John Petersen's  Instablog

Regular readers know that I spend a lot of time focusing on supply and demand issues when I discuss Axion Power International (AXPW.OB) but most don't understand why because many of my theories are based on recollections of other clients who experienced similar problems during a time when I didn't pay as much attention. Since I predicted last summer that Axion would reach an inflection point by early fall, this seems like an opportune time to explain what I've been tracking and why I think it's relevant.

Sometime last year HTL drew my attention with his discussions of the daily short reports published by FINRA. When I looked at the numbers, they seemed to mesh well with recurring SEC reports from Quercus and Special Situations. So I decided to start tracking the FINRA data more closely to see if it might be useful for my purposes. HTL was a priceless research tool in that regard because he was able to give me the daily data all the way back to April 1, 2010, which was about the time the heavy selling started. The following table summarizes the FINRA total volume and short data from April 1, 2010 through October 5, 2012.

(click to enlarge)

On the supply side, I've principally paid attention to seven blocks of shares. Two were legacy holders who owned stock before the December 2009 private placement and the other five were holders or classes of holders that bought in 2009. The following table identifies the blocks.

(click to enlarge)

The thing that makes each of the blocks different from other shares is that they were sold in private placements. Under SEC rules stock that's sold privately is classified as "restricted" until it is sold by the original purchaser. So even if a resale is registered, every resale requires an opinion of counsel that the resale is legal. Even under the best of conditions, that process can't be completed in T+3 so every sale of stock by a private placement purchaser must give rise to a reported short sale.

My last table is part reported fact and part educated conjecture. It takes the annual FINRA short data and matches it up to the time periods when the seven blocks were selling, or at least presumed to be selling. Some holders like Quercus and Blackrock reported their transactions for a time and then fell off the radar when their holdings hit certain levels. Others like the Winner Estate and Mantuck Hill have been invisible all the way along, but the timing seems reasonable in light of the available facts. The little yellow arrows indicate that the timing of likely sales is fuzzier than it is for other holders.

(click to enlarge)

In my mind the most important data point is the yellow balancing entry, which represents reported FINRA shorts that can't be allocated to other holders. For 2010 and 2011, I believe that line represents sales by small 2009 investors and legacy pre-2009 investors. For 2012, it would also include buyers in the February placement that sold immediately after the offering closed who didn't get their stock into electronic form for a week or two after the closing date.

If you assume that there were no sales by legacy pre-2009 investors over the last three years, total sales by 2009 Investors would have been 17.7% of their purchases in the first nine months after closing and 34% in the following year. Those percentages of tie pretty well to patterns I've seen in the past. Once the 2012 investors received their shares in electronic form, they wouldn't be expected to impact short reporting.

When I pull all the numbers together, the only conclusion I can reach is that even under a worst case scenario we're within a million shares of the bottom of the willing sellers barrel. At this point I don't see any remaining 2009 buyers as likely sellers because they've already held on through two years and nine months of misery. I think the same logic holds for the February purchasers. It may have made sense to filp for a modest gain during the first month or two after the offering, but the decision dynamic changes once you've held a stock for nine months.

The bottom line is that I can't see any other large blocks of shares that can flow into the market as the stock starts to perform. They've already been sold. And until the holders who bought those shares over the last two and a half years see an irresistible price, I think we're on the cusp of a major supply and demand inflection because the blocks that accounted for about 60% of sell-side activity are out of stock.

Disclosure: I am long AXPW.OB.

During the Dynamic Charge Acceptance workshop that preceded the European Lead Battery Conference we received a handout prepared by CSIRO that summarized micro-hybrid vehicle offerings in Japan. I was quite surprised at the number of different vehicles that were being offered with stop-start as either standard equipment or an option.

The following table summarizes the number of stop-start vehicle offerings from each of the Japanese manufacturers, and breaks the total down by standard equipment or optional equipment.

The three page handout from CSIRO has been uploaded to my Dropbox. - bit.ly/Qodrmp

Disclosure: I am long AXPW.OB.

On Thursday morning I asked Enders Dickenson, Director of R&D for Axion Power International (AXPW.OB) if I could get a copy of Axion's ELBC 13 presentation directly from him instead of waiting for the conference sponsor to send out the full proceedings "in a week or two." While I'm normally patient enough to wait for things to go through channels, the sponsor took a couple months to get the proceedings out in 2010 so I figured there was no harm in asking. My copy of the 37 slide presentation arrived last night and it's been uploaded to my Dropbox.

bit.ly/UzaLnm

It's my understanding that much of the material in the ELBC 13 presentation was covered at the annual meeting, but slides like these frequently merit more study and thought than you can have in a live presentation.

While I don't want to drill down too deeply in my discussion of the technical issues because many readers know more about the subject matter than I do, I will highlight several slides that presented data and information I hadn't seen before and consider important.

Slide 5 shows a new emphasis on two key terms that Enders focused on heavily in his remarks. The first term, crystals, focuses on the sulfation problem and zeros in on the idea that small lead-sulfate crystals are normal and beneficial in a conventional lead-acid battery but that large insoluble crystals are the primary reason for loss of dynamic charge acceptance. The second term, concave down increasing or CDI, is the reason that strings of PbC batteries tend to self regulate because the weakest battery in the string is always charged first.

Slide 10 has an interesting graph that shows the up and down regulation performance from the PowerCube in New Castle.

Slide 15 is a copy of the DKE test I first saw in Istanbul. What makes it a little interesting is that Axion is finally taking some credit for doing the original development work on the testing protocol in conjunction with BMW. This ties in well with Eckhard Karden's description of the protocol as the "Axion-BMW" test protocol instead of the Ford-BMW protocol. In the 2010 Ford-BMW presentation at ELBC 12 in Istanbul, Axion only got a passing nod at the end of the presentation.

Slide 20 is very important because it shows what happens to a VRLA battery if you push the recharge delay after an engine off event from 10 seconds to two minutes. That kind of abuse drives the DCA down to minimal levels within a week. The reason the slide is important is that the most efficient way to run a micro-hybrid is to push the recharge timing out to the end of a drive interval when the car is decelerating and the engine is not being used to power the wheels. This is the heart and soul of so-called "regenerative braking" in stop-start and it's dreadfully hard on conventional VRLA.

Slide 26 shows what Axion is doing with the bench-top work under the SBIR grant. They'll be using a conventional flooded SLI battery for the cranking load only and a PbC for all the hotel loads. They'll also be increasing the system amperage from 100 to 150 amps. During the recharge intervals, 300 amp seconds of charge will be returned to the SLI battery (9%) and 3,000 amp seconds will be returned to the PbC (91%)

Slides 31 to 35 are dedicated to the difference between conventional lead-acid charging curves and PbC charging curves. Where lead-acid has a much lower voltage change from the beginning to the end of a charging cycle, the PbC has a wide voltage swing. More importantly, the voltage curve for conventional lead acid is slightly convex while the curve for the PbC is quite concave.

Slide 33 shows how concave charging curves tend to bring the batteries in a string into balance without active management of the individual batteries while Slide 34 shows how convex charging curves tend to leave the batteries out of balance.

Disclosure: I am long AXPW.OB.