Minimum wage: Effects on employment?

The United States' experience has been recently reviewed by The Economist (subscription required). An older, more in-depth review of developing nations by the International Labour Organisation can be found in Saget (2001) (.pdf version).

To summarise, the evidence is mixed, partly because historical studies are 'experiments of history', with multiple factors at work. A few facts stand out, though:

- Columbia. During 1977-87, Columbia's minimum wage rose by 15%, and manufacturing sector employment fell by 5%. [Bell 1997]

- Mexico. Here the minimum wage had no effect on manufacturing employment. However a reduction of the Mexican minimum wage between 1970 and 1990 actually increased the overall employment of women aged 15-64, albeit at the expense of older men. [Feliciano 1998]

- United States. While the US Federal Government sets a minimum wage, individual states can choose to set a higher minimum. New Jersey and Pennsylvania are neighbours, but in 1994 only New Jersey increased the minimum wage. One study of fast-food workers (who are paid the minimum wage) showed no disemployment effects; here the data was collected through telephone surveys [Card & Krueger 1995]. Another research group, looking at payroll records, suggested the opposite conclusion: that employment had been hurt by New Jersey's minimum wage [Neumark & Wascher 1995].

The Mexican study is of particular concern. When Mexico reduced their minimum wage, it increased employment of workers who were female, younger and with less education [Feliciano 1998, reviewed in Saget 2001]. If the converse applies, then the minimum wage may disadvantage vulnerable workers, rather than help them.

This is of relevance when applying the minimum wage debate to Singapore, since the test of the policy will be whether it helps the vulnerable. Our most vulnerable workers have limited education and a shortage of skills relevant to the jobs they seek. Many will have been unemployed for some months, and confined to unskilled work even when they do find a job. Some of them will receive assistance from non-profit or charitable organisations, government-supported or otherwise. A few of these organisations may act as an employer of last resort, directly or indirectly offering 'altruistic' job placement to the vulnerable worker.

How might such disadvantages occur? Let's explore some issues:

If implemented, the minimum wage cannot be restricted to Singaporeans. Otherwise we would see a spate of lower-income Singaporeans being retrenched, their jobs taken up by foreigners working for less. Admittedly this loophole can be plugged by enforcing the minimum wage for all workers regardless of nationality, or by raising the foreign worker levy to narrow the difference in cost to the employer. This generates other problems, though...

Minimum wage legislation acts as a tax on employment. The minimum wage does not grow on trees: it is the employer who has to pay the top-up. Forcing an employer to pay more than the market rate for labour has the same effect on employer behaviour as a tax on every worker hired.

Employer behaviour. As one commentator has pointed out, minimum wage policy is not incentive-compatible. The effect of legislation, however altruistic, is shaped by how employers act in response. Unlike workplace safety legislation, where the 'tax' of short-term inefficiency is directly 'paid' in the form of safety measures for employees, the employer may not respond to the minimum wage by raising salaries without changing how the business is run. If the employer has to pay more for every worker hired, the increase in overheads must be offset: productivity must rise or expenses must fall. Or the business model may have to change. Let's look at a few possibilities:

Job Duties and Hiring Profiles. Productivity could be increased by getting more work out of the employee, but the minimum wage means that longer working hours would incur even higher costs. Employers might choose to sack some workers, while making the remaining ones toil even harder. And if higher productivity is demanded from fewer workers, hiring decisions will favour incumbents or people with industry experience. This is not going to help the most vulnerable workers, who may lack training or recent experience in the industry.

Getting caught in an 'inexperience' trap can be very damaging. The longer one is unemployed, the less current experience and training one has, and the lower the wage valuation of one's labour. But what happens if the minimum wage is set higher than this valuation? Will a prospective employer really be willing to pay extra? This dilemma has previously been raised in 1999 by Alan Greenspan and more recently cited by Greg Mankiw, a professor of economics at Harvard University.

Corporations and Automation. An ongoing trend among larger corporations is mechanization and automation. Skilled workers are less vulnerable to this, since they can retrain more easily and may possess specialised ability which a machine cannot easily replicate. But it is not the skilled workers we are trying to protect with minimum wage legislation: it is the unskilled workers.

In America, which has a minimum wage policy, automated checkout counters and distribution networks are supplanting the unskilled human employee. Even the International Labour Organisation has acknowledged in its Information Sheet on Minimum Wages Policy that "the employment of vulnerable categories of workers (those with very low productivity, casual workers) might be at risk and requires special measures."

Small Businesses. Small businesses often work on thin margins and encounter difficulty during an economic downturn. What happens if their labour costs go up? McDonald's can probably afford to pay more per hour for each employee. But what about the HDB coffeeshop owner, or the local grocery store?

Charities. Many charities and volunteer organisations have salaried employees. They, too, will be affected by a rise in labour costs. With the minimum wage there is no room for downward salary flexibility, even if the employee is happy working for less because his employer is a charity. Some non-profit organisations also employ retirees who, while less sprightly than their younger counterparts, still want to work. A minimum wage policy may mean fewer 'altruistic' hires: again, this hurts the more vulnerable workers.

How much wage is a Minimum Wage?

Often neglected in debates on the minimum wage is the actual amount. If the minimum wage is set near the equilibrium wage for the poorest workers, there may be no effect on employment in the short term - but there may be no effect on living standards either! On the other hand, a significant increase in wages will make a big difference to the quality of life for those still employed, but the economic effects will be commensurately greater.

The dilemma of setting minimum wage levels is discussed in a 1996 ILO paper "Combating unemployment and exclusion: issues and policy options". It appears that in those countries who have implemented it without major increases in unemployment, the minimum wage is relatively low.

Let's look at the United States as a case study. Their federal minimum wage of US$5.15 per hour is currently about 30% of the average wage [The Economist, 26 Oct 2006].

Now, it is not clear whether The Economist's figure represents the mean or median. According to the U.S. Census Bureau's historical tables for personal income, for 2004 the median income was US$23,186 (or US$1,932 monthly). The mean was US$33,846 (or US$2,821 monthly).

Many studies look at median income, because a tiny number of extremely rich individuals can skew the mean upwards. (Wikipedia explanations on median and mean averages are available.) Applying the 30% multiplier to the median income gives us US$580 per month as the federal minimum wage for an American.

How much would a minimum wage have to be in Singapore, to have the same purchasing power as the United States minimum wage? If goods cost the same all over the world, then we could simply use the exchange rate. Unfortunately, the cost of living varies with location. One way to approach the problem is by measuring the cost of a combination of daily household items (i.e. basket of goods): such a basket should contain items such as meat, vegetables and starch, while also factoring in transport and employment costs involved in production - e.g. the McDonald's Big Mac. It sounds silly, and it is not a perfect measure, but it is good enough that The Economist has developed a Big Mac Index.

How does this work? Using the Big Mac Index data, a Big Mac in America costs US$3.10 on average. In Singapore, the same product costs S$3.60 (US$2.27 at an exchange rate of 1.586). This means that US$1 will buy 1/3.1 of a burger in America (32%), but 1/2.27 of a burger in Singapore (44%). In summary, the US$ has about 1.37 times its value if buying the same goods in Singapore.

If that US$1 buys more in Singapore, the equivalent minimum wage would be lower. i.e. To buy the same amount of goods as US$580 in America, you'd only need US$423 in Singapore, or S$671.

In short, S$671 per month is what the US$ yardstick for minimum wage would mandate. This is a far cry from the S$1000 per month advocated by some.

Other Pitfalls of a Minimum Wage

Does not specifically help the poor. Not all low-paying jobs go to poor families. Minimum wage legislation also applies to the child of a wealthy family doing a temporary job before going overseas for university. Why should wealthy households benefit from a measure designed to help the poor?

Inflation. What happens when companies, shops and restaurants have an additional tax levied on them? Some of the costs may be passed on to the consumer. So the cost of living will go up. Even the corner shop has to employ people: they, too, will become more expensive. We may end up back at square one, with the new cost of living exceeding what the old minimum wage can provide for. Do we then increase the minimum wage again? And what does this do to our overall economic competitiveness?

Wage Inflexibility. During an economic downturn, employers can reduce labour overheads by lowering salaries or retrenching people. Minimum wage legislation may promote retrenchment during a slowdown, especially for employers who hire numerous low-wage workers - since it is precisely these minimum wage workers for whom salary cuts would have been de facto outlawed.

Policy Inflexibility. It may be politically difficult or impossible to lower the minimum wage amount, even during a recession with mass retrenchments. This is a simplistic thought experiment, but it may prove instructive: Let us assume that employment among lower-income workers has hit 30%. Yet even in this scenario, the remaining 70% of workers would not want a reduction in their ongoing wages, and might yet vote against any reduction in the minimum wage.

An Alternative Approach

So how do we help people who are currently employed, but unable to make ends meet? This in itself is worthy of an entire blog article, but let me outline a few thoughts:

The hardship of poverty comes from unfulfilled basic needs: food, electricity, water, shelter, sanitation, healthcare, education. I make special mention of education, since lack of access to education perpetuates the poverty trap.

Targeting aid to address these basic needs for poor families is a more precise tool. The aid is targeted at the family who needs it. By providing aid in non-convertible form (e.g. rebates on goods for basic needs, rather than cash), we cut out the middleman and help maximise the value of each dollar given as aid.

Returning to the theme of education, it is also crucial that we educate our population with financial literacy. Poverty is linked to income, but there are also households who have suffered from bad investment decisions, or who have been taken advantage of. But for the ongoing burden of debt or mortgage repayments, some households would actually have enough income to make ends meet.

I will look at questions of education and health in subsequent articles. In the meantime, as mentioned at the outset, your comments and suggestions are most welcome.