Owning individual MLPs superior to MLP ETFs

I’m a big fan of MLPs as an asset class and in particular of owning individual MLPs. I’ve posted on KMP, EPD, and on ETP previously. Not too long ago owning individual MLPs was the only way to invest in them. However, in the last couple of years there have been several new MLP ETFs announced such as AMLP and AMJ to name just two. Which leads to the question, what are the pros and cons of MLP ETFs vs individual MLP stocks?

A post this week by Richard Shaw on Seeking Alpha addressed this topic specifically. Richard has a great 8 part series on MLPs which I highly encourage you to read. It addresses most of the common MLP questions. In his post on owning MLP ETFs vs individual MLPs Richard summarizes his position as follows:

The short answer, in our opinion, is that if you have enough money to buy positions of sufficient size to make the commissions and the K-1 paperwork (see our article on MLP Investor taxation) and the increased portfolio complexity worthwhile, then we would tend to favor owning a basket of MLPs.

If convenience, and simplicity is the key, or you do not have enough assets to make multiple positions practical, then we would tend to favor an ETN or ETF.

My take is a little different. I think owning individual MLP names makes the greatest sense to long term income investors unless an investor places an inordinate amount of importance on ‘convenience’. Let me explain.

First, fees. Owning individual names is by far the lowest cost way to invest in MLPs. Consider even a small allocation to MLPs of $10K and owning a basket of ten MLPs. Through even a run of the mill discount broker an investor can create this basket of MLPs for $8 a trade, or $80. That’s 0.80%, one time. The MLP ETFs carry fees of 0.85% of assets annually. Over a long holding period this difference in fees can really add up. Also, remember as your investment increases in value the absolute dollar amount of fees you pay for the ETFs increases while trading your individual MLP positions is independent of the dollar value of your trades (for the majority of discount brokers). And these differences compound over time, as all fee differences do. Again, assuming a $10K initial portfolio,  a fiver year holding period, a 10% return for MLPs in general, 0.85% annual fees for the ETF, and an $80 up front set up fee for the individual basket yields some striking results. The 10% annualized MLP  index return is reduced to 9.07% for the MLP ETF vs 9.82% for the individual MLP portfolio. Fees matter a lot!

Second, and maybe the biggest difference between individual MLPs and their ETFs is taxation. This discussion applies to investments in taxable accounts. The majority of the distributions from individual MLPs are tax deferred – they are considered a return of capital and reduce your cost basis. For the big cap MLPs this seems to be on average about 85%, although for some the deferrals are north of 90%. MLP ETF taxation is a bit dicey and Richard is part 7 of his series does a nice job outlining the issues. For the MLP ETF AMJ the distributions are taxed as ordinary income. How big a difference does the deferral advantage make vs the ETF? Lets assume 85% of the distributions of an individual MLP basket are tax deferred and consider an investor in the 25% marginal tax rate. Starting with an index yield of 6%, the after tax distribution for the MLP ETF investor is 4.5% while for the individual MLP investor is 5.325%. That’s an 18.3% annual advantage for owning individual MLPs!

Combining these two effects together on a hypothetical 10% MLP index return composed of a 6% yield plus 4% distribution growth over a 5 year period yields the following differences. The return after fees and taxes for the MLP ETF would be 7.58% annualized while the return for the individual MLP portfolio would be 9.15% annualized. That’s an annualized return advantage of 21% for the inconvenience and difficulty of owning individual MLPs.

Do these advantages outweigh the convenience and simplicity of owning MLP ETFs? Every investor needs to answer this question for themselves but to me its pretty clear. Of course they do! If you currently pay someone to do your taxes then the convenience factor disappears. Its no more work for you and I doubt an investor would be charged more just for filling out some K-1s. If you do your taxes on your own, I think tax software like Turbotax makes the convenience factor minimal. In other words, I think the convenience argument for owning MLP ETF is over weighted. As far as simplicity there is no question than the decision to buy the MLP ETF is easier than buying a basket of MLPs. But how much harder is it to just buy the 10 largest holdings of the MLP ETF which represents 65% of the market cap of the ETF? Not much.

In my opinion, the only valid argument to owning the MLP ETF would be if owning the ETF was lower risk or somehow provided higher returns. I can’t see the higher return argument given the built in 21% annual advantage in returns I just showed but what about risk? I don’t subscribe to the modern portfolio theory argument that says higher volatility equates to greater risk but lets say that you do. I would make the argument that owning the largest cap MLP names individually is less risky, i.e. less volatile, than owning the MLP ETF. For example, over time KMP has been a lot less volatile than the MLP index. Also, I think the lower risk argument is even truer if you consider some of the structural risk of some of the ETFs products, like AMJ, which is an ETN – a credit product- that is subject to the counter party risk of the issuer.

In summary, owning individual MLP names far is far superior to owning MLP ETFs. I would also make many of the same arguments for owning individual stocks vs ETFs in general. What makes MLP ETFs different is the taxation issues and the unusual structures of the MLP ETFs themselves.