The answer might be: less than you think. A new paper by Allen Drazen and Ari Brender has looked at changes in the composition of government spending across 12 different categories in 68 democracies between 1972 and 2003. And they’ve found something remarkable:
So, why might governments have so little impact on spending? Here are some possibilities:
1. Any party has to appeal to the median voter. If the median voter’s preferences don’t change, nor will the composition of spending.
2. Spending is very hard to change. Governments have to fend off special interests and vested bureaucracies, and the brute power of the status quo.
3. Spending is more sensitive to events than leaders’ ideology. Many Labourites would like to have cut military spending, for example, but this preference has been mitigated by the number of wars the government has felt compelled to fight.
4. “Nixon goes to China” syndrome. It’s widely thought that only a man with Nixon’s strong anti-communist credentials could have improved the US’s relations with China so much in the 1970s; had any more left-wing president attempted to do so, he would have been decried as “soft on Reds.” In this sense, ideology can be a counter-predictor of policy.
This might generalize. Maybe Labour is better equipped than Tories to cut welfare spending, as they are (wrongly?) less vulnerable to charges of being hard on the poor. Similarly, Tories might be better able to cut military spending, as they are (again, maybe wrongly) less vulnerable to charges of lacking patriotism.
5. Maybe macroeconomic aggregates are just to big to detect policy changes. Here’s a challenge for you: show me some time-series data in which we might be able to clearly identify when Thatcher became Prime Minister, and when she left.
If all this is right - or even nearly so - the search for dividing lines between the political parties is, for practical purposes, fruitless. What matters is not politicians’ preferences, but their opportunities. And these are often tightly constrained.