According to phelps, since unemployment tends toward its natural rate the inflation rate:

3.4 The Current Debate Over Keynesian Economics

According to the monetarist-cum-rational expectations schools, Keynesian economics failed the predictive test: it led to inflation, still worse ‘stagflation.’ In the 1970s, Keynesian policies were attacked for ignoring the existence of a ‘natural’ rate of unemployment, and (by the Virginia or Public Choice school) for assuming that politicians wanted to maximize the collective social welfare, rather than their own individual utilities. Taken together, these two attacks offered a forceful argument against the use of discretionary fiscal and monetary policy to balance economies.

The use of models of economies with nominal rigidities is still general, but whether these rigidities are to be taken as given is much more questioned than in 1950s and 1960s. The dominant ‘supply side revolution’ of the 1980s chiefly was concerned to dissolve rigidities seen as institutional by deregulating labor markets. Against this, the ‘new Keynesians’ explained how sticky prices are rational because of transactions and information costs, and how shocks to demand can destroy both physical and human capital. These explanations seemed both to strengthen and weaken the case for Keynesian macroeconomic policy. On the one hand, they gave renewed intellectual respectability to stabilization policy. On the other hand, by explicitly introducing inflation into their analyses, they conceded the existence of a rate of unemployment (the so-called NAIRU or Non-Inflation Accelerating Rate of Unemployment) below which unemployment could not be pushed by manipulating demand. Finally, against the mainstream profession's use of Bayesian statistics and decision theory to model agents’ behavior, a minority school of ‘Post-Keynesians’ continues to assert the fundamental nature of Keynes's attack on the rationality axiom.

An interim judgment on the Keynesian Revolution would be that the main body of classical economics was too well-entrenched to be overthrown by the frontal assault he mounted. The notion that uncertainty was at the heart, rather than at the financial margins, of economic processes, proved too subversive of the science economics claims to be, to be acceptable. Most economists would say today of Keynes what Marshall said of Jevons: ‘His success was aided even by his faults … he led many to think he was correcting great errors; whereas he was really only adding important explanations.’ Whether this will be the final verdict is still questionable.

3.5 The long-run Phillips curve

In the baseline model, without capital, we saw that DM output is decreasing in anticipated inflation, while CM output is independent of anticipated inflation. It is not true that CM output is independent of anticipated inflation in the model with capital in the previous section, because we assumed K enters c(x, K). If this is not the case, and cK(x, K) = 0, then the last term in Eq. (26) vanishes, K drops out of Eq. (25), and the system dichotomizes: we can independently solve Eq. (25) for the DM allocation x and the other three equations for the CM allocation (X, K, H). In this dichotomous case, monetary policy affects x but not (X, K, H). This is why we assumed K enters c(x, K). In this section, without capital, we break the dichotomy using nonseparable utility. In fact, here we take the Phillips curve literally, and model the relation between inflation and unemployment. To make this precise, first, we introduce another friction to generate unemployment in the CM, and second, we re-cast the DM as a pure exchange market, so that unemployment is determined exclusively in the CM.

To give some background, a principle explicated in Friedman (1968) is that, while there may exist a Phillips curve trade-off between inflation and unemployment in the short run, there is no trade-off in the long run. The natural rate of unemployment is defined as “the level that would be ground out by the Walrasian system of general equilibrium equations, provided there is embedded in them the actual structural characteristics of the labor and product markets” (although, as Lucas, 1980 noted, Friedman was “not able to put such a system down on paper”). Friedman (1968) said monetary policy cannot engineer deviations from the natural rate in the long run. However, he tempered this view in Friedman (1977) where he said

There is a natural rate of unemployment at any time determined by real factors. This natural rate will tend to be attained when expectations are on average realized. The same real situation is consistent with any absolute level of prices or of price change, provided allowance is made for the effect of price change on the real cost of holding money balances.

Here we take this real balance effect seriously.

Of the various ways to model unemployment, in this presentation we adopt the indivisible labor model of Rogerson (1988).24 This has a nice bonus feature: we do not need quasi-linearity, because in indivisible-labor models agents act as if utility were quasi-linear. To make the point, we revert to the case where X is produced one-for-one with H, but now H ∈ {0, 1} for each individual. Also, as we said, to derive cleaner results we use a version where there is no production in the DM. Instead, agents have an endowment x―, and gains from trade arise due to preference shocks. Thus, DM utility is vj(x, X, H) where j is a shock realized after (X, H) is chosen in the CM. Suppose j = b or s with equal probability, where ∂vb(·)/∂x > ∂vs(·)/∂x, and then in the DM everyone that draws b is matched with someone that draws s. The indices b and s indicate which agents will be buyers and sellers in matches, for obvious reasons. We also assume here that there is discounting between one DM and the next CM, but not between the CM and DM, but this is not important. What is interesting is nonseparability in vj(x, X, H).

As in any indivisible labor model, agents choose a lottery (ℓ, X1, X0, mˆ1, mˆ0) in the CM where ℓ is the probability of working H = 1, while XH and mˆH are CM purchases of goods and cash conditional on H (if one does not like lotteries, the equilibrium can also be supported using pure Arrow-Debreu contingent commodity markets, as in Shell & Wright 1993). There is no direct utility generated in the CM; utility is generated by combining (X, H) with x in the DM. Hence,

(28)W(m)=maxℓ,X 1,X0,mˆ1,mˆ0{ℓV(mˆ1,X1,1)+(1−ℓ)V(mˆ0,X0,0)}st0≤ϕm−ℓϕmˆ1−(1−ℓ)ϕmˆ0+wℓ−T−ℓX1−(1−ℓ)X0.

As is well known, X and mˆ depend on H, in general, but if V is separable between X and H then X0 = X1, and if V is separable between mˆ and H then mˆ1 = mˆ0. But the function V is endogenous. This is another argument for making the role of money explicit, instead of, say, simply sticking it in the utility function: one cannot simply assume V is separable (or homothetic or whatever), one has to derive its properties, and this imposes discipline on both theory and quantitative work.25

Letting λ be the Lagrangian multiplier for the budget constraint, FOC for an interior solution are

(29)0=V2(mˆH,XH,H)−λ,forH=0,1

(30)0=V1(mˆH,XH,H)−λϕ,forH=0,1

(31)0=V(mˆ0,X0,0)−V(mˆ1,X1,1)+λ(X1−X0−1+ϕmˆ1−ϕmˆ0)

(32)0=ℓ−ℓX1−(1−ℓ)X0+ϕ[m+γM−ℓmˆ1−(1−ℓ)mˆ0].

One can guarantee ℓ ∈ (0, 1), and show the FOC characterize the unique solution, even though the objective function is not generally quasi-concave (Rocheteau et al., 2007). Given V(·), Eqs. (29)–(31) constitute five equations that can be solved under weak regularity conditions for (X1, X0, mˆ1, mˆ0, λ), independent of ℓ and m. Then Eq. (32) can be solved for individual labor supply as a function of money holdings at the start of the period, ℓ = ℓ(m). Notice mˆH may depend on H, but not m, and hence we get at most a two-point distribution in the DM. Also, W(m) is again linear, with W′(m) = λϕ. This is what we meant earlier when we said that agents act as if they had quasi-linear preferences in the model with indivisible labor and lotteries.

In DM meetings, for simplicity we assume take-it-or-leave-it offers by the buyer (θ = 1). Also, although it is important to allow buyers' preferences to be nonseparable, we do not need this for sellers, so we make their preferences separable. Then as in the baseline model, the DM terms of trade do not depend on anything in a meeting except the buyer's m: in equilibrium, he pays d = m, and chooses the x that makes the seller just willing to accept, independent of the seller's (X, H). In general, buyers in the DM who were employed or unemployed in the CM get a different x since they have different m. In any case, we can use the methods discussed above to describe V(·), differentiate it, and insert the results into Eqs. (29)–(31) to get conditions determining (x1, x0, X1, X0, λ). From this we can compute aggregate employment ℓ―=ℓ( M).

It is now routine to see how endogenous variables depend on policy. First, it is easy to check ∂x/∂i < 0, since as in any such model the first-order effect of inflation is to reduce DM trade. A calculation then implies that the effect on unemployment depends on the cross derivatives of buyers' utility function as follows:

if vb(x, X, H) is separable between (X, H) and x, then ∂ℓ―/∂i=0

if vb(x, X, H) is separable between (x, X) and H, then ∂ℓ―/∂i>0 iff vXxb<0

if vb(x, X, H) is separable between (x, H) and X, then ∂ℓ―/ ∂i>0 iff vxHb<0

The economic intuition is simple. Consider case 2. Since inflation reduces x, if x and X are complements then it also reduces X, and hence reduces the ℓ― used to produce X; but if x and X are substitutes then inflation increases X and ℓ―. In other words, when x and X are substitutes, inflation causes agents to move from DM to CM goods, increasing CM production and reducing unemployment. A similar intuition applies in Case 3, depending on whether x is a complement or substitute for leisure. In either case, we can get a downward-sloping Phillips curve under simple and natural conditions, without any complications like imperfect information or nominal rigidities. This relation is exploitable by policy makers in the long run: given the right cross derivatives, it is indeed feasible to achieve permanently lower unemployment by running a higher anticipated inflation, as Keynesians used to (still?) think. But it is not optimal: it is easy to check that the efficient policy is still Friedman's prescription, i = 0.

3.3 Measurement issues and the output gap

One practical issue that affects the implementation of monetary policy is the measurement of variables of interest such as the inflation rate and the output gap (Orphanides, 2001). Many macroeconomic data series such as GDP and price deflators are subject to measurement errors and revisions. In addition, both the equilibrium real interest rate and the output gap are unobserved variables. Potential errors in measuring the equilibrium real interest rate and the output gap result from estimating latent variables as well as uncertainty regarding the processes determining them (Edge, Laubach, & Williams, 2010; Laubach &Williams 2003; Orphanides & van Norden, 2002). Similar problems plague estimation of related metrics such as the unemployment gap (defined to be the difference between the unemployment rate and the natural rate of unemployment) and the capacity utilization gap. Arguably, the late 1960s and1970s were a period when errors in measuring the output and unemployment gap were particularly severe, but difficulties in measuring gaps extend into the present day (Orphanides, 2002; Orphanides & Williams, 2010).

A number of papers have examined the implications of errors in the measurement of the output (or unemployment) gap for monetary policy rules, starting with Orphanides (1998), Smets (1999), Orphanides et al. (2000), McCallum (2001), and Rudebusch (2001). A general finding in this literature is that the optimal coefficient on the output gap in the policy rule declines in the presence of errors in measuring the output gap. The logic behind this result is straightforward. The response to the mismeasured output gap adds unwanted noise to the setting of policy that can be reduced by lowering the coefficient on the gap in the rule. The optimal response to inflation may rise or fall depending on the model and the weights in the objective function.

In addition to the problem of measurement of the output gap, the equilibrium real interest rate is not a known quantity and may vary over time (Laubach & Williams, 2003). Orphanides and Williams (2002) examined the combined problem of unobservable unemployment gap and equilibrium real interest rate. In their model, the unemployment gap is the measure of economic activity in both the objective function and the policy rule. They consider a more generalized policy rule of the form:

(5)it=Et{(1−ρ)(rˆt⁎+ πt)+ρit−1+α(πt−π⁎)+γuˆt+δΔut}.

where rˆt⁎ (uˆt) denotes the central bank's real-time estimate of the equilibrium real interest rate (unemployment gap) in period t, and Δut denotes the first-difference of the unemployment rate.

The presence of mismeasurement of the natural rate of interest and the natural rate of unemployment tends to move the optimal policy toward greater inertia. Figure 3 shows the optimal coefficients of this policy rule for a particular specification of the central bank loss as the degree of variability in the equilibrium real interest rate and the natural rate of unemployment rises. The case where these variables are constant and known by the central bank is indicated by the value of zero on the horizontal axis. In that case, the optimal policy is characterized by a moderate degree of policy inertia. The case of a moderate degree of variability of these latent variables, consistent with the lower end of the range of estimates of variability, is indicated by the value of 1 on the horizontal axis. Values of 2 and above correspond to cases where these latent variables are subject to more sizable fluctuations, consistent with the upper end of estimates of their variability. In these cases, the central bank's estimates of the equilibrium real interest rate and the natural rate of unemployment are imprecise, and the optimal value of ρ rises to near unity. In such cases, the equilibrium real interest rate, which is multiplied by (1 − ρ) in the policy rule, plays virtually no role in the setting of policy.

The combination of these two types of mismeasurement also implies that the optimal policy rule responds only modestly to the perceived unemployment gap, but relatively strongly to the change in the unemployment rate. This is shown in the lower two panels of Figure 3. These policy rules that respond more to the change in the unemployment rate use the fact that that the direction of the change in the unemployment rate is generally less subject to mismeasurement than the absolute level of the gap in the model simulations. If these measurement problems are sufficiently severe, it may be optimal to entirely replace the response to the output gap with a response to the change in the gap. In the case where the value of ρ is unity, such a rule is closely related to a rule that targets the price level, as can be seen by integrating Eq. (5) in terms of levels. See McCallum (2001), Rudebusch (2002), and Orphanides and Williams (2007b) for analysis of the relative merits of gaps and first differences of gaps in policy rules.

3.2 The Standard AD-AS Model

The AD-AS model is a standard tool in macroeconomic analysis. AD represents the aggregate demand, whereas AS stays for aggregate supply. This is explained to students when the macroeconomic theory is introduced, often preceded by the IS-LM model (with fixed prices). Indeed, in an introductory course on macroeconomics, when organized starting from the analysis of the short-run to proceed with the medium- and then the long-run analysis of economic growth, one firstly is taught the IS-LM model, and then the AD curve can be constructed on this basis, corresponding to an IS-LM model with flexible prices. Based on the Phillips curve, which is on the inverse relationship between (wage) inflation and unemployment, typically assuming a constant mark-up, the AS curve is introduced, and the AD-AS model can be used for the macroeconomic analysis of the medium run.

In its simplest form, the AD-AS model is represented as the interaction between two linear curves, though nonlinear relationships are quite commonly employed. In general, however, we have a downward sloping AD and an upward sloping AS.1 Depending on expectations, policy makers can (or cannot) exploit the trade-off between unemployment and inflation because of the different time intervals implied by the adjustment toward the equilibrium. In the extreme (but included in the textbook AD-AS model) case of “rational expectations,” when the agents know the model and are able to anticipate the decisions of policy makers, the AS is vertical at the potential level of output (as if the adjustment was instantaneous), and the AD only determines the price level. The unemployment rate that corresponds to the equilibrium output is the NAIRU (Non-Accelerating Inflation Rate of Unemployment). According to this model, only movements of the AS influence the macroeconomic equilibrium in the long run, whereas a monetary or a fiscal expansion just leads to more inflation, thus suggesting that “structural reforms” are needed to reduce unemployment (i.e., the NAIRU), whereas the Keynesian tools of macroeconomic policy are ineffective (or can have an impact that is limited to the short run). As for stabilization, in such a “natural” equilibrium setting, monetary policy is considered as the primary tool to promote macroeconomic stability and, in general, a growth-enhancing environment [1]. However, neither monetary policy nor fiscal policy aimed at managing the aggregate demand is taught to be useful in affecting the “natural” macroeconomic equilibrium, for example, the NAIRU.

However, it is unlikely that rational expectations are a good approximation of real agents' behavior; thus an upward sloping AS curve seems to be a better representation of the macroeconomic reality, and the sustain of aggregate demand through fiscal or monetary policy can be effective, at least along the adjustment process (thus depending on how long the system takes to go back to the equilibrium). In other words, people are able to adapt, at least partially, to policy changes, but not instantaneously. Adaptive expectations of some sort should be assumed to describe a relatively slow and possibly incomplete adjustment.

Moreover, in many cases, macroeconomic analysis does not involve distributive issues. In fact, considering the distribution of income (and of wealth) can lead to significant results. For instance, even in the AD-AS model, if we assume that agents are characterized by heterogeneous propensities to consume and in particular that the propensity to consume out of income (and wealth) is decreasing with the income (wealth) level, then we simply obtain a Keynesian multiplier that depends on the functional distribution between wages and profits through the mark-up, say z. A simple case is that of an AD-AS model with two social classes, workers and capitalists, in which capitalists have a relatively low propensity to consume, say cK, whereas workers have a relatively high propensity to consume, say cW. In the simplifying case cK=0, that is, with unitary propensity to save for capitalists, the multiplier depends positively on cW and negatively on z, that is, a redistribution from wages to profits reduces the equilibrium output. Accordingly, an increase of inequality can result in lesser consumption and then a decrease of the aggregate demand and of the equilibrium level of output (but for a vertical AS). However, this slight modification seems to have far reaching consequences and a Post-Keynesian flavor, which is not in line with the conventional view neither at the level of textbook macroeconomics or at that of the New Keynesian Dynamic Stochastic General Equilibrium (NK-DSGE) model and its last extensions aimed at including a variety of financial frictions.2

Another factor, which is often leaved out by introductory macroeconomic analysis, is finance, for instance, the role of the “risk premium” and the mechanisms of the “financial accelerator” ([2], which is instead a well-known piece of advanced macroeconomics),3 at least before the new edition of the classical book of [3]. Indeed, [3] has introduced an additional financial channel in the IS curve, where the investment is influenced by the real interest rate (e.g., the nominal interest rate minus expected inflation) plus a risk premium and by income: in particular, an increase of the “risk premium,” for instance due to a negative financial shock, depresses investment, thus leading to a fall in aggregate demand (the IS curve moves leftward) and a decrease of aggregate income. However, finance can play an important role also in affecting the supply side since the availability of bank credit and other forms of financing allow firms to finance production. In particular, firms can be rationed on the credit market due to the lack of collateral, and this effect can depend on different financial conditions faced by heterogeneous firms. This is not considered by the textbook AD-AS model, in which the aggregate supply is based on the labor market and the trade-off between inflation and unemployment (depending on the type of expectations). Credit rationing and, in general, the deterioration of financial conditions, instead, can have an impact on production since a “credit crunch” may reduce the financing of firms' production with a subsequent increase of unemployment and a fall of aggregate income.

In what follows, we explain how an agent-based model is able to highlight the role of heterogeneity and interaction in macroeconomic dynamics by reconstructing both the AD and AS curves from the bottom-up, that is, by simulating how each agent reacts to variations of the demand and supply conditions and what it emerges at the aggregate level.

12.5.6 Models arguing: the Great Inflation debate

In the 1990s and early 2000s there was an active debate about the so-called “Great Inflation” in the US, namely the rise and fall of inflation from the 1960s to the 1980s. While this might not qualify as quite historical (since macro-history seems to start with quarterly NIPA accounts in 1946), it involved some of the most prominent macroeconomists, yet its focus was a single sequence of events and its goal is to understand the process driving them.

Based on a broad narrative (mixing analyses of speeches, writings of policymakers and economists), De Long (1996) argued that the Great Inflation was due to the very high weight assigned to unemployment by monetary policy (“The memory left by the Depression predisposed the left and center to think that any unemployment was too much”). The measures to fight inflation starting 1979 are thus interpreted as a shift in beliefs.

Sargent (1999) took up De Long's idea, deepened the narrative and built a model to account for the change in policymakers beliefs and their impact on economic outcomes. Orphanides (2002) brought “archival material” into play to argue that policy actions can be explained by misinterpretations of the current data. Romer and Romer (2002) used extensively the archives of the Fed to provide evidence of the change in policymakers' beliefs over time. Based on records of the discussions at the Fed, they emphasized that it was not so much a too high weight on unemployment that led the Fed to run inflationary policies in the 1970s (because they adopted early the natural rate of unemployment theory) but an underestimation of the natural rate of unemployment.

In his discussion of Romer and Romer (2002), Sargent (2002) acknowledged the usefulness of their work but criticized the main conclusion based on the narrative approach because it lacked the kind of quantitative analysis that could take into account other parameters, especially the variety of shocks hitting the economy. That being said, Sargent included himself in the “Berkeley” view, together with the Romers and De Long, for emphasizing the importance of the changes in beliefs as an explanation of the Great Inflation. Cogley and Sargent (2005) responded to Romer and Romer (2002) by incorporating model uncertainty. They built a model where the central bank chooses (with uncertainty) at each period between several models of the world/Phillips curve and NAIRU.

Primiceri (2006) modeled and estimated directly the changes in beliefs of policymakers about the NAIRU over time. Sims and Zha (2006) criticized the literature by showing that, when accounting for a wide range of shocks, they did not estimate drastic regime shift in the coefficients of the monetary policy rule over time. In turn, this literature continued to influence economic historians working with a qualitative or narrative approach. For example, Meltzer (2010) in his history of the Fed, Bordo and Eichengreen (2013), Weise (2012) emphasized that, besides internal beliefs and NAIRU estimates of the Fed, the pressures from the Treasury to keep interest rates low were of key importance.

The debate died out with the Great Recession and was left somewhat unresolved. Three views remained: a story based on luck (Sims and Zha, 2006); exogenous changes in policy that moved from a region of indeterminacy to determinacy (Lubik and Schorfheide, 2004); and the learning story (also Orphanides and Williams 2013), either about parameters of a model as in Primiceri (2006) or between different models as in Cogley and Sargent (2005).

The question throughout this literature is a simple, historical one: “Why did X happen?” At first sight this history for history's sake, arguing over a single data point without any attempt at establishing a causal relationship. But the debate was fruitful because models (more or less fleshed out, more or less confronted with observations) were arguing with each other and fighting to claim the data point. Whether or not a winner came out is of little importance: our knowledge and understanding of the (recent) past is deeper, our theoretical insights are sharper. Everyone gained.

b How to eliminate the scourge of unemployment: jobs now at a living wage

It is amazing no one has thought of this before. A dozen years after the GFC began, we have still got up to 15 million people who want jobs but cannot find them. Of course, that is far more than the official unemployment numbers—which do not count anyone who worked just an hour or so, or who gave up looking altogether.

I wonder how on earth we can find a solution to joblessness, or to low pay? It is all so complicated. How can we stroke the business class in just the right manner to get them to create a job or two? How can we prevent our corporations from taking jobs abroad? How can we add a few more jobs without stoking the fires of inflation?

Should we slash government regulations to raise the spirits of our business undertakers? Maybe we should just eliminate minimum wages so that they can afford “expensive” American labor? Then we could compete with Viet Nam's low-wage labor.

Or slash taxes to boost the supply side?

Or maybe we should just throw in the towel and admit that we will never solve the problem of unemployment? Just toss more welfare handouts at the jobless? Expand the dependent classes to include more of the able-bodied. Admit that our Captains of Industry, as well as our Public Stewards, have failed us. That we have created a dysfunctional social system that cannot provide jobs to those who want to work.

As Hyman Minsky put it 6 decades ago, providing welfare rather than jobs is “a conservative rebuttal to an ancient challenge of the radicals, that capitalism necessarily generates ‘poverty in the midst of plenty’”.

Rather than paying people not to work, Minsky proclaimed we must pursue an alternative: “We have to reverse the thrust of the policy of the past 40 years and move towards a system in which labor force attachment is encouraged. But to do that we must make jobs available; any policy strategy which does not take job creation as its first and primary objective is but a continuation of the impoverishing strategy of the past decade.”(Minsky, 1975).

Here is Minsky's idea: why not create jobs with decent pay? Now, why didn’t anyone ever think of that before?

Put people to work doing socially useful things. Take workers as they are, design jobs that they are able to do. Offer a high enough wage with good, supportive working conditions so that no one would take the demeaning and low-paying jobs that the private sector creates. If the private sector wants to compete, it will have to pay more and provide more interesting and fulfilling work.

We must change the system, not the people.

Moreover, a JG program is not just jobs creation, it will act as an automatic stabilizer as employment in the program grows in recession and shrinks in economic expansion, counteracting private sector employment fluctuations. Further, a JG program with a uniform basic wage will also help to promote economic and price stability.

The mainstream view has long been that “full employment” and “price stability” are incompatible goals. They claim that you must have substantial unemployment to keep prices in check. You can call it NAIRU, you can call it the “natural rate” or you can call it the “reserve army of the unemployed”.20 It is a view shared by virtually all economists outside the MMT camp. According to all of them, the unemployed serve as a price anchor; the suffering of the unemployed does the duty of keeping the currency scarce and valuable. Unemployment is the “cost” to achieve the “benefit” of low inflation.

We reject that view as unnecessarily defeatist.

However. And here is the Big However. We do agree with the mainstream that the economy needs a price anchor, or otherwise pursuit of true, full employment probably would, at least much of the time, cause inflation. So, we, too, want a price anchor. We object to the (usually implicit) claim of just about everyone outside the MMT camp that unemployment is the only possible price anchor. Other economists do not have the imagination to come up with any alternative price anchor for a fiat currency.

In our view, that is wrong.

Here is Warren Mosler's response, in what is almost a Haiku in its simplicity:

It comes down to this:

With ‘state currency’

There necessarily is,

Always has been,

Always will be,

A buffer stock policy.

Call that the MMT insight if you wish.

So it comes down to ‘pick one’-

1. Gold

2. Foreign Exchange

3. Unemployment

4. Employed/JG/ELR

5. Wheat

Whatever!

I pick employed/JG/ELR

As it works best as a buffer stock based on any/all criteria for buffer stock.

So yes, it's an option.

You are free to pick one of the others.

So …. You can have full employment but you have got to choose a price anchor. Some want a commodity buffer stock (usually gold). Others want to tie the domestic currency to a foreign currency. Most want unemployment. By contrast, MMTers follow Warren Mosler in choosing an employed buffer stock—the JG (also called the Employer of Last Resort—ELR—program, Tymoigne and Wray, 2013).

We can analyze the JG as a program that uses labor as its buffer stock. Labor goes into all output. It is a domestic resource that can be found in every nation—and it is virtually always in excess, that is, not fully employed. Providing jobs to those who want to work but cannot otherwise find paid work is consistent with an internationally recognized human right. It has many individual, familial, and social benefits that go far beyond earning an income. Keeping the labor buffer stock employed maintains that buffer stock in good working condition. Enforcing idleness on those who want to work is like letting the rats invade your buffer stock of corn—that reduces the value and effectiveness of the buffer stock in controlling the price or wage. A reserve army of the employed is much better than a reserve army of the unemployed as the unemployed are perceived by employers to deteriorate in quality at a rapid pace. In consequence, they prefer to bid already employed workers away from other employers over recruiting the unemployed.

The JG program is quite explicitly a targeted spending program in which government spending is directed precisely to those who want to work but cannot find a job. This places no direct pressure on wages and prices because the workers in the program were part of the “surplus” or “redundant” labor force. And even when they are employed in the JG program, they are still available for private employers (at a small mark-up over the JG program wage—which becomes the effective minimum wage).

For that reason, employing workers in the JG program is no more inflationary than leaving them unemployed. Indeed, the JG lowers recruiting and hiring costs as employers would have an employed pool of workers demonstrating readiness and willingness to work, which should reduce inflation pressures. The JG program will have a work history for each of its participants that can be used for placement purposes. It will be relatively easy for an employer to search through the JG pool of workers to recruit those most suitable to the openings.

Turning to effects on aggregate demand, many critics worry that if, say, 10 million people obtain jobs in the JG and thereby increase their incomes above their pre-employment levels, consumption would increase and drive up inflation. This seems to be a major concern of JG critics. By logical extension, they would also worry about a private sector–led expansion that created minimum wage jobs in, say, the fast food sector. In other words, they should be opposed to any increase of employment on the argument that the employed will spend more. We find such a position to be overly defeatist—a “let the poor eat cake” response to unemployment and poverty.

This criticism is also often combined with the claim that workers in the JG would just “dig holes”, adding nothing to national output. Again, we see that as overly pessimistic—since a jobs program can be designed to produce desirable output, as the New Deal's jobs programs did.

Note also that if the newly employed private workers produced goods for export, the extra wages would increase consumption without producing more domestic goods to absorb the demand. So this would be at least as inflationary as employing workers in a JG. And yet virtually all economists would celebrate increased employment in the nation's export sectors, while most oppose employing more workers in the government sector!

Let us imagine that the JG program is extremely successful at creating jobs and income, so much so that the economy moves from slack to full employment of all productive capacity, resulting in rising prices. The presumed problem is that while JG workers get wages (and thus consume) they do not contribute any production that is sold (hence, does not absorb wages). The “excess” wages from newly employed workers induces spending to rise and beyond some point the producers raise prices rather than increase output. (As noted, increasing employment in the export sector would create the same problem.)

What could government do in that case? It would have at its disposal the usual macroeconomic policy tools: raise taxes, lower government spending on programs other than the JG, and tighten monetary policy. It could also do what the United States did in WWII when the economy operated beyond full employment: wage and price controls, rationing of some key resources, patriot saving (selling war bonds to encourage saving and reduce consumption), and postponed consumption (workers were offered good pensions instead of wage increases). (We will not get into an argument here about the relative effectiveness of these—but they kept inflation manageable during the war when government spending absorbed half the nation's output and a significant portion of the labor force was in the armed forces.)

Indeed, this is what the government would do in the absence of the JG if the private sector achieved full employment through creation of 10 million new minimum wage jobs in the private sector. The only difference is that the government would not be able to fight inflation by increasing unemployment—because the macro policies used to fight inflation would dampen demand but any worker losing a job could turn to the JG program for work.

What this means is that with a JG in place, the inflation-fighting adjustments to spending will occur among the employed rather than by causing unemployment and poverty. In other words, the costs of fighting inflation can be made to be borne at higher income levels. We are surprised that our critics appear to prefer to use unemployment and poverty to fight inflation, which forces the least able to bear more of the costs. With the JG in place, these policies put at least some of the burden of fighting inflation on those who are well-off.

Our position is similar to Keynes's: “No one has a legitimate vested interest in being able to buy at prices which are only low because output is low.” (Keynes, 1964, p. 318) So while those who like an unemployed buffer stock (“reserve army of unemployed”) argue against creating jobs on the argument that those with jobs would have more income, and this could cause inflation, that is not a morally defensible position.

It must be recognized that increasing the number of private sector workers in the fast food industry (or in the export sector) will cause the same sort of inflation, raising prices in the same sectors that consumption by new workers in the JG program would affect. It does no good to argue that hamburger flippers are “productive” (they flip burgers) while JG workers are not (they provide, for example, public services to the aged), because the “semi-inflation” will occur in all sectors where increased spending faces anything less than what economists call perfect “output elasticity” (that is, where output can rise on a one-for-one basis as demand rises, so that prices do not rise at all).

Hence, if our critics were consistent, they would always fight against job creation if any sectors that would experience increased sales to workers had less than perfect output elasticity. Their argument against the JG is a red herring. An employed buffer stock is more effective at constraining wages and prices than is an unemployed reserve army because employment keeps labor in better shape.

Employment also has substantial benefits for the individual, the family, and society as a whole. Unemployment has few (questionable at best) benefits and lots of horrendous costs to individuals, families, and societies, including divorce, deteriorating physical and mental health, abuse of children and spouses, gang activity, and crime. Keeping humans employed avoids tremendous social and individual costs.

4.1.2 The second attempt: The instrumentalist foundations of Friedman’s Phillips curves

The Phillips curve as a framework has a unique position in Friedman’s oeuvre. He devoted his famous presidential address (Friedman, 1968) to disentangling the mechanisms underlying the phenomenal Phillips curve, though there an explicit model played a minor role only. A few years later the Phillips curve got into the limelight: in his Nobel lecture Friedman discussed his accomplishments in the context of the Phillips curve (Friedman, 1977). The Phillips curve provided the analytical framework by which Friedman could combine his views on equilibrium (the natural rate of unemployment) with his thoughts on the limitations to monetary policy. If Friedman is taken as a single-idea economist (De Vroey, 2016, p. 66), the Phillips curve framework is no more than a possible articulation of Friedman’s oft-voiced ideas. However, there is more to it. The Phillips curve provided Friedman with an efficient instrument for combating Keynesian economics24 (De Vroey, 2001), even if the Keynesian missteps he associated with it appear to be his fantasies (Forder, 2010). In the present methodological analysis Friedman’s Phillips curves are important as they aid in highlighting the problematic relationship between the faultily designed microfoundations and macro-level causal understanding.25

Friedman’s ideas on the Phillips curve were born in the debates triggered by Samuelson and Solow’s 1960 paper (Samuelson & Solow, 1960) on the guidelines for monetary policy (Taylor, 2007, p. 136). This was the study in which the authors some 2 years after the emergence of the Phillips curve used it as the analytical framework of economic policy and inflation (Schwarzer, 2013). According to the traditional narrative put forward in Friedman’s Nobel lecture, Friedman and Phelps called into question the stability of the trade-off between the rate of inflation and the rate of unemployment and hence the Phillips curve. As expectations matter, by keeping inflation at a high rate it is impossible to stabilize unemployment under its natural level. The Phillips curve always shifts. Forder (2014) disapproves of Friedman’s account, especially the suggestion that it is Friedman (and Phelps) who called attention to the role of expectations. When Friedman brought in these thoughts, they had already been available in the literature for a long time. As a matter of fact, Samuelson and Solow themselves were also sceptical about the stability of the curve (Hoover, 2015). At the same time the framework had crucial importance specifically to Friedman. He devoted his Nobel lecture to some theoretical problems raised by the curve whilst the economic literature of 1960–70s did not place much emphasis upon it. In those days some attention was admittedly paid to the relationship between wage and price inflation on the one hand and unemployment on the other, but Phillips’s contribution was only one in the array.

As it was argued in Section 4.1.1, Friedman in his F53 by playing down the importance of the real properties of entities placed the sole emphasis upon saving the phenomena; hence he subscribed to instrumentalism.26 Friedman’s Phillips curves provide further striking examples for neglecting the purpose of causal understanding. The role ideal types play in causal understanding properly highlights how Friedman designed his entity-level assumptions in his Phillips curves in order to derive the macro-level conclusions he wished in advance. James Forder (2016) also calls attention to some confusions hidden in Friedman’s AEA address. Forder places special emphasis upon the unclear phrasing through which Friedman portrayed the dissimilar flexibilities of prices and wages. In Forder’s reading, that Friedman (1968) conceives prices as rising faster than wages may mean both that price and wage inflations start at different moments and that there are differences between the rates of price and wage inflations. Friedman, however, apparently neglected to resolve this confusion.

Forder regards it as a simple logical error having no crucial importance as this laxity of Friedman by no means distorted significantly the key message. However, the error is undoubtedly there, and even if we are willing to attribute it to Friedman’s negligence, saying that this minor misstep proved to be neutral as for the overall impact of Friedman’s theory, such a mistake inevitably has an effect on the assessment of Friedman’s oeuvre. As there is no sign of Friedman’s putting forward some undigested thoughts here (as Forder aptly highlights, the same mistake emerges in the drafts), there is no reasoned ground for playing down the mistake. Instead, it is equally easy to argue that Friedman’s explanation is carried by some implicit assumptions that are of crucial importance as for his conclusions. On this showing, these confusions are not logical errors but the results of some arbitrary assumptions and, accordingly, they ought to be assessed in a less generous way. Assuming that Friedman concentrated on his pre-set or pre-given conclusions whilst neglecting the way he got there, these implicit and unnatural assumptions seem to be the perfect manifestations of the instrumentalist principles of F53. At a minimum they effectively highlight how Friedman belittled the importance of careful reasoning.

The system of negatively sloped short-run Phillips curves and a vertical one to signify the natural rate of unemployment, and the possibility for the agents to switch between the curves require an image of employers and employees that cannot be reconciled with the properties their real counterparts have. In these terms the assumptions on the information background are crucial. Friedman’s technique here closely obeys his causal instrumentalist stance in F53. His primary purpose was to generate predictions in line with the observations whilst there is no emphasis upon causal understanding. Hiding behind the idea of descriptive inaccuracy and unrealisticness, Friedman sneaked some suspicious entity properties into his Phillips curves: entity properties that fail to facilitate the sound causal understanding of the labour market phenomena under scrutiny. Friedman’s starting point was the absence of a stable trade-off between inflation and unemployment. As he argued, this trade-off comes to nothing if expectations are taken into account (De Vroey, 2016, p. 102). It is the underlying information structure that Friedman defined via some bizarre assumptions, which is highly problematic as initially he aspired to resolve the confusions over the Phillips curve by considering the role of expectations (i.e. forward-looking behaviour and information processing).

The idea is briefly summarized in the following graphic framework (Fig. 4.1):

The basic mechanism is highly simple, even if its interpretation raises some problems. If the represented economy starts from equilibrium point C (appertaining to the natural rate of unemployment, UN) and if inflation (π) rises from A to B for one reason or another, then the system temporarily relocates into point D and at the same time unemployment lowers below the natural rate (UL). This state, however, is only ephemeral as the system refuses to get stuck in this favourable unemployment situation. The Phillips curve shifts from P′ to P′′ and the economy albeit at a higher level of inflation returns to the natural rate of unemployment (E) where after the temporary swing the system is in general equilibrium (Friedman, 1968, p. 8). In this framework the initial CD movement can be triggered by expansionary monetary policy actions that are designed to set aggregate production and employment to statuses more favourable than the natural level. Friedman, however, was doubtful whether it is possible to permanently reduce unemployment by monetary policy measures.

In Friedman’s (1977) narrative, the narrative Forder (2014) aims to discredit, some empirical studies showed him that the actual (data-based) Phillips curve refused to behave in the way the theory predicted (Friedman, 1977, p. 451). A stable trade-off between unemployment and inflation was undetectable, that is, multiple levels of inflation might belong to the same level of unemployment. Any given rate of unemployment was accompanied by higher and higher rates of inflation.

In the beginning, Friedman came up with a story very similar to the original explanation. When making their labour supply decisions, workers take into account their expectations about inflation and real wages. Taking expectations into account does not mean that workers decide to increase labour supply just because they expect real wages to increase. By contrast, it is an expectation error, a discrepancy between observations and expectations, that triggers the overall effects. Plainly put, the increase in the labour supply is generated by a monetary policy surprise. If monetary policy wants to achieve a rate of unemployment lower than the natural level, then to this end it must boost the money growth rate and hence aggregate demand. One of the channels is the drop in the rate of interest enhancing investment demand, and the other is the higher aggregate demand directly stemming from the increase in the stock of money.

Agents (both producers and employees) in this situation of increased money growth boost their supplies as a response to higher aggregate demand. The worsening of the initial positive employment effect occurs as workers based on their experiences adjust their expectations about prices. As a consequence, unemployment returns to the natural rate. This rebound may take place along two possible mechanisms. In the one scenario, workers finally understand real wage dynamics and as real wage remains the same all along they reduce their labour supply to the initial level. In the other, properly experienced real wages urge workers to ask for higher money wages to be compensated for the higher price level. As a result, this readjustment leads to a drop in the demand for labour force. However, Friedman left this detail unanswered.

To explain the dynamics of the system it is insufficient to refer to some nominal changes experienced as real changes. Two further assumptions are needed. One is the dissimilar reactions of prices and wages as responses to changed demand conditions. Accordingly, stimulated demand increases prices faster than wages, wages are thus stickier.27 As a consequence, the lower rate of unemployment is the effect of dropping real wages, and the only cause of the system’s returning to the natural rate of unemployment is the phase when workers regain their situation awareness which results in an adjustment process. Employees understand the actual dynamics of prices and wages and subsequently they act accordingly. The other arbitrary assumption is a special information asymmetry: what workers do not know about real wages, nominal wages, and prices is completely clear to the employers.

Friedman (1968, p. 10; 1977, p. 456) explains the rigidity of nominal variables (prices and wages) with some contracts concluded on the basis of prior expectations. These contracts hinder prices flexibly adjusting to demand. However, the assumption of dissimilar demand flexibilities of different sets of prices requires a more thorough theoretical underpinning.28 Such contracts only form a special condition that can easily be regarded as responsible in the first place for the stickiness, though assuming their existence is far from obvious. Suggesting some contracts as a theoretical explanation to a subtle system of price changes is only a short cut and requires underpinnings. If such an underpinning is unavailable (as in Friedman’s case), it is easy to assume a radically different pattern of flexibilities where it is the rise in money wages that starts earlier (it only requires the similarly arbitrary assumption that wage and price contracts terminate at different times). However, it is contentious whether the course of events Friedman suggested still emerges. A rise in money wages accompanied by a later increase in prices results in a growth in real wages (dissimilar rates of increase would lead to the same outcome), which leads to a rise in labour supply and, simultaneously, to a drop in the demand for labour force as employers properly perceive price and wage dynamics. Confusing nominal and real changes is still a factor as workers are ill-informed about their actual real wages. They conceive the rise in money wages as a rise in real wages, though in this case they are right. In this scenario the initial expansion of employment thus fails to happen. What is more, the opposite would occur as in case of excess supply it is demand that establishes the level of employment. Only the equilibration of the changes in prices and wages, a restoration of the initial real wages, could take the economy back to the natural rate of unemployment.

Even if for a moment it is left unnoticed that dissimilarity of price and wage flexibilities has remained unexplained, nominal contracts cannot be blamed for workers’ inability to perceive correct price dynamics and hence the changes in real wages. Although wage contracts are formed to apply to a given expected rate of inflation, it is hard to infer that employees cannot perceive actual (and unexpected) inflation. Even under the contracts there is nothing that can stop employees from withdrawing their unnecessarily boosted labour supply when they experience some unexpected changes in real wages. If all it is taken into account, the beneficial employment effect triggered by excess inflation can only be ephemeral, lasting until workers realize the actual albeit unexpected changes in prices.

Lessons become particularly instructive when the same and undelayed flexibility is assumed for both prices and wages. In such a case, excess demand drives prices and wages to start rising at the same time and in the same extent. Workers again, following Friedman’s assumptions, expect price dynamics to keep following the prior trend, so they unsurprisingly experience higher money wages as higher real wages. Employers, however, know that real wages have not changed. As a consequence, the system refuses to leave the natural rate of unemployment even temporarily. Only a vertical shift upwards occurs (from C directly to E). In spite of Friedman’s efforts, a confusion over nominal and real changes in itself is insufficient to explain the CDE route of the economy. The dynamics Friedman suggested rests upon a special system of price and wage dynamics that he left unexplained.

Assumptions of information asymmetry and weird price and wage flexibility are necessary in order that employment effects could be explained. Mistaking nominal changes for real changes and the holding of one of the additional implicit assumptions are insufficient for the dynamics Friedman suggested to emerge. What happens if only one of the assumptions holds? In the first case, only information asymmetry applies. Accordingly, employees fail to perceive the rise in prices. At the same time, however, now wages and prices start changing simultaneously. Workers mistakenly perceive their real wages increasing. Beneficial employment effects do not occur as employers know that real wages have remained unchanged. In the second case, there is a difference only in changes in prices and wages but no information asymmetry. Employees perceiving a rise in real wages err again. However, the same happens to employers (as there is no information asymmetry), so employment can drop at best. By contrast, if assuming information asymmetry away means the correct perception of real wages on both sides, then because of the principle of the short side a boost in employment cannot occur again (for there is a drop in real wages as a matter of fact). All in all, both assumptions, information asymmetry and the flexibility of prices and wages, are indispensable to the emergence of the favourable employment effects.

There is a further case to consider where there is no need for the assumption of information asymmetry, but this is only an irrelevant option. If workers perceive a rise whilst employers a drop in real wages, Friedman’s story on the expansion of employment can be told. In this case it is true that both sides are subject to mistakes, so there is no information asymmetry in the game. However, it is rather an uninteresting scenario as the story is about inflation and sticky wages. The relevant cases are thus confined to either (1) correct perception of a drop in real wages or (2) a misperception.

Friedman seems to be imprecise when saying: ‘[p]roducers will tend to react to the initial expansion in aggregate demand by increasing output, employees by working longer hours, and the unemployed, by taking jobs now offered at former nominal wages’ (Friedman, 1968, p. 10). If workers responded to a rise in aggregate demand with increasing their labour supply at the prevailing nominal wages (that cannot result in higher real wages29), it would be possible only if there was involuntary unemployment in the system. This results in an obvious inconsistency with the ideas on the natural rate of unemployment regarding which Friedman (1968, p. 8) refers to Walrasian general equilibrium. A rise in employment when nominal wages remain unchanged (and when real wages remain unchanged as the best case) is possible only on a disequilibrium labour market, where there are workers willing to take jobs even at the prevailing money (or real) wages, though they cannot because of the insufficiency of aggregate demand.

Another oddity is the implicit information asymmetry underlying the mechanism Friedman suggests. It means that employees’ naïve (adaptive) expectations are accompanied by employers’ correct perception of prices. A rise in prices inflating nominal wages away and facilitating a higher level of demand for labour force is properly perceived by employers. At the same time, on the basis of their expectations, employees acting as buyers on the same market are alleged to still experience former price dynamics living on—erroneously, of course. In terms of real wages, it is the price level of the products employees have in their consumer basket that is relevant, on the basis of which it must be easy to infer the general price level. The idea of a single centralized market, however, makes it impossible to plausibly introduce the incompleteness of information as an assumption.

Shaw (1984, p. 38) traces back this information asymmetry to some explicit assumptions. In his setting the real wage employees perceive is determined by expected inflation, whilst employers make their labour demand decisions on the basis of actual price changes. Friedman refers to this dubious assumption as an evident albeit unimportant circumstance, highlighting that

[b]ecause selling prices of products typically respond to an unanticipated rise in nominal demand faster than prices of factors of production, real wages received have gone down—though real wages anticipated by employees went up, since employees implicitly evaluated the wages offered at the earlier price level. Indeed, the simultaneous fall ex post in real wages to employers and rise ex ante in real wages to employees is what enabled employment to increase. (Friedman, 1968, p. 10)

Confusing nominal and real changes as it is shown here is highly problematic as employees can be deprived of their basic situation awareness only arbitrarily. At the same time, employers’ perception is uncorrupted. This theorizing can hardly be underpinned by the purpose of sound isolation. Friedman seems to have designed the presumptions for this mechanism in order to draw the conclusions he desired in advance. To outline a theory in which unemployment is expected to bounce back to the natural rate after an initial drop, and in which there is no stable trade-off between unemployment and inflation it was necessary for him to introduce a difference of knowledge and the inability to realize the actual real wage. In reality, employees and employers respond to changes in the real wage in different ways, properly depicted even in the simple Marshallian cross of the labour market. Plainly put, what is good for the one is disadvantageous for the other. As the basic market mechanism would never result in the employment effects Friedman wished to see, he was compelled to have a creative play with the underlying assumptions.

In the later version of the model Friedman tried to resolve the problem of information asymmetry by subjecting both sides of the labour market to mistakes (Friedman, 1977, pp. 456–459). Employers face the Lucasian signal processing problem: even though each of them perceives the rise in the price of her own product, she is unable to decide whether it comes from a rise in the general price level or from a favourable readjustment of relative prices. Judging by the possibility of a favourable change in the price structure, an individual producer tries to avoid missing profit opportunities, so she boosts her output. As it is expected, there are two cases of wage rigidity here: nominal wages are sticky and start rising later in time or the rate of growth of nominal wages is lower than the rate of inflation. Any employer when making her labour demand decision only considers the increase in the price of her own product, and on this ground the real wage she perceives is on the decrease. A rise in the demand for labour force is thus plausible. This later version could relax the problematic assumption on price dynamics: the increase in prices naturally starts earlier as this is the change that triggers labour market adjustments and the boost in production. Simultaneously, the problem raised by the confusing word ‘faster’ is also sorted out here.

Friedman’s Marshallian framework implying one centralized market can at best insufficiently suggest the friction between local markets and the knowledge of aggregate-level dynamics. A plausible way of introducing this assumption is definitely beyond its power. Further parts of the story remained unchanged: employees realize the rise in prices with a delay and hence initially conceive higher nominal wages as higher real wages. But for an increase in labour supply, leaving the natural rate of unemployment would never emerge. The swing comes to an end when real wage dynamics becomes properly experienced. Involuntary unemployment does not need to be present in the final version either as it is the rise in nominal wages regarded as a rise in real wages by employees that triggers labour market readjustments.

The theory has remained problematic as it is unclear how the presupposed mechanism leads to a temporary drop in unemployment. The reaction employers show is reasonable. Rising prices naturally inflate nominal wages away; thus under the resulting conditions enhancing employment is possible and profitable. By contrast, the reaction of employees is difficult to understand. They are expected to be aware of their nominal wages. It is also plausible to assume them to perceive the rise in the price level as exactly those prices are considered in the story that they face as consumers on a daily basis: real wage dynamics is driven by the prices of the products they purchase. These two things together should cause employees to experience the drop in real wages. As a consequence, an increase in the labour supply is not a plausible option, whilst a rise in employment presupposes this outcome. Friedman (1977, p. 457) insists on the story that employees only gradually come to understand they have erred on real wage changes. Both sides make mistakes: employers are uncertain about relative prices (i.e. the relationship between product prices and the general price level) whilst employees are unable to realize the real wages.

Confusions worsen if employers are also allowed to show up on consumer markets as buyers. Like workers, now they must perceive the general price level as well. In such a case, neither the information processing problem nor a rise in output emerges: there are no real changes. The problem is that Friedman’s agents perceive multiple prices at the same time (the price of a producer’s own product, nominal wages, and the general price level as on a single market they are inevitably informed about the general price level), and even though there are multiple goods and prices in the picture, information on the general price level cannot be precluded as consumer prices are unproblematically perceived. The story is obscured in one way or another. By contrast, Lucas has placed a single-product economy on his islands where agents acting as employers and employees at the same time are plausibly deprived of information on the general price level. In his framework the general price level consists of the prices of the single product sold on separated markets.

The detail that decreasing real wages trigger a rise in the demand for labour force is clear. This amounts to a shift along the labour demand curve: a drop in real wages enhances the demand for labour. However, lower real wages ought to result in a drop in labour force supplied. To avoid this conclusion, Friedman had to prevent this mechanism from being triggered. To this end, he needed the assumption of employees not perceiving the changes in the prices of the products they consume. Employees can miss cutting down on labour supply in one way only: if they simply do not know they ought to do so. This is exactly the weakest point of Friedman’s theory.

Saying that contracts hinder employees in acting and hence labour supply is constrained seems to be a way out. However, the problem lies in the fact that employees do act—they respond to excess inflation with increasing their labour supply. This cannot be explained by wage rigidity. In order for a boost in employment to emerge a boost in the labour supply is definitely needed to accompany the boost in demand as because of the principle of the short side it is always the smaller of supply and demand that determines the level of employment. If one wants to suggest a rise in employment when real wages are on the decrease, the assumed blindness of employees is indispensable—or else, the power of employers to have the discretion to set employment according to their demand for labour force. We have no other option than attributing the rise in labour supply accompanied by a drop in real wages to the workers’ flawed recognition of the situation. However, this is an arbitrary assumption that efficiently highlights Friedman’s ‘most Friedmanian’ face. No matter how hard we try to resolve the contradictions, we always encounter some new problems.

Is There an Inflation/Output Trade-off?

Milton Friedman American Economic Association Presidential Address was a rebuttal of the Phillips Curve [2]. He argued quite persuasively that there is no output inflation trade-off; hence there is no permanent Phillips Curve. His argument was succinct and elegant; the pricing system was flexible enough to bring the economy to its natural rate of unemployment. Borrowing a page from the efficient markets and rational expectations developments that were taking place around the time of Friedman’s presidential address, some economists including many of Friedman’s disciples began to question the existence of a lagged effect of monetary policy [3]. The argument was as follows: if you know there is a 6 month lag, and you see the money supply going up today, you know that exactly 6 months from now prices will go up. A smart person would at 5 months and 29 days buy goods, hold them for a day and make big profits by selling them when prices go up. A smarter person would anticipate that move and would buy at 5 months and 28 days. The logic of the efficient market is clear and brings us to the first modification of the textbook monetarism. If people know that higher money growth causes higher prices, the lag has to disappear or else some profit opportunities are going unexploited. Hence, if there is any real effect, it has to come from unanticipated money changes. Any attempt to use monetary policy to push the economy beyond the natural unemployment or real interest rates leads to an ever-accelerating inflation rate. Early on the monetary authorities may be able to fool the economy by increasing the money supply unexpectedly. However, over time, the extra money would lead to higher prices and rising inflation expectations. In order for the monetary policy to continue having a real effect requires an ever-higher dose of monetary action/surprises. Ultimately the economy will figure out the game and prices will rise without the output effect. Friedman argued that the long-run Phillips curve is vertical; there is no long run or permanent trade-off.

The American Economic Association (AEA) Presidential Address made clear in no uncertain terms that there was no permanent inflation/output trade-off. Money is only a veil in Friedman’s world. While the supply-siders agree with much of Milton’s address, there is one modification. If the tax system is un-indexed, inflation would push people into a higher tax bracket and thus raise the economy’s marginal tax rate. The higher tax rate leads to lower output and thus a lower demand for money. That, in turn, leads to a higher excess supply of money and thus a higher inflation rate. The process continues until all taxpayers are pushed into the highest tax bracket, at which point the progressive tax system becomes a proportional tax system and bracket creep ceases. The level of output would be permanently lowered as a result of the higher tax rate. From that point on anticipated inflation would not have any effect on the real economy (i.e., there would be no inflation/output trade-off). However, during the bracket creep period as inflation rises so will the unemployment rate. In other words, a perverse Phillips Curve or stagflation will be observed.

The interaction between inflation and the tax rate system offers a simple explanation of the stagflation phenomena that plagued the United States during the decade of the 1970s. That experience fueled an argument that gained much currency in favor of the monetarist view, and it is that price stability reduces uncertainty and promotes long-term economic growth. Therefore the monetarist view leads to a simple conclusion: The Fed should have a single mandate of price stability.

2 Literature review

Central bank behaviour, as a concept, simply denotes the strategy adopted for conducting monetary policy; either as a rule or at discretion. A monetary policy rule specifies future monetary actions as a simple function of economic or monetary conditions and rules out the use of contingent approaches when confronted with economic uncertainties (see Blejer, 1998). Kydland and Prescott (1977) succinctly posited that rules increase the ability of the central bank to remain discipline at its committed policies so as to avoid monetary surprises and, in the end, obtain a lower optimal rate of inflation (see Fischer, 1990). On the other hand, discretion is a ‘zero-based’ strategy that seeks to revise the current monetary policy in tandem with the prevailing economic and/or financial realities. Unlike policy rule that is targeted at price stability only, discretionary monetary policy focuses on economic stability; a combination of price stability and employment generation (Hetzel and Mehra, 1988). However, discretion is dynamically time inconsistent. Hence, considered as a sub-optimal policy outcome (Svensson, 1997; Cukierman, 1986, 1992, 2006; Barro and Gordon, 1983).

Nonetheless, there are many measurement issues attributed to monetary policy rules. Barro (1986) considered it either as quantity or price rules while Van Lear (2000),1 categorized these measurement approaches into five. These are the quantity theory, the McCallum’s (1988) money growth rule, Angell’s (1992) commodity price rule, natural rate of unemployment measure and the Taylor’s (1993) rule. As a hybrid strategy, however, inflation-targeting has been suggested as a valid approach to counteract the effects of both rules and discretionary policy strategies of the monetary authority; especially for developing economies (Walsh, 2009). Since inflation-targeting has been disputed in some empirical studies (Friedman, 2004), Barro and Gordon (1983) suggested good reputation and sound credibility as rules substitute. The theoretical literature on central bank behaviour can be explained either as positive or normative hypothesis (Cukierman and Meltzer, 1986). The former is an exposition about the objectives and constraints that central bankers have to contend with. In this case, inferences are usually drawn from observable variables (such as inflation rate and rate of money growth) while intuition and heuristic implications are obtained from unobservable variables (such as policy credibility). On the other hand, normative hypothesis relates to how the monetary authority would improve the social welfare conditions of all economic agents in a country.

It is along this thread that Kibmer and Wagner (1998) provided three central reasons (such as the employment motive, revenue motive and the balance of payment motive) for the time inconsistency policy; popularly known in the literature as the inflation bias problem. The employment motive is a game-theoretic model which centres on the short-run Phillip curve. This motive predicts that changes in employment is mainly predicated on unexpected inflation and that given perfect forecast as well as rational expectations of private economic agents, output would fluctuate around its natural level on the attendant effects of supply shock. The model presumes that the natural unemployment level can only be lower than the socially desirable output if there exists distortions in taxes or in the presence of distorting wage-setting. The revenue motive presupposes that the government can be motivated to generate unexpected inflation; even in the presence of nominal debt without any effect on the nominal interest rate. In real terms, cost of debt would decline and the real balances for inflation tax would remain unchanged. Regarding the BOP motive, an unexpected devaluation can be undertaken by policymakers in the presence of persistent current account deficits. This is analogous to surprised inflation but explained under the theory of one price. It is indicative to note that one major feature standing out of these motives is the potency of rational expectation of the private agents in shaping the opinion of the policy makers towards a desirable socially optimal welfare function (Blanchard and Fischer, 1989; Schaling, 1995; Kibmer and Wagner, 1998).

The empirical literature on central bank behaviour can be dichotomized into two. The first category has to do with those studies that investigated ex-ante behaviour of central bank; through simulations, while the second relates to those studies that examined the ex-post behaviour of the monetary authority; in the form of reaction functions. Brzozowski (2005) conducted a study to identify central bank’s preferences for the case of Poland. The author tested the hypothesis that the weight on output gap variability in the central bank’s loss function was equal to zero in Poland. The framework adopted was to derive monetary policy reaction function from the central banks optimization problem. The study found that the weights assigned to target variables were not constant over the period 1995–2003. More so, the results showed that the weight attached to inflation stabilization objective in the central bank’s loss function in Poland was equal to the weight assigned to output gap stabilization in the period 1995–1999 but that the output stabilization goal has been abandoned since the year 2000. Besides, Epstein (2009) argued that to get more information on central bank developmental efforts and economic outcomes, case studies will be more useful than cross-country econometrics. The author submitted that there are series of models and institutional structures to learn from and that the cross-country data in the extant literature support central bank’s developmental efforts to promote real investment and economic growth but that the shortfall of these data is that they failed to indicate other broad conclusions of success or failure.

On the other hand, Agu (2011) specified two simple models of monetary policy reaction function for Nigeria with the first being a tracking model based on the revealed preference of the Central Bank of Nigeria and the second as an alternate model which closely followed the Taylor rule. The study could not find evidence for interest rate smoothing and fiscal dominance in the reaction function. No long-run relationship was established among monetary variables and even less so between monetary and real sector variables. The result was found to be consistent with the pronounced policy of the Central Bank of Nigeria to tackle inflation as a priority and the primacy of credit to the private sector as the growth strategies of the monetary authority. However, these results contradicted the findings in the study of Bello and Sanusi (undated) where they found that the Central Bank of Nigeria was conscious of interest rate smoothing. Bello and Sanusi (undated) estimated monetary policy reaction function for the Central Bank of Nigeria through a Taylor-typed rule with quarterly data that spanned 2006Q4 – 2015Q2. The technique of analysis was the Generalized Method of Moment and the results obtained showed that the central bank followed a forward-looking policy rule and committed to an anti-cyclical monetary policy with a forward-looking behaviour of not more than a single quarter into the future.

Iklaga (2007) examined the effect of monetary policy on macroeconomic variables through a Taylor-typed reaction function. The results showed the significance of inflationary pressures in the decision-making process of the monetary authority and that output played the path dependence of interest rate. Also, Siri (2009) analyzed the reaction function of the Central Banks of Ghana, Nigeria and WAEMU and found that the monetary policies for Ghana and Nigeria were not consistent with the Taylor-typed rule or any of its variants. Evidence obtained showed that interest rate weakly reacted to the variations of inflation and output gap. The main observation from the result obtained was that monetary policy applied was different from those announced a’priori and that except for the inflation rate, output appears not to have any influence on the adjustment of the interest rate of the Bank of Ghana (BoG) and the Central Bank of Nigeria (CBN). In addition, the negative sign of the two bank’s reaction coefficient did not adjust their interest rate in response to economic over-activity. Recently, the study of Onanuga et al. (2017) relied on the augmented Taylor’s rule to evaluate the reaction function of the historical path of nominal monetary policy rate in Nigeria for the quarterly period that spanned 1996Q1 – 2014Q4. The main technique of analysis was the Generalized Method of Moments and the reaction function was augmented with the real exchange rate. The study found that real output and exchange rate were both significant in explaining the path of monetary policy in Nigeria.