15
NINETEENTH EUROPEAN CONGRESS OF THE REGIONAL SCIENCE ASSOCIATION SPATIAL VARIATIONS IN UNEMPLOYMENT PROBLEMS: A TYPOLOGY by D. Gleave and D. Palmer* 1. INTRODUCTION Most western industrial nations are experiencing their highest levels of sustained unemployment for over forty years. Events in Britain have been typical of this trend with unemployment exceeding one million workers since 1975. This situation may worsen as the number of school leavers entering the labour market continues to be greater than the numbers retiring, as female participation rates continue to increase and as a consequence of technological innovation. Yet, at the same time certain sectors face labour shortages, particularly in skilled occupations. This paradox lies at the heart of the match- making problem and is succinctly illustrated by the shift in the relationship between unemployment and vacancies shown in Figure 1. Since 1963 three clear outward shifts in the UV curve have occurred, indi- cating that for any given number of vacant jobs more people are registered as unemployed. A continuing debate on the nature and causes of this trend shows that little agreement exists amongst labour economists. Some have proposed explanations which take account of demand factors, others suggest supply factors predominate whilst many take account of changes in legislation on both labour demand and supply (see Palmer and Gleave [30] for a review of the debate but also see -- in chronological order -- [6]; [20]; [1]; [17]; [18]; [2]; [9]; [19]; [24]; [27]; [10]; [7]; [5]; [11]; [8]; [22]; [23]. The significance of this debate is critical in its impact upon suggested policy initiatives. If the balance of causes is incorrectly specified then the efficacy of related policy measures will be in question. Often, the debate has centred around distinguishing and measuring different types of unemployment. Such unemployment typologies can serve as a useful set of pigeon holes into which different causes can be allocated and thereby allows debate about policy mix for different types of labour markets. A sensibly specified unemployment classification also provides a way of breaking down the very complex problem of unemployment causation so that specific types of cause can be tested by examining trends in the appropriate sub-group of unemployed persons. This paper is concerned with developing and describing a new methodology for identifying different unemployment types. In the next section we begin by reviewing the standard methods of classification used at present and suggest why they are inadequate. We then outline a new and more specific system of * The authors are associated with the Centre for Environmental Studies, London, U.K.

Spatial variations in unemployment problems: A typology

Embed Size (px)

Citation preview

Page 1: Spatial variations in unemployment problems: A typology

NINETEENTH EUROPEAN CONGRESS OF THE REGIONAL SCIENCE ASSOCIATION

SPATIAL VARIATIONS IN UNEMPLOYMENT PROBLEMS: A TYPOLOGY

by D. Gleave and D. Palmer*

1. INTRODUCTION Most western industrial nations are experiencing their highest levels of

sustained unemployment for over forty years. Events in Britain have been typical of this trend with unemployment exceeding one million workers since 1975. This situation may worsen as the number of school leavers entering the labour market continues to be greater than the numbers retiring, as female participation rates continue to increase and as a consequence of technological innovation. Yet, at the same time certain sectors face labour shortages, particularly in skilled occupations. This paradox lies at the heart of the match- making problem and is succinctly illustrated by the shift in the relationship between unemployment and vacancies shown in Figure 1.

Since 1963 three clear outward shifts in the U V curve have occurred, indi- cating that for any given number of vacant jobs more people are registered as unemployed. A continuing debate on the nature and causes of this trend shows that little agreement exists amongst labour economists. Some have proposed explanations which take account of demand factors, others suggest supply factors predominate whilst many take account of changes in legislation on both labour demand and supply (see Palmer and Gleave [30] for a review of the debate but also see - - in chronological order - - [6]; [20]; [1]; [17]; [18]; [2]; [9]; [19]; [24]; [27]; [10]; [7]; [5]; [11]; [8]; [22]; [23].

The significance of this debate is critical in its impact upon suggested policy initiatives. If the balance of causes is incorrectly specified then the efficacy of related policy measures will be in question. Often, the debate has centred around distinguishing and measuring different types of unemployment. Such unemployment typologies can serve as a useful set of pigeon holes into which different causes can be allocated and thereby allows debate about policy mix for different types of labour markets. A sensibly specified unemployment classification also provides a way of breaking down the very complex problem of unemployment causation so that specific types of cause can be tested by examining trends in the appropriate sub-group of unemployed persons. This paper is concerned with developing and describing a new methodology for identifying different unemployment types. In the next section we begin by reviewing the standard methods of classification used at present and suggest why they are inadequate. We then outline a new and more specific system of

* The au thors are associa ted with the Centre for Envi ronmenta l Studies, London , U .K .

Page 2: Spatial variations in unemployment problems: A typology

58 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

Vacancies (thousands)

300

iOO

/965j

1973j Key: J = June D = December

FIGURE 1.

1976j

I I I I t I I I I 0.3 0.4 0.5 0.6 0,7 0,8 0.9 i.O 1,1 1,2 1,3

Unemoloymen t (millions)

Source: DEG Bi-annual Data

The UV Relationship, June 1963-December 1978

I, 1 . 4

classification. This is followed by an application of our methodology to quarterly unemployment data from the United Kingdom for the period from December 1973 to March 1979. Finally, we critically examine our method and discuss the implications for policy of the results we have produced.

2. TYPES OF UNEMPLOYMENT AND THE UNEMPLOYMENT PROBLEM There have been many attempts to classify unemployment types by labour

economists. In general three classes have been used most frequently although there has been controversy on how they should be precisely defined (see Lipsey [26]). The classes are frictional unemployment, structural unemployment and demand deficient unemployment. Clearly this classification requires at least two levels of enumeration in order that the structural component can be specified. This is rarely explicitly acknowledged. Implicitly structure is usually a synonym for occupational composition.

Demand deficient unemployment occurs when the total number unem- ployed exceeds the total of vacant jobs. On the assumption of an accurate measure of both unemployment and vacancies, it can be considered to measure the strength of an economic depression. However, the relationship is not simple, for often when a household's single breadwinner becomes unemployed, other members of the household register for work and remain registered until a new job is secured by the main breadwinner. This is known as the added worker

Page 3: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 59

effect (see Stilwell [34]; Hyman and Palmer [23]) and can distort the number who are unemployed due to demand deficiency. If a surplus of vacant jobs occurs this would simply define an aggregate labour shortage.

Structural unemployment is ambiguously defined (see Gleave [11]) and depends, inter alia, on whether demand deficient unemployment occurs. If it does structural unemployment exists if, in some occupations, vacancies exceed the; number unemployed. That is to say structural unemployment exists because an unspecified proportion of the unemployed work force does not possess particular skills for which there is a demand. On the other hand if the total of vacant jobs is greater than the number out of work, i.e. if there is no demand deficiency, structural unemployment exists in those occupations where the number jobless exceeds the number Of vacancies. In this case structural un- employment occurs because an unspecified proportion of the unemployed do

possess skills for which there is no demand. It is this ambiguity which has led to argument about the boundary between structural and demand deficient un- employment.

The 'irreducible minimum' of frictional unemployment obtains when un- ernployment and vacancies occur simultaneously within the same occupation. It is measured as the sum through all occupations of the minimum number of either unemployed workers or vacant jobs. Although this class of unemploy- ment can be attributed to a number of causes, which may also explain the shift in the U V curve, it is usually related to the time lags involved in matching workers to jobs. Thus, frictional unemployment occurs during the search process when unemployed workers and employers with vacancies are seeking to learn of each other's existence and will continue until they agree upon an acceptable wage (and other conditions of employment) to secure a match. The reduction of frictional unemployment therefore defines the 'soft core' of the matchmaking problem at the national level. Some say it must exist to lubricate the labour market and ensure that inflation is kept within politically acceptable bounds.

The measurement of unemployment types can be used to guide policy. For example, during the depression of the 1930s it was noted that "it is usual to attribute a large amount of unemployment to 'frictions' which prevent workers from moving readily from one occupation or locality to another, and remedies for unemployment are sought in training schemes, in providing facilities for transfer and so forth . . . . Lack of mobility can be called a cause of unemploy- ment only when there are unfilled vacancies in some places and idle men in others . . . . Schemes to promote mobility are all to the good, but there is no remedy for immobility so effective as the development of boom conditions" (Robinson [31]).

We agree that the provision of vacant jobs acts as the best catalyst for the matchmaking process, encouraging individual workers to seek better jobs (see Gleave and Cordey-Hayes [12]). Nonetheless, the shift in the U V curve suggests that the amount of unemployment that can be attributed to frictions and struc-

Page 4: Spatial variations in unemployment problems: A typology

60 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

tural causes is increasing. One long run factor which may be affecting the shift is the change in 'life expectancy' of many jobs which seems to be decreasing. This results in an increased need to retrain labour and thereby provide the op- portunity for increased occupational mobility. The rapid restructuring of the demand for labour by occupation cannot be satisfied only by training school leavers in those skills for which there is a high demand and hoping that the retirement process reduces supply sufficiently in declining occupations.

Secondly, changes in the spatial distribution of employment opportuni- ties are continuously occurring at the inter-national, inter-regional and intra- regional scales. Within Britain, on balance, jobs are being lost in the peripheral regions and gained in the south of the country. Furthermore older industrial centres and inner city areas are losing jobs to smaller towns, suburban and rural areas. These changes can be accommodated by longer commuting and labour migration but there exist serious constraints which inhibit spatial readjustment. For example, long distance migration by tenants within the public housing sector is difficult to undertake (see Gleave and Palmer [14]).

Whilst Robinson was correct in arguing that increased demand will resolve many problems pertaining to unemployment it is clear that immobility is a significant cause of unemployment 'frictions.' Monetary and fiscal policies, which are used to generate economic activity through increased demand and reduce demand-deficient unemployment, also create inflationary problems be- cause they affect different areas and sectors in varying ways. In addition, the trend of employment decline in manufacturing industry alongside growth in services has accelerated over the last decade. When these shifts have been spatially coincident, such as in inner cities, serious problems of unemployment arise associated with occupational immobility. Thus the analysis of the match- ing of workers to jobs must be considered at the scale of the local labour market. The unemployment problem is not homogeneous throughout the space economy and therefore the causes of unemployment vary between different local labour markets.

The classification method outlined above does not enable us to consider this problem satisfactorily. Unemployment due to geographic immobility is subsumed within the frictional component and cannot easily be identified. Structural unemployment varies with the level of activity in the national labour market such that a very ambiguous dividing line exists between all three categories. Structural unemployment, for example, is defined to be at a maximum when U 99 is zero, while frictional unemployment varies depending upon the level of the total number of jobless. Furthermore only an approximate distinc- tion may be made between the levels of U F and U s. Quantitative estimates of the relative importance of the types are very difficult to make because of the hazy boundaries and also because they depend upon the reliability of data on notified vacancies as an indication of labour demand. It is well recognised that vacancy statistics are unreliable and may only represent 50 percent of labour

Page 5: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 61

recruitment (see Thomson [35]; Mackay et al. [28]). Consequently estimates of the different types of unemployment are likely to be unreliable as a guide to both national and sub-national economic policy (see Gleave [11]). We now out- line an alternative method which will permit a robust quantitative estimate of different categories of unemployment - - a necessary prerequisite for studies of local labour markets.

3. AN A L T E R N A T I V E C L A S S I F I C A T I O N OF U N E M P L O Y M E N T TYPES The criticisms of the classification system described above are numerous.

We noted that it is difficult to define the boundaries between the classes with precision yet this is what the method does. It does not enumerate persons within the 'unemployables' class but since this was not attempted it cannot be held as a real criticism. (On the basis of official published data this class could only be approximately measured by taking into account unemployment durations). Secondly, we argued that the method of measuring structural unemployment is ambiguous because its specification depends upon the presence or absence of demand deficiency. This shortcoming does not satisfy the ideal requirements of a clear, exhaustive taxonomy (see Grigg [16]; Sokal and Sneath [33]). Thirdly, in our opinion an important cause of structural mismatch is the changing spatial distributions of homes and jobs. In the classification system described above this structural type is subsumed within the class of frictional unemployment. Finally, we think that mismatch, and consequently a typology of unemployment, should be given context by providing information about the number of persons in work. This is important because it relates to explanations of unemployment such as change in demographic structure.

An alternative classification must take account of the above shortcomings, explanations of causality and the variable dimensions in which the latter can be expressed. Starting with a microeconomic perspective we believe that workers, whether unemployed or in a job, possess two main characteristics which constrain mobility and may exacerbate mismatch. These characteristics are spatial location and skill. Of course, other frictions occur within the labour market relating to wage adjustment and job search. However, if changing house and changing occupation are the major constraints upon efficient matchmaking, then the following simple classes of unemployment occur: Frictional unemploy- ment, occupational-structural unemployment, spatial-structural unemploy- ment, spatial/occupational unemployment and demand deficient unemployment.

This simple system, which does not take account of some of the criticisms outlined above, is particularly interesting for one important reason. The order in which the unemployment types are measured affects the outcome. That is to say, the classification implicitly takes account of the fuzzy nature of spatial and occupational structure and provides the policy maker with a choice in deciding how unemployment should be reduced. We suggest two alternative routes for measuring unemployment types:

Page 6: Spatial variations in unemployment problems: A typology

62 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

Method A

Level 1 A (i) Frictional B (i) Level 2 A (ii) Spatial-Structural B (ii)

A (iii) Occupational-Structural B (iii) A (iv) Spatial/Occupational- B (iv)

Structural Level 3 A (v) Demand-Deficiency B (v)

The explanation of this hierarchy is now outlined in particular types of unemployment.

Method B

Frictional Occupational-Structural Spatial-Structural Spatial/Occupational- Structural Demand-Deficiency

a detailed discussion of

Frictional unemployment (U F) Unemployed workers who could be employed in the same labour market

and occupation because there exist sufficient vacancies are defined as fric- tionally unemployed. They define the 'soft core' of the matchmaking problem and may be described as 'mismatch unemployment.' This does not mean that it is necessarily easy to place them in jobs but merely states that in general they do not depend on policies aimed at boosting aggregate economic activity, retraining or allowing inter-regional migration in order to reduce unemployment. Such unemployment which could, theoretically, be solved since sufficient vacancies exist, can arise for a variety of reasons. Employers may only be offering low wages for the job while potential employees may demand a too high reservation wage. Alternatively, such jobs may be inaccessible and involve high cost commuting or require short-range migration which may imply that the employer is located wrongly. Lack of information about such vacancies may be due to employers using the wrong advertising media. Furthermore, the jobs may involve anti-social hours or other poor working conditions whilst employers may complain of the unreasonable demands of workers. Finally, employers may discriminate against some sections of the community possibly due to past experience when similar individuals were unsatisfactory.

Frictional unemployment is the category that is calculated first by the algorithm and is therefore defined as a Level 1 type. It may be formally defined as:

U F = min (Uir, Vir) for all i,r (1)

where U~ is frictional unemployment in occupation i and region r. U~r is total unemployment in occupation i and region r. V~r is total vacancies in occupation i and region r.

Upon completion of this stage we must correspondingly reduce the levels of total unemployment and vacancies by U~ before allocating the remaining unemployment to other categories. Thus:

U~ = U,~- U~ (2)

v~r = v~r - ufr (3)

Page 7: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 63

Structural unemployment can be divided into three groups of which the first we discuss is Spatial-Structural unemployment.

Spatial-Structural unemployment (U ss) Other schemes of categorisation have not normally explicitly identified

those who are jobless due to their inability to adjust to the changing fortunes of different labour markets. We specifically identify this group since geographic mobility is expensive for individuals to undertake within the private housing market and we have provided evidence that within the growing public housing sector migration is particularly constrained (see Gleave and Palmer [14]; [15]). This category separates the numbers who could find a job in the same occupation in a different labour market if they were prepared, or not constrained, from migrating to other areas. If we follow the hierarchy defined as method A (see above) we may define Spatial-Structural unemployment:

u f ~ = u~r , ZrU~r • Z~Vir for all Z~V~ > 0,X~U~r - Z~V~r (4)

= U~ for all ]s > Z~U~ (5)

= 0 for all X~V[~ = 0 (6)

where U sff is Spatial-Structural unemployment in occupation i and region r. We must again adjust the totals of unemployment and vacancies before under- taking the next stage. Hence:

U~" = U~r - U sf (7)

V"r = 0 for all ZrV~r <- ZrU~ (8)

] , [ V~r = Vi~ - [ Z~V~ x X~U~ for all Z~V,~ > Z~U~r (9)

Occupational-Structural unemployment (U ~ This second type of structural unemployment is complementary to the first

and is broadly similar to the definition used by other analysts. Those who are currently jobless but who could gain employment within their region if they could be retrained and learn another skill define U ~ Formally:

u f ~ - U~'r "ZiU~-----~ x "s for all XiV~'~ > O, XiU~'~ -> XiV~" (10)

= U~'~ for all Z~V~" > Z~U~" (11)

= 0 for all ZiV~" = 0 (12)

where U ~ is Occupational-Structural unemployment in occupation i in region r.

We must again reduce the totals of unemployment and vacancies before we proceed with the next stage. Thus:

Page 8: Spatial variations in unemployment problems: A typology

64 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

u~" = u~" - u ~

V}" = 0 for all ~V~" <- iU~"

= Vi'r - [ v~'~ ,,] ,, I ~ x XiUi,. for all XiV~'~ > XiUir t_ i i r

(13)

(14)

(15)

The hierarchy of unemployment types defined previously allowed for two alternatives. Either U ss could be calculated first at Level 2 or U ~ could. This means we assume that both types of unemployment are of equal importance. If one applies Method A, the algorithm will calculate the maximum level of U ss and the minimum level of U ~ Alternatively, Method B calculates the maxi- mum level of U ~ but the minimum of U ss. By using both approaches one is able to generate a range of values for U ~ and U ss, the importance of which will be discussed at the end of this section. Both methods place the third type of structural unemployment at a lower level in the hierarchy.

Spat ia l~Occupat ional -S truc tural u n e m p l o y m e n t ( U s~

This category contains those who would need to change both region and occupation in order to obtain employment. Because it requires a change in both attributes of labour any policies would be more expensive than those that either assist in relocating or retraining workers. Thus U s~ is lower in the hierarchy than the other two types of spatial unemployment. Formally:

u~os _ u~" XiX~uig x "272rVi"

U t t t = ir

= 0

for all 5;~X~V~'~ --- O, (16)

for all X~XrV~ ' > "Z~'s (17)

for all XiXrV}~ = 0 (18)

where U s~ = Spatial/Occupational-Structural unemployment in occupation i and region r.

Again we must adjust the total unemployment and vacancies prior to com- pleting the final stage. Thus:

g~"/ = g~" - g s~ (19)

V~ = 0 for all X~XrV~" -< X~XrU~" (20)

[ V~'~ x 1~i~rU";] for all ~,?~V~" > "2,'2rU~" (21) = V i " - - - , , , k J

D e m a n d - D e f i c i e n t u n e m p l o y m e n t ( U DD) As with previous classifications Demand-Deficient unemployment exists

when the total level of unemployment exceeds the total number of vacancies and is defined as the difference between them. Alternatively, there may be situations when supply deficiency exists, that is when the level of notified vacancies exceeds total unemployment. Such problems of demand or supply

Page 9: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 65

deficiency require efficient policies of aggregate demand management in order to be solved. Our new system allows us to identify both those regions and those occupations where demand-deficiency may exist. Hence:

g ~ : Ui~ for s163 > 0 (22)

VSr ~ = V ~ for "s163 > O (23)

where U~ ~ is demand-deficient unemployment in occupation i in region r.

VSf is excess-vacancies (supply-deficiency) in occupation i in region r.

This simple classification system identifies two conceptually distinct types of structural unemployment which are represented in Figure 2 as two fuzzy sets. Frictional unemployment and Demand-Deficient unemployment are still represented as precise concepts which provide unique estimates when the measurement model is applied. However, it is quite easy to represent the inter- face between frictional unemployment and structural unemployment as a fuzzy intersection. This can also be done regarding structural unemployment and demand deficiency although in both cases the information required for compu- tation is considerable. Nonetheless, we explain how this could be done.

So far, it has been assumed that both labour markets and occupational classes can be defined unambiguously. Thus if, within a given occupation, an unemployed worker and vacant job occur within the same local labour market then frictional unemployment results. However, the worker and the job could well be located at opposite ends of the local area such that the 'employment space' of the former and the 'commuter shed' of the latter do not overlap (see Hyman and Gleave [21]). In this case it would be more realistic to say that Spatial-Structural unemployment was occurring. Had the vacant job been located in an adjacent labour market, but within the employment space of the worker concerned, then our classification system would suggest Spatial- Structural unemployment when clearly it should be frictional. This fuzziness also occurs because some workers can change occupation without retraining.

..::::~.~:;.'.:.:~.:.'.:.:.:.:

~ { g ! :

~ { i ~ u F ~Ai* ~ .......... ~:.'..~t

~?. : :?~: ' . : ! : . : . : . : . : . : ; : . : . : . : . : '.::::::i:.. :.42:.:.:.:.:.:::!~:::::: '.:.:.:.:.:.:~ f:::.:i:i:::_:~ f-:u

- ..'...:::::::::::::%.:: �9 ..~....:::...'.~::

~ . :,:i::.::., ::::i?i:~'!: ~.'.h:i'.'..".:_:::'..::, :::::::::::::::::::::::::::::::

~ ' ~ . : :

i:::.:':'::.%i; u i:i:'.-:h::"

~ i i ~ . . : : : " ~}..;i~.,-.

FIGURE 2. The Sets of Unemployment Types

Page 10: Spatial variations in unemployment problems: A typology

66 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

In such circumstances Occupational-Structural unemployment would be sug- gested instead of Frictional unemployment. A probability approach towards 'suitable' vacancies and based on a method outlined in Gleave [11] is one way to resolve this problem.

The interaction between demand deficiency and structural unemployment is more difficult to explain. The nature and the extent of overlap depends clearly upon the changes through time in the occupation specific demand for labour. For a fuller discussion on this see Cheshire [3] and Palmer and Gleave [30].

Returning to our algorithm two important problems remain. Firstly, it is dependent upon the quality of the data. Overenumeration of unemployment or underenumeration of vacancies will lead to spurious results. In this sense it is no different to other systems and one can only hope that the data will improve. Secondly, the system is sensitive to the numbers of regions and occupations included in the analysis. Ideally we require data collected upon a definition of spatial labour markets utilising self-containment measures such as those proposed by Smart [32]. Furthermore the occupational classification needs to distinguish between those skill groups which it is easy to move between and those which it is not. While the CODOT scheme [4] is a considerable improve- ment upon the Census definition [29], the classification by Gleave and Palmer[13] is probably more useful. Unfortunately the empirical study, reported in the following sections, could only be undertaken using data dissagregated for the standard planning regions and the 'Key Occupations for Statistical Purposes.'

4. APPLICATION TO UK DATA The results reported in this section were based upon quarterly unemploy-

ment and vacancy data for the period 1973-iv to 1979-i disaggregated at the regional scale. The data cover one complete cycle but one in which a rapid worsening of the British economy was occurring. During the period unemploy- ment levels increased dramatically with no comparable increase in vacancies. The UV relationship (see Figure 3) started with a peak in vacancies and finished with a peak but also exhibited a significant rightwards shift. Ideally, we would have analysed data at the scale of the local labour market which would un- questionably have reduced the amount of frictional unemployment and in- creased Spatial-Structural unemployment. For these reasons our findings must be regarded as illustrating our method rather than providing firm conclusions.

The most important type of unemployment for most of the period analysed was demand deficiency (see Figure 4). However, this was in part due to the fact that vacancies are seriously under-notified and for policy purposes cannot be used as a very reliable estimate of labour demand (Leicester [25]) unless cor- recting coefficients are applied to each occupation. Frictional unemployment remained fairly stable in terms of number but varied considerably as a propor- tion of total unemployment and by region. As a proportion it was greatest in 1973-iv and 1974-ii representing 42.7% and 41.2% of all unemployment. It was smallest in the trough of the cycle representing only 9.2% of unemployment in

Page 11: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 67

< Q m m Q

Z

0 rn

0

�9

0

0

1973-iv

< \ ~ % \

% \ \ \\\ _~979-i

'k___

I ! I

0.5 1.0 1.5

Unemployment in millions

FIGURE 3. The UV Relationship 1973-79

1.5

Unemployment (Millions)

i.O

0.5

Total Unemployment/

, / / / /

/ / / /

/ Spatial and Occupational Unemployment

Demand-Deficient Unemployment

J

Frictional Unemployment

1974 1975 1976 1977 1978

FIGURE 4. Types of Unemployment 1973iv-1979i 1979

Page 12: Spatial variations in unemployment problems: A typology

68 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

1975-iv. On a regional basis frictional unemployment always represented the highest proportion of unemployment in the south east ranging from a high of 78.6% (1973-iv) to a low of 15.4% (1975-iv). Conversely in Northern Ireland, where demand deficiency was always a problem throughout the period fric- tional unemployment accounted for between 14.9% (1974-ii) and 2.1% (1979-iii). This last observation was interesting because it represented a steady decline in the situation in Northern Ireland and is particularly significant because we observed a steady relationship between the numbers frictionally unemployed and the number of vacant jobs. In all regions of the United Kingdom the number of vacancies at the second peak of the cycle was much less than at the first (see Figure 3). In most regions there was an increase in the proportion of 'fric- tions' over the period 1977-i to 1979-i but this was not observed in the Northern region nor in Northern Ireland where the proportion declined. Comparing the regions, although in 1979 they all experienced lower proportions frictionally unemployed than in 1973, in most regions a convergence occurred, with the West Midlands declining quite markedly and Scotland improving. The only excep- tions to convergence were the South East and Northern Ireland. These 'best' and 'worst' regions respectively were both diverging away from other regions which, because of the association between frictions and vacancies really indi- cates where the demand for labour is greatest and weakest.

Maximum Spatial-Structural unemployment was examined both in terms of the gross numbers of unemployed who could be found jobs by migrating and as a proportion of the residual unemployed after frictions had been removed. The first point to re-emphasise is that the estimates for this particular class depend very heavily upon the spatial disaggregation. We used only eleven units with an average area of 8727 square miles and a mean diameter of 104.4 miles. Unfortunately these units are larger than the employment space of the long distance commuters and cannot be said to represent labour market areas. None- theless, the results are worthy of comment because of the time trend in spatial mismatch and its regional characteristics.

First, it was observed that spatial unemployment is only of major impor- tance when the demand for labour is high. This suggests that the best time for redistributing workers occurs at the peak of the trade cycle when vacancies are numerous. In the trough of the cycle, from 1976-ii to late 1977, all vacancies were associated with frictional unemployment. That is to say, in every region and every occupation unemployment exceeded vacancies. The interesting question is how does this demand deficiency translate into structural unemploy- ment (both occupational and spatial) when labour demand picks up? From the data we analysed the answer, quite clearly, is that it depends upon the strength of demand at the peak of the cycle. We cannot comment upon this in detail because of poor vacancy data but do confirm that in late 1973 and mid 1974 on the basis of published information the amount of spatial mismatch was consider- able accounting for at least 12.9% of residual unemployed persons. However, the 1979 peak of the cycle indicates little Spatial-Structural unemployment,

Page 13: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 69

largely because frictional unemployment held up, but also because demand deficiency was still enormous. At the peak of the cycle in 1973/4 vacancies exceeded unemployment only in the South East region - - for four quarters - - and in East Anglia - - for two quarters. This did not occur at all in 1979. After allowing for frictional unemployment there were residual vacancies in all regions except Northern Ireland for between four and seven quarters up to late 1975. As the second peak approach there were residual vacancies in the South East for seven quarters and in East Anglia and the East Midlands for two quarters. From later 1975 to mid 1977 no spatial unemployment occurred because of de- mand deficiency.

During 1974 and the first half of 1975 mismatch could have been reduced in all regions except Northern Ireland by both in-migration and out-migration. Northern Ireland had no vacant jobs for in-migrants but could have reduced mismatch among the residual unemployed by up to 41.4% by out-migration. As the downturn set in the first thing to happen was that the more peripheral regions could, like Northern Ireland, benefit only by out-migration. However there were exceptions. For example, Northern England had a surfeit of vacancies in security and protective services! However, on the upturn a completely different pattern emerged. There were few cases where mismatch could be removed by both out-migration and in-migration. Most of the time this could only have been achieved by one way flows into the South East. In two quarters flows into and out of East Anglia and the East Midlands could have partially ameliorated the unemployment problem.

The implication of these observations was that whilst variations in the demand for labour lead inevitably to a shift in the distributions of workers and jobs, at the 1973-74 peak increased efficiency could have been achieved by two-way migration between all regions except Northern Ireland. By 1979 it could only have been achieved by one way flows from all regions into the South East. Quite clearly, improved efficiency in matching workers and jobs appears achievable only by decreased efficiency in the use of fixed capital. More spe- cifically, this situation must exacerbate local pressures in the housing market. However, it is quite possible that some improvement in mismatch could be achieved by two-way migration but this is hidden by poor vacancy data.

5. FURTHER THOUGHTS If it can be assumed that good data support the hypothesis that spatial

variations in the demand for and supply of labour contribute to an inefficient use of resources then there are clear implications concerning those distribu- tions. A centrally planned economy can directly control the distributions con- cerned; a market economy can remove constraints to mobility to ensure that structural rigidities do not prevent workers responding to job opportunities or employers creating work in areas with a plentiful supply of labour. As far as the recent economic history of the UK is concerned, full e m p l o y m e n t - a similar criterion to efficient match m a k i n g - has dominated the minds of

Page 14: Spatial variations in unemployment problems: A typology

70 PAPERS OF THE REGIONAL SCIENCE ASSOCIATION, VOL. 44, 1980

politicians and policy makers alike. However there has been a certain schizo- phrenia as to how this should be achieved. On the one hand, there have been attempts to influence the location of job opportunities, but not (effectively) to influence the location of workers. On the other, full employment has not been considered in a spatial context at all but attacked from a neo-Keynesian demand management position. In our view, the second perspective is concerned with the major problem, that of demand deficiency. As far as the former perspective is concerned it seems quite reasonable to suggest that the spatial pattern of labour supply should in the first instance be affected by removing constraints on mobility and actively supporting voluntary migration between labour markets.

REFERENCES

[1] Bowers, J. K., P. C. Cheshire and A. E. Webb. "The Change in the Relationship between Un- employment and Earnings Increases: A Review of Some Possible Explanations,National Insti- tute Economic Review, No. 54 (November 1970), pp 44-63.

[2] and R. Weeden. "Some Aspects of Unemployment and the Labour Market, 1966-71" National Institute Economic Review No. 62 (November 1972), pp. 75-88.

[3] Cheshire, P. C. "Regional Unemployment Differences in Great Britain" National Institute of Economic and Social Research Regional papers II. Cambridge University Press, 1973.

[4] Department •f Emp••yment Classi•cati•n •f •ccupati•ns and Direct•ry •f •ccupati•nal Titles (CODOT) Three Volumes. London: HMSO, 1972.

[5] Department of Employment ~ Changed Relationship between Unemployment and Vacancies ' ' Department of Employment Gazette Vol. LXXXIV (1976) pp. 1093-1099.

[6] Dow, J. C. R. and L.'A. Dicks-Mireaux. "The Excess Demand for Labour: A Study of Con- ditions in Great Britain 1946-56" Oxford Economic Papers (New Series), Vol. 10 (1958).

[7] Evans, G. J. "A Note on Trends in the Relationship between Unemployment and Unfilled Vacancies," The Economic Journal Vol. 85 (1975).

[8] Evans, A. "Notes on the Changing Relationship between Unemployment and Notified Vacancies: 1961-1966 and 1966-1971" Economica, Vol. 44 (1977).

[9] Foster, J. I. "The Behaviour of Unemployment and Unfilled Vacancies: Great Britain 1958-71 - - A Comment" Economic Journal, Vol. 83 (1973).

[10] - - . "The Relationship between Unemployment and Vacancies in Great Britain 1958-72: Some Further Evidence," in D. Laidler and D. L. Purdy (Editors)Inflation and Labour Markets. Manchester University Press, 1974.

[11] Gleave, D. "Unemployment Types, Vacancy Data and the Demand for Labour." CES Working Note WN 457, London: Centre for Environmental Studies, 1977.

[12] - - and M. Cordey-Hayes. "Migration Dynamics and Labour Market Turnover." Prog- ress in Planning Vol. 8 Part 1, D. R. Diamond and J. B. McLoughlin (Editors) Oxford: Pergamon Press, 1977.

[13] - - and D. J. Palmer. "Labour Mobility and the Dynamics of Labour Market Turnover." Paper presented to the European Section of the Regional Science Association, Krakow, Poland, August 1977.

[14] ~ . "Mobility of Labour: Are Council Tenants Really Handicapped?" CES Review Issue

3, 1978. [15] - - . "Mobility Constraints and Unemployment Mismatch: An Analysis of Structural

Change in the Housing Market and their Consequences in Job Migration." Paper presented to the Fourth Urban Economics Conference, University of Manchester, July 1979.

[16] Grigg, D. "Regions, Models and Classes," in R. J. Chorley and P. Haggett (Editors) Models in Geography. London: Methuen, 1967.

Page 15: Spatial variations in unemployment problems: A typology

GLEAVE AND PALMER: UNEMPLOYMENT PROBLEMS 71

[17] Gujarati, D. "The Behaviour of Unemployment and Unfilled Vacancies: Great Britain, 1958- 1971," The Economic Journal, Vol. 82 (1972).

[18] - - . "A Reply to Mr. Taylor," The Economic Journal Vol. 82 (1972), pp 1365-1368. [19] - - . "A Reply to Mr. Foster," The Economic Journal Vol. 8 (1973), pp 201-203. [20] Hansen, B. "Excess Demand, Unemployment, Vacancies and Wages," The Quarterly Journal

of Economics Vol. 84 (1970), pp 1-23. [21] Hyman, G. and D. Gleave. "A Reasonable Theory of Migration," Transactions of the Institute

of British Geographers Vol. 3 (1978), pp 179-201. [22] Hyman, G. and D. Palmer (1978). "A Regional Analysis of the Dependency between Registered

Unemployment and the Stocks and Flows of Notified Vacancies," Environment and Planning A, Vol. 10 (1978), pp 853-866.

[23] - - . "A Causal Analysis of Unemployment and Vacancies in British Regions 1969-74." CES Working Note WN 535. London: Centre for Environmental Studies, 1979.

[24] Leicester, C. "Vacancies and the Demand for Labour," Institute of Manpower Studies Monitor, Vol. 2 (1973).

[25] - - . "Some Earlier Theories of the UV Curve." Institute of Manpower Studies Monitor, Vol. 3 (1974).

[26] Lipsey, R. G. "Structural and Demand-Deficient Unemployment Reconsidered," in A. Ross (Editor)Employment Policy and the Labour Market. Berkeley: California University Press, 1965.

[27] Llewellyn, D. T. and P. Newbold. "The Behaviour of Unemployment and Unfilled Vacancies," Industrial Relations Journal, Vol. 4 (1973) pp 30-42.

[28] MacKay, D. I., D. Boddy, J. Brack, J. A. Diack and N. Jones. Labour Markets under Dif- ferent Employment Conditions. London: Allen and Unwin, 1971.

[29] Office of Population and Census Studies. Classification of Occupations 1970. London: HMSO, 1970.

[30] Palmer, D. and D. Gleave. "The Dynamics of the Labour Market: Matching Workers to Jobs," forthcoming.

[31] Robinson, J. Introduction to the Theory of Employment. London: Macmillan, 1937. [32] Smart, M. W. "Labour Market Areas: Uses and Definitions." Progress in Planning Vol. 2

Part 4, D. R. Diamond and J. B. McLoughlin (Editors), Oxford: Pergamon Press, 1974. [33] Sokal, R. R. and P. H. A. Sneath Principles of Numerical Taxonomy. San Francisco: Freeman,

1963. [34] Stilwell, F. J. B. "The Regional Distribution of Concealed Unemployment," Urban Studies

Vol. 7 (1970). [35] Thomson, W. "Collection and Use of Job Vacancy Data in Canada," in National Bureau of

Economic Research The Measurement and Interpretation of Job Vacancies, 1966. [36] Thirlwall, A. P. Types of Unemployment in the Regions of Great Britain, Manchester School

No. 4 (1974).