Is the Keynesian multiplier a reliable measure of desirability

Projects such as the HS2 railway in the UK continue to base their macro economic desirability on the Keynesian multiplier but is this measure robust?

For example a combination of potential cost overruns and leakage of expenditure into external economies implies that the domestic multiplier may be less than that calculated.

In the case of HS2 the multiple (varying from report to report) is 1.5 to 2 times the estimated expenditure of £32bn. Negative variations could easily reduce this multiple to +/- 1.

The multiplier was introduced in the 1930’s as a simple way of illustrating the extent to which aggregate demand increased in response to capital spending by government. This was an era of more insular economies and infrastructure projects were then more labour intensive than today.

In the era of mechanisation and globalisation are policy makers deluding themselves that this simple measure of infrastructure expenditure impact remains positive and is there a better alternative?