Grigg v. Robinson Furniture Co.

260 N.W.2d 898, 78 Mich. App. 712, 1977 Mich. App. LEXIS 1244

• Court: Michigan Court of Appeals
• Decided: October 10, 1977
• Docket: Docket 22374, 22375, 22376, 22377, 22378, 22379
• Status: Published

Opinion

Allen, P. J.

I. Factual Background.

These six consolidated class actions are de facto companions of (Sarah) Grigg v Michigan National Bank, 72 Mich App 358; 249 NW2d 701 (1976). Richard Grigg, one of the named plaintiffs in the present series of cases, is the husband of the plaintiff in Grigg v Michigan National Bank, supra. Mrs. Grigg is employed by the attorneys who represent the plaintiffs in the six cases now before us.

Grigg v Michigan National Bank, supra, sought recovery of allegedly excess finance charges 1 paid by users of the Michigan BankAmericard ■ service operated by the defendant bank. This Court affirmed the trial court’s dismissal of the class action in Grigg v Michigan National Bank, supra, primarily because use of the GCR 1963, 208, class action device would have imposed too great a burden upon the defendant and the judicial system as compared with thé benefits which might have been conferred upon members of the plaintiff class.

The trial judge in the present case deferred a ruling on the propriety of a class action format in order to first address the substantive questions raised by the complaints. We have serious doubts about the wisdom of that choice. If we were operating under Federal Rules of Civil Procedure (FR Civ P 23), the trial judge’s failure to consider the class action question would be clear reversible error. Eisen v Carlisle & Jacquelin, 417 US 156; 94 S Ct 2140; 40 L Ed 2d 732 (1974). Our own Rule 208, supra, was copied directly from an earlier version of FR Civ P 23, supra. But Michigan has not yet adopted the subsequent amendment to the Federal rule which Eisen interpreted as mandating that the propriety of the class action be considered before the merits in all cases. That failing at least left the door ajar for the procedure followed by the trial judge in the present cases. Northview Construction Co v St Clair Shores, 395 Mich 497; 236 NW2d 396 (1975), reversing 44 Mich App 614; 205 NW2d 895 (1973), reh granted 395 Mich 924 (1976), aff'd on reh, 399 Mich 184; 249 NW2d 290 (1976). We recognized precisely such a possibility in Grigg v Michigan National Bank, supra.

"We hold that shortly after a class action is filed, and after discovery, a trial court should usually first determine whether a class action can be maintained before any findings on the merits. * * * There may, however, be instances where considerations of judicial economy will dictate that at least a partial ruling on the merits should precede certification of the class.” (Emphasis supplied.) 72 Mich App at 381-382.

But, even if the. opportunity theoretically exists, we believe, for reasons more fully outlined in Eisen, Grigg v Michigan National Bank, supra, and Kass v H B Shaine & Co, Inc, 71 Mich App 101; 246 NW2d 396 (1976), that the trial judge abused his discretion by deferring the class action question in this case. However, in all fairness to plaintiffs we recognize that we have been presented with something of a fait accompli. Given the present posture of the case, we believe that the interests of justice will be better served if we address the merits of the case and resist the inclination to summarily reverse.

The defendants in the present cases are all large retail stores which make credit sales pursuant to revolving charge account agreements. The plaintiffs are seeking recovery of allegedly excess finance charges paid pursuant to those agreements since August 4, 1964. The plaintiffs contend that a monetary recovery is possible under MCLA 445.868; MSA 19.416(118) or one of several equitable causes of action. The trial judge appears to have recognized a combined statutory and equitable cause of action. We affirm that ruling — at least to the extent of authorizing recovery of excess finance charges. The August 4, 1964, date was apparently chosen in recognition of the six-year limitation period imposed by MCLA 600.5813; MSA 27A.5813. The complaints in these cases were filed in August and September of 1970.

Ignoring minor variations, there are four generally recognized methods of computing finance charges on revolving charge accounts. Those methods are ending balance, 2 previous balance, 3 average daily balance, 4 and adjusted balance. 5

The defendants were using the previous balance method on August 4, 1964. At that time, there was no statute speciñcally regulating the amount of finance charges which could be imposed on revolving charge accounts. Michigan’s principal usury law, 1948 CL 438.51 (since repealed and replaced by MCLA 438.31; MSA 19.15[1]) provided that interest rates on loans of money could not exceed 5% for oral agreements, and 7% for written agreements; but the traditional rule has been that the usury laws controlled only "interest” and not "time-price differentials”,* ***5 6 the label applied to finance charges on retail credit sales. Keefe v Bush & Lane Piano Co, 247 Mich 82; 225 NW 585 (1929), Silver v International Paper Co, 35 Mich App 469; 192 NW2d 535 (1971). Since the revolving charge account is a comparatively recent innovation, no Michigan court has as yet expressly decided whether that traditional immunity from the usury laws extends to the time-price differentials charged on revolving accounts. Plaintiffs in these cases argue that the immunity is not so extensive and that on August 4, 1964 — and for some time before that — the defendants’ use of the previous balance method resulted in violations of the usury laws.

The next significant event occurred on March 10, 1967, the effective date of Michigan’s Retail Installment Sales Act (MRISA), 1966 PA 224; MCLA 445.851 et seq.; MSA 19.416(101) et seq. Unlike the old usury laws, MRISA was expressly designed to regulate finance charge rates on revolving charge accounts.

Section 12 of MRISA, MCLA 445.862; MSA 19.416(112), limited time-price differentials to "1.7% of the unpaid balance per month”. Plaintiffs in these lawsuits allege that beginning with the effective date of MRISA on March 10, 1967, the defendants were in violation of that limitation by virtue of their use of the previous balance method.

On August 11, 1970, in response to a legislative inquiry, the Attorney General released an opinion concluding that the previous balance method as then employed by the defendants and other retailers violated the maximum rate limitation in § 12 of MRISA. OAG 1969-1970, No. 4706, p 163 (August 11, 1970). The release of that opinion coincided with the filing of the first of the complaints in these consolidated actions. The Attorney General’s opinion also sparked a lawsuit in Ingham County Circuit Court by some of these defendants challenging the correctness of that opinion. Sears, Roebuck & Co v Attorney General, Ingham County Circuit Court No. 12204-C. That lawsuit was finally ended by a stipulation which (1) reached no conclusion on the legality of the previous balance method, (2) stated that the defendants’ use of that method had been in good faith, but that (3) the defendants would switch to the average daily balance method with a 30-day deferral feature as explained in fn 4, supra. The defendants switched from the previous balance method to the average daily balance method on various dates in 1971. Despite that switch, the plaintiffs contend that even the new method violates § 12 of MRISA because the defendants compound finance charges, i.e., if a finance charge for one month is not paid, it is added to the debit balance on which the next finance charge is calculated. The plaintiffs also contend that the defendants should not be allowed to include purchases made during the current month in their computations of daily balances.

The plaintiffs have consistently characterized these proceedings as equitable in nature, in part because of their requests for an injunction and an accounting, and also because of the analysis by Justice Williams in Paley v Coca Cola Co, 389 Mich 583; 209 NW2d 232 (1973), which traced the equitable origins of class actions. The defendants have also invoked equity jurisdiction. They argue that, even if they are found to be in technical violation of the law, various equitable principles preclude any monetary award. We accept the invocation of both equity jurisdiction and equity principles. Paley v Coca Cola, supra; GCR 1963, 12, 110.3, 203. At least since 1963, equitable and legal claims may be joined in a common complaint and equitable defenses can defeat legal claims.

After extensive preliminary maneuvering, both sides moved for summary judgment in the trial court. The trial judge granted summary judgment for the defendants on the first and third of the plaintiffs’ three theories of recovery; and he granted summary judgment for the plaintiffs on their challenge to the previous balance method. First, he ruled that the finance charges assessed prior to March 10, 1967, were not illegal because they were time-price differentials, not interest, and thus not subject to the limitations of the usury statute. Next, for purposes of evaluating the MRISA-based challenges to the previous balance method, the trial judge interpreted § 12 of MRISA as allowing an "effective yield” 7 of 1.7% per month. He then tested both the previous balance method and the version of the average daily balance method adopted pursuant to the stipulation with the Attorney General by asking the following question: Assuming the worst possible timing of purchases and payments, is it ever possible that the method may produce effective yields in excess of 1.7% per month?

Applying the stated test, the trial judge found that the average daily balance method currently being used by the defendants was legal. However, he also held that the previous balance method was in violation of § 12 of MRISA from the March 10, 1967, effective date of that act until the defendants abandoned use of the previous balance method at various times in 1971. The judgment order provides that the plaintiffs may recover "damages” to the extent that the finance charges actually imposed on their accounts during that period exceeded the amount which would have been imposed if the defendants had been using the present average daily balance method. Since most of the defendants used a 1.5% rate in their average daily balance method computations, it is obviously possible that the damage formula might require them to return finance charges imposed in some months when their charges were legal even according to the trial judge’s effective yield test which used a 1.7% rate. We note the apparent incongruity; however, we believe that, in principle, the trial judge was acting within his equitable powers when he fashioned that remedy.

Neither the propriety of using the class action procedure nor the actual amount of damages has yet been determined. Therefore, the summary judgments granted by the trial court were necessarily partial summary judgments from which there would ordinarily be no immediate appeal. However, this Court granted the defendant’s application for leave to file an interlocutory appeal. GCR 1963, 806.2(2).

II. Issues Involved.

The combined appeals present at least 10 stated questions which we have consolidated into 4 as follows:

(1) Did the previous balance method of computing finance charges used by the defendants between August 4, 1964, and March 10, 1967, violate Michigan’s usury laws?

(2) Does the average daily method of computing finance charges currently being used by most of the defendants pursuant to the stipulation with the Attorney General violate MRISA?

(3) Did the previous balance method used by the defendants between March 10, 1967, and late 1971, violate MRISA?

(4) If violations occurred, what amount of damages should be awarded?

Question #1: Pre-1967 Usury Laws.

In answer to the first question, we affirm the trial court ruling that there has been no usury law violation because prior to the effective date of MRISA, there was no state law limiting the amount of time-price differential which might be charged for the privilege of paying for retail merchandise in one or more delayed installments. We agree with the plaintiffs’ statement that there is no real difference between time-price differentials charged on retail credit sales, and interest charges on loans of money. But this state’s policy has been to recognize a technical distinction in order to encourage the sale of retail goods on credit while simultaneously placing strict usury limits on outright loans of money. The trial judge correctly cited Keefe v Bush & Lane Piano Co, supra, and Silver v International Paper Co, supra, in support of his ruling. It is true that neither of those cases involved revolving charge accounts. But arguments raised by the plaintiffs would have been just as logically sound if made in either of those two cases. Thus, we must reject their proposed distinction.

Similarly, we cannot criticize the logic of the foreign jurisdiction cases cited by the plaintiffs, e.g., State v J C Penney Co, 48 Wis 2d 125; 179 NW2d 641 (1970). But the fact remains that Michigan has always allowed sellers to charge higher finance charges than lenders. Plaintiffs argue that the trend is in the other direction. However, that contention is rebutted by the fact that, simultaneously with the enactment of MRISA, the Legislature amended the general usury statute to clearly exclude time-price differentials from its lower rate limitations. 1966 PA 326; MCLA 438.31; MSA 19.15(1). There can be no doubt that the Legisla ture intended to continue observing the distinction between interest charged on loans of money and time-price differentials charged on retail sales.

For the reasons stated, we hold that the defendants’ use of the previous balance method of computing finance charges between August 4, 1964, and March 10, 1967, did not violate the then effective usury laws.

Question #2: Average Daily Balance Method.

The second question presented is whether the. average daily balance method currently being used by the defendants pursuant to the stipulation with the Attorney General violates § 12(b)-12(c) of MRISA. 8 Specifically, the plaintiffs contend that the statute prohibits including current month purchases in the daily balance computations and prohibits compounding finance charges by adding unpaid finance charges to the balance upon which the next finance charge is calculated.

We hold that the defendants may include current month purchases in their computations of the average daily balance. Including of current purchases is expressly permitted by the Attorney General stipulation and it is not prohibited by § 12 of MRISA. Section 12(c) authorizes imposition of a time-price differential "for the privilege of paying in installments”. The crux of the plaintiffs’ argument is their belief that a customer does not exercise the privilege of paying in installments until he has received the first bill reflecting the current purchases and has failed to make a payment on the bill before the closing date of the next statement. We disagree. We hold that the phrase "paying in installments” refers to any transaction where full payment does not occur simultaneously with the transfer of the purchased goods. If a customer takes possession of the goods without paying, he has received a significant benefit for which the seller is expressly authorized to impose an additional charge.

Because the argument against compounding finance charges is directed against all methods used from 1964 through the present time, it must be analyzed in stages.

The Michigan cases cited by the plaintiffs for the proposition that finance charges may not be compounded were all decided at or before the turn of the century. See e.g., Hoyle v Page, 41 Mich 533; 2 NW 665 (1879). But the age of the authority is not the problem. Rather, the important point is that all of the cases deal with compounding of interest while, for reasons already stated, the finance charges imposed by these defendants are classified as time-price differentials. We again note the inherent logic of the plaintiffs’ argument against distinguishing those terms. However, at least in Michigan, a very real distinction exists. At least before March 10, 1967, there was no prohibition against compounding time-price differentials in this state.

The next question is whether compounding of time-price differentials is now prohibited by MRISA. Section 12(c), MCLA 445.862(c); MSA 19.416(112)(c), authorizes imposition of a time-price differential on "all amounts unpaid thereunder”. We read that language as permitting compounding of time-price differentials. We note that this will only occur if the customer breaches the financing contract by failing to pay at least the required monthly minimum.

Question #3: Previous Balance Method.

The third question posed by this appeal is whether the defendants’ use of the previous balance method between March 10, 1967, and late 1971 was in violation of MRISA. The trial judge ruled in the plaintiffs’ favor on this point. We are reluctant to express an opinion on this question because we are well satisfied with the terms of the stipulation signed by the Attorney General and several of the defendants. Were we seeking only to determine the fairest system for all parties concerned, the true average daily balance method of computation would be our choice. 9 Because of its finance charge deferral and avoidance features, the hybrid method required by the Attorney General’s stipulation is even more favorable to customers. Our resolution of the next issue makes it unnecessary at this time to disturb the stipulated arrangement by expressing an opinion on the validity of the previous balance method under MRISA.

Question #4: The Damages Award.

The final and ultimately determinative question is whether — assuming that the previous balance method violates MRISA — the trial judge nevertheless erred by finding the defendants liable for damages. We conclude that several principles preclude a damage award at this time. Our conclusion is motivated by many of the same factors which led this Court to affirm the trial court dismissal of the class action in Grigg v Michigan National Bank, supra. We suspect that the present actions would have been dismissed for the same reasons if the trial judge had not deferred considering the propriety of these class actions.

The present record does not permit precise computations of the amount of allegedly excess finance charges involved or the cost of computing and paying damages. It seems likely that although the total amount of finance charges was huge, the amount of allegedly excess charges collected was quite small even as measured by the effective yield test designed by the trial judge. 10

The trial judge found — and the Attorney. General stipulated — that the defendants’ use of the previous balance method was done in the good faith belief that it was legal in Michigan. We affirm that finding. This is not a situation like Hays v Regents of the University of Michigan, 393 Mich 756; 233 NW2d 276 (1974), in which the defendant should have been alerted to the clear illegality of its practices by earlier court decisions. In the present case, the most that can be said is that there is a split of opinion among the state and Federal courts which have considered the legality of the previous balance method under statutes similar to MRISA. Contrast Haas v Pittsburgh National Bank, 526 F2d 1083 (CA 3, 1975), and Partain v First National Bank of Montgomery, 467 F2d 167 (CA 5, 1972), with Zachary v R H Macy & Co, 31 NY2d 443; 340 NYS2d 908; 293 NE2d 80 (1972), Seibert v Sears, Roebuck & Co, 45 Cal App 3d 1; 120 Cal Rptr 233 (1975), Federated Department Stores, Inc v Pasco, 275 So 2d 46 (Fla App, 1973), Johnson v Sears, Roebuck & Co, 14 Ill App 3d 838; 303 NE2d 627 (1973).

As previously stated, we are dealing, at most, with excess charges collected during 1967-1971.* 11 Those charges may amount to a significant sum in the aggregate, but they are insignificant when viewed as individual claims. Unfortunately, charges cannot be computed and paid in the aggregate. 12 If damages are to be awarded, it will be necessary to review the activity in each individual account during the 1967-1971 period in order to determine the amount of excess charges. It seems likely that the total cost of calculating excess charges and paying individual claims would greatly exceed the amount of money actually paid to the individual plaintiffs. Since the class action certification question has been deferred, we have no access to the data which would allow precise computations. We are unwilling to blindly affirm a damage award without at least knowing the approximate cost of such an award. For that reason, we must assume the worst until contrary data is presented — preferably in a motion to certify the class and notify potential class members.

Plaintiffs argue that — whatever the cost — it is equitable to require the defendants to bear that cost as a punishment for their allegedly illegal past conduct. That argument is superficially sound; but it fails to trace those costs far enough to see who will ultimately pay them. The answer to that inquiry seems all too clear. Some day, in some way, the costs of calculating and paying damages will be borne by the defendants’ customers, i.e., the individual plaintiffs. 13 And, given the grossly inefficient transaction costs involved in calculating the individual damages, the future costs to the customers will greatly exceed their damage recovery.

III. Conclusion.

All concerned concede that this action is subject to the rules of equity. Paley v Coca Cola Co, supra, GCR 1963, 12, supra. The merger of law and equity and the adoption of the clearly erroneous review standard in GCR 1963, 517.1, have not changed the rule that equity rulings by trial courts are reviewed de novo by this Court and the Supreme Court. For the reasons already stated, we are convinced that it would not be equitable to award damages in these actions at this time.

It might justifiably be argued that the result we have reached destroys the effectiveness of the class action device in precisely the type of case for which class actions were designed. The traditional wisdom is that class actions were created in order to allow recoveries for wrongs in situations where the individual recoveries would be so small that no one person could afford to bring a lawsuit. The lesson we draw from these cases and from Grigg v Michigan National Bank, supra, is that the problem has not yet been completely solved. Where comparatively larger individual claims or comparatively fewer plaintiffs are involved, class actions function effectively as intended. But when it appears that the plaintiffs will not be the present or future 14 beneficiaries of their action, the class action device is being abused. We, therefore, exercise our equitable powers to prevent that abuse.

If the plaintiffs wish to continue this lawsuit, we believe that the best approach would be to resolve the class action certification question first. The principles discussed in Grigg v Michigan National Bank, supra, will require the presentation of evi dence concerning the potential size and administrative costs of possible damage awards. That evidence will necessarily either confirm or refute the assumptions which underly our refusal to sanction a damage award at this time.

Affirmed in part, reversed in part.

1 We will use the term "finance charges” in the generic sense throughout this opinion. So used, the term encompasses several types of charges including "interest”, "time-price differential”, "service charge”, "delinquency charge” and others. The more precise terms are not always fungible. To avoid confusion, we will always speak of "finance charges” unless a choice between two more precise definitions is required.

2 In the ending balance method, the finance charge is assessed against the debit balance in a customer’s account on a designated closing day for the monthly statement. The finance charge is assessed in full even though the debit balance may have been created by a purchase as late as the statement closing date. The finance charge is then added to the debit balance and the total is reflected in the monthly statement as the amount due from the customer. If a company uses an ending balance method in its pure form, the customer cannot avoid paying a finance charge except to the extent that he makes payments on his account before the statement closing date. That means that the customer must pay for the purchases before he receives a bill if he wants to avoid finance charges.

3 The previous balance method is to some extent a modification of the ending balance method. The defendants say that it is the ending balance method with the addition of a 30-day grace period of potential free credit. The customer can avoid all finance charges by paying the previous month’s ending balance in full within 30 days after receiving a statement which includes that balance. The previous balance method imposes a finance charge at the end of month #2 on the ending balance from month # 1 unless a customer pays that balance in full at some time during month #2.

4 A true average daily balance method would assess a finance charge on the amount determined by averaging the debit balances in an account at the end of each day during the month covered by a statement. The daily balance changes from day to day to reflect purchases which increase the debit balance and payments which reduce that balance. Assuming a 30-day month, the average daily balance would be determined by summing the balance at the end of each of the 30 days and dividing by 30. The particular variant of the average daily balance method involved in this litigation includes a 30 day deferral feature like that which distinguishes the previous balance method from the ending balance method. In other words, a finance charge is calculated on the average daily balance at the end of month # 1, but the finance charge is not added to the debit balance in the account unless the customer fails to pay, before the end of month #2, an amount equal to the ending balance (not the average daily balance) from month # 1.

5 The adjusted balance method is a further modification of the previous balance method. Its attraction for the customer is that it allows him to avoid a part of the finance charge by making a partial payment even though his resources do not allow him to pay the previous balance in full. This is accomplished by reducing the previous balance by any payments made during month #2 (without adding purchases) before calculating the finance charge.

6 See in 1, supra.

7 The defendants argue that the 1.7% figure specified by the statute is a ''nominal” rate rather than an "effective yield” rate. See fn 10, infra. If a customer made a purchase on the last day of month #1 and paid almost but not all of his debit balance on the first day of month #2, imposition of a finance charge according to the previous balance method at the nominal 1.5% rate used by most of the defendants would result in an effective yield of approximately 45% per month as contrasted with the 1.5% per month nominal rate. (Debt owed for one day; 1.5% charge assessed; multiplied by 30 days in the month; monthly rate is 45%.) If the purchase were made earlier in month #1, the effective yield would decrease. If the purchase were made on the first day of month #1, the effective yield and the nominal rate would be identical. If the customer waited until the end of month #2 before making the partial payment, the effective yield would be .075%. If he made a full payment at any time in month #2, the effective yield would be 0%.

8 "(b) The buyer under the retail charge agreement shall promptly be supplied with a statement as of the end of each monthly period, which need not be a calendar month, or other regular period agreed upon in writing, at the end of which there is any unpaid balance thereunder, which statement shall recite the following:

"(1) The unpaid balance under the retail charge agreement at the beginning and at the end of the period.

"(2) The cash sale price of each purchase by the buyer during the period and, unless a sales slip or memorandum of each purchase is attached to the statement, the purchase or posting date, a brief description or identification of each purchase.

"(3) The payments made by the buyer and any other credits to the buyer during the period.

"(4) The amount, if any, of any time price differential for such period.

"(5) A statement to the effect that the buyer at any time may pay his total unpaid balance or any part thereof.

"(c) A retail charge agreement may provide for, and the seller or holder may then, notwithstanding the provisions of any other law, charge, collect and receive, a time price differential for the privilege of paying in installments thereunder, in an amount not exceeding 1.7% of the unpaid balance per month. The time price differential under this subsection shall be computed on all amounts unpaid thereunder from month to month, which need not be calendar months, or other regular periods; but if the regular period is other than a monthly period or if the unpaid amount is less than or greater than $10.00, the time price differential may be computed proportionately. The time price differential may be computed for all unpaid balances within a range of not in excess of $10.00 on the basis of the median amount within such range if as so computed such time price differential is applied to all unpaid balances within such range. A minimum time price differential not in excess of 70 cents per month may be charged, received and collected.”

9 See fn 4, supra.

A customer who obtains possession of goods without being required to pay for them immediately has, practically speaking, either borrowed money from the seller or been allowed to retain his own money for some additional period of time. We believe that it is fair to impose a charge for that privilege and that the charge should be directly and proportionately related to both the amount of money involved and the length of time for which it is held. Once a method of computation is selected, it is a simple task to regulate the amount of the finance charge by adjusting the nominal rate at which the charge is assessed.

10 And we do not adopt that test since we believe that the 1.7 percentage rate specified in § 12(c) of MRISA is intended to be a nominal rate, not an effective rate. The effective yield concept does not lend itself to statutory control since it will fluctuate wildly according to the customer’s timing of purchases and payments. See fn 7, supra.

11 Several class action decisions have denied recovery for amounts paid before the complaint was filed — 1970 in this case. See, e.g., Hays v Regents of the University of Michigan, 393 Mich 756; 233 NW2d 276 (1974). But that rule appears limited to the specific cases.

12 We reject any suggestion that the computation difficulties can be avoided by employing some variation of the "fluid class” concept. That idea was emphatically rejected by the Second Circuit Court of Appeals in Eisen v Carlisle & Jacquelin, 479 F2d 1005 (CA 2, 1973). And while that decision does not control an interpretation of GCR 1963, 208, we are persuaded by its reasoning.

13 That inevitable fact might be deferred, but not forever. Complex limitations on price increases or stockholder dividends might possibly delay the impact. But even those limitations would have their own costly consequences for the defendants’ customers.

14 This would not be a concern if the lawsuit conferred some future benefit on the individual plaintiffs by forcing the defendants to switch to a more favorable method of computing finance charges. But Michigan’s elected Attorney General had already begun his ultimately successful effort to force such a change before these lawsuits were started.