Simon Wren-Lewis on fiscal versus monetary policy.

SW-L (emeritus economics prof at Oxford) has just published an article on his blog on the above topic. While I often leave comments after his articles, I have so many comments to make on this article that they could occupy more space than the article itself. So I’ve reproduced his article below, with my comments at relevant points in green italics. He starts…..

One divide between mainstream and many heterodox economists is on whether monetary or fiscal policy should be used for macroeconomic stabilisation (controlling demand to influence inflation and output). What makes a good instrument in this context? As I have argued before, a key difference between the mainstream and MMT involves different answers to this question. I think the following issues are critical.

1. How quickly do changes in the instrument (e.g. increases in interest rates) influence demand?

2. How quickly can the instrument be changed? Are there limits to how far it can be changed?

3. How reliable is the impact of the instrument on demand? In other words how uncertain is the impact of a change in the instrument on demand?

4. How certain can we be that whoever has power over the instrument will use it in the necessary way?

5.  Does changing the instrument have ‘side effects’ which are undesirable?

If we apply these questions to whether to use interest rates or some element of fiscal policy, what answer do we get?

Before doing that, it is worth noting this is all about the quickest and most reliable way to influence demand. It is quite separate to how demand influences inflation (as long as we are talking about underlying inflation).

The first question is important because long lags between changing the instrument and it influencing demand mess up good policymaking. Imagine how good your central heating would be if there was a day’s delay between it getting cold and the heating coming on. It is also perhaps the most interesting question for a macroeconomist. A full discussion would take a textbook, so to avoid that I’m going to suggest that the answer is not critical to why the mainstream prefers monetary to fiscal stabilisation.  

The second question is as important for obvious reasons. If an instrument can only be changed every year, that is like having very long lags before the instrument has an effect. On this question monetary policy seems to have a clear advantage given current institutional arrangements. Some of this difference is difficult to change: it takes time for a bureaucracy to move. As I noted with the fiscal expansion implemented by China after the crisis, about half of the projects were underway within a year. Others delays are in principle easier to change: there is no reason why tax changes need only happen during Budgets in the UK, for example.

First, SW-L obviously has a good grasp of what heterodox economists are thinking, when he suggests they have a preference for fiscal over monetary policy.

Next, SW-L says “Some of this difference is difficult to change: it takes time for a bureaucracy to move.” That is rather contradicted by his next sentence which says that half the fiscal expansion measures implemented in China recent were up and running within a  year.

Re “If an instrument can be changed every year…”, presumably SW-L has in mind the annual UK budget “ceremony” in the House of Commons, and I assume the suggestion is that some fiscal changes can only take once a year (apologies to SW-L if I’m putting words into his mouth). In fact that “budget ceremony” is peculiar to the UK: there is no good reason fiscal changes cannot be made at any time. Indeed during the recent crisis, the UK’s VAT rate was changed twice outside the “budget ceremony window”.

Another point here is that if fiscal changes are difficult and slow to implement, that messes up the Job Guarantee. JG is a system where jobs are supposed to be created VERY QUICKLY given a rise in unemployment. Those jobs can be with existing employers (public and/or private) or on specially set up schemes as was the case with the WPA in the US in the 1930s. (For a discussion of the relative merits of “existing employer versus special scheme” see the several articles I’ve written on that topic. Briefly I argue that the “existing  employer” option is better because one gets a better mix of skilled labour, unskilled labour, capital equipment etc, plus more realistic work experience.)

In fact, it shouldn’t be beyond the wit of man to set up a system where local and city governments, central government departments, state schools, state run hospitals etc can be instructed very quickly to spend more and take on a few extra staff, either in the form of JG people or as regular employees (who might then have to be sacked when the fiscal stimulus is withdrawn).

In contrast, other forms of fiscal spending, e.g. construction projects take much longer to get going.

SW-L continues….

The second part of the second question is a clear negative for interest rates, because they have a lower bound. This is not the case for fiscal instruments: you can always cut taxes further for example. Because this is a critical failure for interest rate policy, effectively the discussion in this post is just about what happens when interest rates are not at the lower bound. Even so, potentially having two different instruments for different situations is a count against monetary policy.

There’s another “count” against monetary policy, which I go into in detail here – “here” being a thesis which will hopefully be published (in updated form) in a journal quite soon. The latter count is thus.

As explained in the latter thesis, most of the arguments for government borrowing do not stand inspection. Thus we are in the strange position where interest rates have been artificially elevated for decades, but the state (i.e. government and central bank) cannot reduce interest rates unless they are first artificially elevated, which is an absurdity. Put another way, if we had a permanent zero interest rate policy (advocated by Milton Friedman and several MMTers and in the above thesis), then cutting interest rates would be impossible.

SW-L continues…

The third question is often not asked, but it is absolutely critical. Imagine raising the temperature on a room thermostat which not only had no calibration, but which acted in different ways each day or even each hour. OMT is a clear example of a poor instrument because central banks have far less idea of how effective it is than interest rate changes, partly because of less data but also because of likely non-linearities.

Are interest rate changes more or less reliable than fiscal changes? The big advantage of government spending changes is that their direct impact on demand is known, but as we have already noted such measures are slow to implement. Tax changes are quicker to makes, but many mainstream economists would argue that their impact is no more reliable than the impact of interest rate changes. In contrast some heterodox economists (especially MMTers) would argue interest rate changes are so unreliable even the sign of the impact is unclear.

The fourth question is only relevant if the power to change interest rates is delegated to central banks. Let me assume we have a UK type situation, where the central bank has control over interest rates but it has to follow a mandate set by the government. A strong argument is that, by delegating the task of achieving that mandate to an independent institution, policy is less likely to be influenced extraneous factors (e.g. there is no way interest rates rise until after the party conference/election) and therefore policy becomes more credible. (There is a whole literature involving similar ideas.)

This advantage for monetary policy simply follows from the fact that it can be easily delegated. However even if it is not delegated, fiscal policy has the disadvantage that changes are either popular (e,g, tax cuts) or unpopular (tax rises). In contrast interest rate changes involve gains for some and losses for others. That makes politicians reluctant to take deflationary fiscal action, and too keen to take inflationary fiscal action. So even without delegation, it seems likely that interest rate changes are more likely to be used appropriately to manage demand than fiscal changes.

That problem with fiscal policy was solved by Ben Dyson, founder of Positive Money. As he explained, decisions on the size of the deficit (or surplus) can in principle easily be delegated to some sort of independent committee of economists. (At least I think Dyson was the first person to solve that problem – I may be wrong.)

That committee could perfectly well be the Bank of England Monetary Policy Committee. Moreover, decisions on the size of the deficit are increasingly being handed over to such committees the World over: for example in the UK there’s the Office for Budget Responsibility.

Note that handing the latter decision to the latter sort of committee does not, repeat not mean the committee has powers over strictly political matters, like what percentage of GDP goes to public spending and how that is split between education, health, defence and so on (as Dyson explains).

The fifth and final issue could involve many things. In basic New Keynesian models the real interest rate is the price that ensures demand is at the constant inflation level. Therefore nominal interest rates are the obvious instrument to use. Changing fiscal policy, on the other hand, creates distortions to the optimal public/private goods mix or to tax smoothing.

I’m baffled. Strikes me it is easy to implement fiscal stimulus (i.e. increase the deficit) while not altering the “public/private mix”. To illustrate if the public/private mix is 50:50, then expand public spending by the same amount as taxes are cut. Though to be realistic it’s a bit more complicated: e.g. taxpayers do not spend 100% of the amount by which their weekly income rises as a result of tax cuts. But it’s not IMPOSSIBLY difficult to get quite near to retaining the 50:50 mix. Plus I don't see why interest rate changes are guaranteed to leave the public/private mix untouched.

SW-L continues…..

So the case against fiscal policy as the main stabilisation tool outwith the lower bound might go as follows: it is slower to change and it cannot be delegated. Even if monetary policy is not delegated politicians may allow popularity issues to get in the way of effective fiscal stabilisation. While government spending changes have a certain direct effect, they are also the most difficult to implement quickly.

A potentially strong argument against monetary policy is the lower bound problem. You could argue that having monetary policy as the designated stabilisation instrument gets government out of the habit of doing fiscal stabilisation, so that when you do hit the lower bound and fiscal stabilisation is essential it does not happen. Recent experience only confirms that concern. I personally do not think mainstream macroeconomists talk enough about this problem.

The fiscal rule that Jonathan Portes and I developed, a version of which is Labour's fiscal credibility rule, does attempt to address this very issue. Switching from monetary to fiscal at the lower bound is a key part of the rule. It is also worth stressing that this rule does not prevent temporary changes in fiscal policy to counteract a downturn outwith the lower bound. (Anyone who says otherwise does not understand the rule.) For example if interest rates are already low, a fiscal expansion that is planned to last less than five years is consistent with the rule, and might be a sensible precautionary measure. (Public investment, which is outside the rule, could also be used in this way.) So Labour’s fiscal rule allows monetary policy to do its job, but fiscal policy is always there as a back up if needed.

And my final comment is that I like the fiscal rule thought up by SW-L and Portes. It’s a good idea. But as a self-confessed heterodox economist (and general oddball) I’d prefer fiscal policy to dominate, with interest rate adjustments being only used in emergencies.