I am really trying to understand the history. I read somewhere that investing a fixed dollar amount was proposed as preferable to buying round lots, if expenses were low enough. That was the genesis of Dollar Cost Averaging.

I have no desire to re-ignite the DCA discussion, but I will appreciate perspectives on the history.

And, not to start a political discussion, but do people understand how difficult and expensive investing used to be?

thank you,

Keith
Déjà Vu is not a prediction

User avatar
bertilak
Posts: 6160
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Re: History: Round lots

Post by bertilak » Sat Oct 22, 2011 5:41 am

I think round lots facilitate trading, perhaps only for places where there is a trading floor (e.g. NYSE as opposed to NASDAQ).

When I do a trade in my Schwab account it is sometimes broken up into a series of round lots + one smaller lot. For example if I put in an order to buy 263 shares of an ETF the order might get filled as three separate orders: 100 shares, 100 shares and 63 shares. Sometimes it may be more like 200 shares and 63 shares. But this is not always true. Sometimes an odd-lot order will get filled as a single order. Since orders can be filled at any of a number of exchanges it might matter where Schwab places the order. Schwab splits the commission across any sub-orders proportional to the number shares.

RE: DCA. I think the discipline of making periodic investments and doing so as cash becomes available is better than trying to pick and choose when and how much to invest. In other words DCA is better than trying to do market timing. I think the DCA "debate" is more about what to do if you have a fairly large lump sum to invest -- do it as a single lump sum transaction (well, simultaneous transactions split across your allocation targets) or to spread investments out over time in a sequence of even dollar amounts.

RE: Costs of investments. It is still expensive and difficult to work with a full-service investment adviser: huge commissions, phone-tag trying to talk to a very busy adviser (aka salesman/broker), long discussions with that adviser/broker as he tries to sell you on load funds or hot stock picks -- sometimes even IPOs -- and generally tries to keep you in the fold. I know. I just bailed out of one of these arrangements less than a year ago.
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker, the Cowboy Poet

User avatar
grabiner
Advisory Board
Posts: 23117
Joined: Tue Feb 20, 2007 11:58 pm
Location: Columbia, MD

Re: History: Round lots

Post by grabiner » Sat Oct 22, 2011 5:33 pm

bertilak wrote:I think round lots facilitate trading, perhaps only for places where there is a trading floor (e.g. NYSE as opposed to NASDAQ).

When I do a trade in my Schwab account it is sometimes broken up into a series of round lots + one smaller lot. For example if I put in an order to buy 263 shares of an ETF the order might get filled as three separate orders: 100 shares, 100 shares and 63 shares. Sometimes it may be more like 200 shares and 63 shares. But this is not always true. Sometimes an odd-lot order will get filled as a single order. Since orders can be filled at any of a number of exchanges it might matter where Schwab places the order. Schwab splits the commission across any sub-orders proportional to the number shares.
The whole order goes to the exchange, but it may be filled in part because other orders are for fewer shares. If you put in an order to buy 263 shares and I want to sell 200 at your price, I will place a market order to sell (or a limit order at your price), and sell my 200, leaving you to buy 63. If I want to sell 300, I will sell all 263 to you, and my remaining 37 will be sold to someone else (market order) or become part of the market (limit order).

The advantage of round lots is that they are displayed in the quote, which makes limit orders more likely to be filled; it doesn't matter with market orders (or marketable limit orders). I once sold 300 shares with a limit order, and my order was accepted in five lots over 20 minutes, only one of them round. The first lot was 24 shares, leaving me with 276; my offer was displayed as the ask but listed as only 200 shares. After the fourth lot, I had sold 278 shares, and the remaining 22 shares were not listed as the ask in the real-time quote. You could find them by viewing the book (a Level 2 stock quote), but anyone just looking at the bid and ask would see a larger spread.
Wiki David Grabiner

djw
Posts: 1456
Joined: Tue Apr 08, 2008 9:34 pm

Re: History: Round lots

Post by djw » Sat Oct 22, 2011 6:10 pm

Since the OP couched his post as a historical question, here's what I think I know about that.

Hard as it is to believe, once upon a time people traded lots and lots of shares of stock and no one had ever heard of a computer, except the clerks who did the math might have the job title "Computer" or "Calculator" or whatever the local language equivalent of those words may have been.

Then, around World War II (c. 1941) someone actually built an electronic computer. IBM's initial sales estimates were that there might actually be a need for a few big ones. They figured that they should be able to sell about one dozen of them per year worldwide. That's not a typo. Don't add any zeros. They were thinking that only folks like the US Census Dept, Defense Dept., and maybe a few REALLY BIG companies, say General Electric for example, might be able to put one to good use and be willing to pay what it would cost to design, build, sell, ship, install, program, service, and operate it at a reasonable profit.

So, it took another few decades for this new computer thing to really catch on and prove that more than 12 per year might be sold at a profit. Meanwhile, for many centuries, right up until the things turned up on Wall Street, people had been trading using chalk boards, abacuses, mechanical calculators, and just plain old pencil or ink on paper and recorded in physical paper ledger-type books.

When you're doing it that way, it's really, really nice if you can convince (or require) everybody to trade in nice round lots of, say 100 shares per lot. Tends to cut down on the math errors too.

Post World War II, brokers were still charging what are now considered usurious commissions, something like 5% per trade. My recollection is that even in the 1970s and 1980s, you might well be charged a HIGHER commission, or perhaps a MINIMUM COMMISSION greater than 5% of the trade if you insisted on buying 27 shares instead of a nice even 100.

Obviously those days have now gone into the history books, but it wasn't all that long ago. In fact, it wasn't very long ago at all that the stock prices were set in eighths of a dollar, not 100ths (by the penny) or even ridiculous additional decimal places (silly unless you're trading, like a million shares at a time, I suppose). Why eighths? Ever heard of "pieces of eight" in an old English tale? They literally used to take a coin and physically cut it into eight wedges. I guess they had really valuable coins in those days, relatively speaking, so that's how they had to make change... If you think about it, while doing things in decimals seems pretty obvious, it's a lot easier to cut a coin (or a pie) into eight equal slices than ten. Give that a try.

I love history. I hear that the reason the English drive on the "wrong" (opposite, left) side of the road is that knights always passed right shoulder to right shoulder so as to better make use of their lances or swords against their opponent...
Love many, trust few, and always paddle your own canoe

User avatar
nisiprius
Advisory Board
Posts: 37077
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: History: Round lots

Post by nisiprius » Sat Oct 22, 2011 6:54 pm

Fixed commission rates were abolished in 1975. It may be difficult for people to believe just how high the fixed rates were. Fixing of commissions goes at least as far back as 1792: the Buttonwood Tree agreement included the provision that
We... do hereby solemnly promise... that we will not buy or sell ... for any person... any kind of Public Stock at a rate of less than one-quarter percent Commissions.
For retail investors in the 1960s and 70s, it was much much much larger than 1/4%! My dim recollection was that Merrill Lynch charged me a bit over $100 commission for a $5,000 purchase, and this jibes well with a Cato Institute report, figure 1, showing individuals paid an average 1.7% commission. Similar numbers are shown in Schwab's 1980 ad, below.

Odd lots were even worse, and there was a sleazy maneuver: your confirmation notice did not have any indication that an extra commission had been paid. Instead, the transaction was simply listed as having taken place at a fictional price that was higher than the real market price. I don't remember how big the extra hidden spread was. It had to have been at least 1/8 per share more because stock prices were only quoted in 1/8-dollar increments! But I think it was more than that.

The abolition of fixed commissions sent shock waves through the industry, and initially the established brokerage houses refused to cut commissions. Reduced commissions were offered only by newly-established discount brokerages. My recollection is that the first ads I saw touted $29.95 commissions on stock purchases, which was considered stunningly cheap. Google Books finds me this Charles Schwab ad from 1980:
Image
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

djw
Posts: 1456
Joined: Tue Apr 08, 2008 9:34 pm

Re: History: Round lots

Post by djw » Sat Oct 22, 2011 7:00 pm

Apropros of nothing, I can't resist adding that computers are actually binary, not decimal, and they do their calculations in hexadecimal (base 16 = 2x2x2x2) NOT base 10. They display the resulting numbers as "rounded-off" decimals only at the very end (on your screen or printout). This explains some otherwise seemingly unexplainable rounding "errors"...
Love many, trust few, and always paddle your own canoe

User avatar
nisiprius
Advisory Board
Posts: 37077
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: History: Round lots

Post by nisiprius » Sat Oct 22, 2011 7:21 pm

djw wrote:Apropros of nothing, I can't resist adding that computers are actually binary, not decimal, and they do their calculations in hexadecimal (base 16 = 2x2x2x2) NOT base 10. They display the resulting numbers as "rounded-off" decimals only at the very end (on your screen or printout). This explains some otherwise seemingly unexplainable rounding "errors"...
Are you really 100.0000% sure about that? There was a longstanding tradition of literally including decimal arithmetic in the processor hardware of the old mainframes. The numbers were stored within core (proper period terminology for RAM) in a "binary-coded-decimal" (BCD) format, and a special set of instructions performed arithmetic operations on numbers stored in that format. And, of course, it is easy for software, running on a binary machine, to perform true decimal operations on numbers stored in BCD format. I never programmed in COBOL, but I'm pretty sure COBOL included decimal numbers and decimal arithmetic as fundamental features of the language.

I really have no idea what programming practices are used for financial and accounting software, but I think it would be wildly inappropriate to use a language like C++ for financial work, or to store values representing sums of money as floating-point "floats" and "doubles."

What a pity that the NYSE went from 1/16 of a dollar to cents! How much better if they'd gone to hundred-and-twenty-eighths!
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
mas
Posts: 1461
Joined: Tue Feb 20, 2007 12:54 pm

Re: History: Round lots

Post by mas » Sat Oct 22, 2011 7:42 pm

nisiprius wrote:I really have no idea what programming practices are used for financial and accounting software, but I think it would be wildly inappropriate to use a language like C++ for financial work, or to store values representing sums of money as floating-point "floats" and "doubles."
I don't think that there is anything inherently wrong with C++ in this regard, but certainly floats and doubles should be avoided.
Depending on the particular requirements, I have seen monetary values represented by integer types with an implied decimal (for example: 125 = $1.25). More common is the use of http://en.wikipedia.org/wiki/Arbitrary- ... arithmetic.

SQL databases can represent numeric values with a defined number of decimal digits (presumably BCD).

umfundi
Posts: 3361
Joined: Tue Jun 07, 2011 5:26 pm

Re: History: Round lots

Post by umfundi » Sat Oct 22, 2011 8:07 pm

As the OP, thank you for the replies.

Arithmetic or computational "accuracy" is now a matter of policy, rather than of computers or programming languages. If your financial institution rounds down that thousandth of a cent, that is their choice. Of course, no one notices a round-off of half a thousandth of a cent, but over trillions of transactions, it could be real money!

In the 1980s I lived through the standardization of computer calculations into what became "IEEE Arithmetic". Geek me out!
http://www.psc.edu/general/software/pac ... e/ieee.php

At that time, it was a real issue. Supercomputers were 4 or 8 bytes, and division was 40 times more expensive than addition. Computer makers cut corners to get approximate results faster.

Anyway, "Rounding off" should not have been a computer issue for at least the last 30 years. Also, I do not think it (intrinsic arithmetic accuracy) has anything to do with the programming language. What your brokerage statements say may be a different thing.

Keith
Last edited by umfundi on Sun Oct 23, 2011 9:21 am, edited 1 time in total.
Déjà Vu is not a prediction

HongKonger
Posts: 1079
Joined: Tue Jun 21, 2011 10:35 am
Location: Deep in the Balkans

Re: History: Round lots

Post by HongKonger » Sun Oct 23, 2011 8:19 am

What do you mean, "...in the old days", here on the Hang Seng, we still have to buy in 'board lots' and these are not fixed...some of the shares I hold must be bought in 'lots' of 100, some in 200, some in 500 - it varies from stock to stock.

Post Reply