Taylor’s rule means the original rule proposed by John Taylor in 1993. Taylor rules means various versions of the rule that have come up since the original rule.

What is the right Taylor rules has been a topic of great interest in the crisis (see this for an overview). John Taylor on his blog has been suggesting to look at his original rule not the other variants which have people have been looking into. Choice of Taylor rule variants is critical as it leads to following policy choices

Taylor Rule says:

Policy rate = 2+ weight (actual inflation – desired inflation) + weight (actual output – potential output)

Weights are assigned based on importance of the target. Taylor original rule proposed it as 0.5 for both implying equal weights to both. Variants have suggested giving more importance to inflation or output and hence higher weight and differences in policy outcomes.

The crisis led to following debates over taylor rules:

- First was fed policy too low before the crisis? Taylor famously said yes. Bernanke defended Fed policy saying if one uses forecasts of inflation instead of actual inflation, Fed policy was right. Mon pol is about expectations and hence Fed should look at future inflation not actual/current one. Taylor said Fed inflation exp were too low to begin with.
- Then post crisis, we are having discussions on whether current Fed policy good enough. Was QE needed? Taylor original rule says Fed policy around 0.5% and QE not needed. However, variants show Fed policy should be -(ve) and as rates cannot go down beyond zero, QE is needed. For instance research shows if Fed buys USD 200 bn of assets = 0.25% of Policy rate cut. Hence till now Fed has bought nearly 2.3 trilluion USD of bonds. This implies Fed has kind of maintained policy rates at -(ve) 2.25%.

With this background, Alex Nikolsko-Rzhevskyy and David Papell write a timely paper.

*We answer the question using the simplest possible methodology. There is widespread agreement that Fed policy was too stimulative following the recession in the mid-1970s, leading to the Great Inflation later in the decade. There is also widespread agreement that Fed policy following the recession in the early 1990s produced strong growth without causing high inflation. We take advantage of these natural experiments by plotting the interest rate implied by various Taylor rules with the actual path of the federal funds rate. *

*Under the stipulation that Fed policy was too stimulative in the mid-1970s and about right in the early 1990s, alternative Fed policies with substantially lower interest rates would have produced worse outcomes. In order to replicate the policy choices actually faced by the Fed, we only use real-time data that was available to Fed researchers and policymakers when the interest setting decisions were made.*

Hmmm…

What are the findings? They discuss findings under original and variants of Taylor rule: 1) Original 2) Taylor rule with a large output gap weight/coefficient 3) Original rule with forecasts of inflation 4) Taylor rule with a large output gap weight/coefficient and forecasts of inflation

*Suppose that Fed policy had followed Taylor’s original rule. The resultant interest rate would have been higher than the realized rate in the mid-1970s, about equal to the realized rate in the early 1990s, and above the realized rate in the mid-2000s. By following the same rule in the 1970s and 2000s that characterized policy in the late 1980s and 1990, the Fed would have avoided contributing to either the Great Inflation of the 1970s or the Great Deviation of the 2000s.*

*Now suppose that, instead of following Taylor’s original rule, Fed policy had followed alternative Taylor rules. We first consider a rule with a larger output gap coefficient. Fed policy much more closely followed a rule with an output gap coefficient of 1.0 than one with a coefficient of 0.5 in both the mid- 1970s and mid-2000s. In contrast, a policy rule with a higher output gap coefficient would have produced a much lower interest rate following the recession of the early 1990s, under 1 percent compared with the actual 3 percent. Given the positive economic performance of the mid-to-late 1990s compared with the late 1970s and late 2000s, this does not auger for raising Taylor’s output gap coefficient. *

*Next, suppose that the Fed had followed a second alternative Taylor rule, this time with Taylor’s original output gap coefficient but with forecasted rather than realized inflation. The results for the mid- 1970s and mid-2000s are similar to those with a higher output gap coefficient and realized inflation, as Fed policy more closely followed a rule with forecasted inflation than one with realized inflation in both periods. For the early 1990s, the implied rules with an output gap coefficient of 0.5 are similar whether realized or forecasted inflation is used. Given the dismal economic performance of the late 1970s and late 2000s, this does not auger for changing the Taylor rule to incorporate forecasted inflation. *

*Finally, suppose that the Fed had followed a third alternative Taylor rule, this time with both an output gap coefficient of 1.0 rather than Taylor’s original output gap coefficient of 0.5 and with forecasted rather than realized inflation. Under this rule, the federal funds rate would have been more than 5 percentage points lower than actual Fed policy and would have hit the zero lower bound at the height of the 1975 recession. Since there is widespread agreement that too stimulative Fed policy in the mid-1970s was an important contributor to the Great Inflation of the late 1970s, it’s frightening to think about how high inflation might have been under this rule. A similar scenario characterizes the mid-2000s, with the federal funds rate implied by the rule one percentage point lower than actual Fed policy and hitting the zero lower bound in mid-2003. This would have added even more fuel to the housing boom and magnified the extent of the Great Deviation.*

So basically, original rule is when the policy works the best, Fed stuck to the rule in 1990s and deviated in 1970s and 2000s leading to crisis.

*If the Fed had followed Taylor’s rule, it would have avoided fueling both the Great Inflation of the 1970s and the Great Deviation of the 2000s, while closely following its successful policy of the 1990s. If it had followed Taylor rules with the “improvements” used in 2009 and 2010 to justify a negative implied federal funds rate, a higher output gap coefficient and forecasted rather than realized inflation, it would have contributed even more to the Great Inflation and Great Deviation while running overly stimulative policy in the 1990s. It is difficult to see why “improvements” to Taylor’s rule which would have produced worse historical outcomes should be expected to produce better outcomes in the 2010s, and so we fail to find any justification for changing the rule*

Even now, Fed seems to be using Taylor variants to defend its ultra loose policy. If history is any guide then Fed is again deviating from the policy suggested by the original rule. And whenever that happens, it is a disaster:

*With Taylor’s original rule, the prescribed federal funds rate for 2009 – 2010 is zero or slightly negative. With a variant of the Taylor rule that doubles the size of the output gap coefficient, it is about negative four percent. This is important because, with the constraint of a zero lower bound on the federal funds rate, large negative prescribed interest rates provide a rationale for the Fed’s quantitative easing in 2009 (QE1) and 2010-2011 (QE2). Our paper does not say whether or not QE1 and QE2 were good policies, a topic that is beyond the scope of our research. It does say that, if you are going to use negative prescribed interest rates to justify quantitative easing, you need to use a rule that can be justified by historical experience. Taylor’s original rule, which can be justified by historical experience, does not produce negative prescribed interest rates for 2009-2011. Variants of Taylor rules with larger output gap coefficients, which do produce negative interest rates, cannot be justified by historical experience. The Taylor rule does not provide a rationale for quantitative easing..*

A nice paper. Though, again using Taylor rule in current situation is always a q-mark. Huge financial sector distress, poor economic situation, zero interest rates etc complicate the whole situation. But still history has shown time and again Fed deviation is costly. And current policy is one such deviation.

May 25, 2011 at 8:07 am

[…] via Taylor’s Rule versus Taylor Rules. […]