New Keynesian Macroeconomics

Transcription

1 84 New Keynesian Macroeconomics Peter Spahn Introduction For any one who is not familiar with history-of-economic-thought debates, it might be confusing to hear of all these different branches of Keynesianism: post-, neo-, new- etc., where single sub-classifications sometimes even seem to convey contradictory messages, and some critics also held that one or the other branch is no "true" Keynesianism after all. This latter assessment particularly can be found with respect to New Keynesian Macroeconomics. Neoclassical Synthesis (IS-LM) Post Keynesianism Fundamental Keynesianism Neo Keynesian Synthesis New Classicals New Keynesian Macro Theory A family tree of macroeconomic thought, adapted and developed further from Leijonhufvud (1981), shows New Keynesian Macro Theory (NKM) as being merged from a Neo Keynesian Synthesis and the Real Business Cycles Approach (RBC). The first mentioned Synthesis, propagated by Mankiw, Stiglitz and others in the 1980s and 1990s, tried to counter the permanent-market-clearing postulate of Lucas's New Classical School by pointing to important wage and price rigidities. These market imperfections could be explained by resorting to traditional neoclassical components like information deficiencies, but helped to maintain the Keynesian agenda of involuntary unemployment and the indispensability of demand policies. Also the RBC Paradigm was a response to the New Classicals. But here, the aim was to strengthen the view of a market economy's dynamic efficiency, particularly in case of real

2 2 supply-side shocks that were regarded as much more important than the monetary demand disturbances analysed hitherto. RBC clearly is a late part of the "Wicksell Connection" (Leijonhufvud 1981) where the issue of intertemporal allocation of spending and saving governs the substance and structure of macro theories, compared to the Quantity Theory of Money where, following Fisher and Friedman, a stable money demand is supposed to guarantee a more or less fixed relation between money holdings and goods demand, irrespective of the latter's structure. RBC describes an optimal allocation of a "representative" agent's labour supply and goods demand over time, i.e. a sequence of per-period saving and consumption, which satisfies an intertemporal utility function in a setting of random productivity shocks. But all this was demonstrated with flexible wages and prices so that hardly any role for economic policy remained. NKM therefore introduced nominal rigidities, which had been the Neo Keynesian key topic, in order to allow for short-run real effects of monetary policy activities. Finally, central bank behaviour was represented as active interest rate policy with the approval of money endogeneity, which on the one hand merely meant to accept the facts, but also conformed to an assumption that Post Keynesians had strived for since decades. From all this, a basic three-equation NK model ensued that offered attractive constituents for economists from various camps: first and foremost, its microeconomic foundations ensured its acceptance among young "mainstream" members. Second, supply-side frictions (that were introduced as emerging from optimising firm behaviour under adverse market conditions) made model simulations look similar to ordinary macroeconomic cycles, particularly if some doses of persistence were added to the array of stochastic shocks, and thus recommended the use of NKM in empirical analyses of booms and depressions. Finally, the active role of interest rate policy that was needed to guarantee the model's dynamic stability pleased "old" Keynesians, and also kept the door open for including institutional peculiarities of monetary authorities, which had been a recurrent theme in New Classical Economics since the 1980s. What Would Keynes Have Thought of NKM? The RBC School re-invented macroeconomics as a version of optimal growth theory: shock-driven deviations from a path that itself is a result of micro decisions on consumption and saving over time. The question of optimal growth was early raised by the young mathematician Frank Ramsey (1928): "How much of its income should a nation save?" Keynes (1933: 335) admired the now famous paper of his friend; he regarded it as "one of the most

3 3 remarkable contributions to mathematical economics ever made. [...] The article is terribly difficult reading for an economist, but it is not difficult to appreciate how scientific and aesthetic qualities are combined in it together." The answer to Ramsey's question, later labelled the Keynes-Ramsey Condition of an optimal path of consumption (Blanchard/Fischer 1989: 41-43, Carlin/Soskice 2006: 216), can be given in the compact expression 1+ δ rt δ C = C cˆ = δ t t+ 1 t+ 1 rt where (calculated from a logarithmic utility function) consumption C t depends on its (expected) future value, and also on the ratio of time preference δ and the real rate of interest r t, so that consumption growth is positive if the latter exceeds the former. Basically, the equation is also used to explain goods demand in RBC and NKM. The derivation of the optimality condition is straightforward from a microeconomic point of view. However, in an indirect comment to the Ramsey model, Keynes (1932: 400) pointed to the interdependence of demand and income in a macroeconomic perspective, on account of which it must appear dubious to base an explanation of consumption demand on time preference and interest rates alone. "For a single individual the notion of time preference is fairly clear. Given all the relevant attendant circumstances which are fixed for me by the actions of others including my income, actual and prospective, and the prices, actual and prospective, of debts, assets and consumables, it is my state of time preference which determines what part of my income I spend on consumables and what part of it I reserve. I say that this is fairly clear, since we do not have to suppose that either the amount of my income or the various relevant price levels appreciably depend on my decision. But when we try to deduce from the general state of time preference, i.e. from the complex of individual time preferences, what part of the community's aggregate income will be spent and what part will be reserved, we are soon in difficulties. For the amount of total expenditure responds immediately to the amount of total income, whilst, for the community as a whole, the amount of total income depends not less directly and immediately on the amount of total expenditure. We are therefore on shifting sands, and must approach our goal more circumspectly and by a different route." The key point of course here is the treatment of income. If, as in the basic Robinson- Crusoe approach of the RBC-NKM world, it is taken to be a choice variable of the "representative agent", there is nothing wrong with deriving consumption behaviour from interest rates (De Vroey 2016: 263, 276). In a modern market economy, this translates into the crucial question whether labour market clearing via a flexible real wage can be taken for granted or not. As a result of his reasoning, Keynes denied that question, but it is answered in the af-

4 4 firmative on the part of NKM as a starting point of the analysis! The model's constituent equation, generalised and log-linearised from above (with r indicating the equilibrium rate, and σ the inverse of the elasticity of intertemporal substitution of consumption with respect to interest rate changes) thus shows how interest rates shift goods demand between periods, but do not induce demand to come forward, as in traditional Keynesianism. ( ) y = E y σ r r + ε d t t t+ 1 t t The equation describes a consumption economy ( Yt = Ct), investment is of low importance, comparable to the demand for intermediate products. Investment is no "strategic" variable as in Keynes, because income is seen to be pre-determined on an efficient labour market. If full employment is assumed as a starting point, the topic of intertemporal co-ordination failures that has characterised macro theories within the Wicksell Connection loses much of its significance. No it does not appear that Keynes would have appreciated NKM. Banking and Finance Maybe the last paragraph's conclusion is premature. With a perfect intertemporal allocation of agents' saving and consumption, the risk of effective demand failures is substantially reduced so that the labour market is not burdened with large adjustment costs. NKM started with the assumption of complete financial markets where all agents are able to place promises and claims for the future delivery of goods and resources, in accordance with the profile of their utility function. The intention to pay respect to one's own life-time budget constraint (i.e. to obey to the transversality condition of preserving individual wealth as non-negative) is essential in the optimisation calculus; this has the implication of turning individual IOUs into "safe assets". "Intertemporal general equilibrium models have solutions that coordinate saving and investment decisions over an infinity of future periods. They do so without fail because they assume the trading plans of all its members to be tied together by a transversality condition way out there at the end of time. [...] Remove the transversality condition from DSGE models and everything unravels. Without it, there is nothing to guarantee that individual intertemporal plans are consistent with one another. [...] Walrasian constructions, even those of recent vintage, take for granted that budget constraints are binding" (Leijonhufvud 2014: 12-13, cf. Goodhart 2007). Of course, the idea of perfect financial markets was not meant to be a realistic assumption. It served to close the model at an early stage of its creation and development where the task still was to elaborate on the synthesis of RBC's demand theory and supply-side pricing

5 5 decisions, and to integrate a central bank's interest rate impulses into an essentially nonmonetary model without any role for banks. Hence, it was a disaster for NKM's reputation that the US financial crisis broke out "so early". Critics poured out vials of wrath on NKM as this theory appeared to have ruled out failures in the banking system on assumption. But actually up to 2007, NKM had not even dealt with the problems of financial intermediation the risks of the banks' securitisation practices were ignored in other branches of economic science. This is not to argue that NKM offered a convincing theory of banking when the financial sector finally was included in the model. The gist of this integrative step can most easily be understood from Woodford's (2010) non-algebraic paper on the role of financial intermediation in macroeconomic analysis. Here, the author designs an IS-LM-like presentation of the NK view where this is the anti-keynesian element investment and saving do not determine the level of income; rather, functions of lending and borrowing are derived from exogenous income variations. Their intersection point establishes, as in Wicksell, a market clearing real rate of interest. The introduction of frictions (emerging from, e.g., asymmetric information) creates a wedge between savers' and final borrowers' interest rates, the mark-up remunerates the banks' services in managing various credit market risks. This approach is in line with modern financial market theory, but it resumes an old neoclassical view of financial intermediation according to which banks collect savings that are then passed on to investors afterwards. "Households deposit funds in financial intermediaries that in turn lend funds to nonfinancial firms" (Gertler/Kiyotaki 2010: 550). Financial intermediation resembles an indirect lending of resources. There is no place for credit creation on the part of the banks, an insight however already central to Wicksell, and which was forcefully taken up by Schumpeter later. Somehow the message "loans create deposits" got lost in the wake of Tobin's (1963) influential "new view" on financial intermediaries: it is of course true that deposits, arising as bookkeeping entries in bank credit contracts, do not necessarily conform to non-banks' portfolio preferences, and therefore will be used in a sequence of transactions. But in this process, the newly created deposits cannot get lost (the exception is paying off bank loans or buying bank assets). Therefore even if, due to a lack of alternative financial assets, at the end of a multiplier process the increase of bank deposits consists of private savings, the neoclassical view "deposits make loans" ignores the stepwise increase of gross nominal money wealth that comes as a by-product of credit creation in the banking sector. The implication is that also the distinction between the roles of financial markets and of banks is

6 6 is blurred: "This is the key feature that differentiates bank lending from nonbank credit. Capital market intermediation, like barter and commodity money or cash-based systems, requires that the creditor have on hand the means of payment to deliver to the debtor before the credit is extended. In modern financial systems, credit transaction between nonbank agents essentially involves the transfer of deposits. Bank lending, on the other hand, involves the creation of bank deposits that are themselves the means of payment. A bank can issue credit up to a certain multiple of its own capital, which is dictated either by regulation or by market discipline. Within this constraint, the growth of bank lending is determined by the demand for and willingness of banks to extend loans. The latter being influenced by funding conditions in the market. There is no quantitative constraint as such" (Disyatat 2011: 716). Macro bank models in NKM start from the balance-sheet budget constraint credit + bonds = capital + deposits where deposits change only slowly on account of non-banks' optimal saving decisions. Thus, bank credit appears as funding-restricted and likewise follows a dampened path. NKM is unable to treat the volumes of bank credit and deposits as "jump variables". As a consequence, the dynamics of the macro cycle after shocks emanating from the bank sector is biased downwards in simulation models because credit and deposit growth as a discretionary step of balance sheet lengthening is ignored (Jakab/Kumhof 2015). It comes as no surprise that NK economists sometimes even hesitate to acknowledge the power of central banks to create base money. Instead, they are also presented as intermediaries who obtain funds from financial markets and channel them into specific sectors (Gertler/Kiyotaki 2010: 567-8). Deeply rooted in their habit to reduce economics to the pattern of constrained optimisation, modern economists want to describe monetary authorities in the same way as private agents, in order to lift the veil of "money mystique". "The central bank [...] must raise new funds, [...] in order to: (i) pay for the outstanding special liabilities and interest-bearing liabilities, [...] (ii) expand the balance sheet by buying new assets [...] in excess of the gross return on last period's assets [...], and (iii) pay dividends" (Reis 2013: 136). A Consensus in Macro Theory? Very rapidly NKM has conquered the scientific departments of central banks and similar agencies, not to mention the editing committees of academic journals. This School has proven to react rather flexible to upcoming needs on the research agenda. The days of a single "representative agent" are gone; models with several sectors in goods and financial markets are now

7 7 standard, often populated by agents with heterogeneous preferences. As microfounded models necessarily are rather complex in "vertical" perspective, because all activities are derived from individual optimisation, multi-sector analysis adds to "horizontal" complexity. Rational, i.e. model-consistent, expectations is a standard assumption, but little attention is paid to the question whether markets agents will understand and calculate the working of a, say, tenvariable dynamic system, given the impression that actual human beings most probably are unable to envisage anything beyond a two-variable structure (supply and demand). Blanchard (2016) argues that even within economic science communication is hampered by too complicated NK models, and calls for the use of more simple complementary theories. With regard to econometric research, NKM is proud not to test single equations only, but models in their entirety, in order to capture the interdependencies between all variables. However, this poses the intractable problem that the number of model parameters exceeds the scope of empirical specification. There are simply too few degrees of freedom. This drawback is sidestepped by "calibrating" key parameters in advance, which then ensures the model "fit" but leaves serious questions as to the significance of NKM (De Vroey 2016: 279, Romer 2016). NK economists often pay lip services to model realism by allowing for small groups of agents who do not optimise but follow rules of thumb, or who suffer from market rationing. But this does not affect the general model architecture that shows a bias towards fully specified solutions where market agents are assumed to be able to calculate each other's net wealth positions and their budget constraint implications, far into the indefinite future. An example is the supposed validity of Ricardian Equivalence; authors have to state their reasons for any deviations from this hypothesis that has often has been rejected on empirical grounds. The NK project "microfoundation of macroeconomics" is a large step towards a reestablishment of Walrasian equilibrium theory. Model agents are assumed to dispose of wideranging knowledge about the implications of each other's decisions, but they do not grasp the endogenous evolution of "natural" rates of output and employment. NKM is subject to Hayek's critical notion of a "pretence of knowledge", and it downgrades an early Keynesian finding: the need for macroeconomic theory arises from the fact that market processes cannot properly be understood from the analysis of the multitude of micro decisions alone.

8 8 References Blanchard, O. J. (2016): Do DSGE Models Have a Future? Peterson Institute for International Economics, Policy Brief, Blanchard, O. J. / Fischer, S. (1989): Lectures on Macroeconomics. Cambridge / London (MIT Press). Carlin, W. / Soskice, D. (2006): Macroeconomics Imperfections, Institutions and Policies. Oxford (Oxford University Press). De Vroey, M. (2016): A History of Macroeconomics from Keynes to Lucas and Beyond. New York (Cambridge University Press). Disyatat, P. (2011): The Bank Lending Channel Revisited. Journal of Money, Credit, and Banking, 43, Gertler, M. / Kiyotaki, N. (2010): Financial Intermediation and Credit Policy in Business Cycle Analysis. In: Friedman, B. M. / Woodford, M., eds.: Handbook of Monetary Economics, Vol. 3. Amsterdam et al. (Elsevier), Goodhart, C. A. E. (2007): Monetary and Social Relationships. In: Giacomin, A. / Marcuzzo, M. C., eds.: Money and Markets. Oxon / New York (Routledge), Jakab, Z. / Kumhof, M. (2015): Banks Are Not Intermediaries of Loanable Funds. Bank of England, Working Papers, 529. Keynes, J. M. (1932): The Parameters of a Monetary Economy. In: Moggridge, D., ed.: The Collected Writings of John Maynard Keynes, Vol. 13: The General Theory and After, Part I: Preparation. London / Basingstoke (Macmillan) 1987, Keynes, J. M. (1933): Frank Ramsey. In: Johnson, E. / Moggridge, D., eds.: The Collected Writings of John Maynard Keynes, Vol. 10: Essays in Biography. London / Basingstoke (Macmillan) 1978, Leijonhufvud, A. (1981): The Wicksell Connection. In: Information and Coordination. New York (Oxford University Press), Leijonhufvud, A. (2014): Economics of the Crisis and the Crisis of Economics. European Journal of the History of Economic Thought, 21, Ramsey, F. P. (1928): A Mathematical Theory of Saving. Economic Journal, 38, Reis, R. (2013): The Mystique Surrounding the Central Bank's Balance Sheet Applied to the European Crisis. American Economic Review, Papers and Proceedings, 103, Romer, P. M. (2016): The Trouble With Macroeconomics. New York University, Stern School of Business. Tobin, J. (1963): Commercial Banks as Creators of 'Money'. In: Carson, D., ed.: Banking and Monetary Studies. Homewood (Irwin), Woodford, M. (2010): Financial Intermediation and Macroeconomic Analysis. Journal of Economic Perspectives, 24, 4,