Wednesday, September 22, 2010

Does Wagner's law make sense?


Wagner’s law refers to the proposition of Adolph Wagner (1893) that there is a positive relationship between the level of economic development and the size of government. The underlying idea seems to have been that the demand for services provided by government tends to rise strongly as average incomes rise.


I think Wagner’s law still has a huge influence on the thinking of many economists. This influence is evident in the tendency of many economists to view big government as the norm for high-income countries. For example, it explains why economists pose questions like: Why doesn’t the US have a European-style welfare system? This is an odd question because there is considerable variation in the size of welfare states even within Europe and Swedish-style welfare systems are certainly not common among high-income countries outside of Europe.

The influence of Wagner’s law on the modern thinking of economists seems to rest on it being an empirical regularity or stylized fact. If you overlook the wide variation in size of government in high income countries, Wagner’s law does appear to fit some of the facts. Looking back at the recent history of individual OECD countries, most of them clearly had smaller governments 50 years ago when their average incomes were much lower. Yet, a recent study for the UK and Sweden from the beginning of industrialization until the present (a period of 177 years for the UK) found that Wagner’s law does not hold in the long run. The data are inconsistent with Wagner’s law in the initial industrialization phase (prior to 1860) and since the 1970s (Dick Durevall and Magnus Henrekson, ‘The futile quest for a grand explanation of long-run government expenditure’, INF Working Paper 818, 2010).

The Durevall and Henrekson paper also rejects a rival theory – the ratchet theory – that government spending ratchets up in times of crisis (wars, social upheavals, recessions) and then tends to remain at the new higher level. The expansion of government spending in the 25-35 years following WW2 cannot be explained in terms of a ratchet effect.

Some people might try to rescue Wagner’s law by arguing that it always applies at some stage during the process of industrialization. Thus it might be argued, for example, that Wagner’s law will result eventually in the development of big governments in jurisdictions such as Hong Kong and Singapore that have been able to restrain growth in government, even though they now have relatively high average incomes. However, there do not seem to be any reasons why governments of high income countries would necessarily find it harder than governments of medium to low income countries to resist political pressures to become more heavily involved in activities such as funding of retirement incomes and provision of education and health services. Nor would they necessarily find it harder to resist arguments for the social welfare safety net funded by taxpayers to rise more than proportionately as incomes rise.

Growing Public: Volume 1, The Story: Social Spending and Economic Growth since the Eighteenth CenturyIf we were desperate to rescue Wagner’s law perhaps we could argue that bigger government is an inevitable response to political pressures associated with the demographic transition - declining birth rates and aging population age structures – associated with economic growth. On this basis Peter Lindert argues that we should expect an expansion of the welfare states in East-Asian countries ‘as they age and prosper’. In OECD countries, including Japan, political systems responded to an increase in the proportion of old people in the populations by providing pensions for aged persons. The further aging of populations has led to increased government spending on pensions - a major factor associated with the growth of government spending in high income countries. Lindert asks: ‘Do we really know that China, Singapore and other East Asians will be more resistant to rising transfer budgets than Japan has been, when they approach Japan’s income level and age structure?’ (‘Growing Public’, Vol 1: 221).

My answer to Peter Lindert’s question is that I don’t know how East Asian governments will respond to an increase in grey power. Perhaps they will see what lessons they can learn from the experience of the big government welfare states of Europe and decide that there is a better way to fund retirement incomes. They might even decide that the compulsory savings approach that has applied in Singapore since 1955 is preferable to the absurdity of providing pensions that are not subject to means tests. Why should people of working age be taxed more heavily in order to add unnecessarily to the incomes of wealthy retirees?

6 comments:

Lorraine said...

Ah, Singapore, representative of the worst of both worlds; inequality and authoritarianism in one package--the exact opposite of 'freedom and flourishing' assuming the words actually mean something. And communist-in-name-only China, which has become Singapore writ large. I'm sure they'll find a way to put harness their elders to some kind of work-house or another, given that cheap labor is the nation's business model.

Winton Bates said...

Thanks for your comment, Lorraine. Singapore is far from ideal in many respects. But I don't think an increase in government spending as a percentage of GDP would necessarily make it a better place to live.
The point I was trying to make in the last paragraph is that it is better to have an effective means tested safety net than a big government.

Lorraine said...

Perhaps not. Some say means tests are mean. Not that I have anything to say in defense of that bogeyman known as Big Government, but in my country whenever the so-called small government party gets in power, it's always the means-tested programs that are first on the chopping block. Non-means-tested programs have a middle-class constituency and therefore get political representation. That, as I see it, is most of the downside to what David Brin called the 'diamond shape' society. That is also why I am inclined to believe that Wagner's law is valid, at least in 'diamond shaped' societies. The snarky right-wing explanation would be the old trope about democracy leading to people voting themselves largesse from the public treasury blah blah blah. A more plausible explanation, I think, is that people realize that the only alternative to safety nets is gated communities and trustworthy bodyguards. Call it a stage in the development of civilization.

Winton Bates said...

Lorraine, I think the diamond shape society does help to explain middle class welfare to some extent. Middle class voters are more inclined to support programs if they are eligible for benefits. That doesn't necessarily rule out means tests, but it may make them very liberal - as in the case of Australia's aged pensions means tests.
The effect of numbers on the power of interest groups is interesting. Increasing numbers tends to make groups more powerful up to a point. However, as groups become larger they have to consider the effect their rent seeking is going to have on the incentives faced by the goose that lays the golden eggs. So, as the proportion of elderly people in the popultaion increases, the poor could end up better off under mean means tests than under universal benefits. I might try to explore that idea further in a future post.

I think history and ideology are also important. We should expect more collectivist outcomes in Europe because Marxism was stronger over there. The East Asians will probably continue to rely heavily on families; Americans will probably continue to lean towards individualism; and Australians will want everyone to get 'a fair go' - whatever that means.

Lorraine said...

If not Wagner's Law, what about the 'one dollar one vote' theory? Does that seem to make sense?

Winton Bates said...

No Lorraine, one dollar one vote theory doesn't make sense to me. Perhaps there is some other mechanism involved. For example, one possibility might be that higher minimum wages result in lower inequality at the lower end of the income distribution, but more unemployment. More unemployment is associated with more redistribution to support the unemployed.