A simple Keen Monetary Circuit simulation¶

Debunking economics¶

Economic crises have become frequent events in the last 80 years. The lack of a coherent narrative to explain them has led to controversy in the academic profession as rival schools challenge or defend the orthodoxy of neoclassical economic theory.

One longstanding opponent of neoclassical theory, an Australian professor called Steve Keen, wrote a book called Debunking economics: The naked Emperor dethroned? (the reference for this example is the 2011 edition, particularly chapters 13-18).

As an Engineer familiar with the mathematics of electronic control systems, Keen saw many deficiencies in the static approach of neoclassical macroeconomics. He presented an alternative; to treat the economy as a natural system capable of dynamic instability.

Professor Keen incorporated ideas from Marx, Schumpeter, Keynes and Minsky, to create what he called the Monetary Circuit Model. Far from being complicated, this approach feels quite natural to someone who has written software to simulate industrial processes.

Keen’s simulation uses simple rules but is nonetheless capable of predicting the kind of phenomena seen in financial downturns. These are not explained by orthodox economic models.

Double-entry paradigm¶

In 2012, the UK-based blogger Neil Wilson wrote a critique of Keen’s new model. Wilson was concerned that Keen’s work had too much Engineering about it.

While supportive of the approach, Wilson felt that the model’s ignorance of accounting concepts prevented its acceptance by the financial community.

In the article Double entry view on Keen Circuit Model, Wilson reviews Keen’s dynamic circuit model. He proposes some modifications to make it more recognisable to a reader familiar with the double-entry accounting system used by banks and businesses.

How it works¶

The tallywallet.common.debunking module presents Keen’s simulation according to Wilson’s perspective of balanced accounting.

You run it as follows:

$ python -m tallywallet.common.debunking

Or, to make use of the advanced options, read the help like this:

$ python -m tallywallet.common.debunking --help

tallywallet.common.debunking.banking_licence(ldgr, val)[source]¶

Wilson’s suggestion is to formalise the ability of a bank to create money. His first step (not present in Keen) is to capture this value as an asset.

1. Grant licence

tallywallet.common.debunking.bank_loan(ldgr, dt, pa=Decimal('0.5'))[source]¶

These two steps create a debit against the licence and a corresponding addition to the loans book. The firms’ account is credited, reducing the balance in the vault.

2. Lend money
3. Record loan

tallywallet.common.debunking.bank_charge(ldgr, dt, pa=Decimal('0.05'))[source]¶

This operation is a significant departure from Keen’s original.

In the real world, a Bank will want to reduce its liabilities at every opportunity. Interest is charged immediately as revenue which reduces the level in the vault. The interest is issued as a loan so the ledger goes up and the value is deducted from the licence.

4. Charge interest
5. Record interest

tallywallet.common.debunking.firms_repayment(ldgr, dt, interest, pa=Decimal('0.1'))[source]¶

The interest calculated in the previous step is debited from the firms’ account. At the same time, part of the principal of the loan is paid off too. The total reappears in the vault, reducing the loan book and accruing value back to the licence.

6. Repay Loan and Interest
7. Record Loan and Interest Repayment

tallywallet.common.debunking.nonbank_interest(ldgr, dt, paF=Decimal('0.02'), paW=Decimal('0.003'))[source]¶

Keen’s original model only paid interest on the firms’ account. Wilson added interest for workers too. In order to maintain the equivalence of the two simulations, I have adjusted the workers’ interest rate so that the steady state output after ten years is close to that described in Keen’s book. This corresponds to a rate of 0.3% on a worker’s current account which, though low, has been observed in the UK in recent times.

8. Pay firm deposit interest
9. Pay worker deposit interest

tallywallet.common.debunking.firms_wages(ldgr, dt, pa=Decimal('3'))[source]¶

Keen models the cost of production entirely as workers’ wages. There is no mention of capital. Clearly this is a simplification which we can elaborate at another opportunity.

10. Hire Workers

tallywallet.common.debunking.nonfirms_consumption(ldgr, dt, paB=Decimal('1'), paW=Decimal('26'))[source]¶

This operation is as Keen defined it, since it doesn’t impact the loan ledger or the licence asset which Wilson introduced.

11. Workers’ Consumption
12. Bankers’ Consumption

tallywallet.common.debunking.simulate(samples, initial=100000000, interval=3600, ledger=None)[source]¶

Run the simulation by repeating the operations described above. At the end of every cycle, the equality of the Fundamental Accounting Equation is tested. A warning is raised if the equality fails.

Parameters:	samples – A sequence of times. At each of these the simulation will print out the state of the Ledger. The simulation will stop after the final sample time has passed. initial – The initial value of the banking licence. interval – The value of the simulation timestep in seconds. ledger – An existing Ledger object. If None is passed, a new one will be created.
Returns:	This routine is a generator which yields RSON strings. The final return value is the ledger object used during the simulation.

Typical output¶

By default the program replicates the behaviour described in Debunking. It begins with a debt supply of $100,000,000 and runs for ten years. The output matches to three significant figures that published by Keen.

{}
    header:version: 0.005
{}
    ledger:columns:    [
            [licence, USD, asset],
            [loans, USD, asset],
            [vault, USD, liability],
            [firms, USD, liability],
            [workers, USD, liability],
            [safe, USD, income],
            [USD trading account, USD, trading]
        ]
    ledger:ref: USD

{}
    note:
        Keen Money Circuit with balanced accounting
    ts:
        0
[ 100000000.00, 0.00, 100000000.00, 0.00, 0.00, 0.00, 0.00]

{}
    ts:
        3600
[ 99994276.62, 5723.38, 99994276.62, 5721.40, 1.96, 0.02, 0.00]

{}
    ts:
        31449600
[ 62400370.51, 37599629.49, 62400370.51, 33389071.04, 3733551.18, 477007.26, 0.00]

{}
    ts:
        62899200
[ 41765575.22, 58234424.78, 41765575.22, 51200784.95, 5832702.36, 1200937.47, 0.00]

{}
    ts:
        94348800
[ 30441134.47, 69558865.53, 30441134.47, 60821254.87, 6966230.99, 1771379.68, 0.00]

{}
    ts:
        125798400
[ 24226245.65, 75773754.35, 24226245.65, 66043064.02, 7581387.75, 2149302.57, 0.00]

{}
    ts:
        157248000
[ 20815495.07, 79184504.93, 20815495.07, 68887106.37, 7916392.69, 2381005.88, 0.00]

{}
    ts:
        188697600
[ 18943664.38, 81056335.62, 18943664.38, 70439795.64, 8099272.33, 2517267.65, 0.00]

{}
    ts:
        220147200
[ 17916397.80, 82083602.20, 17916397.80, 71288870.80, 8199273.09, 2595458.32, 0.00]

{}
    ts:
        251596800
[ 17352630.66, 82647369.34, 17352630.66, 71753705.00, 8254017.40, 2639646.94, 0.00]

{}
    ts:
        283046400
[ 17043233.47, 82956766.53, 17043233.47, 72008380.09, 8284010.14, 2664376.30, 0.00]

{}
    ts:
        314496000
[ 16873435.32, 83126564.68, 16873435.32, 72147986.47, 8300451.11, 2678127.10, 0.00]