The Impact of Labor Contracts: Evidence from Brazil
Dole, Carol A., Denslow, David A., Rush, Mark, Economic Inquiry
Underlying the Keynesian approach to macroeconomics is the assumption of inefficiencies or, as they are sometimes called, frictions. The precise sort of friction to be introduced into the model, however, is controversial. A traditional friction uses sticky nominal wages.(1) One common explanation for sticky nominal wages is labor-market contracts, that is, contracts that fix the nominal wage workers receive and then allow firms to determine the level of employment.(2) Fischer , Phelps and Taylor , and Taylor  present typical models, which assume that contracts make nominal wages rigid to some degree and that the level of employment is adjusted in accord with the firm's labor demand schedule. Nominal wage contracts also are common when developing the textbook IS-LM and AD-AS models. At the professional level, Fischer  has argued strongly in favor of the IS-LM model and Gali  suggests that this model fits the postwar U.S. experience reasonably well.
The standard view of the effects from nominal wage contracts has been challenged. Barro  pointed out that even though a labor contract may fix the nominal wage, firms and workers can equate the marginal product of labor to the marginal value of workers' time over all states of nature and thus jointly insure that employment does not deviate from its efficient, auction market level. Hall  extended Barro's insight by arguing that labor relationships are long tenn so that each side has an incentive to deal fairly with the other by not taking advantage of changes in the price level. Essentially, Barro and Hall show that a contract allowing employment to be determined solely by a firm's labor demand curve is inefficient. Indeed, the Barro/Hall conclusion that fixed nominal wages may co-exist with efficient levels of labor employment is an important justification for the implicit assumption in most real business cycle models that employment is determined in an auction-like labor market.
Some economists, such as Ahmed , Card , and Keane  have directly studied the importance of labor contracts. The results from these papers are mixed. Ahmed investigates the importance of contracts that fix the nominal wage by testing whether the responses of employment and output to changes in the money supply are sensitive to the degree of wage indexation. The basic idea is that if nominal wage contracting matters, the less flexible (that is, the less indexed) is the wage, the greater the impact of a monetary shock on employment and output. Using Canadian manufacturing industries, which employ a variety of indexing schemes, Ahmed fails to find any relationship between the degree of contracting and the extent to which a monetary shock affects real variables. Keane uses data on individuals from the National Longitudinal Survey of Young Men to study the response of real wages to unanticipated inflation and monetary growth. To the extent that contracts fix the nominal wage workers are paid, unexpectedly higher inflation should lower the real wage. Keane, however, finds no evidence in support of this proposition. Finally Card, who also used Canadian data, asserts that it is the unexpected fall in the real wage (namely unanticipated price rises), along with the degree of indexation, that should generate positive employment responses.(3) He thus constructs measures of unexpected inflation and real wage changes. Contrary to Ahmed and Keane, Card's results support a role for surprise inflation in raising employment levels via end-of-contract real wage levels. Card concludes that nominal contracts are instrumental in determining labor and output reactions.
In this paper we follow Ahmed, Card, and Keane by providing some additional evidence on the real effects from nominal wage contracting. We use Brazilian data and exploit the fact that Brazilian firms were required by law to adjust workers' salaries at specified intervals to keep pace with inflation, which was very high, and productivity. From 1965 to 1979, the adjustments were required annually; in 1979 the law changed to require semi-annual adjustments. Because of the high Brazilian inflation, real wages fell until readjustment occurred, when they were raised to maintain purchasing power. According to the standard contracting approach, as the real wages fell prior to the indexing date, firms will add workers and output will respond positively. Conversely, the new classical/real business cycle view of Barro and Hall suggests that there should be no tendency for contracts to influence output.
Relative to Ahmed, Card, and Keane, our data set has two major advantages: First, Ahmed, Card, and Keane were forced to search for relatively small effects because inflation and monetary growth in Canada and the United States were small. Inflation in Brazil, however, was far from small. If there are any effects from nominal contracts, our data may be better able to uncover them. Second, the change in the nature of the labor market contracts is both dramatic - because switching from annual to semi-annual changes obviously had a large impact on the short-run path of real wages - and exogenous - insofar as this change was mandated by the government and firms had no choice in the matter. These advantages, however, might also conceal a related disadvantage: In a highly inflationary environment, workers and firms may make efficiency enhancing adjustments that they otherwise would forego in less inflationary times. In other words, it might be the case that in low inflation countries, nominal wage contracts function as postulated by Keynesian analysis while in high inflation countries, the contracts are altered to be in line with Barro's insight. Of course, the observation that individuals placed in high inflation countries may alter their behavior when compared to residents in low inflation countries is hardly unique to this study. And, given the prominence with which high inflation countries have served as "laboratories" for testing monetary phenomena, most economists apparently think this potential drawback is generally minor.
The next section briefly discusses Brazil's background including institutional arrangements of the period while the second section describes features of the Brazilian data. For more details on these topics, see Dole . The third section presents the specification of the regressions, tests, and their results. The last section highlights the important results.
II. BRAZIL'S ECONOMIC HISTORY
During its 174 years of independence, Brazil's political climate has varied and played an important role in its economic policies, especially in the latter half of the twentieth century. By March 1964, the military's toleration of the country's economic stagnation came to an end, and it toppled President Joao Goulart's government. It was this military regime which instituted the wage indexation scheme that, in some form, remained in force through the 1980s.(4)
To combat uncontrolled wage increases, laws were enacted which dictated that wages were to be adjusted according to a specific formula starting in April 1964. The formula contained three elements: a compensating factor for past inflation, another for anticipated inflation and a third to take account of productivity improvements. By 1965, all wages in the private sector came under the jurisdiction of the law.(5)
The period from 1968 to 1974 was known as Brazil's economic "miracle" and economic growth averaged over 10% annually during the upturn. While the economy boomed, the wage indexation formula was revised in four ways: the government moved to provide standardized indices for determining salary adjustments;(6) second, to prevent underestimation of inflation, the forecasting procedure was refined; third, the base period for calculating the base wage was reduced from two years to one year; and fourth, the wage raise was discounted by multiplying (rather than adding) the productivity coefficient by the average real wage.
From 1974 through 1980, economic growth fell to somewhat over 5% annually, and then rebounded a bit in the 1980s. Against a backdrop of lower growth and rebounding inflation, several additional changes were made to the indexing scheme. A minor change to the indexing formula was in 1976 when the productivity coefficient was redefined using a terms of trade index to account for not only foreign trade but for differences in rural and urban economies.
In November 1979, the indexation policy underwent a much more pivotal change: after November, wages were adjusted every six months rather than every twelve months as had been previously the case. It is this movement toward more frequent indexing that is exploited to provide a test of the Keynesian and New Classical hypotheses.(7) From late 1983 through 1984 wage increases, while still granted every six months, were based on a new formula that related wage brackets and changes in the cost of living index. By 1985, however, under severe pressure from the unions, the government abandoned the indexing formula. More complicated schemes were used until 1987 when quarterly indexing was instituted.(8)
III. DESCRIPTION OF THE DATA
The basic data consist of monthly nationwide output, from Anuario Estatistico do Brasil, for six industries: rubber, plastics, chemicals, transportation, beverages and tobacco. These data cover some of the larger and more important industrial sectors of Brazil as well as secondary sectors.
Our sample period begins in 1975 and ends in 1984. As noted earlier, firms were legally required to adjust workers' nominal wages over our sample period, initially once a year and then, after November 1979, every six months. In 1974 technical aspects of the formula used to index the labor contracts were changed quite substantially, and so we start our sample in 1975 in order to have a consistent wage indexation formula. Then, in 1985 a new administration imposed strict price controls to lower inflationary expectations.(9) The post-1984 changes in these specific programs occurred too frequently to provide enough observations for meaningful results, so our sample concludes in 1984.
The wage adjustment month for workers varied by industry and municipality. Workers within the same industry and municipality received their wage adjustment in the same month. Nonetheless, because municipalities differ in size and because the concentration of an industry within municipalities differs, the real wages paid within the industries have sizeable spikes during certain months, when the majority of workers within the industry received their nominal wage hikes.
To (partially) illustrate this behavior of the real wages, we used monthly wage data from The Fundacao Instituto Brasileiro de Geografia e Estatistica Industria de Transformacao for Brazil's two largest states, Sao Paulo and Minas Gerais; because output in these states effectively dominates the rest of the country, this should be acceptable. In order to compute the real wage, we deflated the nominal wages using the wholesale price index from Conjuntura Economica. Unfortunately, the wage data extend only to 1976. Thus, in Figure 1 we plot the monthly real wage for five years, from 1972 to 1976, for the six industries we study. In all six figures we see a rather consistent pattern in which the real wage tends to have a dramatic increase during one month. This pattern is especially pronounced during 1975 and 1976, which are darkened in the figure. These are the first two years that use the wage indexation formula introduced in 1974 and (hence) are the first two years of our sample period.
IV. TESTS OF THE IMPORTANCE OF CONTRACTING
We conduct stability tests that focus on the switch from twelve-month to six-month indexing.(10) The shift in November 1979 from annual to six-month indexing provides a natural way to examine the importance of nominal contracting. The Lucas Critique (Lucas ) points out that when the rules of the game change, people's behavior changes. According to the Keynesian view, contracts are an important feature of the economic environment. If nominal wage contracts are an important aspect, major changes in them, such as the movement to more frequent indexing, ought to affect the evolution of output, which should be observable through changes in the time-series behavior of output.
Some intuition is illustrated in Figure 2, which displays hypothetical data. The top part of the figure shows the behavior of the real wage, initially with twelve months between indexing and finally with six months between indexing.(11) The bottom shows the effect of this "saw-tooth" behavior of the real wage on employment - and hence output - after making the conventional assumption that firms determine the level of employment along their labor demand curve.(12) Thus, as the real wage declines due to inflation, firms move along their labor demand curves and raise the level of employment. As the figure shows, the behavior of employment - and hence output - changes along with the indexing regime.
Figure 2 is designed to convey only the intuition behind these tests. Thus it is not crucial that the real wage and hence employment or output display the precise pattern illustrated in the figure. For instance, if there are lags in producing output, the peak level of output will not necessarily occur in the month with the lowest real wage. But, the key observation is that according to the conventional labor contract approach, the switch from annual to semi-annual indexing affects the time series behavior of employment.
Testing this hypothesis is straightforward: simply examine whether the evolution of output differs between indexing regimes. To this end, we started with a very simple time series representation for the evolution of output within each industry:
(1) D[Y.sub.t,j] = C + [[Gamma].sub.j] MONTHLY DUMMIES + [[Pi].sub.j] DCARNIVAL + [[Mu].sub.t]
where DY is the first difference of the log of output in industry j at time t, C is a constant, MONTHLY DUMMIES are dummies for January through November and DCARNIVAL is a dummy for the annual month of Carnival.(13) In Brazil, Carnival is a pre-Lenten festival during which there is an obvious decrease in the amount of work performed and the month of Carnival can vary between February or March. DCARNIVAL is 1 in the month of Carnival, 1 in the month after Carnival, and 0 in all other months.(14)
Equation (1) allows only real, seasonal variables to affect output. However, it is clear that other factors may also have an impact. Hence, we also estimated two other regressions that allowed measures of government fiscal and monetary policies to affect the behavior of output. These two regressions are each based on the following equation:
(2) D[Y.sub.t,j] = C + [[Gamma].sub.j] MONTHLY DUMMIES + vj DG + [[Eta].sub.j] DM + [[Pi].sub.j] DCARNIVAL + [[Mu].sub.t]
In equation (2), DG is the first difference of government spending and DM is the first difference of the money supply.(15) In one regression, we allowed all of the coefficients to change, which assumes changing indexing regimes affects the impact government spending and money supply changes have on output. This assumption may be questionable to some, so in the second regression we looked at the stability of the monthly dummies while keeping the Carnival dummy, the constant, DG and DM constant, that is, assuming that changes in the indexing regime have no effect on how fiscal and monetary policies affect real output.
TABLE I Equation (1) Stability Tests Industry F Statistic Beverages 0.342 Chemicals 0.986 Plastics 1.536 Rubber 1.855 Tobacco 1.627 Transportation 0.339 Note: The 5% critical value is F(13,87) = 1.88.
If nominal wage contracts are material in affecting output, changes in the indexing regime should affect the behavior of output. Hence, according to this view, the coefficients in specifications (I) and (2) will not be stable across the switch from annual to semiannual indexing. On the other hand, if output is unaffected by nominal contracts, the coefficients in equations (1) and (2) will be stable across indexing regimes. Thus, our first set of tests looks at the stability of the coefficients in regressions (1) and (2) over the switch from twelve-month indexing to six-month indexing. Table I reports the results of conventional F-tests for the stability of the coefficients. Table II reports the regression coefficients for the entire sample period and for the subsets of the period.
From Table II, it is interesting to note that the simple specification we use fits the data well. Bearing in mind that we are using monthly first differences of output as the dependent variable, it is apparent that the [R.sup.2]'s are relatively high, ranging from approximately 0.50 to 0.75. Moreover, many of the individual monthly dummy coefficients are significantly different from zero. These facts increase our confidence in the results reported in Table II where, strikingly, at the 5% level, none of the industries show significant changes in the coefficients as the indexing regime changes. In other words, none of the results rejects the null hypothesis that the coefficients are the same over the two indexing schemes. These results offer little support for the view that nominal contracts affect the behavior of output. Instead, the results are consistent with the view that labor contracts are unimportant.
Table III reports results for both cases of equation (2), regressions which include a measure of monetary and fiscal policy.(16) Except for one case - the rubber industry - at the 95% significance level, all of the F-statistics again are insignificant. Clearly these results confirm those from Table I, namely that changes in the indexing regime have no effect on the evolution of output.
To this point we have examined the stability of all the regression coefficients. However, it may be more powerful to focus on the stability of the monthly dummies that should show the largest effects from the change in indexing. We can exploit the fact that during the period of annual indexing, the month with the largest (upward) jump in real wages is the base month when most workers received their wage increases. For example, suppose initially the base month for an industry when most of the workers received wage hikes was January; then, starting in November 1979, when the law changed and contracts were indexed twice a year, the industry began to receive readjustments in November and May.(17) If output behaved as proposed in the conventional labor contract framework, when adjustments occurred annually, output would be highest in December, when the real wage is lowest and, once readjustments occurred twice annually, output would peak in April and October when the real wage is lowest. [TABULAR DATA FOR TABLE II OMITTED] [TABULAR DATA FOR TABLE III OMITTED] Thus, when the contract is adjusted once a year, we expect output will rise in May and November as the real wage continues to fall because of the increase in the price level. But, when contracts are adjusted twice a year, after the peak in output is reached in April and October when the real wage is the lowest, we expect output to fall in May and November when the nominal wage is adjusted. Hence, the estimated coefficients for the new adjustment months, May and November, will be significantly smaller when contracts are set twice a year rather than once a year.
To examine this hypothesis, the next test estimates:
(3) [DY.sub.t, j] = C + [[Gamma].sub.j] MONTHLY DUMMIES
+ [[Theta].sub.1][INDEX1.sub.j] + [[Theta].sub.2] [INDEX2.sub.j]
+ [[Tau].sub.j] DCARNIVAL + [[Mu].sub.t]
where INDEX1 and INDEX2 represent dummy variables for the new base month. The INDEX dummies are equal to 1 in the twelve month indexing period and 0 in the six month period. In other words, the "[Theta]'s" reflect the impact from the change in contracts in the months where we expect to find the largest differences.(18)
Again, allowing for a measure of fiscal and monetary policy to impact the change in output, we estimated an additional regression:
(4) [DY.sub.t, j] = C + [[Gamma].sub.k] MONTHLY DUMMIES
+ [[Theta]].sub.1] [INDEX1.sub.j] + [[Theta].sub.2] [INDEX2.sub.j]
+ [v.sub.j] DG + [[Eta].sub.j] DM
+ [[Tau].sub.j] DCARNIVAL+ [[Mu].sub.t]
Table IV reports the results for equation (3). We know from the previous Table II that the fit of these equations will be good, with relatively high [R.sup.2]'s. This fact is confirmed in Table IV. It also is important to note in Table IV that including the INDEX variables for the months with greatest change has little effect on the individual monthly coefficients. More specifically, just as in Table II, many of them still are highly statistically significant. However, the emphasis in Table IV is on the estimated [Theta]'s from the INDEX variables. In order to make it easier to concentrate on these variables, in Table V we present a summary of their coefficients and p-values. The results summarized in Table V are in general accord with our earlier conclusions. In only one industry, rubber, was the estimated [Theta] coefficient significantly less than zero; Indeed, this was the only significant coefficient among the twelve. Hence, once more we are forced to conclude that changes in the nominal wage indexation scheme have no impact on real output.
Finally, Table VI reports the estimated 0 coefficients for the INDEX variables from [TABULAR DATA FOR TABLE IV OMITTED] equation (4).(19) In none of the equations are these coefficients significantly different from zero. Thus, we continue to find little support for the view that nominal wage contracts are important; the results are more in line with Barro's suggestion that firms and workers do not allow the relative fixity of the nominal wage (and hence the flexibility of the actual real wage) to influence the (possibly efficient) level of employment and output.
TABLE V Equation (3) Stability Tests for "Large Impact" Months Industry INDEX1 p-value INDEX2 p-value Beverages .032 .419 -.053 .204 Chemicals .010 .661 -.024 .291 Plastics -.030 .332 -.019 .553 Rubber -.025 .048(*) -.005 .885 Tobacco -.086 .275 .064 .386 Transportation .068 .322 .010 .875 * = significant at the 5% level. TABLE VI Stability Tests for Equation (4) Stability Tests of "Large Impact" Months with DG and DM Variables Industry INDEX1 p-value INDEX2 p-value Beverages -.424 .267 .070 .071 Chemicals -.013 .613 -.027 .260 Plastics -.055 .136 -.013 .674 Rubber .078 .307 .016 .800 Tobacco -.050 .554 .082 .276 Transportation .019 .641 -.027 .512
V. SUMMARY AND CONCLUSIONS
We examined the opposing views about labor contracts of the conventional Keynesian and New Classical/Real Business Cycle approaches. In a Keynesian approach, labor contracts that fix the nominal wage and allow firms to base their employment decision on their demand for labor curve are often an important ingredient. This assumption yields the standard conclusion that higher inflation lowers the real wage, thereby resulting in higher employment and real output. The New Classical/Real Business Cycle theory points out that this conventional approach to labor contracts implies the existence of inefficiencies. These economists suggest that regardless of the behavior of the real wage, firms and workers continue to equate the marginal product of labor to the marginal value of workers' time. In this case, the rigidity of the nominal wage does not imply that higher inflation raises employment or output.
Using Brazilian data covering six industries, where nominal wages were indexed by law, we conducted several tests. These tests gave us consistent results: The impact on output from fixed nominal wage labor contracts seems unimportant: the change in the governmentally mandated indexation scheme, from adjusting money wages in an industry from every twelve months to adjusting them every six months, had no discernible impact on the evolution of industry output. Of course, it might be that other qualitative factors that we cannot include in our regressions, such as a change in the extent of labor repression or a movement to a policy more favorable to workers, changed simultaneously with the switch in the indexetion laws and that this factor(s) offset the impact of the change in indexation. Though possible, such a perfect offset seems unlikely, so we interpret our results as yielding evidence in favor of the New Classical/Real Business Cycle theory, that the behavior of output is unaffected by the presence of labor contracts that fix nominal wages.
We thank Bill Bomberger and Larry Kenny for their helpful comments. We also thank two anonymous referees whose comments greatly improved both the presentation and the substance of this paper. The Financial Institutions Center at the University of Florida helped support this research.
1. For instance, the initial model in Keynes , which occupies approximately the first three quarters of the book, assumes that money-wages are constant.
2. Recently some economists have suggested different sources of inefficiency, such as sticky prices (Mankiw , Koelln, Rush and Waldo ) or asymmetric information in the capital market (Stiglitz and Weiss ). Though we do not directly address either of these approaches in our paper, with the extremely high inflation rates experienced in Brazil during our sample, presumably sticky prices were not a major source of frictions.
3. Note that this does not seem in keeping with traditional Keynesian analysis that assumes any decrease in the real wage leads to an increase in employment.
4. Not surprisingly, military wages were not covered by the policy. Still, its members protested their falling wages. In one instance, an officer stormed the town hall to demand pay raises for his men; he was denied the raise and charged with illegal troop movements.
5. Realizing that labor would oppose this restrictive wage policy, additional legislation passed in 1964 revised the constitution; for instance, strikes, for the most part, were forbidden. The Antistrike Law passed in June 1964 did permit strikes for improved working conditions or wages, once they were authorized by a regional labor court. From 1964 through at least 1985, most strikes that were ruled legal were those where companies had not paid salaries for three months. The strength of the Labor Code combined with the Antistrike Law effectively crushed strike activities: strikes fell from 225 in 1965 to 15 in 1966, 12 in 1970, and none in 1971 and remained relatively low until the middle of the 1980s.
6. The National Council on Wage Policy provided cost of living deflators and expected profit and productivity figures. Labor courts were required to use these to settle wage disputes. Because workers basically were not allowed to strike, workers had little to say in salary adjustments. It should be noted that these "guidelines" were not maximum adjustments; firms and workers could agree to higher terms, but firms could not use this cost to apply for product price increases. Lopes and Bacha  report that increases above the legally mandated ones were unlikely.
7. Wages were to be differentiated by minimum wage levels; and the productivity coefficient was to be negotiated between labor and the firm and could not be passed on in prices. The minimum wage strata criteria were updated within one year. For example, those earning up to three times the minimum wage received 110% of the national consumer price index while those earning between 15 and 20 times the minimum wage earned 50% adjustments.
8. The unions, with regenerated power, pressured for quarterly wage increases which the administration disallowed. In subsequent plans, wages were frozen; at renewal dates, salaries could be raised by at least 60% of the price index (INPC). Wages were only readjusted during the contract period if the INPC rose by 20% when wages were adjusted accordingly. This wage "trigger" was abandoned in 1987.
9. Price controls also were imposed in mid-1983 but they were ineffective and did not cover much of the economy.
10. We also conducted price level tests to examine the effect of a real wage proxy on output. A majority of the coefficients were insignificant lending more support to the view that nominal wage contracts have little influence on output. In an effort to conserve space in this paper, these results are reported in Dole .
11. The crucial factor about this figure is that the real wage falls over the length of the contract, not the fact that the real wage is assumed to fall linearly. Also not crucial is the assumption that the real wage falls to the same level immediately before indexing in both the annual and semiannual indexing schemes. This assumption does, however, seem realistic because the change to more frequent indexing was caused by the expectation of more rapid inflation. Hence, for a given nominal wage, the real wage falls faster in the later part of the figure. Finally, keep in mind that Figure 1 illustrated a gradual upward drift in Brazilian real wages, presumably because of productivity gains. We have drawn Figure 2 without any upward drift only for simplicity.
12. Again, only the fact that employment increases as the real wage falls is important.
13. To make the data stationary, we chose to first difference the data rather than use log levels with a deterministic trend for two reasons: First, in keeping with the work of Nelson and Plosser , it has long been recognized that many time series are better characterized as being difference stationary rather than trend stationary. Second, and more pragmatic, the output series occasionally shows a distinct downward "step" before continuing upward at its usual rate. Such "steps" were rare, no more than one or two per industry, and were likely due to improved measuring of the output series. While rare, these changes were sufficient to severely bias the trend if we used a level specification. Although we could have accounted for this by using a dummy constant to take account of the "step," given the work of Nelson and Plosser, we preferred the difference specification. Of course, the differenced observations for the month in which steps occurred were large. Thus our estimated regressions included dummy variables, called DUMBEV for the beverages industry, DUMCHM for the chemical industry, and so forth, to take account of these months. We also estimated regressions that did not include these dummies and inclusion - or exclusion - had a negligible impact on the results.
14. This specification assumes that Carnival does not lower the level of output in non-Carnival months as would be the assumption if we used a dummy that was 1 in the month of Carnival and 0 during the remaining months.
15. In addition to the government policy variables included in equation (2), output may have affected by other factors, such as taxes, shifts in relative demands, etc. However, some of these data are unavailable, some are unavailable at the monthly frequency, and, in any case, economists disagree with what factors should be included. Moreover, because these other factors are unlikely to be correlated with the variables that we include, particularly the key monthly dummies, our coefficient estimates and F-tests will be unaffected by their omission. Seasonality, though, may affect our tests: if output is dominated by monthly seasonal factors, then the switch from annual to semi-annual indexing may not be discernible. However, if nominal wage contracts play a minor second string role compared to seasonal factors, it seems hard to argue that such contracts can be tremendously important. Moreover, while some of the industries we use may display seasonality - rubber, tobacco, and perhaps transportation - others seem unlikely to be dominated by seasonality - plastics, chemicals, and beverages.
16. The regression coefficients are similar to those for equation (1), and hence are not reported here.
17. Depending on the initial base date month, the new indexing months coincided with enactment of the legislation for some industries while other industries' new dates remained on schedule with their annual adjustments. For details, see Macedo .
18. It might be the case that costs of adjusting their labor force are sufficiently high so that firms do not take advantage of the lower real wage at the end of the contract when it is known that next month the real wage will rebound. This would lead us to find that the INDEX coefficients are zero. Thus, we do not look at this test alone as definitive. Nonetheless, we think this test is useful because it allows another opportunity for wage contracting effects to be displayed.
19. The coefficients from these regressions are similar to those in Table IV and so to save space are not reported.
Ahmed, Shagil. "Wage Stickiness and the Nonneutrality of Money: A Cross-Industry Analysis." Journal of Monetary Economics, July 1987, 25-50.
Barro, Robert J. "Long Term Contracting, Sticky Prices and Monetary Policy." Journal of Monetary Economics, July 1977, 305-16.
Card, David. "Unexpected Inflation, Real Wages and Employment Determination in Union Contracts." American Economic Review, September 1990, 669-88.
Dole, Carol A. "Nominal Wage Contracts and the Business Cycle: An Empirical Investigation." Unpublished University of Florida dissertation, 1992.
Fischer, Stanley. "Long Term Contracts, Rational Expectations and the Optimal Money Supply Rule." Journal of Political Economy, February 1977, 191-206.
-----. "1947, 1963 and 1985," in Macroeconomics and Finance: Essays in Honor of Franco Modigliani. New York: Norton & Company, 1987.
Gali, Jordi. "How Well Does the IS-LM Model Fit Postwar U.S. Data?" Quarterly Journal of Economics, May 1992, 709-38.
Hall, Robert, E. "The Rigidity of Wages and the Persistence of Unemployment." Brookings Papers on Economic Activity, 2, 1980, 301-35.
Keane, Michael. "Nominal-Contracting Theories of Unemployment: Evidence from Panel Data." American Economic Review, September 1993, 932-52.
Keynes, John. The General Theory of Employment, Interest, and Money. New York: Harcourt, Brace & World, Inc, 1964.
Koelln, Kenneth, Mark Rush, and Douglas Waldo. "Do Government Multipliers Decrease with Inflation?" Journal of Monetary Economics, December 1996, 495-505.
Lopes, Francisco L., and Edmar L. Bacha. "Inflation, Growth and Wage Policy: A Brazilian Perspective." Journal of Development Economics, August/October 1983, 1-20.
Lucas, Robert. "Econometric Policy Evaluation: A Critique," in The Phillips Curve and Labor Markets, edited by K. Brunner and A. Meltzer. Amsterdam: North Holland Publishing, 1976, 19-46.
Macedo, Robert. "Wage Indexation and Inflation, The Recent Brazilian Experience," in Inflation, Debt and Indexation, edited by R. Dornbusch and M. H. Simonsen. Cambridge: The MIT Press, 1983.
Mankiw, N. Gregory. "Small Menu Costs and Large Business Cycles: A Macroeconomic Model of Monopoly." Quarterly Journal of Economics, May 1985, 529-39.
Nelson, Charles R., and Charles Plosser. "Trends and Random Walks in Economic Time Series." Journal of Monetary Economics, September 1982, 139-62.
Phelps, Edmund S., and John B. Taylor. "Stabilizing Powers of Monetary Power Under Rational Expectations." Journal of Political Economy, February 1977, 163-90.
Stiglitz, Joseph E., and Andrew Weiss. "Credit Rationing in Markets with Imperfect Information." American Economic Review, June 198 l, 393-410.
Taylor, John. "Staggered Wage Setting in a Macro Model." American Economic Review, May 1979, 108-13.
Carol A. Dole: Assistant Professor, University of North Carolina at Charlotte, Phone 1-704-547-4171, Fax 1-704-547-6442 E-mail email@example.com
David A. Denslow: Professor, University of Florida, Gainesville, Phone 1-352-392-0151, Fax 1-352-392-7860 E-mail firstname.lastname@example.org
Mark Rush: Professor, University of Florida, Gainesville, Phone 1-352-392-0318, Fax 1-352-392-7860 E-mail email@example.com…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: The Impact of Labor Contracts: Evidence from Brazil. Contributors: Dole, Carol A. - Author, Denslow, David A. - Author, Rush, Mark - Author. Journal title: Economic Inquiry. Volume: 37. Issue: 1 Publication date: January 1999. Page number: 13. © 2003 Western Economic Association International. COPYRIGHT 1999 Gale Group.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.