By omitting the earnings of one in seven workers from jobs data, our economic policymakers are operating in the dark

Whether you view the self-employed as the silent victims of our invidious jobs market or emblems of a new spirit of entrepreneurialism spreading through society, what is beyond doubt is that the ranks of those working for themselves are swelling by the day.

The numbers have grown by a staggering 700,000 since the financial crisis. Sure, a big chunk of the recent surge is a result of weakness in the jobs market, with more people turning to self-employment as a last resort rather than as a first choice. But most say they are content with their status, and the trend was upwards well before the recession.

To be self-employed is not to reside at the margins of our economy. It’s increasingly mainstream. So it is right that the Office for National Statistics and the government count the self-employed in our employment measures. If they didn’t, we would be using junk data and have a completely skewed understanding of the scale and nature of the employment recovery.

Yet all our measures of earnings exclude these 4.5 million people. That’s an awful lot to overlook – one in seven workers. This lopsided approach means neither the chancellor, George Osborne, nor the Bank of England governor, Mark Carney, knows what is really happening to overall workers’ earnings.

A Resolution Foundation report published today – All Accounted For: The Case for an All Worker Earnings Measure – has a stab at filling this gap. The findings, necessarily tentative given the paucity of timely and reliable data, are nonetheless striking. They suggest that the official measures understate the size of the fall in earnings since the downturn by a thumping 20% to 30%. This important trend has gone unreported while tiny changes in the regularly published wage figures grab the headlines.

The gap between the self-employed and employees opened up because those working for themselves were hit far harder than employees during the early years of the downturn (more recently the prospects of both groups have declined in synch). Counting the self-employed in the official statistics would have dragged overall earnings down even faster. The implication is that there is even more ground to make up than we thought before the typical worker returns to the pre-2008 position.

But this is only part of the story. There are also reasons for thinking that our incomplete measures of earnings may result in an underestimate of their bounce-back over the next few years as growth picks up. Having crashed so severely, it would be natural for earnings among the self-employed to rebound more sharply than for employees – particularly when the self-employed are well represented in sectors, such as construction, whose prospects tend to ebb and flow directly in line with the economic cycle.

Above all, our skewed approach towards the self-employed means we have to interpret all these statistics with care. Imagine that one of the many self-employed workers making less than the minimum wage manages to secure an employee role at that rate. They – and society – just got richer. But under our current system the average earnings figures would actually fall.

For now, all we can say with certainty is that economic policymakers are operating in the dark. The plight of the self-employed – including their earnings – provides a vital and early insight, like canaries in a mine, into how much slack there is in our jobs market. Strange, then, not to pay close attention. If you don’t really know what is happening to earnings, it’s even more difficult to make the right call on high-wire decisions such as when to raise interest rates.

Given the amount of money spent on measuring the UK economy, it is remarkable that so little effort has been made to understand what’s happening to such a large swath of our workforce. In politics, like it or not, it’s what gets counted that matters. When it comes to work and its reward, it’s time to count everyone.

This blog originally appeared on the Guardian website.