Tim Harford is yet another to consider the problems of macroeconomics. He says that the field hasn't answered or even asked the right questions. I wonder, what questions should be asked? Has anyone asked them? It seems like that's an important step to take. There should at least be a dialogue on the issue about it.
A lot of models look at output, or consumption, interest rates, or exchange rates... those seem to be the answers people want to know. When are we going to get out of this recession? What will the level of unemployment be a year from now? How much money will people lose? Those are questions that we can roughly provide models for, and questions people want answers to.
I think, other questions are much more detailed--and maybe would push the field toward finance. Note that we can try to answer these questions, but not rigorously. There's something to be said for a model, I think, in a world where people look to forecasters, and firms, industries, central bankers, and governments use models. People at least want to see numbers as rough estimates.
I find intro economics courses interesting, because there are a number of things taught that aren't ever covered again in higher-level courses. They seem much more basic. The three functions of money. The flow of money. We make lists and draw diagrams, but to what degree are these things deeply studied? As financial crisis turned into recession, these are issues that could be studied more. How is a crisis in the financial industry different from a crisis in the book industry? To what extent should we care? What models do we have to illustrate this? How often are the circular flows of the economy even discussed? The first cycle that comes to an economist's mind is probably the business cycle. Are we ignoring the money cycle? I learned a more complicated version than I see in Mankiw's book*.
We can use microeconomics to model firms, and even multiple firms, but without regard to a larger effect on the economy. We can model a generalized economy. There's an important disjoint there, particularly when a firm becomes "too big to fail." What happens when a single firm or industry gains price control over the economy? What happens when firms or industries affect each other? You can try to model firms and industries as small open economies or large open economies, but does that take you far enough? What are the effects when one of these diminishes in size by a significant amount? What are important differences between firms, industries, and economies, and how can this be modelled?
I think that rather than refine our current models, macroeconomists of the future will use a bunch of different models. There will be a number of models predicting different things (much like weather forecasting), and we'll report a range of possible outcomes with likelihoods--not unlikely today's forecasters, I suppose, but with more models. We won't have as much, "I use this model, not that model..."
*no offense to Mankiw, or an otherwise good textbook--this is purely a commentary on macroeconomics