The CPU has pretty much trashed its reputation, however, by its eagerness to produce opinions that are just too convenient and servile and so we now have the job done by a semi-independent think-tank, the Bauhinia Foundation which has been going for only about a year and probably still has some time left before it goes the way of the CPU.

But you obviously have to ask just what the Chief Executive's agenda may be when his foundation announces new findings with election time just around the corner. Is this study on household incomes a lead-up to a minimum wage law perhaps or to another round of rent concessions for public housing tenants?

I ask in particular because when I see firm conclusions drawn from faulty use of statistics, I generally suspect that someone has come to a conclusion first and then went on a hunt for facts to fit that conclusion. I cannot say it happened exactly that way in this case of course, but we certainly have faulty use of statistics here.

This study entirely ignores changes in consumer prices over the period it has surveyed. Median household income may indeed have been lower in 2005 than it was in 1997 but prices of goods and services bought with this income had also come down significantly.

The first chart sets this in rough perspective. The red line shows you a four-quarter moving average of median household income set to an index basis of 100 for September 2006 (the latest data available). The blue line shows you the composite consumer price index on the same basis.

The conclusion should be pretty obvious. Consumer prices have declined about as much since 1997 as median household income which means that in real (inflation-adjusted) terms median household income has not really come down at all.

In fact, the story for incomes is even better than that. The average household size has decreased from 3.3 persons in 1997 to three persons at present. Adjust household income for the size of household and you get about a 10 per cent increase in real household income per capita since 1997.

And this is borne out by other data. As the second chart shows, official wage statistics suggest real wages have risen slightly more than 10 per cent since 1997. I think we can take this as conclusive. The Bauhinia Foundation is wrong about falling incomes. Its researchers should have looked more closely at their numbers.

This is not to say that the foundation is wrong about another finding in its study that the gap between high and low income is widening. The evidence of growing income polarity is strong. Government officials do not like it but they have to admit it is a real trend.

Two things need to be taken into account about this. The first is that income polarity does not necessarily mean the rich are getting richer and the poor are getting poorer. It can as easily mean that the rich are getting richer faster than the poor are getting richer which is more likely to be the case in Hong Kong just now.

The second thing worth noting is that an income gap is not the same as a wealth gap. The income gap may have widened since 1997 but the collapse of property prices and of the stock market in that year may mean that the wealth gap between rich and poor now is still less than it was then.

This still leaves us with the question of just why our Chief Executive should want this study, error-ridden or not, and what he wants to do with it. There has to be a vote-sweetening tactic somewhere in this.