Monday, 27 April 2015

High Unemployment: A Failure of the Economic System or Unavoidable Phenomenon?

Use of the Flows model to show why, in my view, unemployment is a failure of our economic system

I think that we have all grown accustomed to the idea of economic cycles. Sometimes we have boom and the economy does well. Sometimes things are bad and we have low growth and unemployment. People also tend to assume that this is the natural way of things and is unavoidable.

I intend to argue here that this is not the case. I argue that unemployment, apart from in exceptional cases, such as skills mismatches due to declines in certain industries, is due to a failure of our economic system and I will also argue that this can be remedied by improving the system.

A simple economic system:

I will start with a simple economic system. It can be argued that this is not realistic but I feel that it shows enough about the real world to be a good demonstration of my point.

Imagine an economy with 1000 people. 1 person is working, 999 are unemployed. The 1 person is producing 1,000 units of food per year. A social security system is in place that unemployed people get half of the share they would get if they were working.

So here, tax is almost 50% on the one worker. Everyone gets half a unit of food, except the worker who has 500.5 units of food, doesn't know what to do with it all, and spends his spare time feeding the ducks trying to get rid of it. 

Now someone else comes in producing 1000 units of shoes. The total tax rate is now a little lower. The unemployed people have half a unit of shoes and half a unit of food. It's not that great for them as they have to wait until next year for the other shoe, but still it is clearly better. It is also better for the food-maker who now has 501 units of food and can swap some of them for shoes. And better for the shoemaker who has 501 units of shoes and can now exchange some of them for lots of food.

This system builds up until everyone has a job. Every worker added increases total production and improves the situation for everyone. On average each worker has 1,000 units of product each per year. In total now, 1 million units are produced per year. A system of money is brought in to facilitate the exchange of goods. We will call this unit of currency a Worldo (as the system represents the whole world). Assume that at the start 1 million Worldos of Central Bank produced cash exist and 1 million are spent per year.

It is not that all units are equal. 1 unit of lawyer service may be exchangeable for 10 units of cleaning services. This is a fully capitalistic system, responding to supply and demand. If no kids want Etch A Sketches for Christmas any more then the Etch A Sketch maker must retrain in making Nintendo Gameboys and this will lead to temporary loss of productive capacity. But overall, full employment is possible and desirable.

But as yet, we have not got any concept of savings. Here I need to divide up the idea of Savings into two strands: a) investment that will increase future productivity and b) savings that do not affect future productivity. 

An investment causes a loss of productivity in the short term. Supposing one worker in the system decides to spend a year building a new factory that could produce 1,200 units of trousers per year instead of the current 1,000. One year's productivity is lost (1,000 units) but the future capacity for making trousers is increased and the total productivity of the whole system would after this be 1,000,200 units of goods and services. This is good for everyone, especially the trouser maker.

An economic decline

But instead let's suppose that in one year, 100 people in this system decide to save 10% of their income (100 units each) - and they do this by putting it under their mattress. Still, in this system, no lending is allowed. Now, instead of a million Worldos being spent, only 990,000 Worldos are spent. In theory, this should not be a problem as the price of the average unit of goods should now go down from 1 Worldo to 0.99 Worldos. In practice, however, due to the concept of 'wage stickiness' this is not the case. Wage stickiness is the idea (empirically shown to be true) that people are very resistant to a cut in their wages. And employers are very reluctant to cut wages in nominal terms. Therefore the process of adapting to the new amount of currency in circulation is slow.

We can see the result using the Flows model where α is the coefficient of spending, d is total dividends (including rentals), i is interest, w is wages, L is new loans created, C is new central bank cash, ES are existing savings and γ is the proportion of existing savings spent in the year:

GDP= αd*d + αi*i + αw*w + αL*L + αC*C + γ*ES    (1) 

Before, d=0, i = 0, αw = 1, L=0, C=0 and ES=0, giving:

GDP= w = 1,000,000 Worldos

Moving on a year, d=0, i = 0, α= 0.99, L=0, C=0 and ES=0, giving:

GDPt+1 = 0.99*w = 990,000 Worldos

This gives Unspent Income of:

UI = 10,000 Worldos

The effect of the reduction of money spent in the system is that there is only enough money to buy 990,000 units. On average 10 people need to be made unemployed. Maybe the government decides that in order to make up the shortfall, it will reduce taxes and cut the least necessary parts of the economy; maybe, for example, the library service and the youth club service workers. 

The 10 people on social security now get a half share (495 units each) so the total for everyone else has now gone down to 985,050 units. So everyone who is working now get an average of 995 units each. The library and youth club workers have now had a cut in income of over half - so they are worse off. But everyone else has also lost out - they have have lost out on libraries and youth clubs for their kids. Maybe the richer ones don't use libraries and youth clubs, but these also may help future economic productive capacity which does help them.

So, everyone is somewhat worse off after these savings, except arguably the people who saved. Let's show this by rolling forward another year on the Flows model with 1% savings level. I will also assume a 10% spending of savings by the savers (so γ  = 0.1, ES is now 10,000):

GDPt+2 = 0.99*w + 0.1*ES = 0.99*990,000  + 0.1*10,000
= 981,100 Worldos

More people will now need to be made unemployed.

By repeating this process, the economy will have a stable savings rate at approx 909,000 Worldos. At this level, the savings made each year are equivalent to the amount of savings spent by exisiting savers.  We are in equilibrium as there are now no net savers. 

GDPinfinity = 909,091 Worldos

If there has been no deflation, then we will now have 9% unemployment. Eventually the deflation will be absorbed by the economy and it will return to full employment. The average price per unit will be 0.909 Worldos per unit and everyone will be happy again. 

But there was a lot of pain on the way there and a lot of lost production. All just caused by an increase in savings rate. 

The main point is that everyone is better off with full employment and yet this system does not provide it. This is a failure of the economic system - the result is clearly sub-optimal.

An economic boom

We can also use the Flows model to model the equivalent, which is people spending more than they are saving. Here, the results are more straightforward. There is more money being spent in the system but, because we are already at full employment, there is no increase in productivity. GDP goes up but it is all inflation. Wages go back up again, prices go up, employment levels and quality of life remain the same. 

Note that the boom and bust are not symmetrically good and bad. The bust is much worse than the boom because in the bust productivity is lost, but in the boom none is gained. Also, the boom leads to inflation which means that next time people decide to save, we will have more deflation and another bust.

How confidence cycles fit in here

I have not included credit in the simple above system, but credit can make the swings even larger. If money is borrowed because of confidence during good times, then forced to be repaid during bad times this cycle is only accentuated. And confidence is an important part of this system. Post-Keynesians are completely correct to stress the importance of what Keynes called 'animal spirits'. If one is more confident about the future, it makes sense to spend more. If one is less confident, one should save more. This is Keynes' 'paradox of thrift'. It makes perfect sense to an individual to save more in an economic downturn, but if everyone were to reduce spending with declining confidence and vice versa (ie. make the α of spending a function of confidence with negative sign),  one can see that GDP in the model suffers from positive feedback in both directions.

How the system is currently controlled

Stabilisers on this system are needed. Typically, in times of downturn, governments will borrow and spend money when the private sector is retrenching. Interest rates are reduced in bad times to encourage more borrowing and spending. Likewise governments should reduce their spending during booms (this is not always done) and increase interest rates.

In this system, looking back to Equation 1, a reduction in wages (w) is compensated for by an increase in new loans, L, both to the government and the private sector.

For the last 35 years, monetary policy has been the main tool for dealing with downturns. Not enough spending in the economy has been balanced out by continually increasing private sector borrowing. Finally, the zero lower bound has been hit, the economy can not be stimulated any longer, and the true cost of the debt overhang is finally being felt.

There is no room left for conventional monetary policy to regulate these cycles as the debt is too high.

More or less everything that I have said so far in this article is straight from Keynes. Keynes viewed government spending as the preferred tool for regulating the cycle. When confidence is high and more money is being spent, the government should withdraw money. When a consensus of gloom appears, it is up to the government to spend more money to balance the reduction of spending from the private sector equation and keep GDP up.

Whilst Keynes was absolutely correct (as he usually was), there is a major problem with this response now. The sheer level of private sector debt is too high. The amount of money lost to the economy through interest payments and dividends is too high and the government's burden is becoming too large.

What Keynes could not have accounted for was the huge savings glut that we have currently seen. Whilst he saw the government counteracting the savings with spending, he was not talking in a world where the savings are so high. 

If the government were to take on the debt necessary to counteract private sector savings then it would sink under the weight of its own debt.

So what is the solution?

Let's look at equation 1 again:

GDP= αd*d + αi*i + αw*w + αL*L + αC*C + γ*ES

To recap, the problem here is that the total of αd*d + αi*i + αw*w + αL*L + αC*C + γ*ES  is not high enough to allow GDP growth. 

In the past, monetary and fiscal policy would counteract this with an increase in new loans, L, but this is no longer possible or desirable. One can try to find methods to increase α and γ, but this is difficult. 

The other option here, and if you have read any of my previous posts, this will come as no surprise, is to increase C - new central bank cash.

C should be positive when the economy needs more money to operate at full employment, and negative when it needs less. This alone, if implemented correctly, can keep the economy at full employment and without the boom and bust cycle?

New cash being printed by the government should really not be a taboo issue. It is actually a far better method of deciding how much money is in circulation than leaving it to private banks.

How could it be implemented?

This is the tricky part. In sustained slump, investment projects should be undertaken. One could print money to fund one-off research projects, infrastructure investment and educational projects which would help future growth. On a fine-tuning basis, it is a little harder.

Tax cuts funded by newly printed money are one option. Similarly taxes could rise in a boom when people have more money. VAT could be altered to go up and down with the cycle, as could corporation taxes.

We currently in the Western world have a sustained deficit of demand which, without government deficits and rising asset prices would, I think result in no growth at all. The situation at the moment is so dire that it would not be ridiculous just to give every citizen of the country, say, £1,000 a year and let them spend it on whatever they want. This could be stopped when a more stable equilibrium is reached.

When would this not work?

One of the only ways to fail if using this framework (short of war, natural disaster or other major economic failing) is to have your debt denominated in another currency. My feeling is that there should be an international framework which rules that if any president of any country decides to borrow in a foreign currency they must, by law, be punched in the nose by their assistant and told to stop being so stupid. Next time they propose it, the same again. And again. Until so battered by punches and verbal abuse they will give up. 

Although the interest rates may be lower, it is a completely obvious way of destroying your economy when times turn bad. By printing money to service the foreign debt, the exchange rate goes down and you just have to print more. The end result is hyper-inflation. 

If the local economy takes a downturn, the cost, domestically, of servicing the foreign currency debt does not go down with the earnings of the people of the country. The result is a higher and higher debt burden that eventually will lead to your citizens paying more than they can afford and domestic GDP suffering.

If only the periphery members of the Eurozone had thought of this before joining the Euro.


Some unemployment is a fact of life. When a pit closes, miners need to find another job. Some people don't want to work and they will be happy on unemployment benefit. 

But unemployment caused by economic cycles is a sign that the system is failing. It is always better for people to be working rather than unemployed.

This framework shows how the government can intervene in the system by adding or taking away freshly printed money in order to keep GDP at the desired level. It can stop booms and avert busts using just this tool.

Application is the difficult part, but the most important stage is to get out of the mindset that the government should not just print and spend money.


  1. Great post. Don't you think a good way of implementing this automatic stabilizer would be the MMT idea of maintaining an employment buffer stock?

    That would deal with unemployment directly AND ensure that the central bank would act countercyclically as you suggest.

    1. Very interesting suggestion - a natural stabiliser.

      I see an issue being the distorting effect it would have on the natural labour market. Especially if the living wage suggested as payment is higher than the minimum wage. Many people might use the service out of choice rather than necessity. And once in, would they then be able to not bother working too hard as they couldn't be fired?

      Maybe these are details that would be ironed out by small scale trials.

    2. When MMT pages on the internet claim the government deficit = private sector savings rate (a flow) they are only counting currency flow. They are not counting real things. Real things and labor goes in the opposite direction of the printed money and credit money!

      Real savings can include real things. For instance you pack food away for next year is savings but you have less cash than if you put just the money away. But, that day you have saved the same amount measured in the currency (unit of account).

      My opinion is MMT on the internet is bunk, and leaves out real costs.

  2. How does this differ from Japanese policy over the last 20 years, and Abenomics recently? Attempt to answer own question: Japan has been debt funding deficits not printing money, and as the debt is mostly domestic, the investment of domestic savings into bonds has negated the deficit spending. Is that why it hasn't worked for Japan? What about their current money-printing?

  3. Nikdow, first my general thoughts on QE. It is something of a confidence trick. It does just about work, but more for second order reasons.

    Inflation is caused when the government borrows and spends money it doesn't have. This is the time that money is created. Exchanging one form of currency (Yen) for effectively another form (JGBs) does nothing in itself.

    The reason it works is that it increases asset prices, and, if you look at the flow model, if gamma stays constant, this will increase spending. It also works by weakening the currency in that it sucks demand away from other countries. But actually the amount of weakening of the currency appears to me to be a huge market overreaction. In Japan the currency has weakened 20% more than the change in interest rate and inflation would suggest it should.

    The way to make QE work with a first order effect is to write off the government debt and allow fiscal spending to compensate.

    Back to the last 20 years; I am not an expert on Japan, but at the end of the 1980s boom, there was a huge debt overhang like we have in the U.K. and the U.S. now. Debts were not written off and the cost of servicing this debt (due to alpha of interest and dividends being lower than that of wages) put a large strain on growth.

    The government responded correctly (conventionally speaking) by increasing spending. This kept gdp per capita growing reasonably well, but the situation that we have now is that there is not much room for any more fiscal spending.

    I hear that they are still trying to increase v.a.t. in Japan and the attempt to reduce the government debt may well quickly negate the one off positive impact of QE.

  4. I assume that by "writing off the debt" you don't mean cancelling the bonds? That would cause a revolution overnight. Do you mean letting the banks default and writing off their debt? Or do you mean printing money to buy back the bonds?

    1. Yes, I just mean buying the bonds with new money and canceling them. Allowing the government to spend more.

      We would not want rioting Japanese pensioners.


Sorry, I have had to moderate comments because of an annoying pest spammer who keeps posting American football matches links.