UBS built their USA Headquarters in Stamford, Connecticut – just 45 minutes by train out of New York City. The architecture is stunning – a contemporary design with huge 5 story windows across the façade of the building, and an interior stairscape that leads up to the fifth story all in one sweep – landings on each floor. Above the fifth story of executive offices, you come up (by elevator and through intense security) to what looks like an airplane hanger of wide open space with a four-story high ceiling complex of beautifully interwoven lighting and air-circulating equipment. This gigantic trading floor is built without columns or interior supports – there is no interference with a view to every periphery of the room twinkling with multicolored numbers and symbols. I met with Bill O’Donnell at his computer in this very exciting environment…
William O’Donnell [WOD]: I’m William O’Donnell — I run the US Interest Rates Strategy Group here in Stamford for UBS. In the beginning years of my career, from 1980 to 1984 I helped to start up a company called Technical Data, which was ultimately bought by Thomson Financials. They published a commentary that I contributed to during ‘82, ‘83 and ‘84 – a technical write-up – and that’s where I learned a lot about technical analysis. I carried those lessons into my career as a Strategist. I work with a wonderful team at UBS in the US Economics Department run by two senior economists.
I’ve always felt that trends are driven by fundamentals, and that technicals are critical at turning points in the market. That’s why I love using a blend of the two — of technical and fundamental analysis — to synthesize a market view. That’s what we do here in Strategy.
Fundamental trends don’t change that quickly — technical trends do. Technicals – looking at support resistance levels, looking at momentum, looking at position sentiment, over-bought and over-sold conditions — help us to read the “weather systems” that evolve continuously over a fundamental landscape.
MS: What tools do you use most frequently?
WOD: That’s an excellent question. If we remove ourselves from the fundamentals for a moment, what we use in technical analysis are a couple of different things… And I would start by saying that my training goes back to pattern recognition. And when I say pattern recognition use two forms in particular: First, there is the “Edwards and Magee pattern recognition” that goes back to that great book, The Technical Analysis of Stock Trends. I look for patterns such as rising wedges, ascending and descending wedges, and rounding bottoms and tops. I also look for formations such as pendants and flags in the context of markets and trends.
Patterns and formations help me to identify where a consolidation is occurring, and when momentum gets overdone. Are we going sideways in a flag? Do we correct slightly downward in a bull trend, in a pendant? Is it a continuation pattern? They help me to know when the next leg of a move is coming.
The second form of pattern recognition I use is candlestick charts – I’m relatively new to candlesticks. I’ve been watching them now for the past five or six years. I trained with Steven Nison’s well-known books on candlesticks, and I have found candlestick observation to be excellent at turning points. I look for dark cloud covers, evening stars, morning stars, hammers, among other patterns, to help identify changes and turning points.
MS: What kind of charts do you look at?
WOD: I typically look at daily charts, because I have to comment every day – or we comment every day as a Strategy Team here at UBS. My experience has shown that traders, people on our trading desk, tend to use short term charts such as 30 and 60 minute charts. They’re trying to define where the next few ticks are coming up. We try to look at the bigger picture. We try to look in the context of the next week or two. And daily charts seem to be the best fit for that. We like to know what season we’re in (in the market) – is it winter, is it summer, is it spring? So we use the weekly charts to try to take a look at even a further step back – kind of like getting a feeling for the weather.
MS: Could you explain your weather analogy?
WOD: Absolutely. This is an analogy that I use with born and bred fundamentalists — those who think technical analysts are shamans – it’s the analogy of the fur trapper in northern climates. The fur trapper has to know the fundamentals of his job. He has to know where his quarry lives. What the idiosyncrasies of his quarry are. In others words, where to set the trap, when to set the trap, what they like to eat, and so on. That’s what I would call the fundamentals of fur trapping.
But at the same time, the technicals of the market are something like a weather system –what is the weather out there – which way are the winds are blowing. What is the temperature, and how is it going to affect his production or yield – as well as his own well-being. Is a storm coming? Based on that, how much food do you take when you’re going out to set your traps, or to recover your game? What kind of protection, clothing — the whole thing. That’s where technical analysis I think has its fit.
In the context of the fundamental nature of the market, it helps to know which way the winds are blowing in the short term. It helps to discover how and when to enter and exit.
For example, learning what today’s Nonfarm Payroll number is (Friday, April 6) is not going to tell you where to get in or out of the market — technical analysis will. Today’s Nonfarm Payroll number will not tell you how over-bought or over-sold the market is or which way people are leaning with respect to positions or sentiment. We’re talking about tools that I use – and on that note – about positions and sentiment — part of our analysis is to look at positions or sentiment proxies along with forecasts for a given number. We keep a weather-eye out to the JP Morgan Survey, the Stone and McCarthy Duration Survey.
MS: A weather-eye?
WOD: We keep a weather-eye out for position proxies, well-known ones, the JP Morgan Survey which comes out on Tuesday; the Stone and McCarthy Duration Survey of portfolios that comes out on Wednesday; the CFTC data which is free, and defines speculative futures positions. We take the CFTC data and aggregate net spec positions in all the financial futures contracts. The aggregated CFTC data has been an outstanding tool to help indicate potential turning points in the market in past years. And it’s been a very, very good tool at turning points – not all of them — but many of them.
People tend to be too long at highs — therefore over-extrapolating economic weakness into the future. And at the same time, the crowd tends to be too short and defensive at market lows. The CFTC data along with other position proxies tend to highlight this phenomenon.
MS: Can you give me an example of that?
WOD: A good example would be at the end of November of 2006 the market had spent basically the whole period from the beginning of July of 2006 all the way to end of November of 2006 kind of beginning to get their arms around the slow down in the housing market and what the ramifications for the economy are. At the same time throughout the period of — I would say really March through June of 2006 when 10 year yields rose from 4.7% to five and a quarter — the market got extraordinarily bearish and very defensively postured and very sure.
And this was highlighted in the CFTC data, the Stone and McCarthy Survey data and my most favorite sentiment index, which is the five day moving average of the Daily Sentiment Index (DSI) that’s put out by Jake Bernstein and MBH Commodities. Jack Bernstein is a well lettered technician who’s been in this business a long time, written a number of different books about sentiment and technical analysis, and seasonality and different things like that.
MS: Is this data available?
WOD: By subscription, but I find it extremely valuable. Because what he does is he goes into the futures pit as I understand it and basically does a survey of percent bulls. I don’t know what his exact methodology is, but let’s say, 48 of the 50 people he interviews in the bond pit are bullish and it’s 96% bulls — it’s an index of zero to 100 — 100% bulls or zero percent bulls.
During the period from March through July (2006) there was excessive bearish sentiment. You had roughly 6% bulls to 15% bulls, if my memory serves me correctly, during that time period. At the same time, speculative positions became extremely short. The market then got a dose of good news at the beginning of July, took off, and bond yields started to fall, from roughly the 4th of July all the way to end of November and the beginning of December.
At the beginning of December the net spec positions of the CFTC showed that speculators were extremely short, in aggregate. What we do is we take all of the contracts, take the duration value of the ‘01 of each of the contracts, aggregate them all back into an equivalent of the 10 year futures, a notional value — and what we’ve seen since 2003 is that aggregate net spec positions and on a net basis have a tendency to swing between being long or short roughly 40 billion in notional ten year contracts. Those seem to be the historic extremes since 2003.
I would add, at that same time since roughly 2003 and when we hit the generation lows and bond yields, we’ve also seen a parabolic rise in aggregate open interest. And this may hint at the growing influence of black box trading, and its parabolic rise. So you can almost see that in the very sharp rise of open interest.
Anyway, that’s sort of a side light. The fact is at the end of November we had extreme bullish fervor, people were really embracing, even over-embracing at that point, the housing slowdown. We had the five day DSI (Daily Sentiment Index) was up in the vicinity of 90% bulls at the end of November, beginning of December, if my memory serves me correctly. At the same time, net speculative positions were roughly about 40 billion long in 10-year contract equivalents. And we went neutral on bonds having been bullish throughout that whole period from March all the way to end of November 2006. Again, the technicals were basically suggesting it was a good place to be short. For us in Strategy, we like to meld the technical view with the fundamental view. So what we did back then was just basically go neutral, because none of the fundamental conditions had really changed. What had changed was sentiment positions and rate levels. But to us, the fundamental underpinnings is in the market were still weakening, so we went neutral, not short. Had we been “technicians only”, we almost certainly would have recommended going short. We did suggest those that wanted to be a little bit more aggressive could certainly do that and here’s where you’d stop yourself out and so on and so forth.
MS: What do you think about Stochastics and momentum indicators?
WOD: My favorites are the market standards — which are the slow Stochastics and the Relative Strength Index. I tend to use the nine-bar RSI and I like to look at the 14/3 Stochastics, this is %K and %D slow Stochastics — as well as the 18/3 slow Stochastics. We ran some regressions back to 1995 for 10 year futures and found that the 18/3 slow Stochastics was actually the best to use in terms of profitability.
We tested the Stochastics by doing the following: we said “Okay, anytime the Stochastics indicator comes from above 80 and falls below 80, and the 5 day daily sentiment index is 85% bulls or higher — and I’m using 10 year prices — so we’ve been in a rising market. Momentum indicators were getting in “over-bought” territory — above 80, this is all on a daily basis. And they get overbought and they come below 80 — and the five day DSI is at 85 or higher — which is sort of what we consider to be the extremes of bullish sentiment — then we go short.
We stay short until we get a slow Stochastics reading below 20 – “over-sold” – that rises above 20 at the same time that the five day DSI is at 1% bulls or lower. That sets up for a buy signal. Oddly enough it generates very few trades, but tends to make a lot of money in those trades over long periods of time.
MS: I was going to say, what’s the time frame?
WOD: We’re going back 15 years.
MS: So these are long calls.
WOD: These are long calls. I think we have 15 trades in 15 years. But just from a pure trading perspective, it’s 15 trades that have made 90 points. This is 90 full points in 10 year futures.
Now, that’s a simplistic way of overlaying — which is what I like to do, — simple known indicators; merging them together and using them in concert. In other words, where the stars align and reliable indicators such as the DSI, the sentiment, the Stochastics and RSIs are all confirming. It is then that you have the genesis of a really good rate call.
And we look at the world order — what’s going on in China, what’s going on in Japan, what’s going on in Europe — beyond what a pure technician would look at. So we’re kind of a hybrid in the technical world, and I think that’s important.
MS: Good word, hybrid.
WOD: As far as the Stochastic Indicator is concerned, George Lane — who revolutionized it back in the 1950s and ‘60s — always said that the best use of Stochastics is in instances of divergence. And over the past 25 years I have found the same.
Divergence is when your underlying price is making higher highs, but at the same time your momentum indicators — if your slow Stochastics 14/3, 18/3, %K and %D — are making lower highs. Suggestive of – and this is an analogy I use — buying a two-story house with a given foundation, and just constantly piling new stories on top of the same foundation that was built for a two-story house. Eventually you see it collapse.
MS: Interesting analogy.
WOD: Another great tool is right out of the Edwards and McGee book (8 the Edition) — I think it’s around page 160 or 161 — on ascending and descending wedges. They talk about the ascending wedge as being somewhat of a bearish warning — a “petering out of investment interest.” And it’s very similar as the divergence — you have an ascending wedge with notable bearish divergence. You have a dominant up-trend — but at the same time you have converging trend lines when you’re growing the trend line off the highs, not the lows.
That idea can be transferred to divergence in Stochastics. When using divergence in Stochastics, I like to look at the Relative Strength Index to confirm the convergence. And I take a step back and hesitate when I see divergence in the Stochastic
Indicator, but not in the RSI. You almost always see the two together — when you don’t, it’s a yellow light. So, again, it’s all about handicapping the future direction of rates.
MS: What about trend lines?
WOD: You can’t beat simple trend lines. They’re almost too simplistic for a market that’s becoming more technical with computerization and programming.
MS: What periods do you use for trend lines?
WOD: It really depends on the points that define them. In other words, if you take a normally scaled chart of five year yields and go back 20 years, these are five year Treasury yields, you will see the most beautiful down-sloping trend line in the yield defined by 3 points — which gives it more power — as opposed to 2 points. Coincidentally, it was that trend line that stopped the market at its lows in May and June of 2006. I’ve watched this chart for 10 years.
The key thing about trend lines beside the fact that they are the most logical way to define a change in trend, is that every piece of technical software, including Bloomberg, allows the technician to draw in his own trend lines and to manipulate and watch them – easily.
I grew up in this business using complicated studies that required cycle tools, and a lot of raw “computational horsepower”. Like many others, I learned by keeping Point and Figure Charts by hand. So keeping my tools simple has always been an important feature of my work. For me, simple trend lines are essential in identifying changes in the market.
MS: You kept Point and Figure Charts by hand?
WOD: All by hand. And I know many people on Wall Street today that still keep Point and Figure Charts by hand. P&F charts kept by hand tend to speak to the chartist in ways that CQG charts do not. George Lane talks about this in his books. A good analogy has to do with playing golf.
A lot of people don’t like to use a golf cart when they play because it increases the pace — it doesn’t allow them to think before the shot. A lot of golfers I know eschew carts for walking the course, because it allows them to come up behind the ball, to line up the shot, to think about the next shot in the context of the hole. And I think that’s a perfect analogy for speaking to those who still keep the Point and Figure Charts, and line charts, and bar charts, or whatever they keep by hand.
MS: It gives you more of a feel for what’s happening…
WOD: Absolutely.
MS: Do you think that Technical Analysis is more of an art than a science?
WOD: It’s all art to me. None of this is an absolute. Things always change, and nothing is constant. A lot of the art of the Technical Analyst has to do with his/her ability to adapt to change. For example, you could develop a killer model — and then something will change in the marketplace. A perfect example of that is that — in terms of the time series of aggregated net spec positions in the CFTC — the world changed in 2003. What was formerly an extreme range — between extreme longs and extreme bulls — dramatically changed in 2003, as open interest climbed, as volatility declined from 2003 to today. Since that time, volatility has declined, and traders appear to be speculating in ever-greater amounts trying to chase ever-smaller profits.
They’re trying to scale-up their positions to try to maintain some sort of semblance of return. You can see that since 2003 volatility has gone from the upper left to the lower right while aggregate open interest, at least in the financial futures, has gone parabolic. And with that, the amplitude and the frequency of the swings for net spec positions have increased dramatically.
So if you’re so dogmatic about sticking with this being over-bought, or this being over-sold, or this being extremely long, or this being extremely short in terms of a specific net spec positioning, you’d be blown out of the water beginning 2003.
And that is the art, things always change, nothing is constant – and you have to adapt to changing conditions. Prices will change, and positions will change, but again, what’s tried and true tends to work.
MS: What about the “art” of using a momentum indicator?
WOD: In terms of using momentum indicators — I need to anticipate that my momentum indicator — as reliable as it is — will have the potential to show (for example) over-bought levels very quickly. If I didn’t know that, I might trade out of a position right when a bull trend is in the initial phases of developing. The art of the technician is to recognize that, and to maybe say – wow, on a larger horizon, do we have a potential sea change here — do we have a major trend change in the larger picture? A technician needs to exercise his art by looking at the bigger picture — in the weekly chart and the monthly chart — which will identify that – yes — you’re probably in the beginning phases of a new trend. That’s the art.
MS: What develops your “art”?
WOD: Experience… and failure and success, you need them both.
MS: You need failure?
WOD: Yeah, it’s like Hogan – Ben Hogan – again, I’m not a golfer, but I’ve heard so many analogies to golfers. I think Ben Hogan said something to the effect that his success was directly related to his time in the dirt, i.e., the more time he spent on the practice tee the better his golf game got. If you apply that idea to Technical Analysis, I think you just have to back test.
Now with computerization you have the ability to back test. I find it very hard to back test divergence, because it’s a fairly complicated methodology. And the problem with divergence is divergence can persist for a long time, so it’s hard to capture that. But with computerization, you can create custom algorithms, which many technical models allow you to do, such as CQG which is what I use for my charts predominately.
MS: CQG is…?
WOD: Commodity Quotes Graphics, pretty well-known technical platform, there are other good ones. I know our foreign exchange people use something completely different. I’ve heard that Trade Station is another good one. People are coming out with ever-more powerful technical tools and that’s being demand-driven by the marketplace. The advent of computerization allows all of this and people are finally coming around to it.
The interesting aspect of the financial markets and technicals is that technicals really didn’t exist in our marketplace before 1980. They were deemed a constant in the commodity markets, they were a constant in the equity markets, as evidenced by the fact that Edwards and McGee some 60 years later are just coming out with a ninth and tenth edition of a book written in 1948. But it wasn’t until financial futures were developed — I think beginning with the bond contract of 1976 — that they started to populate the bond pit with traders from other pits.
So they brought their technical bag of tools with them from the other pits around the financial pit, and specifically the bond pit. And it was almost like a virus. My boss, Jeff Parker who is the founder of Technical Data, along with Jodi Morse, was a technician. He was one of what I would call the freaks of the business back then, he was one of the very few who were willing to even admit that he was a technician in the bond business, and he brought that over with him.
And he basically heard the dog whistle at that time and started Technical Data. And they were somewhat of a leader because they published technicals — even charts — online to subscribers in the financial industry. And the success is well-know. But at the same time that, in effect, inoculated the bond market to the fact that “there is something going on here with technicals.” At the very least because there’s a growing tide between this bond contract and cash bonds. And this is the tail-wagging-the-dog-thing, the guys who were controlling the price action — and the fact that they were driving the price action in the bond pit and creating the bond contracts in Chicago using technical tools. Wow — this is interesting — suddenly the technicians in Chicago were beginning to influence — as the ties between the bond contract and the cash bond market in New York itself began to tighten up through basis and intermediation.
Now all of a sudden people in the cash markets in New York began to take notice. And that’s when things started to take off, they took off right around the time that the IBM PC came out and the Apple IIC — it was the early ‘80s. And that’s the genesis, I think, of where we are today in the bond market. And I grew up in the early days of that.
MS: What was your educational background?
WOD: I was an economics major — schooled in fundamentals. And, I got interested early in life when my dad used to ask me to read off the closing stock quotes for his stockholdings out of the paper to teach me how to use the stock page. A “what is this?” kind of thing. At the same time, I was always fascinated by the financial industry in particular. And I sort of imagined myself going into equities not bonds, given my reading the stock page to my dad when I’m 11. I like it because there’s constant motion — things change every second. So that makes your day go fast. If you’re going to work a long day, you might as well go all the way.
MS: Back to the present, what do you think of the black box trading phenomena?
WOD: I think that it probably is enhancing what I call the delta effect. Option traders would certainly understand this, when you’re long an option you have delta. And so when the market moves you trade what they call their deltas. In other words, as the market rallies you get long, when the market declines you get short.
MS: Would you elaborate on that?
WOD: Sure, what it means is that with people spending ever greater amounts of money and ever higher amounts of open interest to chase return on the market, what there will be a tendency to do as the size of the pool of investors grow, there’s more sellers on up-ticks and more buyers on down-ticks than we would have had before – and this is a supposition on my part. A lot of those black box tools are all probably momentum-based tools, in some form. It’s similar to an options trader selling an up-tick, mining profits, buying on down-ticks, mining profits from their short – they get long in up-ticks, short in down-ticks. Using prices to get them in and out of trades.
And then the delta is just simply trading your deltas — which is mining profits and movement in the market. So that tells me that their growing influence is probably constraining volatility.
MS: When you say it’s constraining…
WOD: It may truncate it –
MS: So it’s a mini-volatility, is that how you would say it?
WOD: Mini-volatility or a macro-volatility. We are starting to see a surge of these black box electronic accounts coming in and even trading our cash Treasury market via a multitude of electronic platforms such as TradeWeb.
MS: Do you think this is a beneficial or a dangerous thing in the market?
WOD: I’ll speak for the US Treasury and they would say it’s a beneficial thing. To the extent that open interest has gone parabolic next to liquidity, and there’s more liquidity, spreads have tightened, daily volumes have risen and that’s all liquidity. From a Treasury perspective, touch down! For those that feed off of disintermediation in markets, their game is going away.
MS: I understand you are publishing a TA book here at UBS?
WOD: We are creating a reference manual for TA – a compilation – the greatest hits from four technicians who have been in the marketplace for a total of 88 man years: Jason Pearl, who runs the technical team in London, Richard Adcock, also in London, and Jim Chorik and I are pooling our best experience — in Elliott Wave, Demark Indicators, cycles, and the various methodologies I’ve talked about here.
MS: When will this be available?
WOD: We’re in the final stages — so very soon.
MS: Thank you so much, Bill — It’s been a pleasure.