Regime Shifts and Expectations - A Move Towards Fragility?

One of the foundations of Market Monetarism is the role of expectations in influencing policy.  The whole goal of a 5% nominal GDP target is so that the economy can expect the same level of nominal growth, thereby allowing the highest efficiency in an economy, as agents can plan for a stable future.  Such a view was recently summed up by Scott Summers in a post calling for a better Monetary REGIME instead better Monetary POLICY:

... If there is some sort of policy that you think needs fixing via monetary policy, then you have the wrong monetary policy regime. You are targeting the wrong variable. Thus you might (wrongly) decide it’s a good idea to target headline inflation at 2%, and then suddenly notice that that target conflicts with your gut instinct that unemployment is too high because of inadequate AD. In that case the right decision is not to pull out the monetary policy tool, but rather to entirely abandon your inflation targeting regime. If you have the right regime, then YOU SHOULD NEVER, EVER, ADJUST MONETARY POLICY.

This concept of a regime makes me think along the lines of the possible impact of large shocks to the macroeconomic system; what if the regime shifts?  While a previous regime shift, such as FDR's decision to depreciate the dollar in the middle of the Great Depression, may have worked to restore the economy, the question is whether a new NGDP targeting regime would be stable.  From a differential equations standpoint, would it be a sink or a source?  The centrality of expectations to the theory seems to suggest that if the expectations were to become unanchored at some point, chaos would ensue.

However, what mechanisms could explain the shift to a more fragile system?  Besides artful comparisons to forest fires or animals' need for exercise, are there actual theoretically justified reasons for an increase in fragility?

Searching and the Role of the Efficient Market Hypothesis
Overall, the EMH is pretty effective at explaining financial markets; it's unlikely that there's some systematic error that causes financial mangers to consistently miss profit making opportunities.  However, some assumptions of the EMH have to be relaxed to make a NGDP monetary regime make sense.  Primarily, information flows are not instantaneous, which means markets do have to go through an adjustment process while economic agents search for the information to form their expectations.

However, what's the sense in searching in a world of stable NGDP growth?  An alternative to the EMH, the Adaptive Market Hypothesis, posits that survival of firms, instead of pure profit maximization, is a stronger driver of moves in the market.  In an aggregate market in which movements are generally stable, survival is not as severe of an issue.  While NGDP stability in theory should have no effect on intra-industry competition and churn, common sense experience seems to suggest that when times are tough, people look for more ways to become competitive to survive.  However, when times are stable, people choose not to use that psychic energy to optimize; they settle with just satisficing.  Admittedly, this requires a more behavioral approach, but considering the vast literature on cognitive biases that reject homo econimus, it seems reasonable.

Impact on Debt
The issue of debt seems to be problematic in a world of constant nominal GDP growth as well.  Such stable expectations allow greater and greater tolerance for debt, as people become more comfortable with the concept of constant growth and start to neglect the possibility of nominal shocks.  This would lead to vastly longer debt chains, which can, in effect, make the economy more fragile and vulnerable to systemic shocks.  While the hot potato effect does not necessarily depend on a function banking system, the flow of money and the repaying of debts can have a significant effect on short term NGDP expectations.  Even if there's a short term lag on repayments, the danger of debt chains breaking down seems like a major issue.  While financial problems really are more supply side issues, how would the government deal with them once the crisis hit?

Conclusions
In the end, my fears, however irrational, are predicated off of a very true, yet also startling quote from one of Nicolas Nassim Taleb's journal articles:

...Because a probability cannot be lower than 0, your expected probability should be higher, at least higher than the expected error rate in the computation of such probability.  Model error increases small probabilities in a disproportionate way...

How does NGDP targeting prepare for this?  If it can not adequately prepare, is there a middle compromise that allows sustainable growth with robustness to error?