Why Government Can't do IT

And why California's unemployment fiasco was both preventable and inevitable

Note - Thanks to the California Globe for running this piece as well. Please check out their website - www.californiaglobe.com

With the rather significant failures of various government technology systems – specifically those involved in implementing proper, fraud-free unemployment benefit disbursement - the question of ‘why can’t government do big tech?’ has lurched back into public view.

If you think government lags woefully behind the private sector in technology standards and is for the most part incapable of performing basic tasks such as buying new up-to-date computer systems, you are correct.

The issue goes beyond the standard bureaucratic inertia (their reaction of “Wait – we have to learn something new and actually work for a living for a while? No thanks!” being a typical driver of inaction), the speed of technological innovation, and the elected and appointed officials all-consuming fear of being made to look specifically inept (generally inept, they already understand is the public’s default view of them but in this case it is the difference between being called a crook at a public meeting by a gadfly and having the beloved long-time local librarian stand up at that meeting and say they saw you rob the convenience store down the street at 6:47 p.m. last Thursday).

There are many factors at play here - first the speed of change.  If I recall correctly, the federal Social Security system was until recently (maybe still is?) still using mainframes they bought from Ross Perot (it’s what he did before he became a bit more well-known) in the 1960s and that when their tech people retire they have to immediately hire them back as consultants to make sure the younger generation knows the coding methods and such to actually make them work. 

So why don’t government agencies buy new stuff?  In part it’s because public agencies all have specific and labyrinthine purchasing systems and regulations.  Initially put in place to keep a governor or mayor or whatever from giving every contract for everything to their idiot brother-in-law, these systems make it impossible to react to the speed of innovation.  A process designed to be extra-super-duper, um, careful simply cannot keep up with the speed of change and/or Moore’s Law.  In other words, if your local Department of Motor Vehicles (DMV) said “we’re buying a new system today,” by the time that system is installed five years from now it would be utterly obsolete and, therefore, be perceived by the public as a giant waste of time and money.

And while the bureaucrats could obviously not care less about wasting time and money, the officials do (or, to be more accurate, they care a great deal if they are seen wasting time and money) because it means stress, irked voters, angry phone calls, and all that rigamarole.  It doesn’t mean they actually listen to, or God help them, actually do anything about the complaints but it really “ripples their pond.”

As to the question “why don’t they just send someone to Silicon Valley?” to garner expertise and find a system that works properly, it is somewhat related.  Because of the current purchasing systems, a company has to have an army of trained lawyers and such to simply navigate the possibility of bidding on any large government contract.  So a nimble and innovative firm would have to spend a ton of money – UPFRONT – to prepare the bid, establish the proper contact networks (and we all know what that means), etc. while they could at the same time have a few folks from sales and marketing create a PowerPoint and hop on a plane to pitch their product to a private company and make the same amount of profit for less costs and hassle.

That is also why you see the same behemothic (and usually not tech-specific) firms getting these types of contracts, time and again, no matter how much they made a shambles of the last contract they had, no matter how much they over-charged for it, no matter how sideways instead of forward-looking it was. They’ve mastered the purchasing system and, even though the system was put in place to deter inappropriate favoritism, the system relies really really really heavily on personal contacts and personal and systemic familiarity (which is also why they lobby hard to keep the system from changing, shouting from the rooftops that modifying it any way would open it up to all kinds of chicanery and self-dealing and idiot brothers-in-law).

It should also be noted that, usually, projects are awarded to the lowest bidder, which we all know is maybe not the best way to get the best outcome.  And the big firms know that and tend to bid projects essentially “at cost” to appear to be thelowest while counting on heavily marked-up change orders and money-making government agency forced delays that are standard in the miasmic fog of such projects for their real profit margin.

Tech is not sexy and new and shiny.  The public expects government computer systems to work, just like they do at home and work.  They understand there will be occasional hiccups and they understand they cost money and they assume government officials understand that, too.  But they do not understand why it takes the government so long to upgrade and adapt since they just go to BestBuy and get a new one if they need it (they may just test drive it at BestBuy and then go and actually get it online cheaper but that’s neither here nor there).

In making major outward-facing systems upgrades there are extremely high political risks with essentially zero immediate political gains, so why bother?

Using California as a jumping off point, you can think of high-speed rail and think of the state’s unemployment insurance system (the Employment Development Department, or EDD) fiasco and you get the picture.

While, for some unfathomable reason, the state’s mind-numbingly over-budget, comically behind schedule, incomprehensibly managed high-speed rail project hasn’t ended any political careers, the EDD scandal is at this very moment causing real political heartache for one particularly well-coiffured governor.  When polled, those who signed the petition to recall Gavin Newsom from office, one of the top five reasons is the gruesome EDD situation (sending billions to fraudsters while simultaneously driving hundreds of thousands of legitimate claimants to despair and, in many cases, homelessness).

So - what is the difference in the public’s mind?  High speed rail is paying for work – using the term work as loosely as possible - with checks that have vapor trails of zeroes.  The EDD has also wasted billions, but in a different way, a way that ended up literally on kitchen tables across the state.

The key is direct personal impact.  Putting aside the fraud issue, millions of individuals were specifically and negatively impacted by gross incompetence of the EDD.  That meant tens of thousands of calls to elected officials (who were made to look foolish by seeming to be caught completely unaware of the situation), hundreds of press reports, lots of uncomfortable legal questions being asked and the engendering a visceral feeling of ‘boy you people really are incompetent, let’s throw the bums out” in the general public.

In other words, the major difference is that the EDD problem was pubic-facing while the high-speed rail is – for the foreseeable future – not.  Also, the public is so inured to reports of a government infrastructure going over-time and over-budget that the public relations teams handling those projects literally factor that numbness into their communications strategies.

So why would anyone knowingly set themselves up for such a catastrophic potentiality by not ensuring tech systems work at the highest possible level?  Because if it is unseen it is an unproblem and if any errors are made while trying to fix things during good calm times electeds know they will suddenly be on the end of those same constituent phone calls, the targets of those same negative press stories.

For example, let’s look at the experience of another California government agency – the Los Angeles Department of Water and Power (DWP).  Some years ago the agency decided it really needed a new billing system so it hired one of those massive “consulting” firms to tackle the problem.  When the system went live, thousands of the utility bills it produced were wildly incorrect, range from negative “we owe you money” amounts to preposterously large amounts ($8,500 for a month for water and sewage for a studio apartment, that kind of thing). 

Needless to say this caused embarrassment, great public – and therefore political – angst, derailed more than one promising local political career, led to news articles detailing the incompetence and the possibly even worse agency response (more than one person who received a multi-thousand dollar water bill was told by customer service to check to make sure their toilet was running – seriously), led directly to multiple lawsuits and, at least indirectly, to an FBI investigation of the agency as a whole.

So why bother?

Even if done as inexpensively and competently as possible, each and every individual transactional “oopsy” fleetingly caused by the upgrade will matter and there will be thousands of those stories popping up all over the state.  Even though they are supposed to do things like this- make sure the computers work - and absolutely know they need to do things like this, for politicians it is simply not worth the political risk.

A related issue, going back to the speed of technological change, is the chance of being made to look really foolish really quickly.

Let’s imagine a scenario - The decision is made to change/speed up the procurement system a bit and to buy whole new EDD system.  Only the best and brightest will be hired and even the bureaucrats are gung-ho!  And it only takes two years to finish! And it comes in under budget! And there were exactly 37 transactional problems statewide and all of them can be traced back to some guy in a cubicle accidentally spilling a Diet Coke on a spreadsheet, blurring the numbers slated for input!

And then, after spending hundreds of millions on the admittedly great new system, about three weeks after it goes on-line, some teenager invents an app anyone can put on their phone for 8 bucks that does EXACTLY AND EVERYTHING the new systems do.  The electeds – without even having got any credit for upgrading the system in the first place because standing in a room full of servers is not exactly a thrilling photo op - will be pilloried for wasting money, for not knowing about the speed of technological change, etc., and that is politically damaging.

When it comes to tech this exact fear is real for everyone (“Hey that TV I bought last year for $1000 I can get for free now if I sign up for Cox cable? Damn!”).  This fear is incredibly magnified and endemic amongst elected officials, at least in part because very few officials have even the remotest clue about or interest in tech issues.

As a (former) elected official myself I actually had, comparably, a better grasp on the technological innovation process than most.  But way too many others had practically none and even less interest (true story – a fellow elected was angry the staff was asking for money to update the agency website because “they had just done that 5 years ago, for God’s sake!!”).

A somewhat more generous interpretation of the matter is that the problem is something of a Gordian knot.  That, in trying to avoid potential political problems, justify the expense and time, fix existing issues, predict the future of innovation, and digest yards and yards of unfamiliar and complicated information, the brain goes into overdrive, realizes that there is no perfect answer, and lurches to a halt resulting in analysis paralysis.  The “right” answer, no matter how close to perfect it may be, is perceived to be unattainable so the question remains unanswered and the decision unmade. But that may be overthinking it.

Trying to convince the vast majority of elected officials that it is vitally important to upgrade tech systems is like trying to convince your dog to buy real estate.  You can stand there and tell your dog that buying an acre of ocean front land that he can run around and play on for $10,000 is a good idea and a great deal but he still won’t buy it and will just look at you in absolute befuddlement and wonder when you are going to shut up and give him a treat.  And I’ve seen that look on many, many, many elected faces in the past.

And that’s why government agencies can’t do tech.