I found this to be the most interesting question explored in ‘Government Size and Implications for Economic Growth’, by Andreas Bergh and Magnus Henrekson. Before I explain, however, I want to provide some comments on the author’s conclusions about the effects of size of government on economic growth.
The reservation I have about the review of the literature by Bergh and Henrekson is somewhat technical – so some readers may prefer to skip this paragraph. My reservation concerns the authors’ enthusiasm for Bayesian Averaging of Classical Estimates (BACE), a technique used to deal with possible sensitivity of parameter estimates to the inclusion of different control variables in regression models. A recent paper by Antonio Ciccone and Marek Jarocinski suggests that margins of error in international income estimates are too large for such agnostic growth empirics to be reliable. In any case, in my view the economic reasoning that tells us that the economic costs of taxation rise approximately in proportion to the square of the tax rate provides a more powerful case against big government than the results of cross-country econometric studies. (The authors appear to attribute this insight to the Swedish economist, Jonas Agell (p.17), although it should more appropriately be attributed to much earlier work by Arnold Harberger, or possibly even to Alfred Marshall.)
It is well known that the economic cost of high tax rates arises in part from the substitution of leisure for income. Some would argue that this is beneficial because many people obtain more happiness from spending time with family and friends than from working. One reason why the argument is spurious is because it may be rational for individuals to sacrifice some current happiness to provide their children with a better education, fund early retirement or pursue any number of other objectives that are important to them.
Another reason why the argument is spurious is that what economists talk about as a choice between income and leisure is often actually a choice between time spent on paid work and time spent on unpaid household chores. It is doubtful whether people obtain much more pleasure from housework, weeding the garden and childcare than from working for pay. Bergh and Henrekson make the good point that high rates of labour taxation provide an incentive for consumers to produce such services themselves in the home rather than to work longer hours in order to purchase them in the market place. The authors suggest that this explains why hours of unpaid work are substantially greater in Sweden than in the US and hours of paid work are correspondingly lower in Sweden than the US.
The authors also provide a graph comparing average hours worked per person in Sweden and the US over the period 1956 to 2003. It shows that average hours worked in Sweden were substantially lower than in the US during the 1950s, when Sweden’s tax rates were much lower, the situation has been reversed in recent decades.
Over the last couple of centuries the ancestors of the vast majority of people in high-income countries have managed to obtain the benefits of participation in a market economy – the benefits of exchange and specialization on the basis of comparative advantage, resulting in much higher living standards and providing greater opportunities for skill development and incentives for further innovation. High taxes associated with big government provide the opposite incentives – encouraging people to shun the market and to produce services for themselves. Self-sufficiency is not without its attractions, but I doubt whether many people would freely choose the poverty experienced by their ancestors, even if that was the only way they could ensure a supply of fresh, organically-grown vegetables.