Abstract
This paper examines how changes in capital gains tax rates affect individual and household realization behaviour, using Canada as a natural setting. Canada taxes unrealized gains at death, partly eliminating the deferral incentive that exists in systems with forgiveness of tax at death. Using theoretical and empirical evidence, this paper shows that the lock-in effect arising from realization based taxes is low in Canada. I implement a difference-in-differences design, using deceased individuals, who realize their entire stock of accrued gains, as a comparison group for living individuals, who choose how much to realize. If the living respond to lower tax rates by increasing the fraction realized out of their stock of accrued gains, differences between the two groups' realization changes capture that effect. This novel identification strategy addresses the challenge that tax reforms are often preceded by changes in capital markets that affect the stock of accrued gains -- an omitted variable, proxied in this paper by realizations of the dead, who are matched to living realizers on observable characteristics. I study the year 2000 reduction in Canada’s capital gains tax using a 20% random sample of administrative tax records. The results show a dynamic realization response: a significant short-term increase in realizations, but no lasting effect in the medium to long run. Overall, the findings suggest that, in the absence of preferential treatment of gains at death, changes in capital gains tax rates only lead to retiming effects and cause no permanent change in realization behaviour.
with co-authors: Kevin Milligan and Michael Smart
Abstract
Social weights are sufficient statistics for distributionally sensitive policy evaluation. In this paper, we ask if Canada’s social planner would prefer the US pre-tax income distribution. We use the US distribution of fiscal income data from the Current Population Survey Annual Social and Economic (ASEC) Supplement and Distributional National Accounts (DINA) from Piketty, Saez and Zucman (2018) . Data for the Canadian distribution of fiscal income comes from the Longitudinal Administrative Databank and the Canadian Income Survey. For the two countries, we calculate equally distributed equivalent income to find whether given Canadian social weights, US incomes are preferred. We find that Canadian and US income distributions are not dominance ranked and there is considerable ambiguity about the bottom of the US distribution that arises due to non-filers and tax treatment of cohabiting couples as single individuals. Several measurement issues affect median incomes and overall income distributions in a non-trivial way. These include the economic unit of analysis i.e. tax unit versus family unit, treatment of non-filers, and evolution of family size across the income distribution. High income Canadian families tend to be larger in size than high income US families, whereas the opposite is true for low income families. Canadian tax units are relatively better off than US tax units for all but the top quartile of income distribution. Much of the Canadian advantage is attenuated when US incomes are augmented to include employer contributions to health insurance. When using census families as the economic unit, the US distribution almost completely dominates the Canadian distribution.
Abstract
This paper shows how responses to capital gains taxes vary with age. I examine heterogeneous responses by age due to a decrease in capital gains tax that occurred in Canada in the year 2000. Because taxes cannot be deferred at death in Canada, the lock-in effect of realization-based taxes differs by age. Using a 20% random sample of individual tax-filing records, I find lower response of older individuals to tax change. Taxation of unrealized gains at death in Canada dampens incentive to defer realizations and I find that this deferral advantage is decreasing in age.
Co-authored with Michael Smart and published in Canadian Tax Journal, Volume 69-4, 2021
Abstract
This article makes the case for increasing tax rates on capital gains income under Canada's Income Tax Act. The current tax preference for capital gains costs $35 billion annually in forgone government revenues, with much of the benefit accruing to high-income families. To address the inequity of the present system, and to reduce tax non-neutralities, the 50 percent inclusion rate for capital gains should be raised to 80 percent. This would constitute a simpler, more efficient way of taxing high-wealth individuals than recent proposals for a novel tax on wealth, and would likely generate far more revenue as well.
Posted on Finances of the Nation website on April 4, 2023 by Sobia Jafry and Michael Smart
We use data on the distribution of capital gains income across and within income groups to simulate the impact of reforms to increase capital gains tax rates. Readers can use our online tool to explore different possible reforms. A reform that targeted only the highest capital gains income amounts would affect very few families, while raising substantial revenue and increasing fairness in the tax system.
Posted on Finances of the Nation website on January 26, 2022 by Michael Smart and Sobia Hasan Jafry
The current capital gains tax preferences cost $35 billion annually – with high-income families accruing most of the benefit. The recent passage of Bill C-208 exacerbates these issues. To fix these problems, the inclusion rate for capital gains should rise to 80 per cent from the current 50 per cent.