The latest chapter in the “why was the original stimulus so small?” story is a memo from December 2008 that reveals Larry Summers’ assessment as to why the stimulus (ARRA) had to be limited to around $800 billion—about half of what was necessary, in Summers’ estimation.  There are various conclusions you can draw from this memo, but the aspect I’d like to focus on is this:  Larry Summers’ suggestion that $225 billion of “actual spending on priority investments” is all that the government could get out the door over a two year time span (and so the rest had to be made up of tax cuts, aid to states, etc.).

Let’s grant for the sake of argument that Summers is correct about this “shovel ready” figure.  The question is:  what can we do about it?  If you’re looking for short-term results, the answer is probably “not much.”  Even things like speeding up environmental impact assessments for infrastructure projects wouldn’t have much effect (at the link, Brad Plumer tells us that only 4 percent of highway infrastructure projects even require such environmental reviews).

But looking ahead, there is more we could and should be doing.  Back in 2009 Martin Shubik sketched out a plan in a Levy Institute policy note for creating a “Federal Employment Reserve Authority“—a kind of Fed for employment (yes, I know:  the Federal Reserve is the “Fed for employment.”  But you don’t need to look very hard to see that the sides of the dual mandate aren’t equally weighted).  Among other things, the FERA would maintain state branches that are charged with keeping updated and prioritized lists of potential public works projects (with a preference for self-liquidating projects) and providing constant monitoring and evaluation so that financing can be put in place as soon as unemployment reaches a particular trigger level in that region.  Regional public investment would respond to objective employment conditions.

The US doesn’t do a great job with conventional “automatic stabilizers.”  One of the biggest social assistance programs, TANF (Temporary Assistance for Needy Families), is far less responsive to poverty than we might want it to be (less responsive anyway than its predecessor, Aid to Families with Dependent Children).  (See also Peter Orszag’s Bloomberg column today comparing US and European fiscal stabilizers.)

Shubik’s FERA-initiated projects wouldn’t be quite as “automatic” as conventional fiscal stabilizers, but this new setup would have the benefits of (1) speeding up the public works process (being able to move more than $225 billion over two years), and (2) responding in a more mechanical way to employment crises, rather than depending entirely on the outcomes of Congressional procedure.  Such automated responses are in many ways preferable to setting fiscal policy and levels of public investment according to the whim and mood of whichever “moderate”* holds the 60^th vote in the Senate.

Now, you might want to say that these sorts of decisions ought to be kept within the confines of the legislative arena, allowing a clash of economic policy visions to determine the course of policy.  But that’s not really the way the US political system is set up.  The way it works (or not) is that the political party in Congress that doesn’t control the White House is given both the incentive and the ability to see that unemployment remains high.  That’s not a recipe for sensible policy.  Getting reasonably coherent policy responses out of the legislature during an economic crisis would require a different set of political institutions from the ones we have right now; it would require political reforms about as dramatic as the creation of a Shubik-style FERA.

* Note to non-US readers: “moderate” in no way refers to equidistance from ideological poles, but rather to the likelihood of going to work for some civically corrosive lobbying firm after leaving Congress.