Credit Accel­er­a­tor Leads and Lags

Flattr this!

A num­ber of blog mem­bers argued that my lead/lag analy­sis of the Credit Accel­er­a­tor and eco­nomic and finan­cial vari­ables (unem­ploy­ment, share and house price indices) appeared erro­neous.

I am the first to admit that–though my math­e­mat­i­cal mod­el­ling is strong–my sta­tis­ti­cal analy­sis is not up to the same level. I long ago reacted adversely to the prac­tice of econo­met­rics in eco­nom­ics, largely on the same grounds that led Ed Leamer to pub­lish his famous paper “Let’s Take the Con out of Econo­met­rics” (AER, March 1983), omit­ted vari­able bias, etc.

How­ever this is one of those issues where some­one with a strong sta­tis­ti­cal basis (as well as math­e­mat­i­cal physics foun­da­tion) using stan­dard sta­tis­ti­cal pro­grams can do bet­ter than I did, so I passed the data on to “Lyon­wiss”. Here are the results of his analysis–and here is the orig­i­nal data, should oth­ers wish to ana­lyze it them­selves.

Steve sent me his cal­cu­lated data for “Credit Accel­er­a­tor” (CA), “Unem­ploy­ment Change”, “Real House Price Change” and “Share Price Change” (Change in % pa) for Aus­tralia and USA over 1993 to 2011 and asked me to check his cal­cu­lated cor­re­la­tions of CA against the other vari­ables. I con­firmed his cor­re­la­tion cal­cu­la­tions and proved that the cor­re­la­tions are all sta­tis­ti­cally sig­nif­i­cant (greater than 99.9% prob­a­bil­ity) over the period. For the period 1993–2011, I also con­firmed that credit accel­er­a­tor leads and lags (in months) the other vari­ables for pro­duce the max­i­mum cor­re­la­tion (or anti-cor­re­la­tion) shown in the fol­low­ing table:

Coun­try

Unem­ploy­ment

House Price

Share Price

Aus­tralia Lead(+)/Lag(-)

0

10

–8

Aus­tralia Cor­re­la­tion

–0.7768

–0.3411

0.7175

USA Lead(+)/Lag(-)

–5

–9

–11

USA Cor­re­la­tion

–0.8516

0.7228

0.5739

 

Note that CA leads another vari­able by +x months, if the data for the other vari­able are lagged by x months, by shift­ing x months of data in the future to the present. Sim­i­larly CA lags another vari­able by –x months if the CA data is shifted x months from the future to the present.

The above table states that CA in the USA lags house price changes by 9 months and share price changes by 11 months. This roughly agrees with Fig­ure 7 and 8 of Steve’s June 11 post, where both house price and share price changes dripped well before CA. The empir­i­cal data sug­gest that the CA is more likely to be a lag­ging vari­able rather than a lead­ing one, as four cases out of six are lags, one leads, while one is con­tem­po­ra­ne­ous.

More­over, if we reject the sta­tion­ary equi­lib­rium world of neo­clas­si­cal eco­nom­ics, the lead-lag rela­tion­ships are not expected to be sta­ble. So, I divided Steve’s dataset into an ear­lier period 1993–2001 and a later period 2002–2011 and per­formed the same analy­sis for each period sep­a­rately. The lead-lag max­i­mized cor­re­la­tions were all sta­tis­ti­cally sig­nif­i­cant. For the ear­lier period, we have:

Coun­try

Unem­ploy­ment

House Price

Share Price

Aus­tralia Lead(+)/Lag(-)

0

4

4

Aus­tralia Cor­re­la­tion

–0.4097

–0.5445

–0.4636

USA Lead(+)/Lag(-)

12

–15

11

USA Cor­re­la­tion

–0.4405

–0.3787

0.4103

 

And for the later period, we have:

Coun­try

Unem­ploy­ment

House Price

Share Price

Aus­tralia Lead(+)/Lag(-)

0

–4

–8

Aus­tralia Cor­re­la­tion

–0.8706

0.5057

0.8007

USA Lead(+)/Lag(-)

–5

–10

–11

USA Cor­re­la­tion

–0.9029

0.901

0.6299

Indeed, the lead-lag rela­tion­ships appear unsta­ble, with four of the rela­tion­ships chang­ing from leads to lags or vice versa, from the ear­lier period to the later period. In the US CA remains a con­sis­tent lag ver­sus real house price changes, while Aus­tralian CA changes remains con­tem­po­ra­ne­ous rel­a­tive to unem­ploy­ment changes. (There are other sug­ges­tive obser­va­tions, not men­tioned here.)

The results are largely what I expected. The key result is that there are sta­tis­ti­cally sig­nif­i­cant rela­tion­ships between CA and eco­nomic vari­ables, sug­gest­ing the impor­tance of pri­vate credit in the real econ­omy and the non-neu­tral­ity of money in the short to medium term (10 to 20 years).

How­ever, the causal­ity of credit appears com­plex, not dis­play­ing the sim­ple time-invari­ant causal­ity of physics. As Steve sug­gests in a com­plex sys­tem where there are non­lin­ear feed­backs rather than lin­ear cau­sa­tion one expects leads and lags to alter over time.

About Steve Keen

I am Professor of Economics and Head of Economics, History and Politics at Kingston University London, and a long time critic of conventional economic thought. As well as attacking mainstream thought in Debunking Economics, I am also developing an alternative dynamic approach to economic modelling. The key issue I am tackling here is the prospect for a debt-deflation on the back of the enormous private debts accumulated globally, and our very low rate of inflation.
Bookmark the permalink.
  • You said

    But the hypoth­e­sis dQ/dm=V/p is wrong, because you are assum­ing V/p is con­stant. And later you asserted that if p is con­stant then V is con­stant. V is not a con­stant. You need to be care­ful about using and inter­pret­ing data from oth­ers and not from orig­i­nal sources. ”

    For­give me if I mis­spoke. What I meant was dQ/dm(1966) = V/p(1966)

    dQ/dm(1966) = 0.75 , p(1966) = 4 , ergo 0.75 = V/4 —> V = 3

    Of course I didn’t meant that V/p was a con­stant.

    It would be inter­est­ing to cal­cu­late V for Aus­tralia from your chart. Do you have Aus­tralian price level data?

  • Lyon­wiss July 5, 2011 at 10:39 am

    From the site http://www.rateinflation.com/consumer-price-index/australia-historical-cpi.php?form=auscpi I obtain some Aus­tralian CPI data which I plot­ted below.

    Your chart has an ordi­nate labeled ‘Veloc­ity of Credit’ which I will call Z.

    Z(1980)= 2.25 and Z(2010)= 0.63 . From my data, I take CPI(1980) = 47.3 and CPI(2010) = 172.6 I believe Z is dif­fer­ent from V of the Fisher equa­tion
    V= pQ/m BUT THAT Z = kV/p

    Solv­ing twice for kV : kV = Z(2010)*CPI(2010) = 108.74

    kV = Z(1980)*CPI(1980) = 106.43 which is vir­tu­ally the same num­ber.

    So, V = 107.6/k = con­stant, prov­ing that V is con­stant.

    Now, if we only knew what k is …

  • Lyon­wiss

    War­ren Raft­shol July 5, 2011 at 2:13 pm

    In my chart, the veloc­ity of credit is V=pQ/m, where p is the GDP defla­tor, Q is the real GDP, pQ is the nom­i­nal GDP and m is money, which I take to be the broad­est mea­sure, credit. CPI is not rel­e­vant here.

  • You wrote

    In my chart, the veloc­ity of credit is V=pQ/m, where p is the GDP defla­tor, Q is the real GDP, pQ is the nom­i­nal GDP and m is money, which I take to be the broad­est mea­sure, credit. CPI is not rel­e­vant here.”

    The ordi­nate in your chart is NOT V , it is V/p as I have demon­strated. To get V, which is a con­stant, one must know p, the price level.

    By the way, Aus­tralian CPI starts in 1948 when US dol­lar was fixed at 1/35 gold oz.

    If one divides Z = 107.6 by 35, one gets V = 3.07 which agrees with my cal­cu­la­tion V = 3 from my analy­sis of my pre­vi­ous chart.

    V is a con­stant, it is not a vary­ing func­tion .

    The point of all this is that V is a con­stant approx­i­mately.

  • Lyon­wiss July 5, 2011 at 4:21 pm

    Cor­rec­tion: Pre­vi­ous post should read

    ” If one divides Z*CPI = 107.6 by 35, one gets V = 3.07 which agrees with my cal­cu­la­tion V = 3 from my analy­sis of my pre­vi­ous chart.”

    If you have data for real GDP, GDP defla­tor, and money sup­ply, then for any date you choose

    pQ/m = V should be con­stant

  • Lyon­wiss

    War­ren Raft­shol July 6, 2011 at 2:52 am

    I have cal­cu­lated V from data and it is not con­stant as in my chart. But V is expected to be approx­i­mately con­stant for nar­row money M1, not credit.

  • Here is the best that can be hoped for.

    Price level relief due to bank­rupt­cies and other debt col­lapse, with infla­tion 1% or less. Note the wage share recov­ery with long cycles with very high unem­ploy­ment