Let's consider Krugman's three main points:
1. This was not an effort at fiscal stimulus; it was a supply policy, not a demand policy. The German government wasn’t trying to pump up demand — it was trying to rebuild East German infrastructure to raise the region’s productivity.
2. The West German economy was not suffering from high unemployment — on the contrary, it was running hot, and the Bundesbank feared inflation.
3. The zero lower bound was not a concern. In fact, the Bundesbank was in the process of raising rates to head off inflation risks — the discount rate went from 4 percent in early 1989 to 8.75 percent in the summer of 1992. In part, this rate rise was a deliberate effort to choke off the additional demand created by spending on East Germany, to such an extent that the German mix of deficit spending and tight money is widely blamed for the European exchange rate crises of 1992-1993.
On #1, policymakers may intend all kinds of results, including a new skat table for Onkel Mathias and bananas for Brandenburg. The scenario nonetheless saw a big boost in debt, spending, and aggregate demand, and in a setting with unemployed resources. That makes it one test case, even if the primary motive was the unification of Germany and not stimulus per se.
On #2, you willl find data on German unemployment rates here, for each half of the country and consolidated. In the West there is unemployment before unification, a big boost in employment during the first two years of unification, and then much higher unemployment.
Let's track those rates, starting in 1989: 7.9, 7.2, 6.2, 6.4, 8.0, 9.0, 9.1 (1995), 9.9, 10.8, 10.3, etc.
You can read those numbers two ways. First, stimulus worked until they lost heart, or "it boosts the economy in the short run but it postpones adjustment problems and in the medium run you end up with a poor and sluggish private sector job market." I was making the simpler historical-explanatory point that, no matter how you slice and dice the data, the poor outcome in absolute terms explains why the Germans are not so impressed by ideas of debt, spending and aggregate demand. And if stimulus proponents keep on losing heart at the critical moments, the citizenry is still right to be skeptical.
The real story requires a look at the tottering East German labor market as well. At the immediate moment of unification, unemployment/underemployment for the consolidated Germany are higher than the above numbers behind the link indicate; read for instance these comments and also here.
The strongest criticism of the example is not one that Krugman considers. If you had access to a more finely calibrated employment index, reflecting the gross underemployment of DDR workers, with continually collapsing jobs, there is at least some chance the true data series might show more effectiveness for fiscal policy than the published data, or the West German mood at the time, would indicate.
#3, like #1, is a distraction. The zero bound might — in some frameworks — make fiscal policy necessary rather than monetary policy. But the zero bound is not required to drive the main arguments that fiscal policy is effective with unemployed resources. So historical examples with a non-zero bound and ineffective fiscal policy do count against fiscal policy.
As an aside, you can find some RBC and cash constraint models where fiscal policy is more effective at the zero bound, but a) I think these papers are deeply flawed, b) the liquidity constraints in those models do much of the work and those constraints held anyway for parts of East Germany, and c) Krugman and others are advocating fiscal policy for a wide variety of economies, not all of which are at the zero bound, even if they have constrained monetary policy for other reasons, such as Ireland.
On #3, Krugman is correct that the Bundesbank actions hurt the German economy (though keep in mind, at the time of tightening in 1992 the rate of German price inflation was five percent and it stayed at 4.4 percent the year after). This implies that tight monetary and loose fiscal policy, taken together, won't work. I endorse that conclusion. It also shows that the monetary authority moves last, which is exactly the point made by fiscal policy critics such as Scott Sumner. Stressing that point does no favors to the argument for fiscal expansion, quite the contrary. Krugman is eager to slap down the historical example but he is losing track of the broader principle he cites to do so.
Overall the pro-ramp-up-the-spending fiscal policy advocates are keener to produce examples which don't count than examples which do count. The positive arsenal is pretty weak, consisting mainly of World War II, an experience which cannot be repeated or replicated.
Krugman also directs his attention at a secondary point in the argument, taking up a single paragraph in the column. He doesn't touch the main point: Germany cut off fiscal stimulus early and focused on the real economy and they're doing fine. Even if you think this prosperity is at the expense of others (and I don't), the pro-ramp-up-the-spending-stimulus view suggests that Germany itself should be ailing, yet it isn't. Here are further good comments on that topic.
On the main point of the column, you can find some good criticisms from Ryan Avent, though I would stress it is exports not stimulus that led the German recovery and that Germany is not obviously due for a spill.