A pointy-head post about issues

by John Quiggin on February 16, 2005

I’d just about finished this lengthy post when I got the news that our readers and fellow bloggers are calling for lots of juicy attackblogging instead of dryasdust issues analysis. But it’s done now, so I’m going to post it anyway.

Matthew Yglesias had a well-argued piece a couple of days ago on Social Security and the Efficient Markets Hypothesis (EMH), in which he quoted me on the (generally left-wing) implications of rejecting the EMH. This spurred me to start on a post (or maybe a series) on the EMH, the equity premium and the implications for Social Security reform. Most of what I have to say is consistent with what Matt and others have said previously, but perhaps there will be a bit of a new perspective.

I’m going to look at three plans[1]. First, the status quo, that is, a scheme where the Social Security fund is invested in government bonds and pays a defined benefit. Next, diversification, that is, investing the Social Security fund in a diversified portfolio of debt and equity, but still with a defined benefit structure. Third, privatisation private accounts personal accounts, that is, allowing individuals to make their own investments and reducing the defined benefit correspondingly.

The main reason the alternatives might yield substantial benefits is the equity, premium, that is, the fact that, on average, returns to equity have been much higher than returns to debt in the long run. So, it might appear that investing Social Security money in equity, either individually or through the Fund could yield better returns than the current policy of holding only government debt.

This post will be about the implications if the EMH is exactly, or approximately true.

The easy case is when the EMH is exactly correct. In this case, the equity premium is simply the market’s aggregate evaluation of the required premium for the additional risk involved in holding equity. The EMH requires that each individual hold their own optimal mix of risk and expected return at the prevailing market prices, and that they take full account of expected taxes, Social Security payments and so on.

Under the EMH, neither of the reforms will make a difference. If the Social Security fund shifts into equities, individuals will reduce their own holding of equity until the desired balance is restored. Moreover, individuals will be entirely indifferent about personal accounts. If forced to open them, they will, initially at least, invest entirely in bonds, so restoring their (by hypothesis, optimal) status quo ante.

One important point about the EMH that’s relevant at present is that the risk associated with the equity premium has to be real long run risk. If equity investors were guaranteed of coming out ahead over, say, 20 years, no significant premium could be sustained. So the US experience, where a 20 year holding period has almost always been enough must be somewhat atypical. Shiller makes the point that other countries have experienced longer periods of poor stock returns, while still exhibiting an equity premium.

A more interesting case is when the EMH nearly holds. Suppose that the market is efficient in most respects but that some people are credit-constrained or face high borrowing costs. As a result they can’t hold as much equity as they want, since they can’t borrow to get it. This gives a case for either of the proposed reforms, since directly or indirectly, this gives credit-constrained households exposure to equity.

Note however, that almost any violation of the EMH leads to interventionist policy conclusions unappealing to free-market types. For example, given credit constraints, it’s generally desirable to increase the consumption of the young, for example by subsidising post-secondary education. something that is commonly stigmatised as “middle-class welfare”. So, as Matt says, invoking EMH violations, implicitly or otherwise, is a very dangerous intellectual strategy for supporters of privatisation.

fn1. In each case, I’ll ignore the underfunding issue, the fact that Social Security taxes at current rates will eventually be insufficient to fully fund promised benefits, under current rules. Whichever scheme is adopted, some combination of lower benefits, higher Social Security taxes or general revenue support will be required.

{ 29 comments }

1

Kevin Donoghue 02.16.05 at 12:20 pm

Some advocates of “reform” seem to have the idea that if the proportion of wealth held in equities rather than bonds were increased, productivity (or something) would somehow rise – there would be an impact on the real economy due entirely to shuffling wealth from one kind of paper into another.

I have never seen the idea spelt out, which leads me to suspect that it has roughly the same tolerance of daylight as a vampire. How anyone who believes such a thing can be described as a conservative is beyond me.

2

Jason Kuznicki 02.16.05 at 1:58 pm

“Under the EMH, neither of the reforms will make a difference. If the Social Security fund shifts into equities, individuals will reduce their own holding of equity until the desired balance is restored. Moreover, individuals will be entirely indifferent about personal accounts. If forced to open them, they will, initially at least, invest entirely in bonds, so restoring their (by hypothesis, optimal) status quo ante.”

Isn’t this assuming an awful lot? Social Security is (like it or not) an intervention into the securities market. EMH would not seem at all to predict that individuals will act to sustain the ratios of investment created by what is, after all, a giant artificial intervention. How on earth can we assume that the present system is the most efficient?

3

Nicholas Weininger 02.16.05 at 2:34 pm

To concretize Jason’s point, suppose EMH holds and suppose that under a totally free system (no SS benefits and no SS taxes), Investor A’s optimal risk/return mix would lead her to put aside 18% of her total income for savings, 8% into bonds and 10% into equities. SS now takes 12% and puts it (effectively) into bonds. So the closest she can get to her optimal mix is to put 6% post-tax into equities and nothing into bonds.

If private accounts let her invest 4% of that 12% as she chooses, she’ll put it into equities, thereby achieving her optimal mix, which she couldn’t do before.

What the post’s analysis seems to be missing is that it’s not only credit constraints which may cause people to be unable to achieve their optimal investment mix; SS itself may do that, by limiting how risk-tolerant they are allowed to be.

4

dave 02.16.05 at 2:37 pm

For example, given credit constraints, it’s generally desirable to increase the consumption of the young…

It’s a modest proposal, but I fear that it might have unbeneficial long-term consequences.

5

jet 02.16.05 at 2:43 pm

Could someone help me, I missed this point. How does moving Social Security from government bonds to investment capital not help the economy. Doesn’t cheaper investment capital make it easier to fund new projects and businesses? And why is it a safe assumption that Social Security would be moved back into government bonds by private individuals? Government bonds have had a real return of close to 0% because of inflation, hardly a good choice.

And I don’t understand the conclusion that total investment in the market would be adjusted back to current rates because of the “imbalance” of new capital. Isn’t it just as likely that we would see higher GDP growth and a quick adjustment the other way, so that current return rates are maintained?

6

robert the red 02.16.05 at 3:14 pm

How does moving Social Security from government bonds to investment capital not help the economy. Doesn’t cheaper investment capital make it easier to fund new projects and businesses?

Unless the Fed govt cuts spending, then the investment in the capital market by the SS system (personal accts or govt fund) will be offset by the govt borrowing the same amount from the worldwide public. So there would be no net increase in the capital stock.

7

BigMacAttack 02.16.05 at 3:15 pm

Can you believe in the EMH and in the EP? (Unless you believe that we are so risk averse that socialism becomes a good idea?) And what about supply and return and belief in the EP and the EMH?

Only SS invests in bonds and pays stock like returns. Not sure how we can achieve that more than optimal result via a free market. Score one for the state.(That was ironic)

8

Kevin Brennan 02.16.05 at 3:20 pm

Jet, yes, you did miss the point. If privatizing Social Security leads to money being shifted out of government bonds into equities, then the EMH has been violated. EMH suggests that the current distribution is the correct one because the markets are the most efficient possible way of determining the correct values.

The preferences of Investor A are irrelevant to the post. EMH doesn’t say anything about whether a particular person has their optimal mix; it just says that on the whole, the markets achieve as close to an optimal mix as theoretically possible.

9

Kevin Donoghue 02.16.05 at 3:45 pm

In any case Investor A can borrow in order to get her desired holding of equities.

(If she can’t then we are back to the case considered by John Quiggin: “people are credit-constrained or face high borrowing costs.”)

10

Nicholas Weininger 02.16.05 at 3:51 pm

Kevin, I don’t see how the preferences of an individual investor are irrelevant. EMH, by John’s own account, says that “each individual hold[s] their own optimal mix of risk and expected return at the prevailing market prices” (emphasis obviously mine). Different individuals have different levels of risk tolerance and will pick different mixes. My point is that if SS constrains the mixes available to more risk-tolerant individuals– and it does– then it is an obstacle to markets’ achieving as close to an optimal mix as possible.

11

Nicholas Weininger 02.16.05 at 3:55 pm

note the last comment was to Kevin B., not Kevin D.

12

Nicholas Weininger 02.16.05 at 4:05 pm

And also, even though in reality credit availability is certainly constrained, John’s final point is still wrong unless you implicitly assume a sort of EGH, an Efficient Government Hypothesis. From the premise that an all-knowing and perfectly public-interested government *could* calculate a level of (e.g.) education subsidy that would improve people’s preference-satisfaction ability, it does not remotely follow that government should be given the power to subsidize education.

13

Donald Johnson 02.16.05 at 5:02 pm

John Q, intellectual inconsistency is a dangerous strategy if the public understands the issues. Most of the time that’s not a problem. I’m not bashing the public, btw, since I don’t follow most of the technical arguments here either. For all I know, the failure of the EMH means we should have American versions of five year plans, but even if that were true, it wouldn’t make the slightest difference politically.

14

Bruce Wilder 02.16.05 at 5:07 pm

It seems to me that Kevin D is right in the first instance, and introducing the EMH is an unnecessary and confusing elaboration.

In 2025 or 2050, there will be a GDP, and the national income will get divvied up. If domestic savings has increased sufficiently to displace foreign ownership of stuff, then domestic claims on national income will be proportionally larger.

If policy has forced larger sums into the equity markets, per se, it is reasonable to suppose those larger sums will have simply disappeared into bad investments (i.e. investments with a negative present value at the time they were made). Unlike the SS trust fund, the NYSE is no lockbox. If you forcefeed it enough money, it just spills the excess.

A reasonable interpretation of Bush’s vague proposals is that he plans to “steal” the accumulated trust fund. The middle class will trade its claims on future tax revenue for the “choice” of purchasing currently overvalued equities from wealthy people, who, in turn, will be given the opportunity to 1.) collect “management” fees from the middle class, and 2.) use the proceeds of their stock sales to buy up trillions of new Federal debt issued to “cover” the transition costs.

All the unnecessary nonsense about the EMT becomes so much smoke to cover a plan for a trillion-dollar heist; a future generation of Americans will wake up in a world where the wealthy are being paid a enormous share of Federal revenue in debt interest, and the retired middle class is eating cat food (but may have the dubious honor of “owning” the cat food maker).

15

Kevin Donoghue 02.16.05 at 5:32 pm

Nicolas Weininger,

Certainly if the choice is between allocation of resources by an inefficient market, or by a corrupt and/or incompetent government, then many of us would sooner take our chances with the market. However the arguments for SS privatisation which I have seen do not include the assertion that the existing system is in the hands of knaves or fools. If it is then reform of the system is in those same grubby hands, so we should be highly suspicious.

Actually, if the debate were about designing a system which minimises the government’s power to do harm, it would be a much healthier debate. But right now that is OT.

16

Nicholas Weininger 02.16.05 at 5:50 pm

On reflection, my Investor A example concedes both too little and too much. In a narrow sense, John and Kevin D. are right: if the EMH is correct and this implies people can borrow costlessly and SS really just forced A to invest 12% of her income in Treasury bonds paying a market rate of interest, she could borrow 4% of her income at that same market rate, use her forced bonds to pay the interest, put the principal in equities and be no worse off.

But of course in any remotely realistic world A can’t do that, and the reason is not just that real markets have borrowing costs, it’s because of the nature of SS itself. Both John’s footnote and my parenthetical “(effectively)” ignore too much.

In fact SS is not the same as people being forced to put 12% of their income into government bonds, not even close. One reason is of course that SS has a large redistributionist component. But even ignoring that, there’s still the fact that SS benefits are set by the grace of Congress and can be changed arbitrarily at any time. When you buy a Treasury bond you’re buying a contractual promise from the government to pay you, the bondholder, a specific rate of return over a specific term. Paying SS taxes gets you no such promise.

Now SS defenders will respond that simply abolishing SS tomorrow is politically unfeasible, perhaps almost as unfeasible as defaulting on the debt. But significant changes in SS benefits and/or taxes, and thus changes to the “effective rate of return” to SS taxpayers, are not only feasible but, as John’s footnote notes, almost certain to occur. And the likelihood of these changes means that SS itself creates uncertainty that raises the cost of getting one’s optimal debt-equity mix.

17

Nicholas Weininger 02.16.05 at 6:02 pm

Kevin D.: indeed, the existing system is in the hands of knaves and fools, and so is the reform process. Don’t mistake me for a supporter of the Bush plan. I favor straight-out cuts in SS over private accounts, in large part because I don’t think the government can be trusted to set up private accounts in a transparent or efficient way. But the validity or invalidity of John’s points doesn’t essentially change, I think, if you substitute cuts for accounts.

18

Kevin Donoghue 02.16.05 at 6:59 pm

Nicholas: cuts are a different kettle of fish. Abolishing both payroll taxes and payments to retirees increases the incentive to work, both for those currently employed and for the luckless retirees. We don’t need the EMH to tell us that. It is a much different proposition from swapping debt for equity, which is just shuffling paper claims.

As for SS creating uncertainty, that’s true of any government policy and it is a useful argument against too-big government. It is also an argument against too-small government; there is a lot of uncertainty in failed states. The debate about the right size for government will never end. It seems to me there is a good case for some form of safety-net but there is a trade-off between generosity and preserving incentives.

19

John Quiggin 02.16.05 at 9:18 pm

“But of course in any remotely realistic world A can’t do that [borrow at the real bond rate to finance her desired equity position]”

which is why EMH is false and you need to look at the credit-constrained case.

20

BIgMacAttack 02.16.05 at 10:52 pm

John Quiggin,

‘“But of course in any remotely realistic world A can’t do that [borrow at the real bond rate to finance her desired equity position]”

which is why EMH is false and you need to look at the credit-constrained case.’

What difference does it make if they can or cannot?

Even at a 1 – 3 % interest rate the return is significantly reduced? I mean how large an EP is being claimed? Presumablely stocks must pay some premium for the risk?

21

Jack 02.17.05 at 12:38 am

bigmac,
the difference it makes is that poor constrained investors might not be able to take appropriate advantage of equity investment if their social security contributions are paid into forward earnings linked annuities.

Equities are expected to earn extra to compensate for their risk but that is not what is being left on the table. Historically equities have earned even more than supposedly necessary to compensate for the risk. This is the equity premium puzzle of Prescott and Mehra.

In fact there are all sorts of complications with the theory — different people have different risk aversions, different liabilities and different fixed investments like house. The point is that there is no argument for private accounts that is compelling in the first order. No direct increase in investment, no obvious free uplift in investment results, nor even any major help on solvency as far as I can work out. Any such arrangement is likely to be very expensive too, perhaps as much as 100 times, certainly an order of magnitude. What there is in the proposal is mostly bizarrely political. Will social security funds become available to creditors? Will unused funds be heritable? Will the poor reveal hitherto invisible Randian heroism when deprived of a safety net for their old age? How much would go on fund management? $5bn a year? $10bn?

22

BigMacAttack 02.17.05 at 1:50 am

Jack,

I don’t believe in much of an EP. (Data I saw did not seem to take transaction costs into account) So for me all investors are credit constrained when taking advantage of a snall or non-existent EP.

So I pretty much agree with you. The argument for privatization is based on secondary fators. Less government spending and other factors. (I am really not a fan of private accounts.)

Which is maybe what John was saying? If you have an EM you cannot have an EP. It seems to me that is true by definition.

23

Nicholas Weininger 02.17.05 at 2:36 am

John: yes, I agree, any version of the EMH strong enough to imply such borrowing ability is ipso facto false. Indeed, I wonder if you can cite examples of people claiming to believe in such a strong version. Certainly you yourself have discussed much weaker and more defensible versions in previous posts, and this one begins to smell of straw.

24

John Quiggin 02.17.05 at 5:17 am

Nicholas, costless borrowing is part of the standard version of the EMH used implicitly all the time, particularly when generic assumptions about the virtues of free markets are used as a basis for policy analysis. If you don’t believe strong versions of EMH, you should be wary of all such analysis (of course, you can still go for second-best arguments).

For a lot of purposes the costless borrowing assumption is not a bad approximation for financial markets, which is where most empirical application takes place. Most direct, large-scale participants in these markets can borrow at low cost (that is, low margin above the market interest rate) so arbitrage arguments and similar work well. On the other hand, as we’ve seen it works very badly in predicting the equity premium.

25

jet 02.20.05 at 6:58 pm

Kevin,
“EMH suggests that the current distribution is the correct one because the markets are the most efficient possible way of determining the correct values.” And that is because the market has so much control over tax rates and government spending? Maybe the current distribution post-government-tax-slurp is at its most efficient, but that isn’t really what we are discussing.

I can’t believe this debate dropped so far down into the nuts and bolts of if government bonds perform the same as investing in the market. Government bonds are just a stubstitue for increased taxes, while investments in the market do real work in that they create wealth.

26

Kevin Donoghue 02.20.05 at 10:06 pm

Jet (to Kevin Brennan):

“Government bonds are just a substitute for increased taxes, while investments in the market do real work in that they create wealth.”

The first part of this statement is true if taxpayers are very far-sighted. They treat a tax cut which is not matched by a spending cut as merely a postponement of the tax-bill (Ricardian equivalence). Hardly anyone believes this is empirically true. Still, there may be a good case for assuming something of the sort when it comes to a topic like SS. It helps to underline the principle that there is no such thing as a free lunch.

The statement that investments in the market create wealth confuses two meanings of the word investment. Stockbrokers refer to the purchases of equities as investment. Economists do not. No wealth is created – all that happens is a transfer of title to existing assets.

27

jet 02.20.05 at 10:25 pm

My head isn’t pointy enough for this discussion.

28

Cleve Blakemore 02.21.05 at 4:51 pm

There is no point to doing thorough and comprehensive widescale experiments in social engineering through central control if we do not follow up the severe problems we create with even more stringent controls imposed on the already overcontrolled marketplace. The point of instituting central control in the first place is to cause enough sheer hell resulting to justify further controls. One of my primary beefs is that people should be willing to accept that if a management class is willing to accept the enormous burden of the greater good of mankind by drawing up virtually infinite lists of rules for others to follow, they have got to have legislative access to the best looking and most fertile young females during their off hours. That’s not me talking, that’s economics, damn it.

29

jet 02.22.05 at 2:57 am

Ol’ Cleve, all over CrookedTimber with the jokes. Why doesn’t he take that fearsome step and actually engage on an issue?

But if all the “most fertile young females” were reserved for the commie socialist, I’d convert tonight.

Comments on this entry are closed.