Skip to content

Multipliers from input-output analysis tend to overstate any additional local employment generated

arrow down

Some months ago, we launched a toolkit on local multipliers. Broadly speaking, local multipliers estimate how many `indirect jobs´ would be created in a local economy if an economic development project generates new jobs. For example, if a government subsidy brings a new car manufacturer to an area, local multipliers aim to estimate how many additional jobs will be created in other local manufacturing firms in the supply chain, local service sector jobs in shops and restaurants, etc.

The evidence we summarise in the toolkit is different to that which underpins most multipliers used in policymaking decisions. Rather than using multipliers derived from `input-output´ modelling (the I-O multipliers) it looks at available empirical evidence on the size of local multipliers that have been generated in practice. That is, it is based on what has happened when employment changed in a given location. This overcomes one of the main caveats of the I-O multipliers – that they might miss general equilibrium effects, such as price changes by suppliers, that may be very important for the size of any multiplier in practice.

In this blog, we provide a summary of the comparison between I-O multipliers from the Office for National Statistics (ONS) and the ones that we provide in our local multipliers toolkit. We used the table of I-O multipliers to calculate tradable-to-tradable multipliers and public-to-private multipliers. Unfortunately, the table of I-O multipliers does not allow us to calculate tradable-to-non-tradable multipliers. Further details can be found here.

How do the two approaches compare? The results are summarised as follow:

  • The average tradable-to-tradable I-O employment multiplier is .73, which means that for each additional job in the tradable sector, it is predicted that .73 jobs would be created in other parts of the tradable sector. This estimate is higher than the average local multiplier of .41 that we report in the toolkit.
  • For high-tech tradable-to-tradable the ONS I-O multiplier predicts that adding a job in that sector leads to 1.4 jobs in the rest of the tradable sector. For our toolkit, we couldn’t find any evidence on employment multipliers for high-tech tradable to tradable, but the finding that I-O employment multipliers are larger for high-tech is consistent with our toolkit.
  • For public-sector-to-private sector multiplier, the ONS I-O multiplier of .37 is again higher than our toolkit estimate of .25.

Why are ONS I-O multipliers larger? There are two possible explanations:

1) I-O multipliers pick up the effect across the UK, whereas our toolkit focuses on local effects. Given that some supply comes from outside the area, such leakage should lead to an overestimate of the local effect when using the I-O multipliers.

2) Price changes might dampen real-world multipliers compared to I-O multipliers.

For both of these reasons, using traditional I-O multipliers might tend to overstate any additional local employment generated. This has important implications for ex-ante appraisals of projects which are likely to report local employment effects that are too large if they rely on I-O multipliers.