Keep Ontario Working (KOW), a coalition of business groups, has released its analysis of Bill 148, the legislation that will increase the minimum wage in Ontario to $15 per hour and introduce important improvements to working conditions. With the legislative window to make any changes to the Bill closing, KOW has opted to make public a presentation of incomplete, “expedited” results. Here are five key things to know about the analysis, which cast serious doubt on its conclusions.
1. The analysis is focused on costs and says little about benefits
The central concern about the analysis commissioned by KOW is that while it has purportedly looked at all impacts of Bill 148, it makes scant mention of benefits, and focuses its attention almost exclusively on costs.
The document released today contains three key, big numbers. It predicts $23 billion in new costs to business spread over two years, 185,000 “current and potential new jobs” that will be put “at risk” over the same period, and $1,300 in new annual cost to households. While these numbers are certain to grab headlines and are clearly intended to sow fear and anxiety among Ontario workers, they appear overblown.
It is true that Bill 148 will impose some new costs on business. But $23 billion is roughly 6% of annual compensation paid by business. Even spread over two years, it is a very aggressive estimate. The estimate of jobs “at risk” is not only out of line with economic research (see the third point in this blog) and unclear---the report does not define precisely what “at risk” means---it is more accurately phrased as a slowdown in job growth. Remember that Ontario added 138,000 jobs just last year. It is more than likely that employment will keep growing even if we allow for these unrealistic and dreary estimates.
But the most egregious "big" number is the report's estimate of $1,300 in new annual spending by households due to a 0.7% price increase. One of these two numbers is the result of a simple math error. If $1,300 was 0.7% of spending then households in Ontario would be spending a whopping average of $185,000 per year on goods and services. The actual average according to Statistics Canada is around $60,000 and this includes $20,000 on average on shelter.
There are two possibilities as to where the authors of the report went wrong. First, they could be off by a factor of 10 in their estimate of household consumption spending on goods and services most affected by Bill 148, intending to include $18,500 and adding $185,000 instead. This would mean that households would be faced with a much, much more modest amount of $130 in new spending per year. Even taking 0.7% of the roughly $40,000 that Ontario households spend on total consumption, excluding shelter, would produce extra spending of $280.
The CANCEA analysts could also have reached the $1,300 figure by taking 29% of $23 billion, the proportion of total costs supposedly passed on in higher prices, and dividing it by the number of Ontario households (around 5 million). Note, however, that $23 billion is the cost over two years, so $1,300 is twice too high. The annual average increase in non-housing expenditure per household would then be $650, or 1.6%, a very high figure.
Either way, without more details about the report (see below), it is impossible to know where exactly the error occurred, but we are certain that this figure is significantly too high.
Nearly left out of the presentation entirely are all the benefits that Bill 148 will create. It will give workers new rights, such as more vacation days and personal emergency leave, increase incomes--substantially for some--and boost the economy as workers spend this new income. Things that look like costs have important benefits: for example, healthier (as a result of sick days) and more committed workers (as a result of lower turnover from higher wages) increase productivity.
Many big numbers are given short shrift in the KOW-funded analysis. For example, over one and a half million Ontarians, or around a quarter of the provincial workforce, currently make under $15 an hour and will therefore be receiving a raise. Full-time workers currently making minimum wage will see after-tax income go up by roughly $5,000 per year. The report also does not touch on the benefits of a more equal distribution of income and lower poverty rates, including decreased social spending by government, that result from an increase in the minimum wage.
In short, the analysis reads like the dog ate half of the cost-benefit analysis homework, namely the half with all the benefits.
2. The model is a black box and the analysis is preliminary
The next thing to say about the KOW analysis is that other than its lack of balance, we know too few details. Today’s release is an “expedited” one of “key” results, produced by an undisclosed economic model.
The Canadian Centre for Economic Analysis (CANCEA), which carried out the analysis for KOW, is known to use a proprietary model of the economy, not subject to peer review, that relies on a complex web of interactions between millions of agents (firms, households and government). However it is unclear whether the results from the impact analysis released today were produced using this black box model, or one of the many other economic models. Either way, without a detailed explanation of the model and methodology, it is impossible for other researchers to do a thorough analysis and corroborate results.
There are many difficulties in using and assessing economic models. Most importantly, all models rely on assumptions, any number of which could be problematic and open to challenge. Each of these assumptions may have an impact on the key results the model produces. Without a transparent process, we cannot tell if some of the results are driven by unrealistic assumptions, or even coding errors.
Low overall transparency raises concerns about the credibility of the key results and makes it difficult to verify them. Modellers normally perform “sensitivity analysis” to see how small changes in assumptions impact results. The presentation suggests this was done for some assumptions, but does not provide details. For example, were other possible estimates of costs and jobs “at risk” significantly lower?
The report provides some details, but these are minimal. For example, we learn that of the total cost to business, 50% will be used to reduce employment, 21% will be absorbed directly and 29% will be passed on to consumers. There is no explanation of which assumptions are driving these particular numbers and how sensitive they are to change. A small change in key assumptions could lead to very different estimates.
(Remember too that many low-wage workers in Ontario are employed by big businesses and in the last weeks we’ve heard CEOs like Galen Weston of Loblaws fretting about being unable to absorb higher costs despite record profits, far in excess of the increased costs.)
A full description of the methodology should have accompanied even these preliminary results. Without this background, there is little ground to put much trust in these numbers.
3. Projected job loss goes against two decades of economics research
The last two decades have seen a consensus emerge within the economics profession that we should expect little to no job loss from typical increases in the minimum wage.
Using improved techniques that carefully isolate the effects of minimum wage increases from the remaining noise in economic data, the weight of evidence from the United States points to job loss effects that are statistically indistinguishable from zero. The few very recent studies from Canada that have used these new economic methods agree, finding job loss effects for teenagers smaller by half than those of earlier studies, and with no effect for workers over 25. Today, the vast majority of workers making under $15 an hour in Ontario are not teenagers; they are working adults from all walks of life, disproportionately women and new immigrants.
This suggests that CANCEA’s estimate of jobs “at risk” is overblown. The report itself states that its results are very sensitive around the assumptions about the economy’s response to minimum wage changes. It would have been good to see the results of this sensitivity analysis.
If all 185,000 jobs “at risk” were lost and half were attributed to the shift to $15 per hour, the CANCEA analysis would imply job loss far outside the norm of modern minimum wage research. This is corroborated not only by individual studies but by "meta-analyses" of minimum wage research, studies that themselves compile multiple individual studies and hundreds of estimates of job loss effects.4. We’ve seen this before
It is not just economic research that should make us question KOW’s numbers; we can also look to recent Canadian experience. Over the years, similar studies with equally dire predictions have been released by business groups and conservative think tanks every time a provincial government plans to raise the minimum wage. And every time, their predictions have been wrong.
The best recent example comes from British Columbia. Starting in May, 2011 the BC government raised the minimum wage from $8 to $10.25 over the course of a year. In relative terms, this boost was almost exactly the same as today’s proposed increase to $15 in Ontario (28% in BC versus 29% in Ontario with a year-long phase in for both). In the lead-up to BC’s change, the Fraser Institute produced a report that also predicted massive job losses, 52,000 in total, disproportionately hitting teenager workers.
In fact, over the course of the year that British Columbia’s minimum wage was increasing and the following year, the province added nearly 50,000 jobs, the opposite of what Fraser predicted, and the unemployment rate dropped by 0.7%. The job losses for teenagers were a mere tenth of what the Fraser Institute had estimated, and even those likely had more to do with a surge in post-secondary education than the minimum wage increase.
5. Effects on government revenues and prices tell a different story
One of the few places where the CANCEA report implicitly acknowledges benefits is in growing federal and provincial government receipts from income and consumption taxes. These figures suggest that, even in the KOW-funded analysis, Ontarians have more income and higher consumption as a result of Bill 148’s higher wages and better working conditions.
Second, as we’ve already pointed out, while the report acknowledges that there will be an increase in prices, its numbers are confused and wrong. Big changes in prices that eat away at wage increases have been a fear consistently raised by the business community. Yet even research funded by its lobbying arm doesn’t support this claim. Indeed, with inflation slowing across Canada, a very modest, short-term rise in prices may be positive---spurring wage growth, demand, productivity and, ultimately, economic growth.
There is strong economic evidence that the provisions of Bill 148, by raising incomes for low-wage workers and contributing to more secure jobs, will lead to a stronger and fairer economy. The self-serving scare tactics of employer lobbyists cannot refute that evidence.
A previous version of this post noted that CANCEA’s estimate of jobs “at risk” would be within the very top bracket of the most pessimistic estimates globally based on a spreadsheet error. The post has been updated to reflect this change in degree, not content. We regret the error.
Zohra Jamasi is an economist with the Canadian Centre for Policy Alternatives’ Ontario office. Michal Rozworski is a research analyst with the Ontario Confederation of University Faculty Associations and a CCPA research associate.