Erosion of purchasing power and labor supply

Erosion of purchasing power and labor supply

The Journal of Socio-Economics 33 (2004) 725–744 Erosion of purchasing power and labor supply Francesco Scacciati∗ Dipartimento di Economia, Universi...

199KB Sizes 1 Downloads 64 Views

The Journal of Socio-Economics 33 (2004) 725–744

Erosion of purchasing power and labor supply Francesco Scacciati∗ Dipartimento di Economia, Universit`a di Torino, via Po 53, Torino 10124, Italy

Abstract This paper presents three studies on the labor market in Italy, carried out using questionnaires over the course of time (1990–2002). The main result is that, if real wages decline to a certain extent, employed workers increase their labor supply, in order to recoup their lost purchasing power and to redress, perhaps unconsciously, the decline in their economic and social hierarchical position. As long as real wages grew at a satisfactory rate, in the early 90s, workers were unwilling to change their work schedules in either direction—exchanging income for leisure time at the current hourly wage or vice versa—whatever were their work time and wages. During the last decade workers—who suffered a decrease in real wages—have been willing to work more hours. As a result, workers did not demand more leisure time despite the fact that it had become cheaper as predicted by traditional theory. Thus, employers may expect a not-insignificant increase in labor supply if real wages decrease. © 2004 Elsevier Inc. All rights reserved. JEL classification: J22 Keywords: Real wage; Labor supply; Target income; Loss aversion; Status quo bias

1. Introduction The first two studies, of the three this paper is based on, were originally carried out independently, in different times and on partially different categories of workers. The first one was run in order to discover if there is any evidence of the status quo bias (see Samuelson and Zeckhauser, 1988) on the valuations that workers place on variations—in either direction—of their work time (see Scacciati, 1990), and Ortona and Scacciati, 1992). ∗

Tel.: +39 011 6702746; fax: +39 011 6702762. E-mail address: [email protected].

1053-5357/$ – see front matter © 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.socec.2004.09.046

726

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

Responding to the questionnaire in the early nineties, workers—after a decade of a constant nominal wage growth equal to inflation plus productivity growth—showed themselves to be reluctant to exchange both leisure for income and income for leisure, given the current price of labor (i.e. the hourly wage). They therefore appeared to consider their own current working time “optimal” in a neo-classical sense. But conventional theory cannot explain the remarkable asymmetries in valuations of marginal changes, shown by the great majority of workers, across occupational categories, despite large differences in work schedules and wages. In other words, why workers are willing to accept only a very high wage increase as compensation for a marginal reduction of leisure time, whereas they are willing to pay so little in terms of wage reduction for an equivalent marginal increase of leisure time. Actually, they value changes very differently, depending on whether they are moving in one or the other direction along the same segment of the “incomeleisure time” balance line. The status quo bias theory is a convincing explanation for this anomaly. The aim of the second study1 —run in 1999—was to record the attitude of businessmen and workers toward some policies aimed at increasing employment (including a shortening of working time), and the predictable effects that it would have on employment. During the ten years that divide the two studies, things have not gone at all well for workers (neither in Italy nor in many other countries). The real wage of many occupational categories has significantly fallen—as nominal wages have grown less than prices—and the share of wages in national income has decreased, sometimes dramatically. In such a situation, it turned out that workers were willing to give up leisure time to be able to recoup fully their lost purchasing power. A prevalence of the income effect, strengthened by the loss aversion (see Tversky and Kahneman, 1991)—especially as regards the target income (as discussed by Altman, 2001)—is a good explanation for their attitudes and behavior. The changes in workers’ revealed preferences were remarkable. But, in spite of numerous elements common to the two studies, the long time that had elapsed between the two of them, the different occupational categories of the respondents and the different dimensions of the samples suggested caution in drawing comparisons and conclusions. So a third study was undertaken (2002) to test the reliability of the comparison of the first two, with the aim of analyzing the determinants of hired workers’ decisions to supply more, the same, or less labor in different situations and economical environments. In particular, to look at how those decisions were affected by recent changes which had occurred in the purchasing power of workers’ wage and in their hierarchical position and status in the society. This paper, in part 3, shows the different attitudes towards supplying over-time, both of blue and white collar workers, employed in two firms. The sample was homogeneous, fairly numerous (323), and almost everybody (88%) answered the questionnaire. The results are statistically very robust, and confirm, beyond any reasonable doubt, the impressions gathered from the results of the first two investigations. 1

Conducted by a group of researchers at the University of Turin: see Fubini (2000).

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

727

2. The preferences of three occupational categories of workers in 1990 In 1990, I ran an experiment designed to measure the asymmetries—if any—between the valuations that workers placed on a small increase of their monthly leisure time and on an equivalent small decrease. The experiment entailed asking two groups of local government employees, a group of high school teachers, and a group of bank employees (altogether 98 subjects): (a) what minimal wage increase they would have asked their employers as compensation for adding one hour (or two) per month to their work schedule; (b) what minimal wage reduction they would have accepted from their monthly wage to decrease their work schedule by one hour (or two) per month. In other words, they were asked their willingness to accept (WTA)2 in terms of income compensation for one hour less of leisure time and their willingness to pay (WTP) for one hour more of leisure time. The sample of the three occupational categories (separately and all together) showed a very strong asymmetry between the two valuations. First of all let’s describe the questions that were put to each of the three samples of workers and the answers they gave. 2.1. Local government employees Forty local government employees were divided into two groups. The members of one of these groups were asked: “Imagine that the Local Government agrees with the Unions to ask each employee if she/he is willing to increase his average monthly labor time from 160 to 162 h, with a corresponding increase in pay. How much is the lowest monthly salary increase you would ask for?” The members of the second group were asked:3 “Imagine that the Local Government agrees with the Unions to ask each employee if she/he is willing to decrease his average monthly labor time from 162 to 160 h, with a corresponding reduction in pay. How much is the greatest monthly salary reduction you would accept?” This experiment yielded the following results: (a) The average amount demanded by respondents to the first question (i.e. the valuation of two hours less of leisure time) was 11.18 times the average amount offered by

2

Among many others discussed by Cursey et al. (1987). There is a slight distortion in the data due to the researchers’ purposes. In reality, the average monthly work time was more or less 161 h. But it is not easy to calculate it on account of summer vacations, holidays, etc., so nobody objected to the data as presented. 3

728

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

respondents to the second question (i.e. the valuation of two hours more of leisure time). (b) The average selling price of leisure time (WTA) was 2.86 times the current average market price (i.e. the net hourly wage). (c) The average buying price of leisure time (WTP) was a 25.58% of the current average market price. 2.2. High school teachers Thirty-five high school teachers were asked what would be the minimal monthly salary increase they would ask for, to be willing to accept a one-hour increase in their monthly labor time (one extra-hour of class time). The following day, the same 35 teachers were asked what would be the maximum reduction of their monthly salary they would be willing to accept, if their labor time were reduced by one hour of class time. Once again the asymmetry showed up. (a) The average valuation of one hour less of leisure time was 6.25 times that of one hour more of leisure time. This shows, again, a marked kink in the indifference curve at the tangential point to the budget line. (b) The average valuation of the selling price of leisure time (WTA) was 4 times the current average market price (i.e. the net hourly wage). (c) The average valuation of the buying price of leisure time (WTP) was much smaller (2/3) than the current average hourly wage. 2.3. Bank employees The same questions were put to 23 bank employees,4 whose status quo in relation to leisure and income was different (the former was smaller, the latter was greater) from that of the teachers. Twenty-one bank employees (91.3%) showed an asymmetry in valuations similar to that of the school teachers. (a) The average valuation of one hour less of leisure time was slightly over 4 times that of one hour more of leisure time. This shows, as in the former cases, a marked kink in the indifference curve at the tangential point to the budget line. (b) The average selling price of leisure time (WTA) was 3.4 times the current hourly wage. (c) The average buying price of leisure time (WTP) was significantly smaller (81.8%) than the current hourly wage.

4 The fact that both questions were asked to the same persons (in two groups out of three) leads probably to less asymmetric results: the first answer might have influenced the second one, reducing the disparities. This seems to be supported by the fact that the greatest asymmetries in valuations were found among the local government employees. Nevertheless, very significant asymmetries showed up also with the teachers and bank employees: this strengthens the results of the research project.

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

729

2.4. Some notes on the results of this experiment The rather small dimensions of each of the three groups of respondents makes comparisons between the valuations of the groups not very reliable (but this was not the aim of the inquiry). But the overall sample (98 respondents) is large enough to tell us something: as a point of fact, all statistical tests showed significance and reliability. This experiment shows a clear and striking asymmetry between the WTA and WTP of workers as regards marginal changes in leisure time in relation to their status quo. These valuations are not predicted by the perfect rationality approach, but can likely be ascribed to a status quo bias, insofar as asymmetries of such a scale are absolutely not ascribable to the “normal” slope of an indifference curve. Of course, everybody is aware that at the equilibrium point—if variations are marginal (as in this study)—the slope of the indifference curve is greater on the left and smaller on the right compared to the slope of the income/leisure-time budget line. But—for the very reason that variations are marginal—the magnitude of the asymmetry described above should be small (obviously the slope is the same at the tangential point). The most striking result comes from the two groups of Local Government employees. If we assume that they form a representative sample, we can see that the typical subject values gains and losses (a lot) differently, albeit moving up and down the same (very small) stretch of the income-leisure time balance line! Moreover, comparing the answers of the three groups, we see that—given current market prices (i.e. the current hourly wages)—a very large majority of workers was not willing to move from the present income/leisuretime situation, even if their employer were to give them such an opportunity. The status quo is considered “normal” and “satisfactory”. It follows that the majority of workers—whatever their initial opinion concerning the optimal income/leisure-time combination (income being a function of the number of hours worked)—after a certain period of activity with a steady and satisfactory pay and a steady work schedule, will consider this the normal and optimal combination. Obviously, they would be willing to work less and/or to earn more, but they declared not to be willing to accept a steady exchange between wage and leisure time at the market price (i.e. at net hourly wage) either in one direction or in the other. This study was run at a time when nominal wages were growing at a rate that was roughly the sum of inflation and the increase of productivity, i.e. real wages were growing and the GNP distribution between labor and capital was more or less constant. We were in what would have been considered “normal times”. So we may conclude that in normal times a normal worker is not willing to give up either leisure time or income at his current wage. If they are generally satisfied with their job, they feel they are in an optimal condition—regardless of the objective differences in the work schedules and in the hourly wages of the different occupational categories. A clear status quo bias appears, which produces in workers’ valuations the asymmetries described above, and which is not predicted by the conventional economic theory. Workers and unions have certainly undertaken many struggles for the reduction of work time, and there is a clear historical trend that shows a decrease in the length of daily, weekly and annual work schedules. But successes in cutting work time have been obtained, in most cases, in periods of robust economic growth and low levels of unemployment—except for

730

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

the early thirties in the US during the Great Depression (it must be said that, nevertheless, those reductions of work time didn’t last long). Thus decreases in work time and increases in real wages have usually been achieved at one and the same time.

3. Workers’ preferences in 1999 In 1999, a group of researchers of the University of Torino and of IRES-CGIL—with substantial cooperation from the Unions—carried out a survey in which two questionnaires were distributed. One to 400 businessmen owning small and medium size firms all over Italy and one to 15,623 workers (blue and white collar) of 90 firms: 6,066 of these were returned. As said above, the scope of this research was to record the attitudes of businessmen and workers towards some policies aimed at increasing employment (including a shortening of work time, something originally included in the center-left government program), and the predictable effects that they might have on employment. A 39% rate of return may possibly imply a selection bias, such as being a member of the Unions: on the other hand, it is conceivable that the removal of such a bias would possibly reinforce the particular results of the study. The results of this part of the research showed that the majority of wage workers—including both women (see Del Boca, 1998) and men, but especially the latter—was not at all willing to give up income for more leisure time. On the contrary, a solid 57% of all workers (and 67% of males) were willing to work over-time. Moreover, they were also opposed to receiving more leisure time for an unchanged wage, if this would mean, as likely, that there would be no wage increase coming for the future or for already achieved increases in productivity. Only 21.8% of the workers that filled out the questionnaire mentioned a reduction of the work schedule as a prominent target for the Unions to pursue, whereas 50.8% put the increase of wages in first place. Among the latter, more or less half would have agreed to split the increases of productivity between a raise in wages and a decrease of work time and the other half decidedly preferred a more substantial wage raise. The majority of the male workers in the metal and mechanical sector said they were willing to work on average 5 overtime hours a week more than the 40 h stipulated by their work contract, so as to increase their monthly income; and they actually managed to do so, because employers were willing to add working hours at current wages. What had happened in the nineties to produce such a change in workers’ attitudes?

4. Worker opinions in 2002 in a booming and in a slumping sector The willingness to accept more money (at the current hourly wage rate) for a reduction of leisure time—something that clashes with the results of the first study presented in this paper—appears clearly in the majority of workers’ (both blue and white collar) responses in the more recent inquiry. This is likely due to the reduction of real wages that has taken place in the last twelve years in Italy, in spite of a not negligible rate of growth in the GNP. Since it could be objected that the distance in time separating the two studies, the different occupational categories of the respondents and the different dimension of the samples do

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

731

not allow for a comparison of attitudes, behaviors and preferences—I ran a third survey in the attempt to remove these elements of weakness in my hypothesis. The sample was composed of all the workers (blue and white collar) employed in two firms located in the Piedmont Region: one firm (X) operating in a booming sector and one (Y) in a slumping sector. The decreasing real wages should have been easier to find in firm Y and vice versa and, actually, so it was. The survey was run by means of a telephone questionnaire. The questions were the following:

(1) Gender (2) Born before September 1st, 1967 (3) Do children below 10 live in your household? (4) Which firm are you working for? (5) Have you ever worked in another firm or company before? (6) Have you ever in the past earned a higher hourly wage, in the sense that it gave you a higher purchasing power than your current one? In other words, with the same work schedule, have you ever enjoyed a higher standard of life? (7) Are you satisfied with your current wage? (8) Do you normally work overtime? (9) Are you more or less willing to work over-time than in past years (“months” if 2002 is your first year of work)? (10) If you answered “yes” to question 8, how many extrahours did you work during last September? (11) And how many during last October?

F Yes

M No

Yes

No

X

Y

Yes

No

Yes

No

Very much Yes

Somewhat

More

Less

Not much

No





About the same

Definitely not

732

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

Altogether 323 telephone questionnaires were filled out among the 367 salaried and wage workers (88%) in the two firms. That gave us 3553 items of information: 170 out of 187 (90.1%) from the firm operating in the booming sector and 153 out of 180 (85%) from the firm operating in the slumping sector. Those above 35 years old turned out to be 199 (61.6%) and those under 35 were 124 (38.4%). Females were 128 (39.6%) and males 195 (60.4%). No statistically significant difference emerged in terms of willingness to answer, neither between the “old” and the “young” nor between males and females, so only the data that are directly related to the objectives of this research project will be dealt with.5 (i) Question 8 is significantly correlated only with question 4. Most of the respondents do work over-time (260, i.e. 84%): 161 (i.e. 94.7%) employed in firm X and 99 (i.e. 64.7%) in firm Y. This seems to have little to do with the level or the dynamics of their contractual wage (i.e. the amount they get for working 40 h per week, or, as the case may be, for their contractual work schedule). Simply, firm X asks for more over-time and obtains it. The same correlation with question 4 holds for question 10. (ii) Questions 6 and 7 are logically linked, and the robust correlation that appeared was expected: Only one out of the 176 that answered “yes” to question 6 (i.e. earned in the past a higher real wage) answered question 7 “somewhat”; 175 responded “not much” or “definitely not”. Those that responded “no” to question 6 (147) were split in their answers to question 7: 55.8% are “very much” or “somewhat” satisfied with their wage, 44.2% “not much” or “definitely not”. It is interesting to note that there is a significant difference between the “old” and the “young” that answered positively to question 6. Of the 176 (54.5% of the whole sample) that answered they had earned a higher real income in the past, 145 (82.4%) were “old” workers. This shows that the “really good times” were quite long ago (actually the late eighties and the very early nineties). The same conclusion is reached if data are analyzed the other way round: 72.9% of the “old” workers are unhappy with their current wage, whereas only 25% of the young ones are. Finally, as regards real wages earned previously, the difference between the workers employed in the two firms was statistically significant, as expected, but not very remarkable: those who had earned higher real wages before represented 50% among those employed in the “booming” firm (50%), and 55.5% in the “slumping” firm. The difference is probably due to the slightly higher percentage of young workers in firm X than in firm Y, rather than to a different wage strategy. (iii) Let’s now analyze the two robust relations between the answers to question 7 and question 9 and between question 6 and question 9. Question 9 was: “Are you more or less willing to work over-time than in past years?” 210 answered “more”, 84 “about the same” and only 19 answered “less”. Of the 210 that answered “more”, 188 responded “not much” or “definitely not” to question 7, whereas only 22 responded they were “very” or “somewhat” satisfied with their current wage. But the most striking 5

Among “other information”, there is a correlation between being a worker (especially female) living in a household with small children and not being more willing than in the past to work overtime.

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

733

result that emerges from the analysis of the survey data is a linked correlation between the answers to questions 6, 7 and 9. Of the 176 workers that answered question 6 “yes”, 99.4% answered question 7 “not much” or “definitely not” and 162 of these (93%) answered question 9 “more”. Therefore, “the representative worker” has earned more in the past, is unsatisfied with his wage and is willing to work overtime. (iv) The opposite attitude is shown—albeit the correlation is less robust—among those who never before earned a higher real wage. Only 32% are more willing than before to work over-time, and this willingness is expressed by only 10 (12.2%) out of the 82 in this set that declared themselves to be “very much” or “somewhat” satisfied with their wage. All these relationships are displayed in Fig. 1. Deeper analysis of the data is presented in the statistical appendix.

Fig. 1. Some elaboration of the answers to questions 6, 7 and 9.

734

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

5. Discussion The third test confirmed that the reasons that I first conjectured for the differences in wage and salaried workers’ behaviors and preferences as regards increasing their labor supply, were correct. Such differences do not depend on the kind of job held by the respondents. Rather, they depend on whether workers feel satisfied with their real wage, or, contrariwise, feel unhappy and underpaid. And they are most likely to fall in the second category when they have suffered a decrease in their real wage. When workers are (or feel) poorer than before, they become intensely unhappy as a result of this change for at least two reasons. (a) The undeniable decline in the quality of wage and salaried workers’ life: This decline in the quality of life might be rather small—if measured in terms of fall in real income—but, nevertheless, valued very negatively because of the loss aversion phenomenon (well known in economic psychology): i.e. their perception of loss in terms of part of their purchasing power. As for this argument, there is something important to point out. If the “goods” at stake were only income and leisure time, there wouldn’t be any reason why the loss aversion for income should outdo the loss aversion for leisure time and therefore justify an increase in labor supply. The fact is that, together with the former, comes the aversion for a loss in the hierarchical position in society and for the economic targets. Let’s now represent a hypothetical worker in relation to two sets of Cartesian axes: real income and leisure time, and how she/he would value them. In Fig. 2 a very stylized version of Kahneman and Tversky’s prospect theory (1979) is presented in relation to the valuation of variations in real income, taking leisure time as a given. From time 1 to time 2 a unitary decrease of real income occurs, measured by the segment It1It2. The valuation of this decrease is measured by the segment Vt1Vt2. As can be easily seen, if there had been an increase of income the same as It1It2 (measured by the segment It1It3) the corresponding increase in its valuation (measured in ordinate by Vt1Vt3) would have been much smaller than Vt1Vt2.

Fig. 2. Prospect theory: valuation of gains and losses from reference state of income.

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

735

Fig. 3. Valuation of gains and losses from reference state of leisure time.

Fig. 3 displays the valuations of variations in leisure time, with real income as a given. If we now compare Figs. 2 and 3, and assume (for purposes of simplification) that there is a ratio of 1 to 1 between income and work time—i.e. one more unit of work time (one hour) is paid for with one more unit of income and vice versa—then we see that: (i) the decrease in real income It1It2 (in Fig. 2) can be recouped by workers if they accept working one hour over-time, moving therefore from Lt1 to Lt2 (in Fig. 3). (ii) the situation described by Fig. 3 will be felt as less negative (and therefore preferred) than the one described in Fig. 2: the reduction in the valuation of the new real income/leisure-time mix is smaller in Fig. 3 than in Fig. 2 (being the segment V t1V t2 in Fig. 3 shorter than the segment Vt1Vt2 in Fig. 2). Therefore, in time 2 workers will choose the situation described in Fig. 3, i.e. less leisure time and an unchanged real income. In the final analysis, does this hypothetical worker reflect reality? The answer is “yes”. This is confirmed by the results of all three studies presented in this paper and fits with what is predicted by both the “prospect theory” and the “target income theory”. Workers definitely wish to climb their income/leisure-time budget line up to a preferred indifference curve by supplying more labor. In this case, the asymmetry between their WTA for a one hour decrease in leisure time (at the current hourly wage) and their WTP for a one hour increase in leisure time (zero, or only a little more) is perfectly explainable by conventional theory. But what does not match conventional theory is that the labor supply curves turn out to be oriented: if the independent variable (real wage) increases, the curve is sloped positively or is vertical, but if the independent variable decreases, the curve is sloped negatively. Therefore, they don’t look at all like the ones predicted by the traditional model. Labor supply curves (such as the Ls) showed a negative slope, regardless of the level of the real wage, once real wages began to decline. (This is what has been going on in Italy and in other industrialized countries during the past decade or so.)

736

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

Fig. 4. A series of oriented labor supply curves.

This phenomenon, presented in the third test, was confirmed both for workers in the booming sector and in the slumping sector in 2002, thus reconfirming what emerged from the 1999 test. Fig. 4 displays the two cases. However the labor supply curve may be inclined for growing wages (i.e. be it “traditionally” positively inclined, as the Ls in Fig. 4, or vertical, as it turned out to be in the inquiry of the early 90s)—if real wages decline, workers will try to recover their purchasing power by supplying more labor. (b) Their standing in the socio-economic scale: As stated before, the asymmetry described above in valuing losses is plausible because of the additional loss in workers’ economic and social hierarchic positions and because of their prevailing will to achieve their target income. So, if workers’ wage (or, more realistically, workers’ purchasing power) declines, they will be willing to work overtime; whereas they would not be willing to do so if their wages were increasing, and if they were satisfied with their income and thought it fair. In this connection, Altman’s model6 predicts that workers will behave this way: that is, they will increase their labor supply when their real income falls short of their “target income”. If we assume that in Fig. 4 the hourly real wage was W* and that afterwards it goes down to W◦ , the labor supply curve changes into the Ls” and probably the OW◦ DL◦ area will be (more or less) equal to the OW*CL* area, so that the overall purchasing power will not change.

6 See Altman (2001), p. 211 and the following: “The effect of a fall in wages is expected to be negative—with a decrease in wages causing an increase in labor supply—when labor supply is initially below its maximum and when maintaining real income is a higher order of need than maintaining the initial level of non-market time. . . Reducing the opportunity cost of leisure would not have the expected substitution effect predicted by the conventional model of reducing labor since working less would reduce income below the target level”.

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

737

6. Conclusions The “conventional theories” state that the labor supply curve is positively sloped. This is not confirmed by any of the three investigations that are presented in this paper as regards wage and salaried workers. When real wages grow at a satisfactory rate (i.e. more or less at the same rate as labor productivity) workers show themselves not to be willing to change their mix of income and leisure time—at the current hourly wage—in either direction. In particular, they are not willing to give up part of their leisure time (though it has become more expensive) in exchange for better paid additional labor time. On the contrary, when decreases in real wages occur, our investigations show that the income effect totally prevails over the substitution effect, again at odds with conventional theory. This happens in most cases and not only when real wages are very low, the latter being a very well-known phenomenon (in Fig. 4, that would be the case shown by the Ls’ labor supply curve). Workers, whose real wages have declined, declare themselves to be willing to work more hours, in order to recover the monthly purchasing power and the level of consumption they were used to and—probably—their economic and social hierarchical position. Consequently, they don’t demand at all more leisure time in spite of its having become cheaper. As appears clearly in the statistical appendix, all the relations analyzed in the third research show very high degrees of statistical significance. As for the economic significance, if we accept the hypothesis—discussed in (a) in the previous paragraph—that workers want to recoup the losses of their purchasing power, it must be pointed out that in the last decade real wages in Italy have decreased by a 4.4% rate (see Eurostat, 2002). This would mean that workers would be willing to supply from 1.75 h of overtime work per week. If we widen the hypothesis, stating that the workers’ desire to recoup their economic and social hierarchical position—discussed in (b) in the previous paragraph—then there is an 11.4% to recover (4.4% + 9% of real national income growth). This would imply that workers would be willing to supply 4.5 h of overtime work per week. The fact that this surplus of labor supply is only partially utilized by employers is probably due to an insufficient level of aggregate demand. If the conclusions I draw by comparing the results of the researches presented in this paper are correct, not only—at odds with the conventional theories—employers may expect a non-insignificant increase of the labor supply if real wages decrease to a certain extent.7 If this occurs concomitantly with a slumping aggregate demand, the overall effect is to further depress real wages in a definitely downward spiral that should be cause for concern. Of course, the surveys conducted in this research do not cover all the determinants of the labor supply. If a great deal has been said here as regards the labor supply of wage and salaried workers, nothing has been argued either about the self-employed or about those who marginally move in and out of the labor force. It may be argued that these workers change their labor supply in a positive relation with the real wage changes in such a way that 7

Fortunately, we cannot tell what would happen in the case of a steep decline in real wages, simply because this did not occur—at least so far.

738

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

the overall labor supply might behave according to the conventional theory. This hypothesis requires further testing, but, in any case: (a) the studies presented in this paper have shown that the changes found in wage and salaried workers’ labor supply, due to changes in real wages, are the opposite of those predicted by the conventional theory, and (b) the economic, social and psychological determinants of the wage and salaried workers’ behavior discussed in this paper (status quo bias, loss aversion, prospect theory, hierarchical position, target income) do not appear to be lacking as determinants of the self-employed behavior. Finally, something must be said about those who may decide to move in or out of the labor force according to a rise or drop in real wages. A great deal of empirical research shows that only very substantial increases in real wages can persuade students to give up school, or induce elderly retired people to go back to work, or incline housewives to leave their children at home or in day-care centers once they had decided to take care of them personally. On the other hand, it is also not very plausible that people drop out of the labor force and give up their job simply because of small decreases in their real wage.

Appendix A. Statistical appendix A.1. Questions under analysis (6) Have you ever in the past earned a higher hourly wage, in the sense that it gave you a higher purchasing power than your current one? In other words, with the same work schedule, have you ever enjoyed a higher standard of life? (variable 6) (7) Are you satisfied with your current wage? (variable 7) (9) Are you less or more willing to work over-time than in past years? (variable 9) A.2. Frequencies • To question 6, 176 (54.5%) answered “yes” and 147 (45.5%) answered “no”. • To question 7, 101 (31.3%) answered “definitely not”, 139 (43.0%) “not much”, 52 (16.1%) “somewhat” 31 (9.6%) “very much”. • To question 9, 25 (7.7%) answered “less”, 89 (27.6%) “about the same” 209 (64.7%) “more”. A.3. Analysis of the relation between the answers to questions 6 and 7 Out of the 176 (54.5%) that answered “yes” to question 6: • 85 (48.3%), answered “definitely not” to question 7; • 90 (51.5%), responded “not much”; • 1 (0.6%), answered “somewhat”.

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

739

Table 1 Frequency table: have you ever in the past earned a higher hourly wage? Frequency

Percent

Yes No

176 147

54.5 45.5

Total

323

100.0

Table 2 Frequency table: are you satisfied with your current wage? Frequency

Percent

Definitely not Not much Somewhat Very much

101 139 52 31

31.3 43.0 16.1 9.6

Total

323

100.0

Out of the 147 (45.5%) that answered “no” to question 6: • • • •

16 (10.9%), answered “definitely not” to question 7; 49 (33.3%), answered “not much”; 51 (34.7%), answered “somewhat”; 31 (21.1%) answered “very much”.

The chi-square test of independence (H0 : independence between the variables. H1 : dependence between the variables), shows a level of significance smaller than 0.05 (Tables 1–3): therefore there is a statistical dependence between the variables 6 and 7 (Tables 4a and 4b). A.4. Analysis of the relation between the answers to questions 6 and 9 Out of the 176 (54.5%) that answered “yes” to question 6: • 14 (8%) answered “about the same” to question 9; • 162 (92%) answered “more”. Out of the 147 (45.5%) that answered “no” to question 6: • 25 (17%) answered “less” to question 9; Table 3 Frequency table: are you more or less willing to work over-time than in past years? Frequency

Percent

Less About the same More

25 89 209

7.7 27.6 64.7

Total

323

100.0

740

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

Table 4a Cross-tabulation: have you ever in the past earned a higher hourly wage? Are you satisfied with your current wage? Are you satisfied with your current wage?

Have you ever in the past earned a higher hourly wage?

Yes No Total

Definitely not

Not much

Somewhat

Very much

85 48.3% 16 10.9%

90 51.1% 49 33.3%

1 0.6% 51 34.7%

0 0% 31 21.1%

101 31.3%

139 43.0%

52 16.1%

31 9.6%

Table 4b Chi-square tests

Pearson chi-square Likelihood ratio Linear-by-linear association Number of valid cases

Value

d.f.

Asymp. sig.

136.808 166.582 120.629 323

3 3 1

0.000 0.000 0.000

• 75 (51%) answered “about the same”; • 47 (32%) answered “more”. Just as for Table 4b, also between the variables 6 and 9 there is a robust statistic dependence (Tables 5a and 5b). Table 5a Cross-tabulation: have you ever in the past earned a higher hourly wage? Are you more or less willing to work over-time than in the past years? Are you more or less willing to work over-time than in the past years? Have you ever in the past earned a higher hourly wage?

Yes No Total

Less

About the same

More

Total

0 0% 25 17.0%

14 8.0% 75 51.0%

162 92.0% 47 32.0%

176 100.0% 147 100.0%

25 7.7%

89 27.6%

209 64.7%

323 100.0%

Table 5b Chi-square tests

Pearson chi-square Likelihood ratio Linear-by-linear association Number of valid cases

Value

d.f.

Asymp. sig.

128.519 144.905 118.62 323

2 2 1

0.000 0.000 0.000

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

741

A.5. Analysis of the relation between the answers to questions 7 and 9 Out of the 101 (31.3%) that answered “definitely not” to question 7: • 4 (4%) answered “about the same” to question 9; • 97 (96%) answered “more”. Out of the 139 (43%) that answered “not much” to question 7: • 37 (26.6%) answered “about the same” to question 9; • 102 (73.4%) answered “more”. Out of the 52 (16.1%) that answered “somewhat” to question 7: • 11 (21.2%) answered “less” to question 9; • 33 (63.5%) answered “about the same”; • 8 (15.4%) answered “more”. Out of the 31 (6.5%) that answered “very much” to question 7: • 14 (45.2%) answered “less” to question 9; • 15 (48.4%) answered “about the same”; • 2 (6.5%) answered “more”. Just as for Table 4b and Table 5b, also between the variables 7 and 9 there is a strong statistic dependence (Tables 6a and 6b). A.6. Analysis of the relation between the answers to questions 6, 7 and 9 Out of the 176 (54.5%) that answered “yes” to question 6: • 85 (48.3%) answered “definitely not” to question 7 and of these: Table 6a Cross-tabulation: are you satisfied with your current wage? Are you more or less willing to work over-time than in past years? Are you more or less willing to work over-time than in past years? Are you satisfied with your current wage?

Definitely not Not much Somewhat Very much Total

Less

About the same

More

Total

0 0% 0 0% 11 21.2% 14 45.2%

4 4.0% 37 26.6% 33 63.5% 15 48.4%

97 96.0% 102 73.4% 8 15.4% 2 6.5%

101 100.0% 139 100.0% 52 100.0% 31 100.0%

25 7.7%

89 27.6%

209 64.7%

323 100.0%

742

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

Table 6b Chi-square tests

Pearson chi-square Likelihood ratio Linear-by-linear association Number of valid cases

• • • • • •

Value

d.f.

Asymp. sig.

189.165 195.461 155.676 323

6 6 1

0.000 0.000 0.000

4 (4.7%) answered “about the same” to question 9; 81 (95.3%) answered “more”; 90 (51.1%) answered “not much” to question 7 and of these: 9 (10%) answered “about the same” to question 9; 81 (90%) answered “more”; only one that answered to question 7 “somewhat”, answered to question 9 “about the same”. Out of the 147 (45.5%) that answered “no” to question 6:

• 16 (10.9%) that answered “definitely not” to question 7, answered to question 9 “more”; • 49 (33.3%) answered “not much” to question 7 and of these: Table 7a Cross-tabulation: have you ever in the past earned a higher hourly wage? Are you satisfied with your current wage? Are you more or less willing to work over-time than in past years? Have you ever in the past earned a higher hourly wage? Yes

Are you more or less willing to work over-time than in past years?

Are you satisfied with your current wage?

Total

Less

About the same

More

85 48.29% 90 51.14% 1 0.57% 0 0%

0 0% 0 0% 0 0% 0 0%

4 4.7% 9 10.0% 1 100.0% 0 0%

81 95.3% 81 90.0% 0 0% 0 0%

Total

176 100%

0 0%

14 8.0%

162 92.0%

Definitely not

16 10.89% 49 33.33% 51 34.69% 31 21.09%

0 0% 0 0% 11 21.6% 14 45.2%

0 0% 28 57.1% 32 62.7% 15 48.4%

16 100.0% 21 42.9% 8 15.7% 2 6.5%

147 100%

25 17.0%

75 51.0%

47 32.0%

Definitely not Not much Somewhat Very much

No

Are you satisfied with your current wage?

Not much Somewhat Very much Total

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

743

Table 7b Chi-square tests Have you ever in the past earned a higher hourly wage?

Value

d.f.

Asymp. sig.

Yes

Pearson’s chi-square Likelihood ratio Linear-by-linear association Number of valid cases

13.311 6.961 4.01 176

2 2 1

0.001 0.031 0.045

No

Pearson’s chi-square Likelihood ratio Linear-by-linear association Number of valid cases

71.587 81.567 55.356 147

6 6 1

0.000 0.000 0.000

• • • • • • • • •

28 (57.1%) answered “about the same” to question 9; 21 (42.9%) answered “more; 51 (34.7%) answered “somewhat” to question 7 and of these: 11 (21.6%) answered “less” to question 9; 8 (15.7%) answered “more”; 31 (21.1%) answered “very much” to question 7 and of these: 14 (45.2%) answered “less” to question 9; 15 (48.4%) answered “about the same”; 2 (6.5%) answered “more.” Tables 7a and 7b show statistic dependence between the three variables.

A.7. Correlation The Spearman’s correlation coefficient is equal to −0.667; this shows that as satisfaction with their current wage increases, workers are less willing to work over-time (Table 8). Table 8 Correlation: are you satisfied with your current wage? Are you more or less willing to work over-time than in past years? Are you satisfied with your current wage? Spearman’s rho

Are you satisfied with your current wage?

Are you more or less willing to work over-time than in past years?

Correlation coefficient

1

Are you more or less willing to work over-time than in past years? −0.667

Sig. (two-tailed) N Correlation coefficient

– 323 −0.667

0.000 323 1

Sig. (two-tailed) N

0.000 323

– 323

744

F. Scacciati / The Journal of Socio-Economics 33 (2004) 725–744

References Altman, M., 2001. A behavioral model of labor supply: casting some light into the black box of income-leisure choice. Journal of Socio-economics 33, 199–219. Cursey, D.L., Hovis, J.L., Schulze, W.D., 1987. The disparity between willingness to accept and willingness to pay measures of value. The Quarterly Journal of Economics CII, 679–690. Del Boca, D., 1998. Intra-households allocations and labor supply. In: Parsson, I., Jonung, C. (Eds.), The Economics of the Family Routledge. Eurostat, 2002. Employment in Europe, European Commission. Fubini, L., 2000. Strategie per l’occupazione, Carocci, Roma. Kahneman, D., Tversky, A., 1979. Prospect theory: an analysis of decision under risk. Econometrica XLVII, 263–291. Ortona, G., Scacciati, F., 1992. New experiments on the endowment effect. Journal of Economic Psychology 2, 277–296. Samuelson, W., Zeckhauser, R., 1988. Status quo bias in decision making. Journal of Risk and Uncertainty 1, 7–59. Scacciati, F., 1990. Evaluation disparities between variations in labor and leisure time. Torino: EXECTO 1, W.P. Tversky, A., Kahneman, D., 1991. Loss aversion and riskless choice: a reference dependent model. Quarterly Journal of Economics 4, 1039–1061.