"Debunking Economics" is a critique of neoclassical economics by a Post Keynesian who likes Sraffa, math and dynamics as much as he dislikes neoclassical economics, equilibrium and statics.
The book has three main parts: Foundations, Complexities, and Alternatives. Foundations includes discussions of neoclassical supply, demand, perfect competition and monopoly, as well as the labour market. Complexities covers capital, assumptions, time, macroeconomics and financial markets. Finally, Alternatives part presents alternative financial theories and heterodox schools of economics, argues against Marxian economics and discusses the use and abuse of mathematics.
The idea behind the book is great. The execution itself, unfortunately, is a mixed bag. I found two chapters - Chapter 6 on capital and Chapter 13 on Marxian economics - to be almost unbearable. All the other chapters are more accessible and have promising arguments, albeit delivered in a mostly tedious manner. In this regard, Chapter 2 is quite representative of the whole book. It points out that in order to derive a smooth downward sloping market demand curve, economists assume homogeneous preferences, which amounts to having only one consumer in the market. Heterogeneous preferences would yield an unpredictable, messy curve with possibly more than one equilibrium. This interesting point is arrived at through a series of boring utility hills and indifference curves. Likewise, Chapter 8 presents a nice discussion of Walras and general equilibrium, but ends with a confused exposition of what temporal analysis in economics should be. Above all, in certain cases when the lay reader (who is supposedly the author's primary audience) might need some easing in (e.g. when Hicks and IS-LM are being discussed), Keen does that very little if at all. Without doubt, the delivery of the arguments could have been much better - much more focused, more accessible. Furthermore, some interesting points are left to be developed on the author's website. Why not use increasing returns to scale (which are otherwise taken so seriously by Keen) possible in this book and develop those issues here as well? I would argue that because of this lacking delivery, many otherwise solid arguments and points will be forgotten soon after being read by the lay reader.
Long story short: this book is a good idea executed in a sloppy manner.