CLASS 8: THE EFFECTS OF CHANGES IN MONETARY POLICY

• In the following two classes we will study both unanticipated and anticipated changes in monetary policy.
• We will emphasize the forward looking nature of prices, which so far did not play much of a role given that the equilibria we analyzed were stationary.
• We will study the actual impact of changes in monetary policy in different countries/time periods (positive lesson) - in particular, we will discuss how hyperinflations were brought to a halt in a number of countries.
• We will also investigate how changes in monetary policy should be implemented - credibility issues (normative lesson).

The effects of an unanticipated change in monetary policy

• Let us consider the following situation: in periods t=0,1,...,T-1 money grows at a rate $\mu$, and people expect it to grow at that rate forever:
  this implies that nominal money supply is equal to $M_t^s=M_0(1+\mu)^t$, real money balances are equal to the equilibrium level $m_t=m^*(\mu)$, and therefore that the price level is equal to:
  
  $$p_t=\frac{M_0}{m^*(\mu)}(1+\mu)^t$$
  
  
  (if we further assume that the level of public debt is zero (B_t^s=0, all t), we obtain that transfers from the government to individuals (which are equal to seignorage) are equal to: $T_t=M^s_{t+1}-M^s_t=\mu M_0(1+\mu)^t$)
• in period T there is an unanticipated regime change in economic policy (let us say that the government unexpectedly changes), with an associated unanticipated change in monetary policy: it is announced (and the people believe it) that from period T onward (for all $t\geq T$) the money supply will be kept constant at the level reached in period T, which is $M_T^s=M_0(1+\mu)^T$
  (if the level of public debt is still kept at zero, B_t^s=0 all t, this means that transfers from the government to individuals will be zero as well in all periods: T_t=0)
• Given that people expect money supply to be constant forever after, their equilibrium real money balances will adjust accordingly to the new stationary equilibrium. In particular, real money balances will be equal to m^*(0) (if $\mu>0$, then $m^*(0)>m^*(\mu)$)
  The price level will also be constant forever after:
  $p_t=\frac{M_T}{m^*(0)}=\frac{M_0}{m^*(0)}(1+\mu)^T$, all $t\geq T$
• notice that, since $m^*(0)>m^*(\mu)$, it may actually be the case that prices decrease from time T-1 to T. In fact:
  $p_{T-1}=\frac{M_0}{m^*(\mu)}(1+\mu)^{T-1}$
  and $p_T=\frac{M_0}{m^*(0)}(1+\mu)^T$:
  if $m^*(0)>m^*(\mu)(1+\mu)$, then p_T<p_T-1 (see Figure 9.1)
• The fact that prices are forward looking has important implications for the impact of the policy change: no matter what inflation was before the announcement, the fact that people perceive money supply to be constant in the future stops inflation immediately. This is because the equilibrium conditions for the price level have only current and future variables in them, and no past variable can affect them.
• The fact that is future, rather than current, monetary policy that matters is exemplified by the fact that prices between T-1 to T may drop in spite of the fact that money supply is rising.
• The possibility of a -sudden- deflation may be a source of concern for the policy maker. Remember that in our flexible prices/wages + representative agent setup a deflation has no cost. But in model with rigid wages it may cause -temporary- unemployment. Moreover if we relax the representative agent assumption an unexpected deflation may have a sizable redistributional impact from debtors to creditors.
• For all these reasons the central bank may consider trying to avoid the deflation by raising money supply from between periods T-1 and T at a rate greater than $1+\mu$
  In particular, since $p_T=\frac{M^s_T}{m^*(0)}$ and $p_{T-1}=\frac{M^s_{T-1}}{m^*(0)}$, for p_T=p_T-1 it must be that $\frac{M^s_T}{M^s_{T-1}}=\frac{m^*(0)}{m^*(\mu)}$. If $\frac{m^*(0)}{m^*(\mu)}>(1+\mu)$ this means that the money supply must grow at a rate greater than $\mu$.
  This way the Central Bank can both stabilize the price level and avoid the deflation, as long it makes it credible to the public that money supply is going to be constant forever after
• Slight conceptual inconsistency: if we are in a world of perfect foresight, how can it be that an event which was not foreseen actually occurs? One can in part justify that by assuming that the agents were putting a very small probability on the event of a policy change. Since the probability was tiny, before the event they were behaving almost as if the event was not going to occur.

Historical evidence on the end of hyperinflations.

• The end of several hyperinflations all display, at least to some extent, the features of the model we just discussed, namely:
• the change in monetary policy brings inflation to a sudden halt: the price level stabilizes -almost- immediately
  • Germany (1923): before October prices were rising at $300\%$ per month. The monetary reform, which set a ceilings on the amount of money issued by the Central Bank, was announced the 15^th of October. Already from the third week of November prices had stabilized, and actually declined slightly. In the whole of 1924 inflation was 4%.
  • Bolivia (1985): prices were growing $60\%$ per month. the 29^th of August the government announced the `Sachs plan': inflation stopped within 10 days.
• Stabilizations are carried out through sudden halts to the rate of growth of money (often, also by the creation of `new' currency -the Real in Brazil, the Rentenmark in the Weimar Germany).
• The money supply does not decrease as much as prices do. In other terms, there is a (large) increase in real balances: by 10 times in Germany from October 1924 to the end of the year. Also, often there is some degree of `reliquification': the money supply keeps rising even after the stabilization (by 20 times in in Germany from October 1924 to the end of the year). The stabilization is still successful because the agents are convinced (in Germany, because of ceilings on the amount of money issued by the Central Bank) that money supply, albeit growing, is never going to grow again at the before stabilization levels.