Introduction: decompression in the development discourse
The US Securities and Exchange Commission is consulting on a proposal to require public companies to report the ratio of top executive compensation to the median compensation of their employees. ‘The gap in pay between chief executives and rank-and-file employees has been growing steadily, and now regulators want companies to tell investors just how wide it is,’ said the New York Times in its recent report.
Yet the essence of the SEC proposal is at odds with the way we have approached public sector pay reform in developing countries over the last twenty years or so. During that time, Bank and Fund reports have routinely focused on the ‘compression ratio’ – that is, the differential between the highest and lowest paid employees – and habitually recommended decompression. See, for instance, this IMF Technical Note which includes the compression ratio among the indicators to be used for evaluating government employment and compensation. Altering the ratio in favour of the highest earners would boost performance incentives for all, it was thought. A World Bank report for Timor Leste from 2011 is a typical example:
The [pay] changes of 2008 … tackled the problem of weak monetary incentives by increasing the dispersion of wages between base grade and senior grade staff. Timor Leste now has a wage compression ratio of 7.5:1.
At the same time, there has evidently remained some unease over assumptions about a ‘right’ ratio, and more broadly about how blunt a tool the compression ratio is for public sector pay reform. In that spirit, this blog argues that the way the compression ratio has been deployed in the development discourse has been unhelpful, at best. It could, however, be used to point us in a more helpful direction.
Problems with the compression ratio and prescription of decompression
There are several problems with the compression ratio, from trivial to fundamental.
The first is that there is no agreement about how to define it. The Bank defines it as the ratio between the pay of the highest and lowest earners on the main salary scale of the public service. For OECD, however, it is the ratio between the median pay of the top and bottom 10 percent of earners. That is presumably in order to avoid a couple of outliers from skewing the overall ratio.
The second problem is that there is no uniformity in what is treated as making up an employee’s pay. Sometimes it is equated with wages only, sometimes with remuneration including all forms of allowances (which can be substantial). In the 1990s, Uganda found that its compression ratio of 6.8:1 increased to 100:1 after non-monetary allowances and benefits were included!
These two problems combine to create a third problem, which is that cross-country comparisons tend to be unreliable, because like is not being compared with like. Schiavo-Campo et al.’s remark that, ‘There is no more hazardous cross-country comparison than in the area of civil service employment and wages,’ applies very strongly to international comparisons between compression ratios. This is not just because the ratio may be defined differently, or pay calculated differently, in different countries. It is also because what constitutes the public service may differ markedly. For example, if all lower-level jobs are contracted out, the ratio will be lower than if they are performed in-house. Further, it is because the broader labour market context can vary enormously – from the existence of few to plentiful alternative opportunities for different skill levels of labor.
This brings us to a fourth problem. If cross-country comparisons are so problematic, what was the basis for the habitual presumption in favour of decompression? Sometimes, these recommendations were based on rules of thumb about a ‘right’ ratio (anecdotally, the IMF was using a ratio of 7:1 for this purpose in central and eastern Europe in the 1990s), or comparisons with a ‘typical’ range of ratios. Yet the Bank’s own survey of compression ratios, in which the ratios for the 34 countries collected ranged from 1.5:1 to 33:1, belies the idea of a ‘right’ ratio. Sometimes, recommendations of decompression seem to have been articles of faith, as in Malawi, where the Bank was advocating decompression at a time when the compression ratio was already 32:1!
Conclusion: where to from here?
Given that employees tend to be extremely sensitive to changes in pay differentials, and the political ramifications of antagonising public servants can be considerable, should we ever have been making recommendations about public sector pay reform on such flimsy grounds? Of course not. The compression ratio is a summary statistic. It should never have been used as an actionable indicator – that is, judged ‘too low’ and thereby used to justify recommendations of higher salaries for the top end or lower salaries for the bottom end.
Is the compression ratio of any use? Yes, but only for what it is: a summary statistic. It can be a useful starting point for asking further questions about pay and grading, questions that will be more incisive if the compression ratio for a given country can be compared with ratios calculated in the same way for a set of countries with comparable public service structures and similar labour markets. Does a low ratio indicate that there are only a few grades in the grading structure, or does it indicate that the differentials between grades are very small? Does a high ratio indicate that workers at the bottom of the salary structure are unskilled, or are earning the national minimum wage? Or does it indicate that workers at the top of the salary structure have managed to capture inordinately high wages and allowances?
Looked at in this way, the compression ratio can direct us to aspects of pay and grading that we need to understand. One of those aspects is the issue of wage differentials between grades, about which we will blog in a week from now. In the meantime, we would love to hear the views of anyone who has an interest in public sector pay, whether as an employee, a taxpayer or just a citizen!
Join the Conversation