A Paris, 1er octobre 2005.



Il y a trente ans, en 1976, le chômage s'était déjà envolé en France: la "barre" du million de chômeurs avait été officiellement franchie au quatrième trimestre de 1975.
Il ne s'est plus jamais posé depuis, au moins jusqu'à aujourd'hui.

                                                  Graphique

                                            Taux de chômage

                                              1967-2001



Source:http://inflation.free.fr/graph_chomage.php

En d'autres termes, fin 1967, le taux de chômage "global" avait passé la "barre" du 1% de chômeurs,
fin 1973 la "barre" du 2%,
fin 1974 celle du 3% et
mi-1975 celle du 4%.

Et des "barres" de plus en plus élevées seront franchies par la suite !


Parallèlement, la durée moyenne du chômage s'était allongée et s'allongera pour atteindre les durées décrites par le graphique ci-dessous.



Source:http://www11.hrsdc.gc.ca/fr/sm/ps/rhdcc/ra/publications/revue_trimestrielle/2002-000061/images/graphique28.gif

Autant de circonstances que la France n'avait jamais connues dans son histoire.

Et cela, malgré... les arrivées aux plus hautes fonctions de l'exécutif, en mai 1974, d'un nouveau Président de la République, M. Giscard d'Estaing, et d'un nouveau Premier ministre, M. Chirac.

En effet, si la politique inflationniste suivie en France depuis la fin de la guerre de 1939-1945 - et contre laquelle Jacques Rueff n'avait pas hésité à avoir des mots très durs dans ses ouvrages intitulés Epître aux dirigistes (1949), L'Age de l'inflation (1963) et Le péché monétaire de l'Occident (1971) -, avait un temps empêché l'envol du chômage, mais non pas les dévaluations du "franc", désormais la France avait non seulement une inflation croissante, mais encore un chômage croissant.


Etant donnée cette rétrospective schématique, on ne peut que s'étonner du livre que Jean Fourastié publiera en 1979, sous le titre Les Trente Glorieuses (Fayard), et qu'on l'ait pris au sérieux, comme référence, jusqu'à aujourd'hui inclus.

En effet, voir tacitement dans la spoliation des créanciers en monnaie "franc" par l'inflation, dans les investissements effectués à des taux d'intérêt manipulés par le gouvernement, et dans les dévaluations successives du "franc" des gloires relève pour le moins du "surréalisme". Pour le pire, cher lecteur, je vous laisse le libre arbitre...


On ne peut aussi que s'étonner de l'article que Maurice Allais écrira en 1994 dans Le Figaro du 10 octobre intitulé "La cassure de 1974".

Non seulement, l'année 1974 n'a pas connu de cassure provoquée par la "politique de libre-échange mondialiste" mise en œuvre par la Communauté européenne ou quelque autre, ni par un "laissez-fairisme monétaire", mais encore il n'y a pas eu de cassure: ce qui devait arriver est tout simplement arrivé, comme l'a souligné Jacques Rueff en particulier dans les articles intitulés
"Ce qui doit arriver, arrive" (1969) ou
"Il m'appelait Cassandre" (1971) ou encore
"Pourquoi la crise" (1975).


Pour faire apparaître plus précisément des conditions de l'envol du chômage passées en général sous silence et asseoir les chiffres donnés, voici ci-dessous des extraits d'un texte que j'ai écrit alors, en collaboration avec le professeur Emil-Maria Claassen, et publié sous le titre " The Effects of Unemployment Benefits on the Unemployment Rate in France" dans H.G. Grubel et A. Walker (eds.), Unemployment Insurance : Global Evidence of its Effects on Unemployment, Fraser Institute, Vancouver (Colombie Britannique), 1978, pp.204-34.

Au terme des extraits - en anglais -, je donnerai un épilogue - en français -.


Introduction.

Until the last few years, unemployment in France has not been considered to be a "structural" problem by policy-makers, but rather as a "random" phenomenon.

The result is that no legislation existed about it and few unemployment statistics were collected (section I).
Since 1967, several political measures had been taken in France to make the rising unemployment socially acceptable (section II).

The hypothesis that these measures have had perverse effects on unemployment itself and induced a new kind of unemployment is formulated and tested econometrically in section III.

The insurance-induced unemployment debate is not a recent one in France and accordingly, section IV is devoted to summarizing the very interesting contributions that were made to the debate in the 1920's and 1930's.


I. Unemployment in France 1967-1975.

The official definition of unemployment in France is based on the concept of "demandeur d'emploi non satisfait" (applicant for a durable job) (see data in table 1, col.2: annual average of monthly unemployed, and col.3: quarterly average of monthly unemployed). There are two reasons why our analysis has been limited to the period 1967-1975.
                        

First, before 1967, the unemployment statistics were very inadequate.
The creation of the National Agency for Employment in July 1967 and of its regional centers marks the beginning of good quality unemployment statistics.
From 1967 to 1972, the Agency centers were gradually established throughout the country and, in December 1972, it was estimated that 95 percent of the unemployed population was recorded. Thus, the unemployment data of table 1 are not comparable with those of previous period.

Second, there have been changes in the definition of the labor force, which complicate analysis over a long period (see data table 1, col.1)

Third, even within the period 1967-1975, there are some difficulties with the statistics, in particular because of the changes in the average duration of unemployment.

The first reason is that no official estimates about average duration of French unemployment are available.

The second reason is that duration can be proxied by information acquired from two different sources and they yield contradictory impressions.

The first approach is to examine the "stocks" of registered, unemployed persons during the quarter (see table 2).

In our analysis, we divided unemployment into two groups : workers who had been unemployed for three months or less – group A – (see data in table 2, col.1) and workers who had been unemployed for more than three months – group B -.

Group A and group B define applicants for a durable job (see data in table 1, col.3 or in table 2, col.2).
In the data thereby obtained, we observe that 258.8 thousand unemployed persons (see data in table 2, col.3).
In 1975 IV, 526.5 thousand unemployed persons belonged to group B, i.e. 48.2 per cent of registered unemployment. From this, we conclude that the average duration of unemployment has been falling.

                               Table 2

                      Applicants for job
      as function of the timing of the demand
    

Source:

The second approach is to look at flows in the labour market and involves analysis of the status of compensated, unemployed persons who found a durable job during the quarter (see table 3).

In this case, we divided compensated, unemployed persons who found a durable job during the quarter into two groups: workers who had been compensated for three month or less – group C – (see data in table 3, col.1) and people who had been compensated for more than three months – group D -.
Group C and group D define "outflow intensity" of indemnified, unemployed persons into employment.

                             Table 3

      Duration of total unemployment
                      1971 I – 1975 IV

Source:

An analysis of the data reveals that 38.8 thousand compensated persons belonged to group C in 1971, i.e. 90 per cent of compensated persons (see data in table 3, col.2). 
In 1975 IV, 36.1 thousand belonged to group C (i.e. 80.5 per cent) after a rise from 1971 till 1973 IV, then a decline from 1973 IV till 1975 II.
Compared to the average stock of unemployed persons in the previous quarter (quarter t-1), group C represented 32.7 per cent in 1971 I, but only 8.9 per cent in 1975 IV after a steady fall. This fall is an indicator of an increase in the average duration of unemployment.

                        Table 4
          Partial unemployment :
                1967 I – 1975 IV

Source:

A final question raised by the statistics concerns the treatment of partial unemployment – behaviour that has never been taken into account by the "unemployment statistics" whatever their sources.
This kind of unemployment which is measured in terms of numbers of individuals and lost hours has not been integrated with the total unemployment picture (see table 4).
Yet, any change in this kind of unemployment has an impact on total unemployment which is difficult to measure because of the lack of relevant information.


II. Evolution and characteristics of the French unemployment insurance scheme.

The scheme has two components: total unemployment compensation and partial-unemployment compensation.

a. Total-unemployment compensation.

All unemployed workers can take advantage, simultaneously, of three unemployment benefit program.
At the moment (1976), these program guarantee, to the "representative" insured unemployed, an unemployment benefit during the first year of unemployment between 40.25 per cent and 90 per cent of the last wage earned.
The three programs are listed in tables 5 and 6 for unemployed workers less than 60 years old) and in table 7 (for unemployed workers of 60 years and more) and they are classified under four criteria: the reason for lay-off (economic motive or not), the insurance program, the eligibility rules, the amount and the duration of compensation benefits.

                         Table 5
 
              Total unemployment:
       workers less thant 60 years old
                      - avril 1976

Source :

                         Table 6
          
               Total unemployment :
     Workers less than 60 years old
                   - avril 1976 -
    amount and duration of compensation
               (special ASS.E.D.I.C.)

Source :

                       Table 7

            Total unemployment:
    workers 60 years old (and more)
                   - avril 1976 -

Source :

(1) The first program, the Public Relief Benefit ("allocation d'aide publique") created in 1967 by the Government and financed by the government budget, consists of a lump sum payment.
 It provides an unemployment benefit of about 30 per cent to unemployed who earn the minimum wage rate (the "salaire minimum interprofessionnel de croissance").
The time during which an individual is allowed to receive the benefit is not limited.

However, after three months, the daily income (from other sources) is not allowed to exceed a maximum level which depends upon the number of people in the unemployed worker's family.
Moreover, for each successive year, there is a deduction of ten percent, except for those over 55 years of age (the cumulated deductions are limited to 30 per cent).
Table 8 describes the evolution of the amount of paid benefits and of the number of beneficiaries (since 1967).

                        Table 8

              Total unemployment :
             compensation benefits
                   – march 1976 -

Source :

Table 9 indicates the evolution of the minimum amount of monthly benefits during the first three months.
Until 1975, the unemployed worker earned the Public Relief Benefit (and all other services of the Social Security) provided that he was less than 65 years old, that he was registered by the local center of the National Agency, and that he had been working 1,000 hours during the last twelve months.


                         Table 9

  Minimum employment compensation benefits
(in FF per month and during the first three months)
     paid to a single eligible unemployed worker
 minimum special ASS.E.D.I.C. benefit (1st line)
      - public relief benefit (2nd line)


Source :

(2) The second program, the Special ASS.E.D.I.C. Unemployment Benefit ("allocations spéciales des Associations pour l'Emploi dans l'Industrie et le Commerce – U.N.E.D.I.C. –) is the result of a private agreement concluded between employers and trade-union in 1958 and financed by employers (80 per cent) and employees (20 per cent).
To receive this unemployment benefit, one has to be less than 65 years old and to have worked 520 hours (as a wage earner) during the twelve last months.
According to their age (see tables 5, 6, 7), unemployed earn, every month, 40.25 per cent or 30 per cent of their last monthly wage over periods of different lengths and the benefit cannot be less than a minimum amount at the beginning (see tables 6 and 9).
Table 8 shows the evolution of the amount of paid benefits and of the number of beneficiaries (since 1967).

The "Income Guarantee Program" ("garantie de resources" since 1967), reserved to the oldest unemployed (see table 7), is an arrangement of the Special ASS.E.D.I.C. Unemployment Benefit and of the Public Relief Program.
This program guarantees to the unemployed worker 70 per cent of the basic wage until his retirement if he has been a wage earner for 10 years and its benefit represents an additional unemployment benefit. Table 10 shows the evolution of the amount of paid benefits and of the number of beneficiaries.

                        Table 10
  
        Income guarantee program


Source:

(3) The third program, labelled the "Additional Waiting Unemployment Benefit Program" ("allocation supplémentaire d'attente") is the most recent one. It was created in December 1974.
To the extent that unemployed are eligible for the first programs, they receive the unemployment benefits of these programs, plus a benefit of the third one (mainly financed by the Government budget), such that the sum of the three kinds of benefits (ASS.E.D.I.C. Unemployment Insurance, Public Relief, and Additional Waiting Benefit) may not exceed 90 per cent of the basic wage for slightly more according to the number of people in the family of the worker.
The additional waiting benefit is only earned by people laid off for economic motive.
They must be less than 60 years old, have been registered by a local center of National Employment Agency and have worked 1040 hours during the last twelve months. Table 11 shows the evolution of number of beneficiaries and of the amount of payments (February 1975).

                           Table 11

            Additional waiting benefit

Source:

The tax treatment of unemployment benefits is differentiated between the various unemployment programs.
The ASS.E.D.I.C. Unemployment Program underlies the general income tax, but the Public Relief Program is free of taxes.
The taxation of benefits resulting from the third unemployment benefit program does not take account of the Public Relief benefit if the payment is less than 1500 FF a month (and increased by 500 FF for every member of the family).

                               Table 12

       Unemployment compensation benefits
                        and net wages 
                            (1967-1975)

Source:

The evolution of the value of unemployment compensation benefits can be compared to the value of wages in two different ways.
It can be seen in table 12 that between 1967 and 1975, the amount of payments increased from 430 millions of FF to 7646 millions of FF (an 18-fold increase) while the amount of wages (net of social contributions) increased only four-fold, from 187 millions of FF to 530 millions of FF. In 1975, unemployment compensation benefits represented 1.42 per cent of the total income of workers (employed and unemployed workers).

                            Table 13 

           Nominal compensation benefits
                 and nominal wages

..............NCB.................W...................NCB/W
.............................................................en %

1967......519.1.............935.8..................55.5
1968......644.3...........1040.2..................61.9
1969......931.4...........1168.0..................79.7

1970......788.3...........1302.6..................60.5
1971......830.6...........1445.8..................57.4
1972......947.8...........1607.5..................59.0

1973....1115.9...........1830.5..................61.0
1974....1345.1...........2168.6..................62.0
1975....1771.0...........2433.3..................72.8
Source:

The average unemployed earned 519.1 FF a month in 1967, and earned 1771 FF a month in 1975 (a 3.4-fold increase, see table 13) while the mean wage earner obtained 935.8 FF a month in 1967 and 2433.3 in 1975 (a 2.6-fold increase).
Thus, in 1975, the mean benefit represented 72.8 per cent of the mean wage. (It should be noted that the ratio has exceeded this value in the past and for example was as high as 79.7 per cent in 1969).


b. Partial-unemployment compensation.


The worker is partially unemployed when the weekly number of hours of work fall below the legal number (minimum number) or when the firm of the worker must be closed during a period of time.
The partially unemployed workers earn a Public Relief benefit for every lost hour (below the minimum number fixed in 1975 to 40 hours) and if his monthly income is less than a given income (see table 14).

                           Table 14

               Partial unemployment

Source :

                          Table 15

                 Partial unemployment
                compensation benefits

Source :

The benefit is, in effect, public assistance, the amount of which changes according to the number of lost hours.
An additional benefit exists which results from employer –employee agreements and guarantees a benefit equal to 50 percent of the hourly wage for the lost hours.
Table 15 indicates the evolution of the amount of paid benefits and that of the number of lost compensated hours.


III. Estimation of induced unemployment.

For the estimation of the effects of unemployment of the unemployment rate (he so-called induced unemployment) in France, we shall use the one-equation model elaborated by Grubel-Maki-Sax (1975, p.183).
 Its empirical formulation for the Canadian economy is as follows:

Log U = - 8.09 + 2.54(UCB/AWW) – 0.04 PCGNP
..(t)......................(3.89).........................(5.01)

.............- 0.05 PCGNP-1 + 0.08 FLFPR + 0.09 MLFPR
...............(7.40)....................(3.37)..............(1.96)

.............- 0.02 INEL
...............(4.65)

.............................................R2 = 0.92, DW = 2.29

where:
* U is the unemployment rate,
* UCB/AWW is the ratio of unemployment compensation benefits to average weekly wages ;
* PCGNP, PCGNP-1 are the percentage changes in current-dollar gross national product of the period and the lagged period ;
* FLFPR and MLFPR are the labour-force participants rates ;
* INEL is the percentage of new and renewals claims for unemployment benefits which are ruled ineligible during the period.

In our study, we kept these variables, but named them differently. Thus, PCPIB is the percentage change in real gross domestic product in 1963 Francs for PCGNP and R is (UCB/AWW). We redefined the variable EL after the following considerations.

The population (P) can be divided into two components : the active population (AP) and the inactive population (people who are less than 16 years old and more than 60-65 years old, or retired).
The active population is the sum mof people (men and women) who have a job (L) and people without work (U).
Consequently:

AP = U + L
  U = Um + Uw
  L = Lm + Lw

Where the subscripts m and w designate, respectively, men and women.

Let us consider the ratio AP/P. We have:

AP/P = (APm + APw)/P
          = (Pm/P)(APm/Pm) + (Pw/P)(APw/Pw)....(1).....

where Pm and Pw are, respectively, the male and female population.
Thus, the male labour-force participation rate is :
                   MLFPR = APm/Pm
and the female rate is :
                  FLFPR = APw/Pw

So we obtain the identity:
AP/P = (Pm/P) MLFPR + (Pw/P) FLFPR

Among people without work, there are unemployed workers who receive compensation (Ue) and unemployed workers who receive no compensation (Une) because they are not eligible:

                         U = Ue + Une

By definition: 
                 
                      AP = L + Ue + Une

The ratio (Ue/P) can be rewritten as

                  Ue/P = (U/P)(Ue/U)

Where (Ue/U) represents the relative part of unemployed who receive the unemployment compensation benefits ; in other words, it represents the variable EL (the eligibility variable)N

In the same way, (Une/U) represents the ineligibility variable (INEL). Thus the relation (1) can be rewritten:

AP/P = L/P + (U/P)EL + (U/P) INEL

And if P is dropped and substituting AP by (1)

Pm MLFPR + Pw FLFPR = L + U.EL + U.INEL

We obtain alternatively

EL = (1/U)( Pm MLFPR + Pw FLFPR - L) – INEL

INEL = (1/U)( Pm MLFPR + Pw FLFPR - L) – EL

Hence :

(U/AP)EL = (Pm/AP)MLFPR + (Pw/AP) FLFPR – (L/AP)
..................– (U/AP)INEL

where we call (U/AP)INEL, EL*........................(2)

By comparing the identity (2) with some elements of the above equation proposed by Grubel-Maki-Sax, it appears that we may use either the variable (U/AP) EL or the right-hand expression of equation (2) as it is proposed. The use of variable EL* permits us to drop MLFPR, FLFPR, L/AP, and INEL. The variable L/AP does not appear in the equation.

We can interpret EL* as the unemployment rate pertaining to labour force participants who are eligible for unemployment benefits (alternatively, (U/AP) INEL, i.e. INEL* is the unemployment rate for those who are ineligible). Thus, the induced unemployment (IU) depends on:

IU/AP = f(UCB/AWW) + g(EL*)

In our econometric estimation, we use quarterly data covering the period 1967 (1st quarter) to 1975 (IVth quarter).
The unemployment data (see table 1) are not seasonally adjusted, they are quarterly averages of monthly data.
The values for the unemployment rate series (table 16) are obtained by dividing the unemployment data by the annual average of the sum "employed workers + applicants for a durable job" (see table 1).
Thus, we do not take account of partial-unemployment.
Table 16 shows that the rate of unemployment (U) was relatively constant until the third quarter of 1969 (quarter of the devaluation of FF) and trebled between the fourth quarter of 1969 and the corresponding quarter in 1975 to reach 4.8 per cent.

                            Table 16

             Total unemployment rate    
                               (%)

.........quarter.....I..........II.......III..........IV
year
1967...............0.95......0.89......0.89.......1.16
1968...............1.33......1.23......1.19.......1.29
1969...............1.27......1.02......0.95.......1.11

1970...............1.21......1.13......1.19.......1.50
1971...............1.68......1.43......1.48.......1.85
1972...............1.90......1.66......1.69.......1.95

1973...............1.99......1.75......1.89.......2.03
1974...............2.30......2.02......2.30.......3.36
1975...............3.62......3.57......4.00.......4.82
Source:

                            Table 17

                  Real wage rate index
                      (1963 I = 100)

Source :

The index of the real wage rate index (w/p) (see table 17) is obtained by dividing the nominal wage rate index by the index of industrial product prices.
We use this indicator of prices because it reflects more the effects of market forces that other price indices (consumer price index, for example).

                              Table 18

                        A comparison

Source :

The times series of the ratio of unemployment compensation benefits to wages (see table 20) is obtained by dividing ASS.E.D.I.C. benefits (i.e. Special ASS.E.D.I.C. + Income Guarantee + Additional Waiting Benefits) (see table 18, col. 2) by beneficiaries (see table 19, col. 3) and by net wages per capita (i.e. net wages of table 18, col.1 divided by applicants for a durable job of table 19, col.1).
No particular trend can be isolated until the second quarter of 1975.
In other words, we can say that the unemployment compensation benefits followed net wages and thus, the productivity of labour.

                            Table 19

                   Another comparison

Source :

                           Table 20

             Quarterly ratios of
   unemployment compensation benefits
                to quarterly wages
 

Source :

Table 20 shows the ratio of quarterly unemployment compensation benefits (paid by ASS.E.D.I.C. to the average unemployed worker) to the quarterly wage of an average worker for the period 1967 I to 1975 IV.
Except for a jump in 1975 II, no particular trend can be isolated: the mean was 0.513 for the period as a whole and the variability (as measured by the standard deviation) was 0.06.

Table 21 shows the ratio of average annual unemployment compensation benefits paid by ASS.E.D.I.C. and Public Relief Programs to the annual wage of an average worker.
Since quarterly data on Public Relief benefits do not exist, we built them up by multiplying the quarterly values for ASS.E.D.I.C. benefits by the ratio of the annual value of "ASS.E.D.I.C. plus Public Relief benefits" to the annual value of ASS.E.D.I.C. benefits.
This had the effect of producing a seasonalized "ASS.E.D.I.C. plus Public Relief benefits" series with the appropriate annual total.

                        Table 21

              Ratio of unemployment
compensation benefits to quarterly wages
                           (in %)

quarter.......I..........II...........III..........IV
year
1967.......57.3 ....54.8......54.0......55.9
1968.......58.6.....81.3......42.5......65.1
1969.......81.2.....79.0......79.8......78.8

1970.......59.7.....61.4......59.0......61.9
1971.......54.9.....59.2......53.6......62.0
1972.......55.3.....58.4......58.0......64.3

1973.......58.8.....61.7......59.1......64.4
1974.......59.2.....62.6......62.7......63.5
1975.......64.5.....73.8......78.0......74.8
Source :

Table 21 shows the ratio of quarterly unemployment compensation benefits to quarterly wages for the period 1967 I – 1975 IV, obtained by the above method.
The mean was 63.3 per cent for the period as a whole, and the variability was 0.09. The value of 1968 II can be explained by the strikes of May and June (exceptional decrease in paid salaries) and the jump of output in 1968 III (exceptional increase in paid salaries) explains the particular value of 1968 III.

                         Table 22

  Percentage of new and renewals claims
        for unemployment benefits
which were rules eligible/public benefit only


Source:

                         Table 23

  Percentage of new and renewals claims
       for unemployment benefits 
which were rules eligible ASS.E.D.I.C benefits

Source:

The time series of the ratio of new and renewals claims for unemployment benefits (see tables 22 and 23) is calculated in two ways. One (table 22) is obtained by dividing Public Relief beneficiaries by total unemployment and the other (table 23) by dividing ASS.E.D.I.C. beneficiaries by total unemployment.

These tables indicate the eligibility variable. Two definitions were used: the first one (ELPu) is obtained by considering only people receiving Public Relief benefits and the other (ELPr) is obtained by considering only people receiving ASS.E.D.I.C. benefits.
As mentioned above, in most cases, people receiving one kind of benefit receive the other. Thus, the two definitions are used separately. For the period as a whole, the mean of these two definitions of the eligibility variables and their variability were respectively.

1967 I – 1975 IV...........Public Relief.....ASS.E.D.I.C.
...........................................ELPu.................ELPr

Mean...................................0.37..................0.42

Variability...........................0.07..................0.04

The data of table 24 are the annual rates of change of real gross-domestic product (in 1963 Francs) between the same quarters of two successive years.

                      Table 24

       Annual percentage change
      real gross domestic product

.......Quarter........I..........II..........III...........IV
Year
1966................8.1........6.9.......5.3..........4.3
1967................4.9........3.8.......5.2..........5.6
1968................6.4.......-5.3.......8.0..........8.5

1969................6.0.......21.2......5.5..........3.4
1970................5.0........5.2.......5.1..........7.1
1971................5.7........4.8.......5.9..........5.2

1972................6.8.........5.5......5.4..........5.7
1973................5.8.........6.7......6.0..........5.7
1974................4.5.........4.0......3.9.........-1.4
1975..............-4.6....... -4.1.... -5.0..........1.5
Source:

Several regressions were eliminated with U and Log U, respectively, as dependant variables (the above equation of Grubel-Maki-Sax is obtained with Log U), and with W/P, PIB, UCB/AWW and EL as independent variables, and by using two definitions of the eligibility variable. The best results are presented in table 25.

                          Table 25

                 Results of regressions :
                   1967 I – 1975 IV

It can be observed that the sign of every regression coefficient is the expected one except the coefficients of eligibility variables. On the other hand, the low value of the D.W. statistics must be noted in equations 1, 2, 3 6 and 7. However, in all cases, we can conclude that variations in the unemployment variable are well explained by the variables in the regressions.


IV. The discussion of the relation between the unemployment benefits and the unemployment rate half a century ago.

As early as 1925, some theoretical and empirical studies referred to unemployment. The first one is the analysis by Jacques Rueff (1925).
It was published simultaneously with Irving Fisher's studies (1925, 1926).
Rueff compares the evolution of the unemployment rate and of the wage/price ratio for England during 1919-1925 and later (Rueff, 1931) for England during 1919-1930.
He indicates a high positive correlation between these two time series.
Fisher analyses the curves of employment and of annual changes in the price level (lagged following a particular statistical rule) for the United States during 1903-1925.
His study shows a certain parallelism of the two curves.

One of the empirical characteristics of Rueff's work is the use of the wage/price ratio as the key economic variable for explaining changes in unemployment.
The other characteristic is the argument that the unemployment insurance benefits (the "dole" in England) were the real cause of the permanent unemployment (see Rueff, 1931 and 1932).

His argument on permanent unemployment runs as follows.
When prices are not changing, the unemployment compensation benefits represent the floor for the money wage and prevent them from failing.
To the extent that the existing nominal wage rate is very near to the level of the unemployment compensation benefits, a fall in the nominal wage rate would not reduce unemployment : workers prefer to go on the dole rather than to work for wages which are only very slightly higher than the unemployment benefits they receive if they are unemployed: a "permanent unemployment" is induced.
When prices are falling, unemployment benefits are not only a fixed floor for the money wages, but also a rising floor for the wage/price ratio. In such circumstances, the fixed unemployment compensation benefits induce a rising permanent unemployment.
These circumstances existed in England from 1925 onwards when prices are falling.
The unemployment insurance program, introduced in 1911, prevented a fall in wages which would otherwise have been provoked by the excess supply in the labour market: wages, which declined following the price fall at the end of 1922, "bottomed out" against the floor provided by implicit wage provided by the dole. […].


References
[…]

Fisher, I. (1925), "Our Unstable Dollar and the So-Called Business Cycles", Journal of the American Statistical Association, June, pp.181-190.

Fisher, I. (1926), "Banking Policy and Unemployment", American Labor Legislation Review, 41, March, pp.24-29.

[…]

Grubel, H.G., Maki, D. and Sax, S. (1975), "Real and Insurance-Induced Unemployment in Canada", Canadian Journal of Economics, 8, May, pp.174-191.

[…]

Rueff, J. (1925), "Les variations du chômage en Angleterre", Revue Politique et Parlementaire, December, pp.425-437.

Rueff, J. (1931), "L'assurance-chômage, cause du chômage permanent", Revue d'Economie Politique, 45, March-April, pp.211-254. Rueff, J. (1932), "Réunion du Comité directeur de l'Association française pour le progrès social", Les Documents de Travail (Bulletin mensuel de l'Association), 16, april-June, pp.8-79. […]


EPILOGUE : 30 ans plus tard.

J'ai eu l'honneur que Jacques Rueff me demande en 1974 d'étudier si la loi qu'il avait mise à jour en Angleterre se retrouvait dans la France contemporaine.
Cette étude qu'il intitula "La fin de l'ère keynésienne" est disponible sur http://www.catallaxia.org/index.php?title=Jacques_Rueff:La_fin_de_l%27%C3%A8re_keyn%C3%A9sienne.

Elle a été publiée dans Le Monde les 19 et 20/21 février 1976. Version anglaise: "The End of the Keynesian Era or: When the Long Run Ran Out", Euromoney, avril 1976, pp.70-7).
Le texte a été réédité sous une forme plus longue intégrant le texte de 1931 sur le "chômage permanent" dans Emil-Maria Claassen et Georges Lane (eds.), Oeuvres complètes de Jacques Rueff, Tome III: Politique économique, livre 1, Plon, Paris, 1979, pp.161-178.

J'ai pu faire apparaître et constater qu'effectivement la "loi de Rueff" s'appliquait en France.

Bien évidemment, ces textes sont des explications de la situation de la France d'aujourd'hui et, en particulier, du texte de 2005 intitulé "Le fardeau de la dette publique et le poids de nos impôts ... qui ont fait basculer la France en queue du peloton européen".


Addendum.

Ce texte sur les "Unemployment Benefits: The Good and The Bad; Robert Moffitt, Johns Hopkins University" de février 2014.




Retour au sommaire