I have been betting on elections for years with mixed
success. Paying Queensland’s “Top Level Ted” Lyons back in 1980
was particularly painful after making the mistake of believing our own ALP
opinion polls when helping Bill Hayden out on the campaign trail. Polls, as I
painfully learned, jump around too much in the weeks before an election to be a
truly reliable guide. Even those on election eve can be spectacularly wrong, as
Jeff Kennett found in Victoria.

Markets are not always right either, especially where
relatively small amounts of money can make large movements in the prices. I was
reminded of this recently after an old friend, who has survived comfortably for
decades betting for a living, sent me some comments on a piece I had written
saying that the bookmakers’ Federal election odds were a better guide to the
real chances than the pollsters.

My friend wrote:

The hundredised market at 11/8 Labor [I had
published the ALP as a 42% chance of winning] and 8/11 Libs (which translates to
being a 58% chance) bothers me. We know the market is hard to beat, but it
can be wrong.

Without seeing those market odds, I would
have said Labor was a big underdog. My reasoning was the old saw: Oppositions
don’t win elections, Governments lose them.

I hark back to the long rule of Menzies. I
knew many traditional Labor voters who said, in effect, “we are doing OK, let’s
not change the winning game”.

I think the voters today are doing too well
to vote for change. And I don’t think the economy and the resources boom (while
doomed eventually to overheat) is likely to crash in the next 15 months or so.
Could happen, but unlikely.

Back to the betting market. What you said
about it being the best guide is right. But it can sometimes be inefficient.
Before the last election, some US academics published some
results, focusing on Centrebet, demonstrating the efficiency of the Election
betting odds.

Centrebet, in my judgment, wrongly took
this to be a tribute to their expertise in forecasting election results. That is
dangerous. They should see themselves as a conduit matching supply to
demand.

When Latham beat Beazley in the Labor
ballot before the last election, Centrebet immediately “blew” Labor several
points. That was a reflection of their opinion and not a response to market
demand. I thought they were wrong for the same reasons I back a lightly-raced
unproven horse to beat a better performed horse whose performance has plateaued.
For the same reasons sporting teams spend a big part of their budget giving
contacts to young, relatively untried guys.

I backed Labor, on the premise that at that
point, whatever the right odds, they were a better chance with an untried new
guy, than with the two-time loser. It was hard to get set for much. Bookies
forever talk up the bets they claim to lay. Getting down for very much is
another matter. I shortened Labor considerably for something like $2,000. That
took several bets, because the limits didn’t allow very much at each odds. The
market did settle for quite a long time, at the point I had moved it to. Was the
efficient market more efficient before or after my small bet?

At this stage, I suspect there is little
betting demand and Centrebet’s odds are a reflection of their opinion – not
tested by money. The betting exchanges will be a much better guide, but they
haven’t yet opened election betting.

With the reservation about markets not being fully
tested by money noted, herewith the opening Crikey Federal Election
Indicator:

Coalition 57.3%
Labor
42.7%