I have recently ordered a copy of “Structuralist Macroeconomics – Applicable Models for the Third World” by Lance Taylor. However, I did not read very far into the book. Let me explain why. On p. 12, chapter 2 – titled “Adjustment Mechanisms – the Real Side” – starts with the sentence:
“MACROECONOMICS begins with the notion that the value of saving generated by all participants in the economy must by one means or another come into equality with the value of investment in the short run.”
While most economists will probably nod their heads, I don’t. Informed by a book chapter written by Basil Moore (download) and my own research, let me point out the fundamental problems with this statement. The first is trivial, the second not so.
First, macroeconomics – for most economists, I assume – does not begin with saving. It should start with employment. Why do we care about macroeconomics? Because we care about unemployment. If we would not care about unemployment, why then do we have central banks that decrease interest rates to spur investment in times of economic recession or depression? Why do governments engage in deficit spending? It is obvious that any social science must first define a problem that it wants to analyze and then, hopefully, fix.
The second fundamental problem is connected to the definition of saving. Moore’s chapter title, which goes back to Warren Mosler, connects saving to investment, with the latter causing the former. Moore writes: “Saving is always identical to investment, irrespective of the time unit or the time period over which they are measured, or how investment is defined” (p. 8). If Mosler via Moore is correct that saving is the accounting record of investment, than the fundamental methodology of structuralist macroeconomics must be wrong.
Moore writes that in the world economy, not distinguishing government or private sector, there are only two types of goods: consumption goods and investment goods. World GDP equals consumption + investment, or – in letters – Y = C + I. From the expenditure side, people of the world can decide to spend their money on consumption goods or … not. So, in other words, world GDP equals consumption plus what is not consumed. If we define what is not consumed as savings, then we get Y = C + S.
Subtracting one equation from the other gives us either 0 = I – S or 0 = S – I. Both can be transformed into S = I. So, there we are! This savings to investment relationship is one of an identity, hence a definition: what we do not consume we call investment from the “real” perspective of use of goods and services and “savings” from the financial perspective. Moore says that “in reality there is no underlying ‘real’ economy that somehow lies below and exists independently of the nominal economy” (p. 20). So, investment equals savings all of the time everywhere, and only problems of accounting give us problems with the data that sometimes seem to show that savings do not match investment.
In summer 2014, I taught a course at Free University Berlin and told my students that ultimately savings depend on investment – except for the case where not spending creates an increase in inventory, which is marked as investment – and it was a popular sport over the term to come up with examples that would imply that savings rise without a corresponding rise in investment or vice versa. We did not find a single valid example.
I know from many conversations I had with students that the investment-savings inequality is very difficult to grasp and had my own problems in the context of planned investment / planned saving, but by now I am very certain. Students of macroeconomics have to understand this identity from the very beginning to save them the trouble of falling for the loanable funds fallacy and other concepts based on misperceptions about he working of a modern monetary system. I’m sure that “Structuralist Macroeconomics” has a lot of knowledge that can be used to improve things, but the way that the models are drawn up will make it difficult to extract the good stuff (on institutions and distributions, as I see it). Perhaps the good stuff can be explained using balance sheets and aggregation at the sectoral level? That might be a worthwhile initiative.