Kazakhstan’s long labor strike: Three misunderstood issues

Is “equal work for equal pay” really the question?


The oil-workers’ strike in west Kazakhstan has persisted for months partly because some underlying disputes are complex. Here’s a look at three.

Taxes. Ostensibly, the firms facing the strike pay the income taxes of their workers. Some observers may think that this reflects the good will of the employer and should elicit the same from employees.

But what matters is who, in the end, really pays the tax. It could fall on any of the three groups in the oil industry – workers, employers or consumers – or on some combination of them.

For example, suppose that consumers are willing to pay almost any price for oil. Then the firm can pass on to them the cost of the tax by raising oil prices without worrying that they will buy less oil than before. In that case, the tax will not reduce wages or profits.

Or suppose that workers in the area are willing to accept virtually any wage, because they cannot find work elsewhere. Then the firm can pass on to them the tax by cutting wages. Oil prices won’t rise, and oil profits won’t fall.

In short, the tax is paid in the end by those who consent to pay it.

Who really pays the personal income tax in Kazakhstan’s oil industry? That question can’t be answered off the cuff. Oil firms are dominant employers in Mangistau Oblast, so they may be able to compel workers to pay the tax in the form of a wage cut. But consumers of Kazakhstani oil may be insensitive to the price, too -- if much of the oil is bought under contract (not on the spot market, where Kazakhstan would have to compete with other countries selling oil); and if the contract permits price increases. Finally, owners of the oil firms may be sensitive to reductions in profit, because they can invest their money elsewhere. In that event, they will not accept the tax in the form of a haircut.

The point is this: That employers initially agree to pay labor taxes need not mean that they will actually pay them in the end.

Likewise, the pension contribution that the employer agreed to pay may actually fall on someone else.

The cost of subsistence. Kazakhstan’s minimum wage is proportional to the amount that a worker needs just to survive, notes an analyst. “Unfortunately, the subsistence minimum does not meet international standards: The percent share of a food basket included in the subsistence minimum compared to non-food products is 60:40 instead of 30:70.” So, the government should double the subsistence minimum – and it should raise the minimum wage accordingly, he contends.

This argument raises two issues – one conceptual, the other mathematical. Let’s take them in turn.

In the West, governments often define poverty as an income that falls below the amount that the household needs to subsist indefinitely. This amount is called the “poverty line.” To determine it, the U.S. government usually estimates: The cost of the cheapest nutritious food-basket; the share of income that a typical household spends on food; and finally the income at which the household would spend this share on the cheapest food-basket. For example, suppose that this food-basket for a family would cost $1,000 a year – and that the typical family spends one-third of its income on food (as Mollie Orshansky estimated when designing the U.S. poverty line in the 1960s). Then the poverty line for families is $3,000 (see the Notes for the arithmetic).

The share of income spent on food is not set in concrete. As the household’s income increases, its spending on food does not rise as quickly, since our stomachs won’t accommodate an indefinite amount of food. The food share of income falls as household income rises – a phenomenon observed so often that economists have come to call it Engel’s Law, in honor of the 19th-century statistician who formulated it.

Thus, in a rich country, the typical household spends a smaller share of its income on food than does a household in a poor country. Astana should probably estimate the food share of income based on statistics for Kazakhstan, not on those for richer or poorer countries.

As the food share of income falls, the poverty line rises, since even a poor household should be able now to buy proportionally more non-food products than before. For example, suppose that the food share of income is now one-fourth rather than one-third. Also suppose, for simplicity, that the food basket still costs $1,000. Then the poverty line will rise from $3,000 to $4,000. If the minimum wage is proportional to the poverty line, then it too will rise by a third.

Were the government to adopt the analyst’s recommendation of a food share of income of 60%, then the poverty line would fall – from $3,000 to $1,667, given a $1,000 food-basket. The minimum wage would also fall.

Hazard pay. In dangerous industries like mining, the government requires what is effectively a higher minimum wage than usual. A typical multiplier is 1.8. “The value of the [Minimum Standard of Pay, which is basically the minimum wage for a particular industry] for all the companies in all the industries should be the same,” argues an analyst. “Otherwise, the principle of social justice as well as the principle of equal pay for equal work will be violated.”

What concerns the worker is not just pay but also his working conditions. Relative preferences over pay and safety differ from one worker to another. Some workers are willing to take risks in exchange for higher pay; others would rather have more safety than more pay. Requiring all firms to pay the same wage for the same kind of work, regardless of the risk that the worker runs at particular firms, would prevent the sort of pay-for-risk tradeoffs that would benefit workers.

People sometimes ask: Why not give the worker high pay and perfect working conditions? The reason is that safety costs. In a competitive industry, the firm doesn’t make much money – its rivals force it to cut its prices – and so it can’t earmark profits to pay for more safety; there are no profits to earmark.  (A happier prospect is that additional safety may enable the worker to produce more and thus pay for itself.)  The worker may have to finance more safety out of his back pocket. It should be up to him to decide whether he’d like to pay for more safety; the government should not force him to do so by prohibiting any wage that is not identical to those paid by other firms.

Anyway, the issue here is not really “equal pay for equal work.” It's market power. When one firm dominates hiring in the region, then it may ride roughshod over the preferences of workers concerning safety and pay. Because the worker cannot simply walk across the street for a new job, the firm can force him to accept a wage that is well below the actual value that he adds to the firm. The firm thus collects a profit from his labor that could indeed pay for more safety.

By mediating labor disputes impartially, the government might be able to compel this firm to take account of the worker’s preference for safety. When the government consistently sides instead with the dominant employer, then the worker may feel that he must “vote,” so to speak, with either his voice or his feet.

In particular, mediators should ensure that employees as well as employers are well-informed about workplace risks so that they can negotiate sensibly.  And mediators should determine whether the wage offered to an additional worker is well below the value of her work.  This would indicate that the firm owns the local labor market, since evidently the worker cannot get a better offer from a competitor.  -- Leon Taylor tayloralmaty@gmail.com


Notes

1.Let s denote the income share spent on food; Y, household income; and C, the cost of the cheapest nutritious food-basket. Then the poverty line is that value of Y – call it Y* -- that satisfies sY = C. Solving, Y* = C/s. In our example above, s is 1/3 and C is $1,000. So, Y* = $1,000/(1/3)) = $3,000.


Good reading

Kanat Berentaev. The strike of oil industry workers: analysis of reasons and ways of solving the situation (short thesis). algadvk.kz

Ronald G. Ehrenberg and Robert S. Smith. Modern labor economics: Theory and public policy. Ninth edition. Boston: Pearson Education. 2006.  Discusses the tradeoff between wages and safety.

Gordon M. Fisher. An Overview of (Unofficial) Poverty Lines in the United States from 1904 to 1965. United States Bureau of the Census. Online.

Gordon M. Fisher. Mollie Orshansky: Author of the Poverty Thresholds. Amstat News. September 2008. Pages 15-18. Online.

Washington University, St. Louis. (Ernst) Engel's Law & Curves. http://faculty.washington.edu/krumme/resources/engel.html.

Minimum wage – maximum unemployment?

Does the minimum wage bite the hand that it is intended to feed?


According to its supporters, the minimum wage – the lowest wage that employers can legally pay – protects workers who are too weak to negotiate a tolerable paycheck on their own with employers. Critics of a wage floor point out that a high wage may attract more workers than firms want to provide. In theory, in an unregulated labor market, this imbalance is temporary. Eventually, the excess workers – the unemployed – will offer to work at lower wages. This will force the market wage to fall until no excess labor remains. But if a law prevents the market wage from declining, then unemployment will persist. The minimum-wage law hurts the workers that it was ostensibly designed to protect.

Applied to Central Asia, this argument may have its flaws, both theoretical and empirical. It assumes that the labor market is competitive: Employers and employees are so numerous that no one of them can affect the market wage. This is determined instead by their collective decisions. In this setting, no employer can pay another worker less than she is worth, since she has lots of other opportunities. But as a former Soviet satellite, Kazakhstan still has many areas that are dominated by a single employer -- so-called “one-company towns,” which may have numbered in the dozens in mining, metallurgy and agriculture during the Nineties, noted two World Bank economists, Martin Rama and Kinnon Scott. Kazakhstan may have even more company towns than is typical in the Commonwealth of Independent States, because Soviet industrial subsidies per capita were larger here. The power of these companies over Kazakhstan’s labor market is suggested by a finding of Rama and Scott that wages rose when a large company moved into a small town and fell when it left.

Bad company

In company towns, given an unregulated labor market, the worker must either accept the employer’s pay offer or not work at all. Thus the employer may set the wage well below the worker’s value in production; to induce her to accept the offer, he needs only to leave her slightly better off than she would be when unemployed.

In addition, the boss may refuse to hire many worthy workers. The reason given by economic theory is that he usually pays the same wage to all workers of a given type (for example, all recent high-school graduates applying for jobs at his plant). To hire another worker, he must raise his pay offer, since she had refused a job at the earlier, lower wage. But if he raises pay for this worker, he must also do so for all of his other workers of the same type. The cost of doing so may exceed the value of the additional worker to the plant; and so he won’t offer to hire her.

In a company town, a minimum wage may raise the workers’ pay without threatening their jobs, since each employee would still add more value to the firm than she cost. Also, the boss may hire more workers, because he no longer perceives a large expense in raising their wage. He must pay the minimum wage (at least) to all workers, so he will hire anyone who is worth that wage.

A simple example may help clear up these ideas. Suppose that the employer dominates the local labor market. A worker now applies for a job. Suppose that she would be worth $10 an hour to the employer but would be willing to work for as little as $3. Then the boss will offer her only $3 and pocket $7 as profit. Now suppose that the government sets the minimum wage at $9. The boss would still hire the worker, since she more than pays her way: But his profit on her labor will drop to $1 ($10 minus $9). The other $6 of profit now go to the worker.

Real wage, real world

Empirically, the claim that the minimum wage creates unemployment is controversial. The most common conclusion is that the minimum wage increases unemployment, holding all else constant. But some statistical studies conclude that the minimum wage has no effect on unemployment – or may even reduce it. One reason may be that governments rarely adjust the minimum wage for inflation. As prices rise, the amount of goods will diminish that, say, a thousand tenge can purchase. If the government does not increase the number of tenge in the minimum wage, then eventually its real value may fall below that of the wage that would clear the labor market. In that case, the minimum wage has no effect on workers; it’s just a wage floor, and the market pay that they actually receive already exceeds it.

In December 2010, the government set the monthly minimum wage at 13,728 tenge (about $92 at the current exchange rate).  This is 2.5 times the “poverty line” (so called because a person making less than this is deemed by the government to be poor), according to the news source Interfax-Kazakhstan. The minimum wage is higher for manual laborers as well as workers in hazardous industries, noted analyst Kanat Berentaev; in that case, it may affect those markets more than others. The labor ministry has said it would increase the minimum wage by 2015 as it restructures the policy, reported neweurasia.net . Perhaps, as officials redesign the minimum wage, they may wish to analyze what, if anything, it has actually accomplished. – Leon Taylor, tayloralmaty@gmail.com


Good reading


Ronald G. Ehrenberg and Robert S. Smith. Modern labor economics: Theory and public policy. Ninth edition. Boston: Pearson Education. 2006. Surveys work on the minimum wage.

Robert H. Frank. Microeconomics and behavior. Third edition. McGraw-Hill. 1998. Discusses the effects of the minimum wage in a labor market dominated by one employer.

Kanat Berentaev. The strike of oil industry workers: analysis of reasons and ways of solving the situation (short thesis). Center for Analysis of Public Reforms. Online at algadvk.kz .

Martin Rama and Kinnon Scott. Labor earnings in one-company towns: Theory and evidence from Kazakhstan. The World Bank Economic Review 13:1, pages 185-209. January 1999. On line.




References



Interfax-Kazakhstan. Kazakh ministry of labor adjusts poverty line for Q1 2011. January 6, 2011. Online.

neweurasia.net . Minimum wages in Kazakhstan and abroad. January 5, 2010. Online.

A wail of two cities

Do the cities of Kazakhstan attract more workers than they need?


Why is the unemployment rate in Kazakhstan’s two major cities – Astana and Almaty – unusually high?

Generally, these rates – which measure the share of the labor force that is looking for work -- have exceeded the national rate since 2005, according to data from the national statistical agency. Almaty's rate was a sixth higher than the national one in 2009 and remained a tenth higher in 2010. Astana's rate is usually lower than Almaty's but was still a twentieth higher than the national rate in 2010.

Among the oblasts, the only unemployment rate that has been as consistently as high as those of the cities is that of Mangistau, in west Kazakhstan’s “oil country.” It was a fifth higher than the national rate in 2005-6 but then dropped sharply, beginning to rise again in the national slowdown of 2009. (For the figures, see the notes below.)

One might expect a low unemployment rate in cities, because they attract employers. If the firm sells in a national or global market, then it will receive the same revenues regardless of where it locates in the country. So it should put down its roots wherever it can produce most cheaply. Typically, that place is a city, for several reasons.

First, a firm can cut its costs by locating near others in its industry. This cluster of producers generates enough demand to support specialized inputs nearby. When two flour mills locate side by side, they attract a pool of flour workers. Both mills will benefit, since they no longer need to recruit workers in distant cities.

Similar reasons may help explain the growth of high-tech industries in the United States, such as Silicon Valley in California or Route 128 near Boston. For example, computer manufacturers sometimes need workers to repair instruments. It usually isn’t worth the factory’s while to simply keep repairmen on payroll for a task that may take just a few weeks each year. But if several manufacturers locate near one another, then they can support a fulltime firm devoted to instrument repair. Such factors may also help explain why, in the Eighties, over half of the steel industry concentrated in Pennsylvania, Ohio and Indiana; and nearly half of the auto industry in Michigan, concluded the American economist Gerald Carlino.


Going steady


Workers may also prefer the cities, since labor conditions may be more stable there than out in the country. The urban economist Arthur O’Sullivan explains why. When firms in the city step up hiring, the wage will begin to increase. This, in turn, will attract immigrant workers, who will vie for jobs by bidding the wage back down. Similarly, when urban hiring declines, wages will begin to fall. Many workers will respond by leaving. Since they no longer bid down the local wage, this will drop only a little. In short, the urban wage is fairly steady over time, whether firms are demanding more workers or fewer. Now consider an isolated rural area, where workers must take any jobs that they can find, almost irrespective of the wage. Changes in firms’ demand for labor can touch off sharp changes in the wage, since there is no moderating change in the number of workers.

So, workers who prefer steady wages may move to the cities. In turn, this expansion of the urban labor force may attract employers, creating jobs.

Why, then, is the unemployment rate so high in Almaty?


Do bright city lights blind?


The traditional explanation of unemployment is that high wages are attracting more job seekers than firms want to hire. In a free labor market, the jobless eventually will offer to work for less than the going wage, eliminating the unemployment. But minimum-wage laws – which ban low wages – may destroy jobs for unskilled workers whom firms would employ only for rock-bottom pay. (Whether minimum-wage laws are actually so damaging is controversial. For a crisp introduction to the debate, see the text by Ronald Ehrenberg and Robert Smith.) However, it is not apparent that wages are too high in Almaty but not in the rest of Kazakhstan. The minimum-wage law is national.

Perhaps Almaty workers are willing to put up with a little joblessness in exchange for the city’s amenities – the “bright city lights.” Unlike the steppes of Kazakhstan, Almaty offers a myriad of shops and schools. This may also help explain why the unemployment rate is higher in Almaty than in Astana, which is not yet as diverse as its sister city.

Some statistical evidence suggests that urban amenities attract consumers. In the past 20 years, cities with more restaurants and live-performance theaters have grown more quickly, in the U.S. as well as in France. The main amenity is good weather, as measured either by January temperature or by the amount of rain and snow. One statistical study finds that, among counties in the U.S., weather may explain more growth in population or in housing than does any other determinant. "No variable can explain [U.S.] state and city growth over the past 80 years as relably as temperature," concluded economist Edward Glaeser in surveying the literature.  This makes sense. You can’t find a substitute for good weather. (Astana, take note.)

To measure the value of amenities, economists look at housing rent. People are willing to pay higher rents in cities because they receive higher wages there and because they enjoy the fruits of urban living. If rents are rising more rapidly than wages, then the city’s amenities must be growing more valuable. Via this reasoning, Glaeser, Jed Kolko and Albert Saiz found that large U.S. cities with higher amenity values grew more rapidly than did other cities over the Eighties.

Or maybe Almaty workers tolerate spells of unemployment in order to have steady wages during their long periods of employment. They may dislike a pay cut more than they would like a pay raise of the same size; that is, they are “risk averse.” Whether this explains Almaty’s high rate of unemployment may depend in part on whether jobless workers here can rely more on some source of income – the family or the government – than can the unemployed in the rest of Kazakhstan. But even if this were true, one would still have to explain why workers in Almaty are more risk-averse than those in the remainder of the country.

Another possibility is that migrants to cities are too optimistic. Hearing of an economic boom in Almaty, they may happily conclude that they will surely find a job there now. But over time, potential migrants should become better informed about employment, if only by keeping in touch with friends and relatives in the cities. The rate of unemployment in the cities (relative to the national rate) should fall. In Almaty and Astana, it has actually risen for nearly a decade. The mystery continues. -- Leon Taylor, tayloralmaty@gmail.com


Notes


The following ratios are for the unemployment rate in the given region, divided by the national unemployment rate. All underlying data are from the national statistical agency.

a. Almaty: 2003, 1.01; 2004, 1.05; 2005, 1.04; 2006, 1.05; 2007, 1.08; 2008, 1.11; 2009, 1.17; 2010, 1.09.

b. Astana: 2003, .95; 2004, .99; 2005, 1.01; 2006, 1.03; 2007, 1.05; 2008, 1; 2009, 1; 2001, 1.07.

c. Mangistaya Oblast: 2003, 1.1; 2004, 1.17; 2005, 1.2; 2006, 1.19; 2007, 1.17; 2008, 1.03; 2009, 1.08; 2010, 1.1.


Good reading


Ronald G. Ehrenberg and Robert S. Smith. Modern labor economics: Theory and public policy. Ninth edition. Boston, Massachusetts: Pearson. 2006. Chapter 4 discusses minimum-wage laws.

Arthur O’Sullivan. Urban economics. Sixth edition. Boston, Massachusetts: McGraw-Hill. 2007.

Edward L. Glaeser, Jed Kolko, and Albert Saiz. Consumer city. Journal of Economic Geography 1, pages 27-50. 2001.

Edward L. Glaeser.  Reinventing Boston: 1640-2003.  National Bureau of Economic Research working paper #10166. 2003.  Online.   Published in the Journal of Economic Geography  5(2), pages 119-153.  April 2005.


References

Gerald A. Carlino. Productivity in cities: Does city size matter? Business Review, Federal Reserve Bank of Philadelphia, Nov.-Dec. 1987.

Alfred Weber. Alfred Weber’s theory of the location of industries. Translated by C.J. Friedrich. University of Chicago. 1929. Weber pointed out in 1909 that firms choose the cost-minimizing location. A related work is August Losch’s The economics of location, 1944.

The great indoors

How important is outdoor work to Kazakhstan?


The government of Kazakhstan is disgruntled about the economy, which doubles the average resident’s income every seven years or so. It has launched a program of “forced” innovation and industrialization (“accelerated” may be a more accurate translation) in order to diversify away from production of oil, gas and other raw materials.

Less publicized is the fact that the economy is already diversifying. Like higher-income economies, it produces more services than goods. Oil and gas still dominate exporting but only because of their high prices.

One gauge of the economy’s maturation is the declining importance to it of outdoor work. A modern economy emphasizes brains, not brawn. Where the latter matters, the season’s weather may affect sharply the number of workers.

In Kazakhstan, weather effects on employment are mild for a transition economy. For 2011, the number of workers and job-seekers – i.e., the labor force -- varied between 8.6 million and 9 million, depending on the season: Lowest in the winter and highest in the fall. But the largest monthly workforce (October) is only 367,000 more than the smallest (January and February), just 4.2% of the mean size for the year. Employment too was steady across the months. In 2010, seasonal effects on the labor force and on just the employed were even smaller than in 2011.


The match game


Similarly, the monthly unemployment rate in 2011 steadied between 5.3% and 5.6%, varying by only a twentieth of the mean annual rate. Unemployment is highest in the winter but probably not by enough (18,000 jobless workers more than in the best month) to justify special benefits for those out of work because of the season. Overall, the seasonal effect on unemployment in 2011 was less than half that of 2010; and the average unemployment rate (for January through November) dropped from 5.8% to 5.4%. Thus the labor market may be improving its matches of the jobless to jobs, although it isn’t clear whether the improvement is due to a strengthening economy or to better information for job hunters.

From only two years of data, we cannot conclude that seasonal employment in Kazakhstan is becoming moot. Some seasonal effects were larger in 2011 than in 2010. But estimates using quarterly data for 2003 through 2008 confirm that seasonality effects have been waning steadily over the long run in Kazakhstan. They raise the possibility that the government’s industrial policy addresses a problem that entrepreneurs already are solving. – Leon Taylor, tayloralmaty@gmail.com


Notes


1. All data are from Kazakhstan’s agency on statistics (www.stat.kz) unless otherwise noted. Quarterly estimates may be more accurate than monthly ones, since the government has actually surveyed the labor market each quarter.

2. I measure the seasonal effect on the labor force as the ratio of the difference between the largest monthly size of the force and the smallest number, relative to the mean size for the year. For example, in 2010, the labor force was smallest in January (8,448,400 members) and largest in June (8,667,900). The variation was thus 8,667,900 minus 8,448,400, or 219,500. The mean size of the labor force over 2010 was 8,611,200. As defined here, the “seasonal effect” was 219,500 divided by 8,611,200, or 2.5%. Seasonal effects for the "employed population" (which includes the self-employed as well as employees) and for the unemployment rate are measured similarly.

We may measure seasonal effects in other ways, such as the ratio of the standard error to the mean. The measure that I use here puts more weight on extreme seasonal effects.

Seasonal effects do not always arise from weather.  Holiday spending may also affect employment. But it is probably safe to say that, in Kazakhstan, weather affects significantly the seasonal effects as measured here, even if it is not the only determinant.

My estimates of seasonal effects in 2003 through 2008 are:

For the labor force: 2003, .082; 2004, .032; 2005, .026; 2006, .019; 2007, .022; 2008, .015.

For employees and the self-employed: 2003, .098; 2004, .042; 2005, .034; 2006, .026; 2007, .029; 2008, .021.

For the unemployed: 2003, .086; 2004, .078; 2005, .063; 2006, .073; 2007, .067; 2008, .060.

For the unemployment rate: 2003, .17; 2004, .107; 2005, .086; 2006, .092; 2007, .094; 2008, .075.