Skip to content

DEHOMAG: IBM's German Subsidiary

Abstract

The Deutsche Hollerith Maschinen Gesellschaft (DEHOMAG), founded in Berlin in 1910, brought Herman Hollerith’s punched-card machines to Germany. After the First World War it came under the control of the American CTR, soon renamed IBM, and in the 1930s it worked for the Nazi state, processing the censuses of 1933 and 1939 and, in the last year of the war, the SS’s registration of concentration-camp prisoners. Its D11 tabulator of 1936 was among the most capable punched-card machines of its time. How much the punched cards contributed to the persecution of the Jews, and how much IBM in New York knew, has been argued over since the 1980s; this article sets out what the evidence shows. In 1949 the company became IBM Deutschland.

From Licensee to Subsidiary

In 1910 the businessman Willy Heidinger took a licence for Hollerith’s patents and founded DEHOMAG in Berlin to import, and later build, punched-card equipment for German business and government (see Herman Hollerith and the Punched Card). It was a year older than its later parent: the Computing-Tabulating-Recording Company (CTR) was formed in New York on 16 June 1911 and renamed International Business Machines on 14 February 1924 (see IBM: The Company). After the First World War DEHOMAG owed licence fees it could not pay, and in 1923 CTR took 90 per cent of the company, leaving Heidinger with ten (other accounts date the takeover 1922 or 1924). In 1934 DEHOMAG moved into a new headquarters at Lichterfelde-Ost station in Berlin, opened with Nazi dignitaries in attendance; the rebuilt building is used today by Deutsche Telekom.

From 1933 the company presented itself in public as a German firm, and its directors as convinced National Socialists; the American parent went unmentioned. Heidinger, who remained its head, was an open supporter of the regime.

The D11

DEHOMAG’s own engineers designed the D11 tabulator, installed for the first time in July 1936; by the end of the war 1,074 had been delivered, and after 1949 it was sold as the IBM 450. Its designer, Hans Gross, who headed the electrical circuit department, died aged 34. The D11 read a stack of cards, accumulated their values in counters and printed the results, following a program wired on a plugboard. It could add and subtract across counters, net balances, and with attachments multiply an eight-digit by a six-digit number, divide, and punch summary cards. It did not execute a list of instructions, but it ran a defined sequence of steps and could switch steps on or off depending on a control hole in a group of cards, a form of conditional branching. Friedrich Kistermann, a historian of IBM Deutschland, considered it a candidate for the first computer. Surviving D11s stand in the Deutsches Museum in Munich, the Deutsches Technikmuseum in Berlin, the Technische Sammlungen Dresden and IBM’s collection in Böblingen, and one in the United States Holocaust Memorial Museum in Washington.

The Censuses

The first census under Nazi rule, on 16 June 1933, asked for religion, as the democratic census of 1925 had, and counted just under 500,000 Germans of Jewish faith, 0.8 per cent of the population. DEHOMAG processed the returns for the Statistical Office of the Reich and advertised the job with a poster of a giant eye: “Übersicht mit Hollerith-Lochkarten.” The punched cards held neither names nor addresses, only the enumeration district and the position on the enumerator’s list.

The census of 17 May 1939 covered the expanded Reich, including Austria and the Sudetenland. Besides the household forms, every resident filled in a supplementary card, handed over in a sealed envelope, giving name, address, date and place of birth, and how many grandparents had been “full Jews.” Some 750,000 enumerators collected the forms; the Statistical Office in Berlin processed them with 80 million cards on 400 electric punches, 300 verifiers, 70 sorters, 50 tabulators, many of them D11s, and 25 reproducers. A preliminary result in March 1940 counted 233,973 Jews in Germany without Austria, 0.34 per cent of the population.

The Question of Complicity

In the debates over the West German census of the 1980s (see Horst Herold and Computerized Policing), the historians Götz Aly and Karl Heinz Roth argued in Die restlose Erfassung (1984) that Nazi persecution relied on modern registration and sorting techniques, Hollerith machines among them; by 1942 at the latest the SS was a DEHOMAG customer. In 2001 the journalist Edwin Black went much further in IBM and the Holocaust: IBM in New York, he argued, knew how its machines were used, and without them the Nazis could not have deported and killed so many people.

Historians accepted much of Black’s material and rejected his conclusions. Reviewing the German edition, Beate Schreiber found that Black could not prove IBM knew of the machines’ use in the camps, that he gave no date for their introduction there because the records do not show one, and that his keyword-driven research, with volunteers photocopying documents in archives, stripped evidence of its context; DEHOMAG also had competitors, Siemens among them. Frederick E. Allen, writing in American Heritage, concluded that Thomas J. Watson and IBM “did not behave in exemplary fashion” but “did no worse than countless other multinational corporations.” Watson had accepted a Nazi decoration in Berlin in 1937; the regime installed new management at DEHOMAG in 1940 and seized it, like other American property, after Pearl Harbor. IBM disputes parts of Black’s account.

On the census itself, the detailed studies, Jutta Wietog’s of 2001 and Stefan Boberg’s dissertation of 2020, point to a narrower role than either the poster or the polemic suggests. According to Schreiber, the 1939 supplementary cards, the only census documents that tied ancestry to names and addresses, were not processed on Hollerith machines at all. After the count they stayed at the Statistical Office, were used in the second half of 1941 to check the population registers and the wartime Volkskartei, and then went to the Reich Genealogical Office; they survived the war and have since been digitised. The HNF concludes that the census data were apparently not used to plan or carry out the murder of the Jews, since the SS had other files: the identity cards Jews were forced to carry from 1938 and the lists the Jewish communities were made to hand over.

Where punched cards did enter the camps is in the prisoner registration of 1944. From the summer of that year until early 1945, the “Maschinelles Zentralinstitut für optimale Menschenerfassung und Auswertung,” a unit attached to Himmler’s personal staff in Berlin, had camp offices fill in thousands of preliminary forms, the “Hollerith-Vorkarten,” which were punched in Berlin. None of the punched cards survives and the institute’s files have never been found; what remains are about 150,000 preliminary forms, digitised in the 2000s, and a Hollerith list sent from Berlin to Buchenwald, probably produced on a D11. The use of a D11 inside a camp has not been proven.

IBM Deutschland

After the war the company moved to Swabia and in 1949 took the name IBM Deutschland, with its works in Sindelfingen; there, from 1956, it began building the IBM 650 (see Small Computers of the 1950s). In the early 1960s IBM returned to Berlin with a typewriter factory in Marienfelde and an office tower at Ernst-Reuter-Platz.

📚 Sources