I'm friends with several current or former affiliates of the Cato Institute, most of whom (along with a few others) have emailed me regarding my Tapped post on the accuracy of the Social Security calculator you can find on Cato's privatization advocacy site. Since my initial post had some problems, it's worth giving this issue a second go below the fold. First, however, I think it's noteworthy that no one has written in trying to defend the Heritage Foundation's calculator, which makes private accounts out to be a much better deal than does Cato.
Cato wants me to believe that when I reach my Normal Retirement Age (67) in the year 2048, I will have accumulated a private account $521,276 in 2004 dollars if we implement the Cato privatization plan [LINK] today. They further allege that with this $521,276 I will be able to purchase an annuity paying $49,639 per year "fully indexed for inflation."
One question I raised was about the legitimacy of this annuity estimate. My proposed counterestimate, however, was also off the mark. Bruce Bartlett, an economist and no apologist for big government, recommended this calculator as a useful source of information in this regard "from a group with no ax to grind."
By their calculations, a 67 year-old man living in the District of Columbia with $521,276 to spend could purchase a "single lifetime income with no payments to a beneficiary" (which I think is what the Cato site is talking about) worth $3,608 per month, or $43,296 per year. I don't know whether or not that's indexed for inflation. If it is, then it suggests that Cato is off by about 15 percent. If it isn't, then Cato is off by more. But it's also important to keep in mind that this is an estimate of what you could buy with $521,276 in 2004 dollars in 2004. By 2048, technology will have improved and life expectancy at 67 should be higher than it is today, meaning that $521,276 in 2004 dollars will buy me a lower annual payment (in 2004 dollars, all the numbers will be much higher, of course, in nominal dollars) than this.
A far more significant source of disagreement stems from the assertion that I'll have $521,276 in my account by 2048. Cato derives this number because (to quote their website) "individual account accumulation assumes the average rate of return on a 60-40 Stock-Bond portfolio (5.27 percent [above inflation]) minus transaction cost (0.30 percent)." Via email, a Catoite notes that "this is considerably lower than the historical return on stocks of 7 percent per year." But since my portfolio is only 60 percent stock, of course the aggregate return is considerably lower than the historical return on stocks. This is a bit of a red herring. Now I think (hopefully someone will alert me if I'm mistaken) that they get the 5.27 percent figure by assuming that stocks will have an average real annual return of 6.5 percent (this is what the president's privatization commission estimated) and bonds an average real annual return of 3.25 percent (6.5 * .6 + 3.25 * .4 = 5.2). A return of 6.5 percent is not considerably lower than the historical return on stocks. But the estimated figure should be considerably lower than the historical figure for two reasons.
One is that price/earnings ratios have reached a level considerably higher than their historical average. This implies either that stocks are overvalued right now, or else that stocks were undervalued in the past. Either way, it implies slower growth in the future.
The other is that the denominator of the P/E ratio -- earnings -- is related to economic growth. But the future economy is very unlikely to grow as quickly as the economy of the past, primarily because the population will be growing much more slowly. This population-driven slowdown in economic growth is the source of the purported "crisis" in Social Security but its effects will be felt in many areas, including the stock market. Dean Baker has used this information to devise his No Economist Left Behind Test as follows:
President Bush's Social Security Commission assumed that stocks would give an average return of 6.5 percentage points above the rate of inflation over the next seventy five years. The commission uses the Social Security trustees projections, which show that profit growth will average 1.6 percent above the rate of inflation over this period. The current ratio of stock prices to corporate earnings is 25.8 to 1, which means that corporate earnings are on average equal to 3.9 percent of the price of a share of stock.
Use this information to show a set of capital gains and dividends that produces the 6.5 percent rate of return assumed by President Bush's Social Security commission.
By Baker's estimate, the return on stocks will be more like 5 percent than 6.5 percent, which would drive the overall rate of return on my private account to well below 5.2 percent which, thanks to the compound effects, would leave me with considerably less than $521,276 in my account when I reach the age of 67. Baker's precise estimate is, of course, open to question. I have not, however, seen anyone offer a serious reply to this argument that's consistent with the 6.5 percent estimate used by the Commission and, unless I'm mistaken, by the Cato Institute.
And all this is without getting into the issue of administrative costs. For more on that see, e.g., the case of Chile.
January 27, 2005 | Permalink
TrackBack URL for this entry:
Listed below are links to weblogs that reference The Calculator:
Tracked on Jan 27, 2005 10:21:56 PM
» Fun with Numbers from The Bit Bucket
Matt Yglesias expresses skepticism about Cato's Social Security Calculator. One objection is over the annuity calculation, which he says is off by about 15%. I don't know that much about annuity markets, but it seems plausible that Cato's using an... [Read More]
Tracked on Jan 28, 2005 9:49:44 AM
» The Tim Lee Social Security Calculator from The Fly Bottle
Ygelesias, among others, is skeptical of the various pro-reform social security calculators. My colleague Tim Lee responds. Here's a bit: More to the point, even if you grant all of Matt's objections, personal accounts still end up doing better. I... [Read More]
Tracked on Jan 29, 2005 5:10:24 PM
Tracked on Jan 30, 2005 8:53:02 PM
» Response to Critics of Cato's Social Security Calculator from The Agitator
There's been quite a bit of criticism from the left side of the blogosphere of Cato's Social Security calculator. See... [Read More]
Tracked on Jan 31, 2005 12:01:28 PM
At your age no calculator is going to give you a result distinguishable from a WAG. The closer you are to retirement the more accurate they are.
Posted by: ej | Jan 27, 2005 3:58:40 PM
Of course, all of these higher rates are higher expected returns and not actual returns. And the only reason expected returns are higher is the additional risk. So Ed Prescott and Robert Barro say the Democrats are just silly when they claim privatization will force folks into riskier portfolios. They are right. People can and WILL select safer positions, which means all of these calculations based on the premise that all of the funds go into stocks violate the sound economic principles put forth by Barro & Prescott. If the calculation were based on these sound principles, the extra risk and extra expected return would be ZERO.
Posted by: pgl | Jan 27, 2005 4:04:32 PM
This whole concept of privatization seems ludicrous to me. What it comes down to is this: capital gains are inherently parasitic. The person earning the capital gains is doing no labor, but he is receiving goods. If you ignore the fact that this is inherently parasitic, it's easy to believe that we can all have a private account with a million dollars in it, we can all earn 5% interest, and we can all live the easy life with nobody ever working again. But that doesn't pass the laugh test. It's a free lunch, it's magical thinking.
In practice, here's how it would play out. The reason anybody gets 5% interest is because capital is scarce. But if the government forces everyone to invest, the supply of capital is going to go through the roof. Simple supply and demand is going to lower the value of capital. That's going to push interest rates much lower. If you create a few accounts, you can earn 5%. If you create 300 million, you'll barely be matching inflation.
Posted by: Josh Yelon | Jan 27, 2005 4:14:58 PM
People can and WILL select safer positions,
That's absurd. People can and WILL bet (correctly) that if they lose their shirts on the market that they will be bailed out by the Federal government. It'll make the S&L debacle look like a liquor store holdup.
The person earning the capital gains is doing no labor, but he is receiving goods.
Think of the capital granted as crystallized labor done previously; what the capital lender is doing is earning delayed returns on his labor.
In practice, here's how it would play out. The reason anybody gets 5% interest is because capital is scarce. But if the government forces everyone to invest, the supply of capital is going to go through the roof. Simple supply and demand is going to lower the value of capital. That's going to push interest rates much lower. If you create a few accounts, you can earn 5%. If you create 300 million, you'll barely be matching inflation.Well said. This is a key point that is not discussed.Cranky
Posted by: Cranky Observer | Jan 27, 2005 4:20:46 PM
Please get one of your cato pals to address the real/nominal issue. I'm pretty sure they're presenting SS benefits in real terms and private account results in nominal terms..
Posted by: Atrios | Jan 27, 2005 4:45:32 PM
Since the stock market is lower than it was 4 years ago, 7% returns seem a bit ... optimistic. Can we instead invest with percentage increases like the debt?
* Stock market
Dow Jones Industrial Average
* S&P 500
* Value of the Dollar
1/19/01: 1 Dollar = 1.06 Euros
1/19/05: 1 Dollar = 0.77 Euros
2000 budget surplus $236.4 billion
2004 budget deficit $412.6 billion
That's a shift of $649 billion and doesn't include the cost of the Iraq war.
* Cost of the war in Iraq
* The Debt
End of 2000: $5.7 trillion
Today: $7.6 trillion
That's a 4 year increase of 33%
Posted by: AlGore | Jan 27, 2005 4:45:57 PM
And all this is without getting into the issue of administrative costs. For more on that see, e.g., the case of Chile.
The administrative cost estimate in the Cato calculator is EXACTLY SPOT ON. I know why you want to look at the Chilean example, but it is just wrong to do so. We can look at a much better model: the Federal Government's Thrift Savings Plan. The TSP has administrative costs just about the same as Cato's estimate of 0.30% per year.
Posted by: Al | Jan 27, 2005 4:47:37 PM
Since the stock market is lower than it was 4 years ago, 7% returns seem a bit ... optimistic.
AlGore's using fuzzy math again. Four years is too short a period from which to extrapolate the expected return from the portfolio. Try 20.
Posted by: Al | Jan 27, 2005 4:55:24 PM
"Think of the capital granted as crystallized labor done previously; what the capital lender is doing is earning delayed returns on his labor."
You can "think of it" that way, but it's still not going to work any better. When grandma and grandpa die and pass their private accounts on to their descendants, and their descendants try to live on the interest of those accounts without ever doing a day of labor in their life, and when everyone in the United States does the same thing because we all have these accounts, you're going to see the accounts devalue very rapidly.
Posted by: Josh Yelon | Jan 27, 2005 4:59:02 PM
Josh Yelon wrote, What it comes down to is this: capital gains are inherently parasitic.
Nope. The real parasites are landowners. See "A Geolibertarian FAQ" and "Are you a Real Libertarian, or a ROYAL Libertarian?".
Al wrote, Four years is too short a period from which to extrapolate the expected return from the portfolio. Try 20.
Of course, there's no sense in extrapolating long-term market trends from short ups and downs.
However, assuming future returns in the domestic equities market isn't playing fair, because of Dean Baker's "No Economist Left Behind Test" that Matt mentions above.
Briefly, past historic stock returns are inconsistent with the projections of future economic growth. Why? Because stock depend on earnings growth, and corporate earnings in the long run can't grow faster than GDP. And the average historic rate of GDP growth---the rate that made these nice, tasty real 7% returns possible in the past---is far higher than the rate assumed in the SS projections. (Not to mention that current P/E and P/D ratios are extremely high.)
Once upon a time, there was a town. The town dreamed up a scheme under which some of its old folks would get some retirement money from people currently working. The old folks loved it! Free money! Since there were so few old folks at first, from time to time the town mayor and council gleefully increased the amount each old person could get, and all the time they spent the surplus on all sorts of goodies.
Now, after the program had gone on for a while, it started to get really expensive. Some of the working folks said, hold on a minute, why am I forced to save so much of my money through you? I think I can do better by investing my money in my business and farm, and lending the money to my neighbors for their businesses. And what’s with this crap about employer contributions? You don’t really expect me to believe that, do you?
The town mayor and council (each and every one of whom, oddly enough, had been booted out of the navy for drunkenness) said, oh, what a silly idea! You can’t get a better deal than giving us your money to spend! Why, we’ll take your money and build tropical forest entertainment centers! We’ll subsidize any one of your businesses that cares to give us money when we run for re-election in our carefully jiggered districts! Blah, blah, blah!
Boy, were those working people chagrined! What logic! Yes, yes, they said. Spend my money! Please!
Posted by: ostap | Jan 27, 2005 5:30:31 PM
Another highly questionable assumption made by the calculator is the 4% REAL increase in income each and every year you are working.
According to the BLS, median weekly earnings of a 25-34 yr old male are $637, while weekly earnings of a 55-64 yr old male are $844. That's a 1% real growth rate.
If you optimistically assume trend productivity growth around 2% for 30 years, and you assume that ALL that real growth ends up in wages, then you're looking at 3% real growth total.
Over 30 years, that lower rate of wage growth will make a big difference to your private account balance.
Posted by: MarkT | Jan 27, 2005 5:32:15 PM
If you optimistically assume trend productivity growth around 2% for 30 years
2% is not optimistic - it probably average. We've been averageing well over 2% for the last decade or so.
Posted by: Al | Jan 27, 2005 5:43:56 PM
Al's right. So, how come the Social Security Trustees are assuming a much lower rate of productivity growth when they calculate the long run solvency of the program? 2% productivity growth would mean that the program stays solvent forever and ever...
Posted by: Atrios | Jan 27, 2005 5:48:40 PM
Atrios, I don't think real/nominal is the issue with these calculators so much as the assumptions on real investment returns and real wage growth. I was able to brew up a spreadsheet that more or less reproduces the Cato and Heritage figures in real terms. Stuff like adding 80 b.p. to the historical average annual real return on Treasuries (as in the Heritage calculator) makes a big difference compounded over a 46 year working lifetime.
Then, as Matt suggests in the main post, the annuitizations are very optimistic.
Another interesting annuity calculator is the TSP's. The TSP calculator has explicit options for Social Security-like joint life annuities with increasing payments and 100% payments to the surviving spouse. Those offer considerably lower payments than the single life annuity scenario (you can also see this to some extent in the calculator suggested by Bartlett).
It's also kind of amusing that privatization is, other things equal, a worse deal for traditional single-earner families.
I (gulp) agree with Atrios. To the extent possible, these calculators should use assumptions that are consistent with the (intermediate) assumptions used by the SS Trustees.
Posted by: Al | Jan 27, 2005 6:22:27 PM
That trend productivity growth has averaged ~2% for the last 10 years, does not mean it will continue to do so for the next 30.
Prior to ~1995, productivity growth was lower. I think ~2% is a plausible number, but it assumes a higher level than has been the average over the last 40 years. It's certainly not a conservative estimate that one should bank on for retirement planning purposes.
Posted by: MarkT | Jan 27, 2005 6:43:48 PM
Kimmitt had a legitimate objection to what I claimed ala Ed Prescott's moral hazard argument. As I noted over at Angrybear when Prescott came out for privatization: (1) the Lindsey Graham style choice plus guarantee invited the S&L style disaster (only thing Andy Biggs ever got right IMO); and (2) Prescott was almost saying this could happen even without a formal guarantee. Kimmitt - well done!
Posted by: pgl | Jan 27, 2005 6:49:07 PM
Yes, the early '90s and all of the '70s were periods of slow productivty growth. The last 10 years and the '50s and '60s had higher growth. As I said, I think 2% would be average, not optimistic. Optimistic would be 2.5% - 3%.
Nonetheless, if we really want to compare what a partially privatized system would look like as compared to the existing system, whatever is used should be consistent with the Trustees.
Posted by: Al | Jan 27, 2005 6:53:42 PM
The problem is that the Trustees themselves are inconsistent - they use implicit high productivity growth when scoring possible privatization plans, and then turnaround and use low productivity assumptions when scoring the solvency.
In other words - hacks!
Posted by: Atrios | Jan 27, 2005 6:58:01 PM
Unless there's a reason to believe that productivity growth would be higher with partial privitization. Maybe due to better access to capital in the private sector? Hmmm...
Posted by: Al | Jan 27, 2005 7:25:18 PM
I think I found where Cato got their bogus annuity calculation. If you check out this link at the federal Thrift Savings Plan it explains how federal employees can convert their savings into an annuity. www.tsp.gov/forms/tspbk05.pdf.
If you go to Table 1 on page 11 you will find the payout rates for various retirement ages. You will notice in column 2 at 67 years old you get about 9.5% which is close to what Cato is calculating. However, the devil is in the details. This number does not include an inflation adjustment. You get the intial value for the rest of your life. If you want inflation protection, you have to go to column 5. In that case you only get about 6.5% initially. The maximum inflation adjustment per year is 3%. If you hit a period like the early 80's with 10% inflation, tough luck. Next, these table values are based on an index of the interest rate on 10-year treasury bonds assuming 7%. Currently the interest rate on the 10-year bond is only 4.25%, and if you do the calculation, you have to reduce your benefit by 20%. So if you retire when interest rates are low, you are stuck with low income for life since the payout is fixed (unless you opt for the inflation protected alternative, in which case you get even less). Finally, the principal sum is forfeited at death, unless you opt for the spousal survivorship benefit, in which case you get smaller initial payments. So much for the idea that you can pass your account to your heirs.
Posted by: Joe | Jan 27, 2005 7:39:26 PM
The comments to this entry are closed.