cancel
Showing results for 
Search instead for 
Did you mean: 
     
Highlighted
Participant ○○○

PCI: a glimpse at the active management behind this income generating machine

I've been tracking NII, UNII & distributions for some of the PIMCO CEFs for a few years to try to get a sense of whether earnings were showing signs of coverage problems. It was evident that money was moving in and out of UNII is ways that could not be directly related to earnings, regular or EOY distributions. We've had some discussions about this in the past. The first chart shows the monthly UNII data reported by PIMCO converted into months of income stored in UNII (Y axis). 


Untitled 2.png

In 2015 PDI had about 8 months of extra income stored in UNII, and made a large EOY distribution. In mid 2016 things got weird. PDI, PCI and some others showed a large jump in UNII, in excess of what could reasonably have been normal income in one month. Presumably, some holding was converted into distributable income. In 2017 UNII drifted down fairly steadily, going negative for a few funds, with PCI being particularly challenged. Things improved in 2018, with another big UNII jump in April followed by a steady increase and then an EOY special. This year, again in July there was a 'too' big jump in UNII. So it seems likely that the recent UNII jump reflects something other than normal income.

In years past, Dick tried to educate us on how the movement of money in and out of swaps etc. can convert unrealized gains to income, or pull income out of the UNII pool into such positions. It might be possible to look back over the holdings from the relevant timeframe and work out what happened, with time resolution limited by quarterly holdings reports with 2-3 month lags (possible for someone with the requisite knowledge - ie not me). 


Over at the Fido site Dick posted on the UNII jump in July and its possible linkage to the NAV drop. His point was that distributable income can be stored in NAV, and that paying attention to unrealized gains might be informative in this regard. It got me curious to take a more detailed look at how month to month change in NAV relate to the monthly PIMCO NII and UNII data. Here are the year to date for PCI:


2019 ytd

income

distribution

UNII

NII excess / shortfall

D UNII

D NAV

jan

0,11

0,164063

0,28

-0,054

-0,05

0,49

feb

0,11

0,164063

0,22

-0,054

-0,06

0,19

mar

0,19

0,164063

0,24

0,026

0,02

-0,06

apr

0,16

0,164063

0,24

-0,004

0,00

0,24

may

0,24

0,164063

0,32

0,076

0,08

-0,07

june

0,23

0,164063

0,39

0,066

0,07

0,39

july

0,39

0,174

0,57

0,216

0,18

-0,08

 

And in chart form to make it easier to see the trends. Y axis is dollars; D UNII is month to month change in UNII; D NAV is month to month change in NAV (sorry the D shows as D below the chart).

Untitled 3.png

Observations:

January:  NII (income, light blue) was less than the distribution (orange). The NII shortfall was approximately equal to the change in UNII. NAV gains were 3x the monthly distribution.

February:  NII was less than the distribution. NII shortfall was approximately equal to the change in UNII (visible in the shorter grey UNII bar). Monthly NAV gains again exceeded the monthly distribution. Nothing special so far - the UNII cushion is still >1 monthly distribution and there were lots of NAV gains available to cover the income shortfall if needed.

March: Now the fun starts. NII increased - and NAV was down by an amount similar to the jump in NII. There does not need to be a cause effect relationship here, but it is certainly possible that some NAV gains were realized into NII. (it goes without saying that NAV goes down or up for other reasons, but I’ll say it anyway so you all don’t have to).

April: Nothing noteworthy except that NII stays high, covering the distribution. Monthly NAV gains were in excess of the distribution.

May: NAV was down this month, and NII increased by a similar amount. Given that NII was already covering the distribution, this move presumably reflects repositioning for other reasons - that also happened to throw off excess NII.  

June: Steady as she goes. NAV increased quite a bit. 


July: Repeat of May – NAV was down, but this time NII increased by considerably more than the change in NAV. Again, repositioning for reasons other than income shortfall. The fact that the NAV changes show up as income might indicate something about the source (to someone who understands how the different types of security are taxed).

 

We cannot know what is happening in real time, but the month to month look-back provides a glimpse of the active management behind the income generation potential of these funds. It doesn’t reveal how they do it, just that they are doing it. Once again, these CEFs can't do what they do without risk. They are not for everyone, and you should know your own risk tolerance. 

PS: let's not revive the whole 'how they misjudged the rate curve discussion and how could they be so dumb as to own Argentina debt'. Whether they blew it big-time will become clear over the next few months, when we can look back to see if (1) they continue to generate the income and (2) whether they do so at the cost of capital preservation. 

 

 

28 Replies
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine

A lot of effort on your part; thank you for passing the info on!  As you note, Dick's commentary is sorely missed here ; so yours is all the more appreciated!

Highlighted
Contributor ○○

Re: PCI: a glimpse at the active management behind this income generating machine

 

Heckuva post, kep.     Kudos.

0 Kudos
Highlighted
Explorer ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

Just a great post. Thank you for the generous use of your time and knowledge in turning this out. Much appreciated.

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine

 

keppelbay...

 

Thanks for info and effort.

R48

 

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine

Much appreciated, Kepp

0 Kudos
Highlighted
Participant ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

Thanks so much, keppelbay!  This kind of effort and generosity are what makes this forum worthwhile.

0 Kudos
Highlighted
Follower ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

Keppelbay, this is the most comprehensive post I have ever seen on a financial blog. THANK YOU. 

0 Kudos
Highlighted
Contributor ○

Re: PCI: a glimpse at the active management behind this income generating machine

Echoing other thank yous and wonderful posts.

I have a lot of PIMCO in my CEF holdings and am continuing to trust that they know what they're doing. 

Not that I'm adding to them at the moment, outside of reinvested dividends, but I'm not selling either. 

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine

Kep,

Thanks for your valuable post.  I tried to "kudo" it, but for some reason the button isn't working.

I'm wondering what importance, if any, you place on the existence of a CEF price premium / discount when buying & selling these products.  TIA.

N.

Highlighted
Participant ○○○

Re: PCI: a glimpse at the active management behind this income generating machine


@norbertc wrote:

...I'm wondering what importance, if any, you place on the existence of a CEF price premium / discount when buying & selling these products.  TIA.

N.


Hi Norbert

A few thoughts:

(1) I’m agnostic about premium vs discount per se. I’m concerned about relative valuation. How well does the fund perform compared to similar CEFs and to OEFs that invest in the same space.  I’m concerned about risk adjusted returns. The fixed income CEFs tend to have higher volatility than PIMIX, so I expect that to be coupled with higher returns. If it is not, I’m not being paid to take the added risk that comes with use of higher leverage and market price fluctuation.

 

The market prices CEFs according to their performance relative to comparable alternatives. So, it can make sense that a fund that delivers higher earned distributions, coupled with reasonable NAV performance might trade at a higher valuation, relative to NAV (this can mean a smaller discount or a larger premium relative to its peer group). There can be good reasons for a fund to trade at a discount to NAV. Trading at a premium implies that the owners have some degree of confidence that the funds management can deliver superior results.

image.png1=PCI, 2= PDI, 3 = PIMIX.

(2) Fixed income CEFs carry equity-like risk. They are not bonds for dampening portfolio volatility.

CEFs are sensitive to market mood. Downside swings can be fast and steep when the market is in risk-off mode. For this reason, I view my CEF allocation as equity-like and I tend to compare risk-adjusted return to SPY.image.png

1=PCI, 2= PDI, 3 = PIMIX.

Volatility tends to be lower than the equity index. Viewing CEFs as equity-like rather than bond-like helps me to get perspective on their role in my portfolio, and helps me manage my reaction to volatility. When I have confidence in fund management I’m more tolerant of price swings.

(3) Volatility is opportunity, but risk tolerance is individual

Downside swings eventually present good buying opportunities.  I have the impression that funds at higher premium levels are more prone to sharp downswings, and those a discounts may be protected a bit – because of investor psychology. I have not run the numbers to see if this is real. I may be imagining it. I really only pay attention to premium and discount because I think it affects other folks behaviour.

I’m fairly tolerant of market price fluctuation, when I have confidence that management will continue to deliver the high earned income with reasonable NAV volatility. I find that I can hold through a downswing without being tempted to sell into weakness (as long as NAV holds), because I’m confident that the market will put a floor under the CEFs when yields become compelling. I happily collect the distributions to use when the time is right.

When things move the other way, I’m not inclined to chase them up. I know that if I’m patient another bout of volatility will come along and offer better prices.  So I wait.  I may sell into strength - watching price relative to resistance levels and if the premium seems to be getting out of line. If prices hit resistance when NAV is dropping I might think about selling, but if they go up along with NAV, I’m more patient to see what develops. 

A bit off topic – as it has nothing directly to do with premium/discount: We’ve had an interesting discussion over at the Fido site on using weekly MACD to identify opportunities when the risk of selling or buying are lower, compared with buy and hold strategies. Capecod pointed out that his use of the MACD indicator allows him to get out when there is a risk of bigger downside moves, and to buy back when the risk of further big downside is reduced. He uses these in conjunction with other information/views to time trades. I think some of his comments were posted here in the usual cc tracker thread.  After modelling Dick’s MACD trading and seeing what he has accomplished relative to buy and hold, I’m studying how to use the indicator and may try it on a small scale.  Let me know if you want me to provide more detail on this.

(sorry - it got a bit long-winded...)

 

 

Highlighted
Participant ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

Hi keppelbay,

More details? Yes, please...

Brian

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine


@keppelbay wrote:

 After modelling Dick’s MACD trading and seeing what he has accomplished relative to buy and hold, I’m studying how to use the indicator and may try it on a small scale.  Let me know if you want me to provide more detail on this.

I'd certainly be happy to see additional information on this if you were so inclined.  I'd like to do more with formalizing my trading and all information helps!  Thanks in advance.

0 Kudos
Highlighted
Contributor ○

Re: PCI: a glimpse at the active management behind this income generating machine

Many look at PCI and other CEFs as belonging in the stock side of their PFs.  I think it is important to note  that it only has a 61% correlation with the Stock market, and several other diversifying stats if you view the metrics tab on this PV link. So while CEFs as most agree should be considered as the stock side, that shouldn't be reason to not give them a closer look as a volatility hedge. PCI for one adds both diversity and good returns.

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine


@archer wrote:

Many look at PCI and other CEFs as belonging in the stock side of their PFs.  I think it is important to note  that it only has a 61% correlation with the Stock market, and several other diversifying stats if you view the metrics tab on this PV link. So while CEFs as most agree should be considered as the stock side, that shouldn't be reason to not give them a closer look as a volatility hedge. PCI for one adds both diversity and good returns.


Correlation r tells only about directionality of the move, not the extent of it.

If you look at Relative_SD, i.e. SD_pci/SD_sp500, that is 0.81. Relative_SD is relative volatility.

Relative_SD = beta/r

 

YBB
0 Kudos
Highlighted
Participant ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

"Correlation r tells only about directionality of the move, not the extent of it."

If a Pearson r with a linear relationship between variables, both direction and strength of relationship can be measured. 

0 Kudos
Highlighted
Contributor ○

Re: PCI: a glimpse at the active management behind this income generating machine

Yogi, Thanks for explaining that. I was focusing more on direction as a diversifier. Your explanation made me realise that I have always assumed that you can take 2 SDs and average them for their combined SD. That would only hold true if they both moved in the same direction all the time.

0 Kudos
Highlighted
Contributor ○

Re: PCI: a glimpse at the active management behind this income generating machine

Thanks, Kep, for the OP.  Excellent observations section.

yes, I too would appreciate you sharing with us what you formalize in terms of trading CEFs.

0 Kudos
Highlighted
Frequent Contributor

Re: PCI: a glimpse at the active management behind this income generating machine


@archer wrote:

Yogi, Thanks for explaining that. I was focusing more on direction as a diversifier. Your explanation made me realise that I have always assumed that you can take 2 SDs and average them for their combined SD. That would only hold true if they both moved in the same direction all the time.


SD1 and SD2 are combined as,

SD = SQRT [ SD1^2 + SD2^2 + 2*r*SD1*SD2 ].

So, if r =1 [i.e. move in the same direction], then SD = SD1 + SD2,

and if r = -1 [i.e. move in opposite direction], then SD = | SD1 - SD2 |

But for a general r, linear combo of SDs are not valid and that explains the diversification benefit [i.e. Portfolio SD < Weighted average of component SDs]. 

PV does find SD of portfolio correctly.

YBB
0 Kudos
Highlighted
Participant ○○○

Re: PCI: a glimpse at the active management behind this income generating machine

@Anitya@racqueteer , @BrianG 

Hi folks,

The modeling CEF trading using MACD thread at Fido started out as an exercise to compare using weekly MACD signals to trade CEFs with a buy and hold strategy. It was based on Capecod’s explanation of his trading strategy.

First: his post summarizing what he actually did:

Hi keppel...thanks again for all this interesting work.  Here's the answer on my trades --- starting just a year ago because 1) I'm lazy, 2) I don't remember all the narratives and macro stuff going on earlier in sufficient detail, and 3) see my later observations below.

My trades in IRA (note I was not very heavily involved at the sell signal):

9/14 and 9/20/18 SOLD 6600 PCI at average 24.20  (aside: sell signal actually came the week before the big swoon)

Then BUYS:

11/28 through 12/3 BOT 5400 PCI at average 22.80

12/19 through 12/31 BOT 3  1000 SHARE LOTS:  19.88, 20.54, 20.72

1/4/19 BOT 2000 21.45  (then slowly built a full position over the next few weeks, accelerating with the subsequent buy signal)

Observations:

  1. Clearly I didn't (in retrospect) optimize or avoid the greatest possible loss/negative mark in this particular PCI downtrend......I did better with some other CEFs.....but the point is that even in this poorly played move SOME realized and/or unrealized losses were avoided, I got to enter the next uptrend with more wealth/shares each at a lower cost basis / higher distribution yieid.
  2. In the thread title, "modeling CEF trading using MACD," you have implied something most folks do automatically but that I need to clarify.  My use of MACDs is not a trading system.  It's very unlikely that I would ever sell on a turn signal and never buy a smaller or larger position back until there is an MACD buy signal.  That isn't how I think about it or use it.
  3. What I have found is that weekly MACDs (and no doubt some other oscillators) provide buy and sell signals at price levels where the opportunity cost of getting out or the risk of loss from being long are very very very small. That is a very valuable bit of information.  Here IS how I think/thought about PCI in mid-Sept 2018.....
  4. The weekly MACD has crossed down, so there is very little immediate future gain that might be sacrificed by selling my position.
  5. This sell signal presents me with an OPPORTUNITY SET.  I believe I have an excellent chance/opportunity to replace the position I just sold at a lower price or prices.  I DON'T KNOW how far PCI might fall, but now I have cash and a statistically very high probability of buying PCI at a lower price.  THAT IS ALL THE SIGNAL GIVES ME.  The extent to which I am advantaged by getting out on the sell signal depends on a different and unrelated set of skills and analysis and patience.
  6. After the fact, it is clear that this sell signal provided a set of loss avoidance opportunities that spread across many weeks.  HOW I EXECUTE WITHIN THIS OPPORTUNITY SET will determine the value I'm able to extract from what the sell signal has provided me.  Here it turns out to have been $1 or $2 or $3 or even $4+.  (I did better with PTY...LOL!)
  7. So then...what was provided by the late-Jan-early-Feb 2019 BUY signal?  Same thing in mirror image: an opportunity set where the risk of buying/being long PCI is negligible and prices are likely to rise -- and here it could have been up to a couple bucks and some distributions, depending on how well I executed within the uptrend opportunity set....and I captured a lot more of this opportunity.
  8. So....buy and sell signals are similar in that they identify low risk opportunity sets, but that's all....and certainly one opportunity set signal --- for example, the buy described above --- does not typically coincide with the end of the opportunities identified/presented by the last sell signal.

Regards, Dick

_____________________________________________________

My follow up:

Here are the numbers comparing Dick’s actual trade data (center group) with the pure MACD signal trade (left) and the buy & hold (right). I used the weekly MACD crosses to dictate when to sell and buy. I chose prices in the midpoint for the week of the cross, but giving the trade the better price where it was ambiguous. The trades are marked by blue arrows on the chart (note this is a 2 year chart, but the numbers below cover 1 year, so ignore the 1st 2 signals):pastedImage_0.png

 

 

pastedImage_1.png

 

 

 

All three series start with 6600 shares at 24.20, based on Dick’s actual trade (cc actual). In cc actual the first 5400 shares were bought back at his price; the next buy used up the remaining cash from the original sale at the average of the 3 prices given for the Dec buys (20.38). Net gain was almost 600 shares. Final sale / accounting for all 3 scenarios was at 23.80.

Summary: Dick did not wait for the MACD buy signal. He was early, but not wrong. His average buyback price was 22.20 - a bit better than the pure MACD signal trade at 22.40 (the first buy was higher, but the lower price of the second made up for that). The major difference between these outcomes was that by being early, he did not miss as many distributions, including the larger EOY special of 0.35. This added a bit over 5K, or 1/5 of the overall gain, compared to the pure MACD signal trade.

In this shorter (1 year ) time frame, the buy & hold did considerably worse then either trade style. But recall, that over the 2-year time frame, the difference was minor (note: this was in an earlier post, not copied here). The MACD trade was better, but not by all that much. The significant factor seems to be how long the trade takes us out of the distribution stream, vs the potential gain by capturing the price moves to increase share count. Evidently, trading around the short, sharp drops can be more beneficial than trading the long slow MACD driven moves in which price changes are not very large. These could keep us out for a longer time, without offering much opportunity for share count gain, as in 2017.

Comparing the pure MACD trade numbers with Dick's more aggressive approach is helping me to get a better understanding of what he means by calling the MACD signals an "opportunity set" that can suggest relatively safe timing for moves out or back in, if one hasn't made the move already. Working through this with real numbers has been most illuminating.

_____________________________________________________

Dick’s response:

Thanks...great stuff.  I'd just add a reminder and an extra datum..

  1. The point of my post on a different thread that got this going:  The avoidance of any realized or unrealized loss, just BEGINS the excess wealth creation process.  Whether it's a $6k or $16k or $2k realized or unrealized loss that's avoided, that amount increases the portfolio "start" value for the rest of one's investment horizon, and THROUGH TIME it can/likely will make VERY SIGNIFICANT additional contributions to terminal portfolio value vs. simply buying and holding.  Loss avoidance savings COMPOUND through time.
  2. Although it wasn't part of the initial discussion that was limited to loss avoidance, we shouldn't forget the benefit of being able to enter and build LONG positions with greater confidence (low likelihood of loss) using MACD buy signals.  Through any long position holding period, my return per invested dollar will be no different from that of the buy-and-hold investor --- BUT the identification of a very low risk upside opportunity set with a MACD buy signal gives me (and maybe others) the CONFIDENCE TO COMMIT MORE CAPITAL to the long position.   (While Keppel's latest spreadsheet follows 6600 shares I avoided some losses on, I ended up selling 11,600 shares at 23.80 a few days ago in mid-August.)

Regards, Dick

(my comment here: the doubling from 6600 to 11,600 involved adding new money - the gain from the trades was to 7196 shares – an increase of almost 10%).

______________________________________

Conclusion: trading around the volatility can add greatly to the total return – if done well. The weekly MACD signal may help to identify times when there is less risk to selling or to buying. But, you need more than just the MACD signal to get the most out of it. Probably not as easy as he makes it look…

 

 

 

Announcements