Seeking Alpha
View as an RSS Feed

Dr. Ken  

View Dr. Ken's Comments BY TICKER:
Latest  |  Highest rated
  • It's not just about Libya, Brett Arends writes. Everyone had become way too bullish and way too complacent. That's why the market downturn is more than just a blip, and caution is the wisest course now.  [View news story]
    Yes, it is not about Libya. It is about Saudi. If this can happen in Libya, Tunisia, Yemen, and Bahrain, KSA is next. This is the fuel behind oil futures. When the protests increase in Riyadh, we could see oil to go $150/barrel or more...
    Feb 23, 2011. 08:29 PM | 5 Likes Like |Link to Comment
  • ETF Correlation Among Asset Classes Is Dropping [View article]
    I agree that correlations are falling (in an absolute value sense).

    Using 63-day correlation periods on daily returns, the correlation between the broad equity markets (seekingalpha.com/symbo...) and bonds (seekingalpha.com/symbo...) is now at -0.20, compared with -0.45 and -0.64 for the prior two 63-day periods. Gold (seekingalpha.com/symbo...) has risen slightly, now at +0.23 versus +0.19 and -0.08 in the prior two periods. Oil has fallen dramatically, from +0.76 to +0.64, and now to +0.37. Even real estate (seekingalpha.com/symbo...) had decreased; from +0.94 to +0.85, and now to +0.66. The exceptions are the dollar (seekingalpha.com/symbo...) and the yen (seekingalpha.com/symbo...) remain at a correlations of about -0.5 and +0.6, respectively over the last 3 quarters.

    However, I disagree with your premise that this indicates an environment of greater opportunity. It does mean that hedges created with older data will work once again. It also means that diversified portfolios based upon older data will have lower volatility. But opportunities exist at all times - provided one is on the correct side of the trade... As for profiting by playing the "trends"; this has little to do with correlation between asset classes; rather trending is dependent (defined by) a high degree of autocorrelation - correlation in returns from one day to the next. And since 1950, daily autocorrelation (an index of trending) has averaged +0.08 for the S&P500; it is now at -0.03, indicating less trending than average. It is slightly higher than the TTM average of -0.06, but not by much.
    Feb 7, 2011. 07:32 AM | 2 Likes Like |Link to Comment
  • *Blue Plate Specials* - Dec. 28 [View instapost]
    Nasdaqczar: Instead of "re-commenting", I wrote an article that I hope will be of some use - on covered call strategies and what they have done historically (Dec 30th):

    seekingalpha.com/artic...
    Jan 6, 2011. 07:24 AM | 2 Likes Like |Link to Comment
  • Gold Price Rising Faster in Euros Than in Dollars [View article]
    A very clearly written and thought-provoking piece. We are often too provincial in our thinking and you have reminded us of the "relativity" of currency and commodity pricing.
    Jan 3, 2011. 09:59 PM | 1 Like Like |Link to Comment
  • What Is the VIX Telling You to Do Now? [View article]
    I agree with your comment that a low VIX is associated with a fall in the market over the subsequent near-term. But this correlation is not robust. The VIX (actually, the daily change in the VIX) is strongly (negatively) correlated (r=0.74) with concurrent (that day's) market returns, but only weakly correlated (now positively, in the range of r=0.08) with future returns (next trading day).

    I totallly disagree with your statement that the VIX goes up when traders buy puts and down when they buy calls. In fact, it goes up when index option buying pressue exceeds selling pressure - either for puts or calls, like any other asset dependent on supply and demand. And, the VIX drops when index option selling pressure (again, puts and calls) exceeds buying pressure.

    That said, the volume of index puts almost always exceeds the volume of index calls (unlike equity options where calls exceed puts), so when the VIX rises, it means more puts are being bought (occurs in a falling market). And, when the VIX falls it means more puts are being sold than bought (occurs in a rising market).
    Dec 29, 2010. 04:15 PM | 1 Like Like |Link to Comment
  • *Blue Plate Specials* - Dec. 28 [View instapost]
    You say that a "steadily rising market is a perfect environment" for covered call writing. I would suggest that it is a perfect environment for leaving your long equity positions uncovered. A range-bound market is the perfect environment for covered calls. But I'm sure that's what you meant.
    Dec 28, 2010. 04:36 PM | 1 Like Like |Link to Comment
  • Global Macro Notes: Massive Topping Process at Work [View article]
    An interesting, well-written article. One quick observation - each of the charts appear to go in the same direction over the time frames depicted (the underlying assets are correlated). Even bonds, more or less. It's frightening when asset classes become correlated - when portfolio diversity becomes difficult to maintain.
    Nov 24, 2010. 05:09 PM | 1 Like Like |Link to Comment
  • Deflation and the Dollar [View article]
    The spot gold price was 1360 at 2am EST on the 23rd. When New York NYMEX opened just after 8am EST, the spot price rose and continued to rise to 1380. So, gold rose as expected with the conflict.
    Nov 24, 2010. 08:46 AM | 4 Likes Like |Link to Comment
  • Friday's Put-Call Index Ratio Hits Yearly Low, But Is It a Valuable Trading Signal? [View article]
    Iggy:

    A valid point. Running averages (for instance, SMA or EWMA) are useful for measures like the PCR that may vary in a non-random fashion (i.e. trend) over time.

    In the article we used the 20d SMA (which was 1.29 at Friday's close) and identified 0.77 as about 2.5 standard deviations below the SMA. If you are inclined to use longer periods for your SMAs, the 63d (one quarter) SMA was 1.37 and the 126d SMA (half year) was 1.43. But the message is the same, the PCR on Friday was quite unusual.

    As you suggest, the PCR was below its 20d SMA each day last week except Nov 17. It was below its 63d and 126d SMA for each of the 5 days. And the one time the PCR was slightly higher than its 20d SMA (Nov 17), the SPX rose big time the following day.

    But sadly over the long run (since 2003), the correlation of PCR with future returns has been very low. The correlation coefficient is only 0.03 using the number of standard deviations from its 20d SMA, and 0.02 for the raw daily PCR.

    As a comparison, even the SPX return one day has a stronger association with the next day's return (coefficient = 0.12) due to its mild degree of mean reversion.

    A long winded answer to a short comment.
    Nov 22, 2010. 07:38 AM | Likes Like |Link to Comment
  • Any Value in Correlation Between the Dollar and Equity Markets? [View article]
    I am not certain that I understand your question, but I will give it a shot and you can tell me if this is what you were asking.

    Correlations are interesting but, unless they can be used to guide trading decisions, they are of academic value alone. It is clear that the daily changes in USDX and AUD are correlated with SPY - currently quite negative for USDX and quite positive for AUD.

    To be of trading value, the current return of one index must predict the future return of another (or of itself). I was trying to say that USDX does have some very mild value in predicting the future direction of SPY. The data table tests the lag, across the top row, and the first column tests the number of days going back. When X=1 there is no lag.

    I apologize in advance for the math, but the intersection of X=1 and Y=19 depicts the correlation of the USDX 19 day prior return (price time t=0 / price t-19 - 1) versus the SPY 22 day future return (price time t+22 / price t=0 -1). The fact that the correlation is best (most negative) at X=1 implies that the best information is obtained when there is no lag.

    And, I don't know whether the underlying cause of all of this is commodity prices or not, since correlations do not imply causation. I would consider the carry trade as one causative mechanism - but will leave this for someone else to figure out.
    Nov 15, 2010. 01:11 PM | 1 Like Like |Link to Comment
  • Any Value in Correlation Between the Dollar and Equity Markets? [View article]
    Djvu is absolutely right on target. AUD has a highly positive correlation with SPY. Using daily returns and a 63-day prior period, the correlation coefficient has been about 0.65 or so for the last 2 months; versus -0.35 or so for the Euro.

    What is very interesting is that the correlation between AUD and SPY was all over the place between 1999 and 2008 - sometimes positive, sometimes negative. In August 2008 the correlation between AUD and SPY was quite negative (-0.45). Immediately following the Lehman collapse in September, however, the correlation rose dramatically to its current highly positive levels. At the same time, the correlation between oil and the equity markets rose as well - identical to AUD. And as expected, the correlation between AUD and oil is quite positive.
    Nov 14, 2010. 09:30 PM | 1 Like Like |Link to Comment
  • Sentiment Overview: Investors Go All In [View article]
    Steve: I agree, it is sad. But, if these indices are highly negatively correlated, they would provide great information for decision making. This is essentially what I asked the author to comment on - so you and I are on the same page.

    My guess is that the correlations are indeed negative, but not significantly enough so to be of any trading value.
    Nov 14, 2010. 03:18 PM | Likes Like |Link to Comment
  • Sentiment Overview: Investors Go All In [View article]
    These are all very interesting indices. And we can all eyeball the charts and come up with a ballpark idea of whether the indices are predictive for future equity market direction, or not. But what is missing from your analysis is an objective assessment of the relationship between the indices and future market returns - even something as simple as a correlation between each index and SP500 return over the subsequent "x" number of months. Can you provide such an assessment for the readers?
    Nov 14, 2010. 02:55 PM | 1 Like Like |Link to Comment
  • Introducing My ETF Portfolio [View article]
    Seems like a simple system; momentum-based, that will work well when momentum is high (e.g. Oct - Dec 2008) but not so well when momentum is low (e.g. Mar - May 2009) . I am guessing you've already checked the performance of your system vis a vis Ken French's data - but if you haven't it would be very interesting to see how the returns correlate with momentum as defined by French's Momentum Factor. Let us know!

    (see mba.tuck.dartmouth.edu...).
    Nov 11, 2010. 05:29 PM | 1 Like Like |Link to Comment
  • Still Interested in Getting Physical With Gold? [View article]
    Interesting information, especially the loss of value by disrupting the "chain of custody". Makes a lot of sense. And, I'm imagining that, to take ownership, one would want to insure against loss (e.g. theft)- and I'm guessing a nonprofessional would pay rates equal to private jewelry- usually 1-2%/yr or more?
    Nov 5, 2010. 09:19 AM | 5 Likes Like |Link to Comment
COMMENTS STATS
26 Comments
39 Likes