Tuesday, 20 April 2010

Is the IMF taxing the wrong things?

I have no objection in principle to higher taxes on banks. But the IMF's surprising proposal (outlined by Robert Peston here) may not quite be taxing the right thing.

If the goal of the tax is to pay for the externality imposed on taxpayers by bank behaviour, that's fair enough. In line with the principles of Pigovian taxation, the tax should fall on the activities which impose those costs. That way, it's fair and also creates the right incentives, to shrink those activities if they do not benefit society enough to justify them.

The first part of the IMF's tax seems to fit this principle. Peston doesn't explain what this "flat rate" would be levied on - I assume it isn't literally a flat rate, otherwise Goldman Sachs or Citigroup would pay the same amount as the Cheltenham Building Society or the Third State Bank of Des Moines. Presumably therefore it will be in line with the recent Obama proposal: a percentage levy either on total assets or on risky, non-deposit (wholesale) liabilities.

The second part, however, is a tax on profits and compensation combined. This implies that, say, a national retail bank in India, which might make $500 million profit and spend $5 billion employing a million Indian clerks, would pay the same tax as an American investment bank making $2 billion profit and paying $3.5 billion in bonuses to a thousand wealthy residents of Manhattan and Staten Island. I am not sure that this would really make a lot of sense.

It's very hard to see how to make a general tax on either profits or salaries effective and fair as a financial regulation mechanism. Last year's one-off UK tax on bonuses is kind of workable, precisely because it is a one-off and therefore gives banks little reason to reorganise or change their practices to get around it. And straightforward taxation of profits and salaries is a fair way to raise general government revenues - but surely not to pay for an IMF rescue fund. Is it really fair that the Indian retail bank, who probably presents almost no risk to the international financial system and makes an important contribution to its local economy, pays the same as the hypothetical American bank, which is likely to be a highly leveraged, high-risk entity?

Now it is not guaranteed that the Indian bank is low-risk. Indeed, some apparently safe-looking state-backed German institutions caused huge problems in the last two years by overinvesting in property-backed financial instruments. But this was not a function of their profits or salaries - I suspect they weren't very profitable at all, compared to the American banks who sold them the assets. Instead, the problem was too much correlation of the financial system's assets - and that is where a tax should fall if we want to discourage risky behaviour.

It's time to write that post I've been putting off, on how to put a price on correlation. I'll update this one when it's published.

Update: I note that the IMF also suggests a limit on the tax deductibility of interest payments. This sounds attractive at first glance - but note that it is not actually an issue about financial distortion, because interest income is already subject to tax which balances out the deductions for interest costs. Instead, it's a way of preventing companies from shipping their profits offshore. At present it's easy to load up a profitable company in (say) Britain with debt, so that all its profits go to service a very expensive loan in the British Virgin Islands. The profits on the BVI loan are not taxed, and voila! No corporate taxation is due. This proposal may therefore be a way to wrap up a bit of tax-haven combat with a popular regulation that's hard to lobby against.

1 comment:

Adam said...

Even within a single jurisdiction, tax relief on debt encourages additional leverage since the marginal cost of capital will be much lower if debt is being raised rather than equity.