Does Management Actually Matter? One Trial Says Yes.

"Better management improves productivity" sounds like something that cannot be tested. It has been, once, properly, and the results are more specific than the slogan.

The trial

Economists worked with large multi-plant textile firms near Mumbai. Fourteen plants received five months of consulting from a large international firm: one month diagnosing 38 operational management practices, then four months of implementation support, totalling 781 consultant hours.

Six control plants received the one-month diagnostic only. That matters: the counterfactual is diagnosis without implementation, not an absence of consultants.

What changed

• Quality defects fell by roughly 43%, significant at the 5% level
• Inventory fell by roughly 22%, significant at the 5% level
• Output rose about 9.4%, significant at the 5% level
• Total factor productivity rose about 16.6%, significant only at the 10% level

That last line deserves attention, because the productivity number is the one everyone quotes. It is the weakest result in the paper. Its confidence interval runs from −0.014 to 0.79, which crosses zero.

The clean findings are the operational ones. Fewer defects and less tied-up inventory are large, well-measured effects. The headline 17% productivity gain is a real estimate with genuine uncertainty attached, and reporting it without that overstates the evidence.

The limitations the authors state

They are direct about it: "The major challenge of our experiment was its small cross-sectional sample size. We have data on only 28 plants across 17 firms." Take-up was 26%, so self-selection is real. And profitability was imputed rather than measured, because they could not obtain the firms’ accounts.

There is also a conflict of interest they raise themselves: the consulting firm delivering the treatment also collected the performance data, mitigated by graduate students on site for six months.

The correlational picture

The World Management Survey has now scored over 13,000 firms across more than 35 countries using double-blind telephone interviews with middle managers, each running 45 minutes or more and costing several hundred dollars.

The relationship between management score and performance is monotonically positive across deciles for productivity, profit, growth, exports and patents. The researchers estimate management accounts for around a third of the total factor productivity gap between the United States and other countries.

They are equally clear about what that cannot show: "non-experimental evidence has limits, including the omnipresent omitted variable bias and reverse causality issues".

What it means for you

If you are considering management, the evidence says the job has real leverage and that it is a distinct skill rather than a reward for doing the previous job well. That second half is the finding in the Peter Principle research.

If you are choosing an employer, management quality is a real and measurable difference between firms, and it is worth asking about directly. The related question of when to take the role is in managers before they are ready, and the individual-contributor alternative in the software engineer guide.