Reconciling Krugman and Keen using nef model

Reconciling Krugman and Keen

Emanuele Campiglio

Giovanni Bernardo

New Economics FoundationLondon

23/07/2012

An analysis of the relation between debt and aggregate demand using nef model

Outline

1. Introduction: the Keen – Krugman debate

2. The relation between debt and aggregate demand

3. Our contribution to the debate

4. The theoretical model

5. Scenario simulations

6. A step further: the supply side

7. Conclusions

Emanuele Campiglio – Giovanni Bernardo – New Economics Foundation

Introduction: the two contenders

Paul Krugman, everybody knows him. Nobel Prize foreconomics in 2008 for his work on trade andinternational economics. Prolific politicalcommentator, he writes a blog in the New York Times(link).

Steve Keen, professor in economics and author of“Debunking Economics”. Keen is developing a formal analysisof the intuitions of American economist Hyman Minsky (1919-1996), focusing on the role of debt in macroeconomicdynamics. A quite recent version of the model can bedownloaded here.


http://krugman.blogs.nytimes.com/

http://debunkingeconomics.com/



http://www.debtdeflation.com/blogs/qed/

The bone of contention

• In March/April 2012 the two have been the protagonists ofa heated debate on money, debt and aggregate demand.

• All was sparked by a paper prepared by Steve Keen for theINET conference in Berlin, in which he asserted thatchanges in the levels of debt add to the economy’saggregate demand.

• Krugman disagreed, and replied with a post on his NYTblog. Keen replied to Krugman, and so on. The feud wenton for a while, involving other economists and generating awide debate on the blogosphere.

• A (slightly pro-Keen) chronological summary can be foundat this link.

Emanuele Campiglio – Giovanni Bernardo – New Economics FoundationEmanuele Campiglio – Giovanni Bernardo – New Economics Foundation

http://ineteconomics.org/sites/inet.civicactions.net/files/keen-steve-berlin-paper.pdf



http://unlearningeconomics.wordpress.com/2012/04/03/the-keenkrugman-debate-a-summary/



Debt and aggregate demand

• What is Keen’s idea exactly, and why has it generated such a passionate reaction fromKrugman?

• A formal explanation by Keen can be found here, but in a nutshell, what Keen argues is thatthe aggregate demand in an economy can be different from income, and that the differencebuilds up to the net change in the level of debt.

• In other words:

Aggregate Demand = Income + Net change in debt

AD=Y+ΔD

• This sounds very different from what Krugman so often repeats: “Your spending is myincome, and my spending is your income”.

• Although both Keen and Krugman can be considered as “demand-side economists”, theirvisions are strongly different when it comes to debt and how this affects the economicdynamics. As Krugman himself wrote in one of his blogs:


“Keen then goes on to assert that lending is, by definition (at least as I understand it), anaddition to aggregate demand. I guess I don’t get that at all. (..) I think it has something todo with the notion that creating money = creating demand, but again that isn’t right in anymodel I understand.”

http://www.debtdeflation.com/blogs/2012/07/03/european-disunion-and-endogenous-money/

http://krugman.blogs.nytimes.com/2012/03/27/minksy-and-methodology-wonkish/







Our contribution (1)

• In this presentation we use nef macroeconomic model to clarify Keen’sanalysis. We do so by explicitly identifying two different variables:1. Realized expenditures (ER);2. Planned (or desired) expenditures (EP).

• We show how the main difference between him and Krugman (and themain source of confusion in the debate) is the way they think of aggregatedemand:– Krugman identifies Aggregate Demand with Realized Expenditure;– Keen instead defines Aggregate Demand with Planned Expenditure.

• Using numerical simulations, we show that, both Keen and Krugman areright in their own logic:– Realized Expenditure (Krugman’s Aggregate Demand) is indeed equal to

Aggregate Income. That is equivalent to say that at the end of each periodrecorded income will be equal to recorded expenditure. What has beenactually spent will result to be someone else’s income.

– Planned Expenditure (Keen’s Aggregate Demand) can instead be differentfrom Aggregate Income and the discrepancy is, as Keen argues, equal to thenet change in the levels of debt.


What is demand?• The question then becomes:

What is the right definition of Aggregate Demand?

• We tend to agree with Keen on this: planned expenditures (what firms,households and governments plan today to spend tomorrow) are thecrucial variable when defining the aggregate demand of an economy.

• As Hyman Minsky put it:

“For real aggregate demand to be increasing,(..) it is necessary that current spending plans,summed over all sectors, be greater thancurrent received income and that somemarket technique exist by which aggregatespending in excess of aggregate anticipatedincome can be financed.”

(Minsky, H. P., 1982, Can "it" happen again? : essays on instability and finance. Armonk, N.Y., M.E. Sharpe)


Banks and credit

• Who finances the gap between the current spending plans and current income? Banks.• The private banking system is responsible for the creation of the overwhelming majority of the

money supply in circulation.

• “By far the largest role in creating broad money is played by the banking sector.. Whenbanks make loans they create additional deposits for those that have borrowed.”(Bank ofEngland, 2007)


Ryan-Collins et al. (2011) Where does money come from?, New Economics FoundationBerry et al. (2007) Interpreting movements in Broad Money, Bank of England Quarterly Bulletin 2007 Q3

The role of banking

• Krugman and Keen differ in the way they look at debt and banking.• As far as we understand, Krugman looks at banks essentially as intermediaries,

moving money from where there is an excess of savings to where credit is needed.• This is a crucial role of banks, but that’s not all: banks are able to indipendently

create new credit and allocate it in the economy, thus significantly influencing itsshape.


Ryan-Collins et al. (2011) Where does money come from?, New Economics Foundation

Our Contribution (2)

• Our model explicitly includes the mechanisms through which bankscreate new money supply. We will here present a simplified versionof the model that assumes no credit rationing, but if you want tohave a taste of the overall structure you can read the presentationwe gave at the Ecological Economics conference in Rio de Janeirolast June.

• Our numerical simulations show that, in the simple model wepresent here, no growth can take place without the creation ofcredit by banks, the net change in the level of debt being the thecrucial variable affecting the dynamics of the economy.

• Finally, we expand Keen’s idea to include the supply side (i.e. theproduction process) into the picture. This allows us to consider thecase in which planned expenditure doesn’t become realizedexpenditure because of a supply bottleneck.


http://www.slideshare.net/emacampiglio/nef-macro-model-isee-rio-2012

But enough with words:Let’s look at the model

• nef has developed a consistent macro framework using system dynamics thatwe use for a variety of research questions.

• The overall structure of the model is pictured below:

• We will here present a simplified version of our macro “core unit”,abstracting from the sectoral accounts.

Aggregate macroeconomic framework

Sectoral accounts

Demand

Government

Production Employment

BanksCentral

BankGilt

sellersHouseholds

Non financial

firms

Every agent in the economy ismodelled using a double-entrybook-keeping representation inorder to ensure consistency inthe model.

Sectoral accounts are connected to anmacro “core unit” where demand,supply, investment and employmentdynamics are modelled.


A word on system dynamics

• The model is built employing system dynamics, a methodologyused to study the behaviour of complex systems and based onthe explicit representation of feedback loops. Its basic units are:

– Stocks and Flows (basically, differential equations)

– Connectors (parameters or simultaneous equations)

See this simple examplewhere Population is a stock,new born and deaths areflows affecting the level ofthe stock, and the rest ofvariables are exogenousparameters or defined bysimultaneous equations.


Output, inventories and sales

First of all, imagine the production ofoutput as a flow of goods and servicethat enters a stock of inventories.Imagine this as firms producing andthen storing their production inwarehouses.

Output is produced according to someproduction function. We are currently usinga standard Cobb-Douglas function ofphysical capital and labour, but thefunctional form is not relevant for the pointwe want to make here.

InventoriesOutput Sales

Goods and services leave theInventories stock once they aresold.

We will for now assume thatsupply is not an issue:everything that is demandedis available to be purchased.This is made to focus on thedemand side dynamics. Wewill relax this assumptionlater on.


Wages and Profits

Sales are distributed amongworkers and firms in theform of wages and profits..

Workers then plan theirconsumption on the basisof their wages. Forsimplicity, we assumethat workers want toconsume their entirewage, without saving norasking for loans.

Wages (W)

Sales (S)

Profits (Π)

Planned Consumption (CP)


..That is, in our model Salesare always equal to Income.

What happens to Profits?

For simplicity, assume thatfirms pay no interests onloans, but just have torepay a portion of the debtalready contracted. Profitsless the repayment of debtare equal to Net Profits.

NetProfits (ΠN)

Planned Investments (IP)

Debt Repayment (DR)

Credit Creation (CC)

Net Profits are then entirelyinvested in new physical capital. Iffirms desire to invest more thantheir net profits they need to askbanks for new loans.

Therefore, Planned Investments are equalto Profits plus Net Credit Creation, whereNet Credit Creation is equal to CreditCreation less Debt Repayment.


Profits (Π)

Planned Aggregate Expenditure

Finally, Planned Aggregate Expenditureis equal to Planned Consumption plusPlanned Investments.

Wages

Profits

Planned Consumption

Planned Investments

SalesPlanned

Aggregate Expenditure

Net Credit Creation


Planned Aggregate Expenditurerepresents the amount that agentstoday plan to spend tomorrow.

Realized Expenditure

Planned Aggregate Expenditure become RealizedExpenditure only in the following period. That is, RealizedExpenditures at time t are equal to Planned Expenditures attime t-1.

Time t Time t+1

Planned Aggregate

Expenditure

RealizedAggregate

Expenditure

Remember that for the moment we are assuming that everything is demandedcan be found on the market, so everything that firms plan to spend (eithercoming from profits or new credit) is actually spent on something.


The overall picture (time t)

Wages t

Profits t

Planned Consumption t

Planned Investments t

Planned Aggregate Expenditure t

Net Credit

Creation t

Sales t

Inventories t

Realized Aggregate

Expenditure t

All this happens during the sameperiod (time t)

At time t+1 the Planned Expenditurebecome Realized Expenditure.

Realized Aggregate

Expenditure t+1


The overall picture (time t+1)

Wages t+1

Profits t+1

Planned Consumption t+1

Planned Investments t+1

Planned Aggregate Expenditure t+1

Sales t+1Inventories t+1

Realized Aggregate

Expenditure t+1

Net Credit

Creation t+1

Plus, plannedconsumption at time tbecomes consumptionat t+1 and plannedinvestments at time tbecome investments att+1.Realized Consumptionand Investments thenaffect other modelsectors, not shownhere.

Realized Aggregate

Expenditure t+2

In the same period (t+1) all the othervariables are determined, including thenew Planned Aggregate Expenditure.

At time t+2 the Planned Expenditurebecome Realized Expenditure, andso on..


In analytical terms:

The model can be described bythe following equations:

1) St = ERt

2) Wt=α St

3) Πt=St-Wt

4) CPt=Wt

5) IPt= Πt+ΔDt

6) Dt+1 = Dt+CCt-DRt

7) DRt=Dt/r8) CCt=η(Πt-DRt)9) EP

t=CPt+IP

t

10) ERt+1=EP

t

Where:S : SalesER : Realized ExpenditureW : Wages Π : ProfitsCP : Planned ConsumptionIP : Planned InvestmentsD : DebtCC : Credit CreationDR : Debt RepaymentEP : Planned Expenditureandα : Labor Sharer : Debt Repayment Timeη : Propensity to Invest


Some scenarios

On the left, you can see how thisreduced version of our model lookson the system dynamics softwarewe’re using, Stella.


For the first run, we set:• Initial Realized Expenditure:

100• Initial level of debt = 0• Infinite inventories

(remember assumption: all demand is satisfied)

• Eta (η) = 1.4• Debt repayment time (r) = 5• Alpha (α) = 0.7

1. Controlled growth

• Realized Expenditure isalways equal to Income(and to Sales of course).

• Planned Expenditure ishigher than Income.

• The difference betweenPlanned Expenditure andIncome is equal to NetCredit Creation (right Y-axis).

• In this case Net CreditCreation converges tozero in the long-term asdebt and its repaymentbecome larger.

• Income (and the wholeeconomy) grows untilreaching a plateau.

0

2

4

6

8

10

12

14

100

110

120

130

140

150

160

170

180

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Ne

t cr

ed

i cre

atio

n

Exp

en

dit

ure

an

d in

com

e

Realized Expenditures Planned Expenditures

Income Net Credit Creation


2. Exponential growth

0

50

100

150

200

250

300

350

400

450

500

0

100

200

300

400

500

600

700

800

0 1 2 3 4 5 6 7 8 9 10

Ne

t cr

ed

i cre

atio

n

Exp

en

dit

ure

an

d in

com

e



• Change of parameters:o Eta = 2;o r= 10.

• In this case theexpansion of Net CreditCreation (CreditCreation growing at afaster pace than DebtRepayment) allows forexponential growth.

• But still the sameapplies:

o Realized Expenditure =Income;

o Planned Expenditure > Incomeo Planned Expenditure – Income

= Net Credit Creation.


3. No growth

-10

-8

-6

-4

-2

0

2

4

6

8

10

50

60

70

80

90

100

110

120

130

140

150

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Ne

t cr

ed

i cre

atio

n

Exp

en

dit

ure

an

d in

com

e



• What happens when eta =1, that is firms invest justtheir net profits andtherefore no new credit iscreated?

• The economy doesn’t grow!• Recalling the Minsky quote:

For real aggregate demandto be increasing, it isnecessary that currentspending plans be greaterthan current received incomeand that some markettechnique exist by whichaggregate spending inexcess of aggregateanticipated income can befinanced.”


4. Decline

-4

-3.5

-3

-2.5

-2

-1.5

-1

-0.5

0

80

82

84

86

88

90

92

94

96

98

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Ne

t cr

ed

i cre

atio

n

Exp

en

dit

ure

an

d in

com

e



• If we allow for apositive value of initialloans (D0=30), debtrepayment can bringincome down, even ifeta>1 (new credit iscreated). In this caseeta=1.1.

• That is, when NetCredit Creation isnegative (look at theright Y-axis) PlannedExpenditures arelower thanincome, and thisbrings the economy ina recession.


A step further: Supply


Sales t

Realized Aggregate

Expenditure t

Output t

Planned Aggregate Expenditure t-1Inventories t

Capital Stock tEmployment t

As we have seen, PlannedAggregate Expenditure at t-1determines Realized AggregateExpenditure at t.

That’s not all. In our macro modelPlanned Expenditure is also usedby firms to decide how manyworkers they want to hire.

It’s reasonable to say that in the modelsupply adapts to demand dynamics.

Together with the Capital Stock (whosedynamics is governed by investments anddepreciation) these two factors ofproduction determine the level of output.

Some adjustments


Sales t

Realized Aggregate

Expenditure t

Output t

Planned Aggregate Expenditure t-1Inventories t

But let’s keep it simple andforget about the adjustment ofproduction factors for amoment.

The point that we want tomake here is that thisrepresentation is also capableof modelling the case in whichin the economy suffers from asupply bottleneck and it’s notpossible to satisfy the entiredemand.

Sales, which are still equal to Aggregate Income bydefinition, are now redefined as the minimumbetween the Attempted Expenditure and Output plusInventories. This allows us to take into account thatit’s not possible to purchase what hasn’t beenproduced. Sales are now the Realized Expenditures.

We rename Realized Aggregate Expenditure asAttempted Aggregate Expenditure. That is, PlannedExpenditure at time t-1 is what the economy triesto spend at time t. If there is enough output, everyattempted expenditure will be realized;otherwise, just a portion of it.

Attempted

= Realized Expenditure

A simple example


• For the first 10 periods of the simulation, output is set to beequal to whatever the economy attempts to spend in thatperiod. We are in the same situation as previoussimulations, where supply always satisfies demand.

• At time 10, we impose a shock: supply is now capable ofsatisfying just 90% of demand (Attempted Expenditures).

• At time 13, the situation goes back to the previoussituation: supply is able to satisfy the entire demand again.

• All the parameters and initial values are set equal toScenario n.1 (Controlled growth). The initial value ofInventories is set equal to 30.

A supply constraint


100

110

120

130

140

150

160

170

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Output Attempted Expenditure

Planned Expenditure Income (Realized Expenditure)

• The shock imposed to outputappears very clearly.

• For a couple of periods, theeconomy doesn’t seem to suffermuch: this is because the stock ofinventories is being depleted.

• When Inventories run out, demand(Attempted Expenditure) can’t besatisfied. Income ( = RealizedExpenditure) decreases.

• Planned Expenditure also godown, as Desired Investmentsbecome lower than Profits, thusleading to a negative Net CreditCreation.

• With a lag of one period, alsoAttempted Expenditure drops.

• At period 13, when the shock isover, there’s little left to do: theeconomy is stuck in a steady statethat is lower than the potential one.

Some conclusions

The model presented here is very simple and characterized by a variety of limiting assumptions:– There are no interest payments;– The demand for loans is completely satisfied by banks;– There is no price dynamics;– etc.

Still, we believe it’s capable of grasping some core features of how modern economies work, and in particular the role of debt in influencing aggregate demand. We have also been able to:• Clarify Keen’s analysis by explicitly separating realized and planned expenditures.• Show how Keen and Krugman have different definitions of aggregate demand. • Show through numerical simulations that both Keen and Krugman are right in their own

logic:– Realized Expenditure (Krugman’s Aggregate Demand) is equal to Aggregate Income. – Planned Expenditure (Keen’s Aggregate Demand) can instead be different from Aggregate Income

and the discrepancy is, as Keen argues, equal to the net change in the levels of debt.

• Argue that Keen’s definition of Aggregate Demand is more appropriate, especially if the role of credit creation by private banks in shaping macroeconomic dynamics is to be understood.

• Show how in our simple model no growth can take place without the creation of credit by banks. The net change in the level of debt (Net Credit Creation) is the single most important variable affecting the dynamics of the economy.

• Expand the framework to include the supply side of the economy, and show how a production bottleneck could complicate the dynamics.

Thank you!

[email protected]

www.neweconomics.org

Reconciling Krugman and Keen using nef model

Economy & Finance

economys aggregate demand

international economics

contenderspaul krugman

steve keen

levels of debt

role of debt

level of debt

aggregate demandwhat