Saturday, March 14, 2009

The neutrality of money (amv)

This post responds to fg’s recent discussion of the debates on who is to blame for the current economic crisis. We share a lot of common ground. Both of us believe that the primary “cause of the mess was NOT insufficient regulation of financial markets,” that rather “unsustainable credit” and thus “the extraordinary possibility of financial institutions to raise massive funds for financing very risky positions” are the root causes of the financial meltdown (even though the monetary base is not at the root of the problem). Both of us blame monetary policy and a too low interest rate, but both of us are aware of the many other factors that have contributed to unsustainable credit expansion. Both of us do think of Lehman’s bankruptcy as a trigger, a symptom, but certainly not as having any explanatory power. We agree that if policy had followed the rules proposed by contemporary monetary theory (“New-Keynesian models”), credit expansion would have been more moderate, perhaps even much more moderate. I even agree that the New-Keynesian workhorse can be modified to the extent that it ‘predicts’ boom-bust cycles (not actual prediction, of course, but in a purely analytical sense).

I slightly disagree, however, with fg’s suggestion that contemporary theory is not a problem in itself. Although the evolution of economic theory is not responsible for the performance of monetary policy in the short or even medium run, there can be little doubt that theory impacts the long-run performance of policy (it is however true that theory usually lags behind financial innovations and operational central bank policy). Just look at our central banks, the way they communicate, their institutional settings, the almost incestuous relationship between monetary theorists and policy-makers, or simply the employment opportunities at central banks. Just consider who heads the FED and who the US Treasury.

I argue that over the past 60-70 years, economic theory heavily impacted monetary policy, the Great Moderation being a benign outcome of this highly subtle process. I also argue that while monetary theory improved policy outcome, it is itself flawed to a considerable extent so as to produce own defects. These defects contributed to the blindness of many economists to the expansionary path of Greenspanian policy (not accidentally, also Milton Friedman). Thus, at the root of the crisis we find not only policy neglecting the rules of conduct suggested by economic theory, but also theory proposing false rules. Before we, as economists, can blame politicians for the current mess, it is crucial to improve on the rules we ‘ask’ them to follow. In concrete, we should restate our understanding of money’s neutrality as a benchmark for monetary policy.

Today, the profession understands policy to be neutral as long as it guarantees price stability (however defined). Due to signal-extraction problems, in face of which rational expectations are helpless, a minimum volatility of the price level or any inflation trend maximizes real economic performance. Malinvestments and improper allocation due to exchange at false relative prices is reduced and real data properly communicated. “The deadweight losses due to relative price distortions can be […] completely eliminated, in principle, by stabilizing the aggregate price level.” (Woodford, Interest and Prices, p. 405) This is argued within the profession since Irving Fisher (an outcast to the economic circles of his time due to his adherence to price-level targeting).

Whereas I side with the view that the neutrality of money means the absence of monetary distortions of the market process, which is driven by changes in real data (preferences, technology, endowments) and enabled by money-prices relations reflecting relative scarcities, I decline the proposition that a stable price level or inflation trend allows for neutral money. Thus, I side with the normative proposition that policy should aim at money neutrality - defined as the absence of relative price distortions due to money. Yet, I cannot see how this can be achieved in the real world, where net savings and net investments are positive, by ensuring a stable price level or inflation trend. Price stability is not a property of equilibrium growth with the rate of accumulation adapting over time to rationally coordinated intertemporal plans.

In the real world, additions to money do not impact on a stationary economy or on steady-state growth. They rather affect a deliberately chosen process of capital deepening (increasing per-capita incomes), that is, they affect saving-in-process. The saving process is endogenous, the saving rate at any point in time determined by the average rate of return on investment, the height of income, and the social rate of time preference. All these variables change over time, making up for the kind of transitional dynamics described by Ramsey (1928) and Wicksell (1928; p.m.). Thus, the public chooses an optimal rate of growth and any deviations therefrom implies consumption being too high or too low as to allow for the growth target of the public. If monetary policy is ask to keep money neutral, it has to support the households’ choice in respect to the optimal growth path of output. Policy has to minimize any such output gap. The central bank has to protect the saver.

Fg may counter that the Taylor-rule, the rule proposed by mainstream theory and not followed by policy, explicitly asks the monetary authority to close the output gap and thus takes account of the saving process and the increase in output. The representative household hinted at above and its intertemporal choice indeed marks the starting point of all modern monetary analysis. But the output gap described by the Taylor-rule states the difference between a given and thus unexplained trend path and the actual path of output that can deviate due to shocks, etc. The Taylor-rule says nothing about the trend path itself.

This is however not the kind of gap I am talking about. If we model an optimum path of consumption and income, we are concerned with the choice among alternative trend paths. As we know at least since Fisher, interest rates matter in the determination of the time shape of aggregate income. His separation theorem tells us that due to the existence of financial markets and intertemporal exchange opportunities among households (savers), producers (investors) are free too improve on the intertemporal exchange with nature (choosing the investment opportunities with the highest time rate of expected yield irrespective of the time preference of the owners of the firm. The existence of interest as a market price, being a yield to savers and a cost- as well as a capitalization-factor to investors, liberates productive choice from the immediate control of the single households’ time preference, amplifying the investors’ degrees of freedom to approach those investments that provide the highest income in the shortest span of time. The loan rate of interest is burdened with the communication between households’ average choice regarding trend growth (net saving) and the growth rate actually initiated by the investors’ choice of the pace of capital formation (net investment).

Money tends to loose this joint. A fall in the rate of interest, other things being equal, reduces expected future income, increases the relative scarcity of future consumption, and induces a higher saving rate. But it also reduces the future return on any present scarifies, discouraging net saving. Which of the two influences will dominate in real life is open to empirical evaluation. In either case, the optimal growth path chosen by the public probably changes. Further, it is clear in principlethat in either case net investment outpaces net saving by the amount of money supply added by credit creation, which pours into capital markets, disguised as voluntary net savings and as such indistinguishable from funds contributed at the cost of present consumption. Investors therefore choose a different time path of income, different from the optimal path determined by the households, thereby violating the neutrality postulate. As a cost-factor, a fall in interest rates increases the scale of investment. As a capitalization-factor, it also changes the structure of investment spending in favor of additions to the stock of durable capital and consumption goods.

No matter if an easy-monetary stance induces a boom-bust cycle or not, monetary policy can easily distort intertemporal coordination, bringing about inconsistencies between individual plans. The impact on output depends on the real state of the economy. If factors of production are idle, or if they are made available by forced saving, output increases with the supply of money. If factors of production are scarce, if their intra-sectoral mobility is low, and the degree of indivisibilities high, output growth increases only by accident. In either case, money is non-neutral. To ensure neutrality, the central bank has to constrain Wicksellian disbursements of credit money so as to offset changes in the level of hoarding or dishoarding, thereby ensuring that the rate of interest is neutral in the sense that the ex ante level of net investments equals the level of net savings, and that the loan rate equates the expected real yield on investment. If these conditions are fulfilled, aggregate expenditures remain constant and all investment financed by consumption foregone and thus constraint by opportunity costs.

This kind of macroeconomic equilibrium is incompatible with the goal of price stability. An economy characterized by positive net savings and net investment is per definition progressive. Given the constancy of aggregate demand as a condition of macroeconomic equilibrium, the price level has to fall as long as the economy has still not reached stationarity or steady-state (Ramsey’s bliss). Any attempt to stabilize the price level or a positive rate of inflation imposes a permanent excess of investment spending over savings, thereby increasing the aggregate level of expenditure. Further, a decline in the interest rate induces savers to shift their portfolio in favor of more risky investments, given the level of average risk aversion, thereby reducing the finance costs of such productive opportunities. At the same time, a decrease in interest rates hikes the demand prices of the most capitalistic processes relative to the prices of labor intensive processes. The financial sector itself fall victim to easy money and overinvests. A boom in the housing industry and other industries producing durable consumer goods reflects the same logic for consumption spending. The actual path of consumption growth deviates from the level necessary to keep the economy on the chosen or optimal pace of capital formation.

Either we ask policy to target macroeconomic equilibrium, or to target the price level instead. We as theorists have to choose first. We cannot have it both ways. To the extent that the interest rate proposed by the Taylor-rule is lower than a neutral rate, that is, to the extent that it allows for "investment over saving" to become persistent, the fault is ours. To the extent monetary expansion has indeed caused the current crisis, the blame is on us.