Tuesday, April 24, 2007

New Eurostat Report About EU Government Spending Burden

Eurostat reports that Euro-zone budget deficits fell from 2.5% of GDP to 1.6%. Unfortunately though, most of it reflected higher government revenues. The burden of government spending fell from just 47.5% to 47.4%.

Combining it with the 2002 numbers in last year's release, you can see that compared to 2002, Euro-zone spending fell by 0.3%: points between 2002 and 2006, from 47.7% to 47.4%. But the change in different countries has been very different. Two countries, Italy and Portugal have fairly dramatically increased spending, by 2.7%: points and 1.8%: points respectively. Not coincidentally, they have had the weakest economies of the Euro-zone. France, Holland and Ireland have also increased the burden of spending. In the case of Ireland though that is less of a problem since it started (and ended) with the lowest burden of spending while France started with the highest burden.

The biggest cuts in government spending were performed in Greece and Germany, with 2.9 and 2.7% points respectively. Both started with government spending somewhat above the Euro-zone average and ended with government spending somewhat below. Not coincidentally, Greece has seen very fast growth while Germany has gone from being the sick man of Europe to have growth above the Euro-zone average. Other Euro-zone countries with significant spending cuts (above 1%: point of GDP) include Slovenia, Austria and Luxembourg, all of whom have had above average growth. So you can see a pattern here. Countries who have significantly reduced government spending relative to GDP have all out-performed the Euro-zone average, while the countries who have increased it have under-performed, with the sole anomaly of Ireland. But Ireland, as previously noted, still has the by far lowest burden of government spending and it has also been helped by extremely favorable demographics. And furthermore, some of its growth has been driven by a housing bubble and related cyclical excesses.

The same pattern is evident if we look at non-euro zone countries. The biggest spending reduction has occurred in Slovakia, where government spending fell a full 6 5: points, from 42.3% of GDP in 2002 to 36.3% in 2006. Other big spending reducers are the Czech Republic (-4.2%: points), Denmark (-3.8%: points), Estonia (-3.3%: points) and Sweden (-2.4%: points). Not coincidentally, all of these countries have had above average growth.

Sweden and Denmark still have relatively high levels of spending, but particularly Sweden's is exaggerated by some statistical distortions. First of all, Sweden and Denmark tax transfer payments much more (page 192 in the pdf-file) than do other European countries-or the United States. In Sweden and Denmark, taxes on transfer payments are more than 4% of GDP, compared to 1-2% in most continental European countries, 0.5% in the U.S. and 0.3% in Britain. Taxes on transfer payments are really nothing but an accounting illusion since the money never leaves or enters the government sector, and so means that both revenues and spending are exaggerated. Moreover, as Eurostat notes, Swedish interest payments are reported on an unconsolidated basis, which means that for example interest payments from the State to State-owned pension funds are included, which it is not in other countries. Taking these two facts into account, both Sweden and Denmark have now been surpassed by France in terms of government spending. And Denmark has also been surpassed by Italy.

The biggest spending increases have occurred in Cyprus (+3.3%: points) and Britain (+3.2%: points). Latvia, Malta and Hungary have also increased government spending relative to GDP. Cyprus, Malta and Hungary have all fallen behind other new member countries, and Britain have lost its previous status as growth leader of Western Europe and now trails the Euro-zone average in growth. Latvia is the one anomaly among non-euro zone countries from the strong negative correlation between changes in the burden of government and economic growth. This is due in part to similar reasons as Ireland's strong performance, namely that Latvia started and ended with one of the lowest burdens of government spending. Latvia also have arguably the strongest cyclical element in its economic growth, with money supply growth running at nearly 40%, which have resulted in a consumer price inflation rate of 8.5% and a current account deficit of more than 15% of GDP.

So, looking at European economies you can see a very strong correlation between economic growth and changes in government spending. The correlation is not perfect, of course, just like all other correlations between causal factors and aggregate social or economic phenomena’s as there are always other factors involved. But given this fact, the correlation is actually unusually strong, with just two real anomalies among 25 countries, both of whom are easily explained. Government spending will hurt the economy regardless of whether it is financed through taxes that reduces incentives or through government borrowing, which has a "crowding out" effect on growth.

2 Comments:

Anonymous Anonymous said...

Although I 99% agree with what you write, there is also a bit of endogeneity in the statistics. The correlation between growth and reductions in government spending (measured as % of GDP) are partly caused by the fact that the economies "grow their way out of high government spending". The reason is a certain lag in fiscal policy - government budgets are set in advance to a large extent and if there is surprisingly high growth, they do not "adjust" the spending upward to keep the same spending/GDP ratio. Similarly when there is a sudden drop. It should balance out over the time, but you are looking at very short time spans.

I have never measured the importance of this factor, but it's certainly nonzero. For instance, when I look at the Czech Republic (my home country), there haven't been any intentional reductions in government spending, it was actually quite the opposite. Nevertheless, we see a strong decrease in the govt. spending/GDP ratio. The country simply grew their way out. But then the causality is reversed.

It would be interesting to disentangle these two effects. My bet is 75% govt.spending->growth, 25% growth->govt. spending. Maybe 80/20.

In the long run (say 10-15 years, at least 2 business cycles) >95% govt.spending->growth, <5% growth->govt. spending.

9:27 AM  
Anonymous Anonymous said...

Actually, this study by Christina and David Romer on the isolation of the impact of "exogenous" tax changes in the U.S. economy may be of interest to you. Here is a nice summary. An "exogenous" increase in tax revenue/GDP ratio by 1% leads to a decrease in GDP by 3% in less than 3 years.

7:49 AM  

Post a Comment

<< Home