Spotting Worstall's Fallacy in the wild

joestiglitz.jpg

In a discussion of Joe Stiglitz's new book in The Observer we see this:

Back in 2008 the top 20% of households in the country were estimated to be worth 92 times more than the bottom 20%. The latest estimate puts the gap at more than 100 times. And a further £12bn of welfare cuts are planned by the new Conservative government. The gap between rich and poor is unquestionably widening.

That conclusion may or may not be right but it's most certainly not supported by the evidence which is given of the contention. That evidence coming from this ONS report:

Total net wealth is defined as the sum of four components: property wealth (net), physical wealth, financial wealth (net) and private pension wealth . It does not include business assets owned by household members, for instance if they run a business; nor does it include rights to state pensions, which people accrue during their working lives and draw on in retirement.

It's known as Worstall's Fallacy simply because our own Tim Worstall bangs on about it so often. We cannot measure inequality (or poverty, any number of other things) without taking into account the things we do to reduce said inequality (or poverty, or any other problem). It's only when we look at the post-attempt to solve the problem situation that we can turn our minds to whether we should be doing more, or possibly less, to try to solve this problem.

So, note that our definition of wealth there does not include that state pension: something we do to reduce the wealth and income disparities between those who can save for a private pension and those that cannot. Note that it includes private housing equity but not the capital value of a below market rate tenancy for life ("social housing"). It does not include the capital value of health care, or education, free at the point of use, for all the citizenry. It does not include whatever capital value we might ascribe to the social insurance policies that will provide us with an income in the case of economic misfortune.

We do all of these things because they make people wealthier. Perhaps not as wealthy as other arrangements might make them, but the essential driving point is that we consider health care, education, social insurance and so on make people wealthier. Thus, in our discussion of wealth we must include them. We cannot look solely at the market distribution of purely market wealth and even attempt to decide whether more or less should be done to try to change that distribution. We must look at the post- all the redistribution we already do situation to be able to make a decision.

To switch from wealth to income to make this point. The TUC has done the calculation about income inequality. Between the top 10% and bottom 10% the market inequality is about 30:1. Maybe that's too high, maybe that's not high enough, your moral choice. When we take account of taxation and benefits that falls, and when we take account of government provided services, that health care and eduation and so on, it falls again. To 6:1. Again, you can say that this consumption inequality is still too much, or not enough, your moral choice. But 6:1 is very definitely different from 30:1.

And what is the relevant ratio to be looking at if we want to make a decision upon whether to do more redistribution or less? Quite, it's obviously that 6:1 one, not the 30:1. And so it is with wealth or poverty or so many other problems. The number we need for our decision is the extent of the problem on the ground, not the extent of the problem before all of the things that we already do.

Assigning reasonable capital values to the effects of both the welfare state and government provided services brings the 10/90 wealth ratio down to anywhere between 20:1 and 5:1. We could and would defend anywhere in that range dependent upon assumptions. Is that too much? Not enough? Entirely up to your moral choices. But it's very different from that 100:1, isn't it, and it's also the relevant number we need to use when thinking about what to do next.