Taxes and asset prices: A cross-country perspective

On November 23, 2005 the Canadian government announced a reduction in the tax on dividends in an effort to neutralize the tax system’s bias in favour of income trusts. Eleven months later, on October 31, 2006, a new government changed direction and eliminated the tax deductibility of income trust distributions altogether. Exempted from this change in policy was the real estate investment trust (REIT) sector.

A 2012 study found that the November, 2005 tax policy change, which followed a period of high return volatility and negative cumulative abnormal returns, had a positive and significant effect on the valuation of REITs. The October, 2006 tax change had different implications for REITs than it did for non-REIT trusts.  Nonetheless, like the rest of the income trust sector, REITs responded negatively on the day of the announcement and the day following. However, within a short period of time the REIT portfolio had regained all of the value it had lost up to and including the event window.

The reason the Minister of Finance gave for exempting REITs was that the government wanted to stay “in sync” with US tax rules, which had exempted REITs from a clampdown on income trusts in the 1980’s. The conventional wisdom at the time was that Canadian real estate would be “gobbled up” by US REITs if Canadian REITS lost their tax exemption. Inasmuch as there is reportedlya significant international component to real estate company returns, this raises the question of whether Canadian REITs got caught up in the overall market reaction to the two tax changes (positive in 2005 and negative in 2006) or whether investors were reacting to other conditions in the North American real estate market.

The current study examines the return behaviour of both Canadian and U.S. REIT portfolios around the time of the tax change announcements. Ordinary least squares with dummy variables are used to estimate market model portfolio returns on the event date and the day after. Both equally-weighted and value-weighted portfolios were created in order to check the robustness of the results. Because the U.S. sample is much larger, randomized subset portfolios were used to compare with the Canadian portfolio. In addition, the relationships between Canadian REIT returns and U.S. REIT returns are examined for each event. Preliminary results indicate statistically significant abnormal returns for both Canadian and U.S. REITs and statistically significant relationships between both countries' REIT returns on both dates.