Actual and desired wealth distributions

Jeff Smith makes an interesting point about this plot from Mother Jones. It depicts what the distribution of wealth actually is in the US, what people think it is, and what they’d like it to be. But as EconJeff points out, the plots depict the total amount of wealth held constant, and it seems doubtful that the question actually specified that. People would likely imagine that their desired distribution involved more total wealth – increasing the wealth of the middle class at no cost to the rich. Practically speaking, a redistribution would involve less total wealth; it can only be achieved through some kind of tax, and taxes are distortionary.

For the record (in light of our earlier discussions of inequality), I would personally support a hypothetical policy that transformed the top distribution into the bottom one so long as it came at no cost in terms of total wealth, or if the deadweight loss were sufficiently small. As noted above, however, such a policy is basically inconceivable.

Advertisements

About Jason Kerwin

http://nonparibus.wordpress.com/author
This entry was posted in Uncategorized. Bookmark the permalink.

5 Responses to Actual and desired wealth distributions

  1. How inconceivable is it? I’m really curious – are there solid empirical papers on this subject? Our top marginal tax rates have dropped precipitously in the past few decades – is there some evidence that we’ve gotten more wealth as a consequence? Beyond that, there are plenty of redistributive policies other than taxes – see Hacker and Pierson (2010) for a lengthy discussion. Finally, isn’t the “distortion” caused by the tax balanced against the potentially larger gains from the production of public goods that would otherwise go underfunded? Why is such a policy inconceivable – is that based on a series of studies or just on your theoretically-informed priors? I’m really curious here as to how one would test/show this.

  2. Jason Kerwin says:

    The claim I was trying to make – which is still up for discussion – is that it would be difficult to go from the top distribution to the bottom one without a substantial loss in overall output. That view is informed by evidence on taxes like this newish article: The fact that I’m assuming taxes are the only option comes from the bits of public finance that have been covered in my theory courses. Basically any policy that’s capable (in general) of moving income across the distribution amounts to a tax, but it’s possible that the US has particular policies that favor the rich at the expense of the overall economy. Fix those and you’ll redistribute income while yielding net gains at the mean.

    There’s a risk of double-counting when you consider whether we might benefit from tax-funded public goods while redistributing income via taxes. It’s true that the value of public goods can offset any distortion from taxation, but that precludes taking tax revenues and handing them to people lower down the income distribution. What you’re going for is the idea that we reduce income inequality via a tax and then the public goods we purchase increase everyone’s welfare by enough to make up for a drop in income. That’s conceivably, and maybe likely, but arguing that it can keep average income constant requires putting a price on public goods, which is the sort of thing that koolaid-drinking economists like myself are happy doing but everyone else is pretty uncomfortable with.

    • Thanks for your prompt reply! I think whatever article you tried to link didn’t actually link?

      And you’re right, the two points I was making were somewhat disconnected. One was a more general point about taxes sort of always being a net negative, the other about the positive consequences of a more equal income distribution. Though, aren’t there (almost surely controversial) macro-models that suggest that, say during a massive depression with widespread unemployment, etc., money in the hands of poorer people (e.g. unemployment insurance) has a bigger multiplier effect and thus might increase the overall pie while simultaneously evening out the distribution?

      Last, like it or not, we put a value on government output already. While it was hotly debated from 1920-1955 or so, ever since, national income accountants have valued government output at precisely the cost of its inputs, no more or no less.* It’s a curious thing – by relying on these data, right and left economists all implicitly assume that government is worth exactly what we pay for it (no more, no less). But that’s a topic for another time!
      * The debate was actually about whether or not to discount some of government output as intermediate – the double-counting problem! – arguing that much of what government does is service to business, not final production (e.g. roads are used by consumers but also by business to get goods to consumers). The solutions for trying to allocate between these two uses were.. hokey and no one liked them, but theoretically it’s a pretty big problem that’s not really resolved (except by some careful hand-waving and changing what you assert the national accounts mean from one moment to the next).

  3. Jason Kerwin says:

    Here’s the link I was trying to post:
    http://aidwatchers.com/2011/03/tax-rates-and-development/

    Actual article is gated but accessible via MLibrary Proxy.

  4. Jason Kerwin says:

    By the way – if we value government services at cost then any non-Pigouvian tax will impose a net deadweight loss. To do what you’re suggesting and have it come out in favor of the policy we would need to accurately price public goods rather than just taking GDP data at face value.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s