The CIOs of most large enterprises can sympathize with New York’s Metropolitan Transit Authority. As described last week, the MTA runs its entire, vast, complex subway system on 1930s technology. Compared to this, the 1970s-vintage boat anchor IMS DBMS so many CIOs can’t get rid of because mission-critical applications run on it seems positively modern and benign by comparison.

As pointed out last week, when depreciation was first incorporated into the world’s Generally Accepted Accounting Principles (GAAP), it was supposed to represent how a capital asset declined in value over time. Were business executives to take depreciation seriously, they’d look at their balance sheets, see that the company’s asset value is declining, and do something about it.

Well, no. It isn’t that simple.

First and foremost, one of the key metrics Wall Street Analysts rely on is Return on Assets (ROA). As most 5th graders can tell you, there are two ways to improve ROA: Increase returns, or decrease assets.

It’s metrics at its finest: decrease asset value as reported on the balance sheet and voila! Without any change in actual competitiveness the company’s financial performance magically improves.

Now imagine you’re in charge, and you have a relentless focus on long-term sustainability instead. What would you do differently?

High on your list might be squirreling away the funds needed to replace obsolete business assets as they become increasingly valueless, ROA be damned.

Nice thought. One minor gotcha: Time, trends, and evolving marketplaces have made some of those assets irrelevant. Or, worse, they’re part of a business infrastructure that impedes your ability to redirect the company in a direction that’s more in harmony with changing competitive demands.

So it isn’t as simple as replacing assets as they become obsolete.

In IT it’s even less simple, because assets that aren’t the least bit obsolete … well-engineered business applications that support important segments of the business quite well … can still require platforms that for one reason or another are marketplace failures.

Like it or not, and it’s usually not, your choices are limited: Redevelop the offending application on and for platforms that seem to have some staying power, or replace it with a commercial package that will do the job and seems to have some staying power.

It’s Hobson’s Choice: Rewrite or convert.

For this case, the underlying principle seems to hold: The business needs to keep money in the bank for use replacing applications and their underlying infrastructure when one, the other, or both are reaching their limits of viability.

It’s downright strange, isn’t it? While you have to replace a perfectly serviceable application because it or the platforms it uses aren’t making it in the marketplace, your buddy down the road has to replace a major application … an ERP suite, perhaps … that’s a major marketplace player, only it’s become a liability because it can’t be configured to effectively support the company’s future business strategies and tactics.

Or, another friend finds herself retiring one application because it supports a business her company is no longer in, while also having to implement an entirely different piece of software that’s needed for the business it’s going to be in.

What do all these permutations have in common?

In every case, dealing with the situation will take time, effort, expertise, and money. It will take them because of a fact known only to a small number of highly insightful individuals who belong to a secret society called, with apologies to Drew Carey, Everyone.

That fact: The future is generally different from the present. More, it’s different in ways that are only predictable in the million-monkeys sense that if enough people make predictions, the odds are pretty good that through nothing more than random chance, a few of them will end up guessing right.

Its future being different from its past, the tools a company will need to deal with the future will be different from those it needed in the past, too, not to mention that the tools available to it will be different from those that were available in the past.

Which gets us to this: Running a sustainable business calls for constant reinvestment, and when it doesn’t that just means the investment that isn’t needed this year will be needed in spades next year or the year after.

Which is yet another reason stock buy-backs are such a terrible idea: They take cash that could be invested in sustainability and get rid of it in order to prop up the price of a share of stock. So when the company finds it needs the money after all, it only has one choice left to it:

Issue a bunch of new shares of stock.