金曜日, 6月 28, 2019

#20? How fiscal policy saved the world – Bill Mitchell – Modern Monetary Theory 2009/10/9



How fiscal policy saved the world – Bill Mitchell – Modern Monetary Theory



How fiscal policy saved the world

Today I read an interview with Richard Koo from the Nomura Research Institute in Japan who is the touring the world promoting his views of why the fiscal stimulus packages are so important. His views are drawn from his extensive experience of the Japanese malaise that began in the 1990s. The interview was published in the September 11 edition of welling@weeden which is a private bi-weekly emanating from the US. I cannot link to it because you have to pay to read. Anyway, much of what he says reinforces the fundamental principles of modern monetary (MMT) and is quite antagonistic to mainstream economic thinking. It is the latter which is now mounting political pressure to cut the stimulus packages. Koo thinks this would be madness, a view I concur with.

In this blog – Balance sheet recessions and democracy – I discussed the concept of a balance sheet recession. Koo was the person who coined this term. The concept is close to a modern monetary conception of a recession that begins with a financial imbalance but quickly impacts on the real economy. There are other ways for recessions to occur, but this sequence is relevant to the current episode.

To summarise, a balance sheet recession follows this trail of cause and effect:

  • When an asset price bubble bursts a large number of private-sector balance sheets have net liabilities.
  • The exposed private sector attempts to save to repair their balance sheets – debt minimisation replaces profit maximisation as the main goal. For households, spending is constrained in favour of an increasing saving ratio.
  • There is a shortage of borrowers seeking funds as the deleveraging process continues.
  • A spending (deflationary) gap emerges and output and income contracts and saving drops in total – defeating the original purpose of the thrift.
  • Private spending will not return until the balance sheets are repaired and sustainable. In lieu, public spending has to drive the economy to allow the private saving to rise and to sustain GDP growth

Koo considers that in this type of event, “the effectiveness of monetary policy goes out the window”. No-one will borrow even at low interest rates. This is what economists called a liquidity trap.

In that situation:

… if the government did nothing to counter this situation the economy would shrink very, very rapidly … the economy will be losing demand equivalent to household savings plus corporate repayment each year … But … if the government comes in and borrows the money which now is just sitting in the banking system and puts it back into the income stream through government spending, then there’s no reason for GDP to fall.

Apart from what I consider to be a misconception of causality, this is totally consistent with MMT. There is a spending gap and it has to be filled.

The misconception arises I think from the implicit assumption that the government is revenue-constrained and utilises the private savings to facilitate spending instead of the private sector drawing on them to invest. There is a two-fold problem here.

First, it implies that the banks need deposits to lend even in good times. That is certainly not the case. Loans create deposits and reserves (where needed) are added after the fact to ensure central bank accounting rules and obligations are met within the banking system.

You might like to read this blog – Money multiplier and other myths – if you are unclear about that.

Second, the government does not “borrow” savings in the sense that there is a finite pool of saving that is sitting in the banking system waiting to be loaned out and it falls on the government to borrow those funds, and, thereby get access to dollars to spend. While it might look like that is how it happens in a causal sense that would lead to incorrect inferences.

The reality is that the government spends without requiring access to prior saving. The fact that it borrows is related to the imperatives of monetary policy to drain the reserves added by the net spending to avoid losing control over its interest rate target via competition between banks to loan their excess reserves into the interbank market overnight. You might like to read this blog – Deficit spending 101 – Part 3 – if you are unclear about that.

Further, they borrow because they are ideologically obsessed with the view that “unfunded” spending will be inflationary. So various legislative and regulative restrictions (depending on which country we are talking about) are put in place – voluntarily – to create a $-for-$ match between spending and borrowing. All of which are unnecessary.

The fact is that any spending can be inflationary under certain circumstances. The so-called funding of government spending is irrelevant and doesn’t lessen the risk of inflation.

The other way of thinking about this misconception is that if the government didn’t spend, then the saving would soon dry up as the economy contracted due to aggregate demand deficiency. So the spending actually ratifies (I use the term “finances”) the non-government decision to increase its saving ratio and reduce its debt exposure.

It is much more conceptually accurate to think of the train of events as being the government spends which creates the reserves in the banking system which it then borrows back. Without that action, there is no increase in non-government saving because GDP falls sharply.

These observations should also condition the debate we had about debt-deflation recently. As I understand Steve’s position and that of a significant number of his regular commentators (who do not regularly commentate on my blog) – debt per se is the problem – private and public. But as I pointed out sometime in my replies last week – you have to be careful what you wish for.

Under current institutional arrangemets (the ideological strait-jackets that governments put themselves in), the only way that private debt can fall as saving rises is if public debt rises (as deficits increase). You simply cannot have the first without the second.

Koo, uses this logic to warn against those who are advocating a withdrawal of the fiscal stimulus. In relation to the dangers of a double-dip, he said:

… a second collapse affects everyone, not just the bubble speculators, and it also suggests to the public that all the efforts to fight the downturn to that point – all the monetary easing, the low interest rates, quantitative easing – they all failed and even fiscal policy failed. Once that kind of mind set sets in, it becomes 10 times more difficult to get the economy going again. So the fact that Larry was talking about “temporary” fiscal stimulus had me very, very worried. The whole Larry Summers idea that one big injection of fiscal stimulus will get the U.S. out of the recession and everything will be fine thereafter probably led to President Obama saying he’s going to cut his budget deficit in half in four years … [but] … fiscal stimulus will be needed for the whole period, if you want to keep GDP from falling.

He reinforces the point that with government borrowing and spending the expenditure multiplier remains positive and that is how Japan in the 1990s kept its GDP growting. The irony for Koo is that:

It’s a complete reversal of what almost everyone alive today learned in school – that monetary policy is the way to go.

The current episode has really reinforced the failure of the mainstream teaching in this regard. The central bank can pump as many reserves into the system as it likes, but this doesn’t facilitate extra borrowing (by making it easier for banks to lend) or increase the number of credit-worthy customers seeking funds.

The only solution is for fiscal policy to plug the spending gap and to allow the balance sheet repair processes to work themselves out. This is a fundamental tenet of MMT and has been ratified, in my view, without doubt by this recession. Once more data is available (longer time series) we will be able to demonstrate this even more emphatically.

He produces a graph (titled Exhibit 12 – Americans spent $1.5trn that should have been saved) which I reproduce here. It is an interesting chart.


The graph shows total savings on the LH-axis and the saving ratio on the RH-axis. The background (white bars) are what the actual savings would be if the US actual saving volume had have maintained its average rate in in the mid-1990s (4 per cent). The shortfall is clear.

Koo says the chart:

… shows just how enormous the amounts involved are: how much savings Americans have to rebuild now.

Where I think Koo misses the point is that he fails to mention the conduct of fiscal policy prior to these events occurring. If you read this blog – Some myths about modern monetary theory and its developers – you will notice that decline in saving in the US coincides with the Clinton budget surpluses. This is no coincidence.

There hasn’t been one time in US history where a period of budget surpluses hasn’t been followed by recession. If you go back to yesterday’s blog you will see the that recessions followed budget surpluses in Australia.

So I agree that the fall in the saving ratio was associated with unsustainable increases in non-government debt positions but you cannot analyse that without also considering the liquidity dynamics that were being introduced by the conduct of fiscal policy. That conjunction is a principle distinguishing feature of MMT, which should always be emphasised.

That is another reason why I think starting with a pure credit economy without a government sector in an attempt to make sense of the actual economy is not a productive way to proceed. It is better to recognise the stock-flow implications of the government’s fiscal position on the non-government sector before you analyse the dynamics within the non-government sector (which net to zero in terms of financial assets). I am not against specialist studies of these dynamics – far from it. But they have to be properly motivated in my opinion.

Koo was asked how GDP can continue to grow if the private sector are taking “massive hits to their wealth”, which is also a common theme among the debt-deflationists – that we have to take our medicine now and excroriate the debt and allow industry to suffer widespread failure as a consquence. Some people even believe you need mass unemployment to assist in this debt elimination process.

Koo doesn’t agree (and nor do I – I actually think it is a fundamental error in debt-deflationist logic):

… Japan’s GDP grew even after the massive loss of wealth it suffered when its real estate bubble burst … That happened even though the private sector was rushing to pay down debt all during that time. And the reason is that the government’s fiscal spending kept incomes growing enough that GDP never fell below its bubble peak, in either nominal or real terms …

And this was over a period that commercial real estate prices fell by 87 per cent from their 1990 peak – and the total wealth effect amounted to “the largest loss of wealth in human history, in peace time.” GDP kept growing because the Japanese government was “highly liberal with public spending”.

Moreover, he rejects the Austrian arguments in this regard (that is, a sharp recession to clear the decks and the debt) because it would be extremely costly in human and economic terms:

… that experiment was tried from 1929 to 1933 and almost half of U.S. GDP disappeared. Unemployment rate went to 25% and bringing the economy back to full employment literally too the Japanese attack on Pearl Harbour. I don’t think that’s the way we want the world to come back to life. People who argue that zombie companies should be allow to go to hell … don’t realise that, first of all, zombie companies with no cash flow cannot pay down debt, and if they cannot pay down debt, they are actually not the source of the problem … Balance sheet recessions are caused by good companies with good cash flow paying down debt.

The problem Koo notes again is that while the first wave of Japanese fiscal injection instantly worked (building roads and bridges) they considered it a temporary measure only – “just like last year Larry Summers was saying the U.S. has to do temporary pump priming”. When the government withdrew the stimulus, the economy double-dipped.

He produced a graph to show the history of fiscal policy over this period (Exhibit 16), which I reproduce next. The graph shows government spending over the period shown relative to tax revenues and the fiscal balance (deficit). 


Koo concludes that:

… what my work shows is that the total additional, or cyclical, deficit that the government created to sustain GDP during Japan’s balance sheet recession … took the fiscal deficit to about 63% of GDP. But I would argue that this 63% of GDP deficit represents the most successful fiscal stimulus in history.

His computation is based on the fact that GDP would have dived (perhaps by more than 46 per cent) if the stimulus was not provided. To be conservative, however, he assumes that GDP returns to its pre-bubble level of 1985 with no stimulus. Given the extra GDP that arose from the stimulus, he concludes that the Japanese “government bought GDP equivalent to 2,000 trillion yen with 315 trillion yen of deficit spending” (which is a “very good bargain”).

So when we are thinking we are in dire trouble with deficits around 3.5 per cent of GDP think again.

I also read during the debate last week comments like “we don’t want bridges to nowhere built” – which was aimed at my advocacy for large public infrastructure projects. This idea that government spending is wasteful is flatly rejected by Koo. He said:

… that basic assumption is just plain wrong … the lesson to be learned is that no matter how huge the asset price collapse may be, if the government takes a meaningful stimulative action from the very beginning and continues it throughout the period when balance sheets are being repaired, there is no no reason for GDP to fall .. Japan is the only country that has managed to keep growing and emerge from a balance sheet recession without fighting a war

The last point is in reference to the fact that the Great Depression really on finished with the onset of military spending (fiscal stimulus) in the late 1930s, after all the neo-classical remedies had been tried and failed.

Further, the Japanese experience supports a fundamental principle in MMT that there is no financial crisis so deep that cannot be dealt with by public spending. This was the title of a paper I wrote last year with James Juniper (the link is to the working paper version which you can get for free).

What about the question of interest rates? Doesn’t the huge deficits drive them up? 

Koo offers this explanation:

… when there is no demand for borrowing from the private sector, government borrowing for fiscal stimulus does not drive up interest rates … When these points are understood by policymakers globally, then people should feel a lot better, even with asset prices collapsing, because it is possible to keep GDP from falling, meaning national income can be maintained and – as long as people can continue to pay down debt – this problem will be over at some point.

I consider this statement to once again be misleading but perhaps that is because he is not being fulsome in his explanation. If he is referring to bond yields then I agree. Under auction type bond issuance systems, which were inspired by neo-liberal thinking that net spending had to be “funded” 100 per cent by the private markets, yields vary inversely with demand for the bonds (because of the nature of treasury debt instruments).

So if there is no demand for alternative financial assets then investors will be queuing up to buy the government debt and the yields will be lower than if public debt demand falls. But, of-course, this is not the basis of crowding out theory.

That theory is based on the notion that there is a finite supply of saving and any competition for it drives up interest rates. The addition assumption is that saving and investment are equilibrated by interest rate changes. Neither of these assumptions is applicable to a modern monetary system.

Saving is a function of income and spending thus creates its own saving. 

Koo was then asked about whether the Japanese experience generalises given that Japan had massive domestic saving … and didn’t have to depend on China or the Middle East like the US does. He replied:

The truth is that Japan was actually in that same precarious position, a decade ago. With Japanese government debt skyrocketing because of massive fiscal deficits, all of the ratings agencies, the IMF, the OECD – they all issued horrendous warnings against Japan. Japanese bond investors remember very well that JGBS were downgraded repeatedly, to the point where Japan’s debt was rated lower than that of Botswana, because the ratings agencies were so sure that at some point the whole thing would come crashing down and that interest rates would soar. But it never happened. And the reason is easy to understand, once you grasp the concept of a balance sheet recession. The amount of money that the government has to borrow and spend to sustain GDP is exactly equal to the amount of excess savings generated within the private sector of the economy. So that money is actually available within the private sector, even in the U.S., even in the U.K.

Again there is this reverse causality implicit in Koo’s reasoning (amount “the government has to borrow and spend”) and I must find out from him whether he intends this causality or whether it is just a recognition of the voluntary institutional arrangements that align net spending and borrowing.

It is clear in Japan’s case that this alignment was not 100 per cent because the Bank of Japan craftily borrowed less reserves than were created by the Ministry of Finance deficits and thus were able to keep the overnight interest rate at zero. In other words, they left excess reserves in the banking system and allowed the interbank competition to bid the rate to zero.

This is a good demonstration of another fundamental principle of MMT – if the central bank is not offering a return on overnight reserves, then it has to issue sufficient debt to drain the excess reserves created by the net spending to meet its interest rate target. If the target is zero, then they do not have to issue any debt. The fact they did still issue debt does not violate that principle.

Further, it demonstrates that the ratings agencies are irrelevant to public debt and yields. You might like to read this blog – Ratings agencies and higher interest rates for a discussion of this point.

But the really important point that Koo clearly advocates and understands is that the government deficit will always be equal $-for-$ to the non-government surplus. If the non-government sector is saving then the government sector has to be in deficit – “the amount of money that the government has to … spend to sustain GDP is exactly equal to the amount of excess savings generated within the private sector of the economy”.

If the government tries to buck against expanding and sustaining budget deficits at that level then the income adjustments that follow will eventually bring planned non-government saving and public net spending into line at much lower levels of activity and higher levels of unemployment. Total saving will also be lower. You simply cannot escape that in a fiat monetary system. So it is better to have “good” deficits than “bad” ones (the latter being forced on the economy by the automatic stabilisers driven by the income adjustments).

The question then came up about the need for fiscal policy caution – it should only be used in small doses “to avoid overdoing it”. Koo replied:

No, because of the political difficulty in trying to maintain fiscal stimulus when an economy is showing signs of life … Its especially fraught in Europe, where the Maastricht Treaty limits the fiscal deficits to a maximum of 3% of GDP. When those economies begin to show signs of life (because all of the European economies are putting in fiscal stimulus), the temptation to cut fiscal spending to reduce the budget deficits will be tremendous … some of the Ministers … [there are] … very pessimistic. Afraid that they are destined to have a second dip because they will cut the stimulus too early. It really is a global problem at this point – except in China, where they don’t worry about an opposition, because there is none.

So the political foibles of an adversarial system of government (driven by parties) where the opposition always has to criticise and can use the ignorance of the population to push neo-liberal arguments that sound reasonable but which, in fact, have no basis is truth. And governments themselves, caught by the same ideology and political constraints.

For MMT to be more widely accepted we need a new breed of politicians that can understand this stuff and explain that when the deficits are saving jobs it is because of x and y … then we might have better times.

In conclusion, the problem Koo notes is that:

… there is nothing in economic literature to suggest that a government should keep spending money even after an economy starts showing signs of life. Conventional wisdom would suggest that with the economy improving, the pump priming worked, so fiscal stimulus is no longer.

Sort of like this bear:


Note: cartoon by John Darkow.


While I do not agree with the way in which Koo constructs some aspects of the monetary system in general he fully understands the basic precepts of MMT and uses that understanding to conclude that even large fiscal stimulus interventions are justified if the non-government sector saving response is large.

He also provides good logic drawn from experience as to why the debt-deflation strategy of reducing all debt would be catastrophic.

The one weakness in his approach (which is also in his book) is that he fails to link the private balance sheets with the conduct of fiscal policy. Once you make that additional step (which really should be the first step) then you have more reason to argue against those neanderthals that are advocating cutting the deficits back now to avoid the destruction of the world as we know it.

Sorry … the Neanderthals were probably smarter than the average mainstream economist.

ABC PM tonight

I did an interview for the national ABC PM program tonight – Transcript – mp3

The interview was about yesterday’s Labour Force data and the RBA decision to hike rates so early. Readers of the blog will already know what I would have said.


Saturday Quiz – more difficult than ever.