Afternoon Letter

The Newsroom - Quick Snaps of the New Dow Jones Newsroom

I was at the Newsroom earlier today to talk about data for our algorithmic products. I took these quick snaps of the newsroom. It really is a cool place, as newsrooms go; much quieter than most of the trading floors I've been on. All I can say is nice job to the designers.

Enjoy!

Signs of Recovery: Our New Economic Sentiment Indicator Thinks So

Today Dow Jones [WE] released a new Economic Sentiment Indicator, affectionately known around here as DESI (now we need a LUCY). The indicator makes use of search strings, word pairings and other qualitative measures while pouring through major newspapers to quantify economic sentiment. The indicator is measured on a scale on 0 - 100. Today's number, for the month of April, came in at 27.6 up from 26.3 in March and off the lows of 22.2 in November of 2008. I've included a chart below that shows the indicator since 1990.

As Die Hard's John McClane would say "Welcome to the party pal" ;-) You can get more information at http://solutions.dowjones.com/esi

Dow Jones Economic Sentiment Indicator

From the impending release:

On April 30th at 9:45AM EST, we will be releasing a new indicator based on sentiment from news stories. Here are the details:

·         The Dow Jones Economic Sentiment Indicator (ESI) is a monthly measure of the relative tone of references to the U. S. economy in 15 major daily newspapers using a proprietary text mining and analysis algorithm through Dow Jones Insight, a media tracking and analysis tool.

·         The ESI is expressed on a numerical scale from 0 to 100, with 0 representing no positive economic references and 100 representing no negative economic references.

So what is it -- a short hand metric to look at economic sentiment using qualitative data. An abundance of words tells you many things either good or bad. Look for it...

Having Fun With The Dow Jones Text Archive

Creating some grins, and giggles with data. I have been playing around with our large Dow Jones text archive. All it took was a few Perl scripts, a spreadsheet and some time. I wrote a quick script just to look for word appearances out of a pool of articles from 1987 to 1994. The counts are based on a certain word appearing at least once in a story. Each success is then counted as a story hit. So all the counts are based on story hits. You can do a lot more, this is very basic but fun. Remember this is very raw and non-scientific but semi-interesting to say the least. I picked the following terms:

collapse
bankrupt
default
upgrade
downgrade

Year	Collapse	Bankrupt	Earnings	Default	Upgrade	Downgrade	Total Articles
1987	155	651	5,261	253	391	483	68,430
1988	132	672	5,022	263	507	540	69,095
1989	177	1,065	7,434	399	644	825	83,934
1990	358	1,826	9,157	694	753	1,300	95,914
1991	492	2,124	10,932	747	1,197	1,526	114,861
1992	461	2,232	15,486	676	1,992	1,900	128,287
1993	451	1,810	18,103	528	2,738	2,022	165,601
1994	1,027	2,508	29,538	739	4,191	2,776	295,397

I also picked 'earnings' as a benchmark term. Why? I thought it might be a good word that appears enough times in a financial publication that would give me an idea of scale. In doing so I created a simple measure for myself to look at word frequency: (term appearing in a X stories / baseline term appearing in X stories) * 100. Example, 1987 "collapse" using "earnings" as a baseline term = (155/5261) * 100 = 2.95 vs. 1987 "bankrupt" using "earnings" as a baseline (651/5261) * 100 = 12.37. Below is a table with calculated results.

Year	collapse	bankrupt	default	upgrade	downgrade
1987	2.95	12.37	4.81	7.43	9.18
1988	2.63	13.38	5.24	10.10	10.75
1989	2.38	14.33	5.37	8.66	11.10
1990	3.91	19.94	7.58	8.22	14.20
1991	4.50	19.43	6.83	10.95	13.96
1992	2.98	14.41	4.37	12.86	12.27
1993	2.49	10.00	2.92	15.12	11.17
1994	3.48	8.49	2.50	14.19	9.40

1990 & 1991 were watershed years for the word "collapse" according to the index. In 1990 the appearances of the word "collapse" jumped from 177 stories to 358 stories, 102% increase and total headlines only increased from 1990 to 1991 by 20%; by that same measure my "earnings" benchmark only grew by 23%. Drexel Lambert "collapsed" in 1990 taking 39 stories in the count with it. There were 388 stories all together on Drexel in 1990. In 1991 "collapse" appeared in 492 stories overall. Remember The Gulf War was a big topic that year with 2,987 stories mentioning Iraq in some way but only 33 of those 2,987 mentioned the term "collapse". As a contrast Russia (remember the collapse of the Soviet Union) was mentioned in some way in 817 stories with 39 mentions of "collapse" and "soviet" appeared at least once in 3,045 stories and in those stories there were 104 that mentioned "collapse". Out of 492 "collapse" stories 9 mention "scandal". And out of the 492 stories mentioning "collapse", "real estate" is mentioned in 39 of them. 1994 would see a huge increase in the stories mentioning "collapse" with 1,027, a 127% increase from 1993. However our number of headlines grew by 78% and the "earnings" term mention grew by 68%. I'll do more digging on that at a later date. Here are some sample headlines from 1991:

• Bank Of New England Collapse May Put Focus On Regulators
• B.F. Goodrich CFO Collapses During Analyst Presentation
• Report Schwarzkopf Says Iraqi Army Near Collapse
• Some Iraq Units Said Close To Collapse; Some In Good Shape
• Iraqi Ridicules Schwarzkopf Comment Iraq Army May Collapse
• Investcorp Sold Intl Leisure Shrs Before Leisure Collapsed
• Vista Chemical Dn; Traders Cite Fear Of RWE Deal Collapse
• Riegle Seeks Probe Of Bank of New England Collapse
• Govt Allows Firm To Cover Losses From Exec Life Collapse
• Dam Collapses In Romania; At Least 66 Reported Dead
• Soviet Communist Party Near Financial Collapse Paper Says

Also another small anecdote, the following headline appeared on September 27, 1991

Wall Street Hired 6,000 In 2Q First Increase Since 1987

That headline coincides nicely with the beginning of an uptick in the word "upgrade". See the the chart above on terms, 1992 shows "upgrade" starting to increase relative to my other terms. Here is a chart of the Dow from 1987 - 1994 for comparison

Bankrupt was also an interesting term. It held sway from 1987 until some point in 1992 when the words "upgrade" & "downgrade" overtook "bankrupt". Here are the Bankruptcy Filings for the period 1987 - 1994 just for amusement.

Year	No. of Filings
1987	88,236
1988	68,482
1989	62,478
1990	64,688
1991	67,714
1992	71,725
1993	64,857
1994	54,425

Look for future updates to this idea as I get further into the data. Enjoy!

Low Latency Trading Is Expanding

One of my colleagues just sent me a story from the site AdvancedTrading.com

The story is on the growing nature of low latency trading. Wall Street & Technology held an event on Wednesday to give people an overview.
Here is a link to the full article.

Low Latency, It’s Not Just for Hedge Funds Anymore

Here is what caught my eye.

CME Group's Goldman also discussed the increasing demand for colocation. The exchange has been offering the service for four years, but, he notes, it's still growing in popularity among customers. One of his concerns is that customers are not creating backup sites to their colocation sites.

This I've seen from personal experience. As the timescales have decreased more and more participants are looking to have data delivered as close as possible to the gateways of the exchange. Especially in the case of trading off of economic indicators.

More TradeTech Observations

Spending a few days with traders can be an eye opening experience. I just wanted to distill a few more things I picked up in the halls and presentations of the TradeTech conference.

How is this market trading? - Short Version

Well it depends on who you ask. Informally I asked quite a few head traders from both the traditional buy side (think big mutual & pension funds) and the alternative side (hedge funds) for their answers. A common refrain; the biggest problem facing mutual funds is the fact that they need to be in invested in the market. Staying in large cash balances is a no-no for any long period of time, given the terms of a fund's prospectus. Add to that the increased volatility and you have a recipe for difficult day to day trading mechanics for the folks in the trenches trying to execute a PMs strategy. So staying in cash is not an option and trading other products may be out of the question. For the hedge fund traders, strategies may shift in order to take advantage of events or volatility. As one trader mentioned to me they were seeing "more of the market as trading with high frequency strategies" Think of this as machine automated tick by tick trading. The Institutional version of day trading. Bottom line - continued craziness.

David Leinweber's Presentation

David gave a talk on the Great Financial Mess or '08. He had some great slides and insights. My favorite part was this Time Magazine Cover from 1994

Time magazine was about 15 years ahead of it's time in reference to the current blow up based on sophisticated products whose pricing and value were difficult to ascertain. These instruments which were mostly confined to the debt side is what killed us. According to David, playing to an equities crowd, "you guys didn't do it this time ... it was those guys on the other floors in your building" He also pointed to the Wired article by Portfolio's Felix Salmon on David X. Li's Gaussian copula function in Recipe for Disaster: The Formula That Killed Wall Street

David is also about to release a new book on the space. Here are the details from Amazon and the link

Nerds on Wall Street tells the tale of the ongoing technological transformation of the world's financial markets. The impact of technology on investing is profound, and author David Leinweber provides readers with an overview of where we were just a few short years ago, and where we are going. Being a successful investor today and tomorrow--individual or institutional--involves more than stock picking, asset allocation, or market timing: it involves technology. And Leinweber helps readers go beyond the numbers to see exactly how this technology has become more responsible for managing modern markets. In essence, the financial game has changed and will continue to change due entirely to technology. The new "players," human or otherwise, offer investors opportunities and dangers. With this intriguing and entertaining book, Leinweber shows where technology on Wall Street has been, what it has meant, and how it will impact the markets of tomorrow.

David Leinweber, PhD (Pasadena, CA), is Haas Fellow in Finance at the Haas School of Business at UC Berkeley, and founder of the Center for Innovative Financial Technology at Haas. He is the founder of two pioneering financial technology firms and has consulted, published, and lectured widely on the use of advanced technology, artificial intelligence, and intelligence amplification in finance.

David is a buddy and he has spoken at Dow events in the past [future]. I didn't get a chance to give him a back of the book blurb -- but here is my shot.

"Dave does us a courtesy by stripping off the varnish of trading with heavy technology. He makes, normally, out of reach topics easy, without skipping around the details. If you want to understand where we've been and where we're going with machine trading on Wall Street, this is the book."

Thoughts, Rumblings, and Insights From My Day at TradeTech

I spent yesterday at the TradeTech Conference in NYC, Dow Jones is a sponsor. TradeTech is a mix of technologists and traders from many buy side and sell side institutions. The mood and tenor of yesterday's conversations were tempered by the downside swing in most indices, so keep that in mind with some of the comments. Overall, trading is alive and difficult in this environment as head traders search for the best ways to execute for their portfolio managers.

The first topic of conversation yesterday was dark pools, those often misunderstood liquidity providers that have become more commonplace for breaking up large orders via algorithmic models. The key feature of dark pools is the anonymity and confidentiality that they provide institutions in masking the intentions on orders. For the most part buy side firms really don't know how many dark pools their sell side brokers may be connected to. It is a concern for the buy side, but the attraction to execution, anonymity, and liquidity out weighs concerns, at this point. About 2 billion trades or 25% of US equity market trades in dark pools everyday.

When the talk turned to the trading environment and volatility, most traders stressed the challenges of externalities. Before March 17th 2008 [the Bear Stearns announcement] trading was still based on company news and overall macro. Once the crisis got into full swing and the government inserted themselves into the market, the shift in trading turned to any little piece of negative news. Most of the impact news is from economic releases and administration officials. Company news, if positive, has little effect. Negative news on the other hand is driving head traders motives on execution. The impact of ETFs is also being felt as traders need to watch the ETF show in the last hours of the day to understand the full move on their positions. With larger positions, he who can wait or hold cash balances longer, seems to be better off, especially when working with a portfolio manager that has a stated view of the market.

When it comes to information flow, the list just grows. Email, TV financial news, broker reports, news feeds, instant messaging, etc. In this information deluge traders see the advantages of research liasions, smart research-sales and sales-traders. These professionals that know their client accounts and can provide a constant "human touch' filter are becoming more valuable. As one trader described it,

"I have to cover so many names, that a good sales/trader, at a broker, can alert me to company calls or to the right news items I may have missed."

Clearly there is an opportunity in this market to have technology amplify and empower human coverage and surveillance in some ways. In all my conversations with traders, remember this is a trading technology conference, I am struck by how much institutional traders value high touch relationships and service.

Lastly when it comes to choices on sell side brokers, portfolio managers and analysts are driving the selections based on research coverage. Traders in contrast are fighting to pay the brokers that they execute through. In most firms the portfolio manager [& analysts] drives 70% percent of the decision input, trading gets about 30%. The ability of the broker's analyst to influence decisions at a buy side shop is a factor of the limited sell side coverage these days. Limited coverage, especially as it relates to Mid Cap stocks. In order to handle the volume of analysts used, each buy side firm is tracking analysts by using internal voting systems to pay brokers for transactions based on rewarding analysts. This is being done outside of Institutional Investor's normal survey -- II just doesn't mean as much to the buy side as it use to. With the seat changes in sell side, the proliferation of payments to 3rd parties for research and execution, will continue to rise.

All great stuff on day one. Today my buddy David Leinweber, of the Berkeley Haas CIFT is speaking on the Great Financial Mess of '08. Should be fun.

We're At TradeTech This Week

My buddy Alan Slomowitz (also from Dow Jones) and I will be hosting a series of buy side panels today at TradeTech in NYC. Here are the topics we will be discussing with a group of buy side participants.

• Current Market Structure And Its Effect On Your Trading
• Low Latency Electronic Trading Solutions
• Navigating US Dark Pools
• Volatility And Its Effect On Your Trading Strategies

I should be able to twitter some of it. I will also try to get a summary post on the event up this evening. Stay tuned.

Marx Quote Is Bogus, Thanks Pimm, But This Hayek One is Real

OK -- I listened to Bloomberg's Pimm Fox and I got some bad mojo. I should have read the IHT first:

A banker in China recently forwarded this quote to me: "Owners of capital will stimulate the working class to buy more and more expensive goods, houses and technology, pushing them to take more and more expensive credits, until their debt becomes unbearable. The unpaid debt will lead to bankruptcy of banks, which will have to be nationalized, and the State will have to take the road which will eventually lead to communism. (Das Kapital, 1867)"

"How prescient Karl Marx was," my friend marveled.

The problem is, Marx never said it - it's a bogus quote that has made its way to Beijing from Wall Street. But the fact that people would look to the father of communism at this time of financial upheaval offers an interesting insight on the state of minds today.
---

Whether Marx said it or not, is not muy importante. The fact that it is being relayed at all through a news organization can tell you much about the state of the world and the behavior of market participants. It is constructive of the current state of the market's psychology and mood. As someone who looks at and appreciates the qualitative side of data -- the words, frequency and meaning -- I find this all fascinating and luckily, it is something I get to work on regularly. Numbers are important(understatement of the century for financials, etc.) but in most cases for finance they are backward looking. Language I believe is becoming very important and can be forward looking and in a growing number of cases predictive.

So if we begin with pseudo-Marx, let's end with some Friedrich von Hayek - from a Meets the Press Interview in 1975

Will:

"Dr. von Hayek, capitalism and particularly American capitalism would seem to have a good record at giving people a rising standard of living. Why are so many intellectuals, and particularly so many economists, skeptical about and even hostile to capitalism?"

Hayek:

"Well, I've been puzzling about it for a long time, particularly about the economists who ought to understand better. It's very difficult to know why they don't. I think it's the intellectual attraction of a system you can deliberately control, which is fascinating to the intellectual."

---

Think on the language of that for a while. Hopefully it starts to make some rounds...

Karl Marx on Bloomberg Radio This Morning

I love the Bloomberg Radio Surveillance show in the AM. You can get some of today's show via itunes as a podcast Host Tom Keene & Ken Pruitt always have great guests with loads of information. The ever popular Dr. Doom Nouriel Roubini is a semi-regular guest. Ken was out so Pimm Foxx was the co-host with Tom. It was pretty funny when during the program a few Karl Marx quotes made their appearance. Here is one in full:

"Owners of capital will stimulate the working class to buy more and more of expensive goods, houses and technology, pushing them to take more and more expensive credits, until their debt becomes unbearable. The unpaid debt will lead to bankruptcy of banks, which will have to be nationalised, and the State will have to take the road which will eventually lead to communism"

Thanks for the chuckle Pimm ;-)