When top-down doesn’t work
Should policy be driven top-down (large units of government creating the program and distributing resources) or bottom-up (local units of government inventing independently)?
It’s a puzzle, Batman
If you ask around the world, the great body of experts will tell you that programs have to originate nationally first. In the last few years I’ve become much more skeptical of that premise.
That premise is gone, baby, gone
Aside from all the time I spend working with housing programs, I spend enormous time designing or analyzing housing programs, always with a pragmatic orientation – do they work? Do they produce quality affordable housing that improves neighborhoods or strengthens families?
I’m on an enduring Quest for knowledge
Because programs are complex, they have four levels of program definition (statute, regulations, administrative guidance, and case practice), and are created and operated across three levels of government (national, state or provincial, and local). Recently the
We whip up zee tasty local concoctions, yes?
Even though we all gravitate to the largest unit of government because that’s where the biggest potential impact is, over time I’ve become much more skeptical about top-down programs, and identified several circumstances when top-down doesn’t work. [Another post, currently contemplated, will cover the situations when top-down does work. – Ed.]
Drum roll, maestro, please!
Keep your shirt on, bro, the post is coming
1. Too much variety of local conditions. Large-scale programs work only when conditions locally are more or less the same – and that requires either largely undeveloped markets.
The Jeffersonian section worked to grid the
But
Program design principle: Programs with a national reach need to focus on matters whose scale is national – like capital markets and income taxation, which apply throughout the country. It is a capital mistake, as Sherlock Holmes the housing finance specialist would say, to specify granularity beyond the point of national consistency.
“It is a capital mistake to over-specify a program”
2. Overly prescriptive while financially insufficient. Money works because the dollar bill doesn’t care for what it is spent – it’s flexible.
Money, like silly putty, can be stretched into any shape you need
Any program designer giving out government money becomes drunk on prescriptivity. Exercise your free choice, most programs say, but do it precisely the way we want you to. That can work when the funding is comprehensive – as in military construction of housing, say, where the Corps of Engineers specifies everything and then pays for everything it specifies. It fails miserably when the resource is essential yet inadequate
Program design principle: If your programmatic resource will fund only (say) 15% of total development costs, don’t prescribe too much and don’t insist that your rules trump everybody else’s, or people will migrate away from your program. [See Section 8 or public housing operating subsidy.]
3. Too many administrative layers. Remember the adolescents’ party time telephone? How messages got garbled along the way? Now add to that game the ability of each link in the chain to add his or her element of personality, self-interest, distraction, or competence.
I’m only changing the message a little
The more administrative layers between program designer and recipient, the greater the ’signal degradation’ (in communications-theory terms), and the greater potential for confusion and corruption to creep in. What gets to the recipients that they can effectively use can be a tiny fraction of what is intended.
Program design principle: The larger the program, the fewer the rules. Unfortunately, this principle is more honored in the breach than the observance.
4. Error in impact radius. In ballistics, the measure of error is ‘impact radius’ – that’s what circle inside which you’re pretty well guaranteed your projectile will land. The analogous financial impact radius is a conceptual measure of how far the properties produced under your program, and their economic results, vary from what the designers envisioned. When you send money down from on high, you often miss the target.
All our misses average out … don’t they?
Program design principle: If you’re targeting niches, use a localized program. If you want national scope, aim for broad income bands. The home mortgage interest deduction works as a national concept; inclusionary zoning will be regional or statewide at the largest.
5. Forbidding customization. Program participants will think of things program designers did not. There is thus a tendency to festoon sweeping legislation with thou-shalt-not’s. Lack of granular knowledge of the local environments means a ‘one-size-fits-all’ approach is the default.
You’ll grow into it
Program design principle: As the programmers say, “it’s not a bug, it’s a feature.” Be tolerant of unexpected uses that nevertheless serve the public interest. The Low Income Housing Tax Credit has funded, by a count I once did, over 30 different specialized tenant population types.
6. Flash flood funding. By the time elected officials have recognized a problem, and the legislation has wended its way through the tortuous journey from concept to law to regulation to actual funding, the situation has always changed. For better or worse – usually, it must be admitted, for worse – the program participants have coped. The cavalry, in short, is always late.
Our funding flashed over an unstable foundation
Program design principle: The most effective programs provide steady funding at a slowly rising rate, not a huge bulge of money thrown at a temporary problem [see Katrina relief – Ed.].
7. Unbalancing the growth of capacity. Too much capacity at the top can cover up too little at the bottom. If the central entity is too strong, it always looks with benevolent disdain upon smaller levels of government, always finds local government wanting, and seldom vests genuine authority – authority to screw up big time – in the lower levels. Excessive centralism, once established, tends to be irreversible (witness
I’ll let it out of my claw when the other one’s strong enough
Program design principle: Even in a national program, use revenue-sharing or block grant to force real estate decisions down a level. It’s no coincidence that the programs Congress now uses – LIHTC, HOME, CDBG – all involve a block-grant or per-capita resource.
8. Enabling dependency. By always providing training wheels, the parent assures the child will never fall, and along the way the child never learns to ride the bicycle. Worse, the child mistakes stable tricycle pedaling for genuine bike-reading.
Skinned knees have a purpose. So do local mistakes, locally corrected.
Program design principle: There’s no such thing as a zero-defects program. Mistakes made by new participants represent valuable learning. Build in failure tolerance – money you can afford to lose.
But I learned something!
As I wrote in a previous blog post, “How much money can you afford to waste?”
The same is true for public policy. Problems are complex; new approaches always involve doing things without knowing how they’ll work. Some things inevitably go wrong — and afterwards, we can see why they went wrong.
The insight of hindsight — my goodness, in retrospect, it’s so obvious! — can be viewed in one of two ways:
Critical. “You should have figured that out beforehand.”
Experimental. “Hey, we discovered something.”
Both views are valid — most things are inferable beforehand — but there’s a vast difference between deducing something works and demonstrating it works.
In other words, “how much can you afford to waste?” is another way of asking, “what is the price of knowledge?“
There’s a reason knowledge is power. Because it’s bought so expensively.
Write a comment