This is an intriguing possibility and has parallels with the situation in Russia after Nov 1917, with the underlying question of why the most ruthless elements in the ruling council came out on top - and the propensity for this to happen in other revolutions too, eg Iran (though in that case the more authoritarian and cohesively ideological religious radicals around Ayatollah Beheshti had the advantage of the favour shown to them by Khomeini, the 'father' / inspiration of the new ruling elite, and moderate individuals like President Bani-Sadr lacked a solid factional backing) or indeed France in 1792-4. Or how Nasser pushed General Neguib aside in Egypt in 1952-4, with the Ethiopian first post-Haile Selassie govt of autumn 1974 headed by a similar well-respected senior general , ie Aman Michael Andom, who was pushed aside (and overthrown/ killed) in a surprise coup within the junta led by a faction of more radical juniors - the first appearance of Mengistu as their co-ordinator and surely a warning to the others of what he was capable of. In Portugal in 1974 the reasonably well-known and respected first head of the junta, General Spinola who had led warnings of the unsustainability of the pre-coup govt's repression in the colonies of Angola and Mozambique in the long term and been ignored, attempted to rein in his more radical juniors in aut 1974 and was pushed aside by them, with them installing a more malleable figure as the new President - General Costa Gomes. (Equivalent in Ethiopia after the fall of Andom in Nov 1974 - General Teferi Bante, who Mengistu then overthrew in a mysterious coup or failed counter-coup in Feb 1977 when most of Bante's supporters ended up shot dead in a bizarre shoot-out at the govt HQ.)
There's a perennial question of how a ruling, military-based new junta at the head of a semi-revolutionary regime with an ideological bent to 'justify' their measures, eg the Ba'ath Party military junta that took over Iraq from a less ideological and more pragmatic milit govt in summer 1968, evolved into a personal dictatorship by Saddam Hussein, first as the strongman behind his ally Pres. General Al-Bakr in the early 1970s and then in 1979 as the leader of a coup and formal Pres. In the case of Iraq, where a ruling group of largely junior or marginal military officers who had removed the previous regime of senior officers in July 1968 broke up in to factional disputes and one-man rule followed in the 1970s, there was also a clearer case of separate regional and family factions who could not get on together - with the winners, led by General al-Bakr and Saddam who were related, based on the regional town of Tikrit. The same problem hindered the new milit-led ruling junta of Syria after 1963, with regional groups and indeed rival ethnic groups within the elite constantly fighting until Assad's Alawite group from the remote NW mountains took over in 1970 and stabilised the situation in a personal regime. This plethora of rival regional or family factions was lacking in the Derg, as the Ethiopian military junior ranks from which they emerged was smaller and there was no parallel established civilian radical or administratively experienced grouping to back up one or more factions - Ethiopia being less urbanised or administratively developed than Iraq or Syria (or post-1952 Egypt). But the more 'outsider' provincials in Iraq (the Tikriti group) and Syria (the Alawites) and Egypt (the working-class and so 'down to earth and populist' Nasser) all won out in internal struggles, as working-class provincial Mengistu did in Ethiopia and the Georgian working-class Stalin did in Russia - arguably as these leaders all had more to do to push their way forward and were both more ruthless and more skilled than those who they faced. The 'street fighter' beats the 'intellectual' who as too many personal enemies - as Stalin did Trotsky.
Does collegiate rule or usually always turn into single-person rule in a 'revolutionary' council's regime, at least if the govt faces a major crisis that one man can combat easier than a quarrelling committee. Does the will to power of a focussed leader who will stick at nothing- be it Mengistu in Ethiopia or Stalin in Russia - usually win out over a disparate band of less bloodthirsty or focussed moderates? Arguably both M and S were skilful at rallying the personal enemies of their leading rivals to back them , then eliminating their rivals by visibly brutal means so that those with a sense of self-preservation gave in and supported them rather than objecting to their brutality and being lined up as the next victims. The Jacobins in France in 1793-4 were more of a 'committee' , firstly two rival factions (Danton vs Robespierre) and then one dominant faction (led by R with Saint-Just as his deputy) which had terrorised the elite by destroying their rivals in spectacular manner (execution of Danton and co compares with the Feb 1977 bloodbath in Addis Ababa) and then launched a mass terror to cow their rivals and the public and the provinces (Ethiopia in 1977 cf France in Jan- July 1794).
But in France the paranoia of Robespierre led to his alarmed next victims removing him in an internal coup in July 1794 - this did not happen in Ethiopia in 1977, unless the fall of Mengistu's deputy Atnafu Abate in late 1977 was designed to forestall this. Was Mengistu better than R at rallying his inner circle, or were they just too scared to act as were Stalin's inner circle after the Great Terror started (1934-8) or were reasonably sure that if they tried anything the entire regime would collapse (which didn't stop Robespierre's rivals in July 1794). Possibly the greater threat of break-up of the Ethiopian state in 1977-8 , due to the war with Somalia over the Ogaden, made the 'internal opposition' , incl in the Derg, more willing to put up with a seemingly uncontrolled purge (incl shooting of much of the civilian, radical Left student opposition in the capital in 1977) than the French 'moderate' Jacobins were - as by July 1794 the threat of break-up in France had abated. An Ethiopia less threatened by civil war and breakup in 1977-8 would arguably have had less of an elite need to stay united and focussed for self-preservation and so given Mengistu's rivals more opportunity to try to remove him, instead of him successfully concentrating propaganda and most of the elite's attention on survival and sending troops and civilian govt supporters alike off to the 'front' while he got on with purging the internal opposition. Cf the more artificial threat of counter-rev. attack that Stalin was able to create in the 1930s, focussed on the alleged threat of exiled Trotsky and his 'allies' - the latter naturally turned out to be anyone who Stalin wanted to get rid of. But keeping the threat of revolution leading to provincial discontent and breakaway attempts - led by conservative enemies of the rev., esp sacked senior officers, is a common theme of any rev. situation in a large and until then centralised state - cf France 1789 ff and Russia 1918 ff. In Ethiopia sacked generals and relatives of the Emperor plus landowners at risk of confiscation joined in the revolts, eg in Tigre - and leaving more of them inside the power-structures and slowing land reform in Ethiopia in 1974-5 was unrealistic and would have only led to a junior officers' revolt.
A more urbanised and industrially developed / wider educated state in Ethiopia by 1974 would probably have led to a greater and more coherent ideological left-wing civilian backing for the overthrow of the fossilised and visibly failing imperial elite (still led by pre-1789 France style courtiers and landowners) and so the possibility of these insisting on partnership with and govt post from the military Derg from aut 1974, arising from their having helped to depose Haile Selassie . They could then help the more moderate Derg figures in 1974-7, possibly focussing on the ideologically articulate Major Sisay Hapte or on ?Alemayhu Haile or Mogus Wolde-Michael, share the kudos of leading radical rural reform and bringing in civilian support to the new ruling party in a more 'bottom-up' than 'top-down' fashion. In real life the new party structure was turned into a Soviet-style 'disciplined' ie obedient and oppressive one-party admin regime answerable to a small 'politburo' group and in due course in effect to Mengistu alone.
A quick settlement with the Eritrean resistance to diminish the focus on the need for the military (and hence Mengistu as its strongman protecting national unity) to lead the state and 'do whatever is necessary to protect it from enemies' would have reduced the prominence and power of M, but that was unlikely in 1975-7 given that the majority of the Derg were determined to fight on and regarded any settlement that gave Eritrea independence as 'treason' to the unitary state - hence probably the reasons for the overthrow of 'compromiser' General Andom in Nov 1974. The war dragging on was probably inevitable given the pride and centralist views of the Ethiopian military - and any sort of civilian-led proposals to create a federal state with a popular assembly representing regional interests and self-govt this early, which could head off separatist threats from the Ogaden Somalis (or , less urgently, the Oromo) would meet a similar truculent anger from the majority of the military. (Cf the quick reassertion of centralism in Iran vs the Kurds , Baluchis and Kuzestani Arabs after Feb 1979, even with a regime led by civilians and their military under control of the new ruling elite.)
So the centrality of the military to the new regime in 1974-7 and its 'weight' in enforcing its strategic priorities is unlikely to change without a more 'developed' Ethiopia; and one-man rule and purges seem to be a common hazard of govts led by a small, embattled elite of ruthless operators. (Cf Mao ' -revolution is not a tea-party'). But a successful counter-coup to remove an already feared Mengistu by his milit rivals in Feb 1977, possibly led by the flexible and capable Sisay (with his potential for civilian student radical support) if he had not been removed already as in OTL, is a possibility. Ditto a more alert and active group of middle-ranking moderate officers, alienated from the Haile Selassie regime by lack of promotion, reacting to the bloodbath of ex-HS elite figures in Nov 1974 - where Mengistu was already seen as the leader in this - rallying to and being accepted by a more politically ruthless and agile General Teferi Bante in 1974-5 as part of the regime and being given important roles. (This would however require the suspicious junior officers agreeing to this in the interest of national unity in wartime rather than fearing them as potential counter-coup leaders who could stop much-needed rural reforms.)
The alliance could then marginalise and hopefully organised the arrest of M as a figure who was too confrontational and was adding to their danger by taking on too many enemies at once - but it would need a flexible and popular 'front-man' or a clever operator behind the scenes to outwit M. Possible, though - the drift ever 'leftwards' or towards dominance by one man or small group in some juntas has been halted where there are capable and supported moderates, eg Portugal in Nov 1975 (Antunes and co, defeating Otelo Carvalho) .
