How to Save Social Security, take IV!
I decided to crank some numbers on my idea to reduce the volatility of the NYSE and improve the return to the funds held in the Social Security Trust Fund by having the US gov't open its own mutual fund that determines its holdings with a simple algorithm that uses 27 weeks of data and medians and log-returns.
http://sodsbrood.com/antimani/2006/07/13/a-simple-thought-experiment-on-us-mutual/
It seems that US Mutual, in order to reduce the volatility of the NYSE and hedge against a worse case scenario of a general downturn, would need to reduce its guaranteed return. However, the system, as described at my main blog, would still be an improvement over the current social security system and have an advantage over other mutual funds with its lower overhead costs. One can also argue that the reduced volatility of the NYSE would be in the public interest for the promotion of greater economic stability.
dlw
What you've done is use single-investor analysis for a huge program with millions of investors. It's invalid for that reason alone. It's like the difference between one person tapping an underground water source and 100 million tapping it.
No matter how the mutual fund is constituted, no matter who runs it, no matter what its policy, the investments at the start will drive up the prices of the stocks in the fund. When stock sales become about equal to purcheses prices will stabilize. When sales outweigh purchases prices will fall. Your single-investor analysis will never show that: you implicitly assume the investments are so small relative to the total market that they have no influence. That is an invalid assumption for a plan that is funding a general retirement security program.
Even saying single investor transactions have no effect is wrong: look at the stock ticker anywhere it's shown. Those are individual sales, those individual sales do affect the prices of the stocks traded.
It's a market. It will respond to market forces. Purchases and sales are market forces.
July 14, 2006 10:02 AM | Reply | Permalink
I replied here to your post.
I fail to see how market power throws the monkey wrench into my idea as the algorithm is more about determining the holdings. Yes, there would be a transitional period after the beginning of US Mutual, but eventually a new equilibrium would emerge.
What matters most is that, in accordance with other reforms, US Mutual would become a permanent fixture, an institution, in the NYSE that would force other agents to change their behavior.
dlw
A blog-activist dedicated to the reduction of the faith-based political acrimony in the United States of America so as to make our political system more democratic and just and to improve our witness to the rest of the world.
July 15, 2006 11:54 PM | Reply | Permalink
You admit you ignore sales and consider only purchases. You have to consider the eventual sales, don't you think?
The "new equilibrium" will be bullish at the start (by design: not letting those who will retire soon participate) and bearish in the long term. The personal account advocates invoke US demographics to point with alarm at the prospects for Social Security and then promptly forget about US demographics when it comes to their scheme. That's not analysis, that's a deceptive sales practice.
SHOW ME A REAL PROSPECTUS. SHOW ME A SPREADSHEET. $X coming in, $Y going out (of the personal account system.) Show me the spreadsheet when Y = X, show me the spreadsheet when Y > X. Show me a stable market at that time. It only works via deus ex machina of some sort.
Also, "tinkering" with the algorithm is just another way of saying "change the algorithm to avoid this kind of loss in the future," isn't it? With "this kind of loss" being a new kind every time.
August 2, 2006 12:24 PM | Reply | Permalink
100% agreement with hewhoasks. And saying that you're going long on preferreds is like saying you're going long on bonds. With your revised estimates, why wouldn't people just buy USG bonds?
What exactly are you trying to fix in Social Security?
Click here for the Users Help Forum.
July 14, 2006 10:30 AM | Reply | Permalink
Well, we can test to see what the effects would be. The buying long would be spread out over 20% of the market. That would mitigate the impact of the market power on the initial purchase of stock some.
And it really is not that hard to improve upon the return that is currently implicit for Social Security. US Mutual would work because it would be held perpetually and its holdings on the buy long side would be relatively stable. It's market power would also thereby work in its favor.
It may need some tinkering in the algorithm, but the combination of economies of scale, with low overhead and protections against corruption by the use of the relatively simple algorithm would make it possible to grow SSTF without raising taxes, or forcing people to pay into Social Security. And it would tend to reduce the volatility of the NYSE, which would encourage more investment in the NYSE and be of help to growing our economy.
dlw
A blog-activist dedicated to the reduction of the faith-based political acrimony in the United States of America so as to make our political system more democratic and just and to improve our witness to the rest of the world.
July 15, 2006 11:48 PM | Reply | Permalink
What you've done is use single-investor analysis for a huge program with millions of investors. It's invalid for that reason alone. It's like the difference between one person tapping an underground water source and 100 million tapping it.
No matter how the mutual fund is constituted, no matter who runs it, no matter what its policy, the investments at the start will drive up the prices of the stocks in the fund. When stock sales become about equal to purchases prices will stabilize. When sales outweigh purchases prices will fall. Your single-investor analysis will never show that: you implicitly assume the investments are so small relative to the total market that they have no influence. That is an invalid assumption for a plan that is funding a general retirement security program.
I make no such assumption. Yes, the initial beginning of US Mutual would cause a turbulent transition phase during which expert assistance in purchasing stocks will be needed. And the market power of US Mutual will result in an increase in the prices of the stocks it buys long on that will spill over to some other stocks. However, it will not simply buy long on stocks, it also sells short some stocks. And so the net effect would not simply be an inflation in the stock market value.
As for my algorithm, it deals more with systematically determining and redetermining the holdings on a weekly basis. The median log-return statistics are designed to ensure greater stability in the holdings. That and the fact that the holdings are distributed over 20% of stocks, weighted for their value, will reduce the impact of buying and selling and add more stability to the market.
Even saying single investor transactions have no effect is wrong: look at the stock ticker anywhere it's shown. Those are individual sales, those individual sales do affect the prices of the stocks traded.
It's a market. It will respond to market forces. Purchases and sales are market forces.
That's why I didn't say it...
duh, if US Mutual were established, it would simply cause the overall equilibrium to change. We can use both historical data and simulations to get at its likely effects. The point you're missing is that this 800 pound gorilla would be using its weight to stabilize further the more stable performing stocks and to bleed the overvalued unstable stocks that are not performing consistently well and to discourage generally stock market buying/selling behavior that tends to increase volatility.
dlw
A blog-activist dedicated to the reduction of the faith-based political acrimony in the United States of America so as to make our political system more democratic and just and to improve our witness to the rest of the world.
July 15, 2006 11:38 PM | Reply | Permalink
The algorithm encapsulates the single-investor strategy, is based on the single-investor strategy. That is the flaw.
"Single-investor strategy" means that the action taken as a result of that strategy is so small compared to the total market that all the other actions taken in that market overwhelm the influence of the action. Once you multiply the action by the number of people in the workforce that can not be the case: what is done with all that retiree money will have an effect. It is not the same market as it was without those billions of dollars of retiree money.
Even individual market transactions have an effect: that's what is reported on a stock ticker. Note what the stock ticker shows: prices change as a result of transactions. If on a particular day a bunch of holders decide to sell "at market" the price will tend to go down. If a bunch of investors decide to buy "at market" the price will tend to go up. With personal accounts there will be, in effect, millions of "at market" buy orders placed. That will drive prices up. That is not a result of any increase in the inherent value of any of the securites purchased, it's just simple supply-and-demand. Your algorithm ignores that. While it looks like it's a really good thing that increase in price only lasts while "at-market" purchases dominate "at-market" sales. Look again at the demographics. The demographics show that in time the sales will outweigh the purchases.
If your algorithm doesn't ignore that effect then if it's a good algorithm it will be, essentially, "buy while prices are going up and sell before they go down or just after they start to go down." Fine, but selling is what makes prices go down. You cannot have millions of investors using that algorithm without creating a market crash. You simply cannot.
Bush, of course, already has told people that the algorithm will be "buy, and hold forever." He said, in Alabama, Arizona, Colorado, and New Mexico that retirees would not be allowed to sell their investments. Details about what those who receive the accounts as a bequest can do are not yet revealed. Do you see why Bush said "hold forever"?
Some will say that other discussions of the personal account scheme provide for the principal to be switched from stocks to annuities when the worker nears or reaches retirement. That guarantees a crash: switching from securites means selling securites, and selling securities drives prices down. It's not a logical investment strategy, it's just propaganda to make the scheme look sound when it is not.
A market is just a market. In the market as it exists for ordinary investors it is required that for each security or mutual fund a prospectus be suppplied. The ones I've seen lately for mutual funds have been filled with descriptions of the different sorts of risk involved. Bush doesn't provide a prospectus (nor do you.) If Bush (or you) did then that prospectus would have to honestly describe the risks. Any general (entire population) retirement plan that depends on investing in financial markets should, if honestly presented, describe the risks involved. With the demographics of the US as they are the biggest risk isn't a "risk" at all: it's a certainty of a crash. In any case, the policy of the US electorate shold be: "no prospectus, no sale." That's smart investor strategy, right?
The mathematical way to deal with that is to say "Well, I'll create an algorithm that ignores such a possibility. I can show that algorithm works under that assumption." Well, sure: you're ignoring the possibility of the personal accounts themselves causing a crash so you don't see the possibility of a crash. Bad idea to ignore that, when the accounts are intended to provide retirement income, wouldn't you say? You have to pay attention to what a market is and to what it would mean to put billions of dollars of retirement money into the financial markets, to be taken out later.
In essence you are claiming that an algorithm based on market experience during the time when there was no general retirement scheme based on investing in financial markets will still be valid for a time when there is a general retirement scheme based on investing in financial markets. Sorry.
This is important enough that you need to actually demonstrate the algorithm will work (which you cannot do.) Hypothesizing the algorithm will work is not the same as showng it will work. Besides, there's been thousands of algorithms for investing. Funny thing: many mutual funds lose money. That there may be a successful algorithm doesn't guarantee that anyone will find it nor does it guarantee that, once found, it will continue to succeed. Even with success there's a companion problem: the algorithm works best when the size of the fund is relatively small. That's why highly-successful funds get closed to new investors: to slow the growth of the successful fund in ordfer to help keep it successful. How are you going to do that with a fund that is the required destination for FICA taxes? Change "the required" to "a required" and the problem is still the same: a bunch of funds is not a solution.
You might get clever and say "Well, how about all those 401(k)s and IRAs held by the baby boomers?" Good point. Check out what the economists at Yale are warning. I'd Google it for you but you can do it yourself.
There is no such thing as a free lunch. Personal accounts are being sold as a free lunch. Every financial commentator should be against them for that alone.
August 2, 2006 11:48 AM | Reply | Permalink