When it comes to ETFs, the Retail HLDRS (RTH) can be a poor proxy for retail in general (too concentrated on Wal-Mart (WMT), Home Depot (HD) and Lowe’s (LOW).

On the subject of poorly-weighted ETF's, RTH is certainly up there, but the biotech holders (BBH), takes the cake in my opinion. One stock, Genentech (DNA) accounts for 43% of the market cap. Three stocks (DNA, Amgen (AMGN) and Gilead (GILD) account for 80%.

RTH, BBH, Semiconductors (SMH), Oil Service (OIH), Merrill Lynch Bank (RKH), Pharma (PPH) and a few other less notable names are a type of ETF known as a Holdr. They were created early in the ETF cycle, and if memory serves me correctly, first to market as sector-specific plays, and thus have become the de facto gold standard in several of these.

Some never really worked, like the Software Holdrs (SWF), as the sector itself is too disjointed. Some were just doomed sectors (remember B2B?) But all are poorly constructed.

Normal indices rejigger over time as new names emerge and old names get woodshedded. But Holdrs just took a fixed portfolio of names and quantities of those names and only adjust when a component gets taken over or de-listed. So take the Internet Holdrs (HHH) for example. It includes Real Networks (RNWK), thanks to the fact that it was popular in the mid 90's and is still a public company, but not Google (GOOG), which was years away from The IPO Heard Around The World.

Despite the flaws, OIH, SMH and RKH work fine. Perhaps it's just the nature of the oil service, semi-conductor, and regional bank sectors that correlation within these groups is fairly high. PPH, I suppose, is fine too, although trading anything Pharma-related is like watching paint dry. For biotech though I prefer iShares Biotech (IBB). It's not perfect in that it only includes Nasdaq names, so no DNA, for example. But it's a more even distribution, and the liquidity is there.

As for retail, I use XRT over RTH. The options have even gotten better as it's now multiply listed.