Hmm... The 8% annualized margin interest charged by some of the more expensive brokerages works out to about $2.11 for each of 365 calendar days or $3.08 for each of 250 trading day for every $10,000 borrowed (because 1.08 ~= 1.000308 ** 250 ~= 1.000211 ** 365). So, yeah, maybe it does matter. Shop around.
Besides offering glib advice to shop around, I will try to answer the original poster's question. Just please keep in mind that I don't have any kind of certification to advise anyone on this. This is just my layman's understanding.
One point that the original poster may be referring to is that the term "margin interest" can be misleading. In a normal margin account, you are not charged for using margin; you are charged for borrowing cash. So, if a brokerage charges a lot, say 8%, to borrow cash, that can easily exceed the leveraged ETF's fees, say ~1% for SPXL, less SPY ETF's 0.1% fee multiplied by the 3X leverage factor for 0.3% (all percentages are per year).
For example, starting with an account consisting entirely of $10k cash, buying $20k of an unleveraged ETF or $10k of a 2X levaraged ETF may have the same margin requirement, but only the former will incur margin interest, in the absence of any other trades.
Regarding the original poster's idea of using margin and leveraged ETF's together, one could avoid borrowing cash by shorting $1 of a -1X ETF for every $5.20 of the corresponding 3X ETF (by solving initial_cash = long - short = (long x 3 x 25%) + (short x 30%)), assuming margin of 25% for long positions (75% for a 3X ETF), 30% margin for short positions, no bid-ask spreads, no commissions, no slippage, avoiding 50% regulation T end of day initial margin requirements (perhaps by not holding overnight), clairvoyance to know that the underlying will not drop by even one tick below the entry until exit, and the really unlikely possibility that I have not botched the math.
Before you embark on a strategy of holding leveraged ETF's across rebalancing periods (usually daily), you should definitely read about decay of leveraged ETF's, often called "volatility drag." The losses from that can be huge. Occasionally, LETF's can outperform their target leverage (for example, SPXL tracks 3X SPY daily, but I believe has gotten almost 4X SPY's total return since election day). However, leveraged ETF's have been shown usually to decay substantially under some reasonable statistical assumptions about the movement of the underlying--according to one paper [1], exponentially at a tempo proportional to, among other parameters, variance (that is, volatility squared) and L*L - L, where L is the leverage factor (L*L-L works out to zero for L=0 or +1, 4 for L=-1 or +2, 6 for L=+2 or -3, and 12 for L=-3).
This is not to say that leveraged ETF's have expected values below their prices, just that they evolve into directional lottery tickets, like some option positions, which, by the way, is another possibility for leverage, along with recession2016's suggestions of considering futures, both of which have their own costs, risks and potentially costly fallacies. Be careful.
[1]
https://www.math.nyu.edu/faculty/avellane/thesis_Zhang.pdf , page 10, equation 2.1.4.
