Brad de Long correctly summarises the argument of my papers with Simon Grant. If you accept that the equity premium (the large and unexplained difference between the rate of return expected by holders of private equity and the rate of interest on low-risk bonds) is explained in large measure by the fact that capital markets do not do a good job in allocating and spreading risk, the the natural solution to all this is the S-World: Socialism: public ownership of the means of production This is because risk can be more effectively through the tax system, and through governments’ capacity to run deficits during economic downturns than through private capital markets. A very robust implication of the observed equity premium is that a dollar of investment returns received during a recession is worth two dollars during a boom – this provides governments with a huge arbitrage opportunity.
But we economists love our ceteris paribus (all other things equal) clauses. At least one commentator noted my qualification that this argument applies “unless there are large differences in operating efficiency between private and public enterprises”. Since, in a wide range of businesses, public enterprises have not performed very well (my own home state of Queensland experimented with state-owned butcher shops in the 1920s) this seems to leave us in the realm of “on the one hand this, on the other hand that”. Fortunately there is a simple empirical test which enables us to balance these considerations, at least in relation to proposed privatisations. If the advantages of privatisation outweigh the difference in the cost of capital, and assets are sold in a competitive market, then the government should come out ahead by selling assets and using the proceeds to repay debt, thereby reducing obligations.
In fact, this is rarely the case. In most cases, the interest savings from selling public assets are less than any reasonable estimate of the earnings foregone. And if you don’t like using estimated earnings you can look at cases where assets were valued for privatisation, but then not sold. Again governments came out ahead from not selling in most cases. It was this empirical observation, rather than theoretical analysis that led me to the conclusion that the equity premium provides a case for public ownership.
On the other hand, the kinds of enterprises where government ownership is common are, in general, those where you would expect the balance of considerations to lean towards public ownership. They are capital intensive, so a lower cost of capital is important and excess labor costs (for example, due to overstaffing) are not. In addition, they are often subject to fairly tight regulation for natural monopoly or essential-service reasons, which reduces the reward to entrepreneurial innovation.
The record of government ownership in other large-scale businesses is mixed (I mean this literally, not as a euphemism for ‘bad’). Brad notes that the US government made a pot of money by rescuing Chrysler in the 1980s, and the British government did the same for Rolls-Royce. But plenty of rescues have turned out badly (from memory, British Leyland didn’t do to well). And in these cases, the cost of acquisition was not great – the case for governments buying profitable enterprises outside the infrastructure sector (broadly defined) is not so strong.
The argument is clear-cut in the case of entrepreneurial businesses that don’t rely on outside equity. For such businesses, the incentive effects of having the residual flow to an owner-manager outweigh any considerations of risk sharing. Hence, as far as the considerations outlined above are concerned, there is no case for public ownership.
So, it turns out that the equity premium provides a case for the mixed economy, rather than for comprehensive socialisation. Given the generally successful performance of mixed economies (most notably between 1945 and 1970), there’s nothing paradoxical or surprising about this.