Why does the size of the government increase with economic development: A theory based on higher household leisure and its prediction of bigger government in the AI era

It is a widely observed fact that after industrialization, the size of the government has increased significantly. This is often referred to as Wagner’s Law. Since the early twentieth century, it grew from around 10% to the current global average of 35%. Existing literature mostly explains this phenomenon from the government side, highlighting factors such as enhanced state capacity and events such as wars. In this paper, we argue that the key driver is the increase of leisure time of households, who are the ultimate payers of government expenditure. We explain that the growth in leisure has raised households’ welfare and becomes a buffer against higher taxation, enabling them to accept significantly higher taxation on their labor income in exchange for better public services. We begin with a simple observation: across seventeen OECD countries, the ratio of government taxes to “augmented national income (ANI)” (GDP plus the opportunity cost of household leisure) did not increase as much as that of tax to GDP and remains relatively modest, at around 7% to 13%. We then construct a simple model to illustrate the mechanism that higher labor productivity leads to higher leisure and therefore higher taxation. One prediction of our theory is that in the AI era, since labor productivity will rise substantially, and household leisure time will further expand, the size of the government relative to GDP is likely to increase by around 2 to 4 percentage points by 2035. In the longer run, if AI sustains at least the pace of productivity growth seen over the post-WWII period, government size could expand by at least a further 10 percentage points by the end of the century.