1. Suppose I have $100,000 in wage income and (for simplicity) suppose the income tax rate is 20%. That $100,000 is subject to income taxes and I pay $20,000. I consume $80,000 worth of stuff.
2. Suppose I sock $1,000 away into a traditional IRA. My taxable income this year is reduced by $1,000 to $99,000. I only pay $19,800 this year. But when I retire and withdraw this $1,000 and all the income that it has generated (say it doubles and earns another $1,000), the whole future sum is subject to income taxes at this later date. I pay $400 in taxes on the $2,000 withdrawn in retirement; $200 deferred taxation on the $1,000 I socked away and $200 on the income it generated. I consume $79,200 worth of stuff today and $1,600 worth of stuff in retirement. The tax rate on the earnings from the IRA is the same as the tax rate on wage income.
3. Suppose instead that I set aside the same $1,000 for investing but it is subject to taxation first. I pay the $19,800 in taxes on the $99,000 in income as before. But since I invest after taxes, I also pay $200 on the $1,000 I set aside for investing. So, I have paid a total of $20,000 in taxes today as in scenario 1 above and consume $79,200 worth of stuff today as in scenario 2 above. But, I only have $800 to invest. It doubles and I realize $800 in income in retirement. (This is a capital gain and is usually subject to a different tax rate, but let’s make it simple and keep the 20% tax rate.) This is taxed at the later date and I pay $160 then and consume $1,440 in retirement. Relative to scenario 2), my retirement income is lower. Why?
Because with an after-tax investments, one round of taxation occurs before I even invest it. It was taxed when I initially earned it. Any additional taxes are – additional taxes. The only way to make scenarios 2 and 3 equivalent is to have a tax on capital gains of 0%. Otherwise we are taxing capital income at a higher rate than wage income.
We may "want" to (read “it may be efficient to”) tax capital income at a higher rate than wage income. But I doubt it. This is not controversial among economists. There is a well established public finance literature that finds that the optimal taxation of capital income is 0%.* I do not know, but I suspect that Romney paid taxes on the income that he then invested. His 15% tax rate (it appears to actually be lower) would not be lower than the rates paid by most people. Because it is was taxed once already, it would be in addition to the rate that other people pay.
*There are simplifying assumptions above and there are some modern public finance results in which, with the right assumptions, it is efficient to have positive capital gains taxes. Even then, they would likely be less than the current 15%.