Prime Minister's suggestion to limit borrowers' liability to their collateral (i.e. real estate) is yesterday's news now. Besides, it was likely a mere negotiating move (as shrewdly pointed by Janis Ikstens), i.e. Prime Minister's Dombrovskis response to the "aggressive" remarks made by Anders Borg, Swedish minister of finance. And yet, another purpose of this proposal is, probably, to test public opinion on such things. So here is my opinion as well. In short, it's hard to be excited about such an initiative - which I will refer to it as 'limiting liability'.
Should we start sending keys to the banks? 12
First, lets start with the simplest observation. If such a decision would apply to all existing contracts (which seems to be the idea), we're really talking about government-authorized default on the loan obligations. From the banks viewpoint, this implies an increase in country-specific risk. They wouldn't turn around and leave (banks are not children), but they would factor that risk into their future decisions. This means higher interest rates for future borrowers and, as a result, less credit. As businesses would find it more expensive to raise capital, this would have a negative effect on future economic growth.
Second, Prime Minister Dombrovskis said such a move would make banks more cautious and future crises less likely. I just don't see what systemic effect this could have on banks' appetite for risk taking. However, limiting liability to the collateral would have the effect of offering borrowers an investment opportunity in which the gains on the upside belong to them, but the losses on the downside are limited to the collateral. This would provide an incentive to invest in real estate for speculative motives. A problem of adverse selection would materialize, in which banks would have to deal with an increased number of high-risk borrowers. Banks would respond by increasing the share of down payment and also by credit rationing, i.e. reducing credit funds available. Both would make it harder for regular people (and firms) to obtain loans.
If the above paragraph was somewhat technical, consider this. It was always possible to borrow against collateral only. All one had to do is to establish a limited liability company (LLC) and have it ask the bank for a loan backed by real estate. This would limit liability to whatever is on the balance of this LLC, i.e. collateral. Anyone would tell you that banks would charge a higher interest rate and demand a greater down payment from an LLC. Naturally, individuals can borrow on much better terms than LLCs precisely because of unlimited liability! What the proposals amounts to is legislating LLC-style borrowing only, which would make it much harder (and more expensive!) for individuals to borrow.
Third, a somewhat more complicated argument. But lets start with acknowledging that limiting liability may have some very positive effects in the context of this crisis. If a large number of people could walk away from the mortgages they cannot afford and start a new life, this would go some way to breaking out of the recessionary spiral. But lets think about the consequences of this. An individual would "send the keys to the bank" (i.e. default) if
total liability - market value of collateral > present value of expected costs of bad credit history
Naturally, individuals who default on their loans would find it difficult to convince the banks of their credit worthiness in the future. So the number of defaults would be smaller than the number of people for whom (liability>market value of collateral). But still, given extremely low rental costs, it is likely there would be many such people. The banks would see a sharp increase in their non-performing loans (NPLs) and have lots of real estate on their assets. Naturally, banks would rush to raise liquidity by selling all this real estate. Such a fire-sale would reduce real estate prices even further, shrinking banks' balance sheets and increasing pressure to raise liquidity. We have seen something very similar in the U.S. during this crisis. As a result, you're looking at either another full-blown financial crisis - with (even more) disastrous consequences for the real sector, or massive government bail out of the banks. Since the latter is more likely, we are really talking about Latvian (and Swedish) taxpayers paying for this policy initiative.
If you accept the above prediction - in the end, the cost of "limited liability" is largely borne by the taxpayers - lets make a shortcut and ask whether taxpayers should offer borrowers the following deal: "We would pay the remainder of your loan if you hand over the collateral (real estate) to us". The benefit (to the taxpayers) is that this may end the crisis faster by resolving the problem of bad loans. What about the costs? There are two. First, the prudent (who didn't borrow, or borrowed but can pay) pay to the reckless (who borrowed more than they could afford). Many of the reckless are probably those who did it for investment purposes, i.e. 'speculated'. By most ethical standards, such a 'socialization of losses' is grossly not fair. Second, you introduce the problem of moral hazard. Such a precedent would mean more excessive risk taking in the future because of the perception that the "gains are privatized, but losses are socialized". This would make future crises more likely.
To sum up, this seems like a very bad idea now (during the recession) and, probably, a bad idea for the future (after the recession). I think that blaming (and punishing) the banks would be a great temptation to all kind of populist politicians in the run-up to the elections. However, this might be equivalent to shooting yourself in the foot. Finally, it is likely that we would need some taxpayer-financed solution to the burden of bad debt. However, such a solution should not let any participant of the 'credit orgy' to emerge unscathed. This also applies to the borrowers.