Thanks to Dr. Kirdan Lees for prompting me to write today’s post. Today’s topic of discussion is nominal income targeting.
What is nominal income targeting?
Nominal income targeting is usually viewed as an alternative monetary strategy to inflation targeting, and has never explicitly been applied in practice by any central bank. However, there is an overwhelming amount of literature discussing this monetary policy tool and how it compares to the flexible (inflation) targeting.
Measures such as nominal gross domestic product (NGDP) and nominal gross domestic income (NGDI) are used to apply this rule practically. As a result in this post we will focus on the idea of NGDP level targeting.
What do we mean by nominal income targeting
When considering where we are, and what NGDP level targeting is, two distinctions must be made here:
- Level targeting between NGDP and CPI (level of output and price level)
- And growth targeting between NGDP and CPI (growth rate vs inflation)
The key distinction we are focusing on between NGDP targeting and flexible inflation targeting is about the LEVEL vs the GROWTH of the measure, not a difference in the actual measure that is targeted by policy makers. The reason for this is that flexible inflation targeting involves a Taylor Rule (looking at inflation and real output variations) while NGDP itself is a combination of these elements.
Greg Mankiw is a big advocate of the nominal income targeting. In “Hall and Mankiw – Nominal income targeting (1994)”, there is a suggestion that growth targeting is generally dominated by level targeting – and that targeting the level of nominal income would have led to lower variability in both inflation and output.
However, things get more complicated. Lawrence Ball (1997) noted that, with backward looking inflation expectations nominal income targets (as opposed to a Taylor rule based on deviations in inflation and output) can generate significant variability in output and prices.
Following authors noted that the Ball result was sensitive to the expectations argument used, and as a result by the turn of the millenium the debate was left relatively unresolved with many seeing a limited role for level targeting.
During the GFC discussions about this were many and varied, with the ECB producing a document that outlines the costs and benefits associated with such a target.
So why do it, or not do it
From the ECB note there are three arguments used for a NGDP level target, and five against (based on a mix of arguments on level vs price, and NGDP vs price level issues):
- Current low inflation and output: Committing to a higher path for NGDP implies that central banks will be committed to increase either prices or output – which will help to stimulate economies that are trapped at the zero lower bound.
- In a world of increasing unforeseen supply shocks flexible inflation targeting will “look past” them, while NGDP targeting will respond if the impact on output is greater than that on prices.
- By stabilising NGDP it stabilised public and private debt paths.
- NGDP targeting and inflation targeting are already close in practice, so any associated gains are overstated.
- With backward looking expectations NGDP targeting will increase variability in inflation and output.
- NGDP is harder to communicate to the public prices.
- The value placed on real output and price growth differ in terms of social value, and NGDP imposes a fixed weight on each.
- It misses financial stability.
However, I am not sure I fully agree with this – and the paper itself is a little inconsistent on a number of these. So let’s take these ideas, but ask ourselves a bit more.
How is it different from what we are currently doing?
To think about it for ourselves it is useful to consider about how a level and growth target differ.
By level targeting of NGDP, historical shocks are taken into account. As a result, if we deviated from a set level in the past, we will implement policy that gets us back to that level path. For example, if we made a mistake and NGDP fell below target we would then push for higher than average target growth until we got back to that level.
With the growth targeting (flexible inflation targeting which includes information on economic output) a central bank committing to targeting a level of growth going forward no matter what happened in the past. As a result, any deviation from a given level is treated as in the past, and any mistake is ignored – and average growth in NGDP in the future will be at the same level no matter what has happened previously.
As a result this is the key – is there a level why we want to commit to a given level of NGDP as opposed to giving certain certainty about the average growth rate going forward?
Is that it?
There is something more fundamental that is missed in all these discussions. But I will get back to this next week.
To understand what is going on we need to ask what expectations are being “set”, what is the “target” and how do these reflect what a central bank can “do”?
Expectations: We know they can be adaptive (backward looking) or rational (forward looking), but what do they refer to? Are people making choices based on expected prices, or are they making choices based on expected incomes? Do they view shocks as permanent or temporary?
Target: Is the goal to anchor the price level, or to anchor the level of expected nominal income growth? Is it to limit variability in prices and output?
What they do: If a central bank wants to increase prices can it, if it wants to increase nominal incomes can it? If they can do both, how does this influence their ability to “close output gaps”?
The answer is usually that the central bank can perform actions to change any nominal variable (NGDP or prices), that the goal is to make as smooth a path for prices and output as possible (limit variability), and that irrespective of the expectations growth targeting is better – as with adaptive expectations level targeting creates variability, while with rational expectations and no lower bound on interest rates inflation targeting can close the output gap without the risk of “unanchoring” inflation expectations.
In this way the only justification for level targeting appears to be when we are at the zero-lower bound, as a way of committing to forward guidance. But is this the whole story? I will share my thoughts next week.