«Volume Title: Politics and Economics in the Eighties Volume Author/Editor: Alberto Alesina and Geoffrey Carliner, editors Volume Publisher: ...»
Committee members with relatively few weak thrifts in their state (such as F’roxmire of Wisconsin) did not have a compelling interest in raising substantial revenues for FSLIC but were interested in other banking issues. Finally, there was the general unwillingness by many congressmen to find the FHLBB warnings of impending doom to be credible, given the longstanding appearance (buttressed by regulatory accounting practices and the claims of the thrift constituency) that the thrift problem was at worst a temporary and regional one.28 A year of legislative deadlock was not without its legacy. Though no legislation had passed, the regulators’ direct interaction with Wright and other members of Congress from states with many problem thrifts (especially Texas, California, Louisiana, Nebraska, and Oklahoma) made it clear that any 195 Political Foundations of the Thrift Debacle future attempts to deal with the thrift problem would involve large doses of forbearance.
6.4.2 Legislation in the 100th Congress (1987) With the Democrats’ recapture of a majority of Senate seats after the 1986 elections, the chairmanship o the Senate committee passed to Proxmire.
f Within a month of the elections, he announced support for a plan drafted by the administration that would have permitted up to $15 billion in borrowing authority for a FSLIC recapitalization. But he intended to include this in an omnibus bill (S 790) that also placed a moratorium on the creation of new nonbank banks.
The U.S. Savings League lobbied aggressively against the $15 billion recapitali~ation.~~ argued against its costs and supported a much weaker They proposal that would have allowed for a temporary $5 billion funding authority.
This proposal would also have continued and even extended regulatory forbearance toward insolvent thrifts in “economically distressed” areas. In this way, the League proposal combined the forbearance measures sought by the sick thrifts and the limited funding that the healthy thrifts preferred.
As approved by the Senate Banking Committee, S 790 limited FSLIC borrowing to $7.5 billion over two years, with no more than $3.75 billion to be spent in either year.30The bill also included forbearance provisions for thrifts in economically depressed areas. Attempts to increase the borrowing limit were rejected by the committee. Most of the debate in the committee focused on issues relating to the nonbank bank matter and other financial sector regulatory questions, such as the entry of commercial banks into the securities industry.
Bill S 790 was approved on the Senate floor 79 to 1 1 without significant change in the FSLIC financing provisions. An attempt by Garn to strike the non-FSLIC-related matters from the bill to provide a “clean” recapitalization bill failed by a vote of 35 to 54. The committee bill also contained a provision that would have exempted bank regulatory agencies (including the FHLBB) from any automatic Gramm-Rudman-Hollings cuts. Approving an amendment by Gramm (R-Texas) by a voice vote, the Senate stripped this provision from the committee bill. In so doing, the Senate made clear that additional funds for FSLIC would be subject to the general fiscal restraints of the time.
On the House side, St Germain had signaled early on that he would refuse to permit a bill to reach the House floor with the nonbank bank provision, and his committee prepared its own version of the legislation (HR 27). As to the message that the bill would send to regulators, Speaker Wright had met with committee Democrats, and, according to news reports, “Mr. Wright in effect said, no forbearance, no bill” (New York Times, 9 February 1987).
As it appeared in the markup session of the Financial Institutions Supervision, Regulation and Insurance Subcommitteeof the House Banking CommitThomas Romer and Barry R. Weingast tee, HR 27 was essentially the bill supported by the U.S. Savings League.
The major issue of debate was the level of recapitalization. On a 23 to 20 vote, the subcommittee approved an amendment to raise FSLIC’s borrowing limit from $5 billion over two years to $15 billion over five years, with a limit of $3.5 billion on borrowing in any given year. The next day (1 April 1987) the full committee reversed this action, and approved, 25 to 24, an amendment sponsored by Stephen Neal (D-North Carolina) to return to the $5 billion, two-year plan. Amendments to increase the recapitalization to $12 billion over four years or $10.5 billion over three years were rejected. The committee then voted 45 to 5 to report HR 27, with the two-year $5 billion plan, and including many forbearance provisions.
FHLBB chairman Gray attacked the committee bill because, in contrast to the measure the committee had passed in 1986, it included many new forbearance provisions. He noted that these were pushed by “some of the worstmanaged thrifts” in order to “hamstring” regulators. Proponents of HR 27 explicitly argued that its main goal was indeed to put regulators “on a short leash” (Washington Post, 29 April 1987).
6.4.3 The St Germain Amendment As FSLIC’s cash woes mounted, toward the end of April, Speaker Wright made the surprise announcement that he would favor increasing the capitalization to $15 billion, though with the forbearance provisions of the committee bill left intact. Speculation about Wright’s motives included his increasing concern about adverse media coverage of his connection with some spectacular Texas thrift failures and indications that the $5 billion package would fail on the House St Germain joined Wright in this reversal and announced that he would sponsor an amendment to raise the FSLIC borrowing limit to $15 billion-but retain the forbearance provisions-when HR 27 came to a House vote.
The U.S. Savings League lobbied vigorously against St Germain’s proposed amendment. Healthy thrifts were particularly concerned about the additional assessments that the higher borrowing limit would impose on them.
When the St Germain amendment came to a vote in the House (May 5 ), it lost by a resounding 153 to 258.32The leader of the floor fight against the amendment was Neal, who put the issue in the following terms: “The argument is between more money and less oversight [of the regulators by Congress] and less money and more oversight” (Wall Street Journal, 6 May 1987).
We conducted a simple econometric analysis of voting on the St Germain amendment to examine our hypothesis that the opposition of healthy thrifts played an important role in defeating the amendment. We obtained data on the number of FSLIC-insured thrift institutions falling into each of five GAAP capital/asset ratio categories in each state.33 We constructed two variables, WEAK and STRONG. The former is the number of thrifts in the state with negative GAAP net worth, while the latter is the number of those with GAAP net Political Foundations of the Thrift Debacle worth above 3% of GAAP assets. Both variables were measured as of yearend 1986.” We then divided each variable by the number of congressional districts in the state to get an admittedly crude measure of the constituency pressure on each member. (Ideally, we would have liked these data by congressional district, but these were not available to us.) The resulting variables are WEAKDIS and STRONGDIS, respectively.
As another thrift constituency variable, we constructed PACMONEY, which measures the campaign contributions received (in thousands of dollars) by each congressman from the largest thrift industry PACs in the 1983-84 and 1985-86 election cycles.35 To capture purely partisan effects, we used party affiliation (a dummy variable PARTY, which equals 1 for Democrats and 0 for Republicans). Another variable, IDEOL, is intended to capture a congressman’s general legislative preference with respect to government intervention. The variable IDEOL is computed from scaling of the roll calls and legislators in the 100th House using the NOMINATE procedure of Poole and Rosenthal (1991). This variable + ranges from approximately 1.3 to approximately - 1.3, with high values corresponding to “liberal” positions and low ones to “conservative” positions. 36 Including IDEOL allows for within-party variation according to individual constituency or representative characteristics that are not captured in our economic variables. Finally, we allowed for the possibility that members of the committee behaved differently than nonmembers by including COM, which equals 1 for committee members and is 0 otherwise.
We estimated the probability of observing a “yea” vote on the St Germain amendment. Using probit analysis, we estimated P = Prob (yea vote) = F(Z), (1) where F (.) is the standard normal cumulative distribution. Our specification
of Z is given by:
z = Po + P, PARTY -t p, IDEOL + p, COM + P, PACMONEY (2)
-I- P, STRONGDIS + p6WEAKDIS.
Based on our earlier discussions, we expect (3, to be negative. To the extent that “liberal” overall legislative preferences are also “proregulation,” we expect P, 0. Since the St Germain amendment was supported by the administration, Republicans would be more likely to be in favor of it, so we expect P, 0.
The committee had, of course, reported the $5 billion package, but this amount had been approved in committee by a one-vote margin. The committee chairman was now sponsoring an amendment to raise the recapitalization level-a switch from the way he voted in committee. This does not imply any clear prediction about P3,but if committee members tend to support the chairman’s position on the floor (as part of an implicit bargain in a continuing relationship), then we should see p, 0.
Thomas Romer and Barry R. Weingast The predicted sign of p, is negative if WEAKDIS is primarily a measure of constituent pressure from insolvent thrifts. But, as we noted in our earlier discussion, having a larger number of insolvent thrifts in one’s district also heightens depositors’ concerns in the district. This may lead to increasing willingness to vote for a larger recapitali~ation.~’ net effect on p, is unThe clear.
As to p4, it would be natural to suppose that higher contributions from the thrift PACs would be associated with a lower probability to vote for the amendment. But studies of the relationship of campaign contributions to voting on individual roll calls have shown no clear indication of such direct association.
Table 6.4 presents our probit estimates.
As expected, Republicans were more likely to vote for the amendment than Democrats-and liberal Democrats (there are no Republicans with high IDEOL values) were more likely to vote yea than their more conservative brethren. Committee members were more likely to favor the amendment, ceteris paribus, than nonmembers (indeed, committee members voted 30 to 18 in favor3*).Money from PACs does not appear to have had a significant, independent association with voting yea. 39 We see that the coefficient on STRONGDIS is negative and quite precisely estimated in all the specifications. Both the estimated coefficient and its standard error are insensitive to which of the political variables are included in the specification. We can say with some confidence, therefore, that the probability of voting against raising the FSLIC borrowing limit was significantly
Note: Computations are based on estimates in column (1) of table 6.4. (a) For all computations, COM = 0. ( b ) Median IDEOL for Democrats = 0.745. (c) Median IDEOL for Republicans =
-0.456. ( d ) Maximum value of IDEOL = 1.34;minimum value of IDEOL = - 1.41. (e) Col. 1:
Economic variables at their means: PACMONEY = 1.112, STRONGDIS = 5.032, WEAKDIS =
1.087. Col. 2: STRONGDIS = 1 and PACMONEY and WEAKDIS at their means. col. 3: STRONGDIS = 10 and PACMONEY and WEAKDIS at their means. Col. 4: WEAKDIS = 0 and PACMONEY and STRONGDIS at their means. Col. 5: WEAKDIS = 3 and PACMONEY and STRONGDIS at their means.
higher, ceteris paribus, for members from states where the average number of healthy thrifts in a district was higher. The effect of the presence of sick thrifts is somewhat ambiguous; though the estimate of p, is positive, it is much less precisely estimated than p,.
To get a sense of the importance of the thrift constituency variables, we used the specification in column 1 of table 6.4to compute the probability of voting yea on the amendment for two types of hypothetical congressmen not on the House Banking committee.40One is a Democrat whose IDEOL has the median value for all Democrats (0.745); other is a Republican at the methe dian of Republican IDEOL (-0.456).For each of these congressmen, table
6.5shows P,the estimated probability of voting yea, as the value of WEAKDIS or STRONGDIS varies.
Not surprisingly, P is generally below 0.5, reflecting the fact that the amendment lost by a 2l majority among noncommittee members.
As noted :
in column (1) of table 6.5,for our hypothetical Democrat, P = 0.305 when PACMONEY, STRONGDIS, and WEAKDIS are at their sample mean values (for the Republican, P = 0.394).Moving from the mean value of STRONGDIS (= 5) to its lowest value in the sample ( = l), increases the probability of a yea vote by about 33%, to P = 0.414.On the other hand, a change from column (1) to a STRONGDIS = 10 (a change of about two standard deviations from the mean) reduces P by about one-third, to P = 0.192.Changes for the Republican are similar, though less pronounced. The table displays similar calculations for changes in WEAKDIS, which go in the opposite direction.
To summarize, the condition of thrifts in the congressman’s state was clearly related to his legislative preferences in voting on the key amendment in the recapitalization debate. Because the number of healthy thrifts exceeded that of weak thrifts in most states (by more than a factor of two in many cases), the effect of STRONGDIS predominated for most congressmen. The net Thomas Romer and Barry R. Weingast effect, by our estimates, was to reduce the probability of a yea vote in nearly all districts.
There were no comparable roll calls in the Senate. Presuming that similar forces operated in that chamber, our findings are consistent with Proxmire’s willingness effectively to veto legislation in 1986, and with his intransigence throughout the FSLIC debate. In a state with only healthy thrifts, any constituency pressure he faced on the issue would be in the direction of delaying recapitalization or reducing its level.