for recent values of "deep"...

[link|http://www.nber.org/cycles/recessions.html|The NBER's Recession Dating Procedure in the Light of Current Developments] says industrial production fell 4.6% (peak to trough) and employment fell around 2%. The peak to trough was around 9 months.

Some alternate data from the Federal Reserve is show in Table 11 of [link|http://www.federalreserve.gov/releases/g17/20010216/g17.pdf|this] PDF (beware - it's over 200 kB). It shows the quarterly industrial production numbers as changing as follows:

1981:
Q1 - +1.9%
Q2 - +2.2%
Q3 - +4.1%
Q4 - -10.5%

1982:
Q1 - -6.4%
Q2 - -5.2%
Q3 - -7.3%
Q4 - -7.5%

In contrast, the 1990-1991 recession was:

1990:
Q1 - +2.0%
Q2 - +0.6%
Q3 - +1.0%
Q4 - -5.8%

1991:
Q1 - -8.3%
Q2 - +1.5%
Q3 - +6.2%
Q4 - +1.1%

I haven't been able to find NBER data on changes in employment and industrial production for the 1980-1981 recession in my brief looking...

So, at least by the standard of the 1981-1982 recession, the 1990-1991 recession wasn't deep. I agree with your statement that it would be a waste of money to push for a NMD system though.

Cheers,
Scott.