Say it isn’t so, George. Here is a classic politician’s statement: factually correct, yet wrong by orders of magnitude. Any fool with the back of a small envelope could tell you the real rate of return on his Social Security payments is massively negative. So yes, massively negative is factually “less than 2 percent,” but that’s like saying someone who’s lost five pints of blood has lost “more than a teaspoon;” it just doesn't quite capture reality, does it? To be fair to poor George, he wasn’t trying to defend the system. Indeed, in his second term, he even proposed a modest reform of Social Security by transitioning a small piece of it to a “my-money-in-the-bank” system. But George’s radical overstatement of my returns was such an affront to my intelligence that I finally decided to calculate my real rate of return on Social Security. Let’s find out just how much below 2 percent it is, shall we?Social Security now offers workers a return of less than 2 percent on the money they pay into the system.
President Bush the Jr.
Address to the Joint Session of Congress
February 27, 2001
My method is simple. I calculate the present value of the:So, if my expected future SS income is $10, and I’ve already paid $5 in SS tax, and I expect to pay $3 in future SS taxes, then my net return is 10-(5+3) or $2, and my net return on investment is 2/8, or +25 percent. Yeah, right. Here’s how the actual calculations came out:
(2+3) is my total investment. 1-(2+3) is my net return, so 1-(2+3)/(2+3) is my net return on investment.
- Future SS income I expect to receive;
- Past SS taxes I’ve already paid, and
- Future SS taxes I expect to pay.
$ 30.8K (present value of the SS income I expect to receive in the future)And the real return on my social security “investment” is . . . -473K/537K or negative 94 percent. I’m shocked! Bush the Jr. was right: my return is less than 2 percent . . . 96 percentage points less.
-407.5K (present value of SS taxes I’ve already paid)
- 99.1K (present value of SS taxes I expect to pay in the future)
-$473.0K (net return)
One interesting result jumps out: the largest line item above, by an order of magnitude, is the value of the SS taxes I’ve already paid. That’s also the result that is least sensitive to changes in my assumptions about interest rates, too. In other words, just as I suspected, I’ve already been screwed and there isn’t a damn thing I can do about it. Another interesting result is that despite the hundreds of thousands of dollars I've already paid into the system, if offered the chance to walk away from the SS system today — get no benefits in exchange for paying no more taxes — I would walk. Why? Because even on a go forward basis, (that is, ignoring my huge historical tax payments), I’m going to pay more than three times in taxes than I will receive in benefits from SS. That’s the first lesson you learn in business school: ignore sunk costs.
I have developed a (DRAFT) Excel spreadsheet I use to calculate the Social Security return on investment for a fictional couple I call “Krista and Bob”. They bear a close resemblance to Katherine and me. If you believe my assumptions are faulty, please substitute your own. [Note: I believe there is an error in the draft model in cell AF:9, which is the 2004 NYSE gain. The cell has 45 percent in it, and that's probably wrong. ] No matter what you do, (up to and including ignoring the time value of money; that is, assuming a dollar received in 1974 is worth the same as a dollar received in 2049), I don’t think you will be able to make the SS ROI positive.
Let’s review the important assumptions in my model:
I. Social Security
1. All my expected future SS income (source: my personalized “Your Social Security Benefits” guide for 2005) from my “normal retirement age” (67) until the age that mortality tables say that I’ll die (79).
1. These are the benefits that my involuntary participation in the SS system has entitled me to.
2. All my wife’s expected future SS income (source: 1. her personalized “Your Social Security Benefits” guide for 2005) from when she retires at her “normal retirement age” (67) until the age that mortality tables say that she’ll die (83). 2. Your spouse gets the larger of A) one-half of your SS monthly benefit or, B) the actual monthly benefit she has earned from the SS taxes she has paid. As of 2005, A is larger than B in my case.
II. SS taxes paid
1. All SS taxes I actually paid;
(sources: 1. my personalized “Your Social Security Benefits” guide for 2005 and 2. my tax records dating back to 1974).
1. These are the actual taxes the SS system has extracted from me so far.
2. All SS taxes my wife actually paid;
(sources: 1. her personalized “Your Social Security Benefits” guide for 2005 and 2. her tax records), and
2. The SS system treats benefits, but not taxes, as a family affair. Because one-half of my benefits is larger than the benefit my spouse has earned with her own SS tax payments, she will get one-half of my benefit. If the SS system were fair, it would exempt all her earnings from SS taxes because she is getting NO incremental SS benefit from the SS taxes she pays. In other words, by virtue of the SS taxes I pay, she will get the same SS benefit she would have gotten if she were unemployed and paid no SS tax whatsoever! If I'm going to give SS credit for the monies they pay my spouse, then I have to include the taxes they took from her as well. This is especially so because the benefit she receives from my SS taxes replaces her own benefit, instead of adding to it. If the benefits she earned were paid to her independent of whether I paid into the system, then one could argue that both her taxes and the benefits they procure are irrelevant to this analysis. But that's just the point: she gets no incremental benefit from the SS taxes she pays. Fairness dictates that if I count all the SS benefits my spouse will get, I must also count all the SS taxes she pays. 3. All SS taxes our employers actually paid on our behalf. 3. If my employers had not been forced to pay into the SS system on my behalf, they would have paid an identical amount directly to me in salary. That's because the value of my services includes this SS tax paid on my behalf. Remember, I get the benefits (meager as they are) that my employer's SS tax contribution procures, so I would demand to be “made whole” if my employer stopped giving me this benefit. My value hasn't suddenly decreased if the tax suddenly goes away, has it?
One could argue that if the employer match were suddenly repealed, employers might simply pocket all or a large fraction of their savings rather than pass it on to employees. I would argue that in that instance, for the SS system to remain solvent, the employer match would have to be made up from somewhere. Most likely it would be recouped via higher taxes on employees. Either way, it belongs in my analysis.
III. Disability income and survivor benefits from SS $0 (zero). Income replacement for spouses and minor children in the event of disability or death is surely one of the more expensive components of SS’s cost structure. So why have I excluded that potential benefit from my analysis? Simply because I’m not dead, I’m not disabled, and I am not likely to become dead or disabled any time soon. In the most likely case, I will work and be taxed until my normal retirement age, and so will my wife. The purpose of my analysis is not to figure out every possible way that my family and I might receive a contingent benefit from SS (or to peg the expected value thereof). It’s to figure out my actual rate of return from the SS system given my historical contributions and the most likely outcome of my future life events. The vast majority of people in my position will never receive any disability benefits from SS. Ditto for survivor benefits to minor children.
IV. Estimate of the dollar value of “psychic” income my family and I derive from knowing that my family will receive benefits if I die or get disabled. $0 (zero). The SS benefits paid to my family in the event of my death or disability are so small that for planning purposes, I ignore them. That's why I have purchased private life insurance and I have private disability insurance: precisely because I have no warm fuzzy feeling (what economists call “psychic income”) about the SS support my family will receive if I die or become disabled. Put differently, if SS’s death and disability benefits were repealed tomorrow, I wouldn't run out and buy more life insurance or disability insurance.
V. Interest rate on historical SS contributions. I have used the historical NYSE index from 1974 to 2005 to calculate the gains my assets would have enjoyed had they been at my disposal. If SS had been repealed (or modified in a way to compel me to save my retirement monies in a private account), and I could have invested my historical tax payments myself, I would have invested it in stocks that trade on the New York Stock Exchange. The NYSE index is a fair proxy for the opportunity cost (the growth I have forgone) of the monies I have paid to SS. This is where I've invested the vast bulk my savings since 1974. In fact, I could credibly argue that the NYSE index understates my opportunity cost, because in recent years, I have put a big chunk of my savings in Pacific Rim indices and other Asian funds, which have enjoyed higher returns than the NYSE index.
VI. Interest rate on future SS contributions. I have used 10 percent, the historical average annual gain on stocks, to calculate the gains my future SS taxes would enjoy if these monies were at my disposal. The historical average annual gain on stocks is a fair proxy for the opportunity cost (the growth I will forego) of the monies I will pay to SS. I could credibly argue that using average historical domestic stock gains understates my opportunity cost, because in recent years, I have put a big chunk of my savings in Pacific Rim indices and other Asian funds, which have enjoyed higher returns than the NYSE index.
VII. Future SS taxes. Taxes will remain fixed at 2005 levels for the duration of my analysis. This has never happened and probably won’t start now. This is a conservative assumption.
VIII. Future SS benefits. Benefits will remain fixed at 2005 levels for the duration of my analysis. SS benefits rise each year at a higher rate than inflation. However, they have also been cut once, in the early 1980’s. I have assumed no change for this variable because: 1. it’s difficult to model COLA’s, 2. benefits have already been cut once, and 3. I believe future benefit cuts for people in my generation are a virtual certainty. Again, I am being conservative in my analysis.
IX. The “normal retirement age” Set at 67 years with no changes. Katherine’s and my normal retirement age has already been raised once (from 65 to 67 years). I think it's a virtual certainty that it will be raised again, probably by more than two years. Again, I am being conservative in my analysis.
X. Taxes on SS income during retirement. $0 (zero). I may need to add this to the model.