Cabinet approves auto-enrolment scheme

[Duke of Marmalade]

So this is where the flaw in your argument is. I don't think that large discrepancies in returns can exist permanently in financial markets like this. If the relative returns to a basket of equities are so great, why aren't they arbitraged away?

There is no scope for arbitrage as in the possibility to make risk free profits, and I don't think Colm is suggesting his proposal is entirely risk free.
All financial models of the market are based in some shape or form on the premise that there is a positive expected annual return on equities relative to bonds but a volatility about that return. Without getting too nerdish expected return increases (log) linearly with time whilst volatility varies with the square root of time. This is not only empirically observable but has a sound theoretical basis as well as being rather intuitive.
What this means is that the risk (volatility) reduces as a proportion of expectation with time.* Or in idiot boy language (needed for some in this parish) short term investors run more relative risk than medium term investors who run more relative risk than long term investors.
So it is not to arbitrage principles that we should have recourse but to supply and demand arguments. If the supply of equities was sufficiently small the demand from long term investors would be overwhelming and then the long term ERP would be pushed way down - though still positive. Shorter term investors would effectively be priced out of the equity market. However, observed ERPs of at least 3% p.a. suggest that long term investors do not dominate the market. Ergo long term investors get a "free lunch" - an ERP calibrated to a market which includes short and medium term investors who demand a higher incentive than long term investors. The key to understanding this phenomenon is to recognise is that all investors pay the same price.

* Example: expected annual return 5%, annual volatility 15% One year relative risk (V/E) = 15/5 = 3. Nine year relative risk = (15 * 3)/(5 * 9) = 1

 

[NoRegretsCoyote]

My biggest holding is Phoenix Group. At the current share price, I'm getting a dividend of 6.6% a year. If I were to put my money in an Irish government bond, I would get 0.06% a year (so I'm told : I don't have any). I'm confident that the Phoenix dividend is safe for many years to come.

This diversion into specifics is unconvincing. A 6.6% dividend for ten years isn't worth much if the share price falls 90% over the period.

However, observed ERPs of at least 3% p.a. suggest that long term investors do not dominate the market. Ergo long term investors get a "free lunch" - an ERP calibrated to a market which includes short and medium investors who demand a higher incentive than long term investors. The key to understanding this phenomenon is to recognise is that all investors pay the same price.

* Example: expected annual return 5%, annual volatility 15% One year relative risk (V/E) = 15/5 = 3. Nine year relative risk = (15 * 3)/(5 * 9) = 1

Many thanks for the long and detailed reply. You've more or less convinced me. I've never had the chance to study finance theory and it's useful to have the concepts explained.

One thing still not clear. The investment horizon of someone starting in AE is 40+years. Relative risk is not eliminated, but declines over time, you say. So shouldn't the ERP decline over long time horizons too? You suggest that it doesn't.

Assume perfect information and freedom of capital. Over a 40+year horizon a rational investor will choose equities over bonds, for all but the most risk averse. All else equal this should drive down bond prices and push up equity prices, but in turn reduce the premium of equities over bonds.

What am I missing?

 

[Duke of Marmalade]

One thing still not clear. The investment horizon of someone starting in AE is 40+years. Relative risk is not eliminated, but declines over time, you say. So shouldn't the ERP decline over long time horizons too? You suggest that it doesn't.

The point is the share price is the same for all and loosely speaking therefore the ERP is the same for all. Short term investors would want a higher ERP as they face a higher relative risk than long term investors but they all get the same so this points to the "free lunch" syndrome for these latter. The ERP is somewhat of a spiritual concept. It can never be measured even with hindsight. It is that margin that investors build in to their pricing of a share when evaluating their guesstimates of future cashflows from that share to compensate for the uncertainty. With hindsight we can see whether a premium was actually enjoyed but this could simply be a random observation.

Assume perfect information and freedom of capital. Over a 40+year horizon a rational investor will choose equities over bonds, for all but the most risk averse. All else equal this should drive down bond prices and push up equity prices, but in turn reduce the premium of equities over bonds.

What am I missing?

Excellent challenge, which can be summed up as "if long term investors have a free lunch available why are some paying for their lunch?"
I am going to wing it a bit here. It is a question of dietary requirements. The investors in 40 year bonds will be insurance companies and banks who will have technical reasons for doing so - matching liabilities of that duration. They will be aware of the equity free lunch and in fact will nibble at that but their digestive system (regulatory capital requirements) prevent them putting on the nose bags and gobbling up the freebies.
On a less metaphorical level it is again a supply and demand thing. The government control the supply of long bonds. They will only chose to supply enough bonds that they can get away at an acceptable price compared to more short term financing. They are certainly not going to try and lure long term equity investors who have access to a free lunch.
So I think my answer to your question is that the long bond market and the equity market are dominated on the demand side by completely different constituencies.
As I said I am winging it a bit here in responding to your excellent challenge, can anyone else do any better?

 

[Colm Fagan]

A 6.6% dividend for ten years isn't worth much if the share price falls 90% over the period.

You obviously missed my subsequent comment:

I'm also confident that I'll get at least my money back if I want to cash it (say) 10 years

Nevertheless, you and Brendan are right to criticise me for arguing from the specific to the general. I do it too often.
No such criticism can be laid at Duke's door!
Thanks, Duke, for your post, which even I could follow. I can even use it to try to answer Coyote's question

All else equal this should drive down bond prices and push up equity prices, but in turn reduce the premium of equities over bonds.

You're right, coyote, but the world is full of irrational investors who take a short-term view. I've even seen it in my own life (here I am again arguing from the specific to the general!) as chairman of the trustees of a DB pension scheme. One would think that a DB pension scheme qualifies as a long-term investor, but the trustees must always look to the short-term solvency position, as revealed at the end of each financial year. If the market value of the assets doesn't meet a certain minimum level specified by the Pensions Authority, the sponsoring employer must put plans in place to pay extra contributions or take other remedial action. This forces the trustees to take a short-term perspective.

 

[elacsaplau]

It should be recognised that the solvency standard (MFS) for on-going DB plans of medium maturity is, by some distance, the weakest in the world. (Imho, unacceptably so!).

It was never intended to be a particularly challenging hurdle.......it only became one principally because liabilities and assets assumed unanticipated trajectories and these divergences remained for very long periods.

Also, irrespective of MFS (i.e. it is unfair to blame the Regulator), trustees are legally obliged to protect the accrued benefits of members.