Arnold Kling  

Remarks on Safe Assets

Some Health Care Reform Discus... Pro-Soviet Neocons?...

The issue of "safe assets" is important and complex. I have a lot of thoughts on the topic, but they do not yet cohere into a whole.

1. People really like safe assets. A lot of effort, both in the private sector and in the public sector, goes into creating assets that are safe or that appear to be safe.

2. Examples of safe assets include insured deposits, expected benefits from Social Security and Medicare, short-term Treasuries, and inflation-index Treasuries. Are there examples of purely private safe assets, or do all assets now considered safe have government guarantees attached to them?

3. When the financial crisis hit, assets that had been considered safe were suddenly deemed to be unsafe. There are people who blame the private sector for stupidity and deception in turning high-risk mortgages into safe assets. There are some who blame the Fed for keeping interest rates too low. There are some who blame government housing policy. My own view, as you know, is that capital regulations explicitly told banks that AAA-rated assets were safe, which means that in my account capital regulations play a significant role. Financial innovation served to manufacture AAA-rated assets out of risky mortgages. The regulators mostly applauded this.

4. Do ordinary investors understand risk? I tend to think not. I find that when people ask me for financial advice, they seem to expect me to be able to provide suggestions for ways in which returns can be high and certain. When I suggest an index fund, a person might ask, "How much interest does that earn?" When I infer that they want to earn a certain rate of interest, I suggest an inflation-indexed bond. When I explain that the return on that is less than 2 percent, adjusted for inflation, they think they are being cheated. Because very few people really understand risk, there tends to be a lot of potential profit to be made by taking advantage of people's desire for assets that appear safe, even if they are not.

5. There is a popular conception that stocks are less risky in the long run than in the short run. This cannot possibly be correct. As Zvi Bodie points out, any option pricing model would break down if stocks were less risky in the long run than in the short run. I notice that the popular conception that stocks are less risky in the long run is rarely challenged by professional economists. Either they do not know any better, or they are not willing to play the skunk at the party.

6. A fair amount of real-world risk is risk of malfeasance. That is, if I invest in Phil's fruit tree, I have to worry about all sorts of things that Phil could do to take advantage of me. A lot of the value created by financial intermediaries involves following customs and adhering to contracts that minimize the costs of preventing malfeasance.

7. Once you have the most cost-effective anti-malfeasance approach given the state of technology, about the only thing you can do to make assets safer is to diversify. It seems that just about everyone is under-diversified. Perhaps this is due to high costs of guarding against malfeasance with a diversified portfolio. Or perhaps it is due to the presence of lots of government-guaranteed assets, which gives people an alternative to diversification if they want safe assets.

8. Does government have a comparative advantage in diversification? One can argue that government is able to guarantee assets because it is able to diversify across more states of the world. So, the government can guarantee its promised benefits under Social Security, while any private annuity would be subject to the insurance company failing under some scenarios.

9. Alternatively, government's ability to guarantee assets rests on its ability to redistribute wealth by force, using taxation. I am inclined to think that this is the more important source of the government's ability to issue credible guarantees.

10. I suspect that the result of government guarantees is to increase the amount of saving and investment. Because banks can invest in risky assets and, with the government's help, issue safe liabilities, we have more capital investment than we would otherwise.

11. What is the optimal amount of safe assets provided by the government? If you want to say "none," that means you have to put up with a lot less capital investment. Maybe that is optimal, but I suspect otherwise. I suspect that, up to a point, creating safe-seeming assets out of risky investments makes us better off. On the other hand, it could be that we typically have way too much risky investment, because government errs on the side of creating too many safe-seeming assets.

Comments and Sharing


COMMENTS (20 to date)
ed writes:

Perhaps stocks could be "less risky" in the long run than the short run if you consider mainly the "risk" of falling behind your neighbors rather than the risk in levels. If the economy grows slowly, your stocks will perform poorly in the long run, but your neighbors will be poorer as well, so in comparison you won't look so bad. (See Eric Falkenstein for much more on this kind of thing.)

kebko writes:

Isn't the problem in #5 just the fact that there are different types of risk? Doesn't volatility account for much of what we call risk? Short of complete societal collapse, wouldn't it be mathematically impossible for broad stock indexes to be more volatile as the time period increased?

drscroogemcduck writes:

maybe the 'stocks are less risky in the long run' is in reference to being less likely to be under the the initial investment the longer you hold the stock even though the variance is much larger.

the poker variance simulator gives you a pretty good idea of long run vs short run when you have a positive edge

if i put in 100,000 hands with the default positive edge i get around 2% falling below zero but there is around $3000 difference between the top 10% and bottom 10% of results. if i put in 100 i get around 50% falling below zero and about $100 difference between the top 10% and bottom 10% of results.

longer run has more variance but more likely to have a positive outcome which is how i think a lot of people interpret risk.

"As Zvi Bodie points out, any option pricing model would break down if stocks were less risky in the long run than in the short run."

Actually I think he's just saying that short term continued rolling over must be equal to long term buy and hold or else there's a pure arbitrage, assuming no transactions costs, etc. This may also be some argument against mean reversion, saying it would allow an arbitrage if (and these assumptions will be far from the truth) there is perfect liquidity, no transactions costs, etc., etc.

But Arnold, I hope you read my long post, but let me say briefly here this intuition:

If you buy stocks when you are young, it's like you are taking the riskier asset when you have time to adjust to how it unfolds, adjusting your savings rate, your consumption rate, what you psychologically grow accustomed to, how many hours you work, even career decisions. You can, for one, be much more likely to prevent a big drop in consumption, and in the real world the direction of consumption, it's context, matters tremendously, even though in typical models the utility function only depends on the absolute level.

If stocks go bad at age 60, it's too late to adjust these things of the last, say, 35 years, and that can be devastating to your utility. Thus, it's a much greater risk.

These factors are extremely important in the real world, but they don't exist in simple models with simple utility functions.

[Serlin's longer post is available at Ed.]

Arnold Kling writes:

You have presented an intuition for why stocks might be less risky for young people. However, it is not the intuition. The intuition goes something like this: yes, stocks go up and down in the short run, but in the long run they only go up. You just have to hold them long enough.

Mike Rulle writes:

It occurs to me I have never seen the total return for a "market weighted index" of AAA securitized debt instruments over any time frame, particularly the last 5-10 years.

An analysis of default rates---total and by geography for real estate--- would be interesting as well. We have seen countless estimates and studies of "marked to market" losses over time (aren't a lot of those coming back into bank P&Ls?---probably some), but never an analysis, after the fact, of just how bad, or not bad, those AAA securities ended up.

Is anyone aware of such a study? Maybe its just a symbol on Bloomberg and I cannot find it.

Phil writes:

"The intuition goes something like this: yes, stocks go up and down in the short run, but in the long run they only go up. You just have to hold them long enough."

Isn't this somewhat true-ish? Suppose a stock's annual return is uniform in [-5%, 15%]. Over 25 years, isn't the probability of an overall negative return really quite low?

guthrie writes:

As an economist, what would you look for in contracts 'that minimize the costs of preventing malfeasance'? What is the language you would want in such a document?

John Bailey writes:

I think that you are missing the problem of "public malfeasance." None of the public guarantee programs, including health, retirement, deposit and unemployment insurance, mortgage guarantees, and public debt are close to being adequately funded to deliver the benefits that were promised.

This is true internationally as well as in the U.S.

The Federal government is projected to run a deficit of between $1 and $2 trillion per year indefinitely. State governments are projected to add between $150 - $300 billion per year as well.

The $2.5 trillion in new debt generated over the last 18 months was largely funded by the Fed combined with around $500 billion in foreign government loans.

Now that the Fed is supposedly going to end its monetization of new government debt, it is now clear to me how the deficit will be funded.

Can anyone suggest a mechanism?

Scott writes:


French and Fama just posted on Point #5 last month on their blog. Not sure if you have seen it, but it is brief and worth taking a look at.

Nick writes:

"The intuition goes something like this: yes, stocks go up and down in the short run, but in the long run they only go up. You just have to hold them long enough."

The real inflation adjusted price gains on the DJIA (ignoring dividends) is close to zero, or possibly negative after taxes and fees, even if you go back to the 40s.

People seem to rely on the idea of price gains now and not income, which to me seems odd, in the past dividends were valued over all.

In my opinion your financial advice is sound, i would consider 2% over inflation to be a good buy (however taxes on TIPS are a nightmare).

I think people expect you to have some great sage wisdom but I believe most people approach investing as something akin to a casino game now, where as in reality the best portfolios for the average investor should be pretty boring.

Jared writes:

Point #5 is assuming that option pricing models are correct. But anyone who looks into the fundamental assumptions underlying Black-Scholes should understand that it's a clumsy model. There is no empirical evidence to show that stock returns follow a normal distribution, that there is constant volatility, or that stock prices follow a strictly random walk pattern.

I wouldn't rely on such an unrealistic theory to base my real-life investment decisions.

Warren Buffet's writing of long-term puts suggests to me that he also disagrees with option pricing theory.

andrew writes:

In the long run, the return on stocks is determined by the future cash flows of the underlying companies. In the short run, the price movement of a stock has almost nothing to do with the future cash flows.

If you buy some version of an index fund that has broad exposure to the entire US economy, you are holding a small slice of GDP. While it is theoretically possible that GDP 30 yrs from now could be lower than GDP today, it's pretty unlikely. At least it's a lot less likely than the possibility that the stock market could decline in the next year.

So if you hold them long enough, stocks are almost certain to go up. Not 100% certain because nothing is, but within the realm of financial forecasting, this is as good a bet as there is. You can be even more certain that the range of outcomes for stocks over 30 years will be tighter than the range of outcomes over the next year. So in that sense, stocks are indeed less risky the longer you hold them - at least in a world that is relatively stable.

In a world that is unstable, then you have to consider things like "the odds of nuclear war are greater over 30 years than over 1 yr", so in that sense there is more risk over the long term than the short term. But if you are considering that then you should also consider the riskiness of alternatives to holding equities, eg what will happen to US treasuries in the event of nuclear war?

I think for the purposes of most individuals who are planning for retirement, it makes sense to think of stocks as being far less risky over the long term than they are in the short term. This is not the same as saying buy stocks at any price, I hasten to add.

As to options pricing, well, good luck trying to buy a 30 yr option on a stock.

andrew writes:

Treasuries are not really "safe" even though people think they are. No fixed income instrument is. The risk lies in inflation. If by "safety" you are talking about protecting your real wealth over extended periods of time, Treasuries are extremely risky.


I'm not sure that's thee intuition.

As I think Zvi Bodie would say, the odds of a negative compounded return go down as the holding period gets longer, but they don't go to zero over human lifetimes; they still stay significant, and even though they do down, the amount of money at risk goes up.

Nonetheless, the intuition you give may still be significant due to things like true human preferences are complicated. People care about more than just mathematical variance. Indeed, it's easy to construct two theoretical assets with equal expected returns and distributions that have equal variance, but one would be considered far riskier to almost everyone. Also, the government safety net and taxes play a role in long run versus short run risk.

But I do think my intuition(s) is a strong one. It's much easier to avoid psychologically devastating consumption drops in response to bad market luck when young than when old.

Master of None writes:

Professors should run through this list on day 3 of Finance 101, and circle back to it throughout the course.

...Anyone who invests should read Arnold Kling's post on "safe" investments...

fusion writes:

Treasuries are not really "safe" even though people think they are. No fixed income instrument is. The risk lies in inflation.

TIPS are "safe." They adjust for inflation (and don't start that CPI is not a reliable gauge of inflation0.

fusion writes:

In the long run, the return on stocks is determined by the future cash flows of the underlying companies.

Companies don't last forever. Consider how many of the largest companies from 100 years ago are still in business.

Returns on stocks are also determined by p/e ratios. If p/e drops, so does return. p/e can drop sharply in a short amount of time.

Look at financial history. It provides a useful antidote to the "stocks always go up" notion.

Max writes:

[Comments removed for supplying false email address. Email the to request restoring your comments. A valid email address is required to post comments on EconLog.--Econlib Ed.]

Dave writes:

Anyone who doubts #5 should think about what happened to a Japanese investor in 1990 to the present.

Comments for this entry have been closed
Return to top