(Warning: this is a large and somewhat geeky post).
I noted in my post on Labour’s tax policy here that there were arguments on both sides for capital gains taxes. This was a euphimistic way of saying that there are economists who I respect who are in favour of capital gains taxes, so I wouldn’t want to dismiss the idea out of hand, but I have a hard time understanding how they could come to that conclusion. So I am going to lay out the case against in a couple of posts. Most of the points below are standard fare but a couple of them I haven’t seen before.
One thing to note in the discussion of any tax is that all taxes created bad effects, and so it is not sufficient to show that a particular tax has bad effects: those effects have to be compared to the alternative source of revenue. Capital gains taxes are taxes on assets, and so part of the general taxation of capital income. Most studies of tax systems have concluded that the long-term distortionary costs from taxing capital income are greater than those from taxing labour income, so that, from the standpoint of efficiency the optimal tax system would have lower tax rates on capital than on labour income. Equity considerations might lead one to favour eschew this result and favour income and labour income being treated identically. It would be difficult, however, to make a case for higher tax rates on capital income than on labour. Note, however, that in an unindexed tax system, this is what we see. Because it is the nominal interest rate that is taxed, not the real rate that subtracts off the rate of inflation, the effective tax rate on capital income is greater than the headline rate in the presence of inflation. In particular, consider a tax system with a 40% tax rate on all labour and capital income and a real interest rate of 5%. Every percentage point of inflation increases the effective tax on capital income by 8 percentage points, so that if inflation were at 2%, the effective tax rate on capital would be 56%. It is against this backdrop that capital gains taxes need to be considered.
First, let’s start with an incorrect, but not necessarily misleading, example given by Rodney Hide showing how we implicitly already do have a CGT. His example was a variant on the following. Imagine that a no-tax world in which the risk-free rate of interest is 5% and you have an asset expected to pay $100 per year in perpetuity. The asset price would be $2,000. Now let there be a surprise increase in the payment of the asset to an expected on-going $200 per year. This will result in the asset price increasing to $4,000, a capital gain of $2,000. Now consider a world with a 40% tax rate on interest income. In Hide’s example, the asset would pay after-tax returns of $70 and $140 before and after the productivity gain, giving before and after asset prices of $1,400 and $2,800. In this example, the tax on the assets income results in the capital gain being only $1,400, and so has effectively been taxed at the 40% tax rate.
The error in this example, as pointed out here by Bill Kaye-Blake, is that it ignores the effect that taxes have on the risk-free interest rate. If the 40% tax rate applies to interest income as well as the flow of earnings from the asset, then the after-tax risk free rate of interest would be 3%, and the before and after prices of the asset would be $2,000 and $4,000, exactly the same as in the no-tax case.
But this doesn’t mean that Hide’s analysis is wrong. It is true that taxes applied equally to fixed interest and asset earnings would not change the capital value of an asset, but it does change the implied income from the an asset of a given value. Consider someone who owns the above asset and will continue to own it. They are earning $100 per year of which $40 per year is paid in taxes. After the increase in earnings, they pay $80 per year in taxes. What is the fairness justification for also taxing them on the capital gain, which is simply capitalises the future earnings, which are going to be taxed anyway? To draw an analogy, consider a tradesman living in Christchurch who suddenly finds that, due to the post-earthquake rebuild, his future wage income deriving from his human capital will increase. As he earns more he will pay more in income taxes; should he also be taxed on the increase in the value of his human capital?
At this point, the riposte might be that capital gains taxes are typically only applied to realised gains, so that the above asset holder would not be taxed unless he sold his assets just as the tradesman is not taxed on the unrealised value of his human capital. So, now imagine that the asset holder were to sell the assets whose value has appreciated in order to purchase an alternative portfolio of assets. He will still in expectation be earning $200 per year paying $80 per year in taxes, up from the $40 per year in taxes before the gain. There is still no compelling argument for taxing that increased earnings twice. Of course, he might choose to consume the gain rather than reinvest, but that is just the distortion introduced whenever capital income is taxed.
Furthermore, taxing just realised but not unrealised gains introduces other distortions: it creates an incentive for investors to hold on to assets rather than adjusting their portfolio mix to suit changing circumstances; if capital gains are taxed but capital losses are not subject to an offsetting negative tax, it implies a higher overall tax on savings, but if capital losses are treated symmetrically, canny investors can structure their investments to ensure a mix of capital gains and losses, and then realise their capital losses while retaining assets that have made a gain.
For these reasons, economists often express a preference for a capital gains taxes that tax both realised and unrealised capital gains. This is one area, however, where the politicians have it right and the economists wrong, in my view. Taxing both realised and unrealised gains is the more efficient policy, but such a tax has nasty implications for basic property rights when it creates situations where asset owners are forced to divest themselves of assets that might have high personal utility value (family holiday homes, businesses that one started, etc.) simply to meet a tax liability. Furthermore, there are the implementation and compliance costs from measuring unrealised capital gains.
So what are the arguments in favour of a capital gains tax? Let’s consider it against the criteria of efficiency (including enforcement and compliance costs), equity (vertical and horizontal), and respect for rights. Working backwards, I noted above why, respect for property rights leads real-world tax systems to only tax realised capital gains. But this creates its own problems with definitions, and potential for tax avoidance. An accountant friend of mine recently described a court case taken by the IRD against a group of accountants, in which the issue to be determined by the court was not what actions the group had taken, but whether their actions did, in fact, constitute a violation of the law. Legal processes where one can be hanged on a comma, are not ideal for a society based on the rule of law.
On equity grounds, since the income stream from assets will be taxed when they occur, at the same marginal rate as labour income, there is no obvious vertical equity imperative for capital gains taxes. It is true that the nature of having traded assets sold in competitive markets is that the benefits of future expected gains and the costs of future expected downturns in value are always captured or borne by the current owners of the assets, so that a comprehensive system of taxing capital gains with negative taxes on capital losses might serve a social insurance role, but there is no clear equity benefit.
Finally, on efficiency grounds, the complications that arise from trying to solve other problems give rise to a tax system with higher compliance and enforcement costs per dollar raised than other taxes, and also introduce other efficiency-reducing distortions.
Against all these arguments is the only serious argument I have seen in favour of capital gains taxes: that they are needed on efficiency grounds to make sure that savings is directed into productive investment rather than into chasing capital gains. This is a complicated issue that merits a separate post. More on that tomorrow.
In the meantime, let me propose a simple alternative to a normal capital gains tax that maximises benefits with none of the above costs. Take a tax system with a 40% tax on labour income and nominal capital income. Introduce a 13% tax that will apply to realised capital gains but also give a negative 13% tax on realised capital losses, and will only apply if the gains (losses) result in an increase (reduction) in consumption rather than being reinvested. (This removes the incentive to cash the losses and run the gains.) Finally, reduce the overall tax rate on nominal capital income so that the effective tax on capital income with an interest of 5% and an inflation rate of 2% is only 43% not the 56% it would otherwise be. And do all of this with a simple tax system with very low transactions and compliance costs. It seems too good to be true, but it can be achieved simply by reducing the tax rate on all income from 40% to 31% and replacing it with a 15%, no-exemption GST! Put another way, we already have a very good capital gains tax in New Zealand without any of the downside costs.