On September 11th, Alerian announced that following the close of business on September 18th, Boardwalk Pipeline Partners (BWP), Columbia Pipeline Partners (CPPL), and Vanguard Natural Resources (VNR) would be added to the Alerian MLP Index (AMZ). VNR and BWP outperformed the index on the day of the announcement, but CPPL did not. The “By the Numbers: Alerian Effect” Series dissects the data of how MLPs entering or exiting an Alerian index perform during the rebalancing cycle.
In Part 1 of our Alerian Effect series, we highlighted how on rebalancing announcement dates, names entering the Alerian MLP Infrastructure Index (AMZI) outperform the AMZI by 362 bps and names entering the Alerian MLP Index (AMZ) outperform the AMZ by 288 bps. In Part 2, we learned that on the rebalancing announcement date, names exiting the AMZI averaged 181 bps of underperformance versus the index and names exiting the AMZ averaged 162 bps of underperformance.
In Part 3, we wanted to evaluate over a longer time period to see if the entering MLP would—on average—be repriced by the market and continue trading at elevated post-announcement levels or if valuations would narrow back to pre-announcement levels. We decided to look at how entering MLPs performed relative to their respective index during the trading period from the announcement date until one trading day after the rebalancing date.
Even within Alerian, we had varying opinions of how MLPs should trade during this period. On one hand, proponents of fundamental and bottoms-up valuation claimed that businesses should be valued on their ability to make money and additions to indices shouldn’t change valuation. Under this belief, the one-day “Alerian Inclusion Effect” outperformance should narrow. On the other hand, if a company’s valuation is based on how much someone is willing to pay for it, and more people are interested in buying the company, then the increased level of fund flows should shift and reprice valuations to higher levels.
Let’s jump into the study.
The “X”s in the below chart represent how an entering MLP performed on the day of announcement. This is the same data set as Part 1 of our series, but updated for data points since September 2014. Between the AMZ and AMZI, the average day-of-announcement outperformance is 220 bps (previous average: 316 bps). AMZ-only represents 180 bps outperformance (previous 288 bps) and AMZI-only represents 318 bps of outperformance (previous 362 bps). It seems the overall Alerian Inclusion Effect is not as magnified as it once was, which could be a function of stakeholders becoming more familiar with the Alerian methodology and doing their own due diligence to predict which names will enter the index. Being added to the AMZI still results in higher outperformance, as there is more AUM linked to the AMZI versus the AMZ.
The diamonds in the below chart represent how an entering MLP performs from the day of announcement through close of one trading day after the rebalancing date. For quarterly rebalancings, this represents seven trading days of performance. For special rebalancings, this usually represents three trading days but depends on when merger approvals are announced. Between the AMZ and AMZI, the average outperformance versus the index during this period is 124 bps. Fundamental and bottoms-up valuation proponents could argue that markets have corrected with an outperformance narrowing from 220 bps on announcement date to 124 bps the day after rebalancing. Not so fast though.
When comparing names entering the AMZ versus AMZI, a completely different story unfolds. On average, names entering the AMZ end up underperforming the index by 15 bps by the end of the time period. Names entering the AMZI, on the other hand, outperform by 465 bps! This left us pretty confused, so we broke the data down further.
We decided to look at how many names outperformed the index on announcement date and continued to outperform the index through the day after rebalancing (“Outperform, Outperform”), how many names outperformed the index on announcement date but ended up underperforming the index through the day after rebalancing (“Outperform, Underperform”), and so on for names initially underperforming.
On an absolute basis by count, 75% of the time, names that enter the AMZI will outperform on the announcement date and that outperformance will magnify through the day after rebalancing (“Outperform, Outperform”). The data seems to be less conclusive for names entering the AMZ, as it was more or less equally as likely for the name to be in one of the four scenarios.
To be thorough, we decided to look at one more variable: quarterly versus special rebalancings. Here again, in both types of rebalancings, names entering the AMZ narrowed in outperformance whereas names entering the AMZI widened in outperformance. But the outperformance was much more magnified for AMZI special rebalancings to show an outperformance of 838 bps versus the index. One possible explanation for the magnified AMZI outperformance is that the time period from announcement date to rebalancing date is much shorter for special rebalancings versus quarterly rebalancings, so it may not be enough time for the market to readjust. But that still doesn’t explain why the AMZ doesn’t perform similarly.
To recap :
- Names entering the AMZI are likely to outperform the day Alerian announces they will be added and are likely to continue outperforming through the day the index is rebalanced. This effect is further magnified for AMZI special rebalancings.
- Names entering the AMZ are likely to outperform the day Alerian announces they will be added but that outperformance narrows through the day the index is rebalanced, in fact below pre-announcement levels. This effect is somewhat muted for AMZ special rebalancings.
Even though these are the results, we’re still perplexed why the results are opposite for the two indices. Yes, AMZI has more AUM tracking it. But if anything, the effect for AMZ should just be a muted version of the AMZI, not the opposite effect. I’m curious to hear your thoughts at [email protected].
In Part 4, we’ll apply this same time period study to names exiting the two indices. (Sneak preview: the results differ from this study on entering names.)