Thanks to the half dozen people who sent me copies of Cowen’s “Does Technology Drive the Growth of Government?”  The paper’s even better than I remember.  Highlights:

The puzzle, courtesy of the great Tullock:

I start with what Gordon Tullock (1994) has called the paradox of government growth. Before the late nineteenth century, government was a very small percentage of gross domestic product in most Western countries, typically no more than five percent. In most cases this state of affairs had persisted for well over a century, often for many centuries. The twentieth century, however, saw the growth of governments, across the Western world, to forty or fifty percent of gross domestic product… I’d like to address the key  question of why limited government and free markets have so fallen out of favor.

Inadequacies of public choice theories of government growth:

Public choice analysis has generated many theories of why government grows and why that growth is inevitable. Special interest groups, voter ignorance, and the pressures of war all are cited in this context. Those theories, however, at best explain the twentieth century, rather than the historical pattern more generally. Until the late nineteenth century, governments were not growing very rapidly. The standard public choice accounts do not contain enough institutional differentiation to account for no government growth in one period and rapid government growth in another period.

Inadequacies of ideological theories of government growth:

According to this claim, the philosophy of classical liberalism declined in the mid- to late nineteenth century. This may be attributed to the rise of socialist doctrine, internal contradictions in the classical liberal position, the rise of democracy, or perhaps the rise of a professional intellectual class. While the ideology hypothesis has merit, it is unlikely to provide a final answer to the Tullock paradox. To some extent ideology stems from broader social conditions. Ideologies changed, in part, because intellectuals perceived a benefit to promoting ideas of larger government, rather than promoting classical liberalism.

I’d add, “Is it really true that in 1890, the typical voter staunchly supported free trade and free migration and staunchly opposed redistributive regulation and taxation?”

Inadequacies of ratchet theories of government growth:

The ratchet effect becomes much stronger in the twentieth century than before. Furthermore most forms of governmental growth probably would have occurred in the absence of war. The example of Sweden is instructive. Sweden avoided both World Wars, and had a relatively mild depression in the 1930s, but has one of the largest governments, relative to the size of its economy, in the developed world. The war hypothesis also does not explain all of the chronology of observed growth. Many Western countries were well on a path towards larger government before the First World War. And the 1970s were a significant period for government growth in many nations, despite the prosperity and relative calm of the 1960s.

Inadequacies of franchise extension theories of government growth:

The hypothesis of franchise extension, however, again leaves much unexplained. First, non-democratic regimes, such as Franco’s Spain, illustrate similar patterns of government growth as do the democracies. Second, much of the Western world was fully democratized by the 1920s. Most governmental growth comes well after that date, and some of it, such as Bismarck’s Germany, comes well before that time. Third, and most fundamentally, white male property owners today do not favor extremely small government, though they do tend to be more economically conservative than female voters.

The last sentence is so obvious yet so ignored.

Tyler then promisingly begins by acknowledging the ultimate power of public opinion:

No matter how incomplete it may be, there clearly must be something to the voter hypothesis. That is, there must be some demand for big government. If all or most voters, circa 2009, wanted their government to be five percent of gross domestic product, some candidate would run on that platform and win. Change might prove difficult to accomplish, but we would at least observe politicians staking out that position as a rhetorical high  ground. In today’s world we do not observe this. Voter preferences for intervention are therefore a necessary condition for sustained large government. Democratic government cannot grow large, and stay large, against the express wishes of a substantial majority of the population.

Next he turns to the many technological changes that pumped up demand for big government.  You can quarrel with the specifics, but his summation is excellent:

[N]o one of these technological advances serves as the cause of governmental growth. Taken as a group, however, these factors made very large government possible for the first time.

To see this, perform a very simple thought experiment. Assume that we had no cars, no trucks, no planes, no telephones, no TV or radio, and no rail network. Of course we would all be much poorer. But how large could government be? Government might take on more characteristics of a petty tyrant, but we would not expect to find the modern administrative state, commanding forty to fifty percent of gross domestic product in the developed nations, and reaching into the lives of every individual daily.

He adds:

Think also about the timing of these innovations. The lag between technology and governmental growth is not a very long one. The technologies discussed above all had slightly different rates of arrival and dissemination, but came clustered in the same general period. With the exception of the railroads and the telegraph (both coming into widespread use in the mid-nineteenth century), none predated the late nineteenth century, exactly the time when governmental growth gets underway in most parts of the West. The widespread dissemination of these technologies often comes in the 1920s and 1930s, exactly when  many Western governments grew most rapidly, leading sometimes to totalitarian extremes. The relatively short time lag suggests that strong pressures for government growth already were in place. Once significant governmental growth became technologically possible, that growth came quickly.

At first glance, Tyler has an extremely functionalist model of government growth: People always had a latent demand for big government; then technology finally made it possible to satisfy them.  But Tyler’s model hardly implies that the public’s latent demand for big government was wise or prudent.  If improved transportation leads to a large increase in consumption of cyanide, it hardly implies that cyanide is part of every nutritious breakfast.

I value papers that make noticeable progress on important questions.  If you share my preference, this could easily be Tyler’s best paper ever.  At least so far; I still hope to see many more like it.  Read the whole thing.