 |
||||||||
|
||||||||
|
|
A New Province for Law and OrderDeregulation and Employment in New ZealandGreg Cutbush1. Introduction The topic I was given for this paper is clearly very broad. Taking advantage of the available degrees of freedom, I have interpreted my duty as:
These findings will be placed in context below.
Environment What deregulation of labour has there been? The Employment Contracts Act (ECA) was just one change of several in recent times:
Furthermore:
Thus there were a few years of significant labour reforms in New Zealand up to 15 May 1991. In a well known paper, presented in July this year2 Alan Jones identified four "pillars of protection" associated with the industrial relations system prior to the reforms: compulsory unionism; monopoly coverage of work (and so workers); blanket coverage; and compulsory arbitration. It is the removal of all four of these pillars in the ECA which was the real achievement. For balance I will mention that there have been other (albeit rare) periods of liberalisation of labour rules in New Zealand. In 1932, for example, as the Great Depression worsened, compulsory arbitration was abandoned after nearly 40 years in operation. (It was, however, reintroduced four years later.)
Table 1: Expenditure on Active Labour Market Programmes as Percentage of GDP (Various OECD countries, 1990/91)
3. Labour Data - Sources 3.1 The NZ Department of Statistics The following table summarises the main NZ Department of Statistics data on labour in New Zealand (Table 2).
The Household Employment and Income Survey is acknowledged to be the most useful source of information. 3.2 Department of Labour The Department of Labour has some additional information of interest. For example, it has surveyed contracts registered (ie those ³ 20 employees). The first survey, after the ECA was passed, looked at 190 contracts. Each survey, apparently, shows much the same story, ie.:
However the proportion of contracts covered by the surveys is less than ten per cent.
3.3 Miscellaneous and secondary studies There are, of course, several additional ad hoc sources of information on the operation of the NZ labour market in recent years. One example has been Harbridge and Moulder's survey of 471 contracts at the end of 1991. They found that 51 percent had enjoyed pay increases. Another is the NZ Employers Federation's survey during 1991 of 1116 employers covering 190,664 employees--- this showed that 34 per cent of workers had stuck to the awards, but that there has been a great increase in the proportion of agreements tailored specially to the enterprise. The New Zealand Treasury has embarked on two main initiatives. One is the construction of a model to predict several security outlays under different social security and wage regimes. It is called TAXMOD.3 Another is the work on the ECA the NZ Treasury has commissioned---it is an econometric study of the impact of the ECA, though is not expected to be ready until midÄ1993.4 Finally there is anecdotal evidence from industry people---such as that included in Douglas Myers' "Tiger on your tail" HR Nicholls Society address in April 1992 or Alan Jones' "What were we afraid of?" July 1992 paper referred to earlier.
The short answer to the question of what the data show is that they do not yet show a great deal in the way of evidence that flexibility is creating benefits---at least not at the aggregate level. Time does not permit a complete survey of the data here. To check the latest facts we are fortunate, however, to have had the release yesterday morning (Friday 13 November) of the September Quarters Household Labour Force Survey results---attesting (as in the case of this Conference being held 18 months to the day from the passing of the ECA) to the impeccable timing of the organisers of this conference. The official statistical media release for Friday 13 November 1992 is attached (Attachment 1). The figures presented can be viewed against some hypotheses we would have about the impact of the opportunity for more flexible wages since May 1991. First look at the graphs on p.2, of Attachment 1, especially the last 2 centimetres of each graph, because this is the period since the ECA was introduced. The seasonally adjusted curves are the broken lines. Note that:
Table A5 thus provides little evidence one way or the other about whether labour deregulation has tended to support employment growth.
I don't know what the papers are saying in New Zealand this morning about these figures. Mostly however the figures are nowhere near as gloomy as some people were predicting them to be. (Some forecasters had been predicting 13½% unemployment by mid-1993.) Caution is required because of the small sample size of the HLFS. The June Quarter's HLFS results were hardly believed when they were released three months ago. These results bring this survey back into line with the other statistics. Nonetheless they are certainly better than the Australian figures, also released yesterday. One point some commentators have been making about NZ employment data over the last year or so is that they show a quicker response to output growth than previously. The well-known case, apparently, is the quick follow up in the December 1991 Quarter employment figures to the 1% September 1991 Quarter output growth. Whereas NZ Treasury people are quick to point out that New Zealand's quarterly GDP statistics are not very reliable, everybody seems to think that NZ is still growing at nearly 1 per cent per quarter. The bottom line is that yesterday's figures are "not inconsistent" with the view that greater labour market flexibility is producing results. However you would probably want to see a couple more periods of employment growth before skiting about it. What other statistics can we add to this information to "fatten up" as it were, the picture of New Zealand's labour market we are painting? We have some further tables here which flesh it out a bit---but there is only time now to look at one chart. Chart 1 depicts the flows underlying the employment and unemployment figures in the December Quarter 1991. The relative sizes of the rectangles are roughly the same as the numbers of people involved. The fluidity of the labour market is striking. Some people would find the quarterly turnover rates of unemployment and non participation surprisingly large. The high turnovers soften the tragedy of the unemployment figures somewhat. I do not yet have figures for later quarters, but there is every reason to expect that the broad magnitudes are the same. How does one assess such data? What does a healthy set of flows between employed/unemployed/non participation look like? It would be interesting to see a long series. Incidentally, not shown in the chart is an estimated 5,000 person quarterly rate of increase in the numbers of working age---2,500 of whom represent immigration. Many of the latter would be New Zealanders returning from Australia (though of course the traffic both ways between Australia and New Zealand continues at a high level).
5. Minimum Wages The absence so far of an unambiguously clear link between the major deregulation of 1991 and the employment figures up to the September Quarter of 1992 provides an appropriate backdrop for consideration of progress in our project on minimum wages in New Zealand. The brief for the project is to investigate as quantitatively as possible, the impact of minimum wage legislation in NZ. Could it be that continuation of the statutory minimum wage has thwarted a response to the abolition of the other wage-fixing instruments?
There is a feeling in informed circles in NZ that statutory minimum wages could defeat the purposes of the ECA, even though the collectively bargained minima known as awards have been abandoned (or are being abandoned). A different argument has been put forward by people who lament the passing of the award system. In particular, authors with Victoria University connections, such as Peter Brosnan & David Rea6 and Phillipa Bascand & Stephen Frawley7, have been among the first since the ECA to write in New Zealand about minimum rates, saying that with the abolition of the central award structure, there is a need to widen and raise minimum wages and indeed to reinforce all of the traditional "minimum conditions" which remain in NZ law. The minimum conditions include: (i) holiday time; (ii) dispute settlement procedures; (iii) protection against employers extracting deductions from wages; and (iv) workplace safety. The ECA itself preserves some of them (esp. (ii)). The NZ Federation of Employers has been suspicious of the Employment Court (which can be called in to review decisions or agreements) because it seems inclined to take upon itself a defence of minimum conditions. The defence of minimum conditions has threatened to become the next growth area for the legal profession. Alan Jones, among others, has argued in the newspapers and elsewhere for the Employment Court to be abolished---ordinary courts are quite adequate, he says. In any case, in New Zealand, the minimum conditions idea seems to have quite a bit of currency. The minimum wage is one of these and the Department of Labour has this listed in its minimum conditions pamphlets to advise workers. 5.2 Some brief history As in Australia, early Arbitration Acts in NZ and their associated award structures go back to the 1890s. The awards were what you might call "centrally bargained minima". The first explicit statutory minimum wage in NZ, that is, a minimum wage set directly by Parliament, seems to have been the Employment of Boys and Girls Without Payment Prevention Act 1899. (It set minima at:
In the recovery period after the Great Depression, a statutory wage floor was introduced to supplement all awards. Using terminology already familiar in Australia at the time, an Industrial Conciliation and Arbitration Amendment Act 1936 established the notion of a "basic wage" when it instructed the New Zealand Arbitration Court:
For male workers the Court was directed to:
There was a great deal of debate about the appropriate foundation for a basic wage. Note that the basic wage applied only to people covered by awards---and at that stage there were still quite a few employees, such as those in agriculture and forestry, who were not part of the award structure. In 1945 an explicit Minimum Wage Act was passed. People inside and outside the award structure were covered. It excluded people under 20 years of age, apprentices and (through another Act) the disabled. In 1952 the Act was amended to allow the minima to be changed by Order-in-Council (instead of Parliament). The minimum wage was 78 per cent of the average wage in 1948, the highest it has ever been. It was allowed to fall to 40 per cent of the average wage by the start of the 1980s. Then it was raised from 30 to 53 per cent in 1987 by the new (Lange) Labour Government. By 1988 it was higher than the minima specified in 20 per cent of prevailing awards. Since September 1990 it has been kept at $6.125 per hour and in the year ended March 1991, for example, relative to the average wage the minimum had fallen back to 1987 levels. In 1992 it has been lower still, in relative terms. As regards the relationship of the minimum wage to welfare benefits, records since 1960 show that:
So notwithstanding recent declines in the real minimum wage, you would still expect it to "bite". 5.3 Postulated effects What do we mean by the term "bite" in this context? Minimum wages have been theorised to cause unemployment, especially among the weak or lowest-paid sectors of society. The precise nature of the impact is not usually discussed in much detail in textbooks. The quantitative literature is extensive, especially in the US around the late 1970s and early 1980s when there was a major government inquiry into the subject.8 Debate in that literature focuses on the right functional form to employ in estimating the effects of minimum wages---there are many complex methodological issues to consider. The simplest statement of the postulated effect is that the frequency distribution of hourly pay rates becomes truncated at about the point of the minimum. The idea is that some people whose marginal product is just below the minimum are pushed either out of employment (to the left) or up the pay scale while retaining employment (to the right) causing a peak at about the minimum rate. There are quibbles about how much the truncation of the left hand side would occur anyway as the product of social security provisions. But New Zealand data showing a trun-cation and peak can be readily presented as follows. Charts 2 and 3 below show the similarity of the shapes of the frequency distribution of wage rates every year over the last decade.
5.4 Eyeballing the averages
Now initially in our study of minimum wages in New Zealand we started looking at the effects by considering "covered" versus "uncovered" categories of people. The feeling was that we might pick up from averages significant employment effects of the minimum wage if we looked at the right categories. The type of data I am talking about are presented in the following diagrams (see Chart 4). They show quarterly averages for several labour statistics from June 1987 to June 1992 for 15-19 year olds on the one hand and 20 to 24 year olds on the other hand. Remember the younger group is not covered by the minimum wage. My one-sentence summary of the diagrams in Chart 4 is that you cannot pick from them the supposed effect of minimum wages on employment. Indeed, paradoxically, the picture with the uncovered group is more indicative of the standard theorised minimum wage effect (ie higher unemployment rate, lower participation rate) than that with the covered group!
The data in Tables 3 and 4 below indicate that average data for categories of people one might think would especially be minimum wage victims (eg. Maoris) are no more revealing. From the data presented in the tables it can be seen that Maoris are more often unemployed, or outside the workforce altogether, than other New Zealanders. But the phenomenon is not obviously any more intense with 20-24 year olds than with 15-19 year olds as one might have expected in the presence of a minimum wage, even if the generally more lowly-paid females are considered.
The above are data for one year only. We now have received data for earlier years and will look at the series. But we are not counting on "better" results. Just as nothing impressive can be seen by eyeballing these data on averages, regressions on such data within groups typically show nothing much either. Nobody seems to have had striking "success" with cross sectional or time series approaches. In the NZ case, one possibility is that junior rates have been prevalent in New Zealand awards and have been carried over into most of the contracts of employment which now prevail. This would mean that 15-19 year olds have been and remain, in practice, part of the covered group, not the uncovered group as we had surmised. We are looking into this. 5.5 Clifford Thies's approach Notwithstanding the possible confusion over which group is covered and which is not, it seems to us that far more subtle approaches are required and one we are in the midst of trying is an approach pioneered by Meyer and Wise.9 It was refined to a practical dimension in a 1991 article by Clifford Thies.10 In Thies's article, Massachusetts brush industry data for the years around 1912 are presented simply and compellingly to show the impact of a statutory minimum wage. The data are frequency distributions of women's wages. Before minimum wage legislation, the wages of women in the brush industry are distributed almost normally. After the minimum wage legislation, the left hand tail of the distribution is cut off. Supporting evidence from other States in the US is also presented. Hoping to repeat Thies's success, we have been comparing the wage distributions of potentially affected NZ groups. For example we are trying it with subcategories of people in the 15-19 and 20-24 year age groups such as the Maori and female sub categories that we introduced in Tables 3 and 4 above. We are hopeful, but not confident. 5.6 Information on other effects We are also gathering supplementary data on effects such as the hypothesised changes in work conditions. It is theorised by more modern writers (and Richard McKenzie11 in particular) that most of the effects of minimum wages are not on unemployment levels. An implication is that average employment levels of covered and uncovered groups will not show very much. Another is that we should not expect to see any pronounced change in the shape of wage rate distribution curves, an hypothesised effect which is also mostly about impacts on employment numbers. Very persuasive commentary (and some limited quantitative evidence) from McKenzie and other authors since the early 1980s indicates that work conditions, and especially on-the-job training, are the principal casualties of minimum wage laws. Let me give a couple of examples:
Our despair with the conventional regression analysis type approach (and our anxiety about even having success with the more subtle Thies approach) are fuelled by the "poor" results obtained by other researchers, notably in France---where as recently as 1991, a very sophisticated study of the minimum wage found nothing significant in terms of employment volume effects of the minimum wage.15 On the basis of such miserable results, the veteran minimum wage researcher Charles Brown was inclined to conclude in a 1988 review article that the minimum wage is much overrated by its critics, as well as by its supporters.16 However, Brown admitted that he did not look beyond the impact on employment numbers. And it is precisely because he did not look beyond employment numbers that his dismissal of minimum wage effects is not an opinion shared by Richard McKenzie. In 1987 McKenzie complained bitterly about the failure of observers to consider working conditions and fringe benefits, warning that:
5.7 Concluding comments Ours is still an unfinished task. There is no guarantee that we will obtain a "positive" finding sufficient to serve as a basis for estimating the social cost of the NZ statutory minimum wage. We need to be prepared to admit, if the data show it, that the legislated minimum wage may be so low that it is redundant. There are some obvious and some not so obvious reasons why this may be so. One of the less obvious types of possible reasons is that in NZ, regardless of which groups the law applies to, ordinary employers may have become used to thinking that wages below a certain level are, somehow, immoral or indecent. Note that in NZ, as in other Western Countries, all the evidence indicates that low wages and low household income are very poorly correlated. The "employers-think-low-wages-are-indecent" theory may have some basis. One New Zealand economist told me of a discussion he had recently with a Nissan executive. The executive's company was suffering, with yards full of unsold vehicles, but he was still granting his workers a pay-rise because he felt they could not keep body and soul together without one. This is not a frame of mind which one sees in the United States, for example. Such findings concerning one of the last remaining NZ wage control instruments would not dissuade me from the general view that wage minima do not improve the lot of poorer people and are, ultimately, the costliest aspects of labour policy. I am hoping that even if our report must present no more than the logic of the damage minimum wages can cause, it will be persuasive on that score. I have to finish without mentioning the interesting literature on the political economy of minimum wages. There may be some time for that later on, during the general discussion.
Endnotes:
STATISTICS
HOT OFF THE PRESS
CATALOGUE No. 05.500 SET No. 92.93/087
HOUSEHOLD LABOUR FORCE SURVEY SEPTEMBER 1992 QUARTER
This Information Release presents key seasonlly adjusted and survey statistics for the September 1992 quarter from the Department of Statistics' Household Labour Force Survey. Each survey estimate is the average for the three months and is subject to possible sampling error.
Figures have been rounded, and discrepancies may occur between sums of component items and totals as a result. All percentages have been calculated using unrounded figures.
|
|
