SENATE ECONOMICS
REFERENCES COMMITTEE
Friday, 9 October 2009
HENRY, Dr Ken, Secretary, Department of the Treasury
GRUEN, Dr David, Executive Director, Macroeconomic Group, Department of the Treasury
Senator JOYCE—Dr Henry, I am going to go through about seven questions and then I will listen to what
you have to say, so if we can just bear that in mind in our answers. I apologise for not being there. Obviously I
am reducing my carbon footprint by trying to do as much of this as I can by phone.
When the Australian dollar was floated in 1983 by the Hawke and Keating government, one reason for the
float was a recognition that a fixed exchange rate made monetary policy ineffective and that the increasing
openness of the Australian economy to international trade made fiscal policy ineffective. Thus the Hawke-
Keating government realised that Australia could only gain effective macroeconomic stabilisation through a
float of the dollar and by placing primary emphasis on the use of monetary policy for macroeconomic
stabilisation. Recognition of these facts has bipartisan political support, and had it until late 2008. So my
question to you, Dr Henry, is: does the Treasury and the Reserve Bank agree that the assignment of fiscal
policy to medium- and long-term goals and monetary policy to short-term stabilisation has served Australia
well for a quarter of a century and, if you do believe that, what justifies the abandonment in 2008-09 of this
policy assignment?
Dr Henry—I do broadly agree with that proposition. I do not consider that that policy assignment, as you
refer to it, has been abandoned in the present period. The consensus that emerged in the 1980s, and perhaps
more particularly following the recession of the early 1990s, was that there were considerable limitations in
using fiscal policy to finetune the economy. Whilst there are considerable limitations also in using monetary
policy to finetune aggregate demand and to finetune the economy, monetary policy is probably better able to
play that role.
But certainly I never interpreted that consensus as implying that, when an economy is hit with an
unprecedented negative shock, one should expect that monetary policy could play all of the role. Indeed,
notwithstanding the consensus that you referred to, this is not the first occasion in the last 10 years on which
fiscal policy has been adjusted in order to support aggregate demand. Senators might recall that in late 2000
there was an adjustment to fiscal policy settings, with a doubling, as I recall, of the first home owners grant
precisely for that purpose. At that time, the economy contracted for one quarter and the policy response was an
expansionary fiscal policy.
Senator JOYCE—So your answer is that this was an exception to the rule.
Dr Henry—That is probably not an unfair way to characterise it. The proposition that I am putting is that,
notwithstanding that rule, there will be exceptional circumstances that warrant fiscal policy adjustment to
assist a monetary policy adjustment.
Senator JOYCE—In the Asian crisis and 2001 US recession, much more reliance was placed on allowing
the exchange rate and monetary policy to play a major role in short-run macroeconomic stabilisation, whereas
fiscal policy in that instance played a minor role. In contrast, fiscal policy has played a major role in the
government’s current Australian macroeconomic policy. What explains the difference in the choice of
macroeconomic policies? Were mistakes made in the Australian macroeconomic management of the Asian
crisis and 2001 downturn? If so, where are those mistakes documented and how did the lessons learnt justify
the greater reliance on fiscal policy?
Dr Henry—No, I do not consider that mistakes have been made on any of those three occasions. I think
what each of those three occasions illustrates is that the appropriate policy response will depend upon the
particular circumstances that the economy is confronting. In fact, my recollection of the period of the Asian
financial crisis is that the exchange rate, indeed, did adjust and adjusted quite sharply and that monetary policy
did not adjust—not immediately, anyway. Neither, of course, did fiscal policy. The exchange rate adjustment
provided most of the cushioning of the economy that was required. In the second occasion to which you refer,
it is true that there was a significant monetary policy adjustment—although, as I indicated earlier, we had
ourselves had an expansionary fiscal policy in late 2000. So I do not think that it is true that in that period of
weakness in the US economy the Australian policy response was left entirely to monetary policy. It certainly
was not. In the present circumstances, where we have been hit with a quite unprecedented negative external shock and a shock that has affected not just aggregate demand directly but also the functioning of credit markets, on this occasion the Australian government—and, indeed, just about all governments around the world—have judged it appropriate for there to be both expansionary monetary policy and expansionary fiscal policy.
Senator JOYCE—I will go into why I believe that they have a different view, and that is their latent
monetary policy settings and interest rates. But there is considerable disagreement in the empirical literature
about the magnitude and even the sign of the response of output to a fiscal stimulus, but a consistent empirical
finding is that the fiscal multiplier is zero for open economies, which I believe Australia is, with flexible
exchange rates. This is suggested by the Mundell-Fleming model. The most recent research that I have found
confirming this empirical finding is in Ilzetzki, Mendoza and Vegh, who in the CEPR Policy insight No. 39
from October 2009 summarised the results of a study of 45 countries—20 high-income and 25 developing.
How does the Treasury square its views on the effectiveness of fiscal policy away with this research?
Dr Henry—I have not had an opportunity to read the paper to which you refer. I am certainly very familiar
with the Mundell-Fleming model as a conceptual framework. My colleague, Dr Gruen, who is sitting right
beside me, has had an opportunity to read that paper—indeed, he has a copy of it in front of him—so I refer
that question to him.
Dr Gruen—Senator, you are right about this recent paper by those three authors you talked about, which
was published by the Centre for Economic Policy Research, I think, just this month. As you say, they have got
a large sample of countries and they are looking at how big the fiscal multiplier is. They make a series of
interesting and relevant points and the one that you drew attention to was that, on average, it is the case that
the fiscal multiplier for countries with fixed exchange rates tends to be significantly higher than the fiscal
multiplier for countries with flexible exchange rates.
Senator JOYCE—That is correct.
Dr Gruen—That is what they conclude. If you read on in that paper, they also come to a further conclusion
that is relevant to Australia. They break their sample into those countries that have a high trade share and those
countries that have a low trade share, and the criteria they use is that your exports plus your imports have to be
greater than 60 per cent. We are very clearly in the category of countries with a low trade share. Our trade
share—exports plus imports—is about 40 per cent. The point they make in the figure immediately following
the figure you reported is that, when you look at countries with relatively small trade shares, you find that they
have significant fiscal multipliers. If you are looking at countries with relatively small trade shares like
Australia, the fiscal multiplier is indeed substantial and the numbers that they quote are larger than the
numbers that we have assumed in the fiscal multipliers that we have applied for the Australian fiscal
expansion. So if you took their results literally, you might well conclude that the fiscal expansions that have
been applied in Australia would have a bigger effect than the Treasury has assumed.
Let me make another point that is relevant to the current situation. The empirical analysis that this paper
presents is over an extended period for a whole lot of different countries, and many of the downturns in which
there has been a fiscal response that they are analysing would have been country specific rather than global.
Why is that relevant? It is relevant because if there is a global downturn and everyone responds with fiscal
policy then the exchange rate effect is much less important because everyone is stimulating their economies. If
you like, you can think of the whole globe as a closed economy and in that world the relevant multipliers are
the ones for a fixed exchange rate because the whole world is providing fiscal stimulus. So the results of this
paper are certainly relevant but, when you have a global downturn or a global negative shock to which just
about every country is responding with large fiscal packages, the relevant multipliers are higher because the
spillovers work in both directions.
Senator JOYCE—Are you saying there that the actions of Australia are more the actions of a good global
citizen, rather than specific actions based on a domestic economic policy? In effect, we are really riding on the
coat-tails of the global effect of stimulatory packages to which we contributed a very minor part. With the
premise of an open economy and floating exchange rates, we had the capacity to lower monetary policy. We
had some reserves in where our interest rate was; we were not at near zero levels like other economies were.
Dr Gruen—I think the governor made it clear in his testimony that cutting interest rates to very low levels
is something that you do if there is no alternative, but that there are attendant dangers in having extremely low
interest rates, certainly for any extended period of time. So it is not as though cutting interest rates is
completely costless and in all circumstances the only sensible response. After all, one of the arguments
presented about what led to this global mess was an extended period of ultra-low interest rates in the US. We
can debate how big a contributing factor that was, but it is certainly the case that a considerable body of
opinion suggests that ultra-low interest rates in the U.S in the early 2000s were a significant contributing factor
in the housing price bubble that developed in the US. So the idea that cutting interest rates further and for
longer only has benefits is, I think, an argument that many people, including the Governor of the Reserve
Bank, would not agree with.
Senator JOYCE—I may come back to that later if I have time. I will move on to another thing. I do not
know if this question is to Dr Gruen or Dr Henry. In the 2009 budget papers there are several dramatic graphs
that show the projected path of the economy with and without a fiscal stimulus. How were those projections
obtained? Were they obtained from a model? If so, what was the model? If a model was not used, what
research based evidence was used to develop the projections? How much uncertainty is in those projections?
What were the standard errors that Treasury would put on those projections?
Senator CAMERON—Senator Joyce, can you point us to where that is—
Senator JOYCE—You would have to have a copy of the budget papers.
Dr Gruen—I have a copy of the budget. The GDP graph is on page 48. There is also an unemployment
graph and a consumption graph, but I guess the GDP graph is—
Senator JOYCE—The one we are referring to.
Dr Gruen—We have provided significant detail on how these results were derived. It is in the paper that we
prepared for the Senate inquiry and tabled a week ago. There is detailed description in there about how these
estimates were derived. We used what we thought were reasonable fiscal multipliers and what we judged to be
reasonable estimates of the time profile over which the spending would impact on the economy. Those
estimates were informed, as Dr Henry said earlier, by a range of studies—by the IMF, the OECD and a range
of academics—on the question of how big fiscal multipliers are.
Senator JOYCE—I am just trying to drill down to exactly what model you used.
Dr Gruen—We use a large number of models of different parts of the economy: consumption, employment
and various other equations. We also have an economy-wide model which we use. At the end of the day, we
take all of these things as inputs and apply judgment, informed by the evidence that we and others collect that
we think is relevant. The crucial input into these calculations is one’s estimate of the fiscal multipliers. We
have provided a table in that paper that we presented to—
Senator JOYCE—I saw the multipliers. So you used a range of models?
Dr Gruen—We certainly use a series of different econometrically estimated equations and, as I said, an
economy-wide model. It is certainly the case that the estimates that are presented in the budget are point
estimates. They do not have associated standard errors, but there would certainly be—
Senator JOYCE—Why don’t they have standard errors? The other question is about the transparency of
your standard error test, but you do not have standard errors?
Dr Gruen—There are a large number of different studies which generate fiscal multipliers. It is not simply
a single statistical test out of which comes a standard error. We can have a discussion about the range of
plausible multipliers for Australia. We have provided different people’s estimates of what—
Senator JOYCE—The Treasurer seems to quote them with great aplomb whenever he gets a chance. I just
want to know exactly how many models were used in obtaining these.
Dr Gruen—I think I have said as much as I can say about that. As I say, we use a large number of single
equation models for different pieces of the economy. We run our own macromodel. We look at the output of
other people who run macromodels and we also look at what researchers and international institutions, such as
the IMF and the OECD, have to say about what they think are the magnitude of fiscal multipliers.
Senator JOYCE—Where do you get your research based evidence from?
Dr Gruen—A range of people are doing this sort of work, including the paper that you quoted to us earlier.
Senator JOYCE—Did you use the CEPR policy insight? It came out in 2009.
Dr Gruen—It only came out this month. Clearly, we did not use papers that had not been published.
Senator JOYCE—So which papers did you use?
Dr Gruen—I have a folder full of papers here. There are very large numbers of papers on fiscal multipliers.
I have 10 or 15.
Senator JOYCE—Which is the predominant one or do they all have equal weight?
Dr Gruen—No. One of the more important pieces of evidence that we used was a couple of academic
studies of the consumption response to the 2001 US tax rebates.
Senator JOYCE—Can you tell me those?
Dr Gruen—Certainly. There is a paper by David Johnson, Jonathan Parker and Nicholas Soules, which was
published in the American Economic Review in the mid-2000s. These researchers added questions to the
consumer expenditure survey in the United States to assess the change in consumption expenditures caused by
the 2001 federal tax income rebate, to see whether the permanent income hypothesis, which would suggest
that very little of that money would be spent, was indeed borne out by the evidence. The amounts of money
that were handed out to individuals in those tax rebates was, I think, US$300 or US$600, so they were of the
same magnitude as the cash payments in Australia. Cash payments in Australia for many households were
slightly more than that. The US handed out money of that magnitude and these researchers came to the
conclusion that households spent between 20 and 40 per cent of their rebates on non-durable goods during the
three-month period in which their rebates arrived. Roughly two-thirds were cumulative during this period and
the subsequent three-month period. That study was published in one of the most prestigious economics
journals.
Senator JOYCE—I will go back and grab the paper and trawl through it.
Dr Gruen—Just let me make one further point. It was on the basis of those estimates, which are similar to
other estimates from similar studies undertaken by people, that we calibrated our estimates of how much of the
December and April cash payments would be spent by consumers in the months following those rebates. We
looked carefully at the evidence that we thought was relevant and we used it to the extent that we could.
Senator JOYCE—Maybe I will pick up with you again during Senate estimates. In evidence the Governor
of the Reserve Bank said, ‘In setting monetary policy, the RBA takes fiscal policy as given and then
determines the monetary policy that is appropriate, conditional on that monetary policy.’ Does the Treasury
agree with this assessment? If so, suppose that the government had chosen a smaller fiscal stimulus; is the
Treasury’s assessment that this would have caused the RBA to provide a larger monetary stimulus in reaction
to the crisis? If there had been a smaller fiscal stimulus and a larger monetary stimulus, is the Treasury view
that the nominal exchange rate would have depreciated by more than it did? If so, would that depreciation
have stimulated the economy and why was it necessary to adopt a fiscal stimulus that was so large as to limit
the action by the Reserve Bank of Australia?
Dr Henry—I probably should take that question. As you know, although I sit on the Reserve Bank board
and therefore have an unusual insight into the question you are asking, my practice has been to allow the
governor to speak for the Reserve Bank board. Indeed, given the independence that the Reserve Bank has in
the operation of monetary policy, I have always considered it inappropriate for me to venture an opinion—
publicly, anyway—on where monetary policy settings would be had other things played out differently. But in
respect of the question you have asked the governor has, indeed, already answered it in two respects. Firstly,
he does not consider that the present stance on fiscal policy has created difficulties for the operation of
monetary policy. Secondly, he has indicated publicly that interest rate settings were taken below the point that
one might judge to be entirely consistent with everything else that was going on in the economy, because there
was a level of insurance built into the monetary policy settings in case events turned out to be not as
anticipated but actually worse than anticipated.
Senator JOYCE—Knowing you are on the Reserve Bank board, when you said, ‘I probably should take
that question,’ I should have said, ‘No, you can’t; I want someone else to answer it.’ But I will continue on.
CHAIR—Senator Joyce, we put you on now because you are going to a meeting, but you have had almost
half an hour.
Senator JOYCE—Can I place any questions I have on notice? If Treasury agrees with the mainstream
view that fiscal stimulus pushes up the exchange rate while monetary stimulus lowers the exchange rate, then
it must be the case that current mixture of monetary and fiscal expansions are working in opposite directions
on the exchange rate. What then justifies the rhetoric that monetary and fiscal policy are supporting each
other?
Dr Henry—I think Dr Gruen has already answered this question, at least, in part. It is not at all clear to us
in present circumstances, and I covered this in my opening remarks. The present level of the exchange rate can
be explained, in large part, by domestic policy settings, either monetary policy or, for that matter, fiscal policy.
Dr Gruen also indicated that whilst in rather simple conceptual models of a small open economy a fiscal
expansion in one economy, when the rest of the world is not engaging in a fiscal expansion, might be expected
to lead to an appreciation of the currency—and that is a standard Mundell-Fleming result, which you referred
to earlier—that same result is not nearly as strong when the small open economy is engaging in a fiscal
expansion in concert with fiscal expansions in the rest of the world.
Senator JOYCE—This is my final question. Assume that Treasury would agree—you can tell me if you do
not—that all taxation involves distortion to price signals that impose a cost to the community. From my
reading, a rough rule of thumb used by economists—and I imagine this would include you—is that the
magnitude of the welfare cost of taxation is proportional to the square of the tax rate. Given that the share of
tax revenue in GDP is 22.5 per cent, or thereabouts, and the maximum tax rate on labour income is 47.5c in
the dollar, the deadweight loss of each dollar raised by taxation revenue is between 5c and 22.5c in the dollar.
So the deadweight loss associated with the fiscal stimulus is between $2.4 billion and $10.7 billion. Could
Treasury provide a more precise estimate of the deadweight loss incurred in raising $47 billion?
Dr Henry—I will make one comment on that. I am not sure that the argument you are running is an
argument against the use of fiscal policy. It sounds more like an argument against the existence of government.
Senator JOYCE—I am all for removing a tier of government, Dr Henry, as you well know. I have written a
paper on it.
Dr Henry—Yes, I am aware of that, Senator. But it is actually a serious question, because it is the case that,
generally, taxes do distort economic activity. Those tax driven distortions either reduce the size of the
economy below the size that it would otherwise be or in other ways impose welfare losses on citizens. That is
generally the case. It is certainly not the case always, and as you would be aware, Senator, it is not the case if
the tax is being used, for example, to correct for an environmental externality—and there are a large number of
other interesting cases in which taxes that raise revenue actually turn out to be welfare enhancing rather than
welfare detracting. But as a general point it is true. What is also true—and this is the reason why people get
interested in tax policy—is that different taxes have different effects. For that reason, it is simply impossible to
answer the question you have raised without knowing precisely which tax rates are going to go up and which
tax rates might even possibly come down.