Seeking Alpha

Leif Peterson

View as an RSS Feed
View Leif Peterson's Comments BY TICKER:
Latest  |  Highest rated
  • Keryx Ferric Citrate (Zerenex) Could Potentially Save Hemodialysis Patients $1B Annually [View article]

    Great comments. Yesterday was approval day (ka-ching!) and someone sent me a list of short interest -- so a huge squeeze. No worries about the warning label - since ESRD and HD patients are under iron management anyhow. The warning is for smaller private clinicians who might prescribe ferric citrate without monitoring iron -- or knowing how to appropriately monitor iron outside of their specialty (need to be a nephrologist or endocrinologist).

    On Monday, I expect a pop in price, then needs to stay above 50-day MA of 15.84.
    Sep 6 03:03 PM | Likes Like |Link to Comment
  • Keryx Biopharma: Approval Delay Does Not Change Zerenex Prospects Or Differentiation [View article]
    Do you think Keryx' PDUFA for Zerenex will delay till Sept. 7, or that it most likely would be approved any time this summer?
    Jun 15 07:14 PM | Likes Like |Link to Comment
  • Keryx Ferric Citrate (Zerenex) Could Potentially Save Hemodialysis Patients $1B Annually [View article]
    Rustal, thanks. I also noticed the 3-month extension of the target approval date of Zerenex to Sept 7. They probably had a lot of updates they needed to make for the NDA, which are not new, but rather long-standing drug design/manufacturing changes based on pk (pharmacokinetics) data from the recently completed phase III trial. But I don't want to second-guess the specifics. The rule of thumb is that any change to a protocol requires a maximum of 3 months to provide adequate time for review by FDA and allowance for correspondence between FDA and Keryx for the review process. The steps would be: 1) Keryx submits an amendment to the NDA, 2) FDA reviews the amendment and gets back to Keryx with any comments and requests, 3) Keryx submits a response letter, 4) FDA reviews the response letter, 5) if Keryx addresses all concerns appropriately, the amendment is approved -- and I would assume Zerenex would be approved at this point as well. Otherwise, the chain mail and reviews continue until all concerns raised by the FDA are appropriately addressed by Keryx. Another main reason for the several-month review is because committees commonly meet once per month -- hence the delay.

    Typically, if a change to a protocol is minor, it can be expedited and approved quickly, but the latest news is that the FDA considered the amended items to warrant full board review by the FDA -- which takes more time. On a positive note, reviews are commonly done quickly, so FDA has probably already reviewed the changes needed to the NDA and is now preparing a response to Keryx. If Keryx responds appropriately and addresses all concerns, then in theory these issues can be hashed out in a week or two. Overall, when both parties truly believe in progressing commerce (business), there are no unnecessary obstacles placed in the way of rapid reviews. If there are any hangups, it would most likely be due to someone at FDA with significant instrument (QC) experience in industry who wants something done that Keryx (and other pharma houses) view as unreasonable -- or an e.g. statistical method that FDA didn't find acceptable. If there is, for example, a stickler at FDA who absolutely has to have his/her way on something that Keryx views as unreasonable, then it can severely degrade progress. At this point, conservatism would be an understatement.
    May 23 03:43 AM | 2 Likes Like |Link to Comment
  • Keryx Ferric Citrate (Zerenex) Could Potentially Save Hemodialysis Patients $1B Annually [View article]
    Agree, see above -- as usual you need earnings growth for a price increase, so one or more quarters for sales/revenue are required to properly value KERX.
    May 22 04:12 PM | 1 Like Like |Link to Comment
  • Keryx Ferric Citrate (Zerenex) Could Potentially Save Hemodialysis Patients $1B Annually [View article]
    Depends, there are results that still need to be published for their recent Phase III trial. Cursory evidence is suggesting positive findings. Also depends on drug pricing and revenue after first quarter of business. What you are suggesting is reasonable however. Thing to remember is that the number of additional patients with ESRD is increasing each year, so the potential for increasing earnings each quarter would be very likely.
    May 22 04:09 PM | 1 Like Like |Link to Comment
  • The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation [View article]
    @Alexkeywest, thanks for the comments. From a quantitative finance (QF) perspective, to understand the comments, you would need to dissect each sentence, one at a time:

    Let's start from the top.

    >There may a bias introduced by assuming the multivariate normal distribution, since most log-return distributions for stocks are either Laplace or logistic distributed, while some are stable distributed and some skew normal.

    First, perform probability distribution fitting of log-returns of the e.g. Dow 30 stocks or SP500 stocks for the last 2 or 5 years. There are 250 trading days per year, so 2 years of data would involve 500 data points, and 5 years would involve 1250 data points. Are we there? When done fitting normal distributions, your results will indicate that normal distributions are not the best fitting distribution -- so MVN is out. Great. What is next? Go through each sentence and perform validation analyses for any analytic comments. When done, you should obtain similar results and everything will fall into place.
    Dec 11 09:06 PM | Likes Like |Link to Comment
  • The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation [View article]
    There may a bias introduced by assuming the multivariate normal distribution, since most log-return distributions for stocks are either Laplace or logistic distributed, while some are stable distributed and some skew normal. Student's t is actually one of the least frequently best fitting distributions to log returns. There may also be a concern with how the correlations between assets are being set prior to MC simulation -- this was not discussed. There are also ways to incorporate stability (persistence) into the simulated returns -- known as fractal Brownian motion. If the Hurst exponent for an asset's log return is 0.6 over the last 5 years, then this can be burned in to the simulation. Other groups would suggest you GARCH filter all of the historical data (at least 5000 days for EOD data) before determining which probability distribution fits *each* return -- but it was observed that removing volatility clustering with GARCH filtering doesn't affect correlation that much -- mostly because volatility induces a widespread "market" effect - so there are few assets that can hide from the clustered arrival of bad news. Spearman rank correlation is another technique to employ with optimization, which overcomes outlier effects due to large spikes when volatility increases.

    It's a monumental task to simulate using MC, so the paper does nevertheless make a strong contribution to the literature. However, the issue is so multifactorial that many assumptions need to be worked out prior to performing the simulation. Copulas do help, and this is why the between-asset correlations assumed before MC should be discussed.
    Dec 8 11:50 AM | Likes Like |Link to Comment
  • Setting Constraints For Asset Classes And Individual Securities [View article]
    This is a nice article Lowell. It warrants noting, however, that Hoadley is based solely on EWMA. For a 2-year historical optimized portfolio (no shorting, dividends included, tangency), the wealth is $140k for these assets after an initial investment of $100k 2 years ago. You would be better off employing a dividend reinvestment program (DRIP) on good yielding assets (see, and not constraining your approach to EWMA only.

    FYI: For readers who may be interested, I contributed an SA article on increasing the stability of a hedging portfolio here:
    Jul 1 08:11 PM | Likes Like |Link to Comment
  • Dynamic Asset Allocation For Practitioners [View article]
    I think the approach using a 10-month SMA or (200-day) SMA based on 200 trading days per year essentially attempts to avoid noise and instability by decreasing the granularity or scale. I recently published an SA article which addresses asset weighting with (a) the Hurst exponent (2 - fractal dimension) and (b) right tail dominance as a way to avoid instability and perhaps decrease left tail risk for daily-based portfolios.

    If each asset in the 10-month SMA-based portfolio were considered to be an airliner, then the amount of daily/weekly instability experienced would be like having extreme wing buffet throughout the duration of a transcontinental flight -- and doing nothing about it. So much for flight stability optimization problems. An engineering approach to this problem would involve taking an FFT of the portfolio to identity the major components that are rattling around in the system, and then removing them or decreasing their weights. Another approach, mentioned in my SA article, is to remove noise from the portfolio's assets, recalculate correlation, then adjust the MinVar (Markowitzian) correlation-based weights by the Hurst exponent and right-tail dominance.
    Jun 4 09:15 AM | Likes Like |Link to Comment
  • Building A More Stable, Low-Risk Hedging Portfolio [View article]
    For "free" accessible explanations, search the internet for publications in e.g. pdf format on "covariance matrix shrinkage methods."
    Jun 3 09:00 AM | Likes Like |Link to Comment
  • SmithOnStocks Opines On Biotechnology Stocks, May 24, 2013 [View article]
    Larry, this is a great article. I must say that I needed to post the following comments before I got past the first several paragraphs. Regarding medical research and clinical trials, a common drawback is the lack of available time among many research clinicians and other key personnel during the typical ~10-year period required for bringing a product to market. In academic medicine (i.e., "get an NIH grant or get lost"), this is due in part to a variety of tradeoffs involving fierce competition for NIH grant support, running a large laboratory and not "seeing" patients for revenue generation, managing the safety for a clinical trial, managing the data, managing regulatory issues, managing progress and funding, managing personnel, etc. These overwhelming career-threatening tradeoffs and the ensuing shortage of available time to concentrate strongly on long-term marketing, and business development tend to force time-lagged goals such as data analyses to a lower-priority level.

    This already suffering management style become exacerbated by clinicians who downplay the complexities involved in other fields or specialties. Since other fields are not considered complex, the expectation is that collaborators can meet same-day deadlines on e.g. evenings, Saturdays, holidays, etc. Thus, normal research and management activities that otherwise should occur through regularly scheduled events become resolved at the last minute. Throw in your ~10-year marketing/business cycle outlined above, and there would be a low probability (<0.1) of being able to pull off such an undertaking.

    There are clinicians (researchers) who have been able to pull off all the above quite successfully. Such researchers have benefited from a personal management style that's amenable to orchestration of large projects with numerous participants. The majority of clinicians I have collaborated with over the last 15 years, however, have had little motivation to participate personally and thrust forward product development – so everything works out to be “academic.” This has led to a large field of medical researchers who have no interest in marketing and business, and who only pursue initial laboratory work, data analyses, publication, to acquire more research support.

    Personally, I think we would be much further in diagnostics and treatment of disease if there was a greater focus on commercialization at the individual level. Time is money, and in medical research there does not seem to be enough time among potential key principals to fulfill product commercialization needs to the extent possible(practical).
    May 27 12:10 PM | 1 Like Like |Link to Comment
  • Modern Portfolio Theory 2.0 - The Most Diversified Portfolio [View article]
    The technique seems interesting, and my guess is that the look back period each month was 60, 90, or 120 days. The Vanguard ETFs are noteworthy -- but I like their funds better than their ETFs. Rerunning an MPT portfolio using the given assets with dividends, MinVar (not tangency), no shorting, no transaction fees, component removal, a 90-day training period and 20-day testing period (rebalancing every 20 days, i.e., one trading month), the result can approach ~$300k over the 7 years. Since this summer (6/1,2012), the weights on this ~$300k portfolio are heavy on TLT (90%) or VTI (90%), with a little VNQ (~10%). Whereas for the first half of 2012 the weight is heavy on VGK (100%). Also, GLD has zero weights in 2012. After looking at the daily price return chart of TLT, it went up ~$30 in July 2011 right after the US debt ceiling with hit with corresponding downgrade -- hence the heavy weight.

    Periodic balancing essentially enriches on higher returns, so even MPT with MinVar will identify those assets and ensure somewhat reduced correlation when rebalancing. As most of the comments suggested, MPT and its variants pick good assets when looking back, and rebalancing catches short-term trends which may or may not last.
    Dec 1 12:53 PM | Likes Like |Link to Comment
  • Top 40 Dividend Stocks for 2011 Based on Concordance Analysis [View article]
    Great comment. In about a year or two, there will be sufficient tracking data for key statistics (fundamentals) of these assets to determine the fundamental parameters that associate with increasing or decreasing dividend and payout ratio. Right now the within-company quarterly changes occurring among the e.g. SP500 stocks from quarter to quarter is quite large -- so they are a little jumpy and require more data to increase the signal-to-noise ratio.
    Jun 2 10:55 AM | Likes Like |Link to Comment
  • Stocks: Why One More Major Correction Still Lies Ahead [View article]
    Great question, but see my comments logged a while ago today. It's mostly exploratory analysis, and not inferential in any way. For inference, you'd have to cut out the time series of price returns over the epochs of time considered, then probably perform an unsupervised analysis to see if indeed the 3 corrections are consistent (vs. noise). You could also correlate price returns over the considered phases using e.g. monthly bars to see if they agree. If you are serious, talk with the Elliott Wave group, since they may have ways to calibrate different time scales (periods) to determine if a pattern is consistently present over different epochs of time. Regarding forecasting, never believe what you hear, and rather, wait and try to determine what patterns are emerging based on your own research -- since it's difficult to know what's around the corner. Also, readers need to realize that the present report is not a prediction of the future, but rather a report of a pattern that has occurred repeatedly and historically. The author is not really speculating that another correction is in order -- since nobody knows what will happen tomorrow. Rather, he is suggesting that a third correction typically has occurred -- so heads up.
    May 27 12:28 AM | Likes Like |Link to Comment
  • Stocks: Why One More Major Correction Still Lies Ahead [View article]
    This is a very good article which is very timely. It's apparent that you have reported an interesting and consistent pattern, but if you invert your charts, the 1,2,3, become 1,3,5 in an Elliott Wave. After leg 5 there typically comes an A leg in the opposite direction (bull phase). The problem is, the path to step 5 (your 3) is typically longer and stronger -- so most Elliott wave theorists would likely not only suggest to batten down the hatches right now, but also: "get this ship in dry-dock." Thus, the next correction could be nasty.

    As far as the southern EU sovereign debt crisis, it's more instructive to learn from this mistake and envision the future for what lies ahead for the US economy. For relevance, when the US baby boomers (born post-WWII) are in their 80's and 90's, when they will really need medical help, the demand for national health care will be 5 times the current entitlements. However, the Medicare entitlement is now 33% of the US budget, so where is the extra 150% (~30% x 5) of Gov't spending going to come from? The current plan is not sustainable, so things are bound to get a little dicey. How will defense be funded if health care needs are a factor of 5 greater then?

    The present sovereign debt crisis in southern EU is the result of lack of stewardship, sustainability, and poor fiscal and monetary policy. Originally, the EU should have been constructed on a term basis, where each member nation is under evaluation for say x number of years. If their sovereign debt is or exceeds y per cent of their GDP, then they're out ("don't ever do business with a friend"). If they become insolvent and are at risk of defaulting on sovereign debt, then they're out. If they suffer from z downgrades, then they're out. Under such a pact, the term-based plan could also pre-program and simulate financial decoupling from less fortunate nations, so that e.g. small sovereigns whose GDP is less than XOM's market cap won't bring down the entire global market.

    At present, the economic crisis in southern EU isn't anything that the host populace are surprised about. These countries have faced 10-15% cuts in every ministry for the last 5 years, so recent events are no surprise. Thus, the news would be filled with events involving some less fortunate and more corrupt neighbor sovereign who simply reached a worse level of austerity and bank runs.

    For positions, I have been monitoring UUP, FXY, SH, PSQ, and think that GIS, MO, GLD, SLV may be useful hedges. The world jumps to the dollar during extreme crisis, so the bull dollar ETF UUP may be rewarding. It's hard to say what will play out over the next few weeks, but the dollar (UUP), short-term treasuries (FDLXX), and then Yen (FXY) are likely to grow in demand. Specifically, the following may be good plays over the summer: ACG, BIP, DUK, GLD, GIS, HCN, JNJ, KMP, LQD, MCD, MO, SH, NLY, NST, PAA, SO, TLT, WMT. Since this correction could be more severe than Sept 2008, LQD (investment-grade corporate bonds) could take a hit. Cash is King, so there is every reason to stay liquid for buying high-quality US corporations after their price gets beat down. Purchasing T and VZ after a price beat-down would be enjoyable to say the least.

    Regarding gold and inverse ETFs, the majority of these instruments are merely part of the equation at any one specific time. During worst-case scenarios, cash or short-term treasuries are king, since everything will likely be in the red. If things turn out very rosy over the summer, then it may be rewarding to stay the course and invest in uncorrelated high-yield assets.
    May 26 11:53 PM | Likes Like |Link to Comment