Evaluating Policy Institutions – 150 Years of US Monetary Policy

Dollar signs and portraits of past Federal Reserve chairs

How should we evaluate and compare the performance of policy institutions? How should we evaluate and compare policy makers after their term in office? These questions are of central importance to the good functioning of democratic and accountable institutions, but there is little consensus on a method for evaluating and comparing performance. 

In the Barcelona School of Economics Working Paper 1410, “Evaluating Policy Institutions – 150 Years of US Monetary Policy”, Régis Barnichon and Geert Mesters propose a new methodology to evaluate and compare the performance of macroeconomic policy makers and institutions; one that takes into account that policy makers face different initial conditions, different structural shocks and different economic environments. They apply their method to evaluate US monetary policy over the past 150 years. 


A new way to measure monetary policy performance

Table 1 shows the Optimal Reaction Adjustments (ORAs), a new statistic introduced by the authors for evaluating and comparing policy performance. In the context of their monetary policy study the ORA measures by how much more or less a policy maker should have adjusted the short-term interest rate in response to a given non-policy shock.

Table 1: ORA Statistics for US Monetary Policy

chart

Notes: Median ORA statistics together with 68% credible sets. Details on the definition of the variables can be found in the WP.

The conduct of monetary policy strongly improved in the past 30 years

A first key finding is the strong improvement in the reaction function of the Federal Reserve (Fed) over the past 30 years (Table 1; Post-Volcker period). In particular, the performance of the Fed improves across all non-policy shocks; financial, government spending, energy, inflation expectation and TFP shocks. Meanwhile, there is no material improvement in the reaction function of the monetary authority in previous periods.

The Fed’s reaction to the Great Depression was highly sub-optimal

Figures 1 and 2 display the impulse responses to financial shocks and the underlying the ORAs in the early Fed period and in the post Volcker period, respectively. In the early Fed period, the Fed raised the discount rate in response to financial shocks (Figure 1). Combined with the decline in inflation caused by the financial shock, the real policy rate increased substantially, and monetary policy was contractionary (Figure 1; black line, lower-right panel). In contrast, in the post Volcker period the policy rate declined following the financial shock and the ORA only slightly adjusted the response of the policy rate, leading to modest adjustments to the ORA-adjusted responses of inflation and unemployment (Figure 2).

Figure 1: Early Fed, 1913-1941, Reaction to Financial Shocks

figure

Notes: The top (resp. bottom) row shows the median responses (thick line) of inflation, unemployment and the Fed’s discount rate to a monetary policy shock ϵ (resp. financial shock ξ). The dotted green lines show the ORA adjusted impulse responses. The 95% and 67% credible sets are plotted as dark and light shaded areas, respectively.

Figure 2: Post Volcker Fed, 1990-2019, Reaction to Financial Shocks

chart

Notes: The top (resp. bottom) row shows the median responses (thick line) of inflation, unemployment and the Fed’s discount rate to a monetary policy shock ϵ (resp. financial shock ξ). The dotted green lines show the ORA adjusted impulse responses. The 95% and 67% credible sets are plotted as dark and light shaded areas, respectively.

The performance of the Fed in the 1920s-1930s was not worse than the performance of the passive Gold Standard monetary regime

Comparing the ORAs in the Gold Standard period with the ORAs in the Early Fed period, the authors find that the excessive passivity in the Gold Standard period continued after the founding of the Fed — the ORAs are similar across the two periods (Table 1; Gold standard, Early Fed). In addition, the early Fed period was excessively passive in their response to financial shocks as well as to other shocks such as government spending shocks. 

The reaction function of the Fed during the post World War II period was almost as sub-optimal as the early Fed’s reaction  

Overall, Fed’s performance during the post-World War II (WWII) period was on a par with the performance during the 1920s-1930s, with ORAs of similar magnitudes, though the nature of the underlying shocks was different (Table 1; Early Fed, Post-WWII). Post-WWII, the Fed’s reaction was too weak following all the different supply-type shocks: energy price shocks, TFP shocks as well as inflation expectation shocks. Notably, the reaction to shocks to inflation expectation over the post-WWII period displays the largest deviation from optimality over the entire 150 year of monetary history.