I apologise in advance for the headline but anyone familiar with academic economics will know that the IS/LM model is well and truly left behind after first year econ 1-0-1. Of course, Paul Krugman disagrees and since the financial crisis he has been championing the use of these models or more specifically, championing the focus on demand side policies with only scant focus on prices/inflation.
In a recent presentation, Krugman cuts to the bone.
IS-LM used to be at the core of macro, as taught and to a considerable degree as practiced. However, it receded greatly in influence during the 80s, 90s, and early 2000s, crowded out from two directions. On one side were the purists, who insisted that macro must be built as far as possible on micro foundations, and that IS-LM was too ad hoc. On the other side were the pragmatists, who pointed out that central banks, after a brief flirtation with monetarism in the late 70s/early 80s, had shifted to interest rate targeting, making the LM curve irrelevant. In the crisis, however, IS-LM has come back into its own.
Krugman's argument I think is slightly tenuous but not completely off-base. If economic slack is high and desired nominal interest rates are significantly below zero (making the zero lower bound binding) then it is very plausible that the workings of the demand side IS/LM model come into their own in the context of policy affecting nominal GDP. We could say that nominal GDP targeting may work as long as we have slack in the economy. The main question is of course whether this holds as a general rule. A corrollary to this argument is of course the notion that the fiscal multiplier is particularly large when the economy is at the zero bound.
The main problem however is that you quickly come full circle and run into some universal problems that neither IS/LMers nor the representative agent/micro foundation crowd have an answer to (in my honest opinion at least).
- Can we measure economic slack?
- Can we measure the output gap?
- Does inflation targeting yield a credible policy rate if the output gap measurement is very uncertain? What happens to financial market and asset prices (booms and busts) as a function of CPI inflation targeting and how do these propagate to the real economy?
- Are there income distribution and inequality effects from pursuing expansive monetary policy at the zero bound?
And I could go on.
Of course, the Rational Expectations school with Lucas, Sargent, Blanchard etc as the founding fathers and their New Keynesian representative agent models with all the bells and whistles  are not exactly providing much better answers.
It is also interesting that Krugman mentions Eggertson. Eggertson was of course Krugman's student at Princeton and is widely held to be responsible for bringing the zero lower bound notion into modern macroeconomics. At least, as far as I know, it was his PhD papers that formed the foundation for the "commitment to be irresponsible" as a policy tool where forward guidance is a derivative of this. Of course, Eggertson's papers are quite difficult to follow mathematically as they are largely built on, guess what, representative agent modelling.
For the record, I am symphathetic to the notion of larger fiscal multipliers at the zero bound but the standard Taylor Rules used to define the zero bound may not be that accurate. No central bank that has ever gotten itself stuck at the zero bound has ever managed to escape. This raises fundamental issues of whether the state dependent multiplier argument holds if a central bank gets permanently stuck at the zero bound for e.g. institutional or other reasons than those merited by an inflation rule.
Finally, after 5 years of extraordinary lose monetary policy I think it is time that we recognize that such policies may also entail costs e.g. in the form of stagflation, income inequality etc.
 - Staggered pricing/sticky wages, rational expectations, general equilibrium features, (simple) financial markets, inflation targeting central banks, overlapping generations etc