Behavioral economics has demonstrated deviations from the perfect optimization depicted in standard models. Some deviations are trivial and irrelevant for macroeconomics. Others, however, are systematic and affect aggregate outcomes, including aggregate household saving. Pervasive uncertainty and the influence of ‘nudging’ on household retirement saving cast doubt on models of saving that are based on strict optimization and rational expectations. Behavioral evidence also points to possible reasons for under-saving. The well-documented existence of ‘present bias’, not surprisingly, can reduce saving, and simple models demonstrate that peer effects on consumption can also have this effect. Macroeconomic relations should indeed reflect microeconomic behavior and macroeconomics must be ‘behavioral’. But the specification of consumption in contemporary macroeconomic models is based on misleading assumptions about microeconomic behavior. The modeling of aggregate consumption must build on behavioral evidence, address aggregation issues, and consider structural constraints, including credit rationing.

Keynesian involuntary unemployment describes a market failure: market forces may be incapable of bringing the economy to full employment. Wage and price stickiness are not the problem. Keynes took prices to be flexible and viewed sticky nominal wages as desirable: flexibility would tend to make the economy violently unstable. The IS-LM model provides a decent representation of the analytical skeleton behind Keynes’s fix-wage equilibrium but leaves out dynamic forces that are central to Keynes’s instability argument. By including one of these forces – the effects of expected inflation on real interest rates – in a formal model, Tobin showed that wage flexibility does not ensure the stability of full employment. The model’s assumption of an exogenous money supply misrepresents the real-world behavior of central banks as well as contemporary theory. Introducing a Taylor rule, stability can be obtained if the zero lower bound does not constrain interest rates – a result anticipated by Keynes: “only a foolish person would prefer a flexible wage policy to a flexible monetary policy”, he argued, while warning about the ineffectiveness of the policy in a liquidity trap.

New information technology has allowed firms to monitor many low-paid workers more closely, putting downward pressure on their wages and increasing work intensity. At the top end of the income distribution, meanwhile, the new technologies have increased firm-level volatility, thereby weakening the ability of owners to monitor managerial performance, increasing the power of top managers, and contributing to an explosion in executive pay. Technological power biases have been reinforced by institutional changes, including declining rates of unionization, falling minimum wages, and changes in intellectual property rights, business regulation and anti-trust policies. The sources of inequality matter. Education will be a primary instrument if skill biases are the main source. But many workers hold jobs for which they are overeducated. This mismatch between workers’ skills and the skill requirements of the jobs can emerge naturally in an efficiency-wage model. In this setting, relative wages are insensitive to changes in the relative supply of high-skill workers; an increase in the minimum wage, by contrast, can raise employment and reduce inequality.

Aggregate demand problems can jeopardize the existence of a steady growth path in mature capitalist economies: fiscal policy may be needed to maintain full employment and avoid secular stagnation. This functional finance approach to economic policy endogenizes the movements in public debt. The debt ratio will converge towards a long-run value that depends (i) inversely on the rate of growth, (ii) inversely on government consumption, and (iii) directly on the degree of inequality. The analysis implies that policies and policy debates on the dangers of public debt have been misguided and that the incipient theoretical redirection following the rediscovery of secular stagnation by Summers and others does not go far enough. Unlike in mature economies, functional finance cannot target full employment in developing economies with high rates of underemployment. Instead, high investment rates are desirable, and functional finance should aim to stabilize the level and composition of demand at values that are consistent with a target growth rate of the modern sector; excessive aggregate demand stimulus can squeeze the modern sector and lead to premature deindustrialization.

The neoclassical aggregate production is ubiquitous in the macroeconomic literature, despite the theoretical weaknesses exposed by the Cambridge capital controversy. This chapter gives an overview of the controversy and subsequent attempts to provide an empirical justification for the aggregate production function. These attempts are unconvincing, and the function is not needed to explain the reconciliation of the natural and warranted growth rates in mature economies. Kaldor's neo-Keynesian models and an older Marxian tradition explained this outcome by endogenizing the distribution of income and assuming that the saving rate is increasing in the profit share. But capitalist economies contain both stabilizing and destabilizing forces. Depending on the balance of these forces, the result may be endogenous cycles, as in Goodwin’s formalization of Marx’s general law of accumulation. This model has weaknesses, but the presence of endogenous cycles is in line with empirical studies which find evidence of local instability.

This concluding chapter offers a few remarks on empirical work and its relation to economic theory. Econometricians have been gaining access to more and better data, and significant progress has been made in the techniques and procedures that are used to extract information from the data. But the data still don’t speak for themselves. Recent contributions on regional fiscal multipliers and state-level inflation in the US are used to illustrate the need for theory to extrapolate from regressions to generalizable implications. Stripped-down New Keynesian Phillips curves, for instance, are used to interpret inflation data, despite the abject empirical failure of these Phillips curves. The imaginative use of new data and careful attempts to identify causal effects clearly can be immensely helpful. But dogmatic adherence to an untenable, scholastically motived theoretical framework can lead to a curious combination of sophisticated econometrics and a willingness to engage in ad hoc handwaving in the interpretation of the data. The ambition in this book is to sketch an alternative framework with behavioral and structural assumptions that are evidence based.

Fairness concerns play an important role in wage formation, with unfair treatment affecting morale and labor productivity. These forces can lead to unemployment and affect the pattern of relative wages and employment. But behavioral evidence shows that perceptions of fairness can be influenced by purely nominal changes in wage rates, which implies that unemployment will be inversely related to inflation, also in the long run; there is no natural rate. Moreover, social norms only persist if they are affirmed by actual behavior while, conversely, outcomes that are well-established will come to be expected and accepted as fair. These endogenous changes in fairness norms can generate employment hysteresis, with the precise implications for economic policy and the relation between unemployment and inflation depending on the details of the model. Models of inflation need modification in developing economies with large amounts of underemployment. Drawing on a structuralist tradition and insights from behavioral economics, the chapter presents a stylized model in which inflation is determined by distributional conflict and cross-sectoral interactions between demand and supply side forces.

The ownership of firms is mediated by financial assets in a corporate economy. The Modigliani-Miller theorem and most macroeconomic models tend to play down the relevance of this financial mediation. Behavioral and institutional factors make it unlikely, however, that households ‘pierce the corporate veil’ and adjust their portfolios to offset changes in firms’ financial decisions and, as shown by (an extended version of) Kaldor’s ‘neo-Pasinetti theorem’, households as a group cannot declare their own dividends by selling shares if corporate retained earnings are raised. As a result, aggregate saving rates will be increasing in the share of profits, the retention rate out of profits, and the share of stock buybacks. Changes in household portfolio behavior also affect aggregate saving, but the direction depends on firms’ retention and buyback policies. In this setting, a tightening of credit constraints can lead to a protracted decline in aggregate demand, feedback effects from capital gains to portfolio choice can fuel stock market bubbles, and Tobin’s q becomes a poor indicator of the return to additional capital.

The baseline model in this chapter combines destabilizing Keynesian-Harrodian mechanisms with feedback effects from the labor market to firms’ output and investment decisions. The aggregation of micro-level output and investment decisions is analyzed explicitly and, following a Keynes-Marshall tradition, prices and profit shares adjust to clear the goods market. The flex-price assumption is empirically motivated: evidence shows that goods prices are much more flexible than commonly believed. The model produces limit cycles around a locally unstable steady growth path. A variant of the model that may fit parts of the service sector takes output as perfectly flexible, while the real wage is fixed. The reduced-form relations and dynamic patterns of this variant are virtually identical to those of the flex-price model, suggesting that these reduced-form equations may be a good starting point for analyzing business cycles in the aggregate economy. Reinforcing this conclusion, simulations of an extended version that uses empirically plausible parameters and includes fiscal and monetary policy shows a remarkable correspondence to cyclical patterns for the US.

The Lucas critique is valid, but the Lucas solution developed by mainstream macroeconomics represents an abject failure. Heroic aggregation assumptions are embodied in the creation of a representative agent: even if individual preferences could be taken as well-defined, exogenous, and stable over time, the celebrated Sonnenschein-Debreu-Mantel results show that microeconomic rationality imposes only very weak constraints on the properties of aggregate excess demand functions. Using simple examples, this chapter illustrates how restrictive assumptions are needed to ensure the existence of a representative agent and discusses the implausibility of these conditions being approximately satisfied. The chapter also questions the utility function of the representative agent as the basis for welfare analysis. Although supposedly ‘natural’ and ‘objective’, this approach imparts a systematic bias against the poor and in favor of the rich. As a corollary, moreover, changes in the distribution of income renders the representative agent’s utility function unstable; the Lucas solution is subject to a Lucas critique.

The existence of a ‘natural rate of unemployment’ informs much of economic policy. But attempts to explain large cross-country differences and time-series changes in unemployment are unconvincing, and strong prior beliefs are required to interpret the evidence as supportive of the natural rate hypothesis. And the natural-rate hypothesis is also fragile on theoretical grounds. Stylized policy games between monetary policymakers and the private sector have suggested that discretionary policy regimes suffer from inflationary bias and that policy makers cannot achieve sustained reductions of unemployment below the natural rate. The results are radically different in the presence of labor unions, even with perfect foresight and well-defined preferences over inflation and output: a central bank that is completely indifferent to the level of inflation may obtain outcomes with high employment and zero inflation, while inflation-averse central banks generate stagflation with positive inflation and low employment. The reduced-form relation between unemployment and inflation following a shift in the central bank’s objective function can be negative or positive, depending on details.

This chapter offers an introductory discussion of the state of macroeconomics. It presents a brief outline of the historical background to traditional Keynesian macroeconomics and its replacement by the current orthodoxy based on microeconomic foundations and intertemporal optimization, with DSGE models at the center. This scholastically motivated paradigm has been a costly detour. A promising alternative structuralist and behavioral approach can draw on behavioral economics as well as rich traditions that have been marginalized within the profession.