The stock-flow consistent approach: background, features, and objectives

1. The stock-flow consistent approach: background, features, and objectives Marc Lavoie University of Ottawa

2. Outline • Background of the SFC approach • Main features of the SFC approach • Main features of the post-Keynesian SFC approach • Limits of the SFC work • Applied SFC spirit: The Levy model

3. Background

4. 1.1 Keynesian and modern (Barro GDE) macroeconomics • Y = C+I+G = W+P • There is no room or no role for banks • What about the central bank, where does it fit? • Individuals and firms are often netted out (representative agent) • Where does personal saving go? • What are the liability counterparts of this saving? • What sector provides the counterparty to the transaction? • How are government deficits financed? • What role play financial stocks?

5. 1.2 National accounting and flow of funds analysis 1940s-1950s • Macroeconomics is based on the system of national accounts of the UN 1953 (Richard Stone) (flow national income and product accounts) • This system left out flow-of-funds and balance sheets • French and Dutch national accountants bitterly complained then (Denizet: irony) • “When total purchases of our national product increase, where does the money come from to finance them? When purchases of our national product decline, what becomes of the money that is not spent?” (Copeland 1949) • The 1968 new System of National Accounts (SNA) remedies to all this (and again in SNA 1993). But to no avail, despite the introduction of Social Accounting Matrices (SAM).

6. 1.3 James Tobin 1960s - 1982 • The New Haven, or pitfalls, school • Introduces balance sheets, with several distinct assets and liabilities • Behavioural equations defining portfolio decisions, based on rates of return • The debts of a sector are the assets of another sector: Financial interdependence • Adding-up portfolio conditions: if you desire less of an asset, you want more of another

7. 1.4 A Wall Street view: the American Post-Keynesians 1960s-1970s • Paul Davidson and Hyman Minsky • One must at least distinguish between money and equities • PK economics in the 1960s is like Hamlet without the Prince • Debts generate flow commitments • “The structure of an economic model that is relevant for a capitalist economy needs to include the interrelated balance sheets and income statements of the units of the economy” (Minsky 1996, p. 77).

8. 1.5 Godley and Cripps 1983 and the New Cambridge • A response to monetarism • Keynesians did not pay enough attention to money and other financial assets • Keynesians did not pay enough attention to inflation accounting • Need to introduce the constraints which adjustments of money and other financial assets impose on the economy. • Money stocks and flows must satisfy accounting identities in sectoral budgets, most notably The NAFA of private sector (net financial saving of private sector) = PSBR + current account balance

9. 1.6 Eichner 1989 textbook • Eichner (1987, pp. 810-838) devotes nearly 30 pages to flow of funds analysis in the chapter on money and credit of his main book, with more than a dozen tables reproducing flow of funds consequences of various decisions by economic agents. • The very first of these tables (Eichner 1987, p. 811) illustrates the quadruple accounting entry principle first put forth by Morris Copeland, the US creator of flow-of-funds analysis.

10. 1.7 Providing a traverse: Godley in the 1990s at the Levy Institute • Ironically, both the Tobin and the Godley research programs had to be abandoned in 1983, when their funding was cut off. • For ideological reasons (Thatcher) • Econometric performance, due to collinearity problems, was a mixed success at best (Buiter 2003) • In the 1990s, W. Godley makes a link between his previous work • which tracks income flows and the money/debt stock through time, • and the work of James Tobin • which focuses on portfolio choice and rates of return. • Godley 1996 Levy Institute working paper, with equities, but still without growth • Godley uses simulations to describe his models and tracks variables through time.

11. 1.8 Similar proposals elsewhere • Various authors propose, at some point or another, to put together NIPA and flow-of-funds. • Bain, A.D. “Survey in applied economics: Flow of funds analysis”, Economic Journal, December 1973. • Roe, Alan R. “The case for flow of funds and national balance sheet accounts”, Economic Journal, June 1973. • Davis, E.P. “A stock-flow consistent macro-econometric model of the UK economy – Part I”, Journal of Applied Econometrics , April 1987. • Patterson, K.D. and Stephenson, M.J. “Stock-flow consistent accounting: A macroeconomic perspective”, Economic Journal, September 1988. • Dawson, John C. (ed.), Flow-of-Funds Analysis: A Handbook for Practionners. M.E. Sharpe: Armonk, 1996. • Klein, L. “Some potential linkages for input-output analysis with flow-of-funds”, Economic Systems Research, September 2003.

12. 1.9 Recent work and research teams • Lavoie – Godley, JPKE 2001-2002 • Godley – Lavoie, “Monetary Economics”, 2007 • Godley, Lavoie, Zhao, Daigle, Greenwood on open economy models • Lance Taylor, Barbosa-Filho etc. on SF models and the business cycle • Taylor, Eatwell, Mouakil on financial markets • Cripps, Izurieta etc. on world model • Jacques Mazier, Clévenot, Guy: financialization, open economy models • Van Treeck, Le Héron, etc. on financialization • Zezza – Dos Santos, growth models • Zezza, SFC and housing • Macedo e Silva and Dos Santos, methodology of SFC

13. 1.10 Otherrelated SAM work • Flaschel, Semmler, Chiarella, Franke, Charpe and others (2000s) withdifferentialequations (Keynes-Meltzer-Goodwin-Tobin model) • Harvard Structuralist General Equilibrium Model (Easterly and L. Taylor) and Computable generalequilibriummodels (S. Robinson, World Bank) • Financial computable generalequilibriummodels (Bourguignon, OECD)

14. Main features of the SFC approach

15. 2.1 No black holes • “The fact that money stocks and flows must satisfy accounting identities in individual budgets and in an economy as a whole provides a fundamental law of macroeconomics analogous to the principle of conservation of energy in physics”. (Godley and Cripps 1983) • Everything must add up. • The simplest way to make sure that nothing has been forgotten is to construct matrices. • This consistency requirement is particularly important and useful in the case of portfolio choice with several assets, where any change in the demand for an asset, for a given amount of expected or end-of-period wealth, must be reflected in an overall change in the value of the remaining assets which is of equal size but opposite sign (cf. Tobin)

16. 2.2 The quadruple entry principle • This principle is attributed to Copeland (1949). • Any change in the sources of funds of a sector must be compensated by at least one change in the uses of funds of the same sector. • But any transaction must have a counterparty. Therefore the above two changes must be accompanied by at least two changes in the uses and sources of funds of another sector. • « Becausemoneyflows transactions involvetwotransactors, the social accountingapproach to moneyflowsrests not on a double-entry system but on a quadruple-entry system » (Copeland, 1949)

17. 2.2A The quadruple entry principle Sources of funds: + sign; Uses of funds: minus sign

18. 2.2B The quadruple entry principle

19. 2.2C The quadruple entry principle and Minsky

20. 2.3 Transactions flow, stock, and revaluation matrices: Full integration • Three matrices are needed to track flows and stocks. • A stock matrix (balance sheet) • A matrix of transactions (flows) • A revaluation matrix (capital gains) • Each stock can be tracked by taking into account both the transactions matrix and the revaluation matrix (full integration)

21. 2.4 An example, a closed economy with capital gains (Model LP)

22. The transactions-flow matrix LP

23. Assessing capital gains in discrete time

24. A full-integration matrix, households only

25. 2.5 Portfolio equations and the adding-up conditions

26. … In matrix form First (Tobin) vertical condition:

27. Other vertical adding-up conditions

28. Horizontal (Godley) adding-up conditions The effect on demand for the asset in question of an increase in its own rate of interest, with all the other rates remaining constant, should not be any different from that of a fall, of the same size, in all the other rates, with the own rate staying put.

29. Symmetry (B. Friedman) conditions An increase in the expected rate of return on bonds will generate a drop in the holdings of bills that will be of the same magnitude as the drop in the holdings of bonds generated by a similar increase in the rate of interest on bills. Symmetry conditions + vertical conditions imply horizontal conditions

30. The importance of adding-up conditions

31. 2.6 SFC models: counting equations and the redundant equation • Each variable must be defined by one equation (a behavioural equation, a definitional equation, or an identity equation). • To track variables (in large models), it is best to put each variable on the left-hand side of one and only one equation. • Each column of the transaction-flow matrix provides an identity, that can be used to define one variable (say m). • Each row that contains more than two terms also provides an identity (say n) {If there are only two terms, the identity is “ordinary”, i.e., obvious, and the two terms need not be distinguished (Gs, Gd)}. • One of these identities must be removed from the simulation model, for otherwise the software will tell you that the model is over-determined. This last equation is the “redundant” or “hidden” equation.

32. 2.7 Constructing the model • Start by assuming that all stocks of the balance sheet add up. • Make sure that the row identities of the transactions-flow matrix are fulfilled. • Make sure the adding-up conditions of the parameters of the portfolio component are verified. • There is no need to start from the equilibrium. • Running the model will get you there, if the equilibrium is stable. • In complexmodels, itmaybequite long and difficult to find a steady state (startfrom a simple model, and add feedback relations afterwards) • Once an equilibrium has been found, parameters can be modified to examine what happens.

33. 2.8 Closures • The same model can be closed in several different ways. • This can involve bumping and inversing several equations • For instance, a 2-country model can be closed with the assumption of a fixed exchange rate, which is held constant because: • A central bank accepts to purchase/sell any foreign asset at the constant exchange rate (endogenous foreign reserves) • Interest rates are let to move freely to keep the exchange rate constant, while foreign reserves stay constant • Government expenditures are let to move freely to keep the exchange rate constant, while foreign reserves stay constant

34. Main features of the Post-Keynesian SFC approach

35. 3.1 Market-clearing vs buffers • In neoclassical economics, markets clear through price changes. • In post-Keynesian economics, markets clear either because quantities supplied are assumed to adjust to demand within the period or because of buffers. The price mechanism in our models only plays a clearing role for stock market equities. • SFC models normally have a buffer for each sector: • Stocks of inventories and loans for producing firms • Money deposits for households • Bills held, or advances from the central bank, for private banks • Bills issued for the government

36. 3.2 Supply-led vs demand-led models • Our models are essentially demand-led. • It is possible to introduce supply-led effects (Phillips curves, and so on) • One could also introduce other supply effects, such as reduced capacity when producing firms default on some of their loans

37. 3.3 Optimization vs reaction to disequilibria • In neoclassical economics, optimization under constraints rules. • In PK SFC models, economic agents often have stock-flow targets (inventories to sales ratio, wealth to disposable income ratio, capital adequacy ratios, …). • Economic agents react to these disequilibria by trying to close the discrepancy

38. 3.4 Contributions of the SFC approach to PK theory • A coherent formalization of Minsky’s financial fragility hypothesis (formalizing the current crisis?) • A coherent formalization of the stock market • A systematic introduction of stock-flow norms • The full integration of real and financial analyses • The integration of money as a flow (as in circuit theory) and as a stock (Keynes, portfolio choice) • The ability to consider historical time, as in Joan Robinson, by tracking variables through time

39. Limits of current SFC work

40. 4.1 It can be cumbersome • As soon as more realistic models are being considered, the number of equations rises very quickly. • A partial way out has been suggested by Mouakil (2006), by ignoring all “ordinary” identities. • But still, models remain large. • Continous time vs discrete time or differentialequations vs differenceequations

41. 4.2 The need for calibration • In our book (Godley and Lavoie 2007), we made (nearly) no attempt at calibration. • More effort could be put in calibration • More effort could be put into empirical work. • Are new time series techniques adequate to deal with the old problems of collinearity encountered by Tobin and his associates?

42. The Levy model of the U.S. economy

43. 5.1 Financial balances • In Godley’s approach, the analysis of financial balances gives clues to where the economy is going in the medium term. • The NAFA of private sector (net financial saving of private sector) = PSBR + current account balance (S – I) = (G – T) + (X – IM + NFY)

44. 5.2 The “New Cambridge hypothesis” • In the 1970s, Godley and his group used this identity as the basis for a model of the UK economy. For the private sector as a whole, the difference between saving and investment is equal to the net acquisition of financial assets (NAFA), which in turn are equal on (the increase in) liabilities of the government or the rest of the world NAFA = GD + CAB • This approach was unconventional, since it merged households and business analyzing the private sector as a whole. • The “New Cambridge hypothesis” was that NAFA was stable, relative to GDP, and this gave rise to a theory for “twin deficits”, i.e. any imbalance in the foreign account was matched by an imbalance in the government account. • In the face of crisis which called for an expansionary fiscal policy on Keynesian lines, it was thus necessary to adopt measures to counter the implied widening of the current account imbalance (some kind of protectionism).

45. 5.3 The “New Cambridge hypothesis” • The “New Cambridge hypothesis” was based on empirical grounds (see Mata, 2006 for a nice reconstruction of the debate at the time), and the evidence in favor of the stability of NAFA/GDP vanished after some years, so that the approach was progressively neglected. • This approach is still at the heart of Godley’s empirical work. It is not, however, grounded in the stability of NAFA, but rather on how financial balances are moving, with a strong emphasis on boundaries for balances. • Since each balance also measures the net increase in a sector net debt, values of the balances above a given level point to instability in stock-flow ratios which may trigger either some adjustment or a crisis. • Similar assessment, BêDuc and Le Breton 2009, ECB

46. 5.4 The fundamental identity in a closed economy (Krugman2009, via GoldmanSachs, via Levy): (S – I) = (G – T) Private Surplus Public Deficit Less investment More saving Private Financial Balance B R D GDP E Public Balance E starting point D without automatic stabilizers (1930s) R with automatic stabilizer B also with stimulus plan Private Deficit Public Surplus

47. 5.5 Levy model projection 2007

48. 5.6 Household Debt One of the reasons for the projection was the unsustainable path of household debt relative to disposable income, along with the dynamics of the price of homes.

49. 5.7

50. 5.8 Levy model – private sector expenditure The crucial equation is the private sector expenditure function, which – under a standard assumption in new Cambridge models à la Godley – implies a long-run stock-flow norm. PXt = c0 + c1YDt + c2FAt-1 + Zt • Where Z is a vector of stationary variables which influence the propensity to spend out of income