Antonio Fatás makes a significant point about countercyclical fiscal policy. The set of economists and policy makers who think favourably of automatic stabilisers (the fact that taxes go down and government benefits go up in a recession, and vice versa) is very large. The set that think discretionary countercyclical fiscal policy is desirable is much smaller. Yet they involve the same mechanisms. It would be quite illogical to claim that discretionary fiscal policy will have no impact on output, and at the same time argue that automatic stabilisers do help stabilise output.
However, there are two good reasons I can think of why you might be in favour of automatic stabilisers but be against discretionary countercyclical fiscal policy. The first is rather dull. Automatic stabilisers kick in fairly quickly: if you become unemployed, you stop paying income tax and start receiving unemployment benefit almost immediately. Discretionary fiscal policy, on the other hand, is subject to important implementation lags. These may be political – a bill has to be passed by politicians – or institutional – projects need to be found to spend the money on. If we really wanted to conduct discretionary fiscal policy on a regular basis then these lags could be considerably shortened by making institutional changes, but in the absence of these changes the lags are important. (They can be partially offset by expectations effects, but clearly there is no point stimulating the economy after it has already recovered.) Campbell Leith and I did some calculations to assess the importance of these lags in some (unsubmitted and therefore unpublished!?) work here.
The second good reason is more interesting. Discretionary fiscal action can be asymmetric. Governments may be very keen to cut taxes and increase spending in a downturn, but less interested in doing the opposite in a boom. Automatic stabilisers, on the other hand, are pretty symmetrical. As a result, discretionary fiscal policy can lead to deficit bias. I argued (see, in particular, section 4) more than a decade ago that, of the many arguments that can be used against discretionary countercyclical fiscal policy, these two were probably the most important for economists precisely because most economists were in favour of automatic stabilisers.
Ironically, that asymmetry in countercyclical fiscal policy may make a good deal of sense in a world of low inflation targets where there is a danger of hitting the zero lower bound (ZLB) for interest rates. When there is not that danger, we rely on monetary policy to do our business cycle stabilisation, for all the reasons I outlined here. If there is a good chance that we could hit the ZLB, on the other hand, it would be prudent to undertake expansionary fiscal policy as a precautionary measure. (It should be undertaken in advance partly because of implementation lags.) That means that in normal times (i.e. when we are not at the ZLB) it makes sense to gradually reduce the stock of government debt – even if it is not thought excessive – so as to allow for expansionary fiscal policy when the economy is hit by large negative shocks. The argument is elaborated here.
For a member of a monetary union, however, countercyclical fiscal policy should be symmetrical, and it was the failure to understand this which was in my view a major cause of the current Euro crisis. In that 2000 paper I argued that one way of avoiding deficit bias would be to give central banks the ability to temporarily change a small number of fiscal instruments. Perhaps not surprisingly, I have found this proposal to be rather unpopular among politicians. However, I think it would have had considerable merit if it had been part of the Euro architecture. If nothing else, it would have given something for all those national central banks to do after the ECB was formed. More seriously, it might have helped reduce the scale of the crisis the Eurozone now finds itself in.