The Australian Treasury has been busy. On top of its usual output, the last 18 months have included the Financial System Inquiry, hosting the G20, the 2015 Intergenerational Report and the tax white paper. All this while eliminating one-third of its workforce!

But today I'd like to focus on the tax white paper, and an underlying assumption Treasury uses to justify some of its agenda.

When modelling and discussing the effects of taxes, Treasury frequently makes the assumption of perfect international capital mobility. This assumption means there is only one worldwide after-tax (risk-adjusted) rate of return on capital. If there were anywhere that offered a better deal, perfect capital mobility would imply that capital would flow into that area, until the return differential was arbitraged away. Similarly, if anywhere offered a worse deal, capital would flow out until, again, returns were equalised. 

Often, this assumption is just for analytic convenience. It is easier to solve models with this assumption, and it generally is of no great consequence.

However, there are times where this assumption is critical.

For example, it is crucial in justifying Treasury's discussion of abolishing dividend imputation. You see, under the assumption of perfect capital mobility, capital flows into Australia determine the amount of investment in the country. Foreigners do not benefit from dividend imputation, but they do have to pay the company tax rate, which considerably reduces the amount they invest. Thus we get the proposition that Australia could increase investment by eliminating dividend imputation and using the proceeds to lower the company tax rate.

The proposition of perfect capital mobility, or at least very mobile capital, has also been important in other work the Treasury has done on company tax. For example, the assumption drove the result that two-thirds of the benefit of a company tax cut went to wage earners. Also, in work released last week, perfect capital mobility was one reason company tax was assessed to be one of the most inefficient taxes we levy.

But how realistic is perfect capital mobility?

In 1980 Marty Feldstein and Charles Horioka published a famous and influential paper that claimed capital was quite immobile. They based this claim on a very tight correlation between a country's saving and its investment. That may sound esoteric, but it is not. As I said before, if capital is freely mobile across country borders, it will seek the highest yielding opportunities. There is no reason for it to be invested in places with the highest savings. The most plausible explanation for the correlation was that savings tended to be invested in the home country, or in other words, there was a high degree of home country bias in investment.

Now, their paper was written 35 years ago, so things may have changed. To consider this possibility, I have plotted Australia's saving and investment as a percent of GDP below. The tight correlation is still there. If we split the sample up between pre- and post-1980, the correlation stays almost exactly the same (correlation of 0.783 pre-1980, 0.786 post-1980).

And by the metric that Feldstein and Horioka used to assess capital mobility (the coefficient of a regression on investment on saving) the influence of domestic saving on investment has actually increased slightly (from 0.75 to 0.79).

So I am wary of interpreting Treasury's modelling that seems to crucially hinge on a high degree of capital mobility.

Photo courtesy of Flickr user Stefan Jürgensen.