Much of the public debate on income inequality focuses on what is happening with market incomes. But most people generally accept that a certain level of inequality in market incomes is inevitable, and indeed necessary in order to provide the sort of incentives that generate economic growth. What really matters is inequality of income net of taxes and transfers. If increasing inequalities in market income are compensated by stronger redistribution, then there's less to worry about.
Unfortunately, that doesn't appear to be happening. A few weeks ago, I noted that Canada's tax-and-transfer system had become less effective in reducing inequality, as measured by the Gini coefficient. In this post, I'm going to look at how taxes and transfers have evolved for the various income quintiles. It reaches the same conclusion, but it also suggests a reason why the system has become less progressive.
- The data are taken from Cansim Tables 202-0701 and 202-0704.
- All data are expressed in 2009 dollars.
- The sample runs from 1976-2009, but since I will want to look at 5 quintiles at once, I've decided to look at the averages for 1976-79, 1986-89, 1996-99 and 2006-2009.
- Unattached individuals and what Statistics Canada calls 'economic families' are different in many ways, so I'm going to look at them separately. The aggregate numbers are - as you might expect - somewhere between the two. Unfortunately, that also means I can't say anything about net transfers between these groups.
- There are 20 - count'em, twenty - graphs. I'm pretty sure that's a WCI record.
Here is market income for families:
The increase in incomes in the top quintile is not matched by corresponding increases in the lower quintiles. The same holds for unattached individuals:
Note that market incomes for individuals in the lowest quintile are essentially zero. For the most part, these people are not working. Here are the same data expressed as income shares:
And for unattached individuals:
There is how governments have used transfers:
Levels of transfers have increased for all quintiles - even at the top, albeit barely. And although the lowest quintile still receives the largest transfers, the strongest growth is in the second and third quintiles, where transfers roughly doubled.
Things are very different for unattached individuals:
I haven't looked into this in any detail, but my conjecture is that the transfers going to the bottom quintile of unattached individuals consist largely of social assistance payments, while higher up, they are things like EI and perhaps pensions. I welcome clarifications in the comments.
When you look at the shares of these transfers, it becomes clear that the trend is one in which resources are shifting away from the lowest quintile in favour of quintiles 2 and 3. This is consistent with the story of my earlier post.
For unattached individuals, the shift in transfers is focused even further to the right of the income distribution:
Taxes are applied after transfers. Since the top quintile is the only one that saw significant wage growth, it's also been the only significant producer of new tax revenues:
The same goes for unattached individuals:
And here is how the tax shares are distributed. The rising share of incomes has translated into a rising share of taxes:
And to an even greater extent for unattached individuals:
Next, I'm going to put the two together to get the net redistribution of income for each quintile
The top quintile is seeing larger amounts of money being redistributed further down the income distribution. Again, while the lowest quintile receives the largest net transfers, the biggest gain is in the second quintile.
This pattern repeats itself for unattached individuals:
Here are similar graphs, showing the differences between the income shares of market and net (after transfers and taxes) incomes:
Once again, the middle of the distribution is seeing larger gains than the bottom quintile.
There doesn't seem to be much in the way of a striking change in how the incomes of unattached individuals are being redistributed:
So this is the conclusion in the title: the top quintile is paying more in taxes and receiving less in transfers, but this money is being redistributed to a very great extent to the middle of the distribution instead of to the lowest quintile.
Why would this be the case? I can think of at least three answers:
- Random chance. These net effects are the results of many different changes across time and across provinces: changes made without any sense of how they would affect the distribution of income as a whole.
- Electoral pandering to win the favour of this person.
But there's another possible explanation. If you scroll up and look at the changes in the distribution of market incomes, you'll see that the biggest losses were concentrated in the second and third quintiles. If you were designing a tax-and-transfer system to reduce poverty and inequality, you'd focus on the lowest quintile. But if you're trying to preserve the distribution of net incomes, you'd focus on those who were suffering the greatest losses.
Thanks to those extra transfers, the losses in market income in the second and third quintiles were significantly attenuated:
The same graph for unattached individuals:
When we look at income share, the declines in the shares of the second and third quintiles are much less visible than was the case for market incomes:
And given the changes in the distribution of market income, the distribution of net income of unattached individuals is remarkably stable:
It's not hard to think of a political economy story in which the government would want to use taxes and transfers to keep the distribution of net incomes relatively stable, especially if the biggest losses are occurring near the median. And especially if people at the middle are thinking more about their situations relative to what they once had, and not relative to those further down the distribution.
But there's one last thing to note: after several decades of decline, the second and third quintiles saw real market income gains in the 2000's. It may finally be politically possible to shift policy attention to those further down the income distribution.
Update: Kevin Milligan informs me that this phenomenon has a name: Director's Law.