A. The response to a shortness of available (soft) capital is economically to destroy part of durable (hard) capital. This raises the possibility of a macro-economic "glitch," (a perverse episode of harmful "positive feedback," often called a "vicious spiral.") This effect, variously described and with varying emphases, has been noted by Ricardo, Jevons, BÃ¶hm-Bawerk, Wicksell, Spiethoff, Hayek, and others. Ricardo's Chapter 1, "On Value," and Chapter 31, "On Machinery," are good introductions. They are nominally well-known, and at the same time treated as non-existent: a feat of compartment-mindedness we find in too much economic writing. As Lionel Robbins points out, micro theory after 1870 became one of acapitalistic production.  Capital theory simply disappears from the picture. 
4. The property tax rate on capital items affects their value just as would a rise in the (real) i.r. of the same percentage. A rise in the rate thus destroys existing real capital; a fall in the rate creates real capital.
5. A rise in i.r.s lowers market prices of land by a much larger factor than it lowers prices of existing capital, because the value of land derives from more remote future prospects, overall. Land prices, accordingly, are hypersensitive to i.r.s.  [ "I.r.s" is used here to subsume all the conditions of availability of both loans and equity funds.] Thus land loans are a most undesirable basis for demand deposits. This was recognized by the English Bubble Act of the early 18th Century, and then alternately forgotten and rediscovered with each succeeding episode of land boom and bust.
6. Changes in the market price of land, when caused by inverse changes in i.r.s, do not represent changes in social wealth. In this respect they differ from changes in the market price, or Discounted Cash Flow (DCF), of depreciable capital. Many potentially useful analyses of our subject are deeply flawed by failure to hew to this difference. 
Land prices are also sensitive to changes in expected growth rates of net income, both real and inflationary. These changes, likewise, do not represent changes in social wealth.
The third major factor determining land prices is the current net income (cash or service flow). This may rise for purely distributive causes, e.g. a fall of the interest charge on financing a new building. ( This is separate from the cap rate applied to the net income of land to find the selling price. Land prices are doubly sensitive to the i.r. for this reason alone.) A fall in wage rates may also raise the residual land rent. These changes, again, do not represent changes in social wealth.
Last, the service flow of land may rise because the land actually becomes more productive, e.g. from the spillover benefits of surrounding urban growth. This may represent a rise of real social wealth - I leave the question moot. The main point here is that most changes in land prices do not represent changes of real social wealth.
A. Land is dangerous to use as debt collateral, because its price is so highly sensitive to i.r. changes. It is even more dangerous to let it become the collateral backing demand deposits. 
B. Selective controls on credit extended by commercial banks may be used to prevent collateralizing land values. Another method would be to make mortgages taxable property, as provided for, for example, in the 1879 California Constitution. Such a provision is enforceable because mortgages (or deeds of trust) are always publicly recorded, along with land titles themselves. Such a provision would also ease the political case for raising property taxes, which otherwise fall solely on equity holders, and appear to exempt lenders (except as they erode collateral security).
Why are banks not lending much in early 1993? Interest rates (at least short-term rates) are low, but collateral requirements are very high. There are 3 problems, at least. 1, Banks are leery of any real estate collateral now. 2, They lack the needed capital. (They also may lack reserves). Both of those result from their recent losses. 3, Real interest rates are higher than they look when we factor falling land prices into the c.o.l index used to deflate nominal i.r.s into real i.r.s. This is a variation on Keynes' perception of a liquidity trap. By "variation" I mean it is the same phenomenon, only differently perceived and expressed.
Here is the sequence of a combined, reverberating land and banking crash. Land boom fizzles. Banks take losses. Their reserves and surpluses (capital) dwindle. They stop making loans and investments. By a process of positive feedback ("vicious spiral") this stoppage aggravates its own cause, viz. the fall of land prices.
7. A rise of i.r.s tends to raise savings rates via a strong wealth (or portfolio) effect. It lowers the current market price of land, especially. To a lesser extent it lowers the prices of items of durable capital.
There is a diminishing marginal utility of total wealth held (for retirement, for business use, for consumer capital, etc.). The fall of asset prices as a store of value thus tends to raise savings rates.
At the same time, a rise of the Marginal Rate of Return (MROR) on new investing raises the reward of saving as vs. consuming income. This is a substitution effect, conceded by all. The traditional counter-argument has been that there is a countervailing income effect: higher income from given sums invested tends to weaken the impulse to save. This counter-argument in turn, however, is offset and more than outweighed by the wealth effect recited above. The wealth effect reinforces the substitution effect, making saving respond positively to i.r. hikes.