In response to my implied criticism of Andrew Leigh for assuming that increases in inequality are bad and decreases good, but never specifying for what level of inequality would satisfy him, commenter Leopold responds:
one could turn the criticism around. Liberals believe in liberty – but how much liberty, exactly?
Leopold’s argument (I am paraphrasing here) is that preferences for greater equality or greater liberty are rules of thumb to be applied to specific circumstances, but there are cases where social democrats could accept less equality and liberals accept less liberty. We can’t always precisely calculate the final overall result of all these complex trade-offs to say what is the exactly right amount of equality or liberty. But this doesn’t invalidate the initial assumption that, all other things being equal, more equality or more liberty (depending on your philosophical position) is desirable.
I think Leopold’s point is reasonable. For example, I say that there should be less tax, and while I have clear pet hates among government spending programmes (eg FTB) that I think should be cut to reduce general tax rates, I never say exactly how much tax I think should be levied or what tax rates I would be happy with.
High-level political abstractions gives us intellectual tools that help organise our understanding of the world, but they don’t necessarily provide answers for specific problems. That requires far more detailed analysis.
But this reinforces my doubt about the value of composite indexes designed to aggregate different liberties or inequalities. There are some of these from a liberal perspective, such as the Heritage Foundation’s index of economic freedom and the rankings of freedom produced by Freedom House.
But these are rare compared to the enormous effort put into measuring overall income inequality. Andrew Leigh alone has about half a dozen academic articles published over the last three years that either measure high-level inequality or (if I recall them correctly) assume it is significant. I was thinking of him in particular when I made my original comment, because I think he is one of Australia’s best social scientists and certainly the best early career social scientist. Being highly productive he does more work on micro issues than most of his peers as well, but there is a lot of intellectual energy being devoted to studies that to me are of uncertain significance even from an egalitarian perspective. He agrees that :
I don’t have a strong sense of what the right level of inequality is. Indeed, I’m not even sure I have the right intellectual framework for answering the question.
This is due to the inherent massive complexity of income inequality. While it is affected by government policies at any given time, it is also the result of billions of individual decisions, actions and transactions taken over very long periods of time (as Andrew L’s work on intergenerational mobility shows). Very few of those decisions would have been taken with income equality in mind, and even those that were could not have more than a vague idea of the precise implications over the long term.
Indexes of liberty are very simple by comparison. Usually there is only one ultimate decision maker (the state), the consequences for liberty are expressly thought about, there are a relatively small number of considerations to be taken into account, and there are a relatively narrow range of likely consequences that can flow from the decision. Generally, we have enough information to classify countries according to whether or not they have satisfactory levels of economic and political freedom.
While I would prefer greater economic freedom in Australia and am concerned about attacks on political freedom via campaign finance laws, the overall situation is good enough to classify Australia as economically and politically free. I don’t think any social democrat could say whether or not Australia could be regarded as having a satisfactory level of income inequality.
So I don’t disagree with Leopold that our basic philosophical views can serve as useful guides to specific issues. But the more of these specific issues you try to build into one aggregate number, the harder it is to incorporate all the relevant trade-offs, and the less meaningful the final result becomes. The number abstracts too far from the underlying reality to be useful in judging whether or not apparent trends are good, bad, or neither.