Why MMT can’t work

Abstract

Using a stock-flow consistent ISLM open-economy model, this article shows that, unless very specific country circumstances hold, modern money theory (MMT) cannot work as an effective and sustainable macroeconomic policy program aimed to achieve and maintain full-employment output through persistent money-financed fiscal deficits in economies suffering from Keynesian unemployment or underemployment. Specific country circumstances include cases where the economy enjoys very high policy credibility in the eyes of the international financial markets or issues an international reserve currency; under such circumstances, the adverse outcomes of MMT policy can be prevented and expansionary demand shocks can be effective. Short of such features, an open and internationally highly financially integrated economy that implements MMT policy would either see its money stock grow unsustainably large or would have to set domestic interest rates to levels that would be inconsistent with the policy objective of resource full employment and that would cause instead economic and financial instability. The article explains why ISLM analysis is used to support the arguments developed in it.

Appendix 1: the optimal demand for money

This appendix develops a modified version of the Allais–Baumol–Tobin (ABT) inventory model for transaction money demand used in the article. The original ABT model is here modified to show that, as discussed in the text, equilibrium in the money market with a growing stock of money requires the government to pay interest to money holders, and that, all else equal, the required interest rate conditional on the government’s policy credibility as perceived by the financial markets. The ABT approach was selected as a simple and yet effective method to incorporate an optimal demand-for-money framework in the economy’s model used in this article. The objective was not to pick the best possible methodology available to derive optimal money but to recognize the role of optimal money demand as instrumental to assess MMT policy tendency toward the creation of "excess" money balances, in a context where money is demanded for its transaction services and based on its opportunity cost.

Suppose households receive nominal income PY at the beginning of each period and spend it evenly during the period. Average wealth is $PY/2$, and, according to the model assumptions, it is held in the form of money balances, M, earning interest ${i}_{M}$, and foreign asset balances, FX, earning interest ${i}_{RW}$. To finance transactions, households must first hold M balances; thus, before wealth held in the form of FX can be spent, it has to be converted into M at transaction cost f per transaction. Suppose each household divides the period into n subperiods initially placing $PY/n$ in money balances and the rest in FX. At the end of each subperiod, FX balances are converted into M balances in n − 1 transactions of equal size $PY/n$. Thus, average money holdings over n subperiods will be $M=\frac{1}{n}\frac{PY}{2}$ and average foreign asset holdings will be $FX=\frac{n-1}{n}\frac{PY}{2}$.

Thus, the net gain, $\Gamma$, from holding wealth in both assets, taking into account the need to finance transactions, is given by:

$$\Gamma =\frac{1}{n}\frac{PY{i}_{M}}{2}+\frac{(n-1)}{n}\frac{PY{i}_{RW}}{2}-(n-1)f$$

(11)

Maximizing $\Gamma$ with respect to n requires:

$$\frac{\partial\Gamma }{\partial n}=\frac{-2{n}^{2}f+PY{i}_{RW}}{2{n}^{2}}-\frac{PY{i}_{M}}{2{n}^{2}}=0.$$

(12)

Therefore, the optimal choice for n is

$${n}^{opt}=\sqrt{\frac{\left({i}_{RW}-{i}_{M}\right)PY}{2f}}.$$

(13)

Replacing Eq. (12) into the equation for M, the optimal demand-for-money equation is

$$\frac{{M}^{opt}}{P}\equiv {m}^{opt}=Y\sqrt{\frac{f}{2\left({i}_{RW}-{i}_{M}\right)}},$$

(14)

which shows that the demand for money varies positively with real income, the transaction cost, and the interest paid on money balances, and negatively with the rate of return on alternative assets (in this case, foreign exchange).

Consider now that in an LC economy (i.e., characterized by a low $\beta$) the expectations that the government mismanages the money supply over the relevant future time horizon induce wealth holders, all else equal, to hold larger shares of their wealth held in FX, as they factor into their portfolio choices the future expected (internal and external) value of money. In particular, they may fear the risk of excess money supply creation and government’s inadequacy or unwillingness to react to threats of currency depreciation and inflation through appropriate policy action. This implies that ${n}_{LC}^{opt}>{n}_{HC}^{opt}$ and, hence, ${m}_{LC}^{opt}<{m}_{HC}^{opt}$. Thus, in order to induce wealth holders in the LC economy and HC economy, respectively, to hold the same level of real money $m$, all else equal, it must be that ${i}_{M}^{LC}(m-{m}_{LC}^{opt})|{\beta }_{LC}>{i}_{M}^{HC}(m-{m}_{HC}^{opt})|{\beta }_{HC}$.

Similarly, the change in the interest rate required by a given rise of excess money balances will be larger in an LC than in an HC economy, and the difference will grow larger with the rise of excess money balances, that is, ${i}_{M}^{{^{\prime}}LC}(\bullet )|{\beta }_{LC}>{i}_{M}^{{^{\prime}}HC}(\bullet )|{\beta }_{HC}$. These relative differences in the interest rate adjustment will vary inversely with the difference in the level of policy credibility of the economies being compared, that is, the policy space available would be higher, the higher the level of policy credibility of the economy concerned (see Sect. 6).

Formally, all these features are formally captured by the following equation for the interest rate on money:

$${i}_{M,j}={i}_{M}\left({m}_{j}-{m}_{j}^{opt}|{\beta }_{j}\right),$$

(15)

with $j=LC,HC$ and ${i}_{M}\left|{\beta }_{LC}>{i}_{M}\right|{\beta }_{HC}>0;\,{i}_{M}^{{^{\prime}}{^{\prime}}}\left|{\beta }_{LC}>{i}_{M}^{{^{\prime}}{^{\prime}}}\right|{\beta }_{HC}>0;$${i}_{M}^{{^{\prime}}{^{\prime}}{^{\prime}}}\left|{\beta }_{LC}>{i}_{M}^{{^{\prime}}{^{\prime}}{^{\prime}}}\right|{\beta }_{HC}>0$, where ${m}_{j}-{m}_{j}^{opt}$ measures "excess" money in country j.

Equation (14) defines the position and shape of the optimal demand-for-money function and indicates that (1) the required real interest rate on money rises with excess money balances and (2) the height and steepness of the demand schedule in the $({i}_{M},m|\beta )$ space are conditional on the economy’s level of policy credibility, all else being equal. All this graphically represented in Fig. 5 below, which appears as Fig. 1 in the text.

Fig. 5

Stylized demand-for-money schedules in economies with different policy credibility

In light of Eq. (14), and dropping the country index, the optimal demand-for-money function given by Eq. (13) can be written in implicit form as:

$$\frac{{M}^{D}}{P}=m\left(Y, {i}_{M}|\beta ,f,{i}_{RW}\right).$$

(16)

With f and ${i}_{RW}$ set exogenously and assumed to be constant (for reasons of simplicity but at no loss of generality), Eq. (15) reduces to:

$$\frac{{M}^{D}}{P}=m\left(Y,{i}_{M}|\beta \right),\mathrm{ with }{m}_{Y }{{^{\prime}}},{m}_{{i}_{M}}{{^{\prime}}}>0,$$

(17)

which enters the money market equilibrium MM schedule of Eq. (3) in the text.

Unlike the conventional LM schedule in ISLM analysis, the MM schedule has a negative slope: for a given stock of M and all else being equal, a higher transaction demand for M driven by Y requires a decrease in interest rate paid on M balances, ${i}_{M}$, required to keep the money market in equilibrium.